m DISCOVERY REPORTS VOLUME XXIII -^ /N ^ /V7 DISCOVERY REPORTS Issued by the Discovery Committee Colonial Office, London on behalf of the Government of the Dependencies of the Falkland Islands VOLUME XXIII <^ W^ 0. / CAMBRIDGE AT THE UNIVERSITY PRESS 1947 Printed in Great Britain at the University Press, Cambridge {Brooke Crutchley, University Printer) and published by the Cambridge University Press (Cambridge, and Bentley House, London) Agents for U.S.A., Canada, and India: Macmillan WOODS HOLE, CONTENTS x^^ASs. THE GUT OF NEBALIACEA (published 17th November, 1943) By Helen G. Q. Rowett page 3 ON A SPECIMEN OF THE SOUTHERN BOTTLENOSED WHALE, HYPEROODON PLANIFRONS (published 23rd March, 1945) By F. C. Eraser, D.Sc page 21 REPORT ON ROCKS FROM WEST ANTARCTICA AND THE SCOTIA ARC (published 27th June, 1945) By G. W. Tyrrell, A.R.C.Sc, D.Sc, F.G.S., F.R.S.E. Foreword, by J. M. Wordie, M.A page 39 I. Petrography of the South Shetland Islands, West Antarctica 41 II. Petrography of Rocks From the Graham Land Peninsula and Adelaide Island, West Antarctica 66 HI. Petrography of Rocks from the Elephant and Clarence Group 76 IV. Petrography of Stones dredged from the .Vicinity of the Shag Rocks ... 89 V. Petrography of the South Sandwich Islands 92 THE DEVELOPMENT AND LIFE-HISTORY OF ADOLESCENT AND ADULT KRILL, EUPHAUSIA superb a (published 20th July, 1945) By Helene E. Bargmann, Ph.D. Introduction page 105 Development 106 Average Growth Rate 120 Factors Influencing Growth Rate 1 28 Conclusions 13° Bibliography 131 Appendix 132 THE ANTARCTIC CONVERGENCE AND THE DISTRIBUTION OF SURFACE TEMPERATURES IN ANTARCTIC WATERS (published 28th January, 1946) By N. A. Mackintosh, D.Sc. Part I. The Antarctic Convergence page 179 Part II. The Distribution of Surface Temperature in Antarctic Waters . . . -194 References 204 Appendix. Table 9 205 Notes on the Plates 211 Plates I-XIV following page 212 61172 vi CONTENTS NEBALIOPSIS TYPICA (published 21st October, 1946) By H. Graham Cannon, Sc.D., F.R.S page 21s REPORT ON TRAWLING SURVEYS ON THE PATAGONIAN CONTINENTAL SHELF (pubhshed 20th December, 1946) By T. John Hart, D.Sc. Foreword, by N. A. Mackintosh page 226 Introduction ^^7 General Account OF THE Fish Fauna 251 Distribution and General Notes on the Species 259 Features of General Biological Interest 3^2 Prospects of Commercial Development 3^7 References 392 Appendices 39° Plate following page 408 [Discovery Reports. Vol. XXIII, pp. 1-17, Octobei- 1943 .J THE GUT OF NEBALIACEA By HELEN G. Q. ROWETT CONTENTS Introduction 3 Methods 3 The structure of the gut of Nebalia bipes (Fabricius) .... 3 The structure of the gut of Nebaliella extrema (f. Thiele) ... 3 I. Fore-gut 3 II. Mid- and hind-gut 6 Musculature 7 The structure of the gut of Nebaliopsis typica (Sars) .... 8 Fore-gut 8 Mid- and hind-gut 8 Musculature 9 The structure of the gut of Paranebalia longipes (Wilemoes Suhm) 1 1 Mode of functioning of the gut of Nebalia bipes 12 Mode of functioning of the gut of Nebaliella extrema . . . . 13 Mode of functioning of the gut of Nebaliopsis typica .... 14 Conclusions 15 Bibliography 17 WOOD' THE GUT OF NEBALIACEA V ^J9if' MASS By Helen G. Q. Rowett, Grisedale Scholar, Manchester University INTRODUCTION MUCH attention has been paid by Cannon, Manton, Lowndes and others to the ' feeding mechanisms ' of Crustacea, but no attempt has so far been made to correlate changes in the structure of the gut with the type of food available and the condition in which it is passed into the mouth. For this purpose it is necessary to compare members of one group which have different habits and habitats rather than isolated examples from different groups. A survey of the Nebaliacea has therefore been made with the object of discovering how far the structure of the gut shows group resemblances and how far it may be associated with the environment and habits of the species concerned. METHODS With the exception of Nebalia bipes, material for this investigation was limited to Discovery specimens of Nebaliopsis typica and Nebaliella extrema kindly made available by Professor Cannon. Reconstructions were made using transverse sections of Nebaliopsis typica specimen E {Discovery Reports, 1931, vol. in) and sagittal sections of half of specimen F2 (op. cit.) and the unsectioned other half of this specimen. From these reconstructions Figs. 4, 5, 6 C and 7 A were made. The single specimen of Nebaliella extrema was sectioned transversely, and the reconstructions shown in Figs. 2, 3, 6 B and 7 C were obtained. Besides sectioned material the cast skins of Nebalia bipes were examined and living specimens were watched in a jar with sea water and some of the mud from their usual habitat in Rum Bay, Plymouth, and also isolated in dishes under a microscope. Carmine was fed to some and, using strong illumina- tion, the passage of the red particles through the gut was easily seen through the semitransparent body. THE STRUCTURE OF THE GUT OF NEBALIA BIPES (Fabricius) The structure of the gut of Nebalia was described in great detail by Claus (1889) and later by Jordan (1909, 19 1 2) in papers comparing the pyloric section with that of Idothea, Ganimanis and Astacus. A detailed description need not, therefore, be given here, but for the sake of clarity in making com- parisons with other types Figs, i, 6 A and 7 B have been made which show the various parts and associated musculature. In one important respect, however (which is not mentioned by Jordan), Claus's description is definitely incorrect. The structures which he describes as chitinous pads with thickened striations are actually rows of very strong evenly set setae (^.^.1 in Fig. i) which with the spines (g.s.2) form a tube in which grinding takes place. THE STRUCTURE OF THE GUT OF NEBALIELLA EXTREMA {i. Thide) I. FORE-GUT The structure of the oesophagus and cardiac region of the stomach of A^. extrema is, as Thiele (1905) says, very like that of Nebalia. Certain important differences may, however, be noted. There are many fewer setae throughout. The anterior median projection {a.m.p.) is much reduced and the lateral fanlike plate of setae which is found on the right side only in Nebalia {l.p. Fig. i) is absent. The spatial relations of the homologous parts are so altered that there is no grinding tube such as is seen in Nebalia. As shown in Figs. 2 and 3 A, the spines {g.s.o) are ventral instead of dorsal ^ s a QiO to 3^fc: Ta'tioiiii o O e C o OJ CA XI SO (1) a ■- C rt ■ ':! "^ a-. , -c; ti3 >- ^ J2 O 1- C .5 P m t* It; -T3 2 ^*H "Ts -^ rt ° .. •' a:; j3 0* bc e ■ c o tx >, o o 1) rt o o tie :r o ^ T3 . ■a 5 ■ - o-S. o S c CO O > THE GUT OF NEBALIACEA S to the regular row of setae (gs.i). Of equivalent functional significance is the grinding organ formed by the setae alone (g.s.i), which slope diagonally backwards and outwards almost parallel to, and rubbing against, a slightly setose horizontal shelf of each lateral wall. The more anterior of these grinding setae are longer than the posterior ones, their tips curve upwards and a pair of small ridges (d.r.') lie medial to the basal thickening on which they are inserted. a.m oes mo. Fig. 2. Diagram of the right half of the fore-gut of Nebaliella extrema. A, region at which Fig. 3 A was cut; B, region at which Fig. 3 B was cut. a.m.p. anterior median projection; d.ca. dorsal caecum; d.gl. digestive glands; d.gl.o. opening of the digestive glands; d.p. dorsal process; d.r. dorsal ridge; d.r.' small ridge on the dorsal ridge; ^.^.j strong setae; ^.y., short stiff spines; i. intestine; l.p. lateral pad; m. mandible; mo. mouth; yn.p. median pad; oes. oesophagus;/), long projections; v.p.ch. ventral pyloric chamber with opening of the digestive glands ; outUne of the lumen of the gut laterally. Ventral to the row of spines {g.s.2) is a slight ridge on each lateral wall behind which is a strong contractor muscle. This ridge marks the division between the oesophagus and the stomach. There is no distinct division into cardiac and pyloric regions, but, posteriorly, where the spines ig.s.i) and setae {g.s.i) cease, the lateral walls approach one another more closely and their surfaces 6 DISCOVERY REPORTS are soft and irregularly corrugated. In this region the dorsal glandular caeca [d.ca.) open into the dorso-lateral angles of the lumen. Slightly posterior to these openings the dorsal ridge [d.r.) becomes free from the dorsal wall and projects a short distance as a soft pad {d.p.). Similarly swellings of the lateral and ventral walls split off together from the gut walls and form a trilobed chitin-covered process {l.p. and rn.p.), each of whose lobes bears a long projection {p.) converging dorsally as shown in Fig. 3 B. The lateral lobes bear strong setae and are undoubtedly homologous with the lateral finger-like processes of Nebalia, but the homologies of the median ventral lobe are uncertain. Thiele suggests that it may be the sum of the two lateral ridges {l.r.) of Nebalia. If this were the case it might be expected that some trace of the double nature would remain, but none could be found. It is more likely that it is homologous with the ventral cardio-pyloric valve, the shifting of which posteriorly is a slight change comparable with the other differences between the two species. A third alternative is that the small pyloric pad of Nebalia (p.p. Fig. i) has been greatly enlarged, but this, like Thiele's suggestion, is a major alteration involving the disappearance of the cardio-pyloric valve. dr. d.r' oesn m. Fig. 3. A. Anterior region of the fore-gut of Nebaliella looking forwards into a piece cut at region A of Fig. 2 to show the relationships of the parts hidden by the median structures, a.m.p. anterior median projection; d.ch. dorsal channel ; d.r. dorsal ridge; d.r.' small ridge on dorsal ridge ;^.s.i strong setae; 0.5.2 short stiff spines; m. mandible ; ow. oesophagus; o«.r. oesophageal ridges. B. Posterior region of the fore-gut and the entrance to the intestine of Nebaliella looking backwards from region B of Fig. 2 to show the spatial relations of the pads and the projections thereon in the pyloric part of the gut. d.gl. digestive glands; /. intestine; l.p. lateral pad; m.p. median pad; p. long projections; v.p.ch. ventral pyloric chamber. II. MID- AND HIND-GUT Besides the dorsal glandular caeca already mentioned, Nebaliella resembles Nebalia in having three digestive gland caeca {d.gl.) on either side. These unite and open into the ventro-lateral corners of the ventral pyloric chamber {v.p.ch.) immediately posterior to the tripartite process. The openings are smaller than those in Nebalia. Ventral glandular caeca were not found. The lumen of the intestine is relatively wider than in Nebalia, and for a considerable distance is roughly triangular, with one angle dorsal and two ventro-lateral, and with walls of highly vacuolated cells. Passing posteriorly the cells become less vacuolated and a striated border appears on them. Then the outline of the gut becomes oval and the striated border deeper. Finally the cross-section THE GUT OF NEBALIACEA y is almost circular, the cells are very dense and closely packed together, and some are elongated and project as ridges into the lumen. As in Nebalia the intestine and digestive glands are embedded together in a loose tissue of highly vacuolated cells. There is a very much reduced rectal gland and an anal chamber comparable to these structures in Nebalia. ^■h.-^ a.m. I n n Fig. 4. Diagrammatic reconstruction of the right half of the fore-gut of Nebaliopsis typica in three sections (I-III). a.h. anterior horn; a.m.p. anterior median projection; d.r. dorsal ridge; gl.r. glandular region; /. labrum; l.th. lateral thickenings; pa. paragnaths ; pl.w. plated walls. MUSCULATURE The similarities between the musculature of the fore-gut of Nebaliella and that of Nebalia are very striking, as the diagrams (Figs. 6 A, B) show. Differences are that the lateral dilator muscles {l.dil.) of the oesophagus have five points of insertion as compared with two in Nebalia. The muscle corre- sponding to the small median projection muscle is greatly enlarged. The anterior dorsal dilators (a.d.dil.) are also enlarged but the posterior dorsal dilators (p.d.dil.) are reduced, though there is a great thickening of the chitin at their point of insertion. The strong circular muscle (cont.) which is so conspicuous in Nebalia is present in Nebaliella also, though slightly reduced. No muscles could be found in the groove between the anterior horns of the stomach where Nebalia has a few thin strands of fibres. 8 DISCOVERY REPORTS In addition to these muscles which have their homologues in Nebolia, Nebaliella has a pair of very strong muscles (t.p.m.) which stretch ventro-laterally from a thickening of the chitin of each lateral lobe of the trilobed process immediately anterior to the point where the process splits from the gut wall. THE STRUCTURE OF THE GUT OF NEBALIOPSIS TYPICA (Sars) FORE-GUT The external features of Neboliopsis (Cannon, 193 1, pi. xxxii) indicate that it is a highly specialized member of the Nebaliacea, and this is confirmed by the internal organization. Even in the gastric mill group resemblances are few. Fig. 4 shows the reconstruction of the right half of the fore-gut in three sections. The molar processes of the mandibles are reduced and do not project into the mouth as in the other forms. The mouth is a transverse slit between the labrum (/.) and the paragnaths {pa.). These can be retracted by strong muscles, thus uncovering a flat plate of chitin with a median antero- posteriorly directed slit leading into the stomach. The latter slit can be opened widely by the dilator muscles [l.dil.), which slope upwards and outwards from the oesophageal wall. Very great variation of both the size and the shape of the gape is thus possible by the combined action of these two slits. There is no distinct separation into oesophagus, cardiac and pyloric parts, but the region surrounded by the horizontal circular muscles, and to which the lateral dilators are attached, may be considered oesophageal in comparison with Nebolia. Setae are entirely absent from the gut. Except in the most anterior and dorsal regions the chitinous lining of the lateral walls has the appearance of crazy-paving owing to the presence of grooves over the junctions between the individual cells of the supporting tissue (Fig. 5 C, D). Posteriorly these grooves are less distinct. The paved parts of the walls lie very close together (Fig. 5 C) and provide a good gripping surface. An anterior median projection (a.m.p.) is present as in Nebalia, but much reduced. There is a dorsal ridge [d.r.) which is strongly chitinized and slightly grooved in the region immediately dorsal to the mouth, but it arises anteriorly as a soft pad and becomes so again posteriorly. The lateral walls have thickenings {l.th.) against which this ridge bites. The thickenings of the chitin are prolonged into short anterior projections of the stomach, and may be homologous with the slight thickenings at the bases of the spines (^-^.2) in Nebalia. There are no grinding tubes, but a strong grinding or biting action probably occurs between these heavily chitinized regions. Posteriorly where the dorsal ridge becomes a soft pad, the side walls open out slightly and they also become soft. The chitinous lining of the fore-gut ends raggedly ; the walls become glandular and lose the thick muscle sheath which encircles them throughout the stomach region. MID- AND HIND-GUT The glandular region mentioned above marks the beginning of the mid-gut. Here the anterior digestive diverticula (d.gl.), which are comparatively small (Fig. 6 C), and extend only a short distance forwards and which are probably homologous with the dorsal caeca of Nebalia, open by irregular apertures. Some of these are small channels passing through the glandular region, but the largest opens below a flap (gl.r. in Fig. 4) directly into an immense digestive sac {d.s. in Fig. 7 A), which widens out suddenly and almost completely fills the body cavity back to the end of the fourth abdominal segment, and which may be homologous with the digestive caeca of Nebalia, though it is difficult to be certain of any homologies when the specialization is so great. It is important to note that this is a plain sac without any convolutions and with only a few thin septa rising from its walls. The enlarge- ment therefore does not provide a great deal of extra surface area for absorption, but it does give a THE GUT OF NEBALIACEA 9 large volume for storage. The walls of the sac are formed of a thin layer of pavement epithelium made up of huge highly vacuolated cells with a striated border and an average diameter of o-i mm. (Fig. 5 A, B). A very thin basement membrane lies behind them. The intestine is a very narrow tube lying dorsal to this sac (Fig. 7 A). In places the lumen is so small that it is hardly distinguishable, but posterior to the end of the digestive sac it opens out into a wider rectal region. A muscular sphincter separates it from a short proctodeum. No rectal gland was found. D Fig. 5. A. Cells of the digestive sac in surface view. B. Same in section showing striated border. C. Section of the plated side walls of the stomach showing the grooves in the chitin as it is laid down over each cell and the closeness of the opposite walls of the gut : ch. chitin ; gr. intercellular groove in the chitin ; /. lumen of the gut ; n. nucleus. D. Surface view of the chitin. MUSCULATURE The musculature of the fore-gut of Nebaliopsis is shown in Fig. 6 C. The similarities to the other Nebaliacea are striking. The oesophagus and stomach are sheathed in strong bands of circular muscles [h.circ. and v.circ). These bands are many times thicker than the corresponding ones in Nebalia, while the tissue between them and the chitin is comparatively much reduced. They cease abruptly at the end of the fore-gut. Acting antagonistically to these circular muscles are the dilator muscles. The dorsal dilators (d.dil.) are probably homologous with the anterior dorsal dilators [a.d.dil.) of Nebalia, as the groove muscles which lie close to the dorsal ridge between the anterior horns of the stomach run between them and DISCOVERY REPORTS are inserted on the dorsal wall more posteriorly. In Nebaliopsis the groove muscles consist of a very thick bundle of fibres passing from the dorsal ridge as mentioned above to the anterior wall of the oesophagus ventral to the anterior median projection, while only four pairs of slender strands were found in Nebalia. The anterior lateral dilators {l.dil.^ and l.dil.^) differ only in that they slope dorsally p.ddii a.d.dil. a.h.m. y. arc. arc. 3. oes. pj.dil. Fig. 6. Diagrams to show the musculature of the fore-gut of: A, Nebalia; B, Nebaliella; C, Nebaliopsis. a.d.dil. anterior dorsal dilators; a.h.m. muscles of the anterior horns of the stomach; a.m. median anterior muscle; a.oes. anterior oesophageal muscles; a.s.m. small anterior muscles; conl. strong contractor muscle; c.p.m. cardio-pyloric muscles; d.ca. dorsal caeca; d.dil. dorsal dilator; d.gl. digestive glands; d.s. digestive sac; h.circ. horizontal circular muscles; /. labrum; l.dil.^, l.dil.^, l.dil.^, l.dil.^, and l.dil.^, points of insertion of the lateral dilators; m. mouth; m.p.m. median projection muscle ;/)a. paragnaths; p.d.dil. posterior dorsal dilators; p.l.dil. posterior lateral dilators; p.m. posterior muscles; p.oes. posterior oesophageal muscles; t.p.ni. muscles of the trilobed process; v.ca. ventral caeca; v.circ. vertical circular muscles ; v.l.dil. ventro-lateral dilators; v.m. small ventral muscles. THE GUT OF NEBALIACEA n instead of ventrally. A small ventro-lateral dilator muscle {v.l.dil), three small muscles (a.s.m.) extending anteriorly and a muscle (p.m.) pulling posteriorly from a triple insertion on the gut were found. The anterior muscles (a.m. and a.h.m.) are small but correspond to similar muscles in Nebalia. The digestive sac has dorsally and ventrally a pair of longitudinal muscle bands. Small segmental muscles which support the thoracic limbs lie in the wall of the sac and cause slight ridges in it. No musculature was found on the intestine and only a small sphincter at the anus {a.sp.). Fig. 7. Diagrams of the right halves of A, Nehaliopsis, B, Nebalia, and C, Nebaliella, showing the position of the gut in the body cavity of each. (The positions only of the limbs are indicated in B and C.) a. anus; a.ch. anal chamber; a.sp. anal sphincter; d.gl. digestive glands; d.p. dorsal process projecting down the intestine; d.s. digestive sac;/.^. fore-gut; i. intestine; m. mouth ; md. mandible ; r. rectum ; r.ca. rectal caecum ; sep. septum. THE STRUCTURE OF THE GUT OF PARANEBALIA LONGIPES (Wilemoes Suhm) No specimens of Paranebalia were available for examination. Nevertheless, to complete the survey of the Nebaliacea Thiele's description may be quoted. He found that the gut is, on the whole, like that of Nebalia, but records these differences: (i) strong spines are present under the long setae on the ventral side of the ' hypopharynx ' (the ventral lip) ; (2) the rows of setae on the dorsal ridge in the pyloric region do not extend far back, and the funnel formed by the dorsal process bears no setae ; (3) two ventral bristle plates take the place of the small lateral ridges {l.r.) of Nebalia. These plates are composed of a transverse row of thick setae which meet each other across the lumen of the gut. The midmost setae are longest. Thiele's diagram does not show an anterior median projection, lateral plate setae (but he has depicted the left side), setae on the walls of the oesophagus or cardiac region, or a ventral cardio-pyloric valve. 12 DISCOVERY REPORTS but he does not mention these points as differences. Therefore either his description is inadequate or his diagram incorrect. In the absence of further material the answer to this question cannot be given. MODE OF FUNCTIONING OF THE GUT OF NEBALIA BIPES The mode of functioning of the gastric mill of Nebalia bipes may be deduced from evidence furnished by the structure of its parts, the distribution of particles within the gut, and also from observation of living animals. Specimens kept in shallow water in a jar, on the bottom of which was mud from their natural habitat, were observed undisturbed. They occasionally swam about, but usually lay on the surface of the mud (often in the shadow of large pieces of seaweed or stones), where it could be seen that the thoracic limbs seldom ceased their regular rhythmic motion even when the animal as a whole was stationar}^ They appeared to burrow only when disturbed. When isolated in small dishes and placed under the microscope they swam rapidly but at times lay quiescent and could then be studied. The thoracic limbs continued to beat unless the specimens were kept long in these conditions when they frequently became completely inactive for considerable periods though often reviving later. When all the movement of the limbs ceased in this manner the rate of heart beat slowed down and large particles were seen floating in the blood stream. This effect has not been studied in detail but it is probably caused by the unnatural conditions in the dishes, as no such long pauses in the motion of the thoracic limbs were noted when watching the animals in the jar. The currents produced by the movements of the thoracic limbs bring particles to the filter apparatus (Cannon, 1927). There is thus a continuous supply of food depending only on the concentration of suspended matter in the water. If excessive amounts are collected the particles are gathered into balls and shot out ventrally in the anterior region of the carapace. This mechanism probably helps to prevent the filter apparatus from being choked with mud when the animal is burrowing and also indicates that the movement of the thoracic limbs serves another purpose besides feeding. It is possible that the continuous current of fresh water is necessary for respiration and must be maintained whether it bears many or few particles. Thus normally there is a constant stream of filtered material being passed to the mouth. Large particles have a preliminary grinding by the maxillary endites (see Cannon, 1927, for details), and are also ground between the mandibles. Rows of setae on the lips prevent pieces from falling off into the grooves on either side of the mandibles and direct them into the oesophagus. Strong contraction of the circular muscles keeps the passage from the oesophagus to the stomach closed most of the time, but periodically these muscles are relaxed, and simultaneously the lateral dilators work actively causing a 'puff' of particles to pass into the stomach and swiftly back into the pyloric region and the intestine. Setae on the walls of the oesophagus point dorsally and prevent backflow. All the gut muscles move violently during this operation. The anterior median projection and the lateral plate setae {a.tn.p. and l.p.) help to direct the current round the angle between the oesophagus and the stomach so that large amounts of material do not pass dorsally and choke the grinding tubes. The setae are, however, not close enough to form a strict filter, and some particles pass up into the grinding tubes and are ground between the setae (^.^.1) and the vertical ridges {v.r.) and spines {g.s.o), which are rubbed across one another by a complex circular and see-saw motion of the dorsal ridge, easily seen in living specimens and probably caused by alternating contraction of the dorsal dilators combined with peristalsis of the circular muscles. Only liquid was found in the dorsal channels (d.ch.), which are open posteriorly, and it is possible that a secretion from the dorsal caeca may flow forwards in them and be poured upon the food as it is being ground up in a manner analogous to Yonge's suggestion for Nephrops (Yonge, 1924). THE GUT OF NEBALIACEA 13 The particles from the grinding tubes are passed back and on to the long setae of the pyloric region. The narrowness of the lumen of the gut in the posterior part of the cardiac section and the presence of the ventral cardio-pyloric valve {v.v.), whose tip moves violently describing an ellipse, causes particles which have been driven directly back without secondary grinding in the grinding tubes to pass up on to these setae also. The latter filter off the larger pieces and bear them back beyond the openings of the digestive glands and far down the intestine in the tubular extension of the dorsal process. The smaller pieces fall through into the ventral pyloric chamber and pass into the digestive glands {d.gl). Muscle bands on the walls of these glands probably cause pumping in and out of fluid bearing small particles as in Nephrops (Yonge, 1924). Certainly in ink- fed specimens grains were found far into these caeca indicating that particles from the stomach are passed into them. Particles appearing like finely ground food were frequently found in them also. No ink grains or other particles could be seen in either the dorsal or ventral caeca {d.ca. and v.ca.). This suggests that absorption as well as digestion probably takes place in the digestive glands, while the dorsal and ventral caeca secrete a digestive fluid only. The arrangements of the gastric mill of Nebalia are such that there is continuous action of the secondary grinding apparatus which increases the number of particles small enough to pass into the digestive glands, while at the same time the animal is able to deal casually with the large quantities of potential food which are automatically available and whose amount depends only on the concen- tration of particles in suspension in the water filtered and the proportion of inorganic to organic matter. MODE OF FUNCTIONING OF THE GUT OF NEBALIELLA EXT REM A Nebaliella is a mud-living form. The eyes and rostrum have been shown by Cannon (1931) to be a mechanism whereby mud is prevented from entering the space within the carapace and choking the filter apparatus as the animal burrows. Variations in the completeness of closure of this apparatus control the current entering the filter chamber. Particles found amongst the mouthparts and also within the gut include large pieces of diatom skeletons, radiolaria, and many unidentifiable broken pieces showing that the animal is an indiscriminate mud feeder and also that it can deal with relatively coarse filtered food. That these particles are present far down the intestine indicates that there is no very efficient grinding of the food. It is probable that as in Nebalia much material is passed through rapidly and a little is more carefully treated. That particles passed on to the mandibles receive only slight grinding before entering the oesophagus is shown by the state of the food within the gut. The sheath of circular muscles probably functions in the same way as in Nebalia and releases particles spasmodically. The angle between the oesophagus and the stomach is more obtuse than in Nebalia, the anterior median projection is much reduced and the lateral plate setae are absent, but clogging of the grinding setae {g-s.-^) is prevented by an entirely different mechanism. In the anterior region the edges of the horizontal shelves of the side walls almost touch the small ridges {d.r.') on the dorsal ridge so that the channels containing the setae (^.^-i) are nearly closed and only relatively small particles can enter them. These particles are ground between the setae and the shelf and when fine enough pass between the former and are found as a ' felty ' layer on the dorsal side of them. This arrangement and the general reduction in the numbers of setae are almost certainly correlated with the coarse texture of the food against the passage of which fine setae would have no effect. Such setae would soon be broken or worn away. The spines (^.^.2) may have some guiding effect on the current, but as they are so short they are probably only a relic of their homologues in Nebalia. There is no filter mechanism in the pyloric region. The openings of the digestive glands are small, and only most minute particles were found within them. In the absence of more material the mechanism 14 DISCOVERY REPORTS which prevents the openings of the digestive glands from being occluded by large particles is uncertain, but the following is a possible interpretation of the structures found. The long projections on the lateral and ventral pads form a triple barrier across the lumen. This barrier is augmented by the long setae on the lateral pads. As a large mass of food is passed back it comes up against the barrier and depresses the projections so that a bridge is formed which guides the particles across the pyloric chamber into the intestine. This movement causes the lobes from which these projections rise to be bent backwards and downwards to fill a large part of the ventral chamber, occlude the openings of the digestive glands and at the same time press digestive secretion from the chamber out on to the food as it enters the intestine. Elasticity for this movement is provided by the large blood sinuses within the lateral pads and below the ventral one just anterior to the point where they become free. When the food has passed, the projections spring back to the vertical position assisted by the powerful muscles in the lateral lobes of the process. The pyloric chamber is thus opened once more and ready to be refilled with secretion from the digestive glands. The structure of the digestive glands is such that they are probably almost entirely secretory, while a little digestion and absorption of the small amount of finely divided material which enters them may also take place. The dorsal caeca are entirely secretory as in Nebalio. The structure of the intestinal wall suggests that besides absorption there is additional secretion of digestive enzymes especially in the anterior region. MODE OF FUNCTIONING OF THE GUT OF NEBALIOPSIS TYPICA The mode of functioning of the various parts of the gut of Nebaliopsis cannot be described with certainty as yet, for in specimen E hardly any particles were present and in specimen F2 the digestive sac was full of an almost homogeneous mass resembling coagulated yolk, but two alternative mechanisms are here suggested, the second being the more probable. I. Fine particles filtered out of the water by the maxilla and first trunk limb (Cannon, 193 1) may be sucked into the stomach by the action of the lateral dilators and the circular muscles. There can be no preliminary grinding owing to the structure of the mouthparts, but once within the gut any large pieces may be ground between the dorsal ridge {d.r. Fig. 4) and the lateral thickenings {l.th.) and also between the side walls which approach each other very closely and are heavily chitinized and grooved (Fig. 5 C, D). There are no setae to hinder direct passage of food into the digestive sac, therefore it cannot remain long in the fore-gut. Digestive secretion is poured on to it as it passes the openings of the anterior digestive diverticula {d.gl. Fig. 7 A). These openings are large and unprotected and particles could easily enter them, but the structure of the glands does not suggest that any absorption takes place within them. There is no possibility of any food passing straight from the fore-gut to the intestine as it does in Nebalia and Nebaliella. Everything must enter the sac where both digestion and absorption probably take place. It is difficult to visualize how the sac does not become clogged with indigestible matter, as there is no apparent means of circulating the material in it. A possible explanation is that a deep pelagic filter feeder will obtain very little particulate inorganic matter such as is so abundant in and near the surface of mud so that digestion will be almost complete. Filterable particles are scarce in this zone, and the blind diverticulum permits the retention of all material until it is thoroughly digested thus preventing waste. This suggested mechanism agrees with Cannon's belief that Nebaliopsis is 'entirely a filter feeder'. His conclusions were reached from a study of the mouthparts alone, particularly important being the facts that 'the whole mouth armature is extremely soft and unsuited for dealing with large food THE GUT OF NEBALIACEA 15 particles', and that 'in addition there is a compUcated structure which, in my opinion, must be a fiher'. The internal organization, however, and other considerations, make probable the following alternative mechanism, in which the first of these facts plays an important part. II. Nebaliopsis is probably an egg sucker. As far as information is at present available eggs of various types have been found in small numbers in all the hauls in which Nebaliopsis has been collected. The mechanism by which it sucks the yolk from these eggs is probably as follows. As Cannon reports, the mandibular palps are long and armed at the tips with stout claw-like setae, which grip the slippery surface of the egg. The eddy currents round the mouth caused by the move- ments of the trunk limbs and mouthparts also help to keep the egg pressed against the mouth. The molar processes of the mandibles are soft and useless for biting ; thus the egg is not punctured outside the mouth, where there would be great risk of the contents being washed away by the water currents in that region. While being held close to the mouth one side of the soft egg is sucked into the oesophagus by the pumping action of the muscles on its walls. It is then gripped tightly by the plated surfaces of the lateral walls, while the biting action of the dorsal ridge against the dorso-lateral thickenings of the chitin makes a hole in the egg membranes. The liquid yolk is then pumped through this puncture into the digestive sac, digestive secretion being poured on to it as it passes the openings of the anterior glandular caeca. The great thickness of the muscle sheath of the fore-gut may be associated with this strong pumping action. The steadying action of the mandibular palps and the eddy current round the mouth are most important during this process. The empty egg case would then be thrown away. The capacity of the digestive sac is sufficient to accommodate the contents of several average-sized fish eggs. As suitable eggs are likely to be found in groups near where they have been spawned, and only at certain times of year, a meal is available only at infrequent intervals. Much food is thus taken at one time and is stored in the immense digestive sac where it is assimilated slowly as required. There is hardly any solid waste in this method of feeding, which agrees well with the observed structure of the extremely narrow intestine and the lack of through current or strong muscles by which solid waste could be evacuated from the blind digestive sac. It also agrees with the fact that no ' structure ' was found in the solidified mass in the sac, as would be expected if particulate matter were collected indiscriminately by a filter mechanism or indeed if Nebaliopsis fed on anything but liquid or semi-liquid food. There is nothing in the structure of the mouthparts or gut to suggest that it is a blood sucker, and the appearance of the food undoubtedly suggests coagulated yolk. In the depths from which Nebaliopsis has been collected there can be very little finely divided material for a filter feeder — only the slow rain of dead plankton from the surface layers. An animal of the size of Nebaliopsis will require a considerable amount of food. The second theory would supply this better than the first. As has been shown, the structure of the gut and especially the presence of the large digestive sac also indicate that an occasional large meal is taken. It is possible that the animal depends chiefly on sucking eggs, but has a filter mechanism which provides a small additional supply of food, alone insufficient, but valuable when prey is scarce. Without intermediate forms it is impossible to tell how this complex and highly specialized mechanism originated. It is undoubtedly, however, well adapted to the environment in which the species now lives. CONCLUSIONS The structure of the gut diff'ers considerably in the different members of the Nebaliacea, and many of the changes may be correlated with the feeding habits. The greatest similarities are found in the musculature. It is obvious that for the efficient working of a complicated chitinized apparatus simple peristalsis of circular muscles is insufficient. Opposing dilators are necessary. The oesophagus of Crustacea almost invariably has lateral dilator muscles i6 DISCOVERY REPORTS and others are associated with the teeth and other grinding parts. In the Nebaliacea the dorsal ridge always forms part of the grinding organ of the gastric mill and at least one dorsal dilator is present in all species. Other muscles are developed in association with special parts or functions ; for example, the muscles {t.p.ni. Fig. 6 B) which move the trilobed process of NebalieUa back to the vertical position after the food has passed, and those (v.m. and c.p.m. Fig. 6 A) which cause the elliptical motion of the tip of the cardio-pyloric valve in Nebalia. The numerous small muscles which are not attached to particular structures in the gut probably function in steadying the whole organ in relation to the other parts of the body. Thus, though the plan of the musculature is simple and constant, the changes may be associated with the structure of the chitinous parts, and these in turn may be correlated with the habits of the species concerned. Nebalia and NebalieUa both live where the bottom deposits are muddy, but observations of the former, when living, show that it lies most of the time above the mud just beneath or amongst larger debris of pieces of seaweed, shells and stones. The particles on which it feeds are thus the small ones in suspension in this zone. NebalieUa, on the other hand, appears to be a true mud dweller. The specializations of the eyes, rostrum, and antennae are adaptations to burrowing, and the food particles found amongst the limbs and in the gut indicate that it feeds indiscriminately on the mud. Many of the particles are too large to stay long in suspension. Therefore it must either allow some to pass into the carapace chamber as it burrows or kicks up the mud and then filter rapidly before it settles (as on occasions does Chirocephalus). In this mud there is a much higher percentage by volume of silica and other inorganic matter than in the suspension of finer particles taken by Nebalia. This means that the material which NebalieUa swallows has a lower food value, and there must be more of it. The mechanism, which is already present in Nebalia, for rapid dealing with much food is elaborated and that for efficient grinding of a little is reduced. The food is largely retained in the through passage of the gut and not passed into the diverticula. In this way the indigestible particles are passed on rapidly, and such nutriment as can be easily extracted by the digestive enzymes is obtained. In morphological association with this, the openings and lumina of the digestive glands, are reduced and the lumen of the intestine increased, and in place of the filter allowing the passage of selected finer particles into the glands there is a mechanism whereby they are almost all excluded and passed straight on down the intestine. Thus the differences between the structure of the gut of Nebalia and that of NebalieUa may be definitely associated with the habits of these animals and the food thus made available. Nebaliopsis, which has so far been found only at great depths in the open ocean, is in very different surroundings from the bottom-living forms. It is only to be expected that adaptations to these conditions would cause specializations, such as are found both in the external and in the internal structures. The form of the gut may be correlated with the difference in food.^ Filterable particles are much scarcer in this zone, and, as has already been shown, the mechanism whereby much useless material is passed rapidly through the gut has disappeared. A special method for dealing with an entirely different type of food has been developed. This food is almost certainly eggs, and in adaptation to the periodic abundance and scarcity of these the large digestive sac has been developed as a store chamber and the lumen of the intestine has been reduced to insure that 1 Since the above was written my attention has been drawn to a description by T. J. Evans (Q.J. M.S. 1922, lxvi N.S. p. 439) of Calma glancoides, an Aeolidiomorph Nudibranch which feeds exclusively on ' the eggs and embryos of the smaller shore fishes'. The amazing similarity between the adaptations of this mollusc to an egg diet and the specialized structure of Nebaliopsis forms additional evidence that the latter also feeds on eggs. This is a remarkable case of parallel adaptive evolution in two animals widely separated in phylogeny, habits and habitats and it is hoped to elaborate the comparison elsewhere. THE GUT OF NEBALIACEA 17 nothing escapes thorough digestion. The fore-gut is adapted to the puncturing and sucking of the eggs and the mandibles to holding them in position during these processes. Interesting parallels to the development of a large storage chamber when an occasional meal is taken are to be seen in the Decapoda and in the Anaspidacea. In the former group there is a swelling of the anterior region of the cardiac portion of the stomach in all the predatory forms examined, while in Porcellana, which has been shown by Nicol (1932) to be a filter feeder, there is no such swelling. Similarly in the Anaspidacea, Koonunga cursoria, which has been shown by Cannon and Manton (1929) to have 'given up fiher feeding completely', has a long tubular storage section of the fore-gut which is absent in Anaspides and Paranaspides, which are filter-feeding forms. In the above three examples the same result has been attained by entirely different means. The gut is in more direct contact with the environment than any other internal organ and is thus more subject to the same influences as act upon the external features. The type of food available not only influences the method of capture and the mouth parts, but also the structures which have to deal with it later on. An attempt has here been made to show how the digestive mechanism of the Nebaliacea may be correlated with the habits and habitats of these animals as far as can be deduced from present knowledge of this rare group in which so many evolutionary links are missing. The greater part of this work was carried out in the Zoology Department of the University of Manchester while holding the Grisedale Research Scholarship. I wish to thank Professor Graham Cannon and Dr S. M. Manton for the loan of fixed material, Mr G. A. Steven for the living specimens of Nebalia, specially collected along with characteristic elements of their habitat, and all three for much helpful advice and criticism. BIBLIOGRAPHY Cannon, H. G., 1927. On the feeding mechanism 0/ Nebalia bipes. Trans. R. Soc. Edinburgh, lv, pp. 355-70. Cannon, H. G., 1931. Nebaliacea. Discovery Reports, in, pp. 199-222. Cannon, H. G. and Manton, S. M., 1929. On the feeding mechanism of the Syncarid Crustacea. Trans. R. Soc. Edinburgh, LVi, pp. 175-89. Claus, C, 1889. Organismus der Nebaliiden und Systematische Stellung der Leptostraken. Arb. Zool. Inst. Univ. Wien, viii, pp. 1-149, pis. 1-15. Jordan, H., 1909. Die Pylogenese der Filtervorrichtungen in Pylorttsmagen der Malacosiraca. Verb. d. Zool. Ges., Leipzig, 19. PP- 255-66. Jordan, H., 1912. Der Magen der hoheren Krebse. Naturw. Wschr. xi. Nicol, E. A. T., 1932. The feeding habits of the Galatheidea. J. Mar. Biol. Ass. U.K. 1932, pp. 87-106. Ohlin, 1901. Arctic Crustacea collected during the Swedish Arctic E.xpeditions 1898 and 1899. Bihang Svenska Acad, xxvi, 4, 12. Thiele, J., 1904. Die Leptostraken. Wiss. Ergebn. d. Tiefsee Expedition 'Valdivia', vii, pp. 1-26, pis. 1-4. Thiele, J., 1905. Ueber die Leptostraken der Deutschen Siidpolar Expedition, 1901-1903. D. Siidpolar Exp. ix (Zool. i), pp. 61-8, pi. 2. Yonge, C. M., 1924. Mechanism of feeding, digestion and assimilation in Nephrops norvegicus. J. Exp. Biol, i, pp. 343-89. [Discovery Reports. Vol. XXIII, pp. 19-36, March 1945] ON A SPECIMEN OF THE SOUTHERN BOTTLE- NOSED WHALE, HYPEROODON PLANIFRONS By F. C. ERASER, D.Sc. CONTENTS Introduction page 21 Lateral view of skull 21 Dorsal view of skull ......•• 24 Ventral view of skull ......•• 25 Mandible 26 Teeth 26 Vestigial teeth 27 Vertebrae -27 Chevron bones 3^ Ribs 32 Sternum ......•••• 33 Scapula ......••••• 34 Hyoids 34 Appendix 34 Acknowledgments 3^ References .....•■••• 3° ON A SPECIMEN OF THE SOUTHERN BOTTLE- NOSED WHALE, HYPEROODON PLANIFRONS By F. C. Fraser, D.Sc. Department of Zoology, British Museum (Natural History) (Text-figs, i-ii) INTRODUCTION THE specimens of Hyperoodon planifrons, the Southern Bottlenosed Whale, of which there are published accounts, are few enough in number to be detailed. The type of the species in the British Museum collection is an imperfect, partly waterworn skull (Reg. no. 1814A) from Lewis Island, Dampier Archipelago, North- Western Austraha, described and figured by Flower in the Proceedmgs of the Zoological Society (1882). In the Anales del Museo de la Plata (1895), F. P. Moreno gives a brief account of three specimens : (i) Skeleton of an adult from the coast of the province of Buenos Aires. (2) Skull of an adult, Chubut Territory, Patagonia. (3) Skeleton of a young animal, Santa Cruz Bay, Patagonia. Finally, the Records of the South Australian Museum, vol. iv, no. 3, 1931, contains an account by H. M. Hale of a male which stranded near Port Victoria, Yorke Peninsula, South Australia. The present paper is concerned with the description of a skeleton from South Georgia, presented to the British Museum (N.H.) by the Discovery Committee, with an appended note about two additional specimens, no part of which has been preserved, from South Georgia and the South Orkneys respectively, in the Falkland Islands Dependencies. The widely separated regions from which the Southern Bottlenose has been recorded indicate the great area of distribution of this species. It may be presumed that its range includes the Southern Ocean generally and extends into the warmer parts of adjacent seas in the southern hemisphere. The Discovery skeleton (Reg. no. 1934.7.23 .3) belonged to an animal 6-5 m. long, a female, which was presented to the Discovery Committee by Capt. Sorlle, Westfold Whaling Co., Stromness, South Georgia. The skull and axial skeleton are in very good condition and almost complete, only the slender zygomatic arches in the skull, one or two of the terminal bones in the caudal series of vertebrae and probably one chevron being lacking. The appendicular portions of the skeleton are missing except the scapulae which are damaged. The sutures of the skulls of the Discovery specimen are all well defined and the epiphyses throughout the length of the vertebral column are not fused to the centra. In the South Australian specimen, which was only 0-4 m. larger. Hale states that the sutures of the skull are more or less ankylosed, and the figured vertebrae show no trace of separate epiphyses. These features suggest that, unlike the northern H. rostratus, in which the physically mature female is appreciably smaller than the male, in H. planifrons the two sexes must be about the same size when fully grown. Recorded dimensions of skulls of //. planifrons, together with the dimensions of a skull of H. rostratus for comparison, are given in Table i. LATERAL VIEW OF SKULL (Fig. i) In the description of the type specimen Flower (1882) drew attention to two features distinguishing H. planifrons from H. rostratus, both of which are most obvious in the lateral view of the skull. The first, the character which gives H. planifrons its specific name, is the relatively low development of a 0^ w O^O^O^r^J^r';0pC^ in Tl- in ON ►-t M t^ -1 00 NO II ON " iri K 0^ f^ob 00 " CO t^ i/-i ON ON N N ON N 'i- fO ON m r^, ^-^hNroO M »- N t^ 00 1- M 00 N i-i « 1^ >j fe is e f^tn m oou-ju-iMOvMroC^ •^ in rn in "H m 1 ^ " N ro ■-■ r^OvOrOt^i-iN« m ON NO On NO ON moo E roN -^ wivO^roK-lOO" N N m " N ' en i-i •^ •■^ " -< M ^ ^ Vi "■^ II 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 C 0^ II 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 •4-> Q. £ rt }_i 03 e E E 1 iri ir^ 00 1 1 1 N m 1 CO 1 1 1 N 00 1 1 1 *-i l^ t~- vO ro " ON NO 1 1 1 2^ " en 10 00 g ' m m 1 1 1 ^ i/^ 1 « N N M ' ' ' ffi" % *>> •2 s &^ = 1^ p oo^f^r^^or^oo OC> " io^66^N"t^" rn r^ CO rn -t W ON N NO in in N rn f^ N N nO ih II ON r^ ^t/^T)-ro-^"-i" N m t^ r-, M M M 00 N M "i^ g HH CU ' a speci 6 mCO tJ- Oi^'-tr*^^OGO'-t hH rr^, CO m ON ONCO N •-' ■g ov ^ t^ \0 N ONOO c^oo -i- CO r^ -jt- NO r-- Ti- r^ « f^ rt 6 w ro ro tosO U-) r^ -i-O "-i " N m ON 00 " N ^ ON m M ^ i \0 0^ II 1 1 II 1 ll-ll 1 1 1 1 1 1 Mill > -s s" s ,0 K-i r^OO 1 CO rn 1 1 1 1 I 1 1 1 E s r^ sO " t^ t^CO r^ r^jNO 00 1 1 III s 'ro r^ tovDiorOTh'n « N en ' 3 CO cl 03 t— 1 E CO u-iOOrot^t^"-lt^ r^ in 1 1 1 1 moo NO u^N r^i r^O>-'OOOt^O-<*- t-^ ON ON t-»CO NO -a E •H ro fO u-isO ^ ro -^^ ►"" ■-• N m ' 1 1 - ' ON N " g 2 " >n 00 ^0 0^ On ro t^ c^ r^ 9^^ ?^ 7*" 9^ 00 ^ t^db ^ " ro i/^ "-1 N r^- ON ^- ON NO 1 1 > p-i ON Ti-u->inc<^Ti-ioi-i N m " 00 N " "So ^ E r«-iio r^row^NmOOOi^j NO m 1 1 1 m 1 N N " C4 42 t-H ro t^ vO ON -^ N r^vO »o (^ ON 1 N ONOO N * E CON N rort--. 1'i U ^ 3 , E^ ^ s g^ '0 V M 00 Og BccS= <- S '0 h c .E -a m ^ 2S-al|g= °gl 1 ^il^S2<^|EEp^ -^ Ej= gj:^^^J=J= f,J= ^ «.S^2-SS^.SfSS-5-S^-SS a _c E J 1 -a tffl t« OX .1^ "Oh ;" 4) U S u THE SOUTHERN BOTTLENOSED WHALE 23 the maxillary crests. These in H. rostratus originate anteriorly approximately equidistantly between rostrum tip and antorbital notch, and ascend in a slope which varies according to age to a summit which overtops the skull vertex in all but the most juvenile specimens. Posterior to the summit there is a decline in level which is generally more abrupt than the anterior slope, the maxillar}' bone attaining normal thickness again before it ascends posteriorly in contact with the vertical portion of the frontal. The crests in H. planifrons originate anteriorly about two-thirds from the anterior end of the distance between rostral tip and antorbital notch. The slope is gradual to a low summit above the antorbital notch, and the decline posteriorly is equally gradual so that there is no horizontal thin portion of maxillary before it rises vertically in contact with the frontal in the occipital crest. In the Discovery specimen the maxillary crests are less massive than in the South Australian specimen. It may be that the difference is associated with sex, but it may equally v/ell be due to difference in age. Fig. I. Lateral view of skull and lower jaw. ( x \.) The second feature referred to by Flower, and visible in the lateral view of the skull, is the much larger size in H. planifrons of the crest formed by the vertex behind the nares. Not only is it much larger than in H. rostratus but it differs somewhat in shape, overhanging the narial area considerably, whereas in H. rostratus the anterior face of the crest viewed in profile is approximately vertical. Other differences will be mentioned when describing the dorsal aspect. In skulls of comparable size the distal portion of the rostrum is more slender in H. rostratus than in H. planifrons. The differences which exist in the proximal portion are associated with the dis- similarity of the maxillary crests. The distance of the maxillary tip from the tip of the premaxilla is alike in both species. The extent to which the lachrymal is seen in the lateral view appears to be equally variable in both species. The type specimen of H. planifrons has the left lachrymal completely separating the malar anteriorly from the orbital process of the frontal posteriorly and it has a wide contact with the maxilla. In the Discovery specimen it hardly appears in lateral view. It does not nearly reach the maxilla, and 24 DISCOVERY REPORTS the malar and frontal are in contact above it. The H. rostraius specimens examined showed a variety of form in the lachrymal. In one it did not appear in lateral view, in another there was a ventral portion separated from a more dorsal portion by a considerable extent where malar and frontal were in contact, and in a third the lachrymal completely separated malar from frontal and was in contact dorsally with the maxilla. This variation in the H. rostratus lachrymal is apparently connected neither with the age nor the sex of the animal. The temporal fossa in the Discovery specimen, like that of the type, is higher and shorter antero- posteriorly than that of H. rostratus. Apart from these differences the general form of the skulls is very similar and confirms the generic affinity of the two species with each other. DORSAL VIEW OF SKULL (Fig. 2) The differences between H. rostratus and H. planifrom are again clearly seen in the dorsal view. The massive prominences over the nares in H. planifrons extend forward so that the anterior boundary of the right one is almost vertically above the premaxillary foramen. The left, smaller in size, does not extend forward quite so far, but both in this aspect shut out any view of the narial apertures. The two bones are separated from each other by a parallel-sided gap bounded by the nasals. Hale states that in the South Australian specimen ' The inner anterior edge of each nasal (at the bottom of the groove) drawn up into a low thin flange'. This is the condition in the type, but in the Discovery specimen the flanges are wanting, so that there is no median ridge at the hinder end of the groove. All the specimens show the internasal gap diverging to the left from behind forwards and con- tributing to the asymmetry which characterizes the whole of this region of the skull. In H. rostratus the narial prominences do not extend forward so as to shut out completely the view of the nares ; the right boss, still larger than the left, has a greater width to length proportion than in H. planifrons. The internarial groove is wider, and has divergent, not parallel, sides. The nasal septum is similar in both species. It is strongly deflected to the left anterior to a pronounced emargination, and overlays to some extent the left premaxillary. Its extension forward in the gutter of the vomer is similar in the Discovery and the type specimen, in both of which it ends in the region of the posterior edge of the maxillary foramina about 100 mm. behind the level of the antorbital notches. The South Australian specimen has this ossification extending forward nearly to the antorbital notch level. In the region of the antorbital tubercle the outline of the skull is consistently different in the two species. In H. planifrons from the apex of the tubercle the external outline of the antorbital region extends posteriorly at an obtuse angle with the external edge of orbital process of the frontal, whereas in H. rostratus it is very nearly a right angle. Fig. 2. Dorsal view of skull. ( x J.) THE SOUTHERN BOTTLENOSED WHALE 25 Flower drew attention to the large size of the premaxillar}^' foramina in H. rostratus as compared with H. plaiiifrom, and this is consistent in all the specimens so far described and figured. The maxillar)' crests of H. rostratus rise vertically from the external margins of the maxillary foramina, and the inner faces are nearly parallel to one another. In old males especially, the crests approximate to such an extent as nearly to touch and thus form an arch over the prenarial portion of the premaxillae. The medial margins of the maxillary crests of H. planifrons overhang gutter-like extensions forward of the maxillary foramina. The inner faces diverge from each other at a very wide angle, and this feature, together with the lesser height of the crests in H. planifrons, provides one of the most conspicuous diagnostic differences between the two species. The vomer, which is without mesorostral ossification, is visible between the overarching anterior portions of the premaxillae. Its anterior tip is nearer the end of the snout in the Discovery specimen than in the South Australian specimen, the distance being 257 mm. as compared with 380 mm. In H. planifrons the greatest width of the premaxillae anterior to the foramina is about midway between the foramina and the premaxilla tip. In H. rostratus the greatest width is at about two-thirds of the distance from the tip. VENTRAL VIEW OF SKULL (Fig. 3) In ventral view such difl^erences as exist between the skull of H. rostratus and H. planifrons are of detail rather than of fundamental structure. The vomer in both species appears as two lenticular areas in the middle line of the rostrum. The anterior area separates the premaxillaries posteriorly and the maxillae anteriorly. There is then a short length where the maxillae are in contact in the middle line before the vomer appears again, when it is bounded partly by maxillae and partly by the palatine and pterygoid bones. The anterior portion of the vomer appears to be consistently shorter in H. rostratus than in H. planifrons. The palatine bones in both species are in two portions, palatal and lateral, separated by the pterygoid coming into contact with the maxilla. The palatal portion is bounded by maxilla, vomer and pterygoid, the lateral part by pterygoid and maxilla. In H. planifrons the palatal portion is a narrow strip with a greatest width, in the Discovery specimen, of less than a centimetre, and a length of about 10 cm. Each palatal portion in H. rostratus is roughly triangular in outline and of greater expanse (width about 4 cm. and length 11-5 cm., in a specimen of size comparable to H. planifrons). This diff'erence appears to be constant. The space between the palatal and lateral portion of the palatine, where the pterygoid anteriorly comes in contact with the maxilla, is much greater in H. planifrons than, in proportion to skull length, in any of the H. rostratus skulls in the British Museum collection. Incidentally it may be remarked that in Berardius arnuxii the palatal and lateral portions come into contact, being Fig. 3. Ventral view of skull. ( X l) 26 DISCOVERY REPORTS separated from each other only by a suture. The lateral portion of the palatine is smaller in H. planifrofis than in H. rostrattis, and in general the impression obtained is that in the former species the pterygoid anteriorly has expanded at the expense of the bones adjacent to it. The pterygoids are of typical ziphoid form in both species, ' large, solid, backwardly produced, meeting in the middle line, not involuted but simply hollowed on the outer surface' (Flower, 1871). The zygomatic process of the malar has its origin much nearer the posterior border of the bone in H. planifroiis than in H. rostratm, in which species it originates only a little way behind the antorbitai notch. Differences in the anterior margin of the antorbitai region involve the malar bone and were referred to in the description of the dorsal view of the skull. The lachrymal, a distinct bone, has the same essential form in both species. The extent to which it appears on the lateral border of the skull has already been referred to. It is long and narrow, extending obliquely backwards from the external margin of the skull to the infra-orbital foramen. It is bounded anteriorly by the malar and maxilla and posteriorly by the orbital process of the frontal. The external margin of the orbital process of the South Australian specimen is more pronouncedly concave than that of the Discovery specimen. No marked differences are discernible in the squamosals either between the South Australian and the Discovery specimens or between either of these and H. rostrattis. The tympanic-periotic bones are very similar in H. rostrattis and H. planifrons, and as in the former species so in the latter they are secured to the skull anteriorly by a slender inward-curving process from the squamosal embracing the periotic, and posteriorly by a rugose wedge-shaped extension from the tympanic between the squamosal and basi-occipital. In the posterior view of the skull all the available specimens of H. planifrons show the characters to which Flower drew attention in his description of the type, namely, the narrowness and greater height compared with H. rostratus and also the inferior size of the occipital condyles in the southern species. MANDIBLE (Fig. i) The jaws of H. planifrons compare closely with those of H. rostratus in general shape and in the extent of the symphysial region. The two rami of the mandible are not ankylosed at the symphysis in the Discovery specimen, whereas in the South Australian specimen Hale describes fusion as pro- ceeding, the two rami being linked by ossified bridges. In the former specimen the tooth alveolus at the tip of each ramus is continuous posteriorly with the dental groove, gradually merging into it. In the South Australian specimen (Hale's Fig. 4) the alveolus appears to be sharply defined from the dental groove. These differences between the jaws of the two specimens are such as might be expected from their difference in age. TEETH (Fig. 4) The outlines of the teeth of the Discovery specimen and of the South Australian specimen show the main differences between the two. Those of the former are conical and slender, and have a widely open pulp cavity. The dimensions are as follows : Right Left (i) Length 50 mm. 50 mm. (2) Greatest diameter 18 mm. 18 mm. (3) Diameter at right angles to (2) 17 mm. 16 mm. THE SOUTHERN BOTTLENOSED WHALE 27 The greatest diameter is just a little distance above the lower edge of the tooth, which has this indication of incipient closing of the pulp cavity. The tip of each tooth, an unworn crown of about 8 mm. length, projects from a thin investing coat of cement. The South Australian specimen has much more massive, fusiform teeth. Their length is comparable to that of the Discovery specimen — 57 and 59 mm. — but the greatest diameter is double. Apart from the difference in the pulp cavity (the root is entirely closed in the South Australian specimen) which is due to age, it is considered that the dissimilarity is associated with sex, and that in this as in other ziphoid whales the teeth of the male are large, massive and projecting above the gum, whilst those of the female are more slender, and, since the crowns are unworn, presumably concealed by the gum. Fig. 4. Teeth of H. planifrons. Upper pair, 9, Discovery specimen ; lower pair, S, South Australian specimen. ( x \.) Fig. 5. Anterior view of atlas. (xi) VESTIGIAL TEETH IN THE UPPER JAW When the Discovery specimen was received the skin and dried flesh on the ventral surface of the rostrum were still attached, and on each side of the upper jaw was a row of teeth commencing at about 24 cm. from the jaw tip and extending along the jaw about 16 cm. The teeth were spaced roughly equidistantly about 8 mm. from each other. All of the teeth were not in situ ; some had either been absorbed or had dropped out, but evidence of their existence was indicated by the fibrous follicles in which they had rested. It was estimated that each row consisted of twenty teeth, but the difficulty of dissection made exact computation impossible. Sixteen teeth were recovered on each side ; most of them projected 2-3 mm. from the dried gum, but whether this post-mortem conspicuousness existed in the living animal is doubtful. Their shape is fusiform and they are slightly to moderately curved. A basal portion consisting of cement envelopes the dentine of the crown to a greater or lesser extent, in some the junction between cement and dentine being clearly defined. The root portion of some of the teeth is drawn out into a needle-like extension. This is considered to be due to absorption in process, and in the shorter teeth, in which the extension has disappeared, it is presumed that the process has gone still further. The length of the teeth ranges from 4 to 14 mm. with diameter up to 2 mm. VERTEBRAE (Figs. 5, 6) Vertebral formula. Cervical 7, dorsal 8, lumbar 11, caudal 17 + . Cervical vertebrae. The Discovery specimen, like the South Australian and H. rostratus, has all seven centra fused together. The posterior epiphysis of the seventh is still distinct. In correspondence with the superior size of the occipital condyles in H. rostratus the anterior articular surface of the 28 DISCOVERY REPORTS I THE SOUTHERN BOTTLENOSED WHALE 29 X 60 30 DISCOVERY REPORTS atlas is also larger than that of//, planifrons. Otherwise the cervical mass is much alike in both species. Such differences as exist between the South Australian and the Discovery specimens may be regarded as coming within the range of individual variation. The former has the lateral process of the atlas fused with the inferior lateral process of the axis, whereas in the Discovery specimen the inferior lateral process of the axis is distinct. Both specimens show a short rugose superior lateral process on the axis, the South Australian specimen having ' an incomplete foramen on the right and complete foramen on the left between it and the inferior lateral process ', whilst the Discovery specimen has this arrangement of foramina transposed. The superior lateral processes of the third to sixth vertebrae are separate and of diminishing size antero-posteriorly in the Discovery specimen. The South Australian specimen has the third ankylosed on the left with that of the preceding cervical. The neural arch of the sixth is not completely fused with the arches anterior to it, and fusion is less on the left than on the right side. The corresponding arch in the South Australian specimen appears to be completely fused. There is a strong forward-projecting inferior lateral process on the sbcth vertebra of the Discovery specimen. Hale (1931) does not mention its presence in the South Australian specimen, and his figure shows that the inferior lateral process of the seventh is of con- siderable size and prominence and similar to that of the specimen of H. rostratus used for comparison with the Discovery H. plmiifrons. The inferior lateral process of the seventh in the Discovery H. planifrons is small and inconspicuous. Between it and the superior process is the articular facet for the head of the first rib. The neural arch is free except at the tip, whereas the South Australian specimen has the ' greater part of right side of neural arch free including apex which does not meet the opposite member of the arch '. Thoracic vertebrae. The Discovery specimen has eight pairs of ribs and therefore eight thoracic vertebrae. As the South Australian animal had nine pairs of ribs the possibility was considered of the ninth pair in the Discovery specimen having been overlooked. However, this is discounted to some extent by the fact that in the La Plata examples eight, not nine, is the number recorded. The reduction to this number represents the extreme reached in any of the Mammalia. The series of thoracic vertebrae in the Discovery Bottlenose commences with one having a slender neural spine, wide neural arch, widely separated zygapophyses, and short metapophyses at the proximal ends of transverse processes, which last are directed downwards and forwards and bear a facet for the tuberculum of the rib. There is a short centrum bearing a postero-lateral facet for the capitulum of the second rib. Proceeding tailwards the neural spines increase in length and width, the neural arches diminish in size, and the zygapophyses are very much reduced. The metapophyses, from being stout and short, are, in the eighth thoracic laminar, almost semicircular in outline and projecting from the anterior edge of the neural arch. The centrum at the end of the series is about double the length of that of the first thoracic. The arrangement of the articular facets for the ribs is interesting, and it is unfortunate that the centra of the vertebrae were damaged by the harpoon which killed the animal just at the point where detailed description is most required. However, enough remains to make some sort of interpretation possible. As far back as the fifth thoracic vertebra the articular facets are conspicuous on the postero- lateral edges of the centra. In the sixth vertebra the surface of the centrum on the left side has been obliterated, but the right side which is entire has only the very slightest indication of a facet, whilst the seventh vertebra has a distinct antero-laterally placed facet. It would appear therefore that as far back as the fifth vertebra the capitular articulation is with the rib of the succeeding vertebra, that the sixth is transitional between this arrangement and one in which the capitulum of the rib articulates with the centrum of the same vertebra with which the tuberculum is associated, and that in the seventh THE SOUTHERN BOTTLENOSED WHALE 31 this process is almost complete, with capitulum and tubercle of the seventh rib having articulation almost completely restricted to the seventh thoracic vertebra. The change in position of the transverse process from the side of the neural arch (upper transverse process of Flower, Osteology, 1870, p. 60) to the side of the centrum (lower transverse process of Flower, op. cit.) takes place in the eighth vertebra. There is not in the Discovery specimen as in the South Australian specimen a vertebra showing the transition from the one to the other kind of transverse process. The H. rostratus specimen used for comparison with H. planifrons showed in the eighth vertebra a condition intermediate between that of the other two specimens. In it the upper transverse process is in the form of a small knob-like and quite vestigial process on the lower margin of the metapophysis. Going tailwards the ventral surface of the centrum shows increasing development of the median ridge which is in the form of a well-defined keel on th. VIII. Lumbar vertebrae. There are eleven vertebrae in the lumbar series of the Discovery H. planifrons. The South Australian specimen has one less, but this discrepancy may be accounted for by the greater number of thoracic vertebrae in the latter specimen. The neural spines increase in length to about the middle of the series and then diminish gradually, so that a line joining their extremities makes a very shallow arcr There is an increasing inclination backwards of the spines going tailwards, a widening of the spine as a whole and of the distal end as well in the more posteriorly situated elements. The metapophyses are laminar, have rounded margins, and show increasing approximation to each other. The neural canal diminishes in size ; the centrum increases so that at the end of the series it is about i J times the length of the first lumbar ; the diameter also is increased. The transverse processes are directed obliquely forward, flattened, beginning to diminish in length, and get wider at the tail end of the series. The first lumbar transverse process is somewhat different from those that succeed it, being disproportionately broad and rather stouter. The hypophysial ridge is of increasing definition to about the middle of the series, whence it diminishes in prominence ; and in the last lumbar it is a low, flattened, inconspicuous keel. No obvious differences distinguish the vertebrae in this region from those of H. rostratus. Caudal vertebrae. The caudal series of vertebrae is incomplete in the Discovery specimen. Seventeen remain and the missing elements are at the posterior end. The South Australian specimen has 20 caudals. The neural spines diminish tailwards and disappear after the tenth caudal. In lateral view they are broad distally with a slight narrowing towards the neural canal. There is a corresponding diminution of metapophyses which anteriorly in the series are laminar with rounded border, and posteriorly are rather stout short tubercles which finally disappear. The neural canal continues the diminution in size observed in the lumbar series. Anteriorly the centra have the massiveness which characterizes the more posteriorly placed lumbars and, going tailwards, although length diminishes gradually, the decrease in transverse diameter is not noticeable until near the end of the column where the diminution becomes more marked and the vertebrae adopt a subcuboid shape unlike the cylindrical form of the more anterior elements. The transverse processes disappear as distinct prominences after the seventh caudal. While still distinguishable they maintain the obliquely forward direction noted in the lumbar vertebrae. The perforation of the transverse process of the seventh, noted by Dale, is represented in the Discovery specimen by a pronounced emargination of the outer edge of the process on each side near its posterior end. This is visible, although much less obvious, on the transverse processes of two vertebrae im- mediately anterior to the seventh caudal. On the lower surface of the centrum anteriorly and posteriorly are the paired facets for the chevron 32 DISCOVERY REPORTS bones. Two longitudinal ridges with concave margin join the anterior to the posterior facets. The concavity is ill defined at the anterior end of the series, and is correlated with the lesser prominence of the facets themselves ; but going tailwards with the greater development of the articular surfaces and the shortening of the length of the mass of the centrum, the emargination becomes increasingly pronounced until on the ninth (in both the South Australian and the Discovery specimens) a foramen is enclosed. CHEVRON BONES (Fig. 7) The nine chevron bones figured are an incomplete series; at least one is considered to be wanting. However, those remaining give an adequate idea of the form these bones assume in H. planifrons. Fig. 7. Chevrons. ( x \.) Only one side of the first chevron is present, a slender lamina of bone which has no evidence of having been fused to the element of the other side. The second chevron, a single bone, has a broad, short, spinous process with obliquely rounded ventral margin. The third has the spinous process greatly elongated with rounded antero-ventral margin, and with hinder and ventral margins meeting at roughly a right angle. From the third tailwards there is a progressive diminution in the spinous process length and a reduction in size of the bone as a whole, in the last of the series the spinous process being only about one-half as long as it is wide. The chevrons show no distinctive difference from those of the South Australian specimen or of H. rostratus. RIBS (Fig. 8) The Discovery H. platiifrons has eight ribs on each side, in this number agreeing with the La Plata Museum specimens. The South Australian specimen has nine pairs of ribs, the ninth pair being small, asymmetrical and obviously vestigial. H. rostratus normally has nine pairs of ribs also, but at least one specimen in the British Museum collection has only eight pairs. In the Discovery specimen the first pair of ribs is short, broad, flattened and with sternal end directed at a slight angle forward from the remainder of the shaft of the bone. The second rib is moderately broad, more elongated than the first and without forward trend of the distal end. The third to the sixth are similar to each other, long, slender and subequal in length. In the seventh, shortening of the shaft has become pronounced, but otherwise the essential features of the four preceding ribs are maintained. The eighth is still shorter, and in the absence of a capitular portion is distinguished from all the ribs that precede it. The first seven ribs have the capitulum defined to a greater or lesser degree. In the first the capitulum and tubercle are almost confluent, in the following five the capitulum is situated at some distance from the tubercle. In the seventh the tubercle and capitulum approximate again and the eighth, as just stated, has no capitulum. THE SOUTHERN BOTTLENOSED WHALE 33 STERNUM (Fig. 9) The sternum consists of three elements, the largest of which is the manubrium. The manubrium is roughly rectangular in outline. The anterior emargination is semicircular and not so pronounced as in the South Australian specimen. There is a small posterior notch, and the bone extends tailwards on the right side of this to a greater extent than on the left. Asymmetry is also displayed on the lateral margins. The facets for the first pair of sternal ribs are equally prominent, but whereas the Fig. 9. Sternum. ( x |.) Fig. 8. Ribs. ( x i.) right side bears a facet a little way posteriorly to the first there is no corresponding one on the left. The external surface of the bone is convex and the internal concave. Anteriorly, the second sternal element has a median notch, on the right side of which the anterior margin is a little way behind that on the left side. This asymmetry is repeated on the posterior margin, in which, however, the notch is wanting. The lateral margins are shallowly concave, and at the antero- and postero-lateral corners are facets for the appropriate sternal ribs. 34 DISCOVERY REPORTS The last sternal bone has again an uneven anterior border, the left side being in advance of the right. It is without anterior notch. The posterior margin has a deep, angular notch extending nearly to the middle of the bone ; in the South Australian specimen it is wide and shallow. There are three facets on each side for sternal ribs, one at each antero- and postero-lateral corner and one midway between these. The ventral surface of the bone is raised into a low, ill-defined tubercle. SCAPULA (Fig. lo) Both the scapulae of the Discovery specimen are damaged posteriorly. Anteriorly the evenly convex dorsal margin meets the straight anterior margin at almost a right angle, not being broadly rounded as in the South Australian specimen. The acromion, as in the latter specimen, is bent upwards and inwards, the superior and inferior margins being parallel to each other and the distal margin rounded. It is shorter than in the South Australian specimen. The coracoid is without the distal expansion noted in the South Australian' specimen, but is otherwise similar in position and shape. Fig. 10. Scapula. ( x ^.) Fig. II. Hyoids. ( x ^.) HYOIDS (Fig. II) The thyro-hyals are not fused to the basi-hyal. The basi-hyal has a short, straight anterior margin and deeply concave posterior margin. The lateral portions of the bone which are convex are rugose, and are completely occupied by the facets for connexion with the thyro-hyals. The thyro-hyals are wing-like in shape, and stoutest at their proximal ends where there is a broad area for attachment to the basi-hyal. The bones diminish in thickness from the anterior to the posterior border, where the upper and lower surfaces meet in a ridge at a very acute angle. The distal tips of the thyro-hyals are truncated and rugose. The tympano-hyals are elongate, flattened and tapering at each end to a truncated rugose tip. The thickness of the bone diminishes from the front to the hinder margin, which last has a fairly acute edge. APPENDIX (i) A male specimen of H. planifrons was measured and examined by Dr L. Harrison Matthews, at Leith Harbour, South Georgia, on 3 January 1927. It was intended that the skeleton should be preserved, but before it could be despatched to England an avalanche, which obliterated part of the whaling station, buried the specimen, and it was not recovered. THE SOUTHERN BOTTLENOSED WHALE 35 The external measurements recorded by Dr Matthews are as follows : m. Total length, tip of snout to notch of flukes 4-63 Projection of lower jaw beyond tip of snout Nil Tip of snout to blowhole 0-74 Tip of snout to angle of gape 0-85 Tip of snout to centre of eye 076 Tip of snout to tip of flipper 1-72 Notch of flukes to posterior emargination of dorsal fin i -27 Width of flukes at insertion 0-39 Notch of flukes to centre of anus i'33 Notch of flukes to umbilicus 2-46 Centre of anus to centre of reproductive aperture 0-38 Vertical height of dorsal fin 0-25 Length of base of dorsal fin 0-37 Axilla to tip of flipper 0-42 Anterior end of lower border to tip of flipper 0-51 Length of flipper along curve of lower border 0-55 Greatest width of flipper 0-17 Length of severed head from condyle to tip 0-697 Greatest width of skull 0-369 The following notes were also made : Colour Black dorsally shading to grey ventrally External genitalia Normal External parasites None Hair None Ventral grooves Two grooves on the throat, one on each side situated under the ramus of the mandible, 22 cm. in length Blubber 5 cm. thick on the side below the dorsal fin Palate Grey Tongue Flesh-pink Food Stomach contained a few crystalline lenses from the eyes of cephalopods Internal parasites None seen Mammary slits Each 4 cm. in length, situated 12 cm. anterior to the anus (2) Mr A. G. Bennett, at one time naturalist to the Government of the Falkland Islands, has provided another record of the occurrence of H. planifrons. He obtained photographs of a specimen killed in the vicinity of the South Orkney Islands in January 191 5. One of the photographs, in which the carcass is floating in the water alongside the factory ship, shows the surface of the skin scored by numerous irregular marks. Similar streaks have been noted in other ziphoids and are presumed to be the teeth marks of other individuals of the same species. In addition to these elongated scratches one or two oval marks can be seen. They are reminiscent of the scars described and figured by Mackintosh and Wheeler (1929) as occurring in various members of the whalebone whales. Other features which can be observed in the photograph are the pronounced ' forehead ' which rises at almost a right angle from the well-defined beak ; and the right flipper which is of typical ziphoid form, having a very shallowly convex lower border and slightly more convex upper edge. A second photograph gives a ventro-lateral view of the anterior portion of the body, lying on the deck of the whaling vessel. The region of the mouth and throat, as far back as the two ventral grooves is of a much lighter colour than adjacent portions of the body. The ' forehead ' appears to be quite darkly pigmented. The rostrum is stout and well defined and the upper and lower lips meet in a line which anteriorly is horizontal but farther back swings obliquely upwards. 36 DISCOVERY REPORTS ACKNOWLEDGMENTS I have to thank Dr L. H. Matthews and Mr A. G. Bennett for the information and assistance they have given me. The figures illustrating the paper are the work of Col. M. St L. Simon, and it is with pleasure that I acknowledge my indebtedness to him, and to my colleague W. H. T. Tams, Esq., who took the photographs from which the figures of the axial skeleton were executed by Col. Simon. I have also to thank Mr E. J. Manly who has helped me with the compilation of the report. REFERENCES Flower, Sir Wm., 1871. On the recent ziphoid whales, with a description of the skeleton of Berardius arnouxi. Trans. Zool. Soc. London, vol. viii, part in. Flower, Sir Wm., 1882. On the cranium of a new species of Hypeioodon from the Australian Seas. Proc. Zool. Soc. London. Flower, Sir Wm., 1885. An Introduction to the Osteology of Mammalia, 3rd ed. Hale, H. M., 1931- Beaked whales — Hyperoodon planifrons a7td Mesoplodon layardii — from South Australia. Records of the South Australian Museum, vol. iv, no. 3. Mackintosh, N. A. and Wheeler, J. F. G., 1929. Southern Blue and Fin Whales. Discovery Reports, vol. i, pp. 257-540. Moreno, F. P., 1895. Nota sobre los Restos de Hyperoodontes conservados en el Museo de la Plata. Anales Mus. de la Plata. Secc. Zool. III. [Discovery Reports. Vol. XXIII, pp. 37-102, Jw/zc, 1945] REPORT ON ROCKS FROM WEST ANTARCTICA AND THE SCOTIA ARC By G. W. TYRRELL, A.R.C.Sc, D.Sc, F.G.S., F.R.S.E. (Lecturer in Geology, University of Glasgow) CONTENTS Foreword, by J. M. Wordie, M.A page 39 I. Petrography of the South Shetland Islands, West Antarctica . . 41 II. Petrography of Rocks from the Graham Land Peninsula and Adelaide Island, West Antarctica 66 III. Petrography of Rocks from the Elephant and Clarence Group 76 IV. Petrography of Stones dredged from the Vicinity of the Shag Rocks . 89 V. Petrography of the South Sandwich Islands 92 REPORT ON ROCKS FROM WEST ANTARCTICA AND THE SCOTIA ARC By G. W. Tyrrell, a.r.c.sc, d.Sc, f.g.s., f.r.s.e. Lecturer in Geology, University of Glasgow (With Geological Notes by N. A. Mackintosh, D.Sc, and J. W. S. Mark, M.A., B.Sc.) (Text-figs. 1-14) FOREWORD By J. M. WORDIE, M.A. In the second volume oi Das Antlitz der Erde published in 1888, and again in more detail in the final volume in 1909, E. Suess put forward the view ' that the Andes are to be seen again in Graham Land '. By this dramatic phraseology he implied that the folded mountain border of the Pacific, as exemplified in the Andes, swings eastward from Tierra del Fuego to South Georgia and then curves back from the South Sandwich Islands through the South Orkneys to Graham Land and the South Shetlands. Suess based his views on a memoir by H. Reiter in 1886,^ who there gave substance to an idea put forward as far back as 1831 by Sir John Barrow."^ Suess characteristically gives the credit for these arguments to Reiter, whose paper I have not seen, but it is not unlikely that it was Suess himself who suggested this work; the first volume of the Antlitz had appeared in 1885, and there can be no doubt but that the ideas of the second volume would already have formed themselves in the author's mind, and this was a problem which required to be examined. Andersson, in his Geology of Graham Land, in fact mentions that Reiter had been stimulated by Suess's first volume. In the interval between Suess's first statement in 1888 and his more detailed advocacy in 1909, Dr Otto Nordenskjold led the Swedish Antarctic Expedition to the east coast of Graham Land in 1901-3, and J. Gunnar Andersson who was with him published his important Geology of Graham Land in the Bulletin of the Geological Institute of Upsala, vol. vii, Upsala, 1906. Nordenskjold himself was also much alive to the problem and has both described the rocks, Petrographische Untersuchungen aus den Westantarktischen Gebiet, Upsala, 1906, and also put forward an authoritative statement of the whole problem in Handbuch der Regionalen Geologie: Antarktis, Heidelberg, 1913. Nordenskjold and Andersson carried out in the field what Reiter had sensed in the study. Andersson, Nordenskjold, and Suess together may , therefore be regarded as the main advocates of 'two groups of Antilles'. 'South Antilles' was the name first given to the islands of the southern arc; but more recently the sea enclosed by these islands has been named the Scotia Sea, and the name South Antillean Arc has now automatically been replaced by the more appropriate title of Scotia Arc. Andersson and Suess could base their arguments only on imperfect data, some of which are now known to be incorrect. Since then many new rock specimens have been obtained and worked on by qualified geologists. The activities particularly of the Discovery Committee have succeeded in providing collections surpassing all previous material. Dr Tyrrell has already dealt with some of the collections in earlier papers on South Georgia, the South Sandwich Islands and the South Shetlands ; and in the present memoirs he is at last able to make authoritative statements on the remain- ing portions of the arc either scantily known or completely unexplored at the time when Suess made his great analysis of the plan of the Earth. 1 H. Reiter, Die Siidpolarfrage imd Hire Bedeutimg fiir die genetisctie Gliedenmg der Erdoherflache, Weimar, 1886. " Sir John Barrow, Journal of tlie Royal Geographical Society, vol. i (1832), p. 62. 40 DISCOVERY REPORTS Dr Tyrrell's main conclusions are as follows : Two dredgings were made from 'Discovery IT in the neighbourhood of the Shag Rocks in November 1930. Of the nineteen specimens obtained fifteen are described as tremolite-epidote- greenstone or greenstone-schist. This is an important find, as it can be paralleled both with rocks from Clarence Island and with specimens from Tierra del Fuego. Fresh material has been obtained in the South Sandwich Islands both in situ at Saunders Island and from dredgings elsewhere in the group. These rocks are all volcanic in origin and of Recent age. The new material, along with earlier collections, shows that the South Sandwich rocks have more in common with rocks from the Antilles of North America than with any specimens so far known from the Andes. Dr Tyrrell considers that the South Sandwich Islands probably lie on a ridge parallel to, but east of, the main Scotia Arc. Elephant Island and Clarence Island and others east of the main South Shetland Islands not only lie at some distance from the South Shetlands proper but also differ from them geologically. A greenstone-greywacke-mudstone association is present, such as is formed in the geosynclinal stage of a mountain-building cycle and is affected as would be likely by low-grade metamorphism. Assemblages of this character are found not only in the Elephant and Clarence Group but also in the South Orkneys. They are paralleled near Ushuaia in Tierra del Fuego, and a somewhat similar assemblage occurs in South Georgia. Dr Tyrrell considers that these types may also be expected to form the at present unknown rock basement of Graham Land. There are extensive collections from the South Shetlands which modify earlier conclusions. The occurrence of sediments of presumed Mesozoic age on certain of the islands has apparently been over-emphasized, and one should now regard the South Shetlands as of preponderatingly volcanic origin, made up either of lavas, mainly andesites, dacites and rhyolites, or of their associated tuffs, breccias and agglomerates. Plutonic rocks may, however, be commoner than so far supposed. There were two lava periods, and the intrusive rocks, such as the diorite on King George Island, are regarded as the underground equivalents of the later period. The Recent volcanoes along Bransfield Strait are still younger than either of the above lava periods, and it is even probable that Deception Island and Bridgeman Island have been active in historic times. The chemical characters of the Deception Island lavas indicate a soda-rich andesite, not readily paralleled in the Andes. Elsewhere the andesites and basalts are of normal circum-Pacific, that is to say undoubted Andean, type. Finally, a fifth section deals with some specimens from Graham Land. These are less numerous as a collection, but they include a quartz-porphyry formation at Adelaide Island of the same nature as the rocks of a belt 400 km. in length already known from Patagonia. No new rocks are to hand either from the South Orkneys or from South Georgia. Both localities are now well known. The importance of the new material lies in the nature of the rocks themselves, and Dr Tyrrell, in these five papers, has provided petrographic arguments for what was up till now not more than a matter of inference. The petrographic evidence is more or less complete. To settle the actual line of the Arc, however, requires that the bottom contours should be better known. Soundings over a wide area are much to be desired, and will decide whether there is a single arc or a series of concentric curves. Meantime one can safely say that Suess's, Andersson's and Nordenskjold's arguments no longer relate merely to a possibility, and that Suess's vision of the Pacific structure advancing into the Atlantic must now be regarded as firmly established. 41 PART I. PETROGRAPHY OF THE SOUTH SHETLAND ISLANDS INTRODUCTION TH I s work is based on two collections of rocks, made during the third and fourth commissions of the 'Discovery II' in 1934 and 1937 respectively. The specimens were accompanied by excellent geological and geographical notes, those of 1934 by Dr N. A. Mackintosh, and those of 1937 by J. W. S. Marr, M.A., B.Sc. Relevant points from these notes have been incorporated, with appropriate acknowledgement, in the following descriptions. Bibliography. A full bibliography of the earlier literature relating to the geology and petrography of the South Shetland Islands (and adjacent lands) is given in my paper listed as (i) below. Only papers which have been published since 1920 are given in the following list: (i) G. W. Tyrrell. 'A Contribution to the Petrography of the South Shetland Islands, the Palmer Archipelago, and the Danco Land Coast, Graham Land, Antarctica.' Travis. Roy. Soc. Edinb. Liii, pt. I, 1921, pp. 57-79. (2) H. H. Thomas. 'On the Innes Wilson Collection of Rocks and Minerals from the South Shetland Islands and Trinity Island.' Ibid. pp. 81-9. (3) O. Holtedahl. 'The Geology and Physiography of Some Antarctic and Sub-Antarctic Islands.' Scientific Results of the Norwegian Antarctic Expeditions, 1927-28 and 1928-29, instituted and financed by Consul Lars Christensen, No. 3, Norske Vidensk.-Akad., Oslo, 1929, 172 pp. (4) T. W. F. Barth and P. Holmsen. ' Rocks from the Antarctandes and the Southern Antilles (Being a Description of Rock Samples collected by O. Holtedahl, 1927 28, and a Discussion of their Mode of Origin).' Ibid. no. 18, 1939, 64 pp. General. The South Shetlands comprise a group of ten large and small islands extended in a north-east to south-west direction parallel to, and at a distance of from 60 to 70 miles from, the coast of the Graham Land peninsula, from which they are separated by Bransfield Strait. From north-east to south-west the islands are Bridgeman Island, King George Island, Nelson Island, Roberts Island, Greenwich Island, Livingston Island, Deception Island, Snow Island, Smith Island, and Low Island. Of these, practically nothing is known of the two last-named. Deception Island, a sea-flooded Recent crater, is the best known. Bridgeman Island, too, is a Recent volcano and may, like Deception Island, have been in comparatively recent eruption. Mr Marr's notes make it clear that Penguin Island, off the eastern horn of King George Bay in King George Island, is also a Recent volcano comparable with Deception Island and Bridgeman Island. The rock specimens collected during the recent Discovery II expeditions number in all 141, of which 81 come from King George Island, 19 from Deception Island, 17 from Roberts Island, 16 from Livingston Island, 4 from Nelson Island, and 4 from Snow Island. The plan of the present paper is to describe the collections from each of these islands in turn, incorporating as much of the geology as can be gleaned from the field notes made by Dr Mackintosh and Mr Marr. The chemistry of the igneous suite will then be studied with the aid of previously published and two new analyses, and finally a conspectus of the geology of the South Shetlands will be attempted from the material now available. 43 PETROGRAPHY KING GEORGE ISLAND Admiralty Bay. The Ullmann Range, a ridge trending north and south, projects into Martel Inlet (north-east arm of Admiralty Bay) and forms the eastern side of Visca Anchorage. Specimens were collected from the western side of this ridge. In his notes, Dr Mackintosh has given an excellent sketch of the Ullmann Range as seen from Visca Bay, and has called attention to a prominent dike which climbs the scarp and culminates in a sharp pinnacle near the central point of the ridge. This view is undoubtedly the subject of Mr Ferguson's fine photograph (Ferguson, op. cit. pi. iii, fig. 1),^ which clearly shows the dike and a series of lava scarps to the left (north) of it. The dike consists of a highly porphyritic pyroxene-andesite with phenocrysts of plagioclase (basic andesine, Ab^j), yellow augite, and chloritic pseudomorphs after orthorhombic pyroxene, in order of abundance. There are also some large irregular masses of magnetite. The ground-mass is fine- grained, but apparently holocrystalline, although somewhat altered. It contains a little quartz. The lavas of which the Ullmann Range is composed are represented by several specimens mainly collected from screes. Alongside the dike occurs a trachytic lava with a very dense fluxional ground- mass, consisting of minute feldspar microlites, apparently orthoclase, in a cryptocrystalline base. There are numerous small phenocrysts of soda-orthoclase and a plagioclase which is now mostly albite, but the presence of epidote suggests that it may originally have been a more calcic variety. The rock also carries numerous euhedral crystals of ilmenite rimmed with a leucoxenic alteration product. Traces of ferromagnesian minerals are present, but are altered beyond recognition. This rock is notable in containing a few crystals of pale blue pleochroic apatite. A coarser textured specimen provides further data. The ground-mass is seen to consist of laths of orthoclase mingled with oligoclase, and contains visible quartz. Still another specimen consists of an angular breccia of fragments similar to the above. Many of the fragments are rich in quartz. The shapes of some pseudomorphs outlined in iron ores suggest that the ferromagnesian mineral in these rocks may have been hornblende. These lavas may be provisionally classed as dacite or quartz-latite according to the amount of quartz or orthoclase present. Similar types have been described from Admiralty Bay by the author ((i), p. 71). They also occur in the Fildes Strait area (p. 44). Near the beach on the western side of the Ullmann Range was collected a lava which maybe described as an altered quartz-andesite. It contains phenocrysts of plagioclase badly carbonated, and chloritized pseudomorphs after pyroxenes. Quartz is comparatively abundant, but is partly of secondary origin. Bluish apatite crystals are abundant, and the lava is therefore regarded as belonging to the same series as those described above. From the screes to the south of this point a silicified and pyritized volcanic tuff was collected. Mr Marr collected three specimens from the western side of the Keller Range along the eastern shore of Mackellar Inlet. He describes this coast as consisting of slopes of reddish brown tuff with frequent outcrops of lava which are also prominent at sea level. While two of his specimens are so highly carbonated and silicified that they can only be described as altered andesites, the third, which is stated to have come from a fan-shaped columnar outcrop, is less altered, and can be described as pyroxene-andesite. Feldspar phenocrysts are numerous and, although badly carbonated, can be identified as plagioclase of composition about k\,An^,. The ferromagnesian constituent consists of chloritized pseudomorphs after pyroxenes, usually found in crystal clots along with feldspar, ilmenite, and large crystals of apatite. The ground-mass is dense, brown, and cryptocrystalline, the only identifiable constituent being feldspar microlites showing straight extinction (.? oligoclase). 1 For full reference see p. 76. 44 DISCOVERY REPORTS Two specimens were collected by Mr Marr from near Point Thomas, Admiralty Bay. One, from the coast a little south of the Point, is a fresh hypersthene-andesite. This rock appears to be identical with the rock called hypersthene-augite-bandaite of the volcanic vent of Three Brothers Hill, Potter's Cove, Fildes Strait, described by the author ((i), p. 68) from Mr Ferguson's collection, and the reader is referred to this full description for petrographical details. In fact, Mr Ferguson actually collected material from the same area ((i), p. 69). The extreme freshness of this rock, as compared with the extensive alteration suffered by the lavas from the interior of Admiralty Bay, suggests that it belongs to the later of the two volcanic episodes on the mainland of King George Island. On the other hand, the rock collected by Mr Marr from the coast of Ezcurra Inlet, one mile west of Point Thomas, is an altered pyroxene-andesite which clearly belongs to the older series of lavas. This occurrence suggests that the boundary between the older and newer series of lavas should be drawn a little farther south than is shown on Mr Ferguson's map (D. Ferguson, op. cit. supra, fig- 2, p. 38). Fildes Strait. Fildes Strait separates King George Island from Nelson Island to the west. Dr Mackintosh collected several specimens from a harbour (St. 1482) near the south end of the strait, which may be identical with the ' Potter's Cove ' of Mr Ferguson, or it may be the ' Marian Cove ' of the same author which is a little farther north. Dr Mackintosh describes the rocks as much weathered, breaking down into screes through which solid rock appears here and there. Three of the specimens from this locality are dark, very compact rocks of basaltic type. They consist mainly of a very fine-grained ground-mass of intersertal type with numerous microlites of a striated feldspar giving extinctions up to 20° (andesine), scattered patches of chlorite and obscure brownish material probably representing pyroxenes, and particles of haematitized iron ore embedded in a reddish cr>'ptocrystalline or glassy base. The few small phenocrysts consist of epidotized plagioclase (originally labradorite), and, in one section, fresh, euhedral, colourless augites of small optic axial angle (.? pigeonite). A chemical analysis (p. 59) shows that these rocks must be regarded as of tholeiitic composition. One specimen from this locality, however, is much more acid than the above, and must be classed as soda-rhyolite or quartz-keratophyre. It is a whitish felsitic rock much reddened by haematitic staining. In thin section it is seen to consist of a dense quartzo-feldspathic ground-mass with an obscure hint of spherulitic structure, which carries numerous large phenocrysts of turbid albite and haematitized biotite. A single specimen was collected from another locality on Fildes Strait near the narrow northern entrance (St. 1483). Dr Mackintosh states that the rock formation here appeared to be quite different from that of St. 1482, an observation which is confirmed by examination of the specimen. One adjacent islet consisted of a dome-shaped mass of rock, ' probably basalt ', with a pronounced columnar structure, but the outcrop from which this specimen was collected was not columnar. This rock turns out to be a feldspathic olivine-basah or olivine-andesite. Large phenocrysts of fresh basic labradorite (Ab^Ang) are very abundant. Calcified and serpentinized olivines are numerous, but a fresh pale augite is quite subordinate in amount. These are embedded in an intergranular ground-mass consisting of plagioclase laths, augite and iron-ore granules, and a dark crypto- crystalline base. North Foreland District. The North Foreland is the tip of a long narrow peninsula springing from the north-eastern corner of King George Island. A shorter peninsula ending in a steep blulT headland called Brimstone Peak occurs a mile or two to the west, and the two peninsulas enclose a deep bay. Still farther west comes the well-known Esther Harbour, which was apparently not entered on this occasion. This district (St. 1949) was visited by Mr Marr. SOUTH SHETLAND ISLANDS 45 Mr Marr writes that ' the cUffs forming the west side of the Foreland ... are composed of a massive grey rock much traversed by cracks and joints, giving it a very shattered appearance '. This is borne out by the three specimens collected here, which are all parts of a plutonic igneous rock of variable grain size. This may be described as quartz-hornblende-pyroxene-diorite, and represents a ver)' abundant type in West Antarctica ((i), p. 6i). Its three principal minerals are plagioclase (core andcsine; outer shell oligoclase) ; pale green hornblende, sometimes with a pale brown tint; colourless diopsidic pyroxene which is altering into a pale green amphibole. The accessory minerals are quartz, filling the interstices between the main constituents; some large flakes of reddish biotite; abundant ilmenite altering to leucoxene ; and a considerable amount of apatite. The amphibole and pyroxene tend to form well-shaped crystals, and to enter into clots with biotite and ilmenite. One of the specimens is a true plutonic type with allotriomorphic texture and comparatively coarse grain. Another is a fine-grained aplitic type poorer in the mafic minerals, which may be styled quartz-microdiorite ; and the third is a porphyritic type in which the feldspars, hornblendes and pyroxenes (including both augite and hypersthene) occur as phenocrysts in a fine-grained granulose ground-mass. A few large crystals of bluish apatite occur in this rock. This type may represent a chilled marginal phase of the intrusion. It is clear that the vicinity of North Foreland is occupied by a large plutonic intrusion of the same type as occurs at Noel Hill, Marian Cove ((i), p. 6i), and at Le Poing on the west side of Admiralty Bay ((i), p. 62). This mass may occupy the whole of the eastern side of King George Island, as Mr Marr states that the cliflFs to the east and south of the Foreland, and probably as far as Cape Melville, are high and sheer, and seem to consist of the same grey massive rock. Brimstone Peak is said to be composed of perpendicular 'basalt' cliffs rising sheer out of the sea to a height of 150 ft. The single specimen obtained from this locality shows, however, that the rock is a fresh hypersthene-augite-andesite of the Recent type so common elsewhere in King George Island. The hypersthene is mostly altered to chlorite or bastite, and often forms the core of an augite crystal. A single crystal of magnetite-rimmed brown hornblende was present in the thin section. Bolinder Beach (St. 1953) is situated a few miles west of Esther Harbour and Brimstone Peak. It is described by Dr Ommanney as a bluff peak crowned by three buttresses of dark grey and light brown rock veined by what, on closer examination, proved to be finely crystalline rose and amber quartz. All the rock specimens collected here were lost in a boat accident except a few from a 100 ft. cliff at sea-level on the northern face of the bluff. This rock proves to be an enstatite-andesite of micro-porphyritic and intersertal texture, consistmg of very numerous feldspar laths (andesine, AbgAn,), and less abundant pseudomorphs in chlorite after enstatite (typical square prisms with truncated corners), in a dense, brown, cryptocrystallme to glassy ground-mass. It probably belongs to the older series of lavas, as it is intersected by mineral veins which may represent the same group of veins (quartz and pyrites) as that described by Ferguson from the islands of Esther Harbour {op. cit. supra, p. 41). These veins run nearly east and west, and might thus probably intersect the region of Bolinder Beach. Pengum Is/and and Adjacent Mainland. Penguin Island is situated off the eastern horn of King George Bay. That Penguin Island is a Recent volcano, one of the line of volcanoes fringing Bransfield Strait, is Mr Marr's important and most interesting discovery. The following is a description of Penguin Island quoted from Mr Marr's report: The southern half of Penguin Island is a volcanic cone. The northern half consists of a long, very low plateau, much of it only about 50 ft. high. The western face of the cone is steep and has a deep brick-red tint. On its south- eastern and eastern sides the cone slopes down to a plateau roughly 100 ft. high, which is continuous in a wide sweep with the lower plateau which forms the northern half of the island. On the southern side the cone ends 46 DISCOVERY REPORTS abruptly in sheer and inaccessible cliffs from 50 to 100 ft. higii which continue round the coast to the eastern side of the island. The rock is lava, at a distance dark in colour, and much broken with cracks and fissures. . . .The island is remarkably free of snow and ice, and although snow may lie thinly on it after a heavy fall it does not remain for long. [This fact strongly suggests that there is still much residual heat in the cone, and that it may only be dormant.] Penguin Island is a volcanic cone in the shaping of which three, and perhaps four, periods of activity seem to have been involved. What seems to have been the earliest and biggest eruption is represented now by the concave section of a very large, but almost entirely cut away crater which occupies nearly the whole of the western face of the cone, from the shingle beach up to the summit. The degree of concavity is not very high, yet it is unmistakable. The sides of the interior of this now almost destroyed cone are composed of rather finely divided volcanic clinker of a rich brick-red colour which gives this side of the island its characteristic tint. The clinker fragments have the even consistency of a coarse gravel. Projecting out of this eroded crater, its base on a level with the beach, is a huge plug [? dike] of lava from three to five feet in width and rising vertically like a wall for nearly a hundred feet. Similar though less conspicuous plugs [dikes] occur elsewhere in this crater. Main summit crater. A later eruption is perhaps represented by this crater, a third of a mile across and about 200-300 ft. deep, which occupies the summit of the cone. Evidently the rim of this crater has crumbled away considerably, for it is highest to the north, but slopes downward towards the south (see sketch, Fig. 2). The bottom is rather damp and shows signs of there having been water lying about. On the east side of the interior of the bowl a gigantic plug of lava sticks up vertically for about 100 ft., the top, however, not projecting beyond the rim of the crater. There is some quite deep snow, which is possibly permanent, inside the bowl on its north-east side. Fig. 2. Penguin Island. Another eruption, subsequent to that which produced the main summit crater, is represented by the small secondary cone which rises concentrically from the bottom of the former. The secondary cone is about 100 ft. high and has a crater less than 80 yards across at the rim, and about 20 ft. deep. Ash beds. Much of the lower part of the cone, and a large part of the 100 ft. plateau to the south-east and east of it, seem to be composed of horizontally stratified ash beds of a light colour. The coastal cliffs throughout are composed of lava often broken by cracks and fissures. On the eastern side of the island the crests of the cliffs are extremely rugged and often twisted into grotesque sliapes, evidently the result of cooling in the surface of an ancient lava flow. At the south-west corner of the island a certain warmth was felt on the lava and inside a fissure. The heat experienced was very slight, but we were of the opinion at the time that it was unlikely to have been due to absorption from the sun. Crater on east side. On the east side of the island, some 60-80 yards from the coast, another old crater occurs in the 100 ft. plateau. Its rim is flush with the general level of the plateau, and it is rather a remarkable sight, strongly resembling an old quarry. It is a perfect circle and about 150-200 yards across at the rim. The sides are steep, descending for at least 50 ft. There is deep water at the bottom in which a few penguins were swimming; the water was not icy cold. On its west side the crater cuts through horizontally stratified, light-coloured beds of volcanic ash at least 30 ft. in thickness. On the eastern rim of the crater there is much glassy lava, obsidian, of various hues. All specimens of the lavas collected from the volcanic cone of Penguin Island represent textural variants of a typical olivine-basalt. The most fully crystallized type comes from the plug in the summit crater. In thin section it is found to be highly porphyritic with numerous phenocrysts of fresh olivine and pale brown augite, sometimes aggregated into clots, and very numerous micro- SOUTH SHETLAND ISLANDS 47 phenocrysts of plagioclase (Ab55An4g) with both chemical and mechanical zoning, embedded in an intergranular ground-mass consisting of feldspar microlites mingled with granules of augite and iron ores. In other specimens the ground-mass contains some glassy matter usually blackened with iron-ore dust, and is of intersertal or cryptocrystalline texture. In one of the rocks olivine is serpentinized and much reduced in amount, but its place is taken by a small quantity of pleochroic hypersthene, illustrating the affinities of these olivine-basalts with the more common hypersthene-augite-andesite lava-type. This association suggests that the olivine- basalts are possibly due to some accumulative process operating in the early stages of the crystallization of a pyroxene-andesite magma from which olivine began to separate. A closely comparable olivine-basalt has been described from Edinburgh Hill, a volcanic vent in Livingston Island on the M'^Farlane Strait coast (Ferguson, op. cit. p. 44; (i), p. 66). Mr Ferguson's fine photograph {op. cit. pi. i, fig. i) illustrates the magnificent columnar structure of this plug. An olivine-basalt also occurs in the Desolation Islands, off the northern coast of Livingston Island (this paper, p. ^i). Olivine-basalts of very similar characters have been described by H. H. Thomas from Roberts Island ((2), p. 86). Basalts have also been described from the volcanoes of Deception Island and Bridgeman Island. The mainland coast opposite Penguin Island, according to Mr Marr, consists of cliffs of lava, fronted by extensive raised shingle beaches. Only one specimen was collected from this locality. This is a typical augite-andesite with a beautiful pilotaxitic texture. The few phenocrysts are small and consist mainly of a colourless augite which is, however, occasionally zoned with cores and bands of a yellowish variety. The remaining phenocrysts are of andesine feldspar (Ab5An4). This lava is quite fresh and no doubt belongs to the younger lava series. Many specimens of the coarse rounded shingle on the beaches of Penguin Island and the adjacent mainland were collected. These consist of the older andesite lavas, together with many of the typical plutonic rocks of the region — granite, adamellite, tonalite, quartz-monzonite, quartz-pyroxene- diorite, etc., and two highly metamorphic types, quartz-chlorite-biotite-schist and hornblende- granite-gneiss. Martin s Head and The Lions Rump. These are conspicuous adjacent headlands on the western side of King George Bay. Mr Marr's report states that the basal portion of both headlands consists of a dark grey columnar 'basalt' about 100 ft. in thickness, and with the columns inclined at a steep angle towards the south. At Martin's Head the ' basalt ' is overlain by a massive rock with a ' twisted appearance' (? confused columns), and from 50 to 60 ft. in thickness. This in its turn is covered by what appeared to be a tuff (Fig. 3). Behind the headlands are tuff slopes characterized by an abundance of angular rock fragments of many different kinds (.'' agglomerate). About 200 ft. above the Lion's Rump there is what appears to be an old volcanic crater, now almost completely filled with dirty stagnant ice (Fig. 4). A little to the north of the headland is a conspicuous lava flow reaching the sea. Near by, perched on the beach, are several gigantic erratics of conglomerate, one of which must weigh more than 200 tons. The conglomerate is exceedingly coarse, containing rounded water- worn stones from a few inches in diameter to some 2 ft. across. The columnar lava of Martin's Head is a fresh hypersthene-augite-andesite of the type common among the younger lava series. An andesite of similar type, but much richer in feldspar phenocrysts, poorer in augite, and apparently devoid of hypersthene, was collected i mile east of the Lion's Rump. From the same locality comes a green mudstone, consisting of finely divided quartz and vermicular chlorite, much of the latter being aggregated into small rounded or ellipsoidal pellets. It is difficult to diagnose this rock in the absence of data regarding its field occurrence, but it may be a muddy sediment made up of decomposed wash from a surface composed of the older andesite lavas. m^t^JtA jmjii. b<^c^c^U.^W*eir 'buv^t^U.tA jMriOTVJMviv Fig. 3. Martin's Head. 7ajA-''iMr N A M. F'g- 5- coarse conglomerate of well-rounded boulders indicating a long period of erosion, suggest that, notwithstanding their freshness, they belong to the older series of lavas. Examination of a series of pebbles from the agglomerate in the vicinity of Martin's Head and Lion's Rump shows that the majority consist of hornblende-augite-andesite lavas and their tuffs. In addition, there is an altered doleritic andesite somewhat similar to that described above, a highly epidotized andesite obviously belonging to the older lava series, and an altered tonalite in which the feldspars have been thoroughly sericitized and epidotized, and the ferromagnesian minerals chloritized. The hornblende-andesite is an unusual type which has not hitherto been described from West Antarctica. In the best-preserved specimen brownish green pleochroic hornblende in well-shaped crystals comes next to plagioclase in abundance as phenocrysts, and is greatly preponderant over augite. The ground-mass is dense and cryptocrystalline. "nelson island Harmonv Cove. Very little is known about the geology of Nelson Island. Mr Ferguson {op. cit. p. 43) visited Harmony Cove, a harbour at the western corner of the island where Nelson Strait joins Bransfield Strait, and collected a quartz-diorite-porphyry which appeared to be intrusive into an igneous breccia. Dr Mackintosh collected four specimens from Harmony Cove. His account is almost entirely topographical, but he has provided an excellent sketch of the rock exposures (Fig. 5). Study of these specimens confirms Mr Ferguson's results. One of them is a fine-grained norite 50 DISCOVERY REPORTS consisting of labradorite (somewhat albitized and epidotized), fresh pale augite with which the feldspar laths are sometimes in ophitic relation, numerous brown pleochroic pseudomorphs after hypersthene, and much diffused chloritic matter. There is also a micro-granular variant of this type with porphyritic feldspars and hypersthene (bastite), and highly epidotized. An outcrop near the glacier (Fig. 5) consists of pyroxene-andesite of a type common among the older lava series. It shows porphyritic feldspars (andesine), pale brown augite, and chlorite pseudomorphs after ortho- rhombic pyroxenes, in a very fine-grained intergranular ground-mass. The fourth specimen, from the shore, is an igneous breccia mainly composed of angular fragments of altered andesite, much epidotized, and peppered with cubes of secondary pyrites. Mr Ferguson's specimen of igneous breccia from the same locality, however, is rich in fragments of the more acid dacitic and rhyolitic lavas. COPPERMINE COVE , GnQLISM STRAiT, SOUTH SHETLAND: *i.tfV^ ->-"*j-L.-i:rta^ii STATION 14-85 TABL€ I .CUM ii.JXJ^a4> X tnajiKd lEl tun«^j.t"iot^;ti^ CI ^jiuv^ ^<.Mo C«_ «^\ii.*. ahaUc. ■DVKE ^ ;^r1^f-?-^--rt-yi^?-rfT-^-.-y;^^Ty^^ Fig. 6. 1^^ " ROBERTS ISLAND Coppermine Cove. This anchorage is situated at the north-western end of Roberts Island, close to the multitude of small islands and rocks which are scattered over the northern exit of English Strait. Specimens were collected by Dr Mackintosh from a small peninsula ending in a flat-topped columnar rock known as Fort William (Fig. 6). Opposite the anchorage (reports Dr Mackintosh) are cliffs of reddish breccia, presumably volcanic, and Fort William appears to consist of columnar basalt. In this respect it resembles Table Island, and many, if not all, of the islets and rocks in the vicinity. Many rock specimens were collected between the anchorage and Fort William. A dike about 5 ft. thick cuts the cliff opposite the anchorage. The only previous description of rocks from Roberts Island is that by H. H. Thomas ((2), pp. 85-7). He describes five specimens from Coppermine Cove, all porphyritic olivine-basalts and all showing considerable variations in the relative abundance of the porphyritic constituents, and in the richness of the ground-mass in ferromagnesian minerals. Most of Dr Mackintosh's specimens are also olivine- basalts of varying composition and texture. Thus the columnar rock of Fort William is a feldspathic olivine-basalt, or rather dolerite, with a ground-mass of excessively coarse intergranular texture SOUTH SHETLAND ISLANDS 51 composed of lathy plagioclase (about AbiAiii), pale brown augite, and iron ores. Both feldspar and augite occasionally attain micro-porphyritic dimensions. The abundant fresh olivine, however, forms large phenocrysts. A 'common type' along the shore is a basalt with numerous small feldspar phenocrysts, and less numerous olivine and augite crystals, embedded in a ground-mass of intersertal texture. This recalls the Dunsapie type of the Scottish Carboniferous, as was also remarked by Dr Thomas. Another type which appears to be abundant in this locality is one with an intergranular ground-mass exceedingly rich in augite. Dr Thomas described rocks of this type but, unlike our specimen, his material contained much olivine. Some of these augite-basalts, as they might be called, carry numerous little prisms of l(iw double refraction and straight extinction which are identified as enstatite, in the ground-mass along with the monoclinic pyroxene. This is an enstatite-basalt. Dr Thomas described a similar rock as hypersthene-basalt. While most of the specimens collected here are basalts, one is an augite-andesite of the common type belonging to the younger lava series. It is accompanied by an andesitic agglomerate. Beach pebbles collected from Coppermine Cove consist of tonalite and granite-aplite. LIVINGSTON ISLAND Livingston Island is the second largest of the South Shetland group, but very little is known of its geology. Mr Ferguson collected an olivine-basalt from a fine columnar exposure forming a small island off the coast in M^Farlane Strait (Edinburgh Hill), and noted tuff's in the vicinity which, beside basalt, contained fragments of quartz-diorite and black mudstone [op. cit. p. 43 and pi. i, fig. i). Desolation Island. Dr Mackintosh collected a few specimens from Desolation island which lies off the northern coast of Livingston Island. He gives no geological details except that the island is mainly composed of a columnar igneous rock. It is noteworthy that on the Discovery Chart (Discovery Reports, vol. vi, 1932, Chart 6) Desolation Island is represented in the shape of an irregular broken ring, suggesting that it may be a breached crater flooded by the sea ; but this resemblance may, of course, be quite accidental. Two of the specimens were collected in situ from columnar outcrops. Both are very fresh and coarse- grained hypersthene-basalts of an unusual type. The major part of both rocks consists of a coarse intergranular admixture of laths of labradorite (Auen-An^o) with granules of pale green augite, prisms of enstatite-hypersthene with faint pleochroism, and iron ores. The feldspar and augite occasionally form somewhat larger micro-porphyritic crystals, but the rock is not conspicuously porphyritic. Both kinds of pyroxene, moreover, tend to build small aggregations or clots, which stand out as a glomero-porphyritic texture. Olivine occurs only sparingly as small pseudomorphs in brownish serpentine. A small amount of dark brown glass fills up interstices in the ground-mass. A basalt with orthorhombic pyroxene in the ground-mass was described by Thomas from Roberts Island ((2), p. 86). OUvine did not occur in this rock, and the augite occasionally formed glomero- porphyritic aggregates. A closely comparable rock from the same locality has been described in this paper (p. 51). These rocks are no doubt closely related to the basic hypersthene-augite-andesites above described, which are so common in the South Shetland Islands. In these rocks, however, the hypersthene is porphyritic and does not occur in the ground-mass. Barth and Holmsen have given an interesting discussion of the petrographical problem involved in the presence of hypersthene in these rocks ((4), pp. 14^17)- Numerous pebbles from the beaches and fragments from the screes of Desolation Island were collected. These include tonalite and a sericitized and chloritized diorite, silicified andesitic breccia, and a series of acidic volcanic rocks including a fluxional rhyolite or dacite with augite, a rhyolitic 52 DISCOVERY REPORTS tuff made up of angular fragments of the fluxional rock, a biotite-rhyolite, and orthoclase-porphyry or felsite with only sparse phenocrystic quartz. Finally, a fragment collected from the scree on the cliffs of a rocky islet near the anchorage turns out to be a crushed sericitic quartzite of a distinctly ancient aspect. DECEPTION ISLAND Deception Island is the best known of the South Shetland Islands. Dr Thomas ((2), p. 81) has commented on the earlier literature of the island. Mr Ferguson added a few details and published two excellent photographs {op. cit. p. 44; pi. iii, figs. 2, 3); but the fullest recent description is that by Holtedahl ((3), pp. 29-47). Deception Island apparently represents a huge breached crater flooded S£. WALL OF OecePTiON HflKBWR , SOUTH SHeTLAMOS. >«i- ^.S^'i OMctc^y . STATION 1484 ttju-46wu^ ^//^<^'>>' Souk .(Sjj SeoLclv dan* I II --^*"^4, ''^'':c'' N ■ A .M. Fig. 7- by the sea, of which the inner diameter is about 8 km. Holtedahl believes, however, that it is not a single large crater, but a volcanic ring mountain built around a caldera subsidence bounded by a circular fault or series of faults. Dr Mackintosh collected material from the cliffs and slopes on the south-east side of the whaler's anchorage near the entrance to Deception Harbour. These form a narrow ridge of land separating the anchorage from Bransfield Strait (see Dr Mackintosh's sketches. Fig. 7). He reports that the whole of the cliffs shown in the sketch, except beyond Neptune's Bellows,^ consist of an 'agglomerate of ashes in a yellowish matrix'. It is possible that the yellow colour is mainly superficial, as freshly broken surfaces generally seem darker. The slopes below the cliffs are mainly of a soft gravel obviously formed from the disintegrated agglomerate, carrying a fair proportion of solid boulders of agglomerate, and here and there boulders of a harder dark rock presumably derived from intrusions in the agglo- merate (andesitic basalt). 1 Apparently the name given to the entrance channel of Deception Harbour. SOUTH SHETLAND ISLANDS 53 A visit was also made to the bluff on the south-west side of Neptune's Bellows (Fig. 7). The lower part consists of conspicuous red cliffs, but higher up there are outcrops of the yellowish agglomerate characteristic of the other side of the channel. The main range of hills in this locality appeared to be composed of ' cindery lava or scoria ' with reddish black tints. It appears to be the weathered surfaces of this rock which impart the striking red colour to the lower cliffs. Three rock specimens were collected from this locality, and a few from localities north of Whaler's Bay (Anchorage?). The petrography of Deception Island has been dealt with by the writer ((i), pp. 67, 71), who described olivine-basalt and basaltic tuffs, ^ and hyalo-dacite (ungaite). Dr H. H. Thomas ((2), pp. 81-5) described ophitic olivine-dolerite, various types of andesite and their tuffs (mostly glassy), and soda-trachyte (oligoclase-trachyte). He also noted the presence of tridymite and iron-olivine (fayalite) in some of the more acid types, and of anorthite in the hyalo-andesites. Barth and Holmsen ((4), pp. 8-17) described andesine-basalt and a vesicular, glassy 'pillow-lava', both of which they regarded as of bandaitic composition, a view which is borne out by their chemical analyses. Further- more, they gave a full description of a rock which seems to be identical with my oligoclase-dacite and Thomas's oligoclase-trachyte. Barth and Holmsen find the closest analogues of this rock in the products of the Santorin volcano in the Aegean Sea, and as it contains 17 per cent of tridymite they call it tridymite-santorinite. From the study of Mr Ferguson's original specimens on which I based my first account of the rocks of Deception Island, of Dr Mackintosh's new material, and of the above literature, it seems clear that four main types of rock have been erupted from the Deception Island volcano, namely, olivine- basalts or dolerites (of which there are no analyses), lavas of bandaitic composition, hyalo-andesites of more acid type, and finally, the trachytic type which has been variously called oligoclase-dacite, oligoclase-trachyte, and tridymite-santorinite. Eight analyses of Deception Island rocks have been published (p. 58) from which it seems clear that they form a perfectly gradual series varying from basic to acid, all of which (except the olivine-basalts) are highly sodic and relatively poor in potash ; and are mineralogically characterized by the presence of calcic feldspars, orthorhombic and monoclinic pyroxenes, and, in the more acid types, by fayalite, tridymite, and sodic feldspars. The following account of the petrography of Deception Island is based on the study of the specimens collected by Dr Mackintosh, and on the re-study of the material collected by Mr Ferguson ((i), pp. 58 et seq.). Olivine-basalt. Only two rocks, both from the Ferguson collection, belong to this type. One is described in the following terms ((i), p. 67): 'A beautifully fresh rock showing more or less rounded olivine phenocrysts in a ground-mass of good fluidal texture, which consists of elongated microlites of labradorite with subordinate granules of augite and magnetite.' The texture can be described more exactly as fluxional intergranular. A few of the augite crystals are of slightly larger dimensions and more euhedral than the granules of the ground-mass, and can be regarded as micro-phenocrysts. The rock has a close resemblance to the Dalmeny type of the Scottish Carboniferous basalts. Its occurrence is as a pebble in a tuff or agglomerate. The other olivine-basalt is flow-banded in the hand specimen, but its ground-mass is not so con- spicuously fluxional as the above. The ground-mass is of coarse intergranular type and consists of laths of andesine, with granules of pale augite and magnetite. Numerous phenocrysts and glomero- porphyritic aggregates of fresh olivine and brown augite, together with smaller and much less numerous feldspar crystals (labradorite) are embedded in the ground-mass. This rock has aflinities with the Craiglockhart and Dunsapie types of the Scottish Carboniferous basalts. Basaltic andesites of bandaitic type. These rocks differ from the basalts described above in not 1 These are now regarded as andesitic tuffs. 54 DISCOVERY REPORTS being conspicuously porphyritic, and in being almost or quite devoid of olivine. All but one of the six specimens available come from Dr Mackintosh's collection, and were obtained from both sides of the entrance channel to Deception Harbour. The ground-mass is of the same type as that of the basalts, that is, composed of andesine laths, and granules of augite and iron ores. A plagioclase of somewhat more basic character forms numerous laths which run in wavy flow-lines through the ground-mass. A few large phenocrysts of augite may occur, but olivine, if present at all, is always in very small quantity, and is altered to brownish serpentine. The ground-mass varies in texture from coarsely intergranular to fine-grained intersertal, with a brownish glassy base blackened with iron-ore dust. These rocks are adjudged to be the same as those described by Barth and Holmsen ((4), p. 9) as andesine-basalt and pillow-lava of bandaitic type, of which they have provided chemical analyses (p. 58). Dr Thomas, too, described what is apparently the same type, in the more basic varieties of his 'hyaloandesites' ((2), p. 82). Both Barth and Holmsen, and Dr Thomas, mention hypersthene as a constituent of this rock type, but the writer was unable to identify orthorhombic pyroxene with certainty in the material at his disposal. Andesite {hyalo-aiidesite). This is the most abundant rock type in both Mr Ferguson's and Dr Mackintosh's collections. As the analyses show (p. 58), there is a continuous series of com- positional types from the basic bandaites to the relatively acid oligoclase-andesites (santorinites), varying chiefly in silica percentage and proportion of ferromagnesian to feldspathic minerals and quartz. As many of the rocks are of glassy facies, these variations are masked, at least mineralogically, by the glassy matrix; in thin section the rocks present a relatively unvarying appearance and, except for one or two more crystalline types, may be grouped as hyalo-andesites. Dr Thomas ((2), p. 82) described several rocks from Deception Island under this heading. In hand specimens these rocks are black or dark grey in colour, usually slaggy, vesicular or even pumiceous, and are obviously of glassy nature. Even the more crystalline varieties are black and of dense texture. From these black slaggy types there are all transitions to dark, non-vesicular, glassy rocks, resembling pitchstones, which are, however, more acid than the majority of the types grouped under the name hyalo-andesite, and properly belong to the oligoclase-andesites or santorinites. In thin section many of these slaggy rocks are found to be composed of a brownish glass, dusted thickly with black specks of iron ores, and often highly vesicular. They always show swarms of plagioclase microlites (oligoclase to andesine), usually in parallel fluxional streams, but occasionally felted together with the production of pilotaxitic texture. Microlites of pyroxene can often be detected in varying numbers by their bright polarization tints and oblique extinction. Some microlites, however, which are indistinguishable from the pyroxenes in their appearance under ordinary light, have a very high double refraction and straight extinction. ^ It is probable, therefore, that these are olivines. Olivine does actually occur in very small amount in a few of the rocks as micro-phenocrysts, and is almost invariably altered with the production of a reddish serpentine. There are also occasional micro-phenocrysts of andesine and augite. From these highly vitreous types there are all gradations to almost holocrystalline (micro-crystalline) types consisting of a very dense intergranular admixture of plagioclase microlites with granules ot augite and iron ore, which carries fluxional streams of plagioclase laths. Dr Thomas detected well-formed ciystals of tridymite lining steam cavities and planes of flow in these rocks ((2), p. 84). Barth and Holmsen ((4), p. 1 1 e^ seq.) found no less than 17 per cent of tridymite lining steam cavities in one of the more acid types. The writer found abundant tridymite in only one of the vesicular hyalo-andesites. It lines and fills steam cavities and fracture cracks m Barth and Holmsen ((4), p. 9) have also noted small elongated crystals of olivine in the ground-mass of these rocks. 1 SOUTH SHETLAND ISLANDS 55 the rock. Associated with and apparently passing into the tridymite aggregates there are a number of small spherulites giving a perfect extinction cross, of which the constituent fibres have straight extinction and a refractive index much lower than that of canada balsam. While these may be tridymite, it is possible that they represent cristobalite. A. G. MacGregor has described both tridymite and cristobalite from the Recent lavas (pyroxene-bandaite) of Montserrat.' He writes: 'The cristobalite, besides obviously replacing tridymite laths and twins, often occurs as innumerable rounded to irregularly shaped spots up to o-i mm. across', but he does not mention any spherulitic structure. Oligoclase-andesite (oligoclase-trachyte — Thomas; santorinite — Barth and Holmsen; oligoclase- dacite (ungaite) — Tyrrell). This rock represents a somewhat more acid development of the magma which gave rise to the hyalo-andesites above described. Its nomenclature presents a rather per- plexing problem, and it has been given various names by different authors as shown above. As indicated by the analyses (p. 58), the free silica works out at between 15 and 20 per cent. The writer has shown that the average andesite contains round about 15 per cent of normative quartz;- and as the principal feldspar in the rocks under discussion is oligoclase, it is thought that oligoclase-andesite is the best name for the type. It is, however, of somewhat unusual composition, as shown by Barth and Holmsen ((4), p. 13), in that the ratio of soda to potash is much higher than in normal andesites. They have marked this distinction by conferring the name santorinite, since the lavas of Santorin are found to be the closest analogues of this rock type. Perhaps the most acid types should be called oligoclase-dacite to mark the presence of as much as 20 per cent of free silica. In hand specimens these rocks vary from light grey compact 'stony' to black pitchstone-like material, which carries scattered whitish crystals of feldspars and often shows marked parallel banding due to flow. In thin section they are seen to contain very sharply bounded micro-phenocrysts of plagioclase, augite, enstatite, olivine (fayalite) and magnetite, embedded in a ground-mass which varies greatly in its proportion of glass to crj'stals. The glass may form at least 50 per cent of the ground-mass ; at the other extreme the rocks may be almost completely crystalline. The glass is usually yellowish brown in colour, but may be colourless; it contains many minute needle-like crystallites. Numerous microlites of oligoclase-albite (and perhaps a potash-soda feldspar) stream through the glass in fluidal fashion, mingled with minute granules of pyroxenes and iron ores. The feldspar micro-phenocrysts were identified in my earlier memoir as anorthite ((i), p. 71). Dr Thomas also found anorthite in his material ((2), p. 82), but Barth and Holmsen ((4), p. 1 1) apparently noted only andesine of composition AbesAnas . The ferromagnesian phenocrysts include augite (probably diopside) in well-shaped prisms and octagonal basal sections, enstatite and fayalite. The micro-phenocrysts often cluster in groups. Only one of the rocks w^as vesicular, and in it was found tridymite lining steam cavities exactly as reported by Barth and Holmsen. These rocks resemble some of the more basic pitchstones of the Tertiary igneous episode in the west of Scotland, notably the types called leidleite and inninmorite,^ especially the latter, which is reported to contain anorthite phenocrysts. Indeed, the text-figures of the microscopic appearance of leidleite and inninmorite (e.g. figs. 47, 48) given in the Mull Memoir cited above might pass for some of the hyalo-andesites and oligoclase-andesites of Deception Island. Tiijf and agglomerate. Every account of Deception Island emphasizes the abundance of fragmental volcanic rocks — tufi^ and agglomerate — in the constitution of the volcano. Five specimens from 1 The Royal Society Expedition to Montserrat, B.W.L: 'The Volcanic History and Petrology of Montserrat, with Obser- vations on Mont Pele, in Martinique', Pliilus. Trans., B, ccxxix, 1938, pp. 58-61. - G. W. Tyrrell, ' Some Tertiary Dykes of the Clyde Area', Geol. Mag. 1917, p. 31 1- 3 'Tertiary and Post-Tertiary Geology of Mull, Loch Aline and Oban', Mem. Geol. Surv. Scotland, 1924, pp. 281-4. 3-2 56 DISCOVERY REPORTS Dr Mackintosh's collection have been sliced, and they are found to be singularly uniform in com- position. They are made up of irregular, angular, and highly vesicular lapilli and scoria, the fragments usually var^'ing in size between a hazel-nut and a walnut. The fragments consist of glassy forms of both the basic and acid andesitic types, the black opaque slaggy form and the clear glassy form being about equally abundant. The glassy fragments are frequently of a bright yellow colour, but some are brown and a few others of a greenish tint. Many of these fragments have a narrow border of the black opaque variety, suggesting that the separation of magnetite dust in the glass which gives rise to the opacity may be due to a reheating or annealing process. There is little or no matrix of finer material between the fragments, and they appear to be welded together along their contacts. This material therefore might be better classed as agglutinate'^ than as agglomerate. Tridymite (and cristobalite?) occurs abundantly in these fragmental rocks, not only lining the vesicles of the glassy fragments, but also as an edging around the individual fragments. This suggests that, in these rocks at any rate, the tridymite is of deuteric crystallization. It has been formed shortly after the consolidation of the fragmental material, and is no doubt due to late emanations derived from the parent magma. SNOW ISLAND This is a small island west of Livingston Island, and west-north-west of Deception Island. It is geologically unknown, and no description and no record of any landing is known to me. Four specimens of rocks from Snow Island, however, were found in the first set of material sent to me by the Discovery Committee, with no record when and by whom collected. Three of the rocks appear to have been collected in situ from actual exposures, but the fourth is a pebble from a raised beach at 50 ft. above present sea-level on the eastern coast of the island. Of the three specimens collected />/ situ on the eastern side of the island one is a quartz-pyroxene- diorite or feldspathic quartz-gabbro of a type identical with other occurrences in the South Shetland Islands and the Palmer Archipelago; the second is an oligoclase-andesite breccia with a tuffaceous matrix containing a good deal of quartz. The third is a quartz-felsite or rhyolite with a scanty crypto- crystalline matrix. The pebble from the raised beach is quartz-augite-microdiorite, identical with the quartz-pyroxene-diorite above mentioned except that it contains patches of fine-grained ground-mass. Even from this scanty material, therefore, the indications are clear that the constitution of Snow Island is the same as that of the other islands of the South Shetlands group, and that rocks of the older igneous series are here represented. DREDGINGS FROM BRANSFIELD STRAIT A few score of stones dredged from two stations in Bransfield Strait were included in the first collection of rocks received from the Discovery Committee. These came from St. 175, about 25 miles south-east of Deception Island, and St. 177, about 27 miles south-west of Deception Island, and were dredged from depths of 200 and 1080 m. respectively. The stones were probably dropped from the ice which formerly occupied Bransfield Strait, and which probably moved from the west and south- west. Some of the material may have been carried by icebergs breaking away from glaciers on the South Shetlands and the Graham Land coast. The specimens range in size from blocks 6 in. across to \ in. pebbles. Most of them are angular and facetted, with corners and edges roughly rounded oflF; only a few appeared to be well-rounded, apparently water- worn pebbles. As was to be expected, the great majority of the seventy-nine stones sliced consist of the older series of andesites, dacites, rhyolites, agglomerates and volcanic breccias, which appear to constitute ^ G. W. Tyrrell, Volcanoes {Ylome. University Library), 193 1, p. 66. SOUTH SHETLAND ISLANDS 57 the main part of the South Shetlands, and perhaps some part of the Palmer x'Vrchipelago and the Graham Land coast. There is also one hyalo-andesite with good tridymite which certainly comes from Deception Island and two others which probably come from the same source. Rocks of plutonic aspect are also well represented in this collection. They include the quartz-pyroxene-diorites and their porphyries which are common in the South Shetlands and adjacent regions. Diorite, tonalite, granodiorite, biotite-granite, and their porphyries, together with granophyric granites and true granophyres, which more probably come from the Palmer Archipelago and adjacent parts of Graham Land, are also fairly abundant. Rarer types are represented by a basic diorite with abundant brown hornblende, biotite, and apatite ; and a serpentine derived from augite-peridotite. The most interesting material, however, is provided bv specimens of sedimentary and metamorphic character, which are unrepresented among the rocks in the Discovery collections obtained from actual exposures. Little is known of these rock types in the South Shetlands and adjacent regions as they have attracted little attention, perhaps owing to the relatively great abundance and conspicuous characters of the igneous rocks. Many of the sediments represented among the dredged stones have suffered a low-grade cataclastic metamorphism by crushing and shearing. Among the unaltered sediments are mudstone, siltstone, greywacke, arkose and sandstone. There are two mudstones, and both appear to represent exceedingly fine-grained washes from the weathered surfaces of basic lavas. Microlites of plagioclase can be recognized in a chloritic and ferruginous clay matrix, and in one of them there is a sparse sprinkling of angular quartz grains of silt grade. Another mudstone of similar type has undergone a little crumpling and shearing with the development of thin quartz-chlorite veins. Seven pebbles appear to represent laminated sediments consisting of alternate beds of grey\vacke and siltstone or slate in various stages of shearing and crushing. The least altered specimen shows angular grains of quartz and subordinate feldspar in a siliceous ground-mass of silt grade in which quartz is mingled with finely divided sericite, chlorite, epidote and iron ores. This material is pene- trated by thin veins of secondary silica, now recrystallized to lines of granular quartz. The other members of this series have undergone severe cataclasis, whereby ultimately quartz-chlorite-schist has been developed from the greywacke bands and phyllite from the slaty bands. Three of the specimens show signs of having first been broken up by crushing into an angular breccia in which, by further shearing, the fragments have been drawn out with the production of a kind of mortar structure, and with the development of much coarse chlorite and white mica. In one specimen, which is relatively poor in quartz and rich in chlorite and epidote, it is probable that basic igneous rock fragments made up the greater part of the original greywacke. The extreme term of alteration is represented by a true schist consisting largely of quartz, biotite and sericite, in which mortar structure is finely developed. One specimen is an interesting arkose consisting of extremely angular grains of quartz, alkali- feldspar and plagioclase, small chips of andesite and keratophyre (?), a few bits of garnet and epidote, and many flakes of unaltered biotite, in a ferruginous clay matrix. This composition suggests the rapid waste of a mixed terrain consisting of granitic rocks, andesitic lavas, and perhaps some meta- morphic rocks. Finally, there is a true sandstone consisting mainly of angular to subrounded grains of quartz, with less abundant grains of alkali-feldspar and plagioclase, a few chips of slate and siltstone and, above all, many large angular fragments of pale garnet. Mudstones, greywackes, quartzites and igneous breccias have been described from the South Shetlands, but especially from the Palmer Archipelago ((i), p. 74; (4), p. 28). Ferguson {op. cit., p. 37) described siliceous and argillaceous sediments interbedded with the lavas and tuffs of the older 58 DISCOVERY REPORTS igneous series in Admiralty Bay, King George Island. The present study of dredged stones from Bransfield Strait has brought out the fact that somewhere in the surrounding region there must be a basement series of greywackes, mudstones and slates, which has undergone severe cataclastic meta- morphism. There is good evidence from contact-metamorphic effects that the plutonic masses of the South Shetlands, the Palmer Archipelago and Graham Land, have broken through this sedimentary basement ((i), pp. 75-7), and also through the older series of andesite lavas. Hence the metamorphosed sedimentary basement must be at least of early Mesozoic age, and quite possibly Palaeozoic. CHEMICAL CHARACTERS For a discussion of the chemistry of the igneous series of the South Shetland Islands there are available twelve previously published analyses and two others made for the present investigation and here published for the first time. Seven of the twelve published analyses were given by E. Gourdon Table i a. Analyses of igneous rocks from Deception Island I 2 3 A 4 5 6 7 8 SiOa 69-01 68-28 67-71 68-33 60-62 56-89 52-93 53-50 49-84 AI2O3 14-21 15-95 14-65 14-94 16-22 16-07 15-86 17-62 19-37 FcOg 2-23 2-00 1-59 1-94 1-76 i-8i 2-01 2-58 3-42 Feb 2-89 1-82 3-29 2-67 5-67 7-08 8-90 6-07 3-69 MgO 0-62 0-09 0-85 0-52 1-62 2-79 3-63 4-39 4-71 CaO 2-II 1-78 2-34 2-08 ' 4-18 5-89 7-60 9-22 12-35 Na^O 6-30 , 7-03 6-09 6-47 : 6-25 5-89 5-03 4-15 2-50 K2O 2-07 1-75 1-99 1-94 1-20 0-94 0-64 0-75 0-87 H,0+) H,0-) 0-09 0-24 o-i6 0-16 0-56 (0-56 1 0-08 0-42 1 0-04) 0-00 1-79 TiO, 0-58 0-70 I -00 0-76 1-54 1-79 2-29 1-65 1-32 P2O5 0-12 0-07 0-16 0-12 , 0-24 0-21 0-35 0-36 o-ii MnO — — — — o-o8 O-II — S — — — — 0-06 0-06 — 100-23 9971 99-83 99-93 99-86 100-14 99-87 100-29 2-0 99-97 Q 2I-I 18-5 20-1 1 20-2 8-1 1-9 — I-O 2-7 F' 64-9 70-5 62-5 65-3 60-5 55-2 46-2 39-0 26-7 M' 14-0 II-O 17-4 14-5 31-4 42-9 54-8 59-0 70-6 link 89-2 84-1 82-6 84-9 71-7 66-4 56-4 43-1 26-3 k 177 j 14-3 17-7 i6-i 1 1-4 9-5 8-0 lO-I 20-0 I. 2. 3- A. 4- 5- 6. 7- 8. Trachyandesite, Deception Island. E. Gourdon, C.R. Acad. Sci., Pan's, clviii, 1914, p. 1906. Tridymite-santorinite, Deception Island. Barth and Holmsen ({4), p. 14). Trachyandesite, Deception Island. Gourdon, op. cit. Average of nos. i, 2, and 3. Andesite,* Deception Island. Gourdon, op. cit. Bandaite,f pillow-lava. Deception Island. Barth and Holmsen ((4), p. 11). Andesine-basalt, Deception Island. Barth and Holmsen {{4), p. 11). Basalt (' Labradorite' — Gourdon), Deception Island. Gourdon, op. cit. Doleritic basalt, block (in tuff or agglomerate.') Deception Island. Gourdon, op. cit. * The alkalis in this analysis are given as recorded in Gourdon 's first paper of 1914, i.e. NajO, 6-25; K„0, 1-20. In Washing- ton's Tables (U.S.G.S. Prof. Paper 99, 1917, p. 466) the alkalis are given as Na^O, 6-67; K2O, 0-78, and as the summation remains the same it seems clear that 0-42 per cent has been transferred from KjO to Na,0. This may have been a correc- tion of the original analysis when it was transmitted to Washington by Gourdon, but it has been thought best to leave the original figures intact, especially as they are repeated in Gourdon's later work published in Deuxieme Expedition Antarctique Franfaise (1908-1910), commande par le Dr Charcot: Mineralogie, Geologic, Paris, 1917, p. 7. The earlier figures for the alkalis are also more accordant with the serial characters of the Deception Island suite than the later. •f- Correct summation, 100-14, given in the table. Barth and Holmsen give 100-08. SOUTH SHETLAND ISLANDS 59 in a short paper, ' Sur la constitution mineralogique des Shetland du Sud ' {sic),' with only exiguous petrographical notes. Four new analyses are given in the 1939 paper of Barth and Holmsen ((4), pp. II, 14, 25). The remaining analysis is a computation made from a Rosiwal estimate of mineral proportions in a quartz-gabbro from King George Island by the writer ((i), p. 65). The two new analyses made for this work are of a tholeiitic lava type from Fildes Strait (p. 44), and of the Recent olivine-basah lava of the Penguin Island volcano (p. 45). Thus there are now available analyses of eight rocks from Deception Island, five from King George Island, and one from Bridgeman Island. Table I b. Analyses of igneous rocks from King George Island and Bridgeman Island SiO., AlA FcaOj FeO MgO CaO Na^O KaO H,0~) CO., Tido P205" MnO 57-30 17-97 2-17 379 2-57 6-72 3-25 0-96 4-26 0-56 0-20 Q F' M' nak k 99-75 17-3 34-7 48-0 35-8 17-5 54-9 15-6 5-4 7-0 27 9-1 2-9 1-7 0-7 II loo-o 10-6 34-7 54-7 42-5 27-7 53-45 19-37 3-37 4-09 4-42 8-iS 3-55 1-35 1-69 0-66 0-04 12 13 100-17 53-02 15-57 4-40 6-58 3-93 8-15 2-38 1-68 (2-02 jo-so tr. i-i6 0-35 o-i6 99-90 48-26 17-42 3-36 5-6i 8-83 11-56 2-44 0-89 0-24 1 o-i6| nil 1-07 0-22 0-14 14 I00'20 3-8 38-5 57-7 37-9 20-8 10-7 30-9 58-4 37-2 31-6 -5-2 25-0 80-2 28-6 i8-4 54-24 17-20 2-81 4-98 5-84 10-19 2-91 0-92 0-09 0-91 0-09 ioo-i8 5-3 29-3 65-4 33-7 17-5 9- 10. II. 12. 13- 14. Hypersthene-andesite, Admiralty Bay, King George Island. Gourdon, op. cit. Quartz-gabbro, intrusion, Le Poing, Admiralty Bay. Tyrrell ((i), p. 65). 'Dolerite',1 dike, Admiralty Bay. Barth and Holmsen ((4), p 25). Tholeiitic basalt, lava, Fildes Strait, King George Island. New analysis by F. Herdsman. , . , Olivine-basalt, lava of Recent volcano, Penguin Island, King George Bay, King George Island. New analysis by F. Herdsman. Basalt, Bridgeman Island. Gourdon, op. cit. X The description of this rock by Barth and Holmsen makes it tolerably clear that it is a porphyritic hypersthene-augite- andesite, practically identical with the rock of the dike in Admiralty Bay described in the present paper (p. 43). As this is a very conspicuous feature in Admiralty Bay, it is very probable that the two specimens come from the same dike. The fourteen available analyses are set out in Tables i a and i b in the above geographical order. The von Wolff normative parameters as modified by the writer are also given.- In these O represents the excess or defect of molecular silica, a positive number giving the amount of normative quartz, and a negative figure representing the amount of olivine. F is the percentage amount of normative alkali-feldspar (orthoclase and albite), and M' the combined percentage of anorthite, pyroxene, 1 C.R. Acad. Set., Paris, ci.viii, 1914, pp. 1905-7. 2 A full account of this method of calculation will be published as soon as possible. 6o DISCOVERY REPORTS iron ore and apatite. The symbol nak represents the percentage of alkah-feldspar in total feldspar, and k the percentage of potash feldspar in total alkali-feldspar. Thus : salicCNa^O-KaO) nak- 100, X I GO. salic(Na20.K20.CaO) , _ salic K2O salic (Na^OTK^O) The geographical arrangement of the analyses in Tables i a and i b shows at once that there is a considerable difference between the Deception Island series at the southern end of the South Shetland archipelago, and that of King George Island and Bridge- man Island at its northern end. The Deception Island series is characterized throughout (except no. 8) by comparative richness in alkalis as against lime, as shown by the high nak ratios. Moreover, in the alkalis, soda is extraordinarily high in relation to potash, as is shown by the low k ratios. The members of this series show regular chemical variations throughout, again with the exception of no. 8, which stands apart in several particulars. This rock is described by Gourdon as ' doleritic basalt '. It is stated to occur as ' blocks ' (.'' in agglomerate or tuff), and is not found in situ} As its analysis agrees fairly closely with those of the Recent basalts of King George Island (no. 13) and Bridgeman Island (no. 14), it is possible that the rock represents a fragment torn from a foundation of Recent basalts through which the Deception Island volcano, of quite different constitution, has burst. It will be so regarded in the present investigation. The serial relations of the Deception Island series are shown in the variation diagram (Fig. 8). The silica per- centages, and the values for F' and M' , were tried as abscissae against which the other constituents were plotted. It was found that F' gave the smoothest curves. In all cases analysis no. 7 (Gourdon 's ' labradorite ') was some- what discrepant from the others. The curves show the same general trends as for other andesitic series. The distinguishing feature of the diagram, however, is the height of the NagO curve and its distance from the K,0 curve. The Deception Island rocks may thus be regarded as an andesitic series of quite abnormal sodic composition (Barth and Holmsen, (4), p. 13). On the other hand, the King George Island and Bridgeman Island suite, together with the block of doleritic basalt (no. 8) from Deception Island, constitutes a quite normal series of pyroxene-andesites ranging to olivine-basalt, with accompanying plutonic types, and belongs to the great circum-Pacific petrographic region of which the characteristic lava type is hypersthene-augite-andesite. 1 E. Gourdon, ' Sur la constitution mineralogique des Shetland du Sud (lie Deception)'. C.R. Acad. ScL, Paris, CLViii, 1914, pp. 583-6. Fig. 8. SOUTH SHETLAND ISLANDS 6i The Deception Island series. It is difficult to match the rocks of the Deception Island series with those of other andesitic fields. Very occasionally one finds soda-rich andesites as, for example, in the Andean petrographic region, and in that of western North America ; but the more normal andesitic types are overwhelmingly predominant in these regions. As a suite the Deception Island rocks are almost unique. The only other series which approaches them in richness in soda is that of the Santorin volcano in the Aegean Sea, as has already been pointed out by Barth and Holmsen. But even among the Santorin analyses only two are closely comparable to the ' santorinite ' of Deception Island. In Table 2, col. B, the closest Santorin analogue of the Deception Island santorinite (Table 2, col. A) Table 2. Deception Island ' santorinite ' and comparable analyses A B C D E SiOa AUOa Fe.03 FeO MgO CaO Na.O K.,6 H.,0+| HoO-j TiOa P2O5 MnO S CI 68-33 14-94 1-94 2-67 0-52 2-08 6-47 1-94 o-i6 0-76 0-12 64-99 14-32 1-30 4-01 1-12 3-94 6-20 1-99 1 0-05 1 I nil J 2-23 o-oi 0-07 65-9 15-8 1-6 3-4 I-o 3-5 5-1 2-1 0-4 I-O O-I O-I 69-00 14-48 1-25 I-OI 0-36 2-34 6-00 2-76 2-19 0-24 66-05 13-29 3-22 S-°7 1-36 0-50 6-67 0-87 (1-88 (0-96 0-49 0-09 ?tr. ?tr. ?tr. 9993 100-23 lOO-O 99-63 100-45 F' M' )iak k 20-2 65-3 14-5 84-9 i6-i 14-2 63-0 22-8 86-4 17-3 20-6 54-8 24-6 67-1 2I-I 21-8 67-7 10-5 89-4 237 20-8 62-8 16-4 93-5 7-8 A. Average santorinite, Deception Island (Table i a). B. Hyalodacite, east lava flow, August 1925, Fouque Kaimeni, Santorin, Aegean Sea. Quoted from H. S. Washington, 'Santorin Eruption of 1925', Bull. Geol. Soc. Atner. xxxvn, 1926, p. 378. C. ' Santorinite', average of eleven analyses of the Recent lavas of Santorin volcano, Aegean Sea. D. Biotite-andesite, Inca-loma, Cotopaxi, Ecuador. A Young, Hochgeb. Republik Ecuador, 11, 1904, p. 256. Quoted from Washington's Tables [op. cit. supra), p. 154. E. Keratophyre, Trevennen, St Goran, Cornwall. Quoted from Cliem. Anal. Ign. Rocks, etc. Geol. Surv. Gt. Brit. 1931, p. 85. is tabulated. It agrees closely with the Deception Island analysis except for silica, which is 3 per cent lower. The von Wolff parameters also show concordance except for 0. Even Santorin is not a very close analogue for the Deception Island volcano, as is shown by the average of eleven accordant analyses of the lavas of that volcano (Table 2, col. C). The Deception Island rock is distinctly richer in soda and silica, and poorer in potash than that of Santorin. Among Andean andesites the biotite-andesite of Inca-loma, Cotopaxi (Table 2, col. D) provides a close comparison with the santorinite of Deception Island. Further, some rocks of the keratophyre- spilite association are chemically similar to those of the Deception Island series, as is shown by an analysis of a Cornish keratophyre (Table 2, col. E) ; but the k ratio of this rock is notably smaller, and the nak ratio higher, than those of the Deception Island rock (see also Table 4). 62 DISCOVERY REPORTS The intermediate rocks of the Deception Island series are even more difficuh to match. The oUgoclase-andesite (Table 3, col. 4) can be paralleled, and that not very closely, by an andesite from the Sincholagua volcano in Ecuador (Table 3, col. F), and by a trachytic andesite from the Recent lavas of the Modoc Quadrangle, California (Table 3, col. G). The bandaitic pillow-lava of Deception Island (Table 3, col. 5) can be most closely compared with a hypersthene-augite-andesite from Grenada, B.W.I. (Table 3, col. H); and less closely, at least in respect of the nak and k ratios, with an andesitic ash from Cotopaxi, Ecuador (Table 3, col. I). It is to be noted that the Ecuadorian Table 3 . Intermediate lavas of Deception Island and comparable analyses 4 F G 5 H I J SiO, AlA Fe,03 FeO MgO CaO Na,0 K„0 H^O t ! H„'o ) TiO„ P2O5 MnO S CI 60-62 l6-22 1-76 5-67 1-62 4-i8 6-25 1-20 0-56 1-54 0-24 58-82 i6-35 5-50 2-36 4-37 4-06 5-31 2-02 1-05 0-36 0-25 59-98 16-71 2-52 5-04 2-22 4-84 5-12 1-63 0-19 1-30 0-43 o-ii 56-89 16-07 1-81 7-08 2-79 5-89 5-89 0-94 1 0-56! I0-08J 1-79 0-21 0-08 0-06 56-51 14-07 4-04 4-65 3 95 8-44 5-32 0-79 1-51 0-19 0-23 tr. 56-89 19-72 4-06 3-65 I-9I 5-87 5-14 1-96 0-62 tr. tr. tr. tr. 54-53 13-06 6-85 4-86 3-14 9-83 4-62 1-59 0-52 0-96 99-86 100-45 100-09 100-14 99-70 99-82 99-44 Q F' M' nak k 8-1 60-5 31-4 71-7 II-4 7-6 56-5 35-9 65-8 19-8 11-2 52-4 36-4 60-4 17-2 1-9 55-2 42-9 66-4 9-S 3-4 49-5 47-1 68-1 9-6 3-8 56-4 39-8 53-9 20-2 3-9 48-3 47-8 71-1 18-7 4- F. 5- H. I. J- Oligoclase-andesite, Deception Island (Table i, col. 4). Pyroxene-andesite, Ceballos-chupa, Sincholagua volcano, Ecuador. A. Young, op. cit. supra, p. 24S. Quoted from Washington's Tables, op. cit. supra, p. 452. Trachytic andesite (Platy Andesite Group), south of Medicine Lake, Modoc Quadrangle, California. H. A. Powers, 'The Lavas of the Modoc Lava-bed Quadrangle, California', Amer. Min. xvii, 1932, p. 292. Bandaite (hypersthene-augite-andesite), pillow-lava. Deception Island (Table i, col. 5). Augite-hypersthene-andesite, Grenada, B.W.I. J. B. Harrison, Rocks and Soils of Grenada, 1896, p. 10. Quoted from Washington's Tables, op. cit. supra, p. 466. Andesitic ash, Cotopaxi, Ecuador. J. W. Mallet, Proc. Roy. Soc. XLii, 1887, p. 2. Quoted from Washington's Tables, op. cit. supra, p. 764. Augite-hypersthene-andesite, Mt Kouragio, Aegina, Greece. H. S. Washington, 'A Petrographical Sketch of Aegina and Methana, Part III', J. Geol. in, 1895, p. 150. volcanoes have provided two of the comparable analyses in Table 3. It would appear that the andesites of these volcanoes are more sodic than the usual run of Andean andesites. It is interesting to find, also, that an augite-hypersthene-andesite from the Aegean region (Table 3, col. J) has some chemical characters in common with the bandaite of Deception Island. It will be noted that all the Deception Island rocks and the comparable types dealt with in Tables 2 and 3 have been characterized by a ratio F'jM' greater than unity. In the remaining rocks of the Deception Island series, the andesitic basalts (Table 4, cols. 6, 7), however, this ratio is less than unity. The andesine-basalt (Table 4, col. 6) is closely comparable with another Ecuadorian lava, a basalt SOUTH SHETLAND ISLANDS 63 from the Ruminahui volcano (Table 4, col. K). Some spilites as, for example, those of Oregon (Table 4, col. L), are also quite similar. The basalt (' Labradorite ' — Gourdon) of Deception Island (Table 4, col. 7) differs from the andesine-basalt only in its positive O. Comparable analyses are those of a hornblende-soda-andesite-basalt, an inclusion in dacite lava from the San Franciscan volcanic field of Arizona (Table 4, col. M), and a hypersthene-augite-andesite from the Czerhat Mountains of Hungary (Table 4, col. N). These rocks, however, are only isolated examples of the type, for in both the Arizona and Hungarian fields the great majority of the andesites otherwise comparable to the Deception Island rocks have a much higher k ratio. Table 4. Andesitic basalts of Deception Island and comparable analyses 6 K L 7 M N SiOa Al.,03 Fe,03 FeO MgO CaO Na^O KjO H2O+ H2O- CO2 TiO, P2O5 MnO SO3 CI 52-93 15-86 2-01 8-90 3-63 ' 7-60 5-°3 0-64 0-42 i 0-04 j 2-29 0-35 52-92 i6-66 476 4-89 7-96 5-71 5-12 0-89 o-8o 078 53-15 14-39 1-28 9-33 4-74 7-04 4-58 I-OI ( 2-02 1 I0-I9) O-IO 1-50 0-19 0-14 53-50 17-62 2-58 6-07 4-39 9-22 4-15 0-75 0-00 1-65 0-36 53-97 1 6-00 4-56 3-63 6-36 7-47 4-38 1-23 (1-31 1 0-03 nil 1-46 o-io nil tr. 52-80 19-44 3-47 5-15 2-33 8-70 471 1-12 1-26 0-21 1-05 0-24 O-II 99-87 100-49 99-66 100-29 100-50 100-59 Q F' M' nak k -i-o 46-2 54-8 56-4 8-0 -3-6 47-5 56-1 56-1 10-9 -1-5 44-9 56-6 60-3 12-9 2-0 39-0 59-0 43-1 lO-I 2-0 43-8 54-2 53-5 15-^ 0-3 47-5 52-2 46-3 13-6 6. K. 7- M. N. Andesine-basalt, Deception Island (Table i, col. 6). Basalt, Panang Hondon, Ruminahui volcano, Ecuador. A. Young, op. cit. supra, p. 243. Quoted from Washington's Tables, op. cit. supra, p. 538. Spilite, Poorman Mine, Oregon. J. Gilluly, ' Keratophyres of Eastern Oregon and the Spilite Problem', Amer. J. Set. XXIX, 1935, p. 235. Basalt ('Labradorite' — Gourdon), Deception Island (Table i, col. 7). Hornblende-soda-andesite-basalt, inclusion in hornblende-soda-dacite, Bill Williams Mt, San Franciscan Volcanic Field, Arizona. H. H. Robinson, ' The San Franciscan Volcanic Field, Arizona', U.S.G.S. Prof. Paper 76, 1913, p. 147. Hypersthene-augite-andesite, Czerhat Mountains, Hungary. A. Vendl, 'Ober die Pyroxenandesite des Czerhat- gebirges (Ungarn)', Min. u. Petr. Mitt, xlii, 1932, p. 516. The Deception Island series has been treated at some length because, chemically at least, it appears to be almost unique among andesitic series, especially in its richness in soda. As a series, only that of the Aegean volcano Santorin approaches it in chemical character, although sporadic examples of similar rocks occur in andesitic regions of the normal type, and especially among the volcanoes of Ecuador. It is not necessary to deal with the King George Island and Bridgeman Island series in such detail, for it consists of perfectly normal andesites and basalts conforming closely in their minerals and chemistry with the great circum-Pacific granodiorite-andesite region, and other similar regions (western 4-2 64 DISCOVERY REPORTS North America, Hungary, New Zealand, etc.). The hypersthene-andesite of Admirahy Bay (Table i b, col. 9) closely accords, except for lower potash, with an average hypersthene-andesite computed by the author from 114 analyses derived from the circum-Pacific region, including the East and West Indies, and certain European fields (Sardinia, Hungary, Aegean Sea).i The quartz-gabbro (Table 16, col. 10) agrees well with an average of 11 analyses of rocks so called taken from Washington's Tables {op. cit. siipra).^ The tholeiitic basalts of the series (Table i^, cols. 11, 12, 14) are accordant with the average Non-porphyritic Central Magma-type of MuU,^ and with as yet unpublished average analyses of tholeiitic types from the Tertiary igneous region of Scotland. They also accord with the sparsely developed basalts which are found in the great andesitic regions. The above-mentioned rocks are all over-saturated with silica (positive O) ; and in this respect the under-saturated olivine-basalt (0=— 5-2) of the newly discovered Penguin Island volcano (King George Island) stands quite apart from the rest. With M', 80-2, it is also the most basic lava type from the South Shetland Islands so far analysed. Its closest analogue appears to be the olivine-basalt or 'plateau-magma type' of the Tertiary igneous series in Scotland,^ although it is richer in alumina and lime and poorer in the ferromagnesian oxides than that type, and is thus richer in plagioclase feldspar and poorer in olivine. It is precisely in these chemical and mineral characters that the comparatively rare basalts occurring in andesitic regions differ from the olivine-basalts which are the most abundant and characteristic types of oceanic regions and of many mildly and richly alkaline regions on the continents. Thus the olivine-basalt of Penguin Island preserves its relationship with the associated andesites, notwithstanding its superficial similarity to the olivine-basalts of quite different petrographical regions. CONCLUSIONS ON THE GEOLOGY OF THE SOUTH SHETLAND ISLANDS A synopsis of the geology of the Danco Land Coast (Graham Land), the Palmer Archipelago, and the South Shetland Islands was given in my memoir of 192 1 ((i), p. 75). The following are relevant excerpts from that summary : The oldest rocks in the region (excluding a possible basement of crystalline schists and gneisses) appear to be a series of folded bluish slates and mudstones, with subordinate fine-grained sandstones and greywackes, and abundant intercalations of coarse breccias made up principally of igneous fragments.. . .The igneous breccias. . .may possibly be as much due to the rapid denudation of an earlier range of porphyry mountains under arid conditions, as to explosive igneous action.. . . Because of the abundance and size of the plutonic masses the sedimentary series is only visible in small fragmentary exposures on the Danco Land coast. It appears, however, to occur in great force on the islands of the Palmer Archipelago, in which the igneous breccias are also especially prominent. The sedimentary series constitutes a large part of the South Shetland Islands, especially King George Island. Blue mudstones are intercalated with the older andesites around Admiralty Bay, and are intersected and metamorphosed by the intrusion of Noel Hill, in Marian Cove. . . . The presumably Mesozoic mudstones are interbedded with an early series of andesite lavas in King George Island, and possibly also in the other islands of the South Shetland group. The plutonic masses of Noel Hill and Le Poing intersect and cause hornfelsing in both sediments and lavas.. . . The next event in the geological history of the region seems to have been the extrusion of a great series of later andesites, which, in King George Island, are regarded by Mr Ferguson as being banked up against the older series and interbedded mudstones to the north-west. An eruptive focus of this period is probably to be seen in Three Brothers Hill, Potter's Cove, a columnar plug of typical fresh bandaite lava. . . . 1 G. W. Tyrrell, 'The South Sandwich Islands. Report on Rock Specimens', Discovery Reports, iii, 1931, p. 195. 2 G. W. Tyrrell, The Principles of Petrology, 1926, p. 120. 3 'Tertiary and Post-Tertiary Geology of Mull', Mem. Geol. Surv., Scotland, 1924, p. 17. * G. W. Tyrrell, 'The Geology of Arran ', Mem. Geol. Surv., Scotland, 1928, p. 121. SOUTH SHETLAND ISLANDS 65 The latest volcanic episode seems to have been the extrusion of olivine-basalt lavas mainly from a series of volcanoes in the north-west side of Bransfield Strait (Deception Island; Edinburgh Hill, Livingston Island; Bridgeman Island). These volcanoes are largely built of basalt tuffs with subordinate basalt lavas and intrusions. Deception Island, however, contains hyalodacites and oligoclase-trachytes, as well as basalts. Nordenskjold (Antarctis, 1913, p. 11) suggests that these volcanoes may have some relation to the subsidences of the Bransfield Strait region.. . .He regards the Bransfield Strait volcanoes also mainly as of early Quaternary age; but Deception Island, and probably Bridgeman Island, continued erupting until recent times.. . . The main addition we have been able to make to Nordenskjold's account of the region is the recognition of folded sediments in the South Shetland Islands, similar to those of the Palmer Archipelago and the Danco Land coast, but here interbedded with, and covered by, typical Andean lavas. It seems probable that a tectonic zone parallel to those of the Palmer Archipelago and Graham Land runs through the South Shetland Islands. It is worthy of note that the intensity of plutonic action diminishes towards the outer (north-western) part of the region. Plutonic rocks build up the greater part of the mainland ranges; they are also abundant in the Palmer Archipelago, but folded sediments are here also very conspicuous, while in the South Shetlands plutonic masses are small and isolated, and very subordinate in bulk to the sediments and lavas. Conversely the volcanic rocks are very largely confined to the South Shetlands, and are rare in the Palmer Archipelago and the Danco Land coast. The new Discovery II collections described in this memoir make it clear that King George Island, at any rate, and probably all the larger islands, are mainly composed of the older series of andesites, dacites, rhyolites, etc., with their tuffs, volcanic breccias and agglomerates, which are interbedded in places (Admiralty Bay; Marian Cove) with argillaceous and arenaceous sediments, all conjecturally of late Mesozoic age. This series is intersected by a number of tonalite, diorite and gabbro intrusions. Although Ferguson {op. cit. p. 37) has tabulated a thick section of the older andesites, tuffs, agglo- merates and sediments in Admiralty Bay, it seems possible that the importance of the sedimentary intercalations has been exaggerated in previous accounts. Ferguson himself collected only a very few of these sediments, and other collections from many localities in King George Island have not included any. If the sediments had been at all prominent in the field, it seems likely that they would have bulked much more largely in the collections, notwithstanding their inconspicuousness in contrast with the more spectacular igneous rocks. On the other hand, the importance of the plutonic intrusions in the make-up of the South Shetland Islands may have been minimized in previous accounts. The Discovery II collections have brought to light the existence of a large mass of diorite on the eastern coast of King George Island ; and diorite seems to form a part of the previously unknown Snow Island. Diorites are also known to occur in Livingston Island, Greenwich Island, and Nelson Island. These rocks are certainly intrusive into the older series of andesites and sediments, as shown by their contact-metamorphic effects. It may be conjectured that these plutonic masses are the underground equivalents of the later and fresher series of andesite lavas which appear to be unconformably banked up against, and superposed upon, the older andesite series. That a long period of erosion succeeded the extrusion of the older series is shown by the occurrence of large erratics of coarse conglomerate at Martin's Head (p. 49), which contain well-rounded boulders of the older andesite, altered tonalite, and comparatively fresh augite- andesite. Since the last-named contains the blue apatites characteristic of the older series of lavas, it is a reasonable assumption that all the boulders and pebbles belong to the older series. The latest volcanic episode is represented by a series of Quaternary or Recent volcanoes along Bransfield Strait, the craters of which are still well preserved. It is probable that the Deception Island and Bridgeman Island volcanoes have erupted within historical times (Ferguson, op. cit. pp. 36, 45). A very notable addition to our knowledge has been provided by Mr Marr's discovery of the Penguin Island volcano (p. 45). The lavas of Penguin Island and Bridgeman Island are olivine-basalts. Olivine-basalt was also erupted at Deception Island ; but the main products from this volcano were slaggy and glassy andesites of peculiar composition (p. 54). 66 DISCOVERY REPORTS Another noteworthy addition to our knowledge made by recent Discovery II expeditions is the existence of several basaltic volcanoes on the north-western side of the South Shetland Islands. Desolation Island, off the northern coast of Livingston Island, consists of columnar basalts of Recent aspect. On M^Farlane Strait, not very far to the east, is the beautiful columnar basalt plug surmounted by agglomerate of Edinburgh Hill, discovered and figured by Ferguson {op. cit. pi. i, fig. i). Then again at Fort William, Coppermine Cove, on Roberts Island, the islands at the northern end of Fildes Strait, and on the mainland of King George Island along Fildes Strait, fresh columnar olivine- basalts were collected which probably mark the sites of Quaternary or even Recent volcanoes. All these volcanic centres on the north-western side of the South Shetlands have obviously suffered considerable denudation, and are therefore somewhat older than those on the Bransfield Strait side. There can be no doubt but that these occurrences will be augmented in number when the geological survey of the South Shetland Islands is carried out in detail. Finally, it is possible that the South Shetlands rest on a basement of crystalline schists and gneisses, with sedimentary rocks in various stages of cataclastic metamorphism. Boulders and pebbles of these rocks are numerous in shore and glacial accumulations, and among the dredged material from Bransfield Strait (p. 57). Quite possibly some of this material has been derived from exposures on the South Shetland Islands, although it is more probable that the bulk of it has come either from the Graham Land peninsula to the south-east or from the Palmer Archipelago to the south. PART II. PETROGRAPHY OF ROCKS FROM THE GRAHAM LAND PENINSULA AND ADELAIDE ISLAND, WEST ANTARCTICA INTRODUCTION Among the material sent me for description by the Discovery Committee during recent years I found small collections of rocks from Cape Roquemaurel, Wiencke Island, and the Marin Darbel Islands, as well as a large collection of stones dredged a few miles off the west coast of Adelaide Island. Dr N. A. Mackintosh kindly provided me with a copy of the short geological notes he had made on Cape Roquemaurel and Port Lockroy in Wiencke Island. These notes have been incorporated with suitable acknowledgement in the following descriptions. The collections, especially that from Adelaide Island, have proved valuable in extending our knowledge of the geology of West Antarctica, and in providing confirmatory evidence in favour of previously expressed views on the relationships of West Antarctic rocks with those of the southern Andes in Patagonia and Tierra del Fuego. PETROGRAPHY STATION 1490 (20 JANUARY 1935), CAPE ROQUEMAUREL, TRINITY PENINSULA, GRAHAM LAND Cape Roquemaurel is situated on the northern coast of the Trinity Peninsula, the eastern termination of Graham Land, in long. 58° 30' W., lat. 63° 30' S. In his notes on this locality Dr Mackintosh states that: 'The headland consists of several high rocks projecting from the ice-sheet of Trinity Peninsula. On the south-west side of the outermost rock is a good boat harbour with a very small beach. The rocks of the headland are said to be about 600 ft. high, and consist of a pale granite-like rock traversed by conspicuous dikes of fine-grained blackish rock. On the south-west side of the head- land beneath the granite (?) a yellowish brown rock could be seen for several hundred yards just showing itself above the water line. This seemed to be a different kind of rock, though its structure GRAHAM LAND 67 65° 60" Smith I. ^ ^Deception (7 Low I Palmer Archipelagro Anvers 65 s '■ -I Victor Hugo \.o I ,^tf ^^><„. , ...„> 70 - 55° W Fig. 9. Graham Land. 68 DISCOVERY REPORTS and cleavage [jointing?] did not look much different from the crystalline rocks above it.' [This may have been a discoloration of the granite due to intensified weathering between tide-marks. — G.W.T.] Dr Mackintosh collected four specimens from this locality, two from the main rock formation (granite), and two from dikes. He remarks that the granite showed some variation within short distances, especially in the proportions of the darker minerals, and that his two specimens may have a smaller proportion of the dark minerals than is typical of the rock as a whole. The main rock is a true granite consisting of quartz, orthoclase, albite-oligoclase, and a very small amount of biotite largely replaced by pale green chlorite. One of the specimens is verv coarse-grained, the crystals ranging from ^ in. to h in. in greatest diameter. The feldspar is pinkish white and the quartz milky blue in colour. The other specimen is finer in grain and shows a white vein of aplite with a knife-edge contact against the granite. In thin section the feldspars are seen to be thickly dusted with kaolinitic and sericitic alteration products. The orthoclase seems to be almost pure, with only obscure traces of albite lamellation. The albite-oligoclase occasionally shows an approach to the typical chequer-twinning, and is sub- ordinate in amount to the orthoclase. The quartz and feldspars are sometimes intergrown in a coarse and obscure graphic structure, especially in the finer-grained specimen. The only ferromagnesian constituents are a very few flakes of chloritized biotite. The aplite vein consists of a very fine-grained base of quartz and sericitized orthoclase with a saccharoidal texture, which carries small micro- phenocrysts of quartz, orthoclase, and albite. It is quite devoid of coloured constituents. Conspicuous dikes of a blackish rock traverse the granite. One of Dr Mackintosh's specimens is ' probably characteristic of all the black dikes in the headland '. In hand specimen it is a fine-grained dark grey rock with a few large fresh phenocrysts of feldspar and a sprinkling of pyrites. In thin section it consists mainly of a panidiomorphic plexus of andesine feldspar and a pale green hornblende in about equal amounts. In addition, there are a few phenocrysts of labradorite (extinction 30 ), a little quartz, and numerous fine-grained irregular aggregates of a reddish brown biotite which, in many cases, are apparently growing at the expense of the hornblende. As these aggregates are invariably associated with pyrites, they are probably of secondary origin, and connected with the ingress of sulphide solutions into the rock. This rock is identical with some of the lamprophyres described by Rosenbusch as spessartite} The remaining dike specimen was taken from the inner portion of what is probably a composite dike. This dike was of the same blackish tint as the others. It was about 8 ft. thick, and had a central part of greenish colour and a foot in width. This, however, is only a surface coloration. When broken, the fresh rock is of a greyish blue colour and is very dense, with a flow-banding delineated by the alinement of small pink feldspar crystals. In thin section it proved rather hard to interpret owing to its denseness and opacity. It appears to consist mainly of straight-extinguishing feldspar microlites ( .^ oligoclase) arranged in a wavy flow-banding, with somewhat larger feldspars (? orthoclase), and quartz in smaller quantity. The feldspars are all highly sericitized. In this ground-mass material there are embedded micro-phenocrysts of quartz, oligoclase, and a few pseudomorphs in pale green fibrous hornblende of what may have been an earlier amphibole. As some epidote is always associated with the oligoclase, the original crystals were probably of a more calcic composition. On the whole, the rock has the mineral composition of a dacite. Perhaps an earlier generation of petrographers would have called it quartz-porphyrite. 1 Osann-Rosenbusch, Elemente der Gesteinslehre, 4th ed. 1922, p. 333. PORT LOCKROY 69 PORT LOCKROY, WIENCKE ISLAND Port Lockroy is a small harbour on the west coast of Wiencke Island, opening out on to the Neumayer Channel which separates the large Anvers Island from Wiencke Island. Rocks from Wiencke Island and Doumer Island, as well as from the islands in the Neumayer Channel, and on the south and west of Wiencke Island, have been collected by several expeditions. Thus Pelikan^ described quartz-diorite and gabbro, the former cut by diorite-porphyry and diabase dikes. Gourdon^ described quartz-mica-pyroxene-diorite, quartz-diorite, and micro-diorite, with numerous ' labradorite ' (hornblende-andesite) dikes penetrating the quartz-diorite massif. Ferguson wrote: 'Wiencke Island is bounded on the side facing Neumayer Channel by almost vertical walls of sedimentary rocks in- cluding bluish black mudstone ; it is, however, largely formed of gray diorite, which is the only rock present in Doumer Island and the Cairn Islands.'^ From Ferguson's collection the writer described tonalite, igneous breccias, and a siliceous mudstone.^ The most recent work on the petrography of this part of the Palmer Archipelago is that of T. Barth and P. Holmsen.* They described eucrite and anorthosite (with chemical analyses) from an islet near Victor Hugo Island (west of Wiencke Island). The Joubin Islands, also west of Wiencke Island, consist mainly of igneous breccias, and an analysis is given of a prehnitized rock fragment from these breccias. From Port Lockroy, Barth and Holmsen described quartz-diorite and adamellite, with analyses. They remark that the whole region from Port Lockroy westward to the Joubin Islands and Victor Hugo Island is penetrated by ' diabase ' dikes. The general picture of the geology of this region is then that of an ancient basement consisting of sediments and igneous breccias, cut by plutonic intrusions of tonalite and adamellite, the whole being penetrated by numerous dikes, especially ' diabase '. Dr Mackintosh collected two rock specimens from an island in Port Lockroy harbour. Both consist of tonalite identical with that described by me from Ferguson's collection, but the larger specimen shows a sharp contact of tonalite with a dike of fine-grained grey micro-porphyritic rock which is a porphyritic micro-tonalite. In thin section the tonalite shows biotite, hornblende, and magnetite as mafic constituents, with very abundant euhedral plagioclase (andesine, Anjo), all of which are embedded in a coarse ground-mass consisting of interlocking crystals of quartz with subordinate orthoclase. Biotite and hornblende are present in roughly equal amounts. The hornblende is variegated in shades of green, the larger crystals breaking up into aggregates of smaller, diflterently coloured grains. The dike rock shows numerous phenocrysts of andesine with heavy mechanical zoning, and somewhat fewer phenocrysts of a fibrous, pale green hornblende, enclosed in a very fine-grained equigranular ground-mass consisting of quartz, orthoclase, andesine, hornblende, biotite passing into chlorite, and cubes of magnetite. It is a quartz-diorite porphyry or tonalite-porphyry ; or, if it be desirable not to use the ambiguous term ' porphyry ', it may be designated as porphyritic micro- tonalite. 1 A. Pelikan, ' Petrographische Untersuchungen der Gesteinsproben ', Resultats du Voyage de S.Y. 'Belgica', Exped. Antarctique Beige; Geologie, Anvers, 1909. " E. Gourdon, 'Geographic physique, Glaciologie, Petrographie', Exped. Antarctique Frattfaise, 1903-5, Paris, 1908. 3 D. Ferguson, 'Geological Observations in the South Shetlands, the Palmer Archipelago, and Graham Land, Antarctica', Trans. Roy. Sac. Edin. Liii, 1921, p. 49. « G. W. Tyrrell, 'A Contribution to the Petrography of the South Shetland Islands, the Palmer Archipelago, and the Danco Land Coast, Graham Land, Antarctica', ibid. pp. 59, 73, 74. ^ ' Rocks from the Antarctandes and the Southern Antilles ', Scientific Results of the Nonvegian Antarctic Expeditions 1927-28 and 1928-29, instituted and financed by Consul Lars Christensen, No. i8, Norske Vidensk.-Akad., Oslo, 1939, pp. 17-33- 70 DISCOVERY REPORTS THE MARIN DARBEL ISLANDS This group of small islands and rocks lies a few miles south-west of Cape Bellue at about long. 66" 20' W., lat. 66 00' S. In a brief note accompanying the specimens, in which the above location is given, they are wrongly allocated to the Biscoe Islands, which form a long chain of islands north-east of Adelaide Island. The above-given latitude and longitude are those of the Marin Darbel Islands. I have been able to find no previous reference to the geology of these islands. The specimens collected are stated to come from a small uncharted island lying to the south-west of Cape Bellue. This island, like all those in the vicinity, consists of an ice-worn mass of igneous rock. Two large specimens of this rock (norite) were taken ; five others represent dikes penetrating it. The main rock of the island is a coarse plutonic type of a mottled, greenish grey tint, consisting of white feldspars and greenish black ferromagnesian minerals. In thin section the appearance of coarse grain is seen to be illusory, for the rock consists of large areas of fresh labradorite (An55) in small crystals, alternating with larger and more isolated crystals of hypersthene, augite, and magnetite. The hypersthene is mainly fresh and distinctly pleochroic, but some crystals are in process of alteration to a pale green fibrous bastite mineral, and a few to brown biotite, both modes of alteration being accompanied by the disengagement of magnetite. There is also some primary iron ore. The hypersthene is apparently slightly preponderant over the pale diopsidic augite, and the periods of crystallization of the two minerals appear to overlap. Thiis the rock is a norite or more exactly a hyperite, since the hypersthene is accompanied by a notable amount of monoclinic pyroxene. In another specimen the hypersthene has gone over completely to bastite. Three of the dike rocks are dark, greenish grey, aphanitic types in which numerous micro- phenocrysts of serpentinized olivine and feldspar can be made out with the lens. In thin section they turn out to be olivine-basalts with very numerous micro-phenocrvsts of bytownite (Augo) and almost equally numerous olivines which are perfectly euhedral but completely altered to pale green serpentine. The ground-mass is very minutely crystalline, and consists of microlites of plagioclase, augite, and magnetite. Numerous spherical steam cavities are present which are usually filled with fibrous, radiating, pale green delessite. A fourth specimen is much coarser and is highly carbonated. It appears to represent a coarse basalt or dolerite. That part of the Graham Land peninsula and the Palmer Archipelago which lies between lat. 64°-67° S. and long. 62 "-66° W. seems to be rich in gabbroic intrusions and basic dikes. Thus Pelikan {op. cit. siipro) described gabbros and dolerite dikes from Anvers Island, Bob Island (off south coast of Wiencke Island), and Cape Anna (Danco Land). Gourdon, likewise {op. cit. supra) described basalt dikes from Wiencke Island and Doumer Island, diabase dikes from Booth Island (Wandel I.), diabase and gabbro from Petermann Island and Cape Tuxen. From the Andvord Bay region the writer {op. cit. supra) described basalt dikes and an intrusion of fresh olivine-gabbro (Bruce Island). Barth and Holmsen {op. cit. supra) commented on the abundance of basic dikes in the region between Victor Hugo Island and Port Lockroy (i.e. along the line of lat. 65° S.), and described eucrite and anorthosite from Victor Hugo Island. ADELAIDE ISLAND Adelaide Island is a large island off the coast of Graham Land at about lat. 67" S., long. 69° W. Geologically nothing is known of the main island, but the French Expedition of 1903-5 collected rock material from three small islands, Jennv, Leonie, and Webb Islands, off its south-eastern coast. Gourdon {op. cit. supra) described them as consisting of gabbro cut by numerous dikes of basalt, diabase, and andesite, and has given no fewer than ten analyses of these rocks. ADELAIDE ISLAND 7, Among the first set of Discovery II material sent me I found a box of stones dredged from St. 599, off the west coast of Adelaide Island at a depth of 203 m. The exact position of the Station is lat. 67 08' S., long. 69 o6|' W. Forty-six of these stones were examined and thin sections made. They ranged in size from boulders 9 in. in greatest diameter to pebbles less than i in. across. As these dredgings were taken only a few miles oif the western coast of Adelaide Island near the central point of the western coastline, it is likely that many, if not all, were derived from this geologically unknown land. Ten of the stones belong to the granite family, including ordinary granite, granophyre, granodiorite, and tonalite. Eight are quartz-diorites, three dioritic lamprophyres, and one quartz-gabbro. No fewer than fourteen of the specimens are quartz-porphyries or allied rocks, all of which show signs of crushing and brecciation, in extreme cases reducing them to ' porphvroids ' and even to types which might be regarded as metamorphic quartzite. Five of the stones are lavas, including rhyolite, dacite (or dellenite), and andesite. Finally, the collection includes five andesitic breccias similar to those which have been described from other parts of the Antarctandes. One of the two true granites consists of a coarse-grained allotriomorphic mixture of quartz, micro- perthitic orthoclase, and somewhat less abundant albite-oligoclase which is much more heavily dusted with clayey alteration products than the orthoclase. The sparse ferromagnesian constituents are mainly chloritized biotite, and there are a few crystals of fibrous hornblende. The second granite, like the first, is of a pale flesh-pink colour, but is of finer grain and obviously richer in dark constituents. The feldspars consist of micro-perthitic orthoclase and oligoclase (Ab,o) in roughly equal quantity. The oligoclase frequently forms well-shaped crystals which are enclosed in the larger plates of orthoclase. Both feldspars tend to be poikilitically enveloped in a mosaic of large grains of quartz, and both exhibit coarse intergrowths with quartz. The chief ferromagnesian constituent is biotite which is mostly chloritized. With abundant magnetite, sphene, and apatite, the chloritized biotite mainly occurs in small clots or segregations which appear to be of cognate origin. Both these granites are, strictly speaking, adamellites, as plagioclase occurs to the extent of more than one-third of the total feldspar. One of the pebbles is a good granophyre consisting almostentirely of a fine micro-graphic intergrowth between quartz and very turbid orthoclase. This encloses a few larger crystals of rounded and embayed quartz. The original ferromagnesian minerals appear to have been biotite, now chloritized, and a few flakes of muscovite; but a later mineralization has brought in some large aggregates consisting of calcite, radial sheaves of muscovite, and irregular masses of pyrites. Next comes a granitoid rock which bears a considerable resemblance to the second adamellite described above, as it carries the same clots of chloritized biotite, but with epidote and pyrites instead of sphene and magnetite. It differs, however, in its more richly ferromagnesian character, and especially in the relation between the feldspars. In this rock oligoclase occurs in distinctly superior amount to the orthoclase. It is therefore to be classed as granodiorite. Another stone is a porphyritic micro-crystalline variety of this type, and may be called granodiorite-porphyry or porphyritic micro- granodiorite. Five stones belong to the tonalite group. Tonalite, in the author's opinion, is a granitoid rock inter- mediate between granodiorite and quartz-diorite, distinguished by its abundant plagioclase relative to orthoclase while retaining an amount of quartz sufficient to exclude it from the quartz-diorite group. Its ferromagnesian constituents are mainly hornblende and biotite. They are more abundant than in the granites and less abundant than in the quartz-diorites. Each of the five stones assigned to this group conform more or less closely to the above definition. Two of them contain biotite, mostly altered to chlorite and epidote, as their sole ferromagnesian 72 DISCOVERY REPORTS mineral, with magnetite and apatite as accessories. In one of these rocks the biotite is interleaved with narrow lenticles of a colourless mineral of high refraction and birefringence, straight extinction, and good cross-fracture, which is doubtfully identified as sillimanite. The remaining three tonalites have a considerable amount of green hornblende in addition to biotite, and sphene is a rather abundant accessory. One of these rocks, however, has a well-marked granulose structure, and the irregular grey-green plates of hornblende are spotted with rounded inclusions of quartz and feldspars. This is the 'sieve structure' which is often taken as a sign of hybridism. The diorite family is represented by eight rocks of which six are typical quartz-mica-diorites, consisting of plagioclase (oligoclase to andesine), hornblende, and biotite, with a small residuum of quartz and occasionally a little orthoclase. Magnetite and apatite are the most important accessory minerals, and the apatite often occurs in some abundance as comparatively large crystals. Pyrites, epidote, and chlorite occur as secondary minerals, the two last-named being the products of alteration of feldspar and biotite respectively. The six quartz-diorites vary among themselves within narrow limits in the proportions of dark to light minerals, and in the relative amounts of hornblende and biotite. The seventh quartz-diorite is distinguished from the above-described by containing a notable amount of colourless augite, which occurs in small clots or segregations with hornblende, biotite, magnetite, and apatite. It is therefore a quartz-mica-augite-diorite of a type approximating to Stelzner's 'andendiorit' from the Argentinian Andes. The eighth rock assigned to the diorite group is a micro-diorite of very fine grain and uniform, allotriomorphic granulose texture, consisting of andesine and green hornblende in about equal quantity. A small amount of biotite is involved with the horn- blende as well as a notable quantity of apatite and magnetite, and there is also a small residuum of quartz. This rock may be regarded as a mesocratic quartz-micro-diorite which shows affinity to the malchite of Osann.^ Only one of the stones in this collection falls in the gabbro family. It is a medium-grained rock consisting of plagioclase, probably labradorite, but now intensely altered with the production of aggregates of epidote and unidentifiable turbid matter; pale augite, and an almost equal amount of faintly pleochroic hypersthene which is largely altered to chlorite. A little brown hornblende occurs as an alteration product of the augite. Magnetite and apatite constitute the only accessory minerals, together with a small residuum of quartz. This rock may therefore be described as quartz-hypersthene- gabbro or quartz-hyperite. It is probably to be correlated with the quartz-gabbros of the Jenny Island group off the south-eastern coast of Adelaide Island. - The three lamprophyres in the collection all belong to the spessartite group, and consist essentially of green hornblende and andesine with typical panidiomorphic texture. The hornblende is somewhat in excess of the plagioclase. One of the rocks contains numerous phenocrysts and crystal aggregates of hornblende in the lamprophyre ground-mass. Another contains patches of a pale bleached biotite and of pale green chlorite, with a few micro-phenocrysts of feldspar. The third has much chlorite and magnetite, and its hornblende is mostly of the brown variety. All these rocks carry a small residuum of quartz. This group of lamprophyres appears to be abundant in the Graham Land peninsula and the adjacent archipelagos. We now come to the most interesting and important group of stones from Adelaide Island, namely, the acid volcanic rocks, including rhyolite, dacite, and igneous breccias which contain a variety of acid types. The breccias consist mainly of quartz-porphyry fragments which have suffered cataclastic 1 Osann-Rosenbusch, Elemente der Gesteinslehre, 4th ed., 1922, p. 321. 2 E. Gourdon, ' Sur la constitution mineralogique de I'lle Jenny (Antarctica)', C.R. Acad. Sci., Paris, 159, 1914, 369-71. ADELAIDE ISLAND 73 deformation of the same kind as that described by Quensel from the ' porphyry formation ' of Patagonia and Tierra del Fuego.^ Sixteen stones belong to this group. Three specimens appear to belong to the rhyolite-dacite group. One is a dense whitish rock mottled with pale green streaks which exhibit a rough parallelism. In thin section it becomes clear that this is a coarse and even contorted flow-banding of alternating lighter and darker streaks, more obvious when the slide is held up to the light than when it is viewed through the microscope. The rock consists of a quartzo-feldspathic paste of variable but always fine grain, mingled with varying quantities of sericite and a colourless to palest green, almost isotropic mineral of higher refractive index than quartz or Canada balsam. This mineral occurs in reticulated areas with a flaky, fibrous, or vermiculate structure under polarized light. These properties may serve to identify it tentatively as a variety of kaolinite. Sericite and kaolinite are much more abundantly developed in the darker bands, although they are not absent from the lighter streaks. The only other identifiable mineral is some secondary pyrites. The rock is intersected by thin, thread-like, discontinuous veins of secondary quartz. The flow structure may be primary and the rock therefore a rhyolite; but there is the possibility that it is a pseudo-flow structure like that of the quartz-porphyries or porphyroids described later, and due to cataclastic deformation. The facts that some of the larger quartz grains show undulose extinction, and the considerable development of sericite, may perhaps be regarded as in favour of this view. Another rock appears to be the same as that described by Quensel- from Patagonia as 'felsite- porphyry'. This shows small phenocrysts of bipyramidal quartz, orthoclase, and oligoclase, in a largely cryptocrystalline, quartzo-feldspathic ground-mass. There is, however, a large amount of recrystallized quartz forming irregular areas which carry inclusions of ground-mass material, and which impregnate feldspar phenocrysts in their vicinity. Both quartz and feldspar phenocrysts are euhedral, and the latter enclose large, well-developed crystals of epidote and zoisite. The only coloured minerals present are a few areas of leucoxene representing altered ilmenite, and some secondary pyrites. Veins of secondary quartz traverse the rock and cut through some of the feldspar phenocrysts, but appear to merge into the areas of recrystallized quartz in the ground-mass. This rock is a quartz- felsite or quartz-porphyry which differs from those later described in its comparative lack of alteration and in its much smaller proportion of phenocrysts to ground-mass. Its mineral composition roughly corresponds to that of adamellite or granodiorite, and it might therefore, if a lava, be styled dellenite. A third member of this group is obviously a fragmental rock of composition similar to the above except that plagioclase feldspar is much more abundant. It contains numerous angular chips of rhyolitic or dacitic composition in a uniform cryptocrystalline ground-mass of quartzo-feldspathic composition. The rock has been heavily impregnated with secondary pyrites which has stimulated local silicification of the ground-mass. It is best regarded as a dacitic tuff. Next come three rocks interpreted as coarse tuffs or igneous breccias consisting mainly of fragments and fine comminuted debris of the rhyolite and quartz-felsite (dellenite) just described. One of them consists mainly of fragments similar in composition and structure to the above rhyolite, but in general of coarser grain. There are nevertheless rapid variations in grain size across barely visible boundaries between adjacent fragments. In fact it was only possible to identify the rock as a rhyolitic breccia through the occurrence of a few angular fragments of a coarse feldspathic type apparently belonging to the granite-porphyry described later. Some of the coarse-grained material may be due to secondary silicification. The two remaining rocks of this group are clearly igneous breccias consisting mainly of fragments of the dellenite above described. 1 P. Quensel, ' Die Quarz-porphyr- und Porphyroidformation in Siidpatagonien und Feuerland ', Bull. Geol. Inst. Upsala, xn, 1913, pp. 9-40. 2 Op. cit. supra, p. 14, and fig. 10, p. 27. 74 DISCOVERY REPORTS The ten remaining stones of the acid volcanic series consist of coarse quartz-feldspar-porphyries and their tuffs or igneous breccias, in which a progressive series of cataclastic deformations have taken place, resulting in the formation of typical ' porphyroids ' and, finally, a completely mylonized rock which can only be distinguished with difficulty from a metamorphic quartzite. While the majority of the porphyroids and igneous breccias consist of quartz-feldspar-porphyry fragments only, three contain fragments of rhyolite, felsite, and oligoclase-andesite in subordinate amount. The series begins with an almost normal, practically unstressed quartz-feldspar-porphyry or granite-porphyry, containing very abundant phenocrysts of quartz, some a centimetre in length, orthoclase not quite so large, and still smaller crystals of albite-oligoclase, in a fine-grained ground- mass of aplitic type which consists of equidimensional crystals of quartz, orthoclase, and albite- oligoclase. A few small crystals of altered biotite and a little iron ore represent the only ferromagnesian constituents. The phenocrysts collectively make up considerably more than half the volume of the rock. Only the large quartz crystals show the beginnings of stress. They are cracked and somewhat rounded, with narrow zones of granulation along the fissures. Next comes a series of rocks which may be described as igneous breccias consisting of shattered fragments of the above quartz-feldspar-porphyry with, in some cases, a few pieces of rhyolite, quartz- felsite, and oligoclase-andesite. These may, perhaps, be best interpreted as explosion breccias, but they may possibly represent scree material at least in part. All these rocks have been subjected to crushing and shearing stress of varying degrees of severity. The quartz phenocrysts have been shattered and ground-mass material has been forced in between the fragments. Sometimes the fragments have not been so far separated that the outline of the original phenocryst cannot be traced, but in more severe cataclasis the fragments have been dispersed far and wide throughout the ground- mass. Where the stress has not been great the feldspars have retained their crystal forms, but have been more or less completely sericitized. With more severe shearing the feldspars have been broken down and may show more or less rounded fragments enclosed in areas of comminuted and sericitized material. In extreme cases the feldspars are represented merely by elongated areas of sericitized material the margins of which fade out gradually into the ground-mass. The ground-mass itself has been sheared and sericitized in the same way, but owing to its finer grain and its consequent greater mobility under shearing stress, it has been forced to flow round the phenocrysts, producing what Quensel {op. cit. supra) has called secondary flow structure. The rocks are then typical 'porphyroids', with elongated strips of felted sericite flakes winding round the broken phenocrysts. Secondary epidote and chlorite have been produced in some quantity, especially in the breccias that contain andesite fragments. What appears to represent the final stage of cataclastic deformation is reached in a quartzite-like rock which, if seen in isolation away from the associated types, would certainly be regarded as a metamorphic quartzite or quartz-schist. It consists of alternating strips of coarse and fine quartz crystals. Some water-clear plagioclase feldspar is mingled with the quartz of the coarse layers, and a very pale green, almost isotropic chlorite with the fine-grained quartz. The larger quartz crystals interlock with their neighbours along crenulated margins. Chlorite and ilmenite decomposing to leucoxene are somewhat concentrated in restricted areas presumably where fragments of andesite occurred in the original breccia. Patches and veins of clear recrystallized calcite also occur. Not a trace of sericitization is left. Presumably the sericite, together with particles of iron oxide, has been reconstituted into chlorite. This rock is somewhat tentatively identified as the mylonized end-product of extreme cataclastic deformation aflfecting a breccia composed of acid igneous rocks. The connected series of rocks above described is thus regarded as a complex of acid lavas, or lavas and intrusions (quartz-feldspar-porphyry, quartz-felsite, rhyolite, dellenite, dacite, and oligoclase- ADELAIDE ISLAND 75 andesite), with their tuffs and explosion-breccias, which has been subjected to extensive crushing and shearing. This complex appears to be identical with that described by Quensel {op. cit. supra) from Patagonia and Tierra del Fuego. The same or a similar complex of acid igneous rocks has also been noted in at least three localities in the Graham Land peninsula and adjacent islands. Thus, O. Nordenskj6ld,i writing of the loose blocks on the land surface and in the moraines, and of the boulders in the Late Mesozoic and Tertiary conglomerates, found in the northern part of the peninsula, says that they include quartz-porphyries of various types, some showing such a high degree of mechanical metamorphism that they have been transformed into sericite-schists. He remarks the similarity of these rocks to the porphyry formations of Patagonia which he had previously investigated. Again, in 1913, Nordenskjold'^ stated that at Hope Bay, within the eastern ranges of Graham Land, there occurred acid porphyries and porphyry tuffs apparently concordant with the folded and metamorphosed Jurassic sediments of that locality. He further remarked that these rocks are probably the same as those that form part of the South American cordilleras. At Hope Bay, on the western side of Antarctic Sound at the northern tip of Graham Land, J. G. Anderson^* described sediments with Jurassic plants overlain, in Mount Flora, by 200 m. of whitish tuffs derived from acid volcanic rocks. Finally, E. Gourdon* described an erratic from the north of Hovgaard Island as a 'rhyolite with globular quartz', which he regarded as an 'ancient facies' of porphyry. This rock carries porphyritic orthoclase and bipyramidal quartz, and the crystals are associated with sinuous flow lines. The quartz is much corroded and surrounded by aureoles of ground-mass material. The rock, he says, has suffered severe mechanical deformation. It obviously has a close resemblance to the porphyroids of Adelaide Island described above. The last remaining group of rocks from the Adelaide Island collection consists of oligoclase-andesite lavas, and coarse tuffs or breccias consisting mainly of fragments of the same type. Eight stones are assigned to this group. Two are normal lava types, two are slaggy and vitreous variants, and the remaining four are coarse tuffs or breccias. The lavas exhibit numerous very small micro-phenocrysts of fresh oligoclase, usually with well-marked parallel flow-orientation, embedded in a fine-grained ground-mass consisting of microlites of oligoclase and orthoclase, with chlorite representing the original ferromagnesian mineral (probably augite). This is peppered with numerous, irregularly shaped particles of iron ore. t^i a- j Slaggy variants of this lava contain much dark glass and are somewhat haematitized. The tuffs and breccias consist of angular fragments of the above-described lava of varying textures, with an occasional flake of mudstone or shale. Furthermore, volcanic mud has infiltrated into the breccias and acts as a scanty cement. These rocks recall the characteristics of some of the older group of andesite lavas which are so conspicuous in the geological make-up of the South Shetland Islands (Tyrrell, op. at. supra and preceding paper, pp. 43 et seq.). CONCLUSIONS The rocks from Graham Land and adjacent islands described in the foregoing pages strengthen the already abundant evidence that the igneous rocks of the region, down to the latitude of Adelaide Island at least, are identical with those of the Patagonian Andes. Of particular interest is the discovery 1 'Petrographische Untersuchungen aus dem westantarktischen Gebiete', Bull. Geol. Inst. Upsala, vi, 1900, p. 241. 2 'Antarctis', Handbuch der Regionalen Geologie, Bd. viii, Abt. 6, 1913, p. 9- 3 'On the Geology of Graham Land', Bull Geol. Inst. Upsala, vn, 1906, p. 24. or * 'Geographie physique, Glaciologie, Petrographie ', Exped. Antarctique Franf.aise, 1903-5, Pans, 190b, p. 103- 76 DISCOVERY REPORTS of a quartz-porphyry formation which has undergone intense cataclastic deformation in Adelaide Island. This formation, which is of Mesozoic age (older than Upper Cretaceous) in Patagonia, and extends in that country over a belt more than 400 km. in length, is thus shown to continue in Graham Land to a further distance of about 1000 km. The evidence of this rock collection thus strongly reinforces the conclusion the writer came to in an earlier study, namely, that ' the Graham Land eruptives are identical down to the smallest chemical and mineralogical details with Andean types as far as we know them. The chemical and petrological similarities are so great that one can have no hesitation in subscribing to Nordenskjold's view that the Graham Land ranges, and those of the contiguous islands, are the continuations in Antarctica of the Patagonian chains In Nordenskjold's expressive phrase, Graham Land is a mirror-image of the southern end of South America. '^ PART III. PETROGRAPHY OF ROCKS FROM THE ELEPHANT AND CLARENCE GROUP 6130 Minstrel , Bat/ Cornwallis I SIS 'C Lookout- ^i ^^'^ ■c'O'Brien I Clarence I.. The Elephant and Clarence Group of islands, comprising Elephant Island, Cornwallis Island, and Clarence Island, in its northern section, and Gibbs Island, Aspland Island, and O'Brien Island to the south, is usually regarded as a part of the South Shetlands archipelago (see map. Fig. 10). But there is a good case for its separation as an independent group, and for regarding it as on a parity with the South Shetlands and the South Orkneys. There is a wide sea gap between Gibbs Island and King George Island (South Shetlands), much wider than the distances between the in- dividual islands of either group; moreover, the Elephant and Clarence Group is geologically quite different from the South Shetlands with their thick coverings of andesite lavas, which are absent from all the visited islands of the Elephant and Clarence Group. , . , , r 1 T-1 I ,. J Fig. 10. Elephant and Clarence Group. Landmgs on the islands of the Elephant and & f Clarence Group have been few, and consequently the geological data up to date are very scanty. In the following pages the available information is assembled and supplemented by the investigation of new material from Clarence Island and Gibbs Island, collected during expeditions ot the ' Discovery II '. C SiO'N ;e5 SCALE OF MAU SI 30 55 S"! w ELEPHANT ISLAND During the Salvesen expedition of 191 3 the late Mr David Ferguson passed close to Elephant Island, but was unable to land owing to stormy conditions. He made a few observations from the ship, however, and has recorded them as follows i^ 'The rocks at the south-east corner of the island [Cape Lookout?] are light grey to dark, and more or less banded. The grey rocks appear to be stratified 1 G. W. Tyrrell, 'A Contribution to the Petrography of the South Shetland Islands, the Palmer Archipelago, and the Danco Land Coast, Graham Land, Antarctica', Trans. Roy. Soc. Edin. liii, pt. i, 192 1, p. 78. 2 D. Ferguson, 'Geological Observations in the South Shetlands, the Palmer Archipelago, and Graham Land, Antarctica', Trans. Roy. Soc. Edin. Liii, pt. i, 1921, p. 35. ELEPHANT AND CLARENCE GROUP 77 as the bedding is uniform, but some of the darker rocks may be bedded lavas. [Mr Ferguson was in error here as shown by Prof. Tilley's observations on the Quest Expedition collection — see below.] . . . Much of the island appears to be formed of stratified sediments. Along the extreme west coast, and some eight to ten miles out to sea, is a series of sea-worn hummocks, roughly banded, with smooth slopes, which resemble dark-coloured, table-topped lavas.' [Seal Islands?] The first landing by a geologist on Elephant Island was made by J. M. Wordie in 1914 as a member of the party marooned on the island during Sir E. Shackleton's Antarctic Expedition, 1914-17. Although living under very difficult conditions Mr Wordie made rock collections at Cape Valentine, the north-eastern point of the island, and at Cape Wild, 6 miles farther west, which were described by the present writer in a section of Mr Wordie 's account of the geology.^ The rocks of the north-east coast consist of dark grey, indigo blue, bluish green and grey-green phyllites of fine texture and glossy cleavage surfaces. Many of them are profusely veined and permeated with secondary silica. The rocks consist of quartz, feldspar (plagioclase), chlorite of three varieties, calcite, and opaque greyish (sericitic?) and black (carbonaceous) matter. The calcite is always, the quartz' frequently, of secondary origin. These minerals are arranged in thin, elongated, parallel lenses representing a small-scale flaser texture indicative of intense pressure metamorphism. These puzzling rocks are difficult to interpret; some may represent ordinary argillaceous sediments, as Tilley believes from a study of the similar rocks of Minstrel Bay on the west coast (see below), but others may have been fine washes from an andesitic terrain, or even andesitic dusts. These rocks are highly folded and tilted. At Cape Valentine Mr Wordie states that they dip south by east at about 30°. South of Cape Valentine the rocks dip uniformly to the south and show no folding. Between Cape Valentine and Cape Wild the dip is to the north and changes rapidly from verticality to between 30 and 40°. At Cape Wild the dip of foliation is about 60° towards N. 15° W. At the foot of Mt Houlder (south of Cape Wild) the most striking feature of the section is a reduplica- tion of the beds by 'concertina' folding. There are thus indications of folding on both a small and large scale; small-scale folding and foliation were probably contemporaneous, but the large-scale folding was probably due to a later set of movements. The Shackleton-Rowett Quest Expedition (192 1-2) landed parties at Lookout Harbour at the extreme south of Elephant Island and at Minstrel Bay on the west coast. Rock collections made by Mr G. V. Douglas- have been described by Prof. C. E. Tilley .» Tilley describes the rocks from Minstrel Bay as dark grey to leaden grey phyllites, much contorted and penetrated by numerous veins of secondary silica. The constituents are essentially quartz and albitic feldspar, with scales and closely packed films of chlorite and white mica, abundant carbonaceous matter and some granules of epidote. These rocks are regarded as normal sediments, and Tilley thinks there is no reason to believe that volcanic material enters into their composition. These phyllites are correlated with those of the Cape Wild area described by me (above). On G. V. Douglas's map (Tilley p 56) signs indicate that the phyllites strike a little south of east and are vertical. Since these phyllites have been found at Minstrel Bay, and in the area between Cape Wild and Cape Valentine, it may be conjectured that the northern coast and perhaps the northern halt of the island consists of these rocks. 1 J. M. Wordie, 'Shackleton Antarctic Expedition, 1914-1?: Geological Observations in the Weddell Sea Area', Trans. ""T'Grolfgtl Re^uUs of?hrShL7e:n:Rowett (Quest) ExpedU.on (Report of lecture)', Quart. Jourr.. Geol. Soc. .xxix, ''3Vp?t;ogra;hicaTNo;:fon Rocks from Elephant Island, South Shetlands', Quest Expedition Report, British Museum (Natural History), London, 1930, pp. 55-62. ^ 78 DISCOVERY REPORTS On the other hand, the rocks of Lookout Harbour at the extreme south are of markedly different mineral composition and metamorphic grade. According to Tilley they are divisible into three petrographical groups: (a) Garnet-hornblende-albite-schists, (b) Amphibole-bearing marbles, (c) Para-amphibolites. The rocks of these three groups are linked by the general presence of hornblende, and, to a less degree, albite. Their study, aided by chemical analyses, leads to the conclusion that 'they form a graded series of related sediments ranging from limestones to impure types giving the amphibolites and garnet-hornblende-schists rich in albite'. The original sediments were of abnormal composition, inasmuch as abundant albite was present, probably derived from detrital plagioclase. The grade of metamorphism is obviously much higher than that of the northern phyllites. No data are given of the attitude or geological structure of the Cape Lookout series, which may occupy the southern half of Elephant Island. CORNWALLIS ISLAND This is a small island lying in the strait between the much larger Elephant and Clarence Islands. There is no record of a landing, and nothing is known of the geology except a brief note by Mr Ferguson (op. cit. supra, p. 35). He says: 'It was not possible to land, but the steamer got very close in. It [Cornwallis Island] rises sheer out of deep water in splintery crests, and is partly covered with snow. The highest point of the island may be 1000 ft. or more above sea level. The slopes are very steep, often quite vertical, and there is consequently much bare rock. ... It is formed of light-grey schistose rocks, the foliation planes having a direction [of strike] about N. 70-80" E., with a nearly vertical dip.' Cornwallis Island is not far to the east of Cape Valentine on Elephant Island, where Wordie recorded the strike as east by north, i.e. about the same as that of the rocks on Cornwallis Island. Wordie also says that 'the mountains along the coast [of Elephant Island], when of bare rock, have precipitous slopes and serrated crests of the "frayed cardboard edge" type', which agrees well with Ferguson's description of the topography of Cornwallis Island quoted above. It may therefore be taken as probable that Cornwallis Island represents an eastern continuation of the same rocks as those of the northern coast of Elephant Island. CLARENCE ISLAND So far as is known, no geologist had landed on Clarence Island until Prof. O. Holtedahl, in January 1928, managed with some difficulty to get ashore near the northern point (Cape Lloyd) during the Norwegian Antarctic Expedition of 1927-8.^ But Ferguson, during the Salvesen Expedition of 1913, passed close enough to Clarence Island to make a few observations {op. cit. supra, p. 36). He says: ' The north-east coast is a wall-like rampart, 500 ft. or more in height, of very regular and well-bedded rocks, light grey, dark grey, and drab coloured. The west coast shows light grey, finely banded rocks with a nearly vertical dip in places, and a broad band of brownish rock, evidently an intrusion, was seen at one place cutting through them.' This description agrees well with Holtedahl's and with photographic views of the north-western coast of Clarence Island published by Holtedahl {op. cit. pis. xxiii, xxiv). As regards the rocks, Holtedahl collected a number of characteristic specimens from the scree at 1 O. Holtedahl, 'On the Geology and Physiography of Some Antarctic and Sub-Antarctic Islands', Scientific Results of the Norwegian Antarctic Expeditions of 1927-8 and 1928-9, instituted and financed by Consul Lars Christensen, No. 3, Norske Vidensk.-Akad., Oslo, 1929, 172 pp. (Clarence Island, pp. 47-8). ELEPHANT AND CLARENCE GROUP 79 the foot of a precipitous mountain wall rising behind the beach where he landed, and from wave- rounded boulders. He gives the following brief particulars: The rocks are rather highly metamorphic, grey or greenish in colour, with a more or less distinct schistosity, rather fine-grained, most of them, however, showing a crystalline texture well already {sic) to the naked eye. A grey rock is, according to Broch, a fine-grained albite-epidote-biotite-schist, with quartz and hornblende, further muscovite, titanite, apatite. A chemical analysis shows an andesitic composition.' A greenish chlorite-schist has a basaltic composition. A grey rock, with hardly any schistosity and less fine-grained, is by Broch found to be mainly made up of albite, epidote, hornblende, biotite. It probably represents a highly altered basic igneous rock. These greenish or greyish rocks show a fairly distinct bedding that may be seen in pi. xxiii, fig. 3. The dip is there rather var^'ing both as to inclination and direction. The main direction of the strike is probably south-west to north- east, parallel to the north-western coast. Such a strike is at any rate typical of the extreme western part of the island. The strike of the rocks in Clarence Island is thus not very different from that in Elephant and Cornwallis Islands, and it is to be expected that the same or similar rock types will recur in Clarence Island. From the above brief description of the rocks it would appear that they are comparable in mineral composition and metamorphic grade with those described by Tilley from the southern point of Elephant Island. In the preface to his memoir Holtedahl says that his rock specimens had been assigned to O. A. Broch for petrological investigation. Eventually, however, the work was taken over by T. F. W. Barth and P. Holmsen.- In regard to Clarence Island, Barth and Holmsen give very brief descriptions of a 'common schistose greenstone' and a chlorite-schist, of which analyses are given. In their Table of Analyses (p. 60, op. cit.) these rocks are designated respectively as: biotite-epidote-actinolitc-albite-schist, and chlorite-actinolite-clinozoisite-albite-schist. These analyses are discussed later (see Table 6, p. 87). DREDGED STONES FROM SOUTH OF CLARENCE ISLAND Among the Discovery II material submitted to me was a box containing numerous stones dredged on 23 February 1927 at St. 170 at a depth of 342 m. The exact position is long. 61° 25' 30" S., lat. 53° 46' W. On Chart no. 6^ a sounding of 342 m. is shown about 7 miles south-west of Cape Bowles, the southernmost point of Clarence Island, but this sounding is shown on the chart at lat. 54° 15' W., the longitude being the same as that given above. The position of this sounding is about 30 miles east-south-east of the eastern coast of Elephant Island. The question of the provenance of the stones is rather difficult. It depends on the prevalent direction of the marine currents near Clarence Island, both as affecting direct transport of the stones, and as influencing the drift of icebergs which may have carried the stones or some of them from Elephant Island, or even from more southern localities. It will be assumed that the majority of the stones came from Clarence Island, some from Elephant Island, and possibly a very few from the south. PETROGRAPHY The stones range in size from about 3 in. in greatest diameter down to half an inch. They are all covered with a thick growth of calcareous marine organisms. When this is chipped or dissolved off it can be seen that most of the stones consist of fine-grained grey and green schistose rocks, often profusely veined with quartz. Thirty-five of the stones were sectioned for petrographic examination. Four were found to be igneous rocks, three sedimentary, and twenty-eight metamorphic. 1 This is presumably the analysis of a 'schistose rock' quoted on p. 109 of Holtedahl's memoir. 2 'Rocks from the Antarctandes and the Southern Antilles', Scient. Res. of the Norwegian Antarctic E.xpeditions, 1927-28 and 1928-29, No. 18, Norske Vidensk.-Akad., Oslo, 1939, 64 pp. (Clarence Island, pp. 59-60). 3 H. F. P. Herdman, 'Report on Soundings taken during the Discovery Investigations, 1926-32', Discovery Reports, VI, 1932. 6-2 8o DISCOVERY REPORTS IGNEOUS ROCKS Porphyritic micro-diorite {quartz-diorite-porphyry). This is a fine-grained rock consisting of diversely arranged laths of plagioclase (oligoclase-andesine), with subordinate chlorite representing an original ferromagnesian mineral, probably hornblende, irregular grains of titano-magnetite, a little interstitial quartz, and an abundance of thin needles of apatite. The porphyritic constituents are few and consist solely of badly altered plagioclase (probably andesine). This rock resembles the quartz-diorite porphyries which are abundant in the South Shetlands, the Palmer Archipelago and Graham Land. Porphyritic honibleiide-micro-granite [Iioniblende-quartz-porpJiyry). This is an interesting and unusual rock with very numerous euhedral phenocrysts of feldspar, quartz, hornblende, biotite, and ilmenite, with apatite in well-formed crystals as an abundant accessory, embedded in a pale brown, glassy to crypto-cr}stalline ground-mass. The feldspars are much sericitized and consist of orthoclase and oligoclase (AbjAnJ in roughly equal proportions. Quartz occurs as large embayed cr^'stals up to 0-5 cm. in greatest diameter, often with edges and corners rounded by corrosion. The hornblende forms prisms and plates of green to pale yellowish brown pleochroism, and is often partially or completely altered to chlorite of high d.r. The biotite is completely altered to a pale green chlorite of anomalous 'ultra-blue' polarization colour, with the disengagement of magnetite. Ilmenite altering to leucoxene occurs in large scattered crystals. The phenocrysts form more than half the rock. Spherulitic quarts-porphyry. This rock contains a few small embayed phenocrysts of quartz, rather more abundant euhedral phenocrysts of very turbid orthoclase and a few of albite, in a micro- crystalline and spherulitic ground-mass. The spherulites are often perfect; they may be isolated in the ground-mass, but more often they are grouped around the phenocrysts. The only ferromagnesian minerals are a few small areas of chlorite with separated magnetite, and one or two large crystals of titano-magnetite. Rhyolite. This rock consists mainly of a crypto-cr^'stalline but obviously quartzose ground-mass, with numerous parallel streaks of micro-granitic material. The latter consists of quartz and turbid orthoclase intergrown with the production of a rough micrographic structure. A few small pheno- crysts of oligoclase, orthoclase and quartz occur, but the only ferromagnesian constituents are represented by ragged patches of titano-magnetite, and a few flakes of chloritized biotite, which are associated with the streaks of micro-granite. This rock may be regarded as a rhyolite with flow structure. It may represent a lava, or perhaps more probably, a small dike. These acid volcanic or dike rocks may have come from the extreme northern tip of Graham Land, where O. Nordenskjold has described a similar series, mostly tuflFs, at Flora Bay.^ Also, at Hoffnungs Bay,- he found acid porphyries and porphyry tuffs, apparently concordant with folded and meta- morphosed Jurassic sediments. SEDIMENTARY ROCKS Only three of the stones can be regarded as unmetamorphosed sediments. These are all greywackes, one of sand grade, and the other two of silt grade. The coarser greywacke is grey-green in colour and quartzite-like in aspect. In thin section it is seen to consist mainly of ver^^ angular fragments of quartz and feldspars, with a little biotite (altered to chlorite and magnetite), pale pink garnet, and some epidote, sericite, and chlorite developed as secondary minerals. In addition to the mineral fragments there are numerous rock chips, including carbonaceous shale, chert, fine-grained quartzite, sericite-schist, and fragments of the ground-mass of trachytic and felsitic igneous rocks. Most of the quartz shows a marked undulose extinction 1 ' Untersuchungen aus dem westantarktischen Gebiete', Bull. Geol. Inst. Upsala, vi, 1900, p. 239. ^ 'Antarctis', Handbiicli dcr Rcgioiialen Geologic, Bd. viii, Abt. 6, 1913, p. 9. ELEPHANT AND CLARENCE GROUP 8i indicative of strain. The feldspars include orthoclase and albite (always turbid), and clear fresh andesine (Ab^Anao). The rock is traversed by thin veins of secondary quartz, epidote and calcite. The remaining tw^o rocks have the same composition as that above-described, but the grain-size is coarse silty. They contain a greater abundance of biotite, chlorite and garnet, but rock chips are not so much in evidence, probably because of the finer grain. A few crystals of apatite occur in these rocks, and in one of them carbonaceous streaks delineate the bedding planes. The same slide shows a plane of shearing along which coarse sericite and chlorite have been developed. These rocks are probably due to the rapid waste of a terrain of miscellaneous rocks, including acid and intermediate volcanic types, shales, cherts, quartzites, and schists. The abundance of quartz with undulose extinction points to the presence of gneisses, or, more likely, of a quartz-porphyry formation which has undergone extreme mechanical deformation, within the area of erosion. A mylonized porphyry formation of this character covers great areas in Patagonia and Tierra del Fuego, and has also been found in West Antarctica as far to the south as Adelaide Island (see this Memoir, p. 74). Greywackes and greywacke-siltstones of ancient aspect are common in Tierra del Fuego in forma- tions of Late Palaeozoic and Early Mesozoic ages;^ and O. Nordenskjold {op. cit. supra, p. 238) has described non-schistose slates and greywackes underlying fossiliferous sediments of Jurassic age in Hope Bay at the northern end of Graham Land. He also asserts the abundance of porphyries and porphyry tuffs in the same area. It is therefore possible that the above-described stones came from this region ; but, from the identity in composition of the stones, and the fact that they were associated together in the same dredging, it is considered to be at least as likely that they were derived from the nearest land, i.e. Clarence Island. METAMORPHIC ROCKS Twenty-eight, or four-fifths, of the dredged stones belong to metamorphic types. The great majority of these are due to the dynamic metamorphism of sedimentary rocks resembling the Scottish ' faikes ', alternate laminae of carbonaceous shales and quartzose siltstone or sandstone. These rocks have been intricately folded, sheared, crushed, and converted into carbonaceous sericite-phyllites alternating with quartzose phyllite and quartz-sericite-schist. Some of the rocks contained a significant amount of calcareous cement which has been recrystallized as calcite. This mineral is occasionally so abundant that the rocks have to be recognized as calc-sericite-schists. Thin flakes of sericite are profusely developed in both the siliceous and argillaceous laminae. Calcite and chlorite are formed mostly in the coarser quartzose bands. The chlorite, developed from ferromagnesian impurities in the original sediments, is usually a pale green variety with ' ultra-blue ' polarization colours. It is often vermicular and then almost isotropic. Epidote is sparingly developed in the earlier stages of metamorphism, and generally in the slaty laminae. Some of the rocks are minutely folded and puckered, even within the limits of a thin section (Fig. 11), and the thicker laminae of phyllite acquire a strain-slip cleavage parallel to the axes of folds in the coarser quartzose layers. Others are sheared and smashed into small fragments with the production of crush-breccias. These crush-breccias are often rolled out and a kind of flaser structure is developed, consisting of lenticular fragments of the brittle quartzose layers around which the phyllite laminae have been forced to wind. The quartz grains grow during this process and uhimately form a coarse mosaic. Similarly the size and amount of the sericite flakes increase with the degree of internal movement. These rocks develop into well-crystallized quartz-sericite-schists at the climax of the metamorphic reconstitution. 1 E. H. Kranck, 'Geological Investigations in the Cordillera of Tierra del Fuego', Ada Geographica, iv, no. 2, Helsinki, 1932, PP- 231- 82 DISCOVERY REPORTS A few of the rocks, which must originally have been rich in calcareous and argillaceous matter, contain abundant calcite and epidote. The latter mineral is no doubt produced by the well-known reaction between calcareous and argillaceous matter during metamorphism. Quartz-calcite-epidote- schists are thus formed. As the degree of metamorphism increases, epidote becomes the dominant mineral with the dwindling or disappearance of calcite and sericite. The final product of this change is a quartzose epidosite. All of these rocks are intersected by a profusion of secondary quartz veins. A rock which may belong to the above series is a saccharoidal metamorphic quartzite which carries scattered and irregularly bounded patches of coarse sericitic material. This may perhaps be interpreted as representing one of the thicker beds of sandstone that may have contained clay galls. Fig. II. Section of phyllite, showing folding. Another specimen shows many points of resemblance to the above-described series, especially in the abundance of argillaceous material and the presence of calcite, epidote and sericite. It differs, however, in that some of the folia are rich in large, angular fragments of alkali-feldspars, including orthoclase and albite, which are still comparatively fresh. This may perhaps be best interpreted as a sheared rhyolitic tuff, intermingled with normal sedimentary material. The remaining three stones of the metamorphic group are quartz-epidote-amphibole-schists which have probably been derived from basic igneous rocks or their tuffs. One is a quartz-albite-tremolite- epidote-schist ; the other two are calcite-quartz-glaucophane-epidote-schists. The first is a fine-grained, apparently bedded rock with a schistosity coinciding with the bedding planes. It consists mainly of a mixture of minute grains of epidote with microlites of albite, and prisms of colourless to pale green tremolite which have a tendency to lie athwart the planes of schistosity. This material carries large and small folia consisting of quartz and albite, both enclosing innumerable needles and thin plates of tremolite. The albite often forms large, simply twinned, blasto-porphyritic crystals developed in a mosaic of quartz and small albites. The largest and coarsest of these folia has a distinct resemblance to an aplite vein. This rock is somewhat difficult to interpret, ELEPHANT AND CLARENCE GROUP 83 but the conjecture may be hazarded that it is derived from a rock of the spihtic suite, perhaps a tuff. It has a considerable resemblance to the slightly metamorphosed spilitic lavas of North Glen Sannox (Arran).i Of the glaucophane rocks, one is a quartz-albite-epidote-chlorite-glaucophane-greenstone devoid of schistosity ; the other is schistose and carries abundant calcite in addition to the above-mentioned minerals. In both rocks quartz, albite, and calcite, form a coarse, even-grained mosaic, within and between the grains of which the coloured minerals are developed. In the greenstone the latter are interspersed among the colourless minerals, and are non-schistose ; in the schist the coloured minerals occur as streams winding through the colourless matrix, or they form folia alternating with broad bands consisting of quartz, albite, and calcite. The chlorite is of the deep green penninite variety with low birefringence and anomalous ' ultra- blue ' interference colours ; it is associated with colourless to pale green muscovite. The epidote is of the normal yellowish green variety and is associated with much leucoxenic material. Glaucophane is abundant in both rocks. It has a striking pleochroism as follows: X=pale yellowish green, F= violet, Z= azure blue. In the schist it appears to be altering to a greenish blue soda-amphibole devoid of the violet pleochroism, and with a rather high extinction angle (up to 20 ). This may be the 'abnormal glaucophane' rich in a lime molecule, which is mentioned by Winchell.'^ These rocks are probably due to the recrystallization of igneous rocks of the spilitic series under dynamothermal metamorphism. The abundance of quartz and calcite, with a little muscovite, may indicate that the original rocks were tuffaceous and mingled with normal sedimentary material. Very similar rocks are mentioned by Harker as forming the prosinite type of the Alps.'' Kranck* has described a glaucophane-garnet-schist from Bahia Pliischow in Tierra del Fuego. Its mineral composition is: garnet, glaucophane, quartz, sericite, biotite, chlorite, calcite, apatite, magnetite. This rock is interbedded with garnetiferous quartz-schists and belongs to the Yahgan or Mt Buckland formation. Kranck regards it as due to the metamorphism of a carbonate-rich sandstone [greywacke .?] . THE GIBBS ISLAND GROUP This is a group of three small islands, O'Brien Island, Aspland Island and Gibbs Island (with Narrow Island joined to it), lying about 20 miles south-south-west of Cape Lookout on Elephant Island. Practically nothing was known of the geology of these islands until 1937 when a landing was made on Gibbs Island by a party from the ' Discovery II '. D. Ferguson, however {op. cit. supra, p. 35), was caught in a terrific gale and had to shelter for some time under the lee of Gibbs Island. He says: ' The steamer was sufficiently near to show that the rocks were mainly stratified sediments. The rocks on the west [south?] side of Gibbs Island are dark grey and banded, and dip about 40° W. A higher horizon is represented by some uniformly and well bedded greyish-white rocks which dip about iS"" W. They extend for about { mile, and look soft and friable in places. Aspland Island, 5 or 6 miles west of Gibbs Island, is evidently formed of the same regularly bedded rocks, but they dip east.' A landing on Gibbs Island and Narrow Island was made by J. W. S. Marr on 2 November 1937, and the following facts concerning the geology of the island have been culled from his report (un- published MS.). 1 G. W. Tyrrell, 'The Geology of Arran', Mem. Geol. Siirv., .Scotland, 1928, p. 26. 2 A. N. Winchell, Elements of Optical Mineralogy, Part II, 3rd ed., 1933, p. 259. ^ A. Harker, Metamorphism, 1932, p. 291. * E. H. Kranck, op. cit. supra, pp. 52-4. 84 DISCOVERY REPORTS Gibbs Island is high and steep, rising abruptly out of the sea which is deep close inshore. The coast almost wholly consists of sheer and inaccessible cliffs reaching a maximum elevation of about I GOO ft. These rock walls are remarkably ice-free, and only a thin mantle of highland ice crowns the rising ground above them. Gibbs Island is joined to Narrow Island by a low shingle and boulder spit, 50-80 yards long, which is probably awash at high tide. In its general features Narrow Island is similar to Gibbs Island. The south coast of Gibbs Island is largely composed of a fine-grained schistose rock penetrated by occasional quartz veins. The planes of schistosity are conspicuous from the sea and dip south-west at an angle of about 30°. Specimens of the rock were obtained from an outcrop near sea level on the south coast near the landing place and from another outcrop about i ^o ft. higher. The steep screes which descend to the sea are almost exclusively composed of slabs of the grey phyllite. Above the screes, starting at 500 ft., is a vertical rock face reaching a height not far short of 1000 ft. As this cliff has obviously provided the scree material it is undoubtedly composed of the same phyllite. About 100 ft. above the landing beach [in another direction?] is an outcrop of a massive, dark olive-green rock [serpentine] which has given rise to boulders on the shore. GIBBS I. W.N.W. E.S.E. NARROW I. SCHIST DUN \TE;- SERPENTINE Fig. 12. Narrow Island, on its south side, appears from the sea to be composed of a massive rock of reddish brown hue, with no sign of the schistosity which characterizes the southern face of Gibbs Island. A landing was made on the south coast near the connecting spit, and a specimen was obtained from the cliff face a few feet above sea level. This rock is the dunite-serpentine described below. From the data given above a tentative sketch section may be drawn showing the probable geological structure of Gibbs Island (Fig. 12). The view is here taken that the serpentine has been intruded parallel to the foliation planes of the schist. PETROGRAPHY The rocks of Gibbs and Narrow Islands comprise two sharply contrasted types, namely, schists and serpentine. Schists. Five of the specimens were sliced for microscopic examination. They can be described in general terms as chlorite-sericite-albite-schists containing, in addition, quartz, calcite, and minerals of the epidote group (clinozoisite, zoisite) in some abundance. Small garnets and a mineral of the chloritoid group are found in one specimen, and the latter mineral also occurs in another rock. In hand specimens the rocks show a fine, parallel schistosity yielding flat cleavage surfaces varying in colour from light silvery grey to lead grey. In thin section the rock containing garnet and chloritoid shows a thin foliation with somewhat larger grains of quartz and feldspar taking part in a minute flaser structure. The garnets are small and sparsely distributed; chloritoid is rather more abundant, and occurs as pleochroic grey-blue prisms with good cross fracture. Another rock consists of a mosaic of small grains of quartz through which wind thin folia of ELEPHANT AND CLARENCE GROUP 85 interwoven flakes of sericite, and folia made up of large crystals of green pleochroic chlorite with ' ultra-blue ' polarization colours. In some of the intervening folia of quartz are remarkable ' trails ' consisting of small euhedral crystals of zoisite, strung out as a line of separate crystals, or occurring in small clots. Both the slide and hand specimen of this rock show that it has been permeated by vein quartz which has separated and isolated the individual folia. A third type is rich in epidote. It shows alternating folia consisting (i) largely of quartz with subordinate albite and calcite, but carrying films or thin folia of chlorite and epidote, and scattered crystals of the same two minerals, and (2) mainly of chlorite flakes interwoven with epidote grains. Sericite may form a notable constituent of these folia, but quartz only occurs as scattered fragments. The most feldspathic type is a comparatively coarse schist consisting of more or less rounded grains of albite, intermingled with smaller grains of quartz and patches of calcite, forming a mosaic through which wind streams of flakes of chlorite and sericite, together with grains of epidote and zoisite, and interwoven folia of these minerals. The albite is fresh and water-clear and is mostly untwinned, but a few crystals show simple twinning or the more usual albite twinning. Many of the albites contain curving lines of inclusions of the above minerals, suggesting their growth by accretion during shearing as in the well-known case of ' snowball ' garnets. This rock closely resembles the albite schists of the south-western Highlands of Scotland.^ As a whole the series of schists from Gibbs Island closely resembles those of Elephant Island and Clarence Island, especially those of Minstrel Bay, but they are coarser, somewhat more highly metamorphosed, and do not possess the abundant carbonaceous matter of those rocks. Dtmite-serpentine. The least altered rock and the only one that contains unaltered olivine, is the specimen which was collected from the south coast of Narrow Island. All of the serpentine rocks collected show signs of intense shearing. They are, in fact, serpentine-schists of apple-green and malachite-green colours and ornamental appearance. Some of the specimens show opaque patches, streaks and veins of a black metallic mineral which turns out to be magnetite. The Narrow Island rock must have consisted almost entirely of olivine crystals, but it is now made up of olivine fragments in a mesh of serpentine. The only other mineral is magnetite, a little of which may be primary but, for the main part, is undoubtedly of secondary origin. The olivine is a highly magnesian chrysolite with 21=90" and positive sign, and therefore with a FeO content of about 13 per cent. About half of it has been transformed to serpentine or allied substances. The alteration proceeds as usual along the fissures and from the peripheries of the crystals. The first effect of alteration is to produce a pale brownish yellow uncleaved mineral which is of very low birefringence or sensibly isotropic (delessite?), shot through with colourless fibres of positive elongation which may be chrysotile. These areas of delessite(?) and chrysotile roughly outline the original hexagonal forms of the olivine crystals, and enmesh fragments of them. The next stage of alteration produces colourless antigorite in irregular sheaves of platy crystals with negative elongation, which can be seen to be growing at the expense of the areas of delessite(?) and chrysotile, with the liberation of iron oxides in the form of ragged grains of magnetite. In the remaining specimens of serpentine, all from the south coast of Gibbs Island, the alteration is completed. Not a trace of olivine is left, nor of delessite (?) and chrysotile. The whole rock consists of antigorite in closely woven felts of plates and prisms, with irregular ragged strings of magnetite which have sometimes segregated into definite secondary veins about i mm. thick. The shearing to which the rocks have been subjected has caused the reformation of the antigorite along the major lines of movement, often with a superposed cross-lamellation. With a more severe crushing stress, 1 A. Harker, Metamorphism, 1932, p. 213. The rock figured on this page (fig. 95 A) strongly recalls the microscopic appearance of the Gibbs Island rock. 86 DISCOVERY REPORTS however, the crystals have been ground to powder, and wind in streaks around larger fragments which have assumed a pseudo-spherulitic form. The dunite-serpentine of Narrow Island has been analysed by F. Herdsman, A.R.S.M., with the results shown in Table 5, col. i. For comparison an analysis of dunite-serpentine from Cornwall is given. The resemblance between the two analyses is obviously very close. The calculated norms of both rocks give about 50 per cent olivine and 40 per cent enstatite. While the Cornish rock is stated to contain some enstatite and tremolite {op. cit. p. 64) not a trace of these minerals can be found in the dunite-serpentine of Narrow Island. It may perhaps be surmised that in the alteration to serpentine there has been some differential abstraction of magnesia and iron oxide relative to silica. This appears to be the first record of dunite and serpentine in the West Antarctic region. Table 5 I A SiO, 41-85 40-12 AUOj 1-37 0-98 FePa 2-62 6-52 I. Dunite-serpentine, Narrow Island, West Ant- FeO 2-l6 I-2I arctica. Anal. F. Herdsman. MgO 39-44 35-78 CaO tr. 0-12 A. Dunite-serpentine, Predannack, The Lizard, Na,0 tr. 0-24 Cornwall. Anal. E. G. Radley. Quoted from K,0 0-13 0-08 J. S. Flett and J. B. Hill, ' The Geology of the H,0+ 11-03 12-17 Lizard and Meneage, Mem. Geo!. Siirv., HoO- 0-45 1-69 England and Wales, Expl. of Sh. 359, 1912, c6. nil 0-15 p. 79. Tid„ tr. tr. p.o; 0-22 o-io MnO tr. 0-52 (Ni, Co)0 0-24 0-15 CuO, 0-19 0-28 V2O3 — tr. BaO — nil FeSa — o-oi 99-70 100-12 CHEMICAL COMPOSITION AND ORIGIN OF THE METAMORPHIC ROCKS OF THE ELEPHANT AND CLARENCE GROUP No new analyses have been made of the rocks described above, since none of them has been collected in situ or located with exactitude except a few from Gibbs Island. Four analyses, however, have been published, two each from Elephant and Clarence Islands, and these are collected in Table 6, together with a few comparable analyses from Tierra del Fuego, South Georgia, etc. Prof. Tilley regards the rocks of Lookout Harbour, Elephant Island, as a ' graded series of related sediments ranging from limestones to impure types giving the amphibolites and garnet-hornblende- schists rich in albite '. The amphibolites are closely associated, and even interbedded, with limestone bands. Tilley surmises that the original sediments were somewhat abnormal inasmuch as abundant albite was present. But there is one type of sediment, quite abundant and by no means abnormal, which is often rich in soda and often rich in albite, namely, the impure sandstones known as greywacke. The most typical greywackes are constituents of ancient fold-mountain ranges wherein they are often associated with mudstones, slates, greenstones, ophiolites, and especially with igneous rocks ELEPHANT AND CLARENCE GROUP 87 Table 6 I A B 2 c ° 1 3 4 E SiOa M"M 48-63 51-56 57-66 53-56 53-75 45-10 71-80 11-87 73-°4 ALO, 16-46 14-85 17-54 16-30 19-32 18-60 14-76 10-17 0-56 FejOg 1-92 1-91 1-80 3-46 1-06 2-04 4-5° 2-21 FeO 7-41 9-47 8-28 2-46 7-44 6-97 9-87 2-30 4-15 MgO CaO 8-64 7-93 5-23 3-95 3-43 2-30 5-95 1-94 1-43 10-19 7-20 11-42 6-01 5-21 6-98 11-59 3-02 1-49 3-56 Na^O 274 2-98 2-18 4-39 3-86 4-06 2-55 3-27 K2O 0-06 0-30 0-33 2-68 1-96 1-32 0-47 1-02 1-37 H,0+ 3-38 4-09 0-34 0-98 2-29 0-76 0-26 I-29I 0-48! 2-36 H2O- o-io 0-21 0-22 o-io 0-06 0-07 O-IO 0-84 CO2 0-21 o-i8 nil 0-12 0-20 0-49 1-38 nil TiO, 1-20 2-34 0-56 0-85 1-02 2-83 2-51 tr. 0-16 0-15 P,0= 0-14 0-15 o-oi tr. 0-55 0-22 tr. 0-21 0-23 0-18 MnO 0-12 0-36 o-ii 0-12 0-18 0-26 0-45 (Ni, Co)0 BaO s 0-02 0-21 — o-o8 0-19 o-o6 0-07 — tr. 0-23 nil 0-08 O-IO SO3 — — — — — — 0-20 "" CI nil — — 0-02 tr. F nil — — nil tr. — — — 0-17 C ~ 99-99 100-43 99-82 99-91 99-88 100-35 99-94 99-89 99-80 I. B. D. 3- 4- E. Chlorite-actinolite-clinozoisite-albite-schist ('very schistose'), Clarence Island. Anal. E. Kluver Quoted fram Barth and Holmsen, op. at. supra, p. 60. This rock is briefly described as containing chlorite actinolitic hornblende, clino- TotiieZd aibite (An,,) Calcite and quartz occurred in fissures. It is stated that the latter minerals were removed before the analysis was made (Barth and Holmsen, p. 59). ,,ttii r^ . a tv„„, F H Kranrk Ophiolitic greenstone, north of Monte Olivia, Ushuaia, Tierra del Fuego^ Anal. L. Lokka. ^^^^^1 °f E" f^^'^'l; Ob cit siJm p III. This rock is stated to be an 'effusive' associated with slates and phyllites of the Yahgan (or Mt BuckIand)'^Formation (probably Lower Mesozoic). It is sheared and mylonized m places. The freshest material 2wfoligoclase (An,,) and augite altering to hornblende. Chlorite, epidote, actmohte, sphene altering to leucoxene quartz, and aibite, occur in the highly sheared varieties. In its geological associations and petrography this rock is T'rlmohL^rr— ne^slones dredged near the Shag Rocks, about 130 miles west of South Georgia. New analysis So';^SSr:;t^iSlSSXr('not very schistose'), Clarence Island. Anal. E. KlU^^ Q^-^;;^- ;S and Holmsen oP cit supra p. 60. This rock is stated to be a ' common schistose greenstone, the constituent minerals Jrwhich are IreenbSit, '^actinolitic hornblende, ferriferous epidote, and aibite (An«). [From the analysis it is toleTably certain that quartz should be added to this list.] The summation of this analysis is incorrectly given as 99-82 in R-irth and Holmsen but is correctly stated in Holtedahl, op. cit. supra, p. 109. Sheared tuff from r;;ra^e, Virik Harbour, South Georgia. Anal. E. Kluver. Quoted from Barth and Holmsen, IplTsu^ra pTo- These uffs contain fragments of keratophyres, trachyandesites and spilites (see G^W. Tyrrell 'Petrtraphv and Geology of South Georgia, 'Quest' Report {Brit. Mus. Nat. //»/.), 1930, PP- 35-7)- This aialys^^ - ^' ^""'','?• P\ T^;t Lnf the Yahgan fMt Buckland) Formation. It shows films of chlorite and mica winding Grevwacke fKulm) Steinbach, Frankenwald, Germany. Quoted from R. Ligenteld, uie rvuimcu g SuSmtz in. Fnt!;kenwalde,':4M. Math.-Phys. Kl. Sachs. Akad. W,ss. xlu, no. i, i933, P- 58- 88 DISCOVERY REPORTS of the spilitic suite. These geosynclinal greywackes are rich in fragments of intermediate, basic and ultrabasic igneous rocks and their minerals, especially spilites and their associates.^ Spilitic lavas are of submarine or at least subaqueous origin. The greywackes formed of their debris may be regarded as due to disintegration by submarine eruptions aided to some extent by subaqueous gliding (Bailey),^ which distribute an enormous amount of ' greenstone ' debris, mingled with sand and mud, far and wide over the oceanic regions affected. In its descent through the water this material would become sorted with regard to grain size and would form graded sediments ranging from greywacke to mudstone. This view would explain the frequent passage of greywackes to siltstones and mudstones on the one hand, and into tuffs on the other. Furthermore, limy material lying on the sea floor, and also the radiolarian cherts and impure limestones which are often associated with spilitic lavas, would be incorporated in these sediments. Moreover, spilitic lavas and their tuffs are very frequently saturated with carbonate of lime, which would reappear as calcite in the greywackes resulting from their disintegration. Towards the deeper parts of the oceans these sediments would merge gradually into the blue carbonaceous and ferruginous muds appropriate to this locus ; and towards the coasts they would pass into the terrigenous sands and muds of the continental shelves. The greenstone-greywacke-mudstone association is generally formed during the geosynclinal stage of the orogenic cycle, and is therefore commonly affected by the low-grade metamorphism which ensues when the later orogenic movements take place. Slates, phyllites, and quartz-sericite-schists are thus formed from the mudstones and siltstones; fine-grained quartzites and qviartz-schists from cherts and other siliceous rocks; schistose grits, quartz-chlorite-albite-schists, and greenstones such as those found in the ' Green Beds ' of the Scottish Highlands, from the greywackes and greywacke- tuffs; epidiorites, greenstones, chlorite-schists, hornblende-schists, amphibolites, etc., with epidote, zoisite, garnet, and other accessory minerals, from the basic igneous rocks and their tuffs. Glauco- phane-bearing schists may be formed from the soda-rich varieties of these rocks, or from greywackes composed of their debris. It is precisely an assemblage of this character which is encountered in the Elephant and Clarence Group and the South Orkneys. South Georgia, too, is composed of greywackes and greywacke-tuffs with slates and phyllites, and an occurrence of spilitic rocks is found at the eastern end of the island. Such an assemblage may also form the basement of Graham Land and the adjacent archipelagos. Above all, it is represented in Tierra del Fuego by the rocks of the Yahgan or Mt. Buckland formations, and by some of the Central Schists of that region. Since radiolarian cherts are abundantly developed here, it is probable that the whole assemblage belongs to the geosynclinal greenstone-greywacke-mudstone association discussed above. It is difficult to read Kranck's descriptions of the petrography of these rocks {op. cit. supra) and not to recognize that in West Antarctica we are dealing with exactly similar groups of sedimentary and metamorphic rocks. The bearing of these considerations in favour of the theory of the tectonic connexion between South America and West Antarctica put forward by H. Arctowski, O. Nordenskjold, and E. Suess, is obvious.^ 1 There are, of course, types of greywacke due to the waste of areas of miscellaneous rocks, including slates, basic igneous rocks, etc. These may be styled continental greywackes, and are strictly equivalent to arkoses, which are derived from the waste of a granitic or gneissic terrain. ^ G. W. Tyrrell, 'Greenstones and Greywackes', C.R. Reunion Internat. pour I'etude du Precambrien et des vieilles chaines, Finland, 193 1, pp. 24-6. E. B. Bailey, 'New Light on Sedimentation and Tectonics', Geol. Mag. lxvii, 1930, pp. 77-92. The writer does not accept Bailey's view that greywackes are merely 'muddy sandstones'. ^ For recent discussions of this problem see G. W. Tyrrell, 'Petrography and Geology of South Georgia', 'Quest' Ex- pedition Report {Brit. Mus. Nat. Hist.), 1930, pp. 51-4; and H. F. P. Herdman, 'Report on Soundings taken during the Discovery Investigations, 1926-32', Discovery Reports, vi, 1932, pp. 214-19. 89 PART IV. PETROGRAPHY OF STONES DREDGED FROM THE VICINITY OF THE SHAG ROCKS INTRODUCTION One of the most remarkable geological features of the West Antarctic region is the existence of an eastwardly-directed loop of submarine ridges and islands which connects Staten Island in Tierra del Fuego, through the Burdwood Bank, Shag Rocks, South Georgia, Gierke Rocks, South Sandwich Islands, the South Orkneys, and the Elephant and Glarence Group, with the Graham Land peninsula and its adjacent archipelagos. It represents an extension of Circum-Pacific orogenic structures for more than I GOG miles into the heart of the alien geological region of the South Atlantic. This loop or arc has been called the Southern Antilles on the basis of a supposed analogy with the Antilles con- necting North and South America; but a better term is the Scotia Arc, coined by J. M. Wordie, since the loop surrounds the Scotia Sea. The geological constitution of the Scotia Arc is con- sistent with the view, put forward by E. Suess and others, that it represents an orogenic tectonic connexion between South America and Graham Land.i Something is known of the geology and petro- graphy of all the connecting links of the Scotia Arc with the exception of the Shag Rocks. It is fortunate therefore, that two Discovery dredgings have been made in the vicinity of the Shag Rocks *'=. .Shag Rocks Fig- 13- (see map. Fig. 13), which have provided sufficient material to enable us to assess the geological character of the Scotia Arc in this hitherto unknown region. These dredgings were made on 12 November 1930 by the 'Discovery 11' at Sts. 474 and 475. The exact positions and depths are as follows: St. 474. One mile west of the Shag Rocks. Depth 199 m. St. 475. Long. 53° 30^ S., lat. 42° 44* W. (about 25 miles west of the Shag Rocks). Depth 748 m. PETROGRAPHY Fourteen stones came from St. 474 and five from St. 475. Of these nineteen stones, fifteen are practically identical and consist of tremolite-epidote-greenstone or greenstone-schist, one is a feldspathic quartzite, and three are quartz-vein rocks. The four last-named stones all came from St. 474, nearest to the Shag Rocks. The overwhelming preponderance of the greenstones in this collection makes it tolerably certain that this rock constitutes the Shag Rocks themselves and the submarine ridge on which they stand to at least 25 miles to the west. The stones range in size from 4 in. to i in. in greatest diameter. Fifteen of them, as above stated, are ' greenstones '—dense, compact rocks of grey-green colour, showing an ill-developed cleavage along which they tend to split. Only two are definitely slaty or phyllitic in aspect. The quartzite is a fine-grained rock of a pale buff tint, and obviously contains much feldspar. The quartz-vein rocks are white and coarse-grained. 1 Recent summaries of the evidence have been given by O. Holtedahl, 'On the Geology and Physiography of Some Antarctic and Sub-Antarctic Islands', Scientific Results of the Norwegian Antarctic Expeditions 1927-8 and 1928-9, instituted and financed by Consul Lars Christensen, No. 3, Norske Vidensk.-Akad., Oslo, 1929, pp. 104-18. G. W. Tyrrell, 'Petrography and Geology of South Georgia', 'Quest' Exped. Report {Brit. Mus. Nat. Hist.), 1930, PP- 5i-4- H. F. P. Herdman, Report on Soundings taken during the Discovery Investigations, 1926-32', Discovery Reports, vi, 1932, pp. 214-19. 90 DISCOVERY REPORTS Tremolite-epidote-gree7istone. The principal minerals, as disclosed by thin sections, are tremolite, clinozoisite-epidote, chlorite, quartz, and albite. They are arranged in bands or elongated folia parallel to an ill-defined slaty cleavage which, in two or three of the sections, develops into a phyllitic or schistose structure. The cleavage planes are frilled and puckered by an imperfect strain-slip. The bands consist of one or two of the above minerals to the almost complete exclusion of the others. Folia consisting mainly of tremolite and clinozoisite or epidote are preponderant. The tremolite occurs as colourless to pale green fibres, needles, prisms and plates, often arranged in parallel position or with a slightly divergent, sheaf-like structure. It has a good cross-fracture and longitudinal cleavage, although the typical prismatic amphibole cleavage is rarely seen. The extinction is at io~20° to the cleavage direction [c). Its elongation is positive in sign, distinguishing it from the colourless variety of pargasite (edenite). Both epidote and clinozoisite are present. Epidote is the most frequent associate of the tremolite. It is of yellowish brown colour, and has usually undergone considerable alteration converting it into a greyish cloudy material (leucoxene.?). This material forms ragged areas or, in the more highly cleaved types, it is drawn out into streaks and lines. It is possible that some of this material may represent altered sphene. Colourless clinozoisite occurs mainly as well- shaped crystals associated with quartz and albite in lenticles which may be partly of secondary origin. Chlorite of the pale green variety with ultra-blue polarization occurs in irregular pods or stout lenticles. It is not abundant and, in a few places, appears to be growing at the expense of tremolite and epidote. Quartz, always with undulose extinction, is abundant in some lenticles and bands, and is associated with a little untwinned or simply twinned albite. Finally, in a few of the less altered rocks, very slender microlites of plagioclase (oligoclase.'') can be detected. Discussion of the original character of this somewhat unusual greenstone is deferred to the section dealing with its chemical composition (p. 91). Qiiartzite. This is a hard, yellowish, well-cemented sandstone or semi-quartzite. In thin section it is found to consist mainly of quartz and feldspars (plus alteration products) in roughly equal proportions. All the grains are angular and fit together like the stones in macadam. Only a few of the quartz grains show undulose extinction. The feldspar is easily distinguished by its turbid appearance. It includes soda-orthoclase and albite in about equal proportions. Many of the grains are com- paratively fresh despite their turbidity, but others are completely altered to sericite and crystalline kaolinite. These alteration products have insinuated themselves into fissures in the quartz grains and between the grains, thus acting as a cement which has filled all open spaces. In addition to quartz and feldspar there are a few small grains of epidote, sphene, and iron ores, and rather more abundant fragments of what appears to be the ground-mass of dense acid igneous rocks like felsite or rhyolite. In fact, the mineral composition of the rock suggests that it may have been derived from the waste of rocks like the quartz-feldspar-porphyries which constitute the major part of a great Porphyry Formation in Patagonia and Tierra del Fuego, and are also found in parts of West Antarctica (see this Memoir, p. 75). The rock may thus be described as quartzitic arkose. Quartz-vein rocks. These are all mainly composed of white quartz with films of a chloritic mineral. In thin section one of them shows quartz, albite, chlorite, and a little calcite, all intensely sheared and crushed. The quartz has marked undulose extinction and in the albite the twinning lamellae are bent and twisted. The chlorite is greyish green, and shows the common ultra-blue polarization; it is occasionally quite isotropic. A second rock consists of intensely sheared and sliced quartz with some large crystals of greenish brown epidote. While clearly later than the quartz, the epidote crystals have also been bent and sliced by a movement in a diff'erent direction to that which first affected the quartz. The resulting fissures have been healed by the infiltration of silica. No albite or chlorite occurs in this rock. SHAG ROCKS 9i A third quartz-vein rock consists of sheared quartz with films and foHa of almost colourless, isotropic chlorite. There is no evidence of the nature of the rocks penetrated by these veins. While there appears to be secondary quartz in the greenstones, there are no sharply defined veins. However, from their mineral composition and associations, it is likely that the quartz veins cut rocks of metamorphic type. CHEMICAL COMPOSITION OF THE GREENSTONE A composite sample from three of the least altered greenstones was analysed with the result shown in Table 7, col. i. This analysis has a characteristically basaltic pattern with its comparatively high lime and alumina which, in the rock itself, is accounted for by the abundance of epidote and tremolite, and in comparable basahs, by richness in lime-plagioclase. The analysis is, for example, much like that of the Porphyritic Central Basalt type of Mull (Table 7, col. A), and like the basalt of the South Shetland Islands (Table 7, col. B). The latter, however, has a much higher k ratio than the Shag Rocks greenstone. Table 7 I A B C SiOg 51-56 48-51 48-26 47-37 AUO3 17-54 19-44 17-42 16-46 FeaOg I -So 5-66 3-36 1-92 FeO 8-28 4-00 5-61 7-41 MgO 5-23 5-12 8-83 8-64 CaO 11-42 12-03 11-56 10-19 Na.,0 2-l8 2-53 2-44 2-74 K.,6 0-33 0-25 0-89 0-06 H.,0^- 0-34 0-48 0-24 3-38 H.O- 0-22 0-04 0-16 o-io CO., nil 0-09 nil 0-21 Tid.. 0-56 1-46 1-07 1-20 P2O5 tr. 0-16 0-22 0-14 MnO 0-36 0-23 0-14 0-15 (Ni, Co)0 — 0-04 — — S — — — 0-02 CI — — nil F — — — nil 99-82 100-04 100-20 99-99 I. Tremolite-epidote-greenstone, stones dredged near the Shag Rocks, 130 miles west of South Georgia. Anal. F. Herdsman. rr, ■ , n A. Porphyritic basalt (Porphyritic Central Type), Mull. Anal. E. G. Radley. Quoted from 'The Tertiary and Post- Tertiaiy Geology of Mull', Mem. Geol. Surv., Scotland, 1924, p. 24. B. Olivine-basalt (Recent), Penguin Island, King George Island, South Shetlands. Anal. F. Herdsman. See this Memoir, p. 59. • at ■ C. Chlorite-actinolite-clinozoisite-albite-schist, Clarence Island. Anal. E. Kluver. See this Memoir, p. 87. The West Antarctic rock to which the Shag Rocks greenstone shows most resemblance is the schist from Clarence Island (Table 7, col. C). There is obviously a close mineralogical similarity, and the chemical analyses have the same pattern, although SiO^ is lower and (Fe, Mg)0 higher, in the Clarence Island rock. The latter, however, is of more advanced metamorphic grade than the greenstone of the Shag Rocks. From Tierra del Fuego, Kranck {op. cit. supra, pp. 43, 47, 54, no) has described several ophiolitic greenstones, greenstone-schists, prasinites, etc., containing chlorite, epidote, actinolite, sphene, leucoxene, and albite, but the only analysis given of these rocks (cited in Table 6, col. A) does not accord very closely with that of the Shag Rocks greenstone. The chemical affinities of this rock clearly accord with those of a common type of basalt, and it may g2, DISCOVERY REPORTS be regarded as due to low-grade metamorphism of basaltic rocks of this type. Its association with quartz veins, and with a quartzite-like rock, and its chemical and mineralogical similarity to the prasinitic schists of Tierra del Fuego and Clarence Island, make it congruous with the whole assemblage of rock types found in the Scotia Arc, and adds confirmatory evidence for the theory of tectonic connexion between South America and West Antarctica favoured by E. Suess and other writers (see this Memoir, p. 89). PARTY. PETROGRAPHY OF THE SOUTH SANDWICH ISLANDS INTRODUCTION The volcanic South Sandwich Islands are situated at the extreme eastern end of the Scotia Arc, and form either a part of it, or a volcanic arc parallel to and in echelon with it. They are fully described in a recent publication to which reference will be fre- quently made in the ensuing pages. ^ It is proposed in this paper to summarize and collate the already published petrographic data, and to supplement them with descriptions of new material from five localities, viz. material collected in situ by Mr G. Rayner on Saunders Island, and dredgings from four stations: (i) St. 363, 2-|- miles S. 80" E. of the south-eastern point of Zavodovski Island, (2) St. 366, off the south coast of Cook Island, (3) St. 368, in Douglas Strait between Cook Island and Thule Island, (4) St. 370, 2 miles north-east of Bristol Island. In addition, there are some stones collected from a piece of floating ice near Bristol Island. Many observations on the volcanology and on the rocks of the South Sandwich Islands as seen from a distance are published in the above memoir, but the only petrographic data so far published are to be found in the following three papers: (i) O. Backstrom. ' Petrographische Beschreibung einiger Basalte von Patagonien, Westantarktika, und den Siid-Sandwich Inseln', Bull. Geol. Inst. Upsala, XIII, pp. 115-82 (1915). South Sandwich Islands, pp. 163-76. (2) G. V. Douglas and W. Campbell Smith. ' Zavodovski Island, and Notes on Rock Fragments dredged in the Weddell Sea ', ' Quest ' Exped. Report {Brit. Mus. Nat. Hist.), pp. 63-7 (1930). (3) G. W. Tyrrell. ' Report on Rock Specimens from Thule Island, South Sandwich Islands', South Sandzvich Islands Memoir, pp. 191-7 (193 1). 1 Stanley Kemp and A. L. Nelson, 'The South Sandwich Islands', Discovery Reports, in, pp. 133-98, pis. xi-xxxi (1931). Hereinafter referred to as South Sandwicti Islands Memoir. 4IZAV0D0VSK 1. t LESKOU I. 5OK0I I. 57' VINOICATIO ^ 1 ,% CANOLtHAS 1, ^' SAUNDERS 1. 58- ^ MONIAOU 1. 5* FRCEZEL ^....■■♦e«lSTOLI. THULE I <^* SEUINGSHAUSCN 1. OOK 1. 29- W . .^r ' 7- 2j6° . , , ; . T^ 1 . r M" Fig. 14. The South Sandwich Island. SOUTH SANDWICH ISLANDS 93 PETROGRAPHY General. The South Sandwich Islands, so far as present observations go, are composed exclusively of Recent volcanic rocks, the products of present-day and recently extinct volcanoes. Five of the islands, Zavodovski, Candlemas, Bellingshausen, Saunders, and Visokoi (map, Fig. 14), show definite signs of volcanic activity and emit vapour and fumes; another three, Leskov, Vindication, and Montagu, show no activity at present, although large areas of ice- and snow-free ground, indicating residual warmth, exist on the islands. The remaining islands, Bristol, Cook, and Thule, are heavily glaciated, and show no signs of volcanic activity or warm ground.^ The South Sandwich Islands Memoir (p. 150) states that the rocks consist mainly of 'reddish tuff and black basaltic lava ', and this is supported by the petrological examination of the collected rocks. The 'reddish tuif' may include reddened slags, and the lavas, while mainly basaltic, include basic andesites and even more acid types such as dacite. A possible exception to this generalization is Freezeland Peak, a small islet to the west of Bristol Island, which is referred to later (p. 99). It is proposed to describe the petrography of each of the islands in turn, starting from the northern end of the chain. Zavodovski Island (South Sandwich Islands Memoir, pp. 156-60). This island is nearly circular in outline and is 9 miles in circumference. It consists mainly of an active volcanic cone which rises from a lowlying plateau of black basaltic lava most conspicuous on the eastern side of the island. There are subsidiary craters on the slopes of the main cone, and to the south of West Bluff there are fumaroles in reddish ground with some patches and streaks of sulphur. At one point horizontal strata apparently consist of alternate beds of ash and tuff. In 1908 the Norwegian, Capt. C. A. Larsen, landed on many of the South Sandwich Islands and collected rock specimens of which, unfortunately, some were lost by accident. The collection was presented to Goteborg Museum, and was later described by O. Backstrom (i). Larsen landed at the north-west end of Zavodovski Island, which was found to consist of a porous lava carrying zeolites in the vesicles. These specimens were lost by the upsetting of the boat. Only a few small pebbles and lapilli were retained, which Backstrom identified as olivine-basalts and their tuflrs. The fragments of which the latter were composed showed fresh phenocrysts [feldspars?] in a ground-mass which had been altered by the action of solfataric gases. Their richness in phenocrysts and in shattered basaltic ground-mass material showed that they represented a common type of 'Aschentuff' which was probably rather glassy. During the Quest Expedition of 192 1 G. V. Douglas saw the island at close range although he was unable to land (2). He does not state from which direction the 'Quest' approached the island, but from the fact that he mentions a cliflF 40 ft. high with a long gentle slope inland, it may be assumed that he saw the low plateau on the eastern side. Douglas states that : ' The lava flows seen on the cliff face appeared to consist of a compact columnar basalt at the base. Above, there was a line of red cinder, and above this again what looked to be rough paehoehoe lava.' Material obtained by dredging at 19 fm. corresponded with the above-described section. The sample consisted of rounded black pellets of diameters between i and 5 mm. Twenty of these were sectioned, and ten of them were found to consist of dense black glassy basalts free from olivine. Some were crowded with minute laths of plagioclase; others showed a few small phenocrysts of plagioclase and augite. Four of the pellets consisted of dense, dark brown, glassy olivine-basalts, some containing many crystals of plagioclase and only a few of olivine and augite. Four others were paler basalts of holocrystalline-porphyritic texture with small phenocrysts of plagioclase and sometimes augite in an 1 South Sandmch Islands Memoir, pp. 151-2. 94 DISCOVERY REPORTS intergranular ground-mass consisting of minute microlites of feldspar and grains of augite and magnetite. Tlie two remaining pellets consisted of basalt glass of a deep olive-buff colour. In one of these microlites were absent, but in the other microlites of plagioclase and augite were abundant, and a little olivine was probably present. The Discovery II material submitted to me was dredged at St. 363 from depths between 278 and 329 fm. at a locality 2| miles S. 80° E. of the south-eastern point of Zavodovski Island. It consisted of two bags, one containing grey scoria or lapilli, very rough and angular, the largest being about I in. in greatest diameter; the other contained a few of the larger stones picked out from the scoria. Five thin sections were prepared from this material. The scoria and lapilli consist of a highly vesicular, opaque, pumiceous glassy basalt. The glass varies in colour from black, even in thin section, to pale brown, and carries minute microlites of plagioclase and pyroxene, the latter being noticeably more abundant in the pale brown glass. A few large crystals are entangled in the glassy sponge; these include plagioclase (bytownite, Angj), pale brownish green diopsidic augite, and olivine, all perfectly fresh. In one specimen the glass is much haematitized, and carries much larger and more numerous feldspar microlites which can be identified as labradorite (Anjo). These rocks are on the borderline between andesites and basalts. Their content of olivine is small and sporadic; and as the glassy ground-mass probably contains much free silica it may be presumed that if the magma had not been so rapidly quenched the olivine would have been made over into pyroxene by reaction, and the rock would then have been revealed as a basic andesite. This description agrees with that of the dredged material off Zavodovski given by Douglas (p. 93). One of the dredged stones, however, the largest, is undoubtedly a sedimentary rock. It is a very dense, dark grey material which looks like cementstone. In thin section it shows a carbonate mineral intermingled with argillaceous matter. The rock effervesces only when powdered and treated with hot concentrated acid, and may therefore be identified as a dolomitic mudstone. Leskov Island {South Smidwich Islands Memoir, pp. 161-2). This island, the smallest of the South Sandwich Group, lies some distance to the west of the arc on which all the other islands are situated. Its circumference measures only about i| miles. There is no record of any landing on this island, but it was observed at close range by Capt. Larsen ((i), p. 166), Lt. Filchner,^ and by members of the Discovery II party. The last-named state that the island is crescentic in outline and is doubtless a fragment of a volcanic cone. Material dredged by Larsen at a depth of 75 fm. proved to consist of basaltic rocks ((i), p. 167). At the south-eastern corner of the island a conspicuous conical rock consists of columnar basalt; the cliffs round the southern and western sides are formed of rugged flows of basaltic lava inclined towards the sea on the south side at an angle of 45°, but gradually becoming vertical towards the west. The rock walls of Crater Bay are reddish and yellowish in colour and apparently consist of tuff which shows no definite bedding and is much contorted {South Sandwich Islands Memoir, p. 162). Visokoi Island {South Sandzvich Islands Memoir, pp. 162-5). This island is one of those that show definite volcanic activity. There is no known record of a landing and most of the information regarding Visokoi was obtained during the visit of 'Discovery II'. The only geological information available is that provided by a sketch of rock exposures on the north coast by Mr F. C. Fraser {South Sandwich Islands Memoir, fig. 8, p. 164), which shows columnar basalt, dark grey rock intersected by dikes and surmounted by light grey stratified rock [tuff?], reddish and grey rocks cut by dikes, and an exposure of stratified rocks [tuffs?] in alternate layers of grey and red tints. The general impression was that the rocks were basaltic lavas and tuffs similar to those seen on Zavodovski. ^ Zum Sechsten Erdteil, pp. 1 14-15, figs. 32-6 (Berlin, 1923). SOUTH SANDWICH ISLANDS 95 Candlemas Group {South Sandwich Islands Memoir, pp. 165-72). This group consists of Candlemas Island itself, and a smaller one to the west which is now called Vindication Island. A full account of the geography and volcanic phenomena is given in the Memoir. A large collection of rock specimens from the southernmost point of Candlemas Island, made by Capt. Larsen, has been described by Backstrom in the following terms ((i), pp. 169-70, translated): [The rocks] are mostly reddish and porphyritic with rounded feldspars which sometimes give an almost white colour to the specimens. Under the microscope they are found to be extraordinarily rich in feldspar of composition Ans5, which is zoned with glassy inclusions and shows both albite and pericline twinning. The main pyroxene is hypcrsthene which is often invested by monoclinic pyroxene, but both pyroxenes may occur as independent cr\'stals. The augite shows the usual polysynthetic twinning, which is also seen in the investments around the hypersthenes. Strongly corroded olivine also occurs but is not common. It is mostly altered to a blackish brown dust, but all the other constituents are fresh. In regard to the systematic position of the rocks, their richness in plagioclase suggests that they represent a transition between the basalts and the andesites. It is difficult to assign some of the rocks to either group, but others which are richer in olivine and pyroxenes should be relegated to the basalts. Another type has an extremely fine-grained but holocrystalline texture. It is, however, little different to the above in mineral composition. Its plagioclase is lath-shaped not equidimensional, its pyroxene is sharply euhedral, and olivine is absent. Fragmental rocks also occur as very fresh, reddish brown, sandy tuffs which consist of lapilli of hazel-nut size. The latter consist of vesicular lavas with a glassy ground-mass full of crystallites, and carrying numerous crystals of plagioclase, augite, and hypersthene. It will be seen how closely comparable these lavas and tuffs are to those of Zavodovski Island and Saunders Island (p. 96). Members of the Discovery II party were not able to land on Candlemas Island, but they made numerous observations at close range, noting rugged flows of black basaltic lava in the northern plateau often showing columnar structure {South Sandwich Islands Memoir, pi. xvii, fig. 3). Mr F. C. Fraser has also provided an excellent sketch of rock exposures on the east coast {ibid. fig. 12, p. 169) showing what are obviously stratified tuffs and a coarse agglomerate. It was found impossible to land on Vindication Island, but the geological structure of the island was well seen in a sheer cliff face on its north-western side. The rocks here consist of irregular masses of red and brown colours, presumably tuffs, cut by dikes of grey rock which run obliquely, vertically, and sometimes horizontally, not infrequently intersecting one another. Two islets, Cook Rock and Trousers Rock, both of which are tunnelled by wave erosion, show horizontal strata of red tuff and hard grey rock. Saunders Island {South Sandzvich Islands Memoir, pp. 172-4). Saunders Island, with a circumference of 17 miles, is one of the largest of the group, and is, perhaps, the best known geologically. At its centre is the glaciated but actively volcanic cone of Mt Michael (2640 ft.). The south-eastern part of the island is composed of bare hills (700-800 ft.) apparently consisting of loose ash or volcanic mud, and with several extinct craters. A very fine photograph of a half-section of a crater on the south coast is given in pi. xx, figs. 2 and 3, of the Memoir. The northern part of the island is a low plateau. All the rock exposures show that the basement of the island consists of columnar basalts similar to those of Candlemas and Zavodovski. Capt. Larsen landed with difficulty on the south-eastern coast ((i), p. 170), and Backstrom describes the rocks collected here as, in the main, different from the type common in the South Sandwich Islands in being very dense and non-porphyritic. Under the microscope these rocks show a well-developed fluidal structure delineated by the alinement of the minute feldspar laths in the direction of flow. The mineral composition is plagioclase (An^s-gs), almost colourless pyroxene in rounded grains which belongs to the enstatite-augite series, and magnetite. This rock is free from 8-2 96 DISCOVERY REPORTS olivine, and a little analcite was found in one of the thin sections. A chemical analysis of the principal type, free from analcite, is published, which is set out with others from the South Sandwich Islands in Table 8 (p. loi) of this memoir. Biickstrom calls the rock a basalt. Owing to unfavourable conditions the Discovery II party was unable to land on Saunders Island, but on 28 November 1937, Mr G. Rayner was able to get ashore for a few hours from the 'William Scoresby'. He made some geological observations and collected a small number of rock specimens which are described below. The observations that follow are condensed from his MS. report. Mr Rayner landed near the penguin rookery on the south side of Cordelia Bay (see Chart in the South Sandwich Islands Memoir, pi. xix). The beach material consisted of a loose black volcanic ash, the size of coarse sand or grit. Behind a low cliff of compressed snow heavily loaded with the same ash was a level area extending back to the hills. This platform consisted of a loose ash-like material to a depth of some inches, with occasional small boulders up to 18 in. in diameter of a heavy dark basaltic rock resting upon it. From this point Mr Rayner walked along the shore eastward until he reached the first outcrop of hard rock which forms the basement of the Nattriss peninsula. Here he ascended the hill to the south near the point marked 800 on the Chart. On its northern slopes there were several outcrops of a soft volcanic mudstone with a sub-horizontal stratification, standing up as buttresses and ridges between steep-sided ravines. Mr Rayner thus gained a ridge which sloped eastward to Nattriss Point. The higher parts of this ridge still consisted of the stratified mudstone, which was undergoing extremely rapid atmospheric erosion. At one place he encountered a remarkable pillar 15-20 ft. high carved out of the soft material (' The Beacon '). Elsewhere along the ridge a light, vesicular, reddish, scoriaceous rock was found. Descending eastward towards Nattriss Point Mr Rayner found that the rock became coarser in texture, and took on the appearance of volcanic tuff, light buff in colour, in which many large fragments of rock were embedded. This series of coarse tuffs rested on the roughly horizontal platform of dark, vesicular, basaltic rock of which Nattriss Point is composed. This rock falls in sheer cliffs to the sea and has a columnar appearance owing to wave erosion along vertical joints. With, as the writer thinks, considerable probability, Mr Rayner concludes that ' a volcanic explosion has occurred at no very distant date, possibly from the crater to be seen to the south-west of our landing-place, and near the junction of the ice-covered main part of the island and the earthy region explored. This explosion has thrown up the clastic material forming the hill now resting on a hori- zontal table of rock of which Nattriss Point is the visible part. The finest material would be the last to settle, and this has formed the upper strata of soft mudstones seen in the fast dwindling ridges and buttresses along the hillside and in the pillar at the summit.' Six thin sections were made from the specimens collected by Mr Rayner. The lava which forms the basement of the Nattriss peninsula is a black, highly vesicular rock which, in thin section, shows an abundant ground-mass of minute microlites of plagioclase with granules of augite and magnetite, within which is set a generation of somewhat larger feldspar laths, and finally a few micro-phenocrysts of feldspar and yellowish augite. Owing to their small size it is difficult to make out the composition of the plagioclase microlites of the ground-mass, but they give extinctions up to about 15'' indicating a composition Aug,,. The larger microlites and micro-phenocrysts are highly zonal, and their com- position ranges about Augs, which is the composition ascertained by Backstrom. The pyroxene, too, is zonal, as shown by an undulatory extinction. It is a pale yellow variety of moderate double refraction, and is probably, as Backstrom surmises, a member of the enstatite-augite series. The larger feldspars and pyroxenes, while occurring independently, are often aggregated into clots of which the feldspar forms the greater part, and the microlites of the ground-mass are stream-lined around these clots. I SOUTH SANDWICH ISLANDS 97 Olivine does not occur in this type which, owing to its feldspathic composition, would be better termed andesite than basalt. Another specimen was taken from what appeared to be an inclusion within the above-described lava It is not so dark in colour, but the thin section shows that it is the same lava with, however, a somewhat finer grain and a few sporadic olivine crystals, most of which are altered to green serpentine. This rock is probably a portion of the same lava, but consolidated slightly earlier than the main mass of the flow, and thus retaining a few of the early crystallized olivines. It may have been carried as a solidified lump of slag on the surface of the moving flow, and have been incorporated in it by over-rolling. . . , ■ n The coarse agglomeratic tuff which overlies the lava basement of the Nattnss peninsula is a well- consolidated material of light buff colour containing numerous fragments of gravel size. In thin section it proves to be a coarse lithic tuff consisting mainly of large angular fragments of the lavas embedded in a matrix of smaller fragments and broken crystals. The lava fragments are vesicular andesitic basalts of the same type as that described above, but they show every gradation of texture from purely glassy to holocrystalline-micro-granular. The broken crystals include plagioclase, augite, and fresh olivine. Conspicuous among the rock fragments are glasses of a bright green colour. An isotropic or very feebly birefringent zeolite with cubic cleavage forms a scanty cement in some parts of the sUde This may be the analcite recorded by Backstrom ((i), p. 171)- This rock must have been formed by an explosion in or under a fully consolidated lava, and it may be suggested that it was produced by renewed activity in a nearby volcano which had been temporarily sealed by a plug of solidified lava. r , m .. • The volcanic mudstone which overlies the lithic tuff and forms the higher parts of the Nattnss peninsula, in contrast to the lithic tuff, is a vitric ash consisting almost entirely of small angular fragments of clear brown glass. The only other constituents are a few small fragments of feldspar, aueite and magnetite. This was undoubtedly formed by explosions within a still liquid lava. Hence the sequence of events pictured by Mr Rayner (p. 96) must be slightly amended. The vitric ash does not represent the finer, and the lithic tuff the coarser, material derived from one and the same explosion- but the lithic tuff probably represents the disintegration by explosion ot a solidified plug, and the vitric ash a subsequent explosion within the liquid lava that welled up into the crater The coarse black sand at the landing-place in Cordelia Bay consists of angular fragments of brown glass often blackened with separated magnetite, and crystals, in about equal proportions. The crystals fnclude plagioclase, augite, and olivine, the last-named being rather more abundant than usual This material may have been formed by explosion in an olivine-basalt magma within whic4i while still liquid crystallization had advanced to a considerable extent. Examination of a small pebbe enclosed in the sample bears out this diagnosis. It is an olivine-basalt with large phenocrysts of labradorite (An ) abundant fresh yellowish olivine, and some magnetite, in a very dense ground-mass consisting of augite granules and feldspar microlites, in which the augite is decidedly predominant. MontcZ Island {South Sandwich Islands Memoir, pp. 174-6). Montagu is the largest island ot he group with a circumference of about 24 miles, and is one of the least well known. It contains wha fs probably the highest summit of the group, Mt Belinda (4500 ft.), almost certainly an ext.nc volcano. Montagu is the most heavily glaciated island of the arc, and has fewest signs of residual warmth in the shape of areas free from snow and ice. • r , • ^u^ The Discovery II party did not land on the island, but they had the opportunity of making the following observations on the rock exposures as seen from a distance: As on other islands the lowest strata seen in rock exposures are usually of black bas^alt, often columnar in structure, and itTs ol blalt that the outlying rocks are formed. Above it red and yellowish tut^s with some hard grey rock are 98 DISCOVERY REPORTS to be found. At several points the rocks are clearly stratified, showing three or more horizontal layers of dark grey rock separated by narrow bands of red tuff. Sometimes yellow tuff with red inclusions was to be seen and frequently the rocks were much contorted and intersected by dykes. At the north-eastern corner of the island are low cliffs formed of a light grey rock, perhaps volcanic ash. {South Sandwich Islands Memoir, p. 175.) Capt. Larsen landed at the south-eastern corner and mentions a crater here, as well as at the north-eastern point of the island. Biickstrom ((i), p. 175) described the rocks collected as rather uniform types of vesicular olivine-basalts in which phenocrysts of olivine, augite, and plagioclase (Ansa) predominate over the ground-mass. The ground-mass consists of small granules of pyroxene, laths of plagioclase, and some magnetite. The resemblance of these rocks to the olivine-basalt of Saunders Island (p. 97) is obvious. Bristol Island (South Sandwich Islands Memoir, pp. 176-8). Bristol Island is an irregular oval in shape and has a circumference of 14 miles. The highest point is Mt Darnley (3600 ft.). Its profile seen from the north has the shape of a horse-shoe, and is conjectured to represent part of the rim of a crater. Bristol Island is heavily glaciated and the Discovery II party were satisfied that all volcanic activity had ceased. Three rocky islets, Grindle Rock, Wilson Rock, and Freezeland Peak, stand in line off the western coast of the island. Capt. Larsen landed on the north-eastern side of the island^ and collected some rock specimens. Biickstrom ((i), pp. 175-6) describes them as of reddish grey tints, and as showing numerous small crystals of feldspar. In thin section numerous micro-phenocrysts of zonal plagioclase are disclosed, of composition An75_85 . Pyroxene is confined mainly to the ground-mass and belongs to the enstatite- augite series. Olivine is only sparingly present. A photomicrograph of this andesitic basalt type is given by Backstrom ((i), fig. 20, p. 176). It conforms closely to the main type of lava erupted from the South Sandwich Islands volcanoes. Although no landing was made, the geological observations made by the Discovery II party (South Sandwich Islands Memoir, p. 177) are important and must be quoted in full: The rocks on Bristol are similar to those on the other islands. At Fryer Point black basaltic lava is to be seen and the rock exposures on the bluff on the south side, at the western headland and in other parts, are of yellowish and red tuff, or tuff conglomerate, sometimes stratified with a grey rock interposed between the layers, but frequently much contorted and with many intrusive dykes. From a geological point of view the three large outlying rocks appear to be more interesting than any other place in the entire group of islands. . . . The great pillar on Freezeland is composed of a pale brown rock of a kind not seen elsewhere. It showed distinct signs of bedding and in the upper part of the column some broad reddish bands. We believe this may be a sedimentary rock. The eastern part of Freezeland, forming the lesser of the two summits, is different ; it is formed of a brownish rock, with vertical fissures and striation, and may be metamorphic. Wilson Rock, nearer the mainland, is a vast mass of black columnar basalt, while Grindle Rock repeats the reddish and yellowish tuff's seen on the adjacent headland of the island. Thus, if our conjectures are correct, the whole succession of rock formations in the Sandwich group is to be found in these three islets. Freezeland shows the only likely exposure of the underlying sedimentary series that we know to exist, Wilson is of the overlying basalt, here seen in far greater thickness than elsewhere, while Grindle is formed of the superposed tuffs which are characteristic of all the islands of the group. Among the material from the South Sandwich Islands submitted to the author there were specimens from near Bristol Island. One of these was a bag of scoria and lapilli dredged from St. 370 at a point two miles north-east of Bristol Island, and a bag of small stones, including lapilli, which were picked off a piece of floating ice near the island. A thin section of the dredged scoria from St. 370 shows that it is a sponge of opaque black glass with minute microlites of feldspar and augite, and a few micro-phenocrysts of plagioclase (Augo) ' The position of the landing-place is mentioned in Biickstrom's memoir ((1), p. 175). Cf. South Sandwich Islands Memoir, p. 178. SOUTH SANDWICH ISLANDS 09 entangled in it. This seems to represent an extremely vitreous phase of the andesitic basalt lava described by Backstrom, and carries the same lime-rich feldspar. Most of the smaller fragments recovered from the piece of floating ice answer to the above de- scription. A larger stone, however, is 2 in. in length and presents a microscopic appearance very similar to that of the ' feldspathic basalt ' described and figured by Backstrom. It shows very numerous micro-phenocrysts of plagioclase (An^j-g,,) with subordinate augite and olivine, in a dark glassy ground-mass carrying microlites of feldspar and augite. All the phenocrysts are perfectly fresh and euhedral. The augite is a yellowish, slightly-pleochroic variety belonging to the enstatite-augite series. In this rock the olivine is much more abundant than in Backstrom's material, and it must be regarded as an olivine-basalt. Two other stones are interesting, as they are non-igneous. One is a fragment from a quartz-vein rock, and the other is an epidote-biotite-gneiss. In thin section the latter shows a coarse mosaic of quartz and oithoclase alternating with folia consisting of straggling crystals of bright yellow biotite and epidote (with some clinozoisite). There is also a little ilmenite altering to sphene, and a few fragments of deep green pleochroic hornblende. It is not possible to say whether this is an orthogneiss or a paragneiss. The mineral composition favours the orthogneiss interpretation, but an arkose would yield this type of gneiss on metamorphism. The label attached to the material from floating ice does not state on which side of Bristol Island it was recovered. As the metamorphic fragments were closely associated with scoria which indubitably came from Bristol Island, it seems probable that they too were derived from that locality. It is possible that the metamorphic pebbles came from Freezeland Peak which the Discovery II party believed to consist of sedimentary and metamorphic rocks. Southern Tlmle Group (South Sandwich Islands Memoir, pp. 178-89). This group consists of three islands, Thule, Cook, and Bellingshausen, in order from west to east. Of these, Cook Island is the largest, having a circumference of 9^ miles; Thule, the next largest, is more embayed than Cook and has a coastline of 10 miles; Bellingshausen, the smallest, is only i^ miles wide. Bellingshausen is still an active volcano, as shown by the steam and vapour rising from it, and by the admirable sketches of Lt.-Cmdr. J. Irving (South Sandwich Islands Memoir, fig. 19, p. 184). Cook and Thule, however, are buried beneath thick ice caps and there are no signs of present volcanic activity. Nevertheless, soundings in Douglas Strait between Thule Island and Cook Island have disclosed a steep-sided basin of elliptical shape and more than 400 fm. in depth. At the north and south entrances to Douglas Strait the depths are less than 20 fm. This has been interpreted, correctly in the writer's opinion, as the inundated crater of a volcano of which Thule Island and Cook Island are the remnants. This view is reinforced by the parallelism of the eastern embayment of Thule Island, and the western embayment of Cook Island, with the adjacent contours of the submerged basin (South Sandwich Ishmds Memoir, fig. 16, p. 179), and by the photograph of the eastern side of Thule Island (ibid. pi. xxx, fig. 4), which shows bedded lavas and ashes dipping westward and outward from the Douglas Strait crater. Of the geological constitution of the Southern Thule Group little is known. On Bellingshausen the Discovery II party noted, as on other islands, black columnar basalt with overlying agglomerate, tuff, and ashes. Cook Island (South Sandwich Islands Memoir, pp. 185-6). Rock faces are exposed in the cliffs bordering Douglas Strait. They are described as of yellow, red, or brown colours, sometimes showing signs of bedding but always much crumpled and contorted, and seamed with dikes of grey rock. Large, apparently intrusive, masses of brown rock showing a vertical striation were also seen. Fortunately, however, some stones were dredged by 'Discovery 11' at St. 366, 4 cables south of :oo DISCOVERY REPORTS Cook Island at depths between 155 and 322 m., and a few small fragments of rock at St. 368 in Douglas Strait, i mile north of the Twitcher Rock, dredged from a depth of 653 m. near the bottom of the great submerged crater. These pebbles, which consist mainly of slaggy and vesicular lavas, one or two being well rounded, range in size from about 2 in. in greatest diameter down to about | in. Sixteen of these stones were sectioned for microscopical examination. All of them were found to be textural variants of an olivine-basalt lava. Nearly holocrystalline varieties are grey and compact, and the glassy types black, vesicular, and slaggy, in hand specimens. In thin section these rocks are found to be highly porphyritic, carrying very numerous small phenocrysts of plagioclase, augite, and olivine, in a ground-mass consisting, when holocrystalline, of minute crystals of plagioclase, augite, and magnetite. In the more slaggy varieties the ground-mass becomes richer in dark glass and the number of microlites diminishes. In fact a complete passage from holocrystalline to a purely glassy ground-mass can be traced. The plagioclase phenocrysts are generally most numerous, with augite and olivine following in that order; but in a few rocks the olivine almost rivals the feldspar in abundance. The plagioclase is both chemically and mechanically zoned and shows complex twinning ; its composition ranges between An7o and Augs . The pyroxene is again the yellowish, slightly pleochroic variety of the enstatite-augite series. The olivine is perfectly fresh and often euhedral, especially in the more glassy varieties of the rock. It gives a dead straight isogyre and therefore contains about 13 per cent of the fayalite molecule. This type is an olivine-basalt which compares closely with that from Bristol Island (p. 98), and with the younger basalts of the South Shetland Islands (e.g. Penguin Island, p. 46). A chemical analysis of one of the more holocrystalline types is recorded in Table 8 (p. 10 1). Thule Island {South Sandwich Islands Memoir, pp. 187-9). The south-eastern plateau of Thule Island appears to be composed of the usual black columnar basalt. Near Cape Flannery on the west coast are beds apparently composed of yellowish tuff and ash, and farther north the rocks are definitely stratified, three layers of ash separated by red tuff overlying black basalt. A landing was made by the Discovery II party on Beach Point at the north-eastern corner of the island. The ridge at Beach Point is composed of hard grey rock with outcrops of red tuff and a soft, crumbling, black rock, perhaps volcanic ash, at its summit. The steep cliffs facing Douglas Strait show contorted masses of red, yellow, and dark brown rocks with intrusive dikes. Rock specimens collected here were described by the writer in an appendix to the South Sandwich Islands Memoir (pp. 191-7). Of the fifteen specimens, eight were obtained from exposures and seven were cobbles from the beach. Six rocks were obtained from an escarpment at 50 ft. above sea-level. Four of these were acid lavas (dacite) with good flow structures, and two were pyroxene-andesites containing both augite and hypersthene. As a black slaggy andesitic lava with red crusts was collected at 100 ft. it is inferred that the upper part of the cliff probably consisted of andesite while the dacites came from an underlying flow. At the top of the cliff, 150 ft. above sea-level, a true andesite-tuff was collected, which may represent the final explosive discharge of this volcanic episode. The beach cobbles and pebbles consisted mainly of dacites and andesites similar to those collected m situ. In addition there was a specimen of olivine-andesite (or andesitic basalt) and one of andesitic pumice. Thule Island is therefore notable as providing the only acid lavas so far known in the South Sandwich Islands. The hypersthene-bearing andesites are also distinctive as they have hitherto only been recorded from Candlemas Island (p. 95). Analyses of dacite and hypersthene-andesite from Thule Island were published in the above Appendix, and are restated in Table 8 below. I SOUTH SANDWICH ISLANDS CHEMICAL COMPOSITION OF LAVAS FROM THE SOUTH SANDWICH ISLANDS Four chemical analyses of the lavas are recorded in Table 8, in order of decreasing silica percentage, along with comparable analyses of lavas from the South Shetland Islands, South America, and the West Indies. In the lower part of the table the von Wolff parameters as modified by the author (see p. 59) are given. Table 8 I A B 2 C 3 D 4 E F SiO„ 69-45 67-71 69-56 54-90 54-24 52-68 52-00 48-34 48-26 48-71 AIP3 14-20 14-65 15-65 17-62 17-20 16-38 19-22 13-45 17-42 18-40 Fe,03 2-83 1-59 1-24 2-70 2-81 3-II 2-73 1-12 3-36 3-70 FeO 3-24 3-29 0-91 6-80 4-98 7-98 5-61 11-34 5-61 5-25 MgO 0-25 0-85 0-82 3-93 5-84 7-47 5-54 6-62 8-83 10-30 CaO 3-05 2-34 2-52 9-05 10-19 8-o8 10-58 11-43 11-56 lO-II Na,0 4-15 6-09 409 2-90 2-91 2-75 2-53 2-22 2-44 2-34 K26 I-5I 1-99 2-19 0-54 0-92 0-44 0-76 0-19 0-89 0-43 H,0+ 0-40 o-i6 1 (0-301 1 0-20) (o-20 294 0-24 0-25 H,0- o-6o - J 2-92 0-09 0-20 |o-i5 0-30 o-i6 CO,, nil — — nil — — nil 0-12 nil — TiO, 0-15 i-oo — 0-70 0-91 0-77 0-63 1-47 1-07 1-08 P2O5 0-14 o-i6 0-13 0-09 0-09 0-02 o-i I tr. 0-22 0-06 MnO 0-07 — — 0-23 — o-i6 o-ii 0-32 0-14 o-o8 (Ni, Co)0 ml — nil — — — — — S tr. — — tr. — tr. — — CI — — — — — 0-05 — — 100-04 99-83 100-03 99-96 100-18 100-09 100-17 99-86 100-20 100-71 34-6 20-1 34-7 lO-I 5-3 3-6 3-5 — i-o -5-2 -3-6 F' 44-3 62-5 49-2 27-3 29-3 25-1 25-5 19-3 25-0 21-7 M' 2I-I 17-4 16-1 62-6 65-4 71-3 71-0 81-7 8o-2 81-9 nak 62-6 82-6 67-9 30-1 33-7 30-4 26-1 28-0 28-6 23-9 k 19-3 17-7 25-9 9-6 17-5 8-2 16-3 5-4 18-4 II-6 1. Dacite (Dacitoid) lava, Beach Point, Thule Island, South Sandwich Islands. Anal. F. Herdsman. Quoted from G. W. Tyrrell, 'Report on Rock Specimens from Thule Island, South Sandwich Islands', Souih Samhvicli Islands Memoir (1931), p. 192. A. ' Trachyandesite ' (Gourdon); Santorinite (Earth and Holmsen), Deception Island, South Shetland Islands. Quoted from E. Gourdon, C.R. Acad. Sci., Paris, clviii, p. igo6 (1914). Also see this Memoir, p. 58. B. Dacite, Guaitara Slope, Loma de Ales, Colombia. Quoted from J. P. Iddings, Igneous Rocks, 11, p. 496 (1913). 2. Hypersthene-andesite lava, Beach Point, Thule Island, South Sandwich Islands. Anal. F. Herdsman. Quoted from G. W. Tyrrell, op. cit. supra, p. 195. C. 'Basalt' (Gourdon), Bridgeman Island, South Shetland Islands. Quoted from E. Gourdon, op. cit. supra, p. 1906. Also see this Memoir, p. 59. 3. ' Olivine-free basalt' (Backstrom), Saunders Island, South Sandwich Islands. Quoted from O. Backstrom, Bull. Geol. Inst. Upsala, xiri, p. 173 (1915). D. Olivine-basalt lava. South Soufriere Hill, Montserrat, West Indies. Anal. F. Herdsman. Quoted from A. G. Mac- Gregor, 'The Volcanic History and Petrology of Montserrat. . . ', Philos. Trans., Ser. B, ccxxix, p. 74 (1938). 4. Olivine-basalt lava. Cook Island, South Sandwich Islands. Anal. F. Herdsman. New analysis. E. Olivine-basalt lava. Recent volcano. Penguin Island, King George Island, South Sandwich Islands. Anal. F. Herdsman. New analysis (see this Memoir, p. 59). F. Labradorite-basalt, Chateaubelair, St Vincent, West Indies. Quoted from A. Lacroix, 'Les caracteristiques litho- logiques des petites Antilles', Livre Jubilaire, Soc. Geol. Beige, pp. 387-405 (1926). There is a wide gap between the dacite (no. i) and the prevalent andesitic and basaltic lavas (nos. 2, 3, 4) of the South Sandwich Islands— a gap which may be filled by future collections, although it seems probable that all of the islands are built mainly of the basic lavas. Comparatively high lime ,02 DISCOVERY REPORTS is characteristic of the whole series, inckiding the dacite in which the nak ratio is only 62-6. Another general feature is the low F' jM' ratio (less than 0-5) in the prevalent basic lavas. This is in agreement with the Recent basalt lavas of the South Shetlands, but is in strong contrast with the equivalent lavas of the Andes, in which the F' jM' ratio fluctuates round about unity. It is a remarkable and perhaps significant fact that West Indian or Antillean lavas agree best with those of the South Sandwich Islands in this respect (cols. D and F, Table 8). The dacite of Thule Island (Table 8, col. i), while of sodic type, does not compare very well with the analogous santorinites of Deception Island (this Memoir, p. 58). Comparing the tiak ratios (Table 8, cols, i. A) it is seen to be much less alkalic than the Deception Island rock, and that entails a much larger amount of free silica (O). It compares rather closely, however, with an Andean dacite from Colombia (Table 8, col. B). The hypersthene-andesite of Thule Island (Table 8, col. 2) compares fairly closely with the Bridgeman Island basalt (Table 8, col. C), but no Andean lava of like silica percentage could be found with even an approximately similar F'l'M' ratio. Backstrom's ' olivine-free basalt' from Saunders Island (Table 8, col. 3) finds its closest analogue in an olivine-basalt from Mont- serrat (Table 8, col. D). The olivine-basalt of Cook Island (Table 8, col. 4) is only slightly undersaturated (^= — i-o), notwithstanding its comparatively large content of olivine. This illustrates its affinity with the more basic types of andesite. It compares well with the olivine-basalt lava of the Penguin Island volcano (Table 8, col. E), with the exception that it is slightly less undersaturated and somewhat more potassic than that rock. Again, the closest analogue of this rock is a labradorite-basalt lava from St Vincent in the West Indies (Table 8, col. F). It would appear, therefore, that the predominant basic lavas of the South Sandwich Islands show closer affinities with the comparable rocks of the Antilles than with those of the Andes. This may, in turn, be regarded as evidence in favour of the view that the South Sandwich Islands do not lie on the main line of the Scotia Arc, but form an easternmost ridge parallel to and in echelon with it. On this view the main line of the Scotia Arc may curve southward from the eastern end of South Georgia and join up with the South Orkneys. The most recent chart of the Scotia Sea^ shows South Georgia trending to the south-east away from the line connecting it with the South Sandwich Arc, and pointing towards a marked northerly projection of the 3000 m. depth-contour which, in turn, leads towards the South Orkney Islands. Of the basement on which the volcanoes of the South Sandwich Islands stand we possess only very exiguous and doubtful scraps of information, namely, a comparatively large piece of dolomitic mudstone dredged off Zavodovski Island (p. 94), and fragments of epidote-biotite-gneiss and vein quartz taken from a piece of floating ice near Bristol Island (p. 99). Any future geological exploration of the islands should therefore include search for exposures of this foundation, and examination of coarse fragmental igneous deposits for non-volcanic material which may be presumed to have been derived from the basement. The latter line of research is much more likely to be fruitful than the former, except perhaps on Bristol Island. Acknowledgements. The author's thanks are due to the Discovery Committee for defraying the cost of the new rock analyses published in this work, to Prof. W. J. McCallien, D.Sc, for re-drawing Fig. 8, p. 60, to J. M. Wordie, M.A., for his valuable introductory Foreword, and to Dr N. A. Mackintosh for his editorial vigilance during the progress of this Memoir towards publication. 1 H. F. P. Herdman, 'Report on Soundings taken during the Discovery Investigations, 1926-32', Discovery Reports, VI, pi. xlv(i932). [Discovery Reports. Vol. XXIII, pp. lo^-ij^, Jntie, 1945.] THE DEVELOPMENT AND LIFE-HISTORY OF ADOLESCENT AND ADULT KRILL, EUPHAUSIA SUPERB A By HELENE E. BARGMANN, Ph.D. CONTENTS Introduction Material and methods Acknowledgements . Development Larval krill Adolescent krill Adolescent males Adolescent females Adult krill Mature males Mature females Pairing . Spawning Average growth rate Factors influencing growth rate Conclusions Bibliography Appendix . . . . page 105 105 106 106 106 108 108 III 114 114 "5 117 118 120 128 130 131 132 THE DEVELOPMENT AND LIFE-HISTORY OF ADOLESCENT AND ADULT KRILL, EUPHAUSIA SUPERB A By Helene E. Bargmann, ph.d. (Text-figs. 1-3) INTRODUCTION ^His paper is an extension of my short one, published in 1937- The stages in the development of Tthe reproductive system described therein have been used here to work out the composition of the euphausian population as a whole and its growth rate. This method was first employed by Ruud, but as he was handicapped by lack of material he could not carry his work quite far enough. I have been more fortunate in having access to the very extensive Discovery Collections; indeed, there has been more material than I could cope with single-handed, and some selection became necessary. My object has been to obtain as complete a series of observations as possible throughout the whoL yelr. Unfortunately, weather and ice conditions in the Antarctic make it difficult to fish nets in autumn and winter. The material for this time of the year is consequenUy very scanty compared with that for the spring and summer months when there was such great abundance that I could not examine it all. The voyages of the two ships R^R.S. D-o-ry 11 and R R S 'William Scoresby', have covered between them the whole of the Antarctic zone, but the r programme of work has kept them so continually on the move that regular observations in definie LaUties are not available in consecutive months. I have therefore had to combine -tenal r m different regions and different seasons in order to obtain records extending over all he months of he year, and even so the material for the month of July is so scarce as to be neghgible. however 1 is reasonable to conclude that, by using material for several seasons, a very fair general idea of the average conditions in which Euphausia mperba grows and breeds is obtained. MATERIAL AND METHODS Material collected over a period of ten years was used. The me.ltod of examining specimens was the same as that described in my previous paper. Each specimen was measured to .he nearest mrllune r from the anterior margin of the eyes to the tip of the telson; the carapace was then opened under a binocular microscope and the stage of development of the reproduct.ve system was determined, external sexual characters also being noted ; 8029 specnnens were measured and dissected ,n this way. The results of th,s intenL investigation are all set out in the appendix. The total catch from each station has been divided into males and females, which are tabulated separately. All particulars of ngth and internal and external development are given, together with the totals of the different stages. Frfser's records of eggs and early larvae have been added .0 the lists of females to show as clearly as possible the correlation between the occurrence of adults and eggs. t.t j-ff„.„, ' No statistical tests of validity have been applied to calculations of the average lengths of the different stages, because the stages are in themselves always anatomically distinct. Nor have any fortnulae been used in working out the curve of growth. There are too many factors involved for any of the existing mathematical tfethods to be applied with any certainty. As Ottestad (1933) writes: "In the course of 1-2 ^°^ DISCOVERY REPORTS our studies of the problems of growth, it has gradually become manifest to us that, with our present knowledge of the numerous factors determining growth, the problem of finding a law that will explain the whole Cham of causes upon which growth depends is for the time being insoluble." ACKNOWLEDGEMENTS I am fortunate in having been able to discuss various problems arising during the course of this work with Dr N. A. Mackintosh and with Dr F. C. Fraser, and I am very grateful for their criticism and advice. My colleagues, Dr T. J. Hart and Miss D. M. E. Wilson, have helped me in many ways the former by his work on the phytoplankton of the Antarctic zone, and the latter by her constant interest and the practical way in which she has helped me to reduce the large body of evidence into manageable shape. DEVELOPMENT LARVAL KRILL During the first year of growth, Enphausia mperba passes from the egg through the successive larval stages of nauphus, metanauplius, calyptopis and furcilia, umil it enters upon its second year of post- larval or adolescent life. Its larval history has been dealt with in detail by Fraser (1936) in his paper on the "Development and distribution of the young stages of krill {Euphamia mperba) ". A summary of his work and a com- parison with the observations of Taube and Lebour on euphausians of the northern hemisphere must be given here, in order to present as complete a record of early growth as possible. Fraser obtained, by analysis of plankton samples, records of eggs and their occurrence extending from the first part of November to the latter part of March, a period of four and half months Just before laying, the eggs of E. superba are so tightly packed that, on the outer surface of the ovary they are approximately pentagonal or hexagonal in shape, while on the inner side they are roughly conical I have measured sixty eggs from two gravid females, and I find that their average diameter is 0-55 mm although their greatest diameter may be as much as o-68 mm. or even 072 mm. (Ruud, 1932), but that! after laying, the eggs assume a spherical shape with a consequent adjustment in size, those' found in the plankton and examined fresh measuring o-6o mm. Fraser states that "eggs occurred in the plankton showing all stages of development, culminating m the clearly distinguishable form of the ist nauplius". Only two free-swimming specimens of the ist nauphus were obtained, one measuring 0-63 mm. in length and the other o-66 mm. These were caught during the second half of December, together with three 2nd nauplii, measuring 0-65 o-68 and 070 mm. respectively. "The rarity of ist and 2nd nauplii and the smallness of numbers where records exist may mdicate that these stages are passed through very rapidly in this species, as in other euphausiids where the development is known". Taube (1915) and Lebour (1926) found that in northern waters the euphausian egg can develop into the metanauplius within a few days. Observations on Nyctiphanes norvegicus indicate that the free- swimming nauplius is hatched from the egg three days after laying, and that by the fifth day the limbs have taken on the metanauplius form, but that the mandible and lower lip characteristic of the fully developed stage do not appear until about the fourteenth day. Metanauplii occurred in the Discovery material in fair numbers from February onwards very big catches being obtained at two stations in March. The average length of the larvae at this stage is approximately 0-95 mm. The measurements of these early developmental forms show that the larvae do not grow very I07 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL rapidly until they begin to feed independently. Sars (1898) thought that this occurred in the meta- nauplius, when the mouth opens to the exterior ; but Taube has shown that at this stage, in Nyctiphanes norvegicus, there is as yet no connexion between the mouth and the mid-gut, and Macdonald (1927) states that, in Meganyctiphanes norvegica, "although an open mouth is present in the metanauplius it was not found to feed ". The internal yolk supply suffices until the gut is fully established. This occurs after the calyptopis stage is reached, when the mouth and proctodaeum become connected with the mid-gut. The cells of the mid-gut still contain a certain amount of yolk, but Sars (1898) writes that the larva now begins to feed actively, " chiefly upon small Diatomeae, the remains of which could be distinguished by microscopical examination of the contents of the intestine". In E. superba, "the more typically oceanic species of diatoms are evidently digested rapidly: recognizable fragments are rather rare even in the crop (Hart, 1934).. • .Two forms that appeared constantly in the stomachs of adult specimens and remained clearly recognizable were Fragillaria antarctica^ and Thallassiosira antarcttca". Table i has been compiled from the data in Eraser's paper and gives the average lengths of the different larval stages. It will be noticed that at the ist calyptopis stage, when the larva begins to feed, its length is at once almost doubled, after which growth proceeds again more regularly throughout the summer. "By the time the euphausiid reaches the 6th furcilia stage, the major developmental changes have been effected and in appearance it is characteristically a euphausian." Table i . Average lengths of larval stages Average Average Average Stage length of Stage length of Stage length of larvae in mm. larvae in mm. larvae in mm. Egg o-6o (diameter) 2nd calyptopis 2-71 3rd furcilia 7-32 I St nauplius 0-65 3rd calyptopis 3-98 4th furcilia 8-01 and nauplius 0-67 I St furcilia 4-50 5th furcilia 9-52 Metanauplius °-95 2nd furcilia 5-II 6th furcilia 11-34 I St calyptopis 171 By plotting half-monthly average lengths of the larvae for the period of one year, Fraser found that from November to March (the period of spawning) growth was slow, but that it increased steadily from March to June, was retarded during the mid-winter months and began to increase again at the end of August, by which time the first adolescents had made their appearance. Evidently, growth from the egg of the adolescent occupies an average period of about nine months, although under optimum conditions it can proceed more rapidly. Eraser's work on larval krill shows clearly that spawning in E. superba is not restricted to one short period, but is spread over most of the southern summer, with the result that new broods of larvae are continually being hatched out, and the stock is constantly replenished. Taube (1915) and Ruud (1936) found that in northern waters, Nyctiphanes and closely allied euphausians had a similarly extended spawning season. Consequently, eggs, larval forms, adolescents in every stage of development, and adult individuals can be, and frequently are, found to exist side by side, and the euphausian population presents a very heterogeneous appearance. The larvae of E. superba after one year of growth have attained by the following November an average length of 13 mm. Their subsequent development from adolescence to maturity forms the subject of this paper. ' Revised by Hendey (1937) and now called Fragillariupsis antarctica. io8 DISCOVERY REPORTS ADOLESCENT KRILL i Of the 8029 specimens of E. superba which I have examined, 6006 were adolescent and of these 3073 were males and 2933 females. The youngest adolescents first make their appearance in any number in August ; they show no trace of external sexual characters, but internally the reproductive system is recognizable, and by dissection under a low-power binocular microscope the sex of each individual can be determined. I have described the development of the reproductive system in the short paper forming an introduction to this one, which has been published in vol. XiV of the Discovery Reports. It will therefore be sufficient to summarize this development here, before discussing how growth proceeds during adolescence. Ruud (1932) has drawn attention to the fact that, in E. superba, investigation of the testis and ovary is the only reliable method of determining maturity, and that the reproductive system of each in- dividual must therefore be examined before the composition of any specific population can be estimated. After examining the euphausian material obtained during the cruise of the S.S. ' Vikingen', he dis- tinguished four stages of maturity in both males and females. These are listed in Table 2. Table 2 (after Ruud). Stages of maturity Males Females 1 No spermatophores visible in the ejaculatory duct 1 Ovary small and immature. Eggs o-i 0-0-25 mm. (depository) in diameter 2 Visible spermatophores : not loosened by touching 2 Ovary large but immature. Eggs o-26-o-5o mm. with a needle in diameter 3 Visible spermatophores : loosened when lightly 3 Ovary large and mature. Eggs 0-5 1-070 mm. in touched diameter 4 Empty ejaculatory ducts. Mating has recently 4 Ovary small, mainly germinal layer. Eggs 0-54- taken place 0-65 mm. loose in thorax. Spawning has taken place Ruud states that he does not know of " any practical method by which the degree of maturity of the testicle can be ascertained". Consequently, he included within stage 1 all those male euphausiids which were not fully adult (i.e. all those with no visible spermatophores in the ejaculatory ducts), and he found that the specimens showed a very wide range in length: i6-6-44-4 mm. As a criterion of development in the females, he used the diameter of the egg, and again found great variation in length in the specimens included within stage 1. It is clear that this first group of Ruud's, comprising as it does males and females of such difi^erent size, covers the whole period of adolescence, during which time the reproductive system becomes mature. A closer investigation of these adolescent forms has thrown more light on the development of E. superba. ADOLESCENT MALES Although during adolescence the individual growth rate varies very considerably, five stages can be distinguished in the development of the male sexual organs, both internal and external. These stages Table 3. Growth stages in the male Stage 2 3 4 5 Internal structures Primitive condition. Small testis, simple uncoiled vas deferens Small posterior flexure appears on vas deferens Lateral pocket appears on posterior fle.xure Anterior flexure appears on vas deferens Coiling on vas deferens near posterior flexure Stage B C D E External structures Undifferentiated ist pleopod Petasma appears as an undivided lobe Petasma becomes divided into two lobes Wing develops above petasma Wing grows and curves over petasma THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 109 1> „ C3 •A9p "ixg 1 oooooooo 1 1 Mil tooooo 1 1 1 •luiu ijiSuaq 1 r^u-iLnt^oooooK-il 1 '^ 1 u-i u-t u-t yj~t ui\o 10 10 1 1 "•I 1 1 1 1 vS vO mvD mill CiO - CD 4-. C! •A3p -ixg loooooooo 1 1 1 MM ooooo 1 1 1 « 1'S " ■uim ijiSuaq; 1 N Ln o^ "^ :i :i'=2.°2. 1 1 1 - 1 1 1 lu-iLn-ct-rhTl-l 1 1 bD - B CO to JJ •A3p -1X3 E - [iHpHtafeOlt^OtHl 1 ! 1 1 1 OOfeOOO 1 1 1 M.S - •turn m§u3q •< ' 1 1 lOOOOsOOOOl 1 1 ■A3p -jxa WfcOOfet^f^t^f^l 1 1 1 1 1 owooow 1 1 1 ■uiui qiSuaq OS r- i^ CO i2'=2.°S. 1 1 1 E ' ■♦-» < 1 1 |vDMOOC^J^;-C2| 1 1 1 1 |Tt-r'>c^'l--^T)-| 1 1 IT) is " •Asp ixg CO E ■4-* a 8 CO "o -a <1 WfeWMfeP^feWpH 1 |W WWOt^feOC^I 1 1 1 1 •urni qjSuaq 00 w 10 N ro "^00 1 1 00 ^ ^ -"l- ^ "1 "^^ in ir> 1 1 ^ ly-l osvO M CO " m 1 1 1 1 1 «5 (U •A3p 'ixg QPlJnWWQt^Wta |WW W|W00QQQ| 1 1 1 t»3 "a ••Kt •UIUI ij:iSu3q rorOfOc^-^-^-^f^'+l •^-*- N| M o m^^ "-100 o 1 NNNNNMNNNNcnl "3 •a < M ro iri M O « ■* "^^ sO O sD (SiNNNNC^NNNNC^N ^—1 *-> u p i; " CO •Aap •jxg pq- t^ b^^'^ ^ !i, ^ 1 (SJ o o V rt ^ -I *« G m ^d £ Ji ^ oi ^ ■■ •a o c G C3 -i J3 " , — ^ T- Ln U •_^ 'e u O- ^ U-t o w s s < -^ 3 « * 60 r C _c .C "^ u ^ [« Q u s o a. G o JS J3 ; o - J3 v 1> 60 J, 60 60 y rt rt V S u u > o ^ < < ex G 2 o 5 ^~' ■-^ iS U c5 < m ; h; o s 60 V b 3-2 DISCOVERY REPORTS ■-5 5a O ^ > 3 S? 1 1 8 S §- 1 S 1 t^ m 1 1 1 1 1 N r^ 1 1 1 1 o 1 1 „ „ N 1 « m 1 '^°° 1 1 1 1 »H u C 3 1—1 55 1 « N vO II ■* o O 1 1 1 1 3 o 1 ^2 ;t" 1 1 ■+00 "111! N O 1 1 1 1 oo OS s? M N CO CO 1 1 1 N lO -H III so -J- O I 1 1 1 r*-; »n M 1 1 1 1 'a +-> o - =^N ^^ M 1 %■ N Tl-O 1 1 1 1 N ro 1 1 1 1 u < sO r'. O -t- 1 M vD oo -M o to lO u-l i-i N t-H -^ N " N O ^ ■-. m oo m OO M o O 1 OO r^ m N 1 vO so oo M 01 in CO Ln -:j- N in ^ M i o^ o in en I -1- N 1 1 "^ 1 N "(3 o H m t^ O " C^ -^ "-. vO CO M -i- " "- mo O oo M r^ M m M N M N as N in so m \0 O ^ M O " N lO ■o m Th 1 M N m M 1 Tt- OS 13 o N \0 "^ -t- N " m M N Tj- ^ [ o « m o N 1 - ^s oo 00 d E r" N M M M o mH O M m 01 01 o H O Ti- N M M ro N in in N m M 01 ^O OS Os oo d S^ C> m Ti- Th W M CO -i- N O M M sO N O M ■M o >-' N m r^ -^ r^ CT^ O oo N rJ \o N OO M N rt- ■+ t-H 1— 1 m M in m ^ < > 1 s ■g. cQ < UJ -!3 Lu C C3 ;7 a < ii 1 2 Q fa ■T3 c ^ -a ■-n ^ c s •^ c3 o. H « n ( ) c/5 < 2 1) -o c a 'r?5 o a. D:i 2 Q_ d u < 6X3 -1 Qi C C (1) < 3 2 3 OJ ■u j:: CD (U UJ -Q 45 U_ h- T" 1; n 6 z 1- Q ( J , o feb 1— fa Q_ h U) J1 in en Cvl in in in in S3yx^^Nl"^ll/^ S3yi3i^nniiAi 126 DISCOVERY REPORTS eggs, therefore, the distribution in time of each developmental stage may be spread over a period of three months. Thus, eggs spawned in February may become adolescent in the foUow^ing September, October or November, that is, specimens at stage 1 met with in these months may have originated from eggs laid in February. Similarly, eggs spawned in March may become adolescent in the following October, November or December, or alternatively, specimens at stage 1 in these months may have developed from eggs laid in March and so on. The months, in which the other developmental stages may be theoretically expected to occur, can also be worked out, by assuming that each stage lasts 2 months in the male and zh months in the female. In this way, an explanation of the heterogeneous composition of the euphausian population is obtained, and in Fig. 2 I have attempted to give some idea of this complexity at any given time by a diagrammatic representation of the generations arising from the three batches of eggs, spawned in Table 17 I Stage 1 Stage 2 Stage 3 Stage 4 Stage 5 Stage 6 Stage 7 Males: time interval : 2 months End of Sept. End of Nov. End of Jan. End of Mar. End of May End of July End of Sept. 14 mm. End of Oct. 27 mm. End of Dec. 42 mm. End of Feb. 47 mm. End of Apr. End of June End of Aug. 54 mm. End of Oct. 15 mm. End of Nov. 28 mm. End of Jan. 43 mm. End of Mar. 46 mm. End of May End of July 44 mm. ■ End of Sept. 44 mm. End of Nov. 20 mm. End of Dec. 33 mm. End of Feb. 44 mm. End of Apr. 41 mm. End of June End of Aug. 48 mm. End of Oct. 45 mm. End of Dec. 23 mm. End of Jan. 33 mm. End of Mar. 40 mm. End of Mav 41 mm. End of July 40 mm. End of Sept. 43 mm. End of Nov. 51 mm. End of Jan. 23 mm. 33 mm. 38 mm. — 41 mm. 41 mm. 51 mm. Females : time interval: 2k months 1 End of Sept. Beg. of Dec. End of Feb. Beg. of May End of July Beg. of Oct. End of Dec. 18 mm. End of Oct. 31 mm. Beg. of Jan. 36 mm. End of Mar. Beg. of June End of Aug. Beg. of Nov. 57 mm. End of Jan. 15 mm. End of Nov. 34 mm. Beg. of Feb. 43 mm. End of Apr. Beg. of July 46 mm. End of Sept. 47 mm. Beg. of Dec. 54 mm. End of Feb. 20 mm. End of Dec. 32 mm. Beg. of Mar. 40 mm. End of May Beg. of Aug. 49 mm. End of Oct. 46 mm. Beg. of Jan. 52 mm. End of Mar. 23 mm. End of Jan. 38 mm. Beg. of Apr. 41 mm. End of June 39 mm. Beg. of Sept. 45 mm. End of Nov. 50 mm. Beg. of Feb. 50 mm. End of Apr. 25 mm. 33 niTi- 41 mm. 42 mm. 43 mm. 50 mm. 47 mm. February, March and April. The average length per month of each stage has been plotted, and the points marking the maximum average values have been joined up, as well as those marking the mini- mum average values. The space between has been hatched in colour: blue for the generation arising from February eggs, red for March and green for April. The months in which the stages have been calculated to occur, and the values of the average lengths are set out in Table 17. It will be noticed that in the winter months no values are given for stage 5 in the males and stage 4 in the females. In the scanty material available from this time of the year, these stages did not occur, although theoreti- cally they should be present. In the figure the actual period of time, in which each stage appears, is shown by a solid black line, the theoretical period by a broken line. The fate of early or late spawned eggs can be seen at a glance from the diagram. February eggs, which have reached stage 1 at the beginning of the following season (i.e. September to November) will have grown sufficiently to be at stage 4 or 5 before the winter sets in (i.e. June), and will be mature by October or December of the succeeding spring ; they attain rather greater lengths than the later generations. On the other hand, April eggs may not reach stage 1 until the following January, and will THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 127 not therefore be approaching maturity, that is at stage 5, until the succeeding September, nor will they be fully mature before the January (or April) after that. The overlap of the stages in the three batches of eggs and the resulting mixed composition of the euphausian population is clearly shown. The picture could be made even more complex, if the batches of eggs from every month in the whole spawning season were represented, but I decided not to attempt this, because the diagram would lose in clarity. (2) Grozvth of the etiphmisian population as a whole. By the second method mentioned on p. 120, the population was sexed, measured and divided into adolescents and adults, males and females again being treated separately. The average monthly lengths of adolescents and adults were then calculated (Table i8). Table i8. Showing average length per month of larval, adolescent a?id adult Euphausia superba Month Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. Average length per month of larvae and adolescents compiled from Fraser's paper Larvae Av. length mm. I I 4 4 S 6 8 10 10 II 13 13 16 No. of speci- Adolescents Av. length mm. 5 5 56 lOI 182 177 18 3 29 27 33 49 No. of speci- Average length per month of adolescent and adult males Adolescents Total 13 14 15 18 20 23 24 Av. length mm. No. of speci- mens 686 14 12 33 54 4 5 2 13 14 19 23 26 29 36 38 39 33 39 40 34 34 Adults Av. length mm. No. of speci- mens Average length per month of adolescent and adult females Adolescents Av. length 124 19 19 356 316 437 304 778 330 368 40 39 4 146 41 3197 44 47 44 45 51 51 52 52 SI 12 44 147 39 176 133 164 19 15 13 18 18 21 23 25 31 35 34 31 39 35 32 27 No. of speci- mens Adults Av. length No. of speci- Month 749' 28 30 356 293 351 252 660 415 309 62 93 II 176 21 3057 40 42 41 41 45 48 51 50 47 36 34 200 96 182 142 221 150 212 1273 I adult in July neglected. Larvae = nauplius to 6th furcilia. Adolescent males = stages 1-5 ( + Fraser's adolescents). Adolescent females = stages 1-3 ( + Fraser's adolescents). Adult males = stages 6-7. Adult females = stages 4-7. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. Total 128 DISCOVERY REPORTS To obtain a complete growth curve, monthly averages of the larvae must be included. Using Eraser's measurements, I recalculated these averages, which had been originally worked out on a half- monthly basis, and I also made monthly frequency tables of his measurements of young adolescents. These larval averages show the rate of growth during the first six or seven months, but when the adolescents make their appearance in August, it is not sufficient to work out the average length of all adolescents per month. Some selection is necessary, because early in the southern spring, that is, in August, September and October, the overlap of generations brings about the co-existence in the catch of young adolescents of stage 1 with late adolescents of stages 3, 4 and 5 of earlier generations, and the inclusion of these larger adolescents in the calculations gives a wrong idea of the growth rate in these particular months. Later in the season, the population becomes more sharply divided into adolescents and adults, and the question of selection does not arise. I have, therefore, included in the calculations for August, September and October, only the measurements of Eraser's adolescents and of my own specimens at stage 1 . In calculating the average lengths of the adults, I have used all specimens which could be expected to mature within the southern summer, that is stage 6 as well as stage 7 in the males, and stages 4-7 in the females. In Eraser's original graph of larval growth, there is a marked decrease during the winter months, June, July and August. This tends to disappear when his results are combined with mine (Fig. 3), and may have been due, in part, to scarcity of material. I think, too, that the apparent slowing-up of growth during the second winter, in the transition period between adolescence and maturity, can also be partly explained on these grounds, although the colder temperatures and less abundant food almost certainly have some retarding effect upon the growth rate. Before one year's growth is over, that is, as soon as the adolescents appear in August, it becomes possible to distinguish between males and females, and the curve can therefore be divided into two parts (Fig. 3). The rate of growth in the two sexes is very similar. Although the females are con- sistently smaller than the males, the two curves follow approximately the same course. The period of adolescence occupies, at a minimum, a whole year and is shorter in the males than in the females. In the males, true adults, carrying fully formed spermatophores, appear for the first time in September. In the females, true adults, fully gravid, appear three months later in December. The total period of growth from the egg to the adult occupies a minimum of twenty-two months in the male, and twenty- five months in the female. FACTORS INFLUENCING GROWTH RATE Obviously, the main factor which influences the growth rate of E. superba is the supply of food. Hart (1934) writes that this "consists very largely, if not entirely of diatoms and other phytoplankton organisms". He found that the most strongly silicified diatoms could be identified with certainty in the stomach contents, but that those with thinner cell walls were too rapidly digested to be easily recognizable. In a later paper (1942), Hart discusses the factors which control the production of phytoplankton in the Antarctic zone as a whole. He states that chief among them are the physical influences of "light, the degree of stability of the surface layers and the (interrelated) effects of the pack-ice", and that these three agents are "certainly the prime causes in determining the time of the onset of the main increase" in the abundance of phytoplankton. This time, " falls later in the year as one proceeds southwards ", as much as two months elapsing between its occurrence in the northern and the southern regions of the Antarctic zone. However, Hart considers that none of these factors " adequately accounts for the vastly greater richness of the neritic areas as compared with the oceanic regions". Recent THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 129 work strongly indicates that this is due to minute traces of organic compounds, iron and manganese, derived from the land, which exert a "strongly favourable influence on phytoplankton production . The importance of the pack-ice in this connexion is strongly emphasized. Hart regards it as "giving rise to what might be termed pseudo-coastal conditions at vast distances from land, where neritic species maintained by the ice flourish for short periods when the latter disperses ". All these factors, since they influence the food supply, must have a bearing on the growth rate of E superba but it is not a simple matter to produce evidence in support of this. Only an unbroken series of observations extending over several seasons and made at short regular time intervals at the same stations would provide reliable data. Unfortunately such a series is not available, the material collected being too scattered and interrupted, so that there are many gaps in the chain of evidence and attempts at correlation are always breaking down. 1 •so 50 40- "30 20 ..^^.•^- \-':> \ U''^ V^o',.. Larvae -* A: — a — ^ — Males — ■ — ■ Adolescents ■ ■ Adults IS 12 <-J Females Adolescents a a Adults B a B 50 40 ■30 E 2Q Notf First Year ot browth I Fig. 3. Growth curve showing average length per month of larvae, adolescents and adults. As the diatom maximum occurs earlier in the northern region of the Antarctic zone than in the southern the average development of the northern E. mperba should be correspondingly more advanced. But Fraser found no clear indication that larval development begins sooner in one area than in another, though he obtained some suggestions that local variations in the abundance of food may directly affect the average larval length. He did not feel justified however in concluding that within a restricted area, food was the only factor involved, but decided rather that the effect was the cumulative result of several factors acting locally. ^ ' ■ a . Comparison of the size of larvae, adolescents and adults fron, the different Anrarcc regions do not give a satisfactory result either. This may be due to the fact that length alone ts not a rehable criterion of development, but even if the developmental stages are taken mto account the evidence s n"t more definite From the material available, it cannot be shown that older adolescents or fuHy mature aduks occur any earlier in the northern Antarctic region than in the southern. But on Ae other hand, there is evidence to show that those E. ^perbo, which are hatched early anywhere wthm "arc ic zone, are directly intfuenced by the abundance of the food supply. Hart points out h he summer decrease in phytoplankton may be due in part to a ^^-^^-^^ ^^^\°^XXX this is probably brought about to some extent in the oceanic areas, anyhow, by intensive^ graz ng down by the herbivorous zooplankton". This occurs during and ™™*ately after the period o he spring maximum, in December, January and February. These months "made with he fi.t ha^ o^ the spawning season in E. superba, and it would appear that those generations hatched early, which I30 DISCOVERY REPORTS are able to benefit fully from the spring maximum, reach greater average lengths than those developing later, when grazing down has brought about a decrease in the food supply. This variation in size has already been mentioned (p. no), and Fig. 3 shows that it is maintained throughout the life-cycle, these larger adolescents giving rise to the very big adults, which occur at the beginning of the breeding season. The influence of the spring diatom maximum on the average lengths of the stages is also apparent, though it is perhaps more consistently marked in the males than in the females. If plotted graphically, the average lengths show a fairly rapid rise from October or November to a peak in February or March (Figs, i and 3), after which they generally tend to decrease slightly or to remain almost sta- tionary. The period of increase in length corresponds roughly with the time of the phytoplankton maximum, and the succeeding period of slackened growth corresponds with the time of the post- maximal decrease. A similar rise and fall at the same time of the year is seen in the maximum lengths of the stages. The onset of the southern winter is doubtless also a factor which comes into play at this time and influences the rate of growth, for Deacon (1933) has shown that the difference between the summer and winter temperatures of the Antarctic surface water is as much as four degrees. The other factors mentioned by Hart, light, surface conditions and pack-ice, except in so far as their broad seasonal variations will certainly influence the growth rate, more properly affect the dis- tribution of E. superba, and are outside the scope of this paper. This distribution was being worked out by my colleague, J. W. S. Marr, but unfortunately its completion has been interrupted for the time being by the war. CONCLUSIONS This investigation extends Fraser's work on the growth of E. superba from the egg to the beginning of adolescence, and amplifies Ruud's sketch of a two year life-cycle. In order to estimate accurately the composition of the euphausian population, a method was devised, by intensive study of the reproductive system, for determining the degree of maturity of each in- dividual. It was found possible to distinguish between males and females immediately the larval state was left behind, and to divide the period of their growth to maturity into 7 stages. This method gives a convenient way of checking deductions based on measurements of length alone, and has shown clearly that individual length is not necessarily a reliable criterion of development, since there is evidence to show that length may be the first thing to be influenced by variations in the factors affecting the rate of growth. Division of the population into growth stages, combined with estimations of the average length of these stages, however, gives a good idea of the life history. The spawning season, which extends over 5I months, begins in November or December. Eggs spawned then are probably adolescent by August, and mature about thirteen months later in Sep- tember and October. The males grow more rapidly than the females, attaining slightly greater average lengths on the whole, and requiring a probable minimum of 22 months to reach maturity, as against 25 months in the female. It seems likely that each state lasts 2 months in the male and 2\ months in the female, though these times are not definitely established as yet, lack of material at certain periods of the year making the evidence too scanty. Pairing was first found to take place in October, before the females were fully adult. The sper- matophores are therefore carried for some time before fertilization can occur, the evidence showing that this is effected externally, while the eggs are being laid. Gravid females are present in surprisingly small numbers. This fact, coupled with Fraser's records of eggs and early larval stages in the deeper water layers, seems to indicate that the females go down deep to spawn. More evidence on this point is needed, before this can be definitely established. THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 131 The prolonged spawning season, which is characteristic also of euphausians from the northern hemisphere, gives rise to a very heterogeneous population, the stock being continually replenished by the addition of new generations. Adults were found between the months of August and April, but when the breeding season is over, they appear to die off, being absent from the catch during the autumn and winter months. It should be mentioned, however, that after spawning, females were found in April at the surface feeding actively, a fact which does not suggest lack of vitality, and therefore, since after this month the available material becomes very scanty, their apparent absence may simply be due to lack of evidence, and not to a holocaust consequent on exhaustion after breeding. The material available is not enough to show a correlation between the appearance of the spring phytoplankton maximum in the different regions of the Antarctic zone and the precosity of develop- ment of the euphausian population, but there is evidence to show that generations hatched early in the season anywhere in the zone benefit directly from the abundance of food and the rising tem- perature of the surface layers, while later generations develop more slowly, partly no doubt because the food supply is becoming reduced by grazing-down, and partly because of the onset of the colder weather. BIBLIOGRAPHY Bargmann H E 10^7. The reproductive system of Euphausii superb^. Discovery Reports, xiv, pp. 325-50, 5 pis., 26 figs. Son; G. E: R., 1933- A general account of the hydrology of the South Atlantic ocean. Discovery Reports, vii, pp. 171-238, Eraser, R^C.r"9T6. On the development and distribution of the young stages of Mil (Euphausia superba). Discovery Reports, Hart, i:^]'.^J^2lon'the phytoplankton of the South-west Atlantic and the Bellingshausen sea, 1929-31- Discovery Reports, viii, pp. 1-268, figs. 1-84. A A <; ^^ IQ42 Phytoplankton periodicity in Antarctic surface waters. Discovery Reports, xxi, pp. 261-350, hgs. i-9- Hendey N I., 1937. The plankton diatotns of southern seas. Discovery Reports, xvi, pp. 151-364. Pls-/™. Lebou? M. v., 1926. A general survey of larval Euphmmids, with a scheme for then tdenttficatwn. J. Mar. Biol. Assoc. N.S. MacdonaTd' R^" lltr''Food% habits 0/ Meganyctiphanes norvegica. J. Mar. Biol Assoc. N.S^ xiv, pp. 753-84, 2 figs. Ottes?ad P 1933 A mathematical method for the study of growth. Hvalradets Skrifter, Nr. 7, Oslo, pp. 30-54, Ags. 24-35- RuuD T T 10^2 0«fAefo-o/oavo/w«/te-«Euphausiidae. Hvalradets Skrifter, Nr 2, Oslo pp. 5-105, 37 %». — ^^6 £«?Wac... Report on the Danish Oceanographical Expeditions 1908-10 to the Mediterranean and adjacent Sars, G. 0?ir98."b^/°S^o?a^a;fo« and early development of Euphausiidae. Arch. Math. Natur. Kristian.a, xx, Nr. 11, Taube, l^'i^ls^^Beftrdge zur Entwicklungsgeschichte der Euphausiden. Zeit. wiss. Zool. cxiv, pp. 577-656, 7 pls-, 7 figS- 132 APPENDIX Table 19. Measurements of all Specimens of Euphausia superba examined MALES FEMALES Date August 28, 1928 Locality S. Georgia St. No. WS264 Position |53°l3'-3pS Net N7oB97-om. position ^ 34°5i'.ooW Surface T. -1-65° C. Date August 28, 1928 Locality S. Georgia St. No. WS264 Position (53°"3'-3pS Net N7oB97-om. i-osition -^ 34°5i'-ooW Surface T. - lbs" C Length in mm. Stages 1 Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G II 12 13 14 22 23 24 25 26 27 28 30 32 33 34 35 36 11 39 41 42 43 45 46 47 51 • A . a . 3 1 1 1 2 . 2 . I 1 4 ■ • 5 2 . 3 5 • 2 I . 2 3 I I . I . 2 2 I I I I 4 3 I I 5 4 2 2 5 2 I I I 4 2 2 I I I 10 II 12 13 14 21 22 24 25 26 27 28 29 30 31 32 33 34 11 39 40 41 42 43 45 55 2 2 4 . 4 . I . IC 1 . . . . 2 . . . . I . . . . 21... 22... 21... I . 21.. 2 . I 2 2 4 4 I I 2 . . . 4 • ■ • 2 . . . 3 • • • 12.. I . . . I . 3 . . 2 I . . 2 2 4 4 I I I 3 10 I I I 2 3 3 I 3 4 3 3 3 I I I 3 2 I Total Av. length 5 25 4 8 13 9 • 14 25 32 32 37 44 • 18 13 5 5 14 9 • 22 26 32 32 37 44 • 64 Total Av. length 13 23 10 14 5 ■ • 12 26 33 37 46 13 8 21 13 10 . 12 25 28 35 43 . . 6s Date August 16, 1938 Locality S. of Bouvet L St. No. 2391 Position (55° 03'-3 S, Net NiooHs-om. fosmon | 00° 21' E N 100 B («°-22S m. Surface T. -132 C. \700-0 m. Date August 16, 1938 Locality S. of Bouvet I. St. No. 2391 Position |"°°o'''?.% Net NiooHs-om. I 00 21 E NiooB ,f43c^225m. Surface T. -i-32''C. (700-0 m. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 25 35 38 42 43 44 47 1 . I . . . . . 2 . . I I I I I . I . . I 2 2 I I I I 29 31 35 36 37 38 39 40 44 45 I I . . . . 1 . . . . 2 . . . . 1 . . . . 2 . . . . 1 . . . . 2 . . . . I . . . . 1 . . . . I I I I I 2 I 2 I 2 I I 13 Total Av. length . I I . 7 • • . 25 35 .41 . • .1116.. ■ 25 35 38 42 . . 9 Total Av. length I 12 . . 29 38 . . . . .265... . 30 37 41 • ■ • Date August 17, 1938 Locality S. of Bouvet L St. No. 2393 Pnsitinn /56" 42'-3 S, Net N 100 H 5-0 m. Position | ^^, ^g,.^ j. NiooBi28-om. Surface T. -l-8i°C. Date August 17, 1938 Locality S. of Bouvet I. St. No. 2193 Position |56°42'-3S, Net NiooHs-om. Position ^ 00° 38'-3 E NiooBi28-om. Surface T. -i-8l°C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 2S 30 31 32 33 34 35 36 39 I 1 2 2 2 I I I 4 4 I I I I I I 26 27 28 30 31 32 33 34 35 38 39 40 I I I 51.... 3 1 . . . . 2 II.... 12.... 2 . . . . I . . . . 2 . . . I 1 I I 6 3 I 2 2 3 2 I I Total Av. length . 3 5 6 3 • • • 30 33 34 44 ■ • • 4553- • • 30 33 34 44 ■ • 17 Total Av. length . 15 9 • • • • . 31 36 . . . . 2 12 7 3 ■ • ■ 27 31 34 39 • • • 24 ' Date August 18, 1938 Locality S.E. of Bouvet L St. No. 2396 Pn.lfmn ,;„n /54°23'ioS, Net N7oHo-5m. Position |='%5"-'44'.oo W Surface T. -108° C. Date October 2, 1928 Locality S. Georgia St. No. WS 200 „ .. f';4° 2-!''io S Net N7oHo-5m. P<>^'"°" { 35°4]'ooW Surface T. - 1 08° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 13 15 16 17 22 28 29 30 31 32 33 35 38 41 . I I 2 I 3 2 ; I I 2 4 7 ■ I 5 2 1 2 2 4 4 8 3 I I 13 14 IS l6 22 25 26 11 29 30 31 32 33 34 35 37 I 2 •1 3 2 . 2 5 ■ 4 s . 2 1 2 3 I I 2 . 4 . . 3 • . I . 2 I I 2 • • s • • s s I 1 I 2 4 3 I I 3 3 5 5 7 2 I 2 3 I I Total Av. length 9 10 10 4 3 15 30 30 34 38 . . II 7 13 2 3 ■ • 17 26 32 35 38 . . 36 Total Av. length 10 6 28 I . . . 15 28 30 37 . . . 4 22 8 I . . 5 26 29 33 37 • • 45 Date October 4, IQ28 Locality S. Georgia St. No. WS295 Pn.,;Mnn f55° 23'-40 S, Net N 100 B 97-0 m. Fosition | 34°4i'ooW Surface T. -110° C. Date October 4, 1928 Locality S. Georgia St. No. WS295 P„»;tion fss'23'-4oS. Net N 100 B 97-0 m. Position |= 34" ^r'oo W Surface T. -110° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 13 14 15 16 17 25 26 27 28 29 30 32 11 35 36 37 ( ^ ) I 4 5 4 3 . I 4 5 4 3 2 4 I 3 I 7 2 2 2 3 I 13 14 15 16 17 26 27 28 29 33 37 I I I 2 2 . 3 ■ 7 ■ 15 . 7 • 2 . I I 2 I . . 3 7 15 7 2 2 4 2 2 I I I I 2 • 3 1 I . 2 2 I Total Av. length 34 4 7 I . . . 15 27 29 37 . . . !4 I 5 5 I • ■ IS 26 27 29 37 . . 46 . . . Total Av. length 17 18 2 8 I . . «5 28 33 34 37 . . 19 15 6 5 I ■ ■ 17 28 33 34 37 ■ ■ 46 Date October 5. 1928 Locality S. Georgia St. No. WSzgS Pr„;,;„„ (S5°27'-3oS, Net N 100 B 94-0 m. 1 osition | 32°2r-4oW Surface T. -1-76° C. Date October 5, 192S Locality S. Georgia St. No. WS298 Pn^i.inn (SS°27'-3oS, Net N 100 B 94-0 m. Position -^ 32°2i'-4oW Surface T. -1-76° C. Length in mm. Stages Total in sample 1234567 A B C D E F G Length in mm. Stages Total in sample 25 11 29 30 31 32 33 34 35 36 11 39 40 41 42 43 45 46 SO 51 52 S6 I I 3 I 2 I I 2 2 2 2 2 I I 2 I I I I I 2 2 I I I 1234567 A B C D E F G 2 I 2 I I 11 30 31 32 33 37 38 39 41 43 44 46 48 49 50 51 52 54 55 i [ [ I [ I 2 I I 2 . 4 3 • 2 I I 2 I . i 3 I I I I 2 . 4 • 3 ■ . . 2 1 I . 2 I 4 8 I I 2 I 2 I I 2 I 2 4 3 2 I 1 3 Total Av. length 5 II I 3 19 3 28 30 33 38 47 52 • 7 9 I 3 19 3 ■ 29 30 33 38 47 52 42 Total Av. length . 2 14 9 10 . . 26 31 38 47 . . . . 5 10 iS . 2 . . 28 32 42 . 53 35 136 DISCOVERY REPORTS MALES FEMALES Date October 6. igaS St. No. WS 304 Net N 100 B llo-o r Locality Position Surface T. S. Georgia f54° 54'-40 S, (^ 30^* 2l''2o W -l-58°C. Date October 6, 1928 St. No. WS 304 Net N lOQ B 1 lo-o m Locality S. Georgia /54° 54'-4p S t 30'^ 21 -20 W Surface T. -1-58° C. Position Length in mm. 25 26 31 35 43 45 47 48 Total Av. length Stages 34567 ABCDEFG 26 31 35 46 26 31 2 4 35 46 Total in sample Length in mm. 26 29 30 3t 34 40 44 Total Av. length Stages 234567 ABCDEFG I 3 4 26 30 37 332 28 32 42 Total in sample Date October 16-17, 193° St. No. 453 Net N 100 B 164-0 m. Locality Position Surface T. Bouvet I. to S. Georgia /S4°05J' S, I 03° 57t' E -l-6o°C. Date October 16-17, 1930 St. No. 453 Net N 100 B 164-0 ni. Locality Position Surface T. Bouvet I to S. Georgia /54° osi' S, I 03= 57i' E -l-6o°C. Length in mm. Stages 34567 ABCDEFG Total in sample Length in mm. Stages 234567 ABCDEFG Total in sample 34 35 Total Av. length Total Av. length Date October 17, 1930 St. No. 454 Net N 70 B 192-0 m. Locality Position Surface T. Bouvet L to S Georgia /53°42'oo S, (. 04^ 42'-oo E -1-38° C. Date October 17, 1930 St. No. 454 Net N 70 B 192-0 m. LocaUty Position Surface T. Bouvet L to S. Georgia /53°42'-oo S, (^ 04^ 42'-oo E -1-38° C. Length in mm. 13 14 15 16 17 39 40 41 43 44 45 46 49 51 55 Total Av. length Stages 234567 ABCDEFG 18 2 4J 53 18 2 41 53 Total in sample Length in mm. 13 14 15 16 17 35 36 37 38 39 40 42 43 44 45 46 48 49 50 51 Total Av. length Stages 234567 ABCDEFG 2 26 I 39 41 51 3 26 36 42 Total in sample 15 4 7 Date October 18, 1930 St. >Io. 455 Net N 100 B ii6-om. Locality Position Surface T. Bouvet I. to S. Georgia f53° 55i' S, \ 04° 47' E -fS9°C. Date October 18, 1930 St. No. 455 Net N 100 B 116-0 m. Locality Position Surface T. Bouvet I to S. Georgia |53° 55*' S, I 04° 47' E -1-59° C. Length in mm. 13 14 15 16 17 32 39 41 42 45 Stages 234567 ABCDEFG Total in sample Length in mm. 13 14 15 i6 17 19 35 38 44 45 Stages 234567 ABCDEFG Total in sample 13 4 3 Total Av. length 1 1 5 I 3 39 43 Total Av. length 5 40 5 40 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 137 MALES FEMALES Date October 19, 1930 Locality St. No. 459 Net N too B 183-0 m. Position Surface T. Bouvet I. to .S. Georgia ;S5°09i'S. 1 t 02' 00' E -i-38°C. Date October 19, 1930 St. No. 459 Net N 100 B1 ,0, „„ N-oB 1 '83-0 m Locality Position Surface T. Bouvet I. to S. Georgia t 02 00 E -1-38° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 123456 7 A B C D E F G 13 14 15 16 \l 19 37 43 3 I 3 5 1 2 2 I . . . 2 3 I 3 5 3 I 3 5 I 2 2 I 2 12 ii 17 18 19 37 38 39 40 2 4 II ... . . . . 4 ■ • . . . 3 . . 2 2 . . 4 ■ ■ II . . 5 • ■ 2 2 2 I 2 1 2 . . 2 4 II 5 2 I 1 4 3 2 I 2 . . . 2 . . . I . 2 Total Av. length 17 . . I . . 2 16 . . 37 • -43 17 . I ■ 16 .37 • . 2 ■ . 43 20 I Total Av. length 26 . . 10 . IS • ■ 38 . 26 . . 5 15 . . 3 5 . ■ 36 8 38 . . Date October 20, 1930 LocaUty St. No. 460 Net NlooBlS5-om. Position Surface T. Bouvet I. to S. Georgia /56° 46' S, t oo°4irW -i-29'C. Date October 20. 1930 St. No. 460 Net N 100 B 155-0 m. Locality Position Surface T Bouvet I. to S. Georgia /56° 46' S, I oo°4iil'W -1-29° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G II 12 13 14 \i 11 I 2 I 7 3 3 I I 2 . . . 7 . . . 3 • • • 3 ■ • • I . . . I 2 I 7 3 3 I 12 13 14 IS i5 17 18 2 5 5 7 2 4 I 2 5 5 7 2 4 1 2 s 5 7 2 4 Total Av. length 26 ... . 15 ... . 26 15 26 Total Av. length 18 I 14 38 18 . . . 14 . . . I . . 38 19 Date October 22, 1930 Locality St. No. 461 D Net N 100 B 490-385 m. Position Surface T. Bouvet I. to S. Georgia f56"4l'ooS X 02" 24 -co W -1-72° C. Date October 22, 1930 Locality St. No. 461 D Net N 100 B 490-385 m. Position Surface T Bouvet I. to S. Georgia |56^ 4l'-oo S, \ 02° 24'- 00 W -1-72'' C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 12345 5 7 A B C D E F G 33 . . . . I . . ' ■ I 36 37 40 49 I . . . I . . . I . . . I . . . I I I I 1 I I Total Av. length . . . . I . . . . . . 33 ■ ■ I 33 • I Total . . . 4 ■ 3 I ■ 4 Av. length . . . 41 38 49 • Date October 22, 1930 St. No. 461 G Net N 100 B 700-560 Locality 3'5)m. Position Surface Bouvet 1. to S. Georgia /56° 441' S I 02"'2lJ'W T. -I-74°C. Date October 22, 1930 Locality St. No. 461 G Net N 100 B 700-560 (315) m. position Surface T Bouvet I. to S. Georgia /S6°44rs I 02"2li'W . -I-74°C. Length in mm. Stages Total in sample 12345 6 7 A B C D E F G Length in mm. Stages Total in sample 32 33 34 11 11 39 40 41 42 43 44 45 46 47 48 50 . . . 2 . 2 . . . 3 . . . . 4 ■ . . . 4 . . . . 8 . . I . 7 I . 12 . . . . 13 ■ ... 8 3 . 12 2 . . . 7 I . I . 13 I . . . 5 I . . . 5 2 2 I I I 2 . 2 . 3 • • 4 . . t : : 8 . . 12 13 . . 10 I 't : : I 14 . . 6 . . 6 I . 2 2 I 2 2 3 4 t 9 12 '3 II 't 7 2 2 I 1234567 A B C D E F G 35 11 39 40 41 42 43 44 45 46 tl 49 SI . . . . 2 . I 6 I 52 5 3 . . . . I 3 3 I 7 4 7 1 10 : : : : : 1 8 I 4 2 I 31 I 2 2 I 7 7 2 6 7 8 II II I 4 10 I 4 3 4 3 I 3 7 7 8 7 8 II II 5 10 5 3 4 3 I Total Av. length .... 3 38 52 . . . . 37 42 43 6 87 40 42 93 Total Av. length . 3 . 107 14 . . 41 . 40 44 I . . 38 4 I 118 3 . ♦ 41 46 . 123 5-2 138 DISCOVERY REPORTS MALES FEMALES Date October 23, 1930 Locality Bouvet 1. to St. No. 462 S. Georgia Net N 100 B go-o m. d„..-»- ("56° oi'-oo S, Position f 07° 28-00 W Surface T. -1-55° C. Date October 23, 1930 Locality Bouvet L to St. No. 462 S. Georgia Net N too B 90-0 m. p^^-^-^^ {^'^o7°'28°oo W Surface T. -1-55° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 12 13 14 IS 16 \l St 2 2 s 11 2 2 2 I 2 2 5 II 2 2 2 2 2 5 11 2 2 2 I 12 14 17 18 19 40 2 I 6 5 2 I I . . . 2 6 '.'.'.'.'. '. 5 I 2 I 2 I 5 5 1 2 I I Total Av. length 26 I 15 51 26 I 15 51 27 Total Av. length 18 . . I . . . 16 . . 40 , 18 ... I . . 16 . . . 40 • ■ 19 Date October 23, 1930 Locality Bouvet L to St. No. 463 S. Georgia Net N too B. 32-0 m. p^^.,^^^ {''°J^''sTooW Surface T. -i-8o°C. Date October 25, 1930 Locality Bouvet L to St. No. 463 S. Georgia Ne, N,ooB,32^m. p^^.^^^^ {"".o^^'s^^oo' W Surface T. -180° C. Length in mm. Stages Total in sample Length in mm. Stages Toul in sample 1234567 A B C D E F G 1234567 A B C D E F G II 12 13 14 IS 16 17 19 33 I 2 I I 9 6 8 2 I . . . I 2 I I 9 6 8 2 I I 2 I I 9 6 8 2 I 11 13 14 IS 16 17 19 44 I 2 2 10 4 2 I I . . . I 2 2 10 4 2 I . . . . I . . I 2 2 10 4 2 I I Total Av. length 22 . . I . . . IS . . 44 • ■ ■ 22 ... I . . 15 ■ • . 44 • • 23 1 Total Av. length 30 . . I . . . 16 . . 33 . • • 30 ... I . . i6 . . . 33 . . 31 Date October 26, 1930 Locality Bouvet I. to St. No. 465 S. Georgia Net N 100 B. 13-0 m. p^^-^-^^ {''',4»'o2°' W Surface T. -i-68°C." Date October 26, 1930 Locality Bouvet L to St. No. 46s S. Georgia Net N.ooBii3-om. p^^;,;^^ {"^^^''ojvw Surface T. -i-58'C." Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234367 A B C D E F G 1234567 A B C D E F G 47 50 I I I I I I 39 40 . 2 . . . I . . . . 2 . 1 2 I Total Av. length 2 49 2 49 2 Total Av. length . . . 3 . . . . . . 39 • . . . . . . 3 . ■ . . . . 39 . . 3 Date October 31, 1930 Locality Bouvet L to St. No. 471 S. Georgia Net N 100 B .6S-0 m, p^^i,;^^ |54°|7^-oo S^^ Surface T. -I-62°C." Date October 26, 1930 Locality Bouvet 1. to St. No. 464 ' S. Georgia Net N too H 67(-o) m. Pnsitinn f 56° 03'oo S, fosition -J ,3=,8..ooW i Surface T —1-75-0. 1 Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 38 I I I 48 I . . . I I Total Av. length I 38 38 I Total Av. length I . . . . . . 48 . . . I .... 48 . . 1 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 139 Date St. No. Net November 19, 1929 W.S. Alongside Deception I. N 100 B 0-5 m. Locality S. Shetland Is. FEMALES Date November 19, 1929 St. No. W.S. .Alongside Deception I. Net N 100 B 0-5 m Locality Bransfield Strait Surface T. circa 0-15° C. Length in nun. Stages 234567 ABCDEFG Total in sample Length in mm. 1234567 ABCDEFG Total in sample 42 44 47 34 38 48 Total Av. length S 44 5 44 Total Av. length 4 I 37 48 38 . 41 Date November 13, 1930 St. No. 480 Net N 100 B i6i~o m. Locality S. Georgia Position {"39°54'^W Surface T. -0-58' C. Date November 13, 1930 St. No. 480 Net N 100 B 161-0 m. N 70 V 1000-750 m. Locality S. Georgia Position {";*.i;?w Surface T. -058° C. Length Stages Total Av. length 234567 ABCDEFG Total in sample Length in mm. 13 46 Total Av. length Stages 234567 ABCDEFG 46 . 46 Total in sample Date November 16, 1930 St. No. 484 Net N 100 B 73-0 m. Locality Position Surface T. S. Georgia /S3° 52i' S, I 37° os4' W -056° C. Date November 16, IQ St. No. 484 Net N 100 B 73-0 m. Locality Position Surface T. S. Georgia f53= 52I' S, l 37° osi' W — 0-56° C. Length in mm. 13 15 16 17 18 Total Av. length Stages 234567 ABCDEFG Total in sample Length in nun. Date November 18, 1930 St. No. 492 Net N 100 B 148-0 m. Locality Position S. Georgia /53° 12}' S. I 37°04i'W Surface T. — 0-35° C. Length in mm. 14 15 16 23 24 Total Av. length Stages 34567 ABCDEFG 23 5 19 24 27 I 20 24 Total in sample 13 14 15 16 17 18 Total Av. length Stages 1234567 ABCDEFG Total in sample 36 Date November 18, 1930 St. No. 492 Net N 100 B 148-0 m. S. Georgia /53°i2rs, I. 37°04i'W Surface T. -035° C. Locality Position Length in mm. 14 15 16 17 Total Av. length Stages 234567 ABCDEFG sample 140 DISCOVERY REPORTS MALES FEMALES 1 Date November 19, 193 Locality S. Georgia Date November 19, 193 Locality S. Georgia St. No. Net N 100 B 160-0 m. Position {5^;i4;f.v Surface T. -0-85° C. St. No. Net 494 N 100 B 160-0 m. Position {5^;5°i'^S,^ Surface T. -0-85° C. Stages Total Stages Total Length in mm. m sample Length- in mm. in sample 12 3 4 5 6 7 A B C D E F G I 2 3 4 S 6 7 A B C D E F G 33 I I . . . I 36 I I I 35 I I 37 . . . . 2 I . I 2 38 I I 38 2 2 . 2 39 I I I 39 2 2 . 2 40 I 2 • I 3 4 40 .... 7 ....6.1 7 41 1 1 I 2 41 . . . . 3 . . . . 3 . . 3 42 3 ■ 3 3 42 .... 6 . . . . 5 ■ I 6 43 2 2 2 43 .... 8 .... 5 I 2 8 44 I I . 2 2 44 .... 7 . . . .3.4 7 45 6 b 6 45 .... 8 • • • • 5 - 3 8 46 4 • 4 4 46 . . . . 5 . . . . I 1 3 5 47 2 2 2 47 . . . . 3 I 2 3 48 4 4 4 48 I I I 51 I I I 49 . . . . 5 ....14. 5 52 I I I 51 I 1 I 55 I I I — — Total .... 61 . . . -35 9 J7 61 Total • 3 29 . I I 2 32 36 Av. length ■ ■ ■ ■ 43 . . . . 42 48 44 Av. length 4< 3 45 1 Date November 29, 193 D Locality S. Georgia Date November 29, 193 3 Locality S. Georgia St. No. Net 523 N 100 B1 ,„_ N70B 1 '57 Position {553°«rS,^ St. No. Net 523 N 100 B 157-0 m. Position {";°-'^9rW Surface T. -030° C. Surface T. -030° C. Stages Total Stages Total Length in mm. — in sample Length in mm. in sample 1234 f > 7 A B C D E F ( 1234s 6 7 A B C D E F G 18 19 2 . . . 4 . ■ • 2 4 2 4 15 17 2 . . . . 2 2 20 15 . . . 15 19 20 7 . . . . 20 . . . . 14 ... . 7 7 20 21 22 16 . . . 12 . . . 16 12 16 12 14 8 . . . 8 8 13 • • ■ . 12 ... . 13 24 10 . . . 10 10 12 24 13 ... . :8 . . . . ]l 26 3 • • ■ 3 3 18 27 9 . . . 9 9 26 10 ... . 10 10 28 5 . . . 5 5 I 27 28 6 . . . . 6 6 30 I . . . I 4 - . . . 4 4 Total 95 • • ■ 95 95 41 I I Av. length 23 . . . 23 45 48 I I 50 I ^ • I Date November 6, I 1)32 Locality Bransfield Strait Total 121 ... 5 I 121 ... 2 4 , 127 St. No. Net 1009 N 100 B iio-o m. Pos.,ion {%irll'^ Surface T. -0-85^ C. Av. length 23 . . . 45 48 . 23 . . . 45 46 . Length in mm. Stages Total in sample Date St. No. Net November 6, 1932 1009 N 100 B iio-o m. Locality Bransfield Strait Position {^^;|=5VS,^ Surface T. -0-85'' C. 12 3 4- « 7 A B C D E F G 16 17 18 20 21 2 . . . 4 . • • 2 2 4 4 Stages Total 2 . . . I 2 I 2 I Length in mm. in sample 1234567 A B C D E F G 7 3 ■ • 10 23 2 2.. 4 4 17 2 . . . . 2 2 24 I . . . I 19 I . . . I . . . I ^1 . 9 . . 3 6 . 9 20 3 ■ • • 3 . . . 3 26 5 . ■ 1 4 . 5 21 71.. 7 I ■ . 8 ^l 8 . . 3 5 • 8 22 5 • ■ . 41.. 5 28 4 ■ • 4 • 4 23 44.. 5 3-. 8 29 I 1 1 24 21.. 12.. 3 30 2 . I 3 ■ 3 25 153. 153- 9 31 I 1 26 2 2 2 . 2 4 32 I 27 • 4 I I .312 6 34 I 28 I 2 I 2 3 36 I 5 3 3 6 29 2 2 2 37 I I 1 30 . I I 2 2 2 4 38 • 3 1 2 3 31 ... 2 2 . 2 39 2 - 2 I 2 5 32 2 2 . 2 40 • 4 • .42. 6 34 I I I 41 2 : .4.1 5 37 . . . 3 2 . . . . 5 . . 5 42 2 . 2 . I 3 38 2 2 2 43 , I I 39 2 2 2 44 . I I 40 I I I 45 . . 2 I 3 43 I I I Total 19 35 3 21 1: . : 32 23 7 18 7 s 92 Total 25 18 8 15 8 24 17 7 10 16 . . 74 Av. length 22 26 34 38 4 4' \ • 22 27 35 39 42 4J Av. length 21 25 27 31 39 21 25 27 28 36 . . THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 141 MALES FEMALES Date December?. 1926 Locality S.Georgia St. No. Govt, jfetty, Gritvyken Surface T. circa 4-65 C. Net N 100 H o-i m. Date December 7, 1926 Locality S.Georgia St. No. Govt. Jetty, Gritvyken Surface T. circa 465 C. Net N 100 H O-I m. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 11 27 28 29 30 31 32 34 22 I 2 4 3 26 • 3 I I . 2 22 . I II 3 4 3 5 .3 I I . 2 4 I 2 7 8 3 z I 2 23 24 25 26 27 28 29 30 31 33 35 2 2 5 2 3 6 4 1 I I . 2 I 11 • 5 . 2 • 3 .6 . 22.... I I 2 I 2 5 2 I 4 I I I 28 Total Av. length : 9 20 7 29 ' 9 20 29 8 29 Total 28 Av. length 28 . 24 4 . . . . . 28 29 . . . . Date December II, 1926 Locality S Georgia St. No. King Edward's Cove Surface T. circa 4-65 C. Net N 100 H o-i m. Date December 11, 1926 Locality S.Georgia St. No. King Edward's Cove Surface T. circa 4-65° C. Net N 100 H 0-1 m. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 23 25 26 V, 29 30 31 32 33 34 35 36 37 38 39 I 2 . I I 4 4 ■ 1 2 . J 3 • 2 I I 3 • • 7 • ■ 3 I • 3 • 2 . 3 5 • . 3 • • 4 • I 2 . 4 • 7 ■ • 4 I ■ 3 • . 4 ■ • 3 I I I I 2 2 8 3 4 3 4 7 5 3 4 3 I I I 24 25 25 27 28 29 30 31 32 34 37 2 4 6 9 4 5 8 3 3 11 I II.... . 3 I • ■ • • . 5 I . • • ■ .36.... 22.... . 4 I . • • • ,26.... 12.... . 21.... II.... I . . . . 2 4 6 9 4 5 8 3 3 2 I I 2 1 Total Av. length 46 I 28 34 . 24 23 . . . . . 28 29 . . . . 47 I . . . . Total Av. length 13 30 7 2 . . . 27 31 36 35 • • • 6 45 1 ■ • • ■ 27 31 33 . ■ • • 52 Date December 19, 1926 Locality S. Georgia Ifet'"'- ^^LoH7om, P-«- {''l^zl^-fo^ Surface T. 1-45° C. Date December 19, 1926 Locality S. Georgia St. No. 125 Position (53 28 30 S, Net N 100 H 70 m. l-osmon | 36" 2o'-3o W Surface T. 145 C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 40 41 43 44 45 46 % 50 51 52 53 54 55 56 57 I I 2 . . • 4 I . . ■ 5 . • . 5 . I . 5 . . . 6 . . ■ 4 . ■ . 3 . . . 3 . 2 I . . I . I I I I . . . . 5 . . . I 5 . . . . 6 . . . . 4 . . . . 3 . . . . 3 . 2 2 I I I 3 2 4 1 5 5 6 6 4 3 3 2 43 44 45 46 47 48 49 50 51 52 54 55 S6 57 58 60 61 I . . . I . . . . 3 • ■ I I 2 . . . . . 4 . . . . . .41- 2 . . 2 1 . . . . 6 . . . . . .49- I 2 . I . I 1 I I 1 I I I . I . . . . 1 . 3 . . . . I . 4 2 I . 2 . . . . I . 5 2 II 3 2 I I I 3 I I 2 4 5 2 3 6 13 3 2 1 I Total Av. length . . I 3 I 4 40 ■ ■ 40 43 52 43 52 . . 2 . 2 3 42 . .41 .45 46 51 49 Total Av. length . . . I 33 14 2 . . . 43 51 55 59 . . . . 7 5 38 . . . . 48 55 53 50 Locality S. Georgia Date December 11, 1930 Position {'O'lor'w St No s,zn *■ 34 29J w Net N 450 H 122 (-0) m. Surface T. 0-35° C. Length in mm. Stages Total in sample Date December II, 1930 Locality S.Georgia &'et''°- ^n5oH.32(-o)m. P--" {'^^"I^VW Surface T. 0-35° C. 1234567 A B C D E F G 14 15 i5 17 18 19 20 21 22 24 26 28 42 I . . . . I I I 3 4 4 4 2 3 21 I 1 I 1 I 3 4 4 4 2 3 3 I I I Length in mm. Stages Total in sample 1 . . . . 3 ■ . ■ • 4 . . . . 4 . . . . 4 . . . . 2 . . . . 3 ■ • • ■ 3 . ■ • • 1 1234567 A B C D E F G 16 18 19 20 21 23 24 5 2 4 2 I 4 I 5 2 4 2 I 4 I 5 2 4 2 I 4 Total Av. lengtl 19 1 19 19 19 19 Total Av. lengtl 28 .... I • 20 . . . . 42 . 27 I .... I 20 24 . . . .42 29 142 DISCOVERY REPORTS MALES FEMALES Date December 13, 1930 Locality Net"'" ^'°ooB,6S-om., P°-'-" 450-168 m. Surface T. S. Sandwich Is. f 55° 32J' S I 33"I4'W -O^S'C. Date December 13, 1930 Locality ^'et^°- ^^!ooB,68-om., "-'•- 450-168 m. Surface T. S. Sandwich Is. /55'32rS, I 33= I4'W -0-95° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 16 21 23 25 26 28 29 30 31 I I 3 '■ I 2 I 12 I 1 I 2 I I I I 3 I I 1 2 1 20 25 26 27 28 30 32 I I 2 4 I I I I I 2 31 I I I 1 2 4 I I 1 Total Av. length II 27 7 4 26 28 II Total Av. length 12 26 48 21 28 12 Date December 14, 1930 Locality gfet"""- 'N'tooBz64-om. P^""" Surface T. S. Sandwich Is. fS7° 27' S, \ 34°25'W -0-90° c. Date December 14, 1930 Locality Net'"'- ^^ 100 B, 64-0 m. P--" Surface T. S. Sandwich Is. /57° 27' S, 1 34°2S'W — 0-90° C. Length in mm. 1 Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 25 27 30 I I I I I I J 21 22 24 25 2 2 2 I 2 r I 2 I 2 2 2 Total Av. length 3 27 • 3 .27 3 Total Av. length 7 23 61 23 22 7 Date December 17, 1930 Locality St. No. 534 Position Net N 100 B 172-0 m. l^osition Surface T. S. Orkney Is. f6o° 08' S, 1 47°53'W ois" C. Date December 17, 1930 Locality Net"""- ?l1ooB,7.-om. ?--" Surface T. S. Orkney Is. /6o'> 08' S, 1 47°S3'W 0-15° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 47 49 51 53 I I I I I I I I I I I I 41 I I Total Av. length I 41 Total Av. length 4 50 4 50 4 Date December 18, 1930 Locality St. No. S.J5 Position Net N 70 B m. Surface T. S. Orkney Is f6o° I3i' S, \ 50° 51 J' W o-ds" C. Date December 18, 1930 Locality Net''"- ?."7oBom. ?--" Surface T. S. Orkney Is. (60° I3i' S, I SO°5li'W 0-65" C. Length in nmi. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 41 42 44 45 tl 49 50 51 52 S3 54 11 I I I 2 I I 9 I 3 6 I 7 5 10 1 I I I 2 I I 9 I I I 2 I I 9 4 6 I 7 6 10 I 37 44 47 2 I I I I I Total Av. length 4 :::::: ^ I 7 6 zo I Total Av. length 2 49 SI 51 SI 51 51 Date December i8, 1930 Locality fj^t^"- 'N^':ooBt.2-om. P-"™ Surface T. S. Orkney Is. f6o° 43' S, 1 s2°29i'W -0-30° C. Length in mm. Stages Total in sample Date December 18, 1930 Locality ^'et"""- ?.^'ooB 122-0 m. P-'- Surface T. S. Orkney Is. f6o° 43' S, 1 52°29rW -0-30° C. 1234567 A B C D E F G 35 42 50 51 54 55 Length in mm. Stages Total in sample 1234567 A B C D E F G 36 I Total Av. length 6 48 6 48 6 Total Av. length I . . . . 36 . . 36 1 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 143 MALES FEMALES Date St. No. Net December 19, 1930 Locality N''.ooB.37-om. P°^'"°" S. Shetland Is. /6i"o7i'S. I 54°26-W Date St. .Xo. Net December 19, 1930 537 N 100 B 137-0 m. Locality Position S. Shetland Is. f6i°07*'S. 1 54°26'W 001° C. N 70 V 500-250 m. Surface T. o-oi' C. N 70 V 500-250 m Surface T. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G ■ 2 3 4 5 6 7 A B C D E F G 17 6 6 . . . 6 16 6 . . . . 6 . . 5 19 6 6 6 20 12 ... . 12 . . . 20 24 24 24 21 36 ... . 36 . . . 36 24 24 24 22 30 ... . 30 . . . 30 36 22 23 18 18 18 18 18 23 24 36 ... . 30 ... . 36 . . . 30 . . . 24 42 • 42 42 25 12 ... . 25 36 18 36 18 . . 54 29 I . . . . I . . . 3t 12 6 6 6 6.. . 6 . . . 6 . . 12 6 6 Total 163 ... . 163 .. . 163 44 49 I I I Av. length 22 ... . 22 . . . . * Total 198 18 . . . .2 180 36 . . . 2 218 Av. length 23 25 . . . .47 23 27 . . • • 47 Date December ig, 1930 Locality S. Shetland Is. Date December 19, 1930 Locality S. Shetland Is. St. No. Net nLoB 137-0 m. P°^'*i°" f6i°2g'S, \ S4°44rW St. No. Net 538 N 100 B 137-0 m. Position (•6l°29'S, \ 54° 44i' W Surface T. -0-25° C. Surface T. -025° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 12 3 4 5 5 7 A B C D E F G 17 I I . . . I 18 I . . . . J 18 I I I 19 I . . . . I 19 I I I 20 I . . . . 20 2 2 2 22 I . . . . 21 2 I 2 23 3 . . . . 3 22 3 2 3 48 I I 24 27 46 I I 1 1 I I Total 7 ... I 7 . . . . . I 8 50 S3 S6 I Av. length 21 ... 48 21 . . . . . 48 I I Total II I .... 4 9 3 • • 16 Av. length 21 27 . . . .51 21 23 . . • • SI Date December 19, 193 Locality S. Shetland Is. St. No. Net flooM-"-- Position Surface T. ;6i°48'S, I 54° 514' W -0-30° C. Length in mm. Stages Total in sample Date St. No. Net December 19, 1930 Locality S. Shetland Is. /6i°48'S, I 54°Sii'W 12 3 4 5 5 7 A B C D E F G N 100 B/ '37-0 m. Surface T. 12 I . . . . , I -030° C. 17 z . 2 2 18 4 2 Stages Total 19 2 . 2 Length in mm. in sample 21 22 23 4 . 9 . 6 . I 4 9 6 I 4 9 6 1234567 A B C D EEC 12 I I 28 I I 16 2 2 30 I 2 18 3 3 31 I 2 19 I I 32 : I 3 20 21 3 3S 2 21 2 2 36 I 22 2 2 4 37 I 23 62 8 38 5 24 I I 39 2 27 I 40 I 28 I I 41 2 35 I 1 42 4 n I . . . . I I 43 I 38 I I 44 I 4S . I . . . . I 45 I *1 2 2 46 I 48 . . I . . I 2 47 I 49 3 3 48 I SI 2 . . 2 2 50 I 52 3 • • 3 3 51 I 53 I I I 52 . 2 2 54 2 . 2 2 55 2 . 2 2 Total 20 7 2 I .16 13 15 • I . I 16 46 Total 28 2 I I 8 25 2 29 2 I A 17 14 07 Av. length 20 23 41 48 . . 49 20 23 . 45 • 48 49 Av. length 20 27 28 38 38 41 55 20 29 38 3! 40 4S 144 DISCOVERY REPORTS MALES FEMALES 1 'Date December 19, 1930 Locality N 70 V 500-250 m. Surface T. S. Shetland Is. /62° o6i' S, 1 S5°o8i'W -0-48° C. Date December 19, 1930 St. No. 540 Net N 100 I! 155-0 m. N 70 V 500-250 m. Locality Position Surface T. S. Shetland Is. (62° o6i' S, l. 55°o8i'W -048° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567ABCD E F G 123456 7 A B C D E F G 21 48 51 S3 I I 2 I . . . I 2 I I I 2 I 20 32 34 2 I 2 . . . 2 1 I I I Total Av. length 2 ... I I . 20 . . . 32 34 . 2 . . . 20 . . . . 2 • • 33 4 Total Av. length I 4 21 SI I . . . 21 . . . • • 4 • . 51 5 Date December iQ-20, 1930 LocaHty Net""" ^*",ooB,o8-om. P-"- Surface T. S. Shetland Is. (bz° 22' S, 1 55°23'W -0-85° C. Date December 19-20, 1930 Locality St. No. 541 „ . . Net N 100 B 108-0 m. Posmon Surface T. S. Shetland Is. f62° 22' S. I 55°23'W -085° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 123456 7 A B C D E F G 23 I I I 18 19 21 23 25 27 32 44 Total I I 23 1 Av. length .23 I I Total Av. length 6 ... I I . 22 . . . 32 48 . 6 . . . 22 . . . . 2 . . 38 S Date December 20, 1930 Locality Surface T. Bransfield Strait /62° 16' S, \ 57°2o'W 030° C. Date December 20, 1930 St. No. 543 Net N 100 B 178-0 m. Locality Position Surface T. Bransfield Strait /^62° J 6' S, I 57" 20' W 0-30° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample ■ 2 3 4 S 6 7 A B C D E F G 123456 7 A B C D E F G 20 I I I IS I I I Total Av. length I 20 I 20 I Total Av. length IS I 15 I Date December 20, 1930 Locality St. No. 546 n ■.■ Net N 100 B 164-0 m. P°^'^'«" Surface T. Bransfield Strait f62° 46i' S, l S7°lli'W -o-6o° C. Date December 20, 1930 St. No. 546 Net N 100 B 164-0 m. Locality Position Surface T. Bransfield Strait /62°46i'S, I 57° III' W -o-6o° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 123456 7 A B C D E F G 48 53 I I I I I 46 I . I Total Av. length 46 . I 46 I Total Av. length 2 SI 2 51 1 Date December 20, 1930 Locality Surface T. Bransfield Strait /62° 59i' S, I 57°03'W -1-02° C. Date December 20. 1930 St. No. 547 Net N 100 B 37-0 m. Locality Position Surface T. Bransfield Strait f62° 59i' S, I 57-03'W -102° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 123456 7 A B C D E F G 16 I I I 20 I 1 1 Total Av. length 16 I 16 I Total Av. length I 20 I 20 1 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 145 FEMALES Date December 21, I93< St. No. 548 Net N 100 B 102-0 m. Locality Position Surface T. Bransfield Strait /62°36!'S, 045° C. Date St. No. Net December 21, I93< S48 N 100 B 102-0 m. Length in mm. 42 43 45 46 47 48 49 SO 51 52 S3 S4 Total Av. length Stages 67 ABCDEFG 4 4 4S 48 I 34 45 50 7 35 47 50 Total in sample Date December 21-22, 1930 St. No. 549 Net N 100 B II 5-0 m. Locality Position Surface T. Bransfield Strait f63° ooi' S, 0-41° C Length in mm. 23 31 32 33 35 36 37 38 39 40 41 42 43 44 45 46 47 48 Total Av. length Stages 234567 ABCDEFG 7 6 9 IS 3 • 3 '-t 35 40 40 46 • 46 Total in sample 7 7 7 14 3 2 3 43 21 34 37 42 44 47 46 Date December 29, 1930 St. No. 558 Net N 100 B 146-0 m. Bellingshausen Sea f5s" 31' S, t f,r 07 r W Surface T. -092° C. Localit>' Position Length in mm. Stages Total in sample 12345 6 7 ABCDEFG 49 52 I I I I ; Total Av. length I . . • 49 . 1 ■ 52 I I 49 52 2 Date December 30. 1930 St. No. 55Q Net N 100 Bi 13-0 m. Locality Bellingshausen Sea Position {^*;i.-»;f.w Surface T. -o-8i° C. Length in mm. 40 47 Total Av. length Stages 234567 ABCDEFG Total in sample Bransfield Strait f62' 36i' S, i S8°58'W Surface T. 0-43^ C. Locality Position Length in mm. 39 40 41 43 45 46 47 48 49 50 52 53 55 Total Av. length Stages 2 3 4 5 6 viABCDEFG I 32 3 39 46 47 12 12 12 42 49 46 Total in sample 36 Date December 21-22, 1930 St. No. 549 Net NiooBli5-om. Locality Position Surface T. Bransfield Strait f63° oof S, { 6i°i6i'W 041° C. Length in mm. 13 15 17 32 33 35 36 37 38 39 40 41 42 43 44 45 46 47 48 52 Total 6 4 Av. length 17 34 Stages 567 7 24 38 42 ABCDEFG 6 I 2 3 21 I 7 17 36 33 36 40 45 47 Total in sample Date December 29, 1930 St. No. 5S8 Net N 100 B 146-0 m. Bellingshausen Sea /6s°3i'S, (. 67= 07J' W Surface T. -092' C. Locality Position Length in mm. 23 25 40 Total Av. length Stages 234567 3 40 ABCDEFG 3 ■ 27 ■ 3 40 Total in sample Date December 30, 1930 St. No. 559 Net NiooBii3-om. Locality Bellingshausen Sea ,, ... f66''2i}'S, Position I 68''55rW Surface T. -o-8i° C. Length in mm. 33 43 45 46 47 Total Av. length Stages 6 7 5 43 ABCDEFG Total in sample 33 46 45 6-2 146 DISCOVERY REPORTS MALES Date December 30, 1930 St. No. 560 Net N 100 B 155-0 m. Bellingshausen Sea (66°47rS, i 69°I9'W Surface T. -069° C. Locality Position Length in mm. 29 49 Total Av. length Stages 1234567 A B C D E F G FEMALES Date St. No. Net December 30, 1930 560 N 100 B 155-0 m. Locality Position Surface T. Bellingshausen Sea f66° 47J' S, I. 69° 19' W -0-69° C. Total in sample Date December 31, 1930 Locality Bellingshausen Sea St- No. 56. _ p„,itio„ i^^°Jji'3 Net N 100 B 137-0 m. \ 72° ogi' W Surface T. — i-35°C. Length in mm. 30 37 42 45 47 50 Total Av. length Stages 234567 20 . 40 46 50 A B C D- E F G 3 47 Total in sample Date December 31, 1930 St. No. 561 Net N 100 B 137-0 m. Locality Bellingshausen Sea „ .,. f66°47j'S, Position I 72»o9j'W Surface T. -135° C. Length Stages Total in sample in mm. 1234567 A B C D E F G 52 I I « Total Av. length 52 I 52 • Date December 31, 1930 St. No. 562 Net N 100 B 113-0 m. Locality Bellingshausen Sea Position {"7° '54' S Surface T. 75° 27' W o-62''C. Length Stages Total in mm. 1234567 A B C D E F G sample 47 50 I I I I Total Av. length 1 1 . . • • 47 50 • I . 1 . . . .47 -50 2 Date December 31, 1930 St. No. 562 Net N 100 B 1 13-0 m. Bellingshausen Sea I 75"_27'W Surface T. — o-ba" C. Locality Position Length in mm. 48 Total Av. length Stages 234567 48 A B C D E F G 48 Total in sample Length in mm. Total Av. length Stages 1234567 A B C D E F G 53 Total in sample THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 147 MALES FEMALES Date January 22, l92g Locality S. Georgia Date January 22, 1020 Locality S. Georgia St. No. Net WS373EE p„^;,;„^ IN 100 B 70-0 m. /54° 10' S. \ 35'40'W St. .N'o. WS 373 EE „ . . Net N 100 B 70-0 m. Position /S4° 10' S 1, 35'40 W Surface T circa 2-99- C. Surface T circa 2-99° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 36 I I 33 , I I 37 t I 34 I 38 r I 36 2 . . . 2 42 I . . . I 45 . I I I 2 43 1 2 . . . 2 47 1 1 I 44 I 48 I 1 I 45 2 . 2 49 6 . . I S 6 46 • . . 3 I • 3 4 50 3 I* ■ ■ 4 4 H I . . 3 I . 3 4 51 10 . . I 9 10 48 2 I . 2 3 52 6 . • I 5 6 49 2 2 S3 2 2 . 2 2 4 50 . . . 5 • . s 5 54 7 . . I 6 7 SI • ■ ■ 3 ■ . 3 3 55 2 . . 2 2 52 . . ■ 3 I • 3 4 56 I I I S3 . . . 6 . . 6 6 S7 1 I I S4 2 2 2 S8 1 I I 55 . ■ ■ 3 I . 3 4 59 I I 56 58 I I I Total Av. length 2 2 .. I 43 3 34 36 . . 45 52 52 I 3 • 34 35 . ■ 8 39 ■ 52 52 51 Total Av. length • I 13 I • -33 • 43 45 47 • ■ SI • 3 • 37 4 ) 9 • 33 48 . 51 48 Date January 21-22, 1930 Locality S. Georgia Date January 21-22, 1930 Locality S. Georgia St. No. Net N ^00 B loo-o m. P°^"'°" fS3° 17' S, ^ 37°lo'W Net''" r.ooB 100-0 m. P-"- {";,-nf-w Surface T 330" C. Surface T 3-30° c. Stages Total Stages Total Length in mm. in sample Length in mm. in sample 1234567 A B C I D E F G 1234567 A B C D E F G 35 2 1.... - 3 ■ 3 37 36 2 38 37 I 39 38 I 40 4« 3 41 42 I 42 2 45 I 45 1 51 I I 50 I 54 57 I I 1 I 51 I Total . 8 4 ... 3 . 12 . . . 3 15 Total Av. length 8 .... 2 . 41 .... SI . • 4 4 • 40 42 . 2 • ■ 51 10 Av. length ■ 39 39 . ■ .54 - 39 . . . 54 1 Date January 24-25, 1930 Locality S. Georgia Date January 24-25, 1930 Locality S. Georgia St. No. Net nIooB 150-0 m. P°^"'°" /S3°39r S, l 35°37i'W Net'"'- ^'looBlso-om. ?-"- ("°39rs,,, I 35 37rW TSJ -7r. V / 1 00-50 'n. Surface T ' \sSO-250 m. 2-4S° C. N 70 V 100-50 m. Surface T 2-45" C. Stages Total Stages Total Length in mm. m sample Length in mm. m sample 1234567 A B C I ) E F G 1234567 A B C D E F G 35 I 30 I 1 • . 41 1 I 50 45 . 51 I 47 57 I* 50 SI 52 2 I I 62 Total I ... I . 3 I . . . . 2 2 5 54 SS • ■ ■ 3 I . . 3 I Av. length 30 ... 51 • 56 30 . . . . 57 54 S6 2 . 2 2 57 4 • . 4 4 Total . I 3 ... 25 I I 2 . . 25 29 Av. length . 35 44 • . .54 . 35 41 At • . 54 Date January 29, 1930 Locality S. Georgia St. No. N^iooBro4-om. Position ;S3°39'S, Date January 29, 1930 Locality S. Georgia Net t 3S°24i'W St. No. 318 Position .J 53° 39' S Surface T 3lo°C. Net N 100 B 104-0 m. rosition Surface T. t. 3S^24A'\V 310° C. Stages Total Length in mm. in sample Stages Total in ABC! ) E F G Length 1214567IABCE E F G sample 35 2 . . . . . 2 . 2 37 I I I 37 4 4 39 I I 1 39 z 43 . . I . . . . I 1 40 2 2 Total . 2 3 ■ • • • .41- 5 Total 7 .43. 7 Av. length . 38 38 ... . . 37 43 . Av. length 38 . 38 38 . . . . Note. Females marked with an asterisk have spawned. 148 DISCOVERY REPORTS MALES Date Januari' 8, igii St. No. 575 Net N loo B 97-0 m. Locality Bellingshausen Sea Position |*7°53i'S. 1 91 23 W Surface T. — i-47°C. Length in mm. 31 33 34 36 37 38 40 41 45 46 47 48 50 Total Av. length Stages 34567 226615 32 34 37 43 48 47 A B C D E F G 356215 33 36 41 46 48 47 Total in sample FEMALES Date Januarys, 1931 .St. No. 575 Net N 100 B 97-0 m. Locality Bellingshausen Sea Pos,tion {%pi^/ij, Surface T. -1-47° C. Length in mm. 30 31 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 Total Av. length Stages 234567 ABCDEFG 3 31 33 27 IS 39 45 2 3 2 I 7 32 33 31 34 43 43 42 Total in sample Date January 8, 193 1 St. No. 576 Net N 100 B 132-0 m. Locality Bellingshausen Sea Position {%^°\%.^ Surface T. -1-15' C. Date Januarys, 1931 St. No. 576 Net N 100 B 132-0 in. Locality Bellingshausen Sea Position {^'8<5°;S^.w Surface T. — 1-15' C. Length in mm. Stages 234567 ABCDEFG Total I Av. length 49 Total in sample Length in mm. Total Av. length Stages 1234567 ABCDEFG Total in sample Date January 9, 1931 St. No. 578 Net N 100 B 128-0 m. Locality Bellingshausen Sea Position {^'8 54,fj.w Surface T. -1-20° C. Date January 9, 1931 St. No. 578 Net N 100 B 128-0 m. Locality Bellingshausen Sea Position C^^gS/ij-w Surface T. -I-20°C. Length in mm. 21 28 50 Total Av. length Stages 34567 ABCDEFG Total in sample Length in mm. Total Av. length 1234567 ABCDEFG Total in sample Date January lo, 1931 St. No 580 Net N 100 B 128-0 m. Locality Bellingshausen Sea Position (*7°4ji'S. (, 75 56i' W Surface T. -o-lo° C. Length in mm. 24 46 47 48 50 51 52 53 54 Total Av. length Stages 1234567 ABCDEFG Total in sample Date January 10, 1931 St. No. 580 Net N 100 B 128-0 m. Locality Bellingshausen Sea Position {^'„^-i,f.-w Surface T. -o-lo° C. Length in mm. 45 48 Total Av. length Stages 1234567 ABCDEFG Total in sample THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 149 MALES FEMALES Date January 10, 1931 Locality S. Georgia St. No. Marine Station Jetty „ „ ,„ Net NH m. Surface T. 2-40° C. (?) Date January 10, 1931 Locality S. Georgia St. No. Marine Station Jetty , r^ ,,■. Net NHom Surface T. 240' C. (?) Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G «9 20 21 22 23 24 li 27 28 30 31 I 12 3° 26 27 25 26 2 12 I 31 I I 12 30 26 27 25 26 2 12 I 31 I I 12 30 26 27 25 28 13 4 I I I 18 20 21 22 23 24 25 26 29 30 I 17 25 24 41 30 13 3 I I I 17 25 24 41 30 13 12 I I I 17 25 24 41 30 13 3 I I Total Av. length 56 23 152 4 23 28 156 Total Av. length 63 6 163 6 23 27 23 27 1 169 . Date January 12, 1931 Locahty Bellingshausen Sea Surface T. -o-ig'^ C. Date January 12, 1931 Locality BelUngshausen Sea Surface T. -o-ig' C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G SI I I I 3I 48 1 I I . . . I . . . I I I Total Av. length I 51 I 51 I I Total Av. length I . . . 2 . . 25 . . . 43 • • I . . I . I . 25 . .38 .48 . 3 Date January 13, 1931 Locality Bellingshausen Sea Surface T. -0-72° C. Date January 13. 1931 Locality Bellingshai St. No. 584 Position -f*'^^,^' ; Net N too B 165-0 m. Position ^ ^^^ j Surface T. -0-72° C. isen Sea 5, ' W Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 46 I I 1 23 48 I I I I I I Total Av. length I 46 I 46 I Total Av. length I .... I • 23 .... 48 • 1 I 23 48 2 Date January 14, 193 1 LocaUty Bellingshausen Sea St. No. 590 , Position l''5°^„°*' w',„ N" Nioob|9<^°'"- TT\- ^ 7,\.3oi'W \310-om. Surface!. IS7 L. N 70 B go-o m. Date January 14, 193 1 Locality Bellingshausen Sea St. No. 590 „ Position |*S,^o°*f.'w Net N IOC B 90-0 m. Surface T. W^'a"* "^ Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 49 50 52 54 55 I 1 2 2 I I I I 2 2 I I 2 2 2 I I 48 50 I I I I I 1 Total Av. length 2 . 49 ■ 2 49 2 Total Av. length ....1.7 . . . .49-52 I 7 49 52 8 Date January 16, 193 1 Locality Bellingshausen Sea N7oB/'70-om. Surface T. 1-51° C. Length in mm. Stages Total in sample Date January 16, 1931 Locality Bellingshausen Sea St. No. 596 ^ Position 1 'fi^='c?i'w Net N 100 B 170-0 m. I (f SSi W Surface T. 1-51 C. 1 1234567 A B C D E F G 45 tl 49 50 51 52 53 I 2 I I . I I I . . . I . . . I 2 I I I I I I I I 2 1 2 I Length in mm. Stages Total in sample 1234567 A B C D E F G 43 45 46 I . I I . . . . J . . I I 1 I Total Av. lengtl ...22.6 . . . 46 51 . 51 ...1.36 . . .47 . 49 51 10 Total Av. length 2 I . . . . 46 43 • 1 I I . . . . 45 43 46 3 ISO DISCOVERY REPORTS MALES FEMALES Date January 17. 1931 St. No. S99 Net N 100 B 142-0 m. Locality Bellingshausen Sea „ ... f67° 08' S, Position I 'bg-obVW Surface T. -0-71° C. Date January 17, 193 1 St. No, 599 Net N 100 B 142-0 m. Locality Bellingshausen Sea Position {*';°»^;fi-w Surface T. -0-71° C. Length in mm. Total Av. length 34567 A B C D E F G Total in sample Length in mm. 47 48 49 Total Av. length Stages I z 3 4 5 6 7 3 48 A B C D E F Date January 19, 193 1 St. No. 602 Net N 100 B\ ,,„ „„ N70B ; "o-o" Locality Bellingshausen Sea Position {^^66»^2S' W Surface T. — 0-02° C. Date January 19, 193 1 St. No. 602 Net N 100 Bl ,,„ „„ N70B ; ''°-°" Bellingshausen Sea (66° 03i' S, l 66' 25' W Surface T. — 002° C. Locality Position Length in mm. 41 42 45 46 47 48 49 50 51 52 S3 54 Total Av. length Stages 234567 ABCDEFG 47 50 52 50 6 I 8 18 46 45 51 50 Total in sample Length in mm. 25 37 40 42 43 44 45 46 47 48 50 52 53 54 55 57 58 Total Av. length Stages 34567 ABCDEFG 2 3 7 18 25 • 39 44 44 53 2 7 19 39 • 43 51 49 Total in sample Date January 20, 1931 St. No. 603 Net N too B 140-0 m. Locality Position Bellingshausen Sea /65° 04*' S, I. 67 5l*'W Surface T. i-o8° C. Length in mm. 47 50 53 55 Total Av. length Stages 1234567 ABCDEFG Total in sample Date January 25-26, 193 1 St. No. WS 537 Net N 100 B 67-0 m. Locality Approaching S. Sandwich Is. n ■.■ (56° 10' S, Position |5 ,j, ^^.-^ Surface T. 0-57'' C. Length in mm. 23 24 25 26 27 28 29 30 31 32 33 34 35 40 43 49 50 Total Av. length Stages 1234567 ABCDEFG 33 21 2 25 35 39 42 13 • 25 33 43 Total in sample Date January 20, 1931 St. No. 603 Net N 100 B 140-0 m. Bellingshausen Sea |6s' 04*' S, I. 67° 5 1 *' W Surface T. 108° C. Locality Position Length in mm. 45 53 56 Total Av. length Stages 1234567 3 51 ABCDEFG Total in sample Date January 25-26, 193 1 St. No. WS 537 Net N 100 B 67-0 m. Locality Approaching S. Sandwich Is. Position {">;;o°;f.v Surface T. 0-57° C. Length in mm. 23 24 25 26 27 28 29 30 31 32 33 36 37 47 51 Total Av. length Stages 3 4 5 6 7 ABCD'EFG 33 26 I 25 29 37 Total in sample THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 151 MALES FEMALES Date February 22, 1928 Locality S. Georgia St. No. WS 152 Position ^53° 12 00 S Net NiooBiit^om. Position ^ 34' 52'-oo W Surface T. circa 130 C. Date February 22, 1928 Locality S. Georgia St. No. WS152 Position (5^° 'o^''°? ^',.r Net NiooBlio-om. Position ^ 34° 52'oo W Surface T. circa i-30° C. Length in mm. Stages 1 Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 34 41 43 44 45 46 :^ 49 50 51 52 S3 54 55 57 I I . 2 . 2 . ■ ■ 7 . . . 4 : : : t . . . 14 . . . 9 . . . 6 . . . 5 I . • • 3 I I . . I 2 I . 2 I • 7 ■ . 4 • . 4 . . 6 ■ • 14 . • 9 . . 6 • ■ 5 I I • ■ 3 I I I I 2 4 4 6 14 9 6 5 2 3 I 42 44 45 46 47 49 50 51 52 53 54 55 56 59 60 I I . . 2 2 . . 2 2 . 2 . 2 . I I I 31 12 1 I I I I I 2 . I 1 I I 4 4 2 I I 3 4 I 3 I I 2 I 2 I I I I I I 4 4 2 2 3 4 4 2 2 3 2 I Total Av. length . . . . 7 23 6 . . . . 43 50 S3 5 31 53 50 36 Total Av. length . . 4 . . .66 . . 45 • • -so I 3 . 66 70 ■ . 34 46 . 50 Date February 26, 1928 Locality S. Georgia St. No. WS 156 Position •f53°4p'ooS Net N 70 V 100-50 m. losition -^ 36' jz'-oo W Surface!'. circa 2-31° C. Date February 26, 1928 Locality S. Georgia St. No. WS 156 Position /53°4p'ooS Net N 70 V loc^so m. Position ^ ^g. ^^,.^^ ^ Surface T. circa 2-31° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 24 25 26 28 29 30 31 32 33 34 11 37 38 39 40 41 42 1 2 4 I 2 3 2 2 2 10 . 6 I 4 . 7 . 6 . 2 • 4 I 1 I 1 I . . . . I . . . . I . I I . 3 • 5 5 ■ 4 . II . 6 . 5 . 5 . 6 . 2 I 2 2 2 r . . . I I 3 5 5 4 12 6 5 7 6 2 4 2 I I 2 25 26 27 28 29 30 31 32 33 34 35 36 37 38 2 2 I 3 5 6 s '.'.'.'. '■ 32 13 26 22 . 2 12 II.... 2 2 I 12 • 5 II 16 • 5 .4 .8 13 II.... ■ 3 2 . . . . 2 2 I 3 5 2 7 5 4 8 4 2 3 2 , 2 . I I I Total Av. length 29 20 I . 30 34 38 . . . . 8 39 3 • ■ ■ • 28 32 37 ■ • • • 50 Total Av. length 16 so 2 . . . . 30 34 42 . . . . . 58 9 I • • • . 33 37 41 ■ • • 68 Date February 8, 1929 Locality S. Orkney Is. St. No. WS 376 Position 157° 23 -op S, Net N 70 V 750-500 m. Position ^ 42^52'ooW Surface T. i-45 C. Date February 8, 1929 Locality S. Orkn St. No. WS376 Position (57^^.3 Net N 70 V 750-SOO m. „ ^ ,, '• "tV Surface T. 145 C ey Is. ■00 S, 52'-oo W Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 40 46 50 5« . . . I . . . I . I I . . ■ I I I I I I I 50 53 I I I I I I Total Av. length I I S3 50 2 52 2 Total Av. length . . . I . I 2 . . .40 .46 51 . . . I . I 2 . . .40 . 46 51 4 - Date February i, 1930 Locality S. Georgia St. No. 325 Position (54° 53 S. Net N 70 V 750-Soo m. position ^ 39° 57' W Surface T. 3-32 C. Date February 14, 1929 Locality ^. Orkr St. No. WS381 Position (*' Ao Net N7oV5tw>m. = ' -r *-fi/»P Surface T. 165 C ey Is. '00 S, l9'-00 W Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 1234567 A B C D E F G 41 I . . ' . I I 48 I . ■ • • ' . I . . Total At. length . . 1 . . . . . . 41 . . . . • I • 41 I Total Av. length I .48 . . 48 . . ■ • I IS2 DISCOVERY REPORTS MALES FEMALES Date February 8, 1930 Locality S. Georgia Date February 8, 1930 Locality S. Georgia r55° 01' s, \ 35°27i'W St. No. 349 Position Net N 100 B 60-0 m. f55°oi'S, l 35°27i'W St. No. Net N'ooB6c«m. P°^'"°" Surface T. 300° c. N 70 V 50-C m. Surlace 1 . Stages Total Stages Total Length in mm. in G sample Length in mm. in sample 1234567 A B C E E F 1234 5 6 7 A BCD E F G I 1 36 4 . . . ■ 4 • 4 I I 3Z 4 . . . I . 2 2 I S 40 21.... . 3 ■ • 3 38 I . . . I 41 I I 1 39 3 • • • • • 3 ■ 3 42 . 22.... • ■ 3 I 4 40 2 . . . 1 2 . 3 I I I 41 I . . . 44 45 49 I . . . . . . I . . . . I . . . . I I 1 I tl 3 • ■ • 1 . . . -- • 3 ■ I 3 I Total Av. length 19 . . . 39 • • ■ Total Av. length .85.... . 40 44 . . . . • 743- • • . 39 43 45 • ■ ■ 14 39 • • . 37 39 37 . • 40 Date February 9. 1930 Locality S. Georgia Date February 9. 1930 Locality S. Georgia (54°2li' S, I 35°42'W St. No. 351 Pn^itinn Net N 100 B 48-0 m. Position (54°2li'S \ 35° 42- W St. No. Net 351 Position N 100 li 40-0 m. Surface T. 3-78° C. N 70 V 100-50 m. Surface F. 3-78- C. Stages Total Stages Total Length in mm. in sample Length in mm. m sample 1234567 A B C D E F G 1234567 A B C D E F G 42 2 . . . . 2 • 2 38 I I I 43 2 I I 2 3 39 2 I I 2 40 2 2 4 2 2 4 41 ■ 3 I ■ 31. t 2 4 42 ■ 2 3 ■ 2 3 . 47 I 2 . 2 3 43 ■ I 4 113. 5 48 I 2 3 44 5 113. 5 1 45 5 I . 4 I 50 I 5 46 I 52 . I 47 2 2 S3 . 2 48 I I I 2 49 . 3 • 3 . • 3 SS 56 I J 55 57 2 . . . I* I 2 58 • . 5* 5 is I 60 . . 2« . . 2 1 2 59 3 60 I Total • 9 24 . 8 8 4 14 15 6 . 10 49 64 I Av. length • 41 43 • 49 58 42 42 44 • 47 ■ 58 Total . 22 10 7 2 I . 4 11 I 772 42 Av. length . . 47 SI 56 60 so 1 ■ 43 46 4* ! S3 58 53 Date St. No. Net February 9 352 N 100 B s8 1930 Locality Position -0 m. Surface T. S. Georgia (54°I9'S. I 35°24'W 2-95° C. Date February g, 1930 Locality gfet""" ?l'.ooB58-om. P-""" S. Georgia (54° 19' S \ 35° 24 W Surface T 2-95° C. Stages Total Total in sample Length m mm. I 2 3 4567 A B C D E F G sample 1234567 ABC D E F G u J 37 . . I . . . . . I I . I 38 3 . . . ■ 3 • 3Z 2 . 40 • 3 4 • • • 7 ■ 38 2 I 41 3 4 ■ • . 6 I 39 42 43 12.. . . 3 ■ • 1 2 2 I 40 41 2 I 5 I I 5 6 4S 3 • ■ ■ • 3 42 3 • • 3 46 2 I 43 2 2 I 5 47 52 I . . I I I 45 47 I '. '. !' ^ I 54 I I 55 Total . 10 20 1 1 . 21 8 I . I I 32 Total 17 It 2 2 9 17 3 I 2 32 Av. length . 40 42 . 52 54 ■ . 41 44 4 5 ■ 52 54 Av. length 40 40 44 . . . SI 38 41 42 43 . • 51 THE DEVELOPMENT AND LIFE-HISTORY OF KRLLL 153 MALES FEMALES Date February 9, 1930 Locality S. Georgia Date February 9, 1930 Locality S. Georgia St. No. Net nIoo B 96-0 m. ^°'"'°" (54° ISJ'S, \ 34= 47r W St. No. Net N 100 B 96 -0 m. Position \ 34 47* W Surface T 2 35' C. Surface T. 235' c. Stages Total Stages j Total Length in mm. in sample Length in mm. ] in sample 234567 A B C D E F G 234567 A B C D E F G 35 I I I 34 I I . . . I 38 I I I 36 2 . I 1 39 2 2 . 2 37 2 2 40 I I I 38 I 1 41 2 1 2 I 3 39 42 8 10 4 14 18 40 3 2 3 43 I 4 3 5 41 3 I 3 4 44 . 2 2 2 42 45 9 6 . 9 43 2 ' I 3 46 I I 44 2 . 2 . 47 4 2 : 4 45 3 I 2 3 48 1 I 46 I I I 49 3 2 4 47 1 50 2 I 2 49 51 2 5 ■ •Si 7 SO I" 52 . 3 7 2 I 6 1 10 SI 3* I 2 3 53 . . 6 .33! 6 52 3* } ■ • 3 3 54 55 • 2 7 2 14 -.25, 9 I 4 10 16 53 . I 8* . 10 10 56 • I 5 - • 4 I 1 7 54 . 6» 2 5 7 57 . . 6 . 3 2 6 SS 4* • I 3 4 58 • 2 s 2 4 7 56 3* . . 3 3 59 I I I 2 57 S* I 4 5 60 . 2 2 2 59 ■ 1* . 2 Total IS 38 I . II 61 13 30 I< ) 3 20 41 126 Total 18 7 • 7 35 6 II 6 6 . 6 32 67 Av. length - 41 45 56 . SS S5 - 41 44 S 53 55 S4 Av. length - 40 44 ■ 48 S3 37 41 45 44 • 54 55 Date February 10, 1930 Locality S. Georgia Date February 10, 1930 Locality S. Georgia St. No. Net N 70 V 250-100 m. P°^'"°" f54°ll'S. 1 33° 49' W St. No. Net N 70 V 250-100 m Position /S4° 11' S. 1 33' 49' W Surface T 2-02° C. Surface T. 202° c. Stages Total Stages Total Length in mm. • Length in mm. in sample J 2 3 4 5 6 7 A B C I D E F G sample 1234567 A B C D E F G 40 I . . . . 1 I 38 I I I 41 2 2 . 2 39 43 I I 42 I I I 44 2 2 2 49 45 3 3 3 SO I 46 4 2 4 52 2 48 6 I . a 3 I 7 53 2 50 SI I I . I I 1 2 I 54 55 I* 6 6 6 54 1 I S6 I I I 56 I 1 57 3 i 3 57 I I 1 58 S b 59 .... I 1 I 59 60 • I 3 1 9 4 10 Total 20 3 . I 3 6 II i 3 4 27 Av. length 46 50 . 57 s6 44 46 4 3 SO 57 61 63 I* 1 . • 5 5 64 I I ■ Total 3 J 42 I 2 2 . 3 40 48 Av. length 40 . 5 3 58 39 39 45 • 57 58 Date Februarys, 193 1 Locality Bransfield Strait St. No. 1 62" oSi' S. Date February 8 1931 Locality Bransfield Strait Net n'.oo B 128-0 m. P°^'"°" 1, 62' 57i' W St. No. 609 /62° o8i' S, 1 62°57i'W Surface T 203° C. Net N 100 B 12 i-o m. Surface T. 2-03° C. Stages Total Length in mm. in sample Length in mm. Stages Total in sample 1234567 A B C I 3 E F G 123^ 1 5 1 J 7 A B C D E F G 47 I I I 48 I I I 49 52 3 • • 3 3 SI I 53 3 . . 3 3 55 Total 8 . . 8 8 Total i ■ 2 I 3 Av. length SI . • SI Av. length . 5 ! . . 52 51 7-2 IS4 DISCOVERY REPORTS MALES FEMALES Date February 10, 193 1 Locality Bransfield Strait Date February 10, 193 1 Locality Bransfield Strait St. No. Net n'.ooB 160-0 m, P°''"°" (62° 42' S, I 57°io'W St. No. Net n'ioo B .60-0 m. P°="'°° \ 57" 10'' W N 50 V 1 00-0 m. Sm-face T 052" C. Surface T. 0-52" C. Stages Total Stages Total Length in mm. in sample Length _ in mm. m sample 1 I 2345671 \ B C D E F G I 234567 A B C D E F G zo : 1 1 I 16 I I . . . . I 21 2 2 . 2 19 I I I 22 I I I 21 I I I 23 6 . . 6 6 22 3 3 3 24 3 4 ■ 7 7 23 5 5 5 25 I ■ 5 . I IS 16 24 20 20 20 26 9 20 . . 29 29 25 41 41 41 27 I 21 . 31 31 26 31 ■ 3 31 28 3 13 . . 16 16 27 . 34 • 3J 34 29 I 10 2 . 13 13 28 . 18 . 18 18 30 I .12 3 I 35 36 29 19 . 19 19 31 I 13 I 1 14 15 30 • 42 42 42 32 . 12 2 . 14 14 31 . 12 12 12 33 6 I 7 7 32 24 24 24 34 2 3 4 I 5 33 IS '1 IS 35 . 7 6 2 8 34 6 6 36 . 2 I 2 3 35 20 19 37 4 2 3 36 6 6 6 38 5 6 2 8 37 3 3 3 39 . 3 2 I I 4 40 2* 40 . 3 I • I 3 I 5 41 I* 41 I 3 • i ■ 4 42 42 . 3 . ■ 4 4 43 3; 3 3 43 3 • J 3 • 4 44 V* 7 7 44 I 2 3 • 3 45 14* 1 13 14 45 8 2 I I 10 tl 46 7* 7 7 46 I I 2 47 4* 4 4 47 9 7 • 692. 17 48 6» b 6 48 2 3 • 221. 5 49 i 5 49 6 3 . 2 7 ■ ■ 9 SO 6 6 50 4 4 ■ .71. 8 51 51 2 2 I • 321 ■ 3 I • 131- 6 4 5 S3 I* 52 53 I 4 ■ Total 03 113 86 ■ 58 185 116 I 1 57 360 54 12... I 2 . 3 Av. length 24 29 33 ■ 47 27 31 . 35 45 47 Total « 146 77 39 4 2 3 212 16 27 47 II I 317 Av. length 25 29 41 48 48 49 . 29 29 38 4 5 48 50 51 Date February 12, 1 93 1 Locahty S. Orkney Is. Date February 12, 193 1 Locality S. Orkney Is. St No. Net N^ooB 182-0 m. P°^'"°" |6o° 59}' S, 1 50°42i'W St. No. Net N^ioo B .82-0 m. P°''"°" f6o° 59}' S, 1 50°42i'W Surface T -021° C. Surface T. -0-2I° C. Stages Total Stages Total Length in mm. in sample Length in mm. in sample I 234567 ABC D E F G I 234567 A B C D E F G 45 I I I 22 23 26 35 42 I ... I . . . Total I . I I I ... I Av. length 45 ■ • • 45 ■ i< > _ Total 3 I .... I 3 ■ I ■ . . I 5 Av. length 24 35 ■ • ■ -42 24 • 35 • • ■ 42 Date Februarv 18-19, 1931 Locality S. Orkney Is. St. No Net N^iooBii9-om., P°^"'°" N 70 V 50-0 m. Surface T. (59° 42}' S I 43°57rW 1-29° C. Date St. No. February 18-19, ^931 Locality N''iooBil9-om., P°^'"°" N 70 V 50-0 m. Surface 1 S. Orkney Is. 159° 42}' S \ 43 57? W r. 1-29° C. Stages Total Net Length in mm. in sample 12 34567 A B C D E F G Length in mm. Stages Total - in sample 26 27 28 2 2 ... I ... I 12.. 2 I 3 I 234567 ABC D E F G 27 2 2 2 30 4 3 7 30 21... • 3 • 3 31 I I 2 , 2 32 42... 4 2 . 6 32 ■ 3 . 3 ■ • 3 33 15... I 5 6 33 2 4 24.. 34 9 . . . 9 . 9 34 2 35 4 ■ ■ ■ 4 • 4 35 I II II I 12 36 3 • • • 3 . 3 36 3 4 ■ • 4 37 4 . . . 4 • 4 37 5 5 . ■ 5 38 3 . • • 3 ■ 3 38 4 4 39 2 . . . 2 . 2 40 1 40 1 . . . I I 45 3 41 2 . . . 2 . 2 50 I 45 I 51 2 51 I I 52 54 I . . . . I 54 4 4 Total 9 36 I . . 2 . 7 38 I 48 Total 15 37 I ... II 7 42 4 ■ . . II 64 Av. length 31 35 45 • -53 . 31 35 45 ■ ■ • 5 Av. length 29 35 36 . . . 5C 29 34 38 . ■ ■ 5C THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 155 MALES Date February 20, 1931 St. No. 6zi Net N 100 B loo-o m. Locality S. Orkney Is. Position {=';|?53'''w Surface T. 0-25' C. Length in nun. 39 42 Total Av. length Stages 1234567 A B C D E F G Total in sample FEMALES Date February 20, 1 93 1 Locality S. Orkney Is. St. No. 621 p„^,>i„n /58° 5oi' S, Net N 100 B 100-0 m. Position | ^go 53. yj Surface T. 0-25° C. Length in mm. Date St. No. Net February 20, 1931 622 N 100 B 155-0 m., N 70 V 50-0 m., N 50 V loo-o m. S. Sandwich Is. /59° 05*' S, I 36°25'W Surface T. -0-89° C. Locality Position Length in mm. 25 26 27 30 32 33 34 35 36 37 38 39 40 41 42 43 49 Stages 234567 ABCDEFG Total Av. length 29 24 6 32 37 41 3 51 5 26 35 41 Total in sample 60 Date St. No. Net February 21-22, 1931 624 N 100 B 137-0 m., N 70 V 50-0 m. Locality Position Surface T. S. Sandwich Is. /58° 34}' S, I 3i°2ii'W 022° C. Length in mm. 23 24 25 26 27 28 29 30 31 32 35 36 37 41 Total Av. length Stages 234567 ABCDEFG 41 3 2 27 35 38 16 28 2 24 30 38 Total in sample 46 Date February 22, 193 1 St. No. 626 Net N 100 B 158-0 m. S. Sandwich Is. /57° 22' S, 1 25° 29i' W Surface T. -009° C. Locality Position Length in mm. 23 24 25 29 Total Av. length Stages 34567 ABCDEFG 5 4 II 28 28 Total Av. length Stages 1234567 28 ABCDEFG 28 Total in sample Total in sample Date February 20, 1931 St. No. 622 Net N 100 B 155-0 m. Locality Position Surface T. S. Sandwich It /59°o5S' S. , I 35° 25' W -089° C. Length in mm. 23 25 27 28 29 30 31 32 33 34 35 36 38 30 40 41 50 S6 Total Av. length 27 34 Stages 1234567 ABCDEFG 13 29 7 25 10 26 32 36 Total in sample Date February 21-22, 1931 St. No. 624 Net N 100 B 137-0 m. Locality Position Surface T. S. Sandwich Is. (58° 345' S, '(. 3l°2lJ'W 022' C. Length in mm. 23 24 25 27 29 30 31 32 35 36 48 Total Av. length Stages 34567 ABCDEFG 10 15 27 30 48 9 15 I 27 29 36 48 Total in sample Date February 22, 193 1 St. No. 626 Net N 100 B 158-0 m. S. Sandwich 1 f57° 22' S. Locality Position {'■26° 291' W Surface T. -009° C. Length in mm. 18 19 23 24 26 28 30 32 Total Av. length Stages 234567 ABCDEFG 15 4 21 30 14 5 21 29 Total in sample 156 DISCOVERY REPORTS MALES Date St. No. Net February 23, 193 1 627 N 100 B 1 18-0 m. Locality Position Surface T, S. Sandwich Is. f56°S3rS, \ 23° 474' W 110° C. Length in mm. 23 24 25 26 27 28 29 30 31 32 33 34 35 35 37 38 40 45 51 Total Av. length Stages 3456 31 18 5 27 33 40 A B C D E F G 23 26 4 . I 27 33 37 • 45 Total in sample Date February 24, 193 1 St. No. 628 Net N 100 B 126-0 m. Locality S. Sandwich Is. Position {"aa'l+'^W Surface T. -015° C. Length in mm. Stages 1234567 A B C D E F G Total in sample FEMALES Date February 23, 1931 St. No. 627 Net N 100 B I iS-o m. Locality Position Surface T. S. Sandwich Is. 156° 538' S, I 23'47rW llo" C. Length in mm. 25 26 27 28 29 30 31 32 33 34 35 36 38 Total Av. length Stages 234567 ABCDEFG 2 44 23 29 13 32 I 27 30 38 Total in sample 46 Date February 24, 1931 St. No. 628 Net N 100 B 126-0 m. Locality S. Sandwich Is. Position {'Qo^^^/i^ Surface T. -015° C. Length in mm. Stages 34567 ABCDEFG Total in sample 26 30 Total Av. length Total Av. length 26 30 Date February 25, 193 1 St. No. 629 Net N 100 B 152-0 m. Locality Position Surface T. S. Sandwich Is. — S. Georgia [55° 331' S, X 30° 01' W oii°C. Length in mm. 23 29 32 34 36 Stages 6 7 ABCDEFG Total in sample Date February 25, 193 1 St. No. 629 Net N 100 B 152-0 m. Locality S. Sandwich Is. — S. Georgia Position {";o3o^i;?w Surface T. o!i°C. Length in mm. 25 30 Stages 234567 ABCDEFG Total in sample Total Av. length 5 4 25 35 3 6 21 33 Total Av. length THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 157 MALES FEMALES Date March 14, 1926 Locality S. Georgia Date March 14, 1926 Locality S. Georgia St. No. 23 Position Cumberland Bay St. No. 23 Position Cumberland Bay Net N 100 H 6o-(o) m. Surface T. 267° C. Net N 100 H 6o-(o) m. Surface T. 2-67' C. Stages Total Stages Total Length in mm. in sample Length in mm. m sample 123456 7 A B C D E F G 123456 7 A B C D E F G 26 I 35 3 I 2 . 3 37 I . . . I . . I 36 3 . 2 I . 3 38 1 I I 37 2 . I I 2 40 • • 5 . . .23. 5 38 22.... . 2 2 4 4J I 1 I 39 3 I 2 . 3 42 ■ ■ 4 ■ . 2 2 . 4 40 3 . . 3 • 3 43 t- z . . . . . 3 ■ 3 41 4 . . 3 I 4 44 ..63.. . . I 8 . 9 42 4 . I 3 • 4 45 . .32.. . . I 4 . 5 43 3 3 • • ■ . • • 3 2 6 46 . .42.. . . . 6 . 6 44 I . . . . 1 I 47 2 . . . 2 . 2 48 l» 1 48 . . 2 2 . . . . 3 1 4 49 :• I 49 I 1 I Total 27 6 ... . 2 . 6 21 5 I . 2 35 Total I . 30 12 . . I . 9 31 2 43 Av. length 39 42 . . . . 49 . 38 40 41 43 ■ 49 Av. length 26 . 43 45 ■ • 26 . 41 44 4? . Date March 14, 1926 Locality S. Georgia Date March 14, 1926 Localitv- S. Georgia St. No. 24 Position Cumberland Bay St. No. 24 Position Cumberland Bay Net N 100 H 6o-(o) m. Surface T. circa 2-67° C. Net N 100 H 6o-(o) m. Surface T. circa 267° C. Stages Total Stages Total Length in mm. m sample Length in mm. m sample I 2 3 4 5 f 7 A B C D E F G 123456 7 A B C D E F G 40 I I I 35 2 . I I 2 42 I I I 36 2 . 2 2 45 . . I I . I I 2 37 2 . 2 . 2 46 I . . . I 1 38 5 5 ■23. I 4 5 5 Total . .41. . 122 5 40 41 14 12 . I 9 4 • 552 14 12 Av. length • ■ 43 45 ■ • 40 44 46 42 ■i . 2 10 I 13 8 44 42.... • • 4 2 6 45 81.... ■ 2 3 4 9 46 I I I 47 1* . . 3 3 48 1* I I 49 I* . . I I 55 I* 1 I Total 76 3 ... . 6 . 18 45 16 . . 6 8s Av. length 41 44 • ■ ■ . 49 . 41 41 43 • . 49 Date March 19, 1926 Locality S. Georgia St. No. 38 Position Cumberland Bay Net N 100 H 50-(o) m Surface T. circa 2-85° C. Stages Total in sample Date March 19, 1926 Locality S. Georgia Length in mm. St. No. Net 38 N 100 H so-(o) m. Position Surface T. Cumberland Bay circa 2-85' C. 1234s 6 7 A B C D E F G Stages Total 36 11 39 I I I 3 Length in mm. in sample ': 3 : 123456 7 A B C D E F G 40 I 2 • • 3 • 3 35 I . . . . I I 41 . I 4 • I 3 I 5 37 2 . . . . . 2 . 2 42 2 I I 2 38 I 3 . • • ■ . 2 2 . 4 43 ■ ■ 3