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 0 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 0 « 1^ >j fe is e f^tn m oou-ju-iMOvMroC^ •^ in rn in 0 "H m 1 ^ " N ro 0 ■-■ r^OvOrOt^i-iN« m 0 ON NO On NO 0 ON moo E roN -^ wivO^roK-lOO" N N m 0 0 " N ' 0 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 0 E 1 iri ir^ 00 1 1 1 N m 1 0 CO 0 1 1 1 0 N 00 1 1 1 *-i l^ t~- vO ro " 0 ON NO 1 1 1 2^ " en 10 00 g ' m m 1 1 1 ^ i/^ 1 « N N M ' ' ' ffi" 0 % *>> •2 s &^ = 0 1^ p oo^f^r^^or^oo OC> " io^66^N"t^" rn r^ CO rn 0 -t W ON N NO in in 0 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^, 0 CO m 0 ON ONCO N •-' ■g ov ^ t^ \0 N 0 ONOO c^oo -i- 0 CO r^ -jt- NO r-- 0 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 > 0 -s s" s ,0 0 0 0 K-i r^OO 0 1 CO 0 rn 0 1 1 1 1 I 1 1 1 E s r^ 0 sO " 0 t^ t^CO r^ r^jNO 00 1 1 III s 'ro r^ tovDiorOTh'n « N en ' 3 CO cl 03 t— 1 0 E CO u-iOOrot^t^"-lt^ r^ in 0 1 1 1 0 1 moo NO 0 u^N r^i r^O>-'OOOt^O-<*- 0 t-^ ON ON t-»CO NO -a E •H ro fO u-isO ^ ro -^^ ►"" ■-• N m ' 1 1 - ' ON N " g 0 2 " >n 00 ^0 0^ On ro t^ c^ r^ 9^^ ?^ 7*" 9^ 0 00 ^ t^db ^ " ro i/^ "-1 N r^- 0 ON ^- ON NO 1 1 > p-i ON 0 Ti-u->inc<^Ti-ioi-i N m " 00 N " "So ^ E r«-iio 0 r^row^NmOOOi^j NO m 0 1 1 1 m 1 N N " C4 42 t-H ro t^ vO ON -^ N r^vO »o 0 (^ ON 0 1 N ONOO N * E CON N rort--. 1'i U ^ 3 , E^ ^ s g^ '0 V 0 M 00 Og BccS= <- S '0 h c .E -a m ^ 2S-al|g= °gl 0 1 ^il^S2<^|EEp^ -^ Ej= gj:^^^J=J= f,J= ^ «.S^2-SS^.SfSS-5-S^-SS 0 a 0 _c E J 1 -a 0 0 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 0 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 0 S"! 0 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 0 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 0 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 0 r- i^ 0 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 0 w 10 N ro 0 "^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 0 'r?5 o 0 a. D:i 2 0 Q_ d u < 6X3 -1 Qi C 0 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 . . . 0 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 0 Locality S. Georgia Date November 19, 193 0 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 0 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 0 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 0 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 0 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 . • 3 I 4 39 . 3 ■ • ■ • ■ ■ 3 • 3 44 • • 4 2 2 4 40 . 7 . . . . . I 5 I 7 45 . . 8 ■ • 4 5 9 41 12... I 2 3 46 • ■ 3 . 2 I 3 42 ■ 3 3 • • . ■ ■ 5 I 6 47 • • 4 . . I 4 5 43 . 2 . . . . 2 . 2 48 • . 4 . • I 3 4 44 . 6 . . . . ■ ■ 5 I 6 49 I 2 . 2 I 3 45 .36... • I 5 2 9 50 I I I 46 . 2 5 . • • .241 7 Total . 7 36 6 . . 7 21 20 I 49 Total I 33 16 • . • • 6 35 8 I . . 50 Av. length • 39 44 47 • • 39 43 46 4 9 • . Av. length 38 38 44 . . ■ . 42 42 43 45 . . IS8 DISCOVERY REPORTS MALES FEMALES S. Sandwich Is. Date March 8, 1930 Locality S. Sandwich Is. St. No. 368 Position Southern Thule St. No. 368 Position Southern Thule Net N 100 B 146-0 m. Surface T. 011° C. Net N 100 B 146-0 m. Surface T on" C. Stages Total Stages Total Length in mm. m sample Length in mm. in sample 1234567 A B C D E F G 1234567 A B C D E F G 38 I . . I . • • I 36 . I . I 40 2 . . . . r 2 37 2 . . 2 42 I . . . r I 39 43 22... 3 4 40 44 . . 4 . . . 2 2 4 41 2 45 . 24... 4 I 6 42 46 III.. 2 I 3 43 2 2 4 47 . . s • • • 4 I 5 44 2 2 I - 3 4 48 II.. . 2 2 +1 121.. 4 4 • • ■ 2 2 4 46 III.. I . I I 3 I 3 ■ ■ I I I 4 47 112.. 4 I 1 48 I 3 • • I r I 4 52 55 56 2 . . . . 2 . I I 2 49 SO 51 I' . . 2» . . I* I 2 I 57 . I 1 52 2 S8 .... 1 I I 53 . 2* 54 • • 3* . . 2* . . 2* I 2 2 . 2 3 2 2 Total . 8 25 5 • 32 . 3 21 12 2 4 43 Av. length . 43 47 49 • 55 57 ■ 43 45 48 50 51 57 60 . ' . . I I Total 3 12 10 7 . .16 6 8 9 6 2 3 14 48 Av. length 37 43 44 47 ■ . S3 41 41 46 46 47 51 54 Date March 9, 193 1 Locality S. Orkney Is. Date March 9, 193 1 Locality S. Orkney Is. St. No. 638 Po 7 A B C D E F G 22 I . . . . I . . . I 25 z . . . . 1 . . . I Total Av. length 2 24 2 24 2 Date St. No. March 26, 193 WSS75 I Locality S. Georgia 152° 35' S, Net N 100 B 78-0 m. \ 38°09'W Surface T 282° C. Date March 26, 193 1 St. No. WS 575 Locality S. Georgia (52° 35' S,„, Stages Total Net N 100 B 78-0 m. Surface T. I 38° 09' W 2-82° c. Length in mm. in sample 1234 5 6 7 - \ B C D E F G Stages Total ' I . . . * in mm. 12345 6 7 A B C D E F G sample 29 I . . . I I 29 31 2 . . . 2 . 2 30 32 32 33 . 3 . . . 2 . ■ 3 • . 2 3 2 11 It . I . . I I I I I Total . 5 • • ■ • 5 . ■ 5 Total 67.. 2 II 13 Av. length . 32 • • • . 32 . ■ Av. length 29 33 • . 25 32 . THE DEVELOPMENT AND LIFE-HISTORY OF KRILL i6i MALES FEMALES Date April I, 1926 Locality S. Georgia ' St. No. 42 Position Cumberland Bay Net N 7-T 120-204 m. Surface T. circa 2-08° C. Date April i, 1926 Locality S. Georgia St. No. 42 Position Cumberland Bay Net N 7-T 120-204 m. Surface T. circa aoS" 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 35 38 39 40 41 42 43 44 n 47 I . 2 . . .41. . .91. . . 5 ■ • ..71- • -73- . .91. . .42. 2 I . . 4 . . i I : 4 I 2 6 . 2 8 . I 2 5 10 5 8 10 10 6 3 4 36 37 38 39 40 42 43 44 1 . 2 .41. ■ ■ • II.... 4 . . . . . . 6 . . . . . . 6 . . . . I . . . . I . . . . . 2 . . . . .32.... 2 . . . . . 22.... . 3 3 • • • • .42.... I . . . . 2 5 2 4 6 6 I I 21.. . 4 - • • Total Av. length . 7 20 . . . . . 37 41 . • • ■ . 12 15 . . • • . 40 39 • • • ■ 27 Total Av. length ■ • 54 10 . ■ . 42 43 . 2 26 35 I • • • 37 4J 43 46 ■ . 64 Date April 7, 1927 Locality Bransfield Strait St. No. 207A Position f62° 54'oo S, Net N7oHc^5m Position | sg^jo'-joW c -c T" jFrom —078' Surface T | ,0 -o-86° C. Date April 7, 1927 Locality Bransfield Strait St. No. 207A Position 1''^° 5/°,° S. „ Net N 70 H 0-5 m. i-osmon -^ jg» 5° -30 W c.._f„.-. -r (From -078 Surface T. | ,0 -o-86» 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 32 33 35 35 11 39 40 48 49 50 51 52 • • 3 1 1 I 1 I I ■ I 4 2 I 3 I 2 4 2 1 3 \ I 30 31 32 33 11 39 40 41 43 44 45 46 47 48 49 50 51 53 1 I 2 ;: ;: 4* I* 1 2 2 • 3 . 6 ■ 4 I I 2 2 Total Av. length . I 18 4 3 4 . • 35 38 47 50 50 . .557661 . 35 36 40 46 49 49 30 ... I* I* . . . . I I Toul Av. length 11 35 ■ . . 5 22 . ■ 46 47 . 3 8 . 6 . 21 . 32 36 .46 . 47 38 Date April 7, 1927 Locality Bransfield Strait St. No. 207B Position C"^"'.*'"?^',,, Net N7oHo-5m. fosmon -^ g. ^^..j ^ Surface T. {'^^^-'^1' C. Date April 7, 1927 Locality Bransfield Strait ^'et''°- N^7oH<.5n,. P-''^™ i^'l^^^^V^ Surface T. {^^?-_-^,fc. Length in mm. Stages Total in sample 1234567 A B C D E F G Length in mm. Stages Total in sample U 30 31 32 33 34 11 11 39 40 41 42 44 45 46 47 50 52 I I 4 5 7 8 6 16 7 4 3 2 4 3 - 2 I 4 I 2 I I 1234567 A B C D E F G I I 2 . 1 I I I z i» 2* I* 4* I» . z* . 1* . I* I . 2 I • 4 I . 2 I I 26 11 30 31 32 33 34 35 36 37 38 39 40 41 43 45 52 . 4 ■ 3 I 2 I I I I 3 3 I 2 I 3 I . . : 3 2 2 3 2 I I 2 2 I I 2 I 2 . 2 . . I I I I I 1 2 4 3 3 5 2 I 2 3 3 4 4 2 I I I 2 4 • ! 5 ■ ^ 4 • 2 I : 12 2 ! 5 ■ 2 . Total Av. length 4 13 18 6 I . . 30 32 39 4' 52 . . 7 15 9 4 5 2 . 30 34 38 +1 43 49 • 42 Total Av. length 54 6 5 5 ■ 33 37 40 37 • . 13 . 45 . 23 39 5 3 '3 ■ 34 35 37 38 . 45 83 8-2 l62 DISCOVERY REPORTS MALES FEMALES Date April 7, 1927 Locality Bransfield Strait Date April 7. 1927 Locality Bransfield Strait St. No Net 207 C N 70 H 0-5 m Position _(62" S4''00 S, I S9=So'-3oW St. No. Net 207 C N 70 H 0-5 m. Position /62^ 54''00 .S. I 59° So'-30 W Surface T (From —0-78^ \ to -o-86° C Surface T. /From —0-78° \ to -0-85° C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 1234567 A B C D E F G 12345 6 7 A B C D E F G 25 I . . . , 26 I . . . . I I 28 I I 2 . 2 28 I . . . . I 1 29 I 29 3 . . . . . 2 I . 3 30 I 2 30 6 . . . . •24. 6 31 1 3' I . . . . I I 32 I 32 3 • • . ■ 2 I 3 33 3 I 2 I 4 33 5 . . . . • 23. 5 34 3 34 2 . . . . 2 2 36 I I 2 35 5 . . . . • 23. 5 37 2 2 . 3 37 • • 3 ■ 3 38 3 2 . 4 38 .31.. 1 I 2 . 4 39 3 3 39 3 I • ■ 2 2 4 41 3 2 I 4 5 40 ■ ■ 3 ■ 2 I 3 42 2 2 41 I I I 44 1 2 3 43 . I* . . . I I 45 I 1 44 2* 2 2 47 ■ 4 ■ . 4 . 4 45 1 50 I ..21. 3 46 • 3* • . 3 3 SI 2 . I • 2 47 ■ 4* 4 4 53 . . 5 .33. 6 48 . 3* • • 3 3 49 • 3* • ■ 3 3 Total I 16 14 14 i i . . 3 16 10 7 12 5 . 53 5° I 4* • I 4 5 Av. length 28 32 39 46 5- 29 33 39 41 49 52 . 52 . I* I I Total 30 6 2 5 . I 22 . 11 20 5 7 I 22 66 Av, length 32 39 39 41 • so 48 • 3J 33 39 40 so 48 Date April 7, 1927 Locality Bransfield Strait Date April 7, 1927 Locality Bransfield Strait St. No. Net 207 D N 70 H 0-5 m. Position (62' 54'-oo S, I S9°So''3oW St. No. Net 207 D N 70 H 0-5 m. Position J62" 54' 00 S, i S9"5o'-3oW Surface T. 1 From —0-78° t to -o-86° C. Surface T. /From —0-78° 1 to -086° C. Stages Total Stages Total Length in mm. in sample Length in mm. - in sample 1234 5 6 7 ABC D E F G 12345 6 7 A B C D E F G 30 2 . I 2 27 I . . . . . I . . I 33 I I 2 30 2 . . . . I I 2 34 I 31 3 . . . . . 1 2 . 3 36 I 32 I . . . . I I 38 I 33 2 . . . . 2 . 2 47 I I I 34 2 . . . . I 2 48 1 I I 36 I . . . . I I 49 2 I I 2 42 . * I I 51 2 . 2 . 2 45 • 4* ■ • 4 4 53 1 I I 46 • *: ■ • 4 4 54 J . . 2 . 2 '*2 • 4* • • 4 4 55 J . 2 . 2 48 • 9» • • 9 9 56 . I I 49 . IO» . 10 lO 50 . 8* . . 8 8 Total - 3 4 4 J 2 241. ! 9 ■ 19 Av. length • 31 35 48 5 > 51 • 32 33 38 .4 i S3 . Total Av. length 12 ... . 32 ... . • 40 .48. • ■ 40 • • 49 52 Date April 7, 1927 Locality Bransfield Strait St. No Net 207 E N 70 H o-s m. Position (b2° S4''00 S, (. s9°5o'-3oW Date April 7, 1927 Locality Bransfield Strait Surface T. ( From — 0-78° St. No. 207 E Position (62" 54'-oo S, (. to -0-86'' C. Net Surface T. /From —0-78° (, to -o-86° C. Length in mm. Stages Total in sample 6 7 A B C D E F G Length in mm. Stages Total in sample 42 44 2 2 : : : I 2 1234 ! 6 7 A B C D 1 1 F G . 40 . . I . I I 46 . I I 52 I I I 47 I' I Total I I I I 2 Total . . 2 . . ■ 4 . . . 2 2 . 2 6 Av. length . . 40 52 . . . 40 52 . Av. length ■ ■ 44 ■ • • 45 • • • 44 44 • 46 1 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 163 MALES FEMALES Date April 7, 1927 Locality Bransfield Strait Date April 7, 1927 Locality Bransfield Strait St. No. Net 207 F ^ 70 H 0-5 m. Position Surface T. ("62° 54'-oo S, \ 59°5o'-3oW /From -0-78° \ to -086° C. St. No. Net 207 F ^i 70 H 0-5 Position Surface T. (62° S4'oo S, I 59= 50'-30 W ( From —0-78' I to -o-Sd" C. Stages Total Suges Total Length in mm. in sample Length in mm. m sample I 2 3 4 £ t 7 A B C D E F G 1234 5 6 7 ABODE F G 29 • 29 1 . . . . . I . . 1 30 . I 30 4 • ■ • . . 4 . . 4 31 . I 31 2 . . . 2 . 32 32 3 ■ ■ • 2 1 3 35 2 • 3 " i 2 . . . }(• I 34 2 . . . 1 I 39 1 35 4 • • 3 I 4 40 I 3' 2 42 I 38 48 2 2 39 SI I 40 52 I 2 3 42 53 2 . 2 43 I* 54 I 3 ■ 4 44 . . 2* . 2 I I 45 • • ^l . 5 5 56 57 I I I I 46 47 ■ ■ 4* . . 6* ■ 4 . 6 S 6 58 I I 48 ■ • '! . 7 7 49 50 51 : : f. . . I* • 4 • 7 1 4 7 1 Total Av. length I 4 5 5 I 32 31 37 49 5 i 5 I 253. 31 33 38 .4 4 13 . 9 54 • 27 Total 23 • I • • 37 . II II 2 • 37 62 Av. length 34 . 46 42 . . 47 . 34 33 44 39 ■ 47 Date April 7, 1927 Locality Bransfield Strait Date April 7, 1927 Locality Bransfield Strait (62° 54'-oo S, ^ 59 So'-3oW St. No. Net 207 G N 70 H 0-5 m Position (■62° 54'oo S, \ 59°So'-3oW St. No. Net 207 G N 70 H 0-5 Position m. Surface T. (From —0-78^ 1 to -086° C. Surface T 1, to -0-86 ' C. Stages Total Stages Total Length in mm. in sample Length in mm. in Sample 1234 5 6 7 A B C D E F G I 2 3 t 5 6 7 A B C D E F G 27 X I . . . • 40 42 : : ; I I I I Total I I . . . I 43 45 . . I I I I 1 Av. length . 27 . . 27 . . . Total . . 4 . . . 3 I 4 Av. length . • 43 . . • 43 42 . . Date April 7, 1927 Ixjcality Bransfield Strait St. No. 207 H /62°54-opS Net N 70 H 0-. m. Surface T. 1 59° 50 30 W /From —0-78" t. to -086° C. Date April 7, 1927 Locality Bransfield Strait Stages Total St. No. 207 H Position Surface T. f62° 54'oo S, Length in Net N 70 H 0-5 m I 59" 50 30 W /From —0-78 \ to -086° C. in mm. I 2 3 4567 A B C D E F G sample 30 31 I Stages Total I Length in mm. in sample 33 34 1 1234 5 6 7 A B C D E F G 35 I . I . . I 43 I . . . 47 . . I . I I 44 3 . . . • 3 48 I 45 50 1 I 3 46 47 ^ : : 3 . 2 2 . 2 . . 3 52 3 I 48 49 . . ■ 4 . . . 2 • • 4 . . 2 2 55 57 1 . . . I I I 50 51 . • • 4 . . . 2 . . 4 . 2 4 2 Total . I I 6 5 . I I I 10 13 Total 7 • • 9 . . 19 .526 3 • 19 35 Av. length . 35 47 52 5 2 • 35 • 47 1 8 52 . Av. length 32 . . 4 5 . . 48 . 33 32 45 4 6 . 48 164 DISCOVERY REPORTS MALES FEMALES Date April 7, 1927 Locality Bransfield Strait Surface T. {•^r-'o^sl? C. Date April 7, 1927 Locality Bransfield Strait Net''" l?'7?Ho-sm. ^-tion \;'^l^^^^^ Surface T. {•'r--o.86'?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 25 28 29 30 31 35 11 49 SO 52 S3 I . . I I I I 2 2 2 I 2 I I I 29 32 33 34 35 36 38 39 42 43 45 46 47 48 49 . I . 2 X 2 I I 1 3 I -I 2 I 3 3 3 3 2 2 I I 2 ■ ' 2 . . . I . . 2 . I I . . . I 2 . . . . I . . . . 3* 3* ::;::: ^: 2* . . 11.. I . . . 3 3 3 3 2 TotaJ Av. length 26233. • 27 32 35 49 52 . . 46. .42. 29 33 . . 50 51 . 16 Total Av. length 10 I 3 ... 14 34 39 42 . . . 47 . 4622 .14 ■ 32 36 43 40 . 47 28 Date April 7, 1927 Locality Bransfield Strait Net''"- ^°^7oHo-5m P°-- r^S9^5o°3oW Surface T. {^^^^S^lT^. Date April 7, 1927 Locality Bransfield Strait Net"""- N'7oHo-5m. P°-ion {^^.^^^^^ Surface T. {^'"^^S^lfc, 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 18 24 I I I I I I 31 32 I I I I I Total Av. length 2 21 2 21 2 Total Av. length 2 • 32 2 32 2 Date April 7, 1927 Locality Bransfield Strait ^'et'^" N'7o''Ho-5m. ^osi^io. {"^i?'^^^ Surface T. {^Z"'--^l'c. Date April 7, 1927 Locality Bransfield Strait lie^- N'7o^Ho-sm. Pos.t.on {'^H^^^^ Surface T. {^r--o°86'='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 32 34 35 38 47 50 51 . 2 . z . 2 2 I 2 28 30 31 33 35 40 43 44 46 47 48 2 2 2 . 2 2 2 4 2 I I I 2 4 5 I 2 ...... 1* I 1* ::.::: 4* 5* I* I . . . 4 5 ' Total Av. length .4122. . . 33 35 43 SI ■ ■ . 5 . I I 2 . .33 -38 47 51 . 9 Total Av. length II . 2 , . .12 31 .42 . . .46 . 562. .12 . 29 32 42 . . 46 25 Date April 7, 1927 Locality Bransfield Strait ^'et^°- ??^7?Ho-5m. P-^'io" {'^^'^'^SO^'o W Surface T. {^-^"sI^C. Date April 7, 1927 Locality Bransfield Strait lfet''°- N^7?Ho-5m. P-"- {'ni-^'so^a'c' W Surface T. {^'^?^S^lfc, 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 26 30 32 34 35 2 i ■ 2 2 I 2 I so 53 I 2 . I I 2 I 2 I I I I Total Av. length . I ... 3 . . I ... 3 . .28 . .11 4 Total Av. length . 61.... 7 ^' .-.-.... •^ THE DEVELOPMENT AND LIFE-HISTORY OF KRILL i6s MALES Date April 14, igzS St. No. WS 196 Net N 100 B 103-0 m. Locality Position Surface T. S. Georgia fS4°36'S, \ 38'38'W 2-73' C. FEMALES Date April 14, 1928 St. No. WS 196 Net N 100 B 103-0 m. Locality Position Surface T. S. Georgia /S4°36' S, 1 38=3S-W 2-73' C. Length in nun. Stages 34567 ABCDEFG Total in sample Length in mm. Stages 234567 ABCDEFG Total in sample 24 26 27 28 29 30 31 32 33 34 36 40 46 48 Total Av. length 4 24 4 2 26 30 34 43 48 23 25 26 27 28 29 30 31 34 Total Av. length 6 22 26 29 8 17 3 28 29 32 2S 48 Date April 28, 1929 St. No. WS 427 Net N 100 B 140-0 m. Locality Position S. Georgia /sa" 34' s, ( 40" 10' W Surface T. 1-94° C. Length Stages Total in sample in mm. 1234567 ABCDEFG 48 50 54 I I z . . . . I . . I I I I Total Av. length I ... 48 . 2 . 52 I . 2 . . . . 48 . 52 3 Date April s. 1931 St. No. 663 Net TYF V 250-0 r Locality Position S. Georgia /53° 34i' S, (_ 30°2srW Surface T. 051° C. Date April 5, 1931 St. No. 663 Net TYF V 250-0 m. S. Georgia .fS3° 34i' S, l 30° 255' W Surface T. 051' C. Locality Position Length in mm. 25 27 29 31 33 34 35 36 37 41 44 45 49 50 55 Stages 1234567 ABCDEFG Total in sample Length in mm. 23 27 28 29 30 31 32 33 35 36 39 41 44 48 51 Stages 234567 ABCDEFG Total in sample Total Av. length 8 10 I 3 2 I I 5 13 30 35 45 45 SO 55 55 30 31 48 43 49 53 55 26 Total Av. length 29 19 2 I 31 42 41 Date April 17, 193 1 St. No. 665 Net TYF B 250-0 m. Locality Position Surface T. S. Georgia /5i°4ii' S, I 29° 581' W 250° C. Date April 17, 193 1 St. No. 665 Net TYF B 250-0 m. Locality Position Surface T. S. Georgia f5i°4iJ'S, t 29°58i'W Length in mm. 26 27 28 29 30 3t 32 33 34 35 37 38 49 Total Av. length Stages 234567 ABCDEFG 3 29 I 25 31 38 .49 3 29 1 25 31 35 Total in sample Length in mm. 25 26 27 28 30 3t 32 33 34 35 36 37 38 40 55 Stages 1234567 ABCDEFG Total 6 27 Av. length 26 33 5 26 2 127 32 39 Total in sample 1 66 DISCOVERY REPORTS MALES Date April 17-18, 1931 St. No. 666 Net TYF B 320-0 m. Locality Position Surface T. S. Georgia 149° SH' S, 1, 29° 52J' W 271= c. Length in mm. Stages Total in sample 1234567 A B C D E F G 28 30 41 I 1 I . . . . I I I . . . . I I I Total Av. length . 21.... . 29 41 . . . . . 21.... . 29 41 • ■ - ■ 3 _ Date May 18, 1929 St. No. WS 434 Net N 100 B 91-0 m. Locality S. Georgia Position /53° 10' S, Surface T, /5. \ 34°o8'W 3-26° C.(?) Length in mm. Stages 1234567 A B C D E F G Total in sample FEMALES Date May 18, 1929 St. No. WS 434 Net N 100 B 91-0 m. Locality Position Surface T. S. Georgia (53° 10' S, 1 34°o8'W 326° C.(?) Length in mm. 567 A B C D E F G Total in sample 30 31 33 34 35 35 39 43 46 28 29 31 34 37 39 40 Total Av. length 4 3 3 32 37 41 424 32 36 40 Total Av. length 9 I 33 40 5 4 I 31 37 40 Date May 6, 1934 St. No. 1359 Net N too H 5-0 m. Locality Position Surface T. South of Cape Town /63°4s'-2 S, (, 36° 4i'-i E -1-37° C. Date May 6, St. No. 1359 Net N 100 , 1934 H 5-0 m. Locality Position Surface T. South of Cape Town (-63° 45'-2 S, ( 36°4l'lE -1-37° C. Length in mm. Length in mm. 26 27 28 29 30 31 32 33 36 37 39 40 41 Stages 34567 ABCDEFG Total in sample 23 24 25 26 28 30 31 32 33 34 35 36 37 39 40 41 43 45 34567 ABCDEFG Total in sample Total Av. length 9 17 4 28 32 40 4 24 2 29 32 41 Total Av. length 22 25 5 28 31 41 7 38 5 2 27 35 37 44 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 167 MALES FEMALES Date June 2, 1936 St. No. 1780 Net N 100 B 114-0 m. Locality Position S.E. of Bouvet I. /56° lo'-i S, I. 00' o8'-8 W Date June 2, 1936 St. No. 1780 Net NlooBli4-or Locality Position S.E. of Bouvet i. (56° lo'-i S, \ 00° o8'-8 W Length in mm. Stages Total 1234557 ABCDEFG| sample Length in mm. Stages I 234567 ABCDEFG Total in sample 33 38 39 42 43 46 Total Av. length I 2 4 33 39 44 33 38 43 27 35 43 44 Total Av. length 27 35 44 27 35 44 Date June 2, 1936 St. No. 1781 Net N 100 B 128-0 m., 500-150 m., 800-450 m. Locality Position S.E. of Bouvet I f57°4l'-8S. ( 00° l9'-8 W Length in mm. 30 31 34 35 36 38 40 41 42 43 45 49 Total Av. length Stages 34567 2 6 15 I 30 36 42 40 ABCDEFG sample 2 10 10 2 30 37 43 45 Date June 3, 1936 St. No. 1782 Net N 100 B 500-150 m., 700-400 m. Locality Position S.E. of Bouvet I. 758°44'-6S \ 00° 01 -5 E Length in mm. 33 35 48 Total Av. length Stages Total 33 35 44 Date Jime 2, 1936 St. No. 1781 Net N 100 B 128-0 m., 500-150 m., 800-450 m. Locality Position S.E. of Bouvet I. ^57°4„-'-8S (. 00° 19 -8 W Length in mm. 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 49 Stages 34567 ABCDEFG Total Av. ength 3 21 51 31 35 41 16 32 22 5 32 38 42 47 Total in sample Date June 3, 1936 St. No. 1782 Net N too B 500-150 m., 700-400 m. Locality Position S.E of Bouvet I /58° 44'-6 S, \ 00^ ol'-5 E 4 5 6 7 ABCDEFG, sample Length in mm. 36 38 40 41 43 49 Total I Av. length Stages 34567 ABCDEFG 39 42 . • • • ] • . 36 40 46 Total in sample Date June 6, 1936 St. No. 1786 Net TYF B 800-400 m. Length in mm 37 42 44 Locality Position Total Av. length S.E. of Bouvet I. \ H°I4-7E Stages I Total -[ in 6 7 ABCDEFG! sample Date St. No. Net Length in mm. 39 40 42 June 6, 1936 1787 TYF B 800-400 m. Total Av. length Locality Position Surface T. S.E. of Bouvet I fsS" o5'-9 S, I 12° 48'-6 E -i-5o'-C. Stages 1234567 ABCDEFG Total in sample 1 68 DISCOVERY REPORTS MALES FEMALES Date June 7, 1936 St. No. 1788 Net N 100 B 270-0 m. Locality Position Surface T. S.E. of Bouvet 1. /S9° ir-7 s, 1. 17° 01 -9 E -i-8o°C. Date June 7, 1936 Locality St. No. 1788 n . . Net N 100 B 270-0 m. P'""""" Surface T. S.E. of Bouvet I. fS9°ll'-7S \ 17° 01 -9 E -l-8o°C. Length in mm. Stages Total in sample Length in mm. Stages Total in sample 123456 7 A B C D E F G 1234567 A B C D E F G 40 . . . 1 . . . . . . I . . . I 38 2 Total Av. length . . . I . . . . . . 40 . . . . . . I . . . . . . 40 . . . I Total Av. length . . . 2 38 . . 38 38 • 1 Date June lo, 1936 Locality St. No. 1794 Net N 100 B 360-200 m. Position Cape Town to Ice Edge /52° 37'-5 S, I 18' 22'-7 E Length in mm. Stages Total in sample 1234567 A B C D E F G 47 I . . . . . . . . I . . I Total Av. length . . 47 ■ • • ■ I .... 47 ■ ■ I Date July 12, 1938 St. No. 2362 M., /N 100 H o-s m., Net 1 TYFB4S0-430 Locality ^ Position Cape Town to Ice Edge f54° 59'-3 S, \ 01 10 -2 E Date July 12, 1938 Locality St. No. 2362 Net r-i-°Y°F^r5i-«om. P°-- Cape Town to Ice Edge fS4' 59''3 S, (, oi 10 -2 E Length in mm. Stages Total in sample Length in mm. Stages Total in sample 123456 7 A B C D E F G 1234567IABCD E F G 30 35 54 I . . . I I I , . . . . . I . . . . I 1 I I 26 29 30 31 35 40 11. . . . . 2 . I 2 I I I Total Av. length II.. . . 30 35 • • I 54 2 . I ■ • 54 3 I . . . . . . I . . . . . I . . Total Av. length • 3 4 ■ ■ . ■ . 28 34 . . . . 151. 26 31 40 . . . . ! 7 ... 1 Date July 13, 1938 Locality St. No. 2365 Net TYFB35C^20om. p^^^^.^^ Cape Town to Ice Edge (53" 23'4S, I 04" 5o'-5 E Date July 16, 1938 St. No. 2372 Net N 100 B 102-0 m. Locality' Position Cape Town to Ice Edge fSi" io'-6 S, (. 15° 47'-5 E o,,„„ Total in sample Length in mm. Stages Total in sample Length in mm. ^ ^ . -A™lAD/-•^^ E F G 123456 7 A B C D E F G ^0 1 ■ ' I 2 I 48 . . . . I . . I I I I 41 43 • 2 2 . . I I Total Av. length . 2 . ■ 47 • • . . . . 2 . . .... 47 • ■ 2 Total Av. length ■ • 4 4 . . . . . 41 .... J ... 41 .. . 4 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 169 Table 20. Egg Measurements from 839 Females Date August 28, 1928 | St. No. • WS264 j Depth N 70 B 97-0 m. Date September 5, 1928 St. No. WS 277 Depth N 70 B 124-0 m. Length in mm. Egg classes Total Length in mm. Egg classes 1 ) I 2 3 4 I i 3 4 Total 005-012 0-13-0-24 0-25-0-48 0-49-0-70 0-05-0-12 0-13-0-24 0-2S-0-48 0-49- 0-70 33 34 39 40 41 42 43 4S 55 005 o-os (2) 0-07 0-05 0-09 O-O7-0-09 0-07 0-09 (2) 0-09 0-09-0-I4 0-09- 014 (2) 0-18- 0- 023 18 2 3 31 33 35 36 40 41 43 44 45 ■*5 48 49 oos 0-05 (2) 0-05 0-05 0-05-0-07 0-05-0-07 0'05-oo7 005-007 0-05-0-07 0-07-0- 09 0-05-0-09 009-0-14 0-09-0- 1 1 009-0-18 J 2 2 0-18-0-25 0-23 Total in sample Average diameter of eggs (mm.) 15 007 Total in sample Average diameter of eggs (mm.) 19 0-12 Date September 17, 1928 St. No. WS 282 Depth N 70 B 137-0 m. Date August 18, 1938 St. No. 2396 Depth N too B 109-0 m. 33 35 36 11 39 42 tl 51 55 0-05 0-05 009 o-os 0-05 0-05 009 (2) 0-05 0-07 0-09 O-II o- 09-0- 14 012 (3) 0-09 t 2 I 3 I 40 009 I Total in sample Average diameter of eggs (mm.) I 0-09 Date August 19, 1938 Sj. No. 2399 Depth N 100 H 0-5 m. NiooB-l'**-""' 1,300-150 m. Total in sample Average diameter of eggs (mm.) 17 33 35 36 38 40 42 0-05 005 O'og 009 0-07 009-0- 14 Date September 24, 1938 St. No. 2430 Depth N 100 H 0-5 m. N 100 B 1 17-0 m. Total in sample 7 Average diameter of eggs (mm.) • o-o8 46 0-09-0-14 0-09-0-14 I I Total in sample Average diameter of eggs (mm.) 2 012 Date August 22, 1938 St. No. 2408 Depth N 100 H 5-0 m. Date October 2, 1928 St. No. WS 290 Depth N 70 H 0-5 m. 39 40 41 42 0-05 oos o-os (4) 0-07 (2) 1 4 2 37 0-05 . - 1 I Total in sample Average diameter of eggs (mm.) 8 0-06 Total in sample Average diameter of eggs (mm.) I 0-05 Date August 24, 1938 St. No. 2412 Depth N 100 H 5-0 m. N 100 B 107-0 m. Date October 4, 1928 St. No. WS 29s Depth N 100 B 97-0 m. 42 0-09 1 1 I 37 0-05 1 1 1 I I 005 Total in sample Average diameter of eggs (mm.) I 0-09 Total in sample Average diameter of eggs (mm.) g-a I70 DISCOVERY REPORTS Length 33 34 35 36 38 39 40 4' 42 43 45 46 5° 51 52 56 34 35 35 36 38 39 40 42 43 44 45 46 48 49 50 51 35 38 44 45 Date St. No. Depth October 5, 1928 WS 298 N 100 B 94-0 m. Egg classes 005 007 0-09 0-07 007 O'og o-og 009 0-25-0-48 0-14- 14 •14 -o 18 ■18 ■18 0-23 0-18-0-2:? 0-14-0-18 0-23 Total in sample Average diameter of eggs (mm.) Date St. No. Depth October 6, 1928 WS 304 N 1 00 B II 0-0 m 30 0-05 34 0-05 40 0-05 44 007 Total in sample Average diameter of eggs (mm.) Date St. No. Depth October 16-17, 1930 453 N 100 13 164-0 m. 0-05 0-05 Total in sample Average diameter of eggs (mm.) Date St. No. Depth October 17, 1930 454 N 70 B 192-0 m. 0-05 0-09 o-os (3) 0-09 0-05 0-07 0-09 (2) 0-09 (2) o-og (3) 009 (3) O'll o-ll 0-09 o-ll o-og 0-09 0-09 o-ii Total in sample Average diameter of eggs (mm.) Date St. No. Depth October 18, 1930 455 N 100 B 1 16-0 m. 0-07 0-07 0-09 0-09 0-09 Total in sample Average diameter of eggs (mm.) Total 19 0-I3 4 0-06 27 0-09 5 0-08 Length in mm. 38 39 40 JS3 34 35 36 37 38 39 40 46 48 50 35 37 40 49 Date St. No. Depth October 19, 1930 459 N too B) ,0, „„ N70B / '83-0 m. Egg classes 0-05 0-07 0-09 o-ll 0-07 0-09 0-09 0-07 0-25-0-48 Total in sample Average diameter of eggs (mm.) Date St. No. Depth October 22, 1930 461 G N 100 B 700-560 (315) m. 0-05 0-07 0-07 (2) 0-05 0-09 o-il 00s 0-07 (2) 0-09 0-05 0-07 0-09 o-il 0-07 (3 009 (3 o-ii {2. 005 (3 007 (2, O-II (2' o-os (2; 0-07 (4; 0-09 (6' 0-05 (2! 0-07 (5: 0-09 (5, o-ii 005 (3 0-09 (4 O-II 0-05 (3 007 (3 0-09 (4; O-II 0-05 007 0-09 (5: 00s (2 0-07 (3 0-09 (5 O-II (3 o-os (2 0-07 0-09 O-II 005 0'07 0-09 (2) 0-09 O-II 0-14 (2) o- 1 1-0-23 14(2) 14(2) Total in sample Average diameter of eggs (mm.) Date St. No. Depth October 22, 1930 461 D N 100 B 490-385 m. Total 0-07 S 14 121 0-09 0-07 o-ii O-II O-II Total in sample Average diameter of eggs (mm.) 4 O'lO Date St. No. Depth October 23, 1930 462 N 100 Bl , B / 9°"° ■"■ N70I Total in sample Average diameter of eggs (mm.) THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 171 Date October 25, 1930 St. No. 463 Depth N 100 B 132-0 m. Date November 29, 1930 St. No. 523 Depth N 100 B 157-0 m. Lengtli in mm. Egg classes Total Length in mm. Egg classes Total I 2 3 4 I 2 3 4 0-0S-0-I2 OI3-0-24 0-25-0-48 0-49-0-70 0-05-0-12 0-13-0-24 0-25-048 0-49-0-70 44 0-O9 I . 41 44 ti so 0-14-0-32 0-18 0-18-0-27 023 0-18-0-23 0-14-0-45 I 2 I I I Total in sample Average diameter of eggs (mm.) I 009 Total in sample Average diameter of eggs (mm.) 6 0-23 Date October 25, 1930 St. No. 464 Depth N 100 H 67 (-0) m. Date November 6, 1932 St. No. 1009 Depth NiooBiio-om. 48 O-II I Total in sample Average diameter of eggs (mm.) I o-ii 27 0-07 0-14 (2) 0-14 0-14-0-18 0-14 018 ■ I 2 2 2 2 Z I 5 z 2 I I 29 30 31 32 34 37 38 39 40 43 007 (2) 007 0-09 0-07 0-09 0-09 (2) O-II 0-09 (2) o-li Date October 26, 1930 St. No. 465 Depth N 100 B 1 13-0 m. 39 40 0-05 (2) O-07 2 I Total in sample Average diameter of eggs (mm.) 3 0-06 0-14 Total in sample Average diameter of eggs (mm.) 23 oil Date November rg, 1929 St. No. WS Alongside Deception I. Depth N 100 B 0-5 m. It 48 o-l8 014(3) 0-27 1 3 I Date December 19, 1926 St. No. 125 Depth N 100 H 70 m. Total in sample Average diameter of eggs (mm.) 5 0-I7 43 44 45 46 47 48 49 50 SI 52 54 S6 11 60 61 0- 09 0-18 1 1 Eggs recorded at WS 477-WS 487 at depths between 1000-500 m. 2nd nauplii recorded at WS 480-WS 486 at depths between 750-200 m. 0-23 0-23 o-i8 0-23 0-18 (2) 0-23 (2) 0-14 0-18 (2) 0-23 0-18 0-23 0-23 (2) 0-18 (2) 0-23 (4) 0-18 0-23 (3) 032 0-27 0-27 (4) 032 (3) 0- 0-36- 49 -0-64 I 2 4 S 2 3 6 13 3 2 I I Date November 13, 1930 St. No. 480 Depth N 100 B 161-0 m. N 70 V 1000-750 m. 46 041 I Total in sample Average diameter of eggs (mm.) I 0-41 o-i8 0-41 0-4S 0-27 Date November 19, 1930 St. No. 494 Depth N too B 160-0 m. 0-27 36 37 38 39 40 41 42 43 44 45 46 47 48 49 SI 014 o-\t 014 (2) 0-14 01 4 (4) o-i8 (3) 0-14 (2) o-i8 014(6) 014 (3) 018(5) 014 (6) 018 0-14 (2) 01 8 (6) 014 018(4) 014(2) 018 0'14 01 4 (4) 018 014 I 2 2 2 7 3 6 8 7 8 5 3 1 5 I Total in sample Average diameter of eggs (mm.) SO 0-24 Date December 11, 1930 St. No. 527 Depth N 450 H i22(-o) m. 42 0-23-0-55 1 I Total in sample Average diameter in eggs (mm.) I 0-39 Date December 17, 1930 St. No. 534 Depth N 100 B 72-0 m. 41 0-23 1 I Total in sample Average diameter of eggs (nun.) 61 0-15 Total in sample .Average diameter of eggs (mm.) I 0-23 172 DISCOVERY REPORTS Date December 18, 1930 Eggs and 2nd nauplius recorded at St. 540 at (500-100) m. St No. 535 Depth N 70 B 0 m. Date St. No. December 19-20, 1930 541 Egg classes Depth N 100 B 108-0 m. Length in mm. 2 3 4 Total 1 Length in mm. Egg classes Total 0-05- O.I2 0- 13-0-24 0-25 0-48 0-49- 0-70 I 2 3 4 37 023 0.23-0-45 1 2 0-05-0-12 0- 13-0-24 0-25-0.48 0-49-0.70 44 47 0-32 0-27 1 • I 32 44 0-23 027 I I Total in sample 4 Average diameter of eggs (mm.) 0-29 Total in sample 2 Average diameter of eggs (mm,) 025 Date December 20, 1930 St, No. 536 St. No. W^ r Depth N 100 B 122-0 m. Depth N 100 B 1(14-0 m. 36 0-23 I 46 0.27 I Total in sample I Total in sample , Average diameter of eggs (mm.) 0-23 Average diameter of eggs (mm,) 0-27 Eggs and 2nd nauplius recorded at .St, 546 at (500-100) m. Date December 21, 1930 Date December 19, 1930 St. No. 548 St. No. 538 Depth N 100 B 102-0 m. Depth N 100 B 137-0 m. 39 40 41 0-14 0-14 018 0-23 (2) 0-14 48 023 1 . 4 2 Total in sample I Average diameter of eggs (rr m.) 0-23 43 o-i8 0-14 0-18 (3) 5 0., 45 0-14 (3) 0- 27 6 Eggs and ist r auplius recorded from .St. 537 at (1000- 500) m. 0-18 (2) Eggs recorded at St. 538 at (750-500) m. 46 47 48 0-14 0-18 (2) 0-23 0-14 0-18 I 3 5 Date December 19, 1930 St. No. 539 0-23 (3) Depth NiooBI ,,, „„ N 70 B 1 '37-° m. 49 0-i8 023 2 SO 0-18 0-23 (2) ■" 3 30 023 31 32 023 (2) 0-14- 0-36 (2) 2 3 52 53 55 0-1 8 0-23 o- 27 I 35 0-23 0-27-0-41 2 Total in sample 36 36 0-23-0-50 I Average diameter of eggs (mm.) 0-19 37 o-ls-0-45 38 o-OQ- .011 0-23- 0-27 (2) 5 0-41-0-45 Date December 21-22, 1930 39 0-23-0-41 0-23-0-45 2 St. No, Depth 549 N 100 B 115-0 m. 0-27-0-55 41 0 23 (2) 2 33 0-09 (2) 1 2 42 0- 1 8-03 2 0-23-0-55 0-27-0-41 0-27-0-45 4 ^1 0-09 0-09 0.14 0-I4 (3) 014 (2) I 2 3 2 43 0-32-0-45 I 39 0.09 014 3 44 0-23 0-18 45 0-32-0-45 I 40 014 3 46 0-23-0-32 I 0.18 (2) 47 018-0-27 I 41 0-18 2 48 0-27 I 0-23 SO 0-27-0-45 I 42 0-14 I SI 0-23-0-55 I 43 0-09 I 52 0-18-0-32 0-36 2 44 0-09 0-I4 0-18 3 55 0-36- 0- 64 (2) 45 0-14(2) 0-18 3 46 0-23 I Total in sample 35 '^l o-iS I Average diameter of eggs (r nm ) 0.32 48 52 0-14 023 (2) 2 Total in sample 31 Eggs and 1st nauplius recorded at St. 539 at (looo-ioc m.) Average diameter of eggs (mm.) OIS Date December 19, 1930 St. No. S40 Date December 29, 1930 Depth N 100 B 155-0 m. St, No, 558 N 70 V 500-250 m. Depth N 100 B 146-0 m. 32 0-23 I 40 0-18 I 34 027 I 0-23 (2) 2 Total in sample 2 Total in sample 3 Average diameter of eggs ( -nm.) 0-25 Average diameter of eggs (mm.) 0-21 1 1 Eggs and 2nd nauplius recorded at St. 558 at (200-50) m. THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 173 Date St. No. Depth December 30, 1930 559 N 100 B 113-0 m. Egg classes Length in nun. 33 43 45 46 47 37 42 45 47 SO 47 50 45 47 48 49 50 51 53 54 56 57 S8 59 013-024 o-l8 0-09-0-I8 018 0-09-0-I8 018 0-25-0-48 0-49-0-70 Total in sample Average diameter of eggs (mm.) | 016 Total Date St. No. Depth December 30, 1930 560 N 100 B 155-0 m. 0-09 0-09 o-i8 023 028 Total in sample Average diameter of eggs (mm.) Date St. No. Depth December 31, 1930 561 N 100 B 137-0 m. Date St. No. Depth December 31, 1930 56a N 100 B 113-0 m. 032 Total in sample Average diameter of eggs (mm.) Date St. No Depth January 23, 1929 WS 373 EE N 100 B 70-0 m. 045 0-23-0-45 0-4I-0-45 027 ' 0-32 0-4S 0-36-0-4S 0-32-0-45 0-4I-0-S0 0-27-0-45 036 0-32-0-50 0-27-0-32 0-4S-0-50 03 5-0-41 0-27- 045 (2) 0-27-0-32 0-45 (4) 01 8-045 0-45 (2) 0-4I-0-50 0-35-0-50 0-36-0-59 032- 0-45 (2) 0-4S (2) 0-27-0-45 036- 0-4S (2) 0-36-0-59 0-32-o-so 032 0-36-O-45 0-23-0-45 o-27-0'36 0-35-0'SO 0-32-0-45 Date St. No. Depth January 21-22, 1930 N 100 B loo-o m. Length in mm. Egg classes 50 51 5 0-2O Total in sample 2 Average diameter of eggs (mm.) 0-28 50(2) Total in sample Average diameter of eggs (mm.) 46 0-40 31 33 35 36 37 38 39 40 41 42 44 45 46 47 48 45 48 38 48 0'2S-o-48 032 032 0'49-0'70 Total in sample Average diameter of eggs (mm.) Total Date St. No. Depth January 24-25, 1930 312 N 100 B 150-0 m. Total in sample Average diameter of eggs (mm.) Date St. No. Depth January 8, 193 1 N 100 B 132-0 m. Total in sample Average diameter of eggs (mm.) Date St. No. Depth January 8, 1931 575 „ N loo B 97-0 m. 018 018 0-23 (2) 0-23 o-i8 (2) o-i8-0'23 o- 14-0-27 0-23 0-27 0-14- 0-23 (3) 0-23 (3) 023 0-14-0-27 0-09-0-23 0-14- 0-23 (2) 0-23 (2) 0-23 (2) 0-23 0-14-0-27 0-32 0-23-0-27 0-27 (2) 0-27 0-35-0-45 027 0-27 0-27- 0-45 (2) 0-36-0-45 0-23-0-32 0-32 Total in sample Average diameter of eggs (mm.) 41 0-2S Date St. No. Depth January 10, 1931 580 N 100 B 128-0 m. o-l8-0'36 0-18-0-45 Total in sample Average diameter of eggs (mm.) Date St. No. Depth January 12, 1931 583 N 100 B 139-0 m. 0'I4 0-09-0-18 Total in sample Average diameter of eggs (mm.) 174 DISCOVERY REPORTS Date January 13, 1931 St. No. 584 Depth N 100 B 165-0 m. Date January 25-26, 193 1 St. No. WS 537 Depth N 100 B 67-0 m. Length in mm. Egg classes Total Length in mm. Egg classes Total I 2 3 4 I 2 3 4 005-0-I2 o- 13-0-24 0-25-0-48 0-49-0-70 0-05-0-12 0-13-0-24 0-25-048 0-49-0-70 48 . 0-23-048 I 47 SI 023 0-23-0-25 1 1 Total in sample Average diameter of eggs (mm.) I 0-34 Total in sample Average diameter of eggs (mm.) 2 0-25 Eggs recorded at St. 585 at Cso~o) m. Date January 14, 1931 St. No. 590 Depth N 100 B Qo-o m. Date February 22, 1928 St. No. WS 152 Depth N 100 B iio-om. 48 50 , 0-18-0-45 0-18-0-45 I 2 0-32 42 44 45 46 47 49 50 51 52 53 54 55 56 59 60 0-18 0-18-0-23 0-18 0-18-0-23 0 18 023 0-23 0-27-0-55 0-32-0-45 0-23-0-55 0-27-0-45 0-23-0-36 0-23-0-45 0-23-0-45 0-36-0-55 0 32- 0-45 (2) 0-27-0-SS 0-18-0-45 0-45 0-36-0-45 0-32-0-45 0-41-0-55 0-32 0-45 0-36-0-55 0-23-0-45 0-23-0-55 0-36-0-55 0-23-0-55 I I 4 4 2 2 3 4 4 2 2 3 2 I I Total in sample Average diameter of eggs (mm.) Date January i6, 1931 St. No. 5g5 Depth N 100 B 170-0 m. 0-45-0-64 0-4S-0-55 0-45-0-68 0-36-0-68 0-4S-0-55 43 45 46 • 0-23 009-0-18 0-23-0-45 I I I Total in sample Average diameter of eggs (mm.) 3 0-24 1 Total in sample Average diameter of eggs (mm.) 36 0-38 Date January 17, 1931 St. No. 599 Depth N 100 B 142-0 m. Eggs recorded at St. WS 147 at (250-100) m. Date February 8, 1929 St. No. WS 376 Depth N 70 V 750-500 m. 49 • 032 0-18-0-45 0-23-0-45 1 I I 50 53 • 0-45 0-55 I I Total in sample Average diameter of eggs (mm.) 3 0-33 Total in sample Average diameter of eggs (mm.) 2 0-50 Eggs recorded at St. WS 377 at (100-50) m. Eggs recorded at St. WS 3S0 (250-100) m., WS 382 (400-250) m.. WS 385 (500-250) m., and WS 394 (200-100) m. Date Januar\- 19. 1931 St. No. 602 Depth N 100 B1 ,,, N70B ) '"°-°'"- Date February 8, 1930 St. No. 349 Depth N 100 B 60-0 m N 70 V 50-0 m. 42 43 44 45 46 47 48 50 52 53 54 55 57 58 0-09 0-09 0-23 (2) 0-23 (2) 0-23 (2) 0-23 0-23- 0-36 (2) 0-32-0-55 0-23-0-41 0-36 0-27-0-45 0-36-0-45 0-27-0-50 0-36-0-45 0-27 0-32 0-36 0-23-0-50 0-23-0-45 0 23-0-50 0-09-0-45 0-23-0-68 0-Z3-0-4S 2 I 2 5 I I I 4 2 4 2 I I 37 40 0-18 018 I I 39 Total in sample Average diameter of eggs (mm.) 2 018 Eggs recorded at St. WS 505 at (250-100) m. on Feb. 4. 1930. Date February 9, 1930 St. No. 351 Depth N 100 B 48-0 m. N 70 V 100-50 m. 42 45 48 49 55 0-41 0-45 0-41 0-4S (3) 0-45 (2) I I I 3 2 Total in sample Average diameter of eggs (mm.) 28 0-31 Total in sample Average diameter of eggs (mm.) 8 0-44 ■ Date February 9, 1930 St. No. 354 Depth N 100 B 96-0 m. Date January 20, 1931 St. No. 603 Depth N 100 B 140-0 m. 42 ti 53 54 59 0-32 (2) 0-4S 032 0-32 0-45 0-4S 0-50 0-50 2 I I 2 2 I 45 53 56 \ 0-18-0-32 0-18-0-50 0-18-0-64 • I I 1 Total in sample Average diameter of eggs (mm.) 3 0-33 Total in sample Average diameter of eggs (mm ) 9 0-44 THE DEVELOPMENT AND LIFE-HISTORY OF KRILL 175 Date February lo, 1930 St. No. 356 Depth N 70 V 250-100 m. Eggs recorded at St. 644 at (250-50) m. Metanauplius recorded at Sts. 647, 648 at (1000-500) m. and (250-0) m. Length in mm. Egg classes Total Date April 7, 1927 St. No. 207 A I 2 3 4 Depth N 70 H 0-5 m. 0-O5-O-I2 0-13-0-24 025-0-48 0-49-0-70 Length in mm. Egg classes Total 42 49 50 52 53 S5 56 57 58 59 60 6t 63 64 0-41 0-64 0-55 0-64 0-55 0-59 o-SS (2) 0-59 0-64 (3) o-SS 0-50 0-5S 0-64 0-55 0-59 0-63 0-64 0-S9 0-68 0-50 o-SS (3) 0-S9 (4) 0-68 0-55 (2) 0-59 (2) 0-68 0-64 I I I 2 2 6 I 3 6 3 10 4 1 I 2 3 4 0-4 5 O-0S-0-I2 o-l3-o*24 0-25-0-48 0-49-0-70 0-4 I 43 tl 47 48 0-4S 0-45 045 0-45 (2) 0-50 1 I I I 2 Total in sample Average diameter of eggs (nmi.) 6 0-46 Date April 7, 1927 St. No. 207 B Depth N 70 H o-s m. 35 11 40 o-os oos (2) 0-05 0-05 I 2 X I Total in sample Average diameter of eggs (mm.) 5 OOS Total in samp Average diamt 42 0-S9 ter of eggs (mm.) Eggs recorded at a depth of (250-100) m. Date February 8. 1931 St. No. 609 Depth N 100 B 128-0 m. Date April 7, 1927 St. No. 207 C Depth N 70 H o-s m. 49 51 55 0-27-0-45 0-18-0-45 0-23-0-45 I I I 40 4' 45 SO 0-05 (3) o-os 005 0-45 3 I 1 I 3 0-34 Total in sample Average diameter of eggs (mm.) Total in sample Average diameter of eggs (mm.) 6 0-I2 Eggs and metanauplius at St. 618 (Feb. 18-19, 193 1) at (750-100) m. Metanauphus at St. 620 (Feb. 20, 1931) at (1000-500) m. Date February 21-22, 193 1 St. No. 624 Depth N 100 B 137-0 m. Date April 7, 1927 St. No. 207 E Depth N 70 H 0-5 m. 48 0'l8-0-23 I Total in sample Average diameter of eggs (mm.) I 0-21 46 • 050 0-59 I I Total in sample Average diameter of eggs (mm.) 2 0-55 Date March 8, 1930 St. No. 368 Depth N 100 B 146-0 m. 47 48 0-05 0-07 o-os 0-07 o-os (3) I I 2 3 Date April 7, 1927 St. No. 207 F Depth N 70 H o-s m. Total in sample Average diameter of eggs (mm.) 7 0-C5 42 0-05 . 1 1 I Total in sample Average diameter of eggs (mm.) I o-os Eggs recorded at St. 637 (March 8, 1931) at (250-100) m. Metanauplii recorded at Sts. 636, 637, 638, 639 (March 8-9. 1931) at ( 1 000-0) m. Date March 10, 1931 St. No. 643 S;o°B''}''3-om. Date April 7, J 927 St. No. 207 H Depth N 70 H 0-5 m. 40 41 42 43 44 45 46 47 48 0-05 o-os o-os o-os 0-05 (3) 0-05 005 0-05 (2) 0-05-0-23 3 I 2 3 c 0-2 '■45 3-0-45 43 44 45 46 0-05 1 oos (3) 0-05 0-05 (4) I 3 I 4 Total in sample Average diameter of eggs (mm.) 9 o-os Total in sample Average diameter of eggs (irmi.) 14 0-05 Metanauplii recorded at WS 197 (April 17, 1928) at (1000-750) m. 176 DISCOVERY REPORTS Table 21. Totals of Males and Females taken in each month, including Eraser's Adolescents Aug. Sept. Oct. Nov. Dec. Jan. Feb. Mar. Apr. May June July Males Stage I 5 7 234 182 297 199 190 63 25 9 — — 2 37 17 35 40 82 41 316 87 "5 21 4 — 3 24 3 23 3 23 27 204 120 125 7 10 I 4 37 4 17 24 27 18 54 41 61 3 24 I 5 48 17 14 13 4 14 12 19 42 — I 2 6 12 39 80 5 7 3 21 14 II — — — 7 — 5 67 34 169 130 143 5 4 — — I Total 163 92 470 301 609 432 940 349 383 40 39 5 Stage A 18 7 238 199 252 208 60 58 30 4 — — B 33 18 31 24 137 42 478 81 135 28 4 — C 30 3 30 7 10 8 102 78 60 4 13 2 D 29 3 8 19 17 18 66 85 58 4 20 — E 43 6 12 8 5 12 64 26 39 — 2 2 F 10 49 49 7 14 14 41 7 5b — — — G — 6 102 37 174 130 129 14 5 — — I Total 163 92 470 301 609 432 94° 349 383 40 39 5 Females Stage I 14 18 254 213 338 240 230 275 188 22 4 — 2 66 17 15 18 8 5 304 no 71 34 28 3 3 no 4 54 8 I 2 124 30 50 6 61 8 4 31 33 175 15 12 3 — 18 20 — — — 5 5 I 25 78 114 41 10 I — — — — 6 — — — 3 52 92 45 2 6 — — — 7 — — — — 4 6 166 129 186 — — — Total 226 73 523 335 529 389 879 565 521 62 93 II Stage A 20 8 254 212 259 191 267 74 15 7 I I B 41 19 5 17 58 43 323 162 146 43 17 5 C 65 10 36 7 29 13 64 121 123 9 38 I D 71 10 35 10 4 3 II 58 28 3 29 4 E 29 26 187 56 SI 4 6 20 25 — 8 — F — — 4 13 37 27 17 3 I — — — G — — 2 20 91 108 191 127 183 — — — Total 226 73 523 335 529 389 879 565 521 62 93 II IDiscovery Report,. Vol. XXIII, pp. m-"^. ««'« '-^'^- 1"""y 'M'-l THE ANTARCTIC CONVERGENCE AND THE DISTRIBUTION OF SURFACE TEMPERATURES IN ANTARCTIC WATERS By N. A. MACKINTOSH, D.Sc. CONTENTS Part I. The Antarctic Convergence page 179 Introduction j-q Data j8o The mean position of the convergence 181 The change of temperature at the surface 186 Part II. The distribution of surface temperature in Antarctic waters . . 194 Introduction .... ,„, Treatment of the data j„. Method of drawing the isotherms iqg Notes on the distribution of temperature 201 References _„, Appendix. Table 9 2or Notes on the plates 2ii Plates I-XIV following 212 THE ANTARCTIC CONVERGENCE AND THE DISTRIBUTION OF SURFACE TEMPERATURES IN ANTARCTIC WATERS By N. A. Mackintosh, D.Sc. (Plates I-XIV; Text-figs, i-ii) PART I. THE ANTARCTIC CONVERGENCE INTRODUCTION r-p HE Antarctic convergence, which constitutes the northern hmit of the Antarctic surface water, i was first observed by Meinardus (1923), and it is a boundary of far-reaching importance in the Southern Ocean. It may be regarded as the fine at the surface along which the Antarctic surface water sinks below the less dense sub-Antarctic water, and it is distinguished by a more or less sharp change of temperature at the surface. The change of temperature is much less clearly defined in some longi- tudes than in others, but it is probably correct to say that the convergence is continuous around the Southern Ocean, even though there are regions where it can sometimes hardly be traced. It is essen- tially a feature of the surface, but where it is ill-defined, or when surface temperature records are insufficient, it can usually be assumed to lie in the latitude at which the coldest part of the Antarctic surface layer sinks below 200 m. For further particulars reference should be made to Deacon (1933, pp. 190-3, and 1937, pp. 20-4), and to Bohnecke (1938, p. 201, etc.). The importance of the convergence does not lie only in the fact that it is the boundary between the two principal water masses at the surface of the Southern Ocean. Its position is also related to the distribution and movements of the deeper water masses, and there are reasons for believing that it has a connexion with meteorological conditions in the Southern Ocean. Furthermore, it has a special biological significance, for the Antarctic and sub-Antarctic zones, between which it forms the boundary, have in certain respects, a distinct fauna and flora. That is to say, the convergence marks a limit (though not always a very rigid one) to the range of certain species in the plankton (see Hart, 1934, X937 1942; Mackintosh, 1934, i937; John, 1936), the fishes (Norman 1938) and the benthos (Hastings 1943) The distribution of diatom ooze and diatomaceous mud as shown on Admiralty charts, and by Neaverson (1934, plate xvii) indicates a relation also between the convergence and the bottom deposits. , r 1 • *• *• ^ A number of oceanic species seem to be unaffected by the convergence, but further investigations will probably increase the list of those whose distribution is influenced by it. It is probable that it separates certain species which characterize different water masses or different systems of circulation, but the simple effect of temperature also needs consideration, for the distribution of some species whose upper or lower limit of tolerance approximates to temperatures at the convergence, may be sharply bounded by the abrupt change of temperature. • j u .u w. ^. Deacon (1937 p. 23) concludes that the latitude of the convergence is determmed by the latitude reached by the Antarctic bottom water. He drew the line of the convergence on a circumpolar chart fiQ.7 fi/4) and this was based on a number of positions fixed in the Falkland sector, and rather more scattered positions in other sectors. Since then a very considerable number of additional positions have been fixed. These show that Deacon's original line must lie very near the actual mean position, i8o DISCOVERY REPORTS but that the position in any given longitude is subject to rather more variation than the earlier data suggested. Before the mean monthly surface isotherms in Antarctic waters can be drawn it is necessary to locate the mean position of the convergence as correctly as the available data permit. Furthermore, records of both the mean position and the actual position at different times and places, and the change of temperature at the surface, are likely to be required for various purposes in the future, and it therefore seems worth while to publish particulars of every occasion on which the convergence has been crossed by the Discovery Committee's ships, and to indicate the degree of accuracy with which it could on each occasion be located, and the extent of the change of temperature. DATA The 'Discovery', 'Discovery II', and 'William Scoresby' have crossed or located the convergence on 139 occasions. In nearly every instance at least some indication of its position was obtained, but the degree of accuracy with which it was located varied considerably according to the observations taken, to the extent to which the convergence was itself recognizable, and to the angle at which it was crossed. The best indication of a crossing of the convergence is a sharp change of temperature shown on the thermograph (see p. 195) an instrument which was in operation at all crossings by the ' Discovery ' and by the ' Discovery II ', except for a period in the fourth commission (1935-7) of the latter ship. The 'William Scoresby' did not carry a thermograph except on her last commission (1937-8). The other important method of locating the convergence is by vertical stations taken on each side of it, whereby the level of the minimum temperature of the Antarctic surface stratum can be determined. This is a reliable method but is naturally of little value if the stations are very far apart. Sometimes it is evident that the convergence lies between two stations at which only surface temperatures are read, but there must be a clear difference, and of the correct range, if this method is to be trusted. Normally the sea temperature is recorded every four hours as a routine by the ship's personnel. These records are not altogether reliable as exact readings of the temperature, but in the absence of a thermograph they may be of considerable value, especially when they show a sharp change of temperature at about the expected position of the convergence, or when they show where the convergence lies when it is known to have been crossed between two stations some distance apart. The best determinations of the position of the convergence are obtained when it is crossed approxi- mately at right angles, and the thermograph shows an abrupt change of temperature between two vertical stations of which the more southerly shows a minimum temperature above 200 m. and the more northerly a minimum below 200 m. Such complete indications are not ver)' frequent, though they have commonly been obtained in the Scotia Sea. In many parts of the Southern Ocean it may often be found that there is no clearly defined change of surface temperature, or the change may be obscured by an oblique crossing of the convergence. Sometimes two successive stations may show a minimum temperature at about 200 m., and it is then difficult to know where the true convergence lies. A frequent source of difiiculty is the winding course of the convergence. At the surface the junction of Antarctic and sub-Antarctic water seldom lies in a straight line, probably because it is an unstable boundary. It forms twists and loops that may extend as much as 100 miles north or south, and it possibly even forms isolated rings. The line is perhaps comparable to the edge of the pack-ice, and may be even more tortuous. The latitude in which it is found may thus depend partly on whether a ship happens to cross it where it bends to the north or to the south. Another consequence is that a ship steaming in a straight line obliquely across the convergence may pass from Antarctic to sub- Antarctic water and back again several times. It may also happen that before or after a crossing of THE ANTARCTIC CONVERGENCE i8i the convergence, the thermograph indicates that the ship's track passed near to a loop of the convergence without actually cutting it. , i • i It will now be seen that at many crossings of the convergence its position can only be approximately determined Indeed the exact position in which it is assumed to be is sometimes little more than a matter of opinion. However, even approximate positions are helpful in mapping out the mean position. The latter must always be liable to modification (though in a diminishing degree) in the light of new '^Table 9 (p 205) is a list in chronological order of all crossings of the convergence by the Discovery Committee's ships. The position of the convergence is given as nearly as it can be estimated. This may be taken as the middle of a sharp or comparatively gentle temperature gradient at the surface, as hal way between two stations at which vertical observations were made, etc. Since some positions could be determined with a good deal more accuracy than others, an indication of the extent to which they can be relied upon is given in the column headed ' Degree of accuracy'. The quality of the evidence on which they are based varies considerably, and it is difficult to assign a very exact meaning to the terms entered in this column. As a rule 'V. good' implies that there was a sharp rise or fall in the thermograph record, clearly identified as the convergence. The possible error in the estimated position is not likely to be more than about 5 miles either way. ' Good' may indicate a thermograph gradient which was well defined, but which extended over some hours of steaming ; it may mean that the crossing was unmistakably indicated by the ship's 4-hourly routine temperature readings; or ^^ -ay be used when there was no definite change at the surface, but when two vertical ^^at-ns, not too far apart showed that the convergence lay between them. The maximum error here might be ^b-^ 2° -1 s either way. 'Approx.' generally indicates that the convergence may have been crossed anywhere between two pofnts abo.' 50-150 miles apart; or it may be used, for example, where the only evidence was a vertical station with a minimum temperature at or about 200 m., for this does not necessarily mean ^h the station lay exactly on the convergence. 'Probable' generally implies that there was rv^ltlhaAhe converU hid been crossed at a position which could be d-mmed^^^^^^ precision but that it was not completely certain that it was in fact the convergence The erm is of en ariicabk when there are several sharp changes in temperature which may indicate oops in th clergenci 'Uncertain' is applied to any of the less satisfactory records, and may indicate doubt as to the identification of the convergence or as to the accuracy of the estimated position. In normal times this work might have been undertaken by Dr G E. R. Deacon^J-J^-S- D-^^^^ the war he has been in service with the Admiralty, but we have together reviewed the data on which each position L Table 9 is based, and I am indebted to him for this assistance and for reading through the manuscript of the paper. THE MEAN POSITION OF THE CONVERGENCE Many of the positions in Table 9 are from isolated "bse-ations but sometinaes Ae convergence^^^^ erossed more than once at positions not far apart and w.thm a few days or weeks The convergence does no. appear to change its position very quickly, and such points can often be jomed up so as ,0 :°::a:Lr:.onofitsLrse^ta.ive„..™^^^^^^^^ Llralso F r, P .0.). Nos. 88 and 90 were a few weeks later and are jomed by a separate hne. The tetf the convergence on these occasions may be compared wrth the mean hne ,n Plate I. l82 DISCOVERY REPORTS Figs. 2-4 include the positions listed in Table 9 except No. 20 which was vague and unreliable. Different positions have been joined up, wherever it seemed reasonable to do so. This seems to provide the best basis for estimating the mean position of the convergence, and it gives an impression of the extent to which its position is liable to vary. The majority of positions are in the Atlantic sector. Data are still rather scarce elsewhere, and the points plotted in Fig. 4 are joined by lines drawn simply in what appear to be the most probable positions. In Plate I the convergence is drawn to represent as nearly as possible the mean of the lines in Figs. 2-4. Wherever possible the average latitude has been calculated within narrow limits of longitude, and those records which are marked ' Good ' or ' V. good ' in Table 9, were counted twice in the averages so that some bias should be given to the more reliable data. Fig. I. The Antarctic convergence in the Falkland Islands Sector, September to December 1934. Numbers refer to the serial numbers in Table 9 (p. 205). Black dots denote well-defined positions, and rings approximate or uncertain positions. Pecked lines show the ascertained position of the convergence (apart from unknown minor irregularities), and dotted lines show its probable course where it is not checked. Continuous lines show the tracks of the 'Discovery 11' (see also Fig. o, p. 201). The mean line of the convergence, shown in Plate I, like most of the results set forth in this paper, is an estimate which is open to adjustment in the light of any additional data which may be obtained in the future. It is obvious from Figs. 2-4 that it is based on far better material between 80° W and 30° E than elsewhere. In the Falkland region the estimated position must be very near the true mean position, but in the Pacific sector, between no and 160? W it may be far from accurate. Some comments are needed on the extent of variation from the mean. I have made a rough measure- ment of the distance of each plotted point from the estimated mean position, and the resulting figures given in Table i are perhaps worth noting. This table includes all the positions in Table 9 for which a measurement could be made, and the possible extent of the displacement may not be so great as it seems to suggest. Most of the larger deviations were at places where either the actual record of the convergence was not very certain, or where the mean position and course of the convergence are based on inadequate data. Furthermore, it is known that sub-Antarctic surface water is sometimes thrust a considerable distance south of the normal position of the convergence, and the extensive loops in the convergence (seen in Figs. 2, 3) show that there is an element of chance in the latitude at which it is found at an isolated crossing. This last factor may indeed account for many instances of apparent displacement from the mean position. Probably a displacement of 50 miles or so either way is not uncommon, but it may be that the extreme displacement does not exceed about 100 miles. THE ANTARCTIC CONVERGENCE Table i. Deviation from the mean position of the convergence 183 Miles north or south of mean position Well-defined positions All measurable positions Number 0/ /o Number % 0- 24 25- 49 50- 74 75- 99 ioa-124 125-149 150-174 31 15 7 I I I 0 55 27 12 2 2 2 0 65 38 14 8 4 I I 49 29 II 6 3 I I Total 56 100 131 100 The latitude of the convergence is not much affected by the time of year. Table 2 shows the average deviation for each month. The figures for June, July and August are based on so few records that they should probably be disregarded, and in the other months the average deviation is less than 20 miles except in February. There is a suggestion of a small northerly displacement in the cold months (September-November) and a southerly displacement in the warm months (January-March), but if the latitude of the convergence really varied with the temperature we should not expect the average for April to be so far north. The average monthly deviations are small compared with the actual range of deviation, and the convergence may be found in very different positions even in the same month. For example, Nos. 60 and 109 in Table 9 were both in November in different years and in about the same longitude, but the former crossing was over 50 miles north, and the latter about the same distance south of the mean position. Although there may be a minor seasonal oscillation, it seems probable that variations in the latitude of the convergence depend mainly on local and temporary factors, such as shifting loops and irregularities in its course, and perhaps to variations in winds and currents. Table 2. Average monthly deviation from the mean position of the convergence Month Miles north or south of the mean position Number of instances Month Miles north or south of the mean position Number of instances September October November December January February 13 N 10 N 16 N 3N 19 S 28 S 9 II 16 17 9 17 March April May June July August 11 S 15 N 12 S (34 N) (15 S) (10 N) 16 16 9 4 2 5 Total 131 As already noted, the mean position of the convergence, drawn in Plate I, does not differ much from that arrived at by Deacon (1937, fig. 4). It is farthest south about 80° W and 180° W, and farthest north between 35° W and 70° E. It is possible that some of the minor bends in its course will be smoothed out in the light of additional data. There is little doubt that the S-shaped turn between the Falkland Islands and South Georgia is a normal feature, but it does not always exist, for there has been at least one occasion when the 'Discovery 11' has crossed this region on a meridional course and found no trace of the southern limb of the loop. There is good evidence for the sharp southward turn in 60° W. i84 DISCOVERY REPORTS New data may suggest a modification of the loop in 30° E, but some distortion of the convergence and isotherms here is probably normal. Deacon (1937, pp. 34 and 92, and plate xliv) refers to a steep ridge on the sea floor at this point, which may affect conditions at the surface. The slight northward bends shown about 140, 97, 75, 67 and 35° W and 110° E need confirmation, and the whole of the con- vergence between 1 10 and 160° W, and between 50 and 90° E is based on very little material. Fig. 2. Positions of the convergence between 10° W and 80° W. Where two or more positions were ascertained within a short period they are joined by a continuous line if the convergence was well defined, and a pecked line if the positions were approxi- mate or uncertain. Single positions are indicated by a short line, continuous or pecked, drawn approximately parallel to the mean line of the convergence. Fig. 3. Positions of the convergence between 10° W arid 40° E. See legend to Fig. 2. It is noteworthy that the convergence divides the Southern Ocean into two almost equal zones; in fact it lies very nearly half way between the Antarctic coastline and the extremities of each of the southern continents. From Cape Agulhas (South Africa) the distance to the convergence is rather less than half the distance to the nearest part of the Antarctic continent, but from Western Australia, Tasmania, New Zealand and Cape Horn the convergence is very close to the half-way point. It is difficult to say whether this is of any special significance. The fact that the convergence lies in about THE ANTARCTIC CONVERGENCE i8s Fig. 4. Positions of the convergence other than those between 80° W and 40° E. i86 DISCOVERY REPORTS the same relative position at each of these points rather suggests some balance of pressure between water masses drifting through those parts of the Southern Ocean which are hmited to the north as well as to the south by land masses. Sverdrup, Johnson and Fleming (1942, p. 607) say: 'The south- ward displacement of the Antarctic Convergence to the south of Australia and New Zealand can probably be ascribed to the relative narrowness of the passage between these regions and Antarctica, and the similar displacement off South America can be attributed to the southerly location of Drake's Passage separating South America from Graham Land.' It is also worth while to compare the position of the convergence with the mean northern limit of the pack-ice. The pack extends farthest north about September and October, and the line in Plate I is a reproduction of the September-October line shown by Mackintosh and Herdman (1940, plate Ixix). It is evident that the position of this line bears some relation to the position of the convergence. The extreme limit of the ice will of course lie nearer to the convergence, but it is doubtful whether the pack ever reaches the convergence itself. The distance between the convergence and the mean northern limit of the ice shown in Plate I varies from 120 miles in the Scotia Sea to about 550 miles in the Indian Ocean sector, with an average of 370 miles. This may be taken as a belt which is practically free of pack-ice throughout the year. There is some reason to suppose that it has a characteristic plankton fauna and flora. On the basis of the distribution of macroplankton in the Falkland sector I distinguished a ' northern zone ' imme- diately south of the convergence (1934, p. 150), and Hart (1942, p. 280 and Fig. 2) found that the phytoplankton could be suitably divided into certain biogeographical areas, one of which he took as the 'northern region' lying between the convergence and a line 330 miles to the south of it excepting certain special areas. This region was in fact distinguished largely on the grounds that it is normally free of pack-ice at all times. This is a matter to which I hope to return in a future paper. THE CHANGE OF TEMPERATURE AT THE SURFACE The convergence is generally distinguished by a sharp change of temperature at the surface, which usually appears as a steep gradient on the thermograph, and it is important to consider what variations occur in the position of the gradient on the temperature scale, and the range of the gradient. This is best done by examining first the variations in the middle temperature of the gradient, and then the variations in the range. The aim of the present section is to estimate (as nearly as the data permit) what change of temperature we may expect to find in any longitude at any time of year, and hence how to place the isotherms at the convergence when drawing charts of surface temperatures. Deacon (1933, fig. 10, p. 190) gives an example of the gradient as shown on the thermograph. This was No. 59 in Table 9, and it is an exceptionally well-defined crossing. Some further typical examples are shown here in Fig. 5 A-E. The middle temperature is taken as the mean of the temperatures at the beginning and end of the gradient representing the convergence, or, if the gradient is not very clearly defined, whatever seems best to represent the central temperature of the convergence. It is easy enough to decide the middle temperature and range in such examples as Nos. 33 and 106 (Fig. 5). In such as No. 88 it is difficult, but these small and ill-defined gradients cannot be excluded from calculations of the average. In this case the middle point was taken as 3-0° and the range as o-8°, on the assumption that the rise of temperature immediately to the right of St. 1476 represented the convergence. Table 3 is a list of all the middle temperatures and ranges which I have been able to measure. It includes some rather doubtful measurements, such as No. 88, but should serve for provisional estimates of the changes of temperature to be found in different months and positions. Usually both the middle HON I DAY a .1 * TUESsOAY THE ANTARCTIC CONVERGENCE * ■» 8 i "> o '• ' _" ~ . iq H---dor-t--f t^THi-/ ; r''333Eirn m IrZS'fr ::50 ::,1S 10 • 451 1 . 5"^ 0 -■'■■ v. ''■ -T^^.—n 452: \ \ A, : „S- i. i:;;,.,-0.., :^A... :l_^..i.,i.I.i,.^i.a...l...x...,t...x„i 187 SATU I RSAV n ? 4 4 8 [, 1 1 i [:::|i#H4^ * ; ♦■■-?■■■ iElESE.^ I i ; £ ^-■•tESEs =p$:U ' ; ' r '"^^T^t^-I—^ 320 is I A.C.4:Z 4 8^4 T7-rTTTT7 i. t i-t 35 'r , ' ' .UHlcm ..ON.I.OAY . THUKlSSAr ^ «; 4 " 30 25 20 « a 5^ ■< 8 ' ■» 35 - , ' V ' ' /---r-T";-; 30 25 20 15 ' r I I 7 AC^yV AC 106 ^?l:, \ 1357 f\ 185& \ \ 4M36 i-5 ^ ^ .7^ 00 r \ \ 5X1%, V V \ ^ \ J Fig. 5 A-E. E..a.p,.s .r .he™o,„ph records cro.!^, the A„.a.>^^^^^^^^^^^ tu? J ^iTry-t So" 33 Kbr.'C.f.n'i^Ir^f sr„Sler ,he periods dur.ng which ,he sh.p w.s .topped or towing nets at slow speed. 2-2 i88 DISCOVERY REPORTS temperature and the range are given, but sometimes the former could be distinguished while the beginning and end of the gradient were too ill defined for measurement of the range. In some cases where the change of temperature was small, and the convergence and its gradient could not be dis- tinguished, the middle temperature is omitted, but, for calculating the average, a small figure is entered as an estimate for the range. In one or two other instances an approximate figure could be given for the range though the middle temperature was very obscure (e.g. No. 49). Table 3. The change of temperature markmg the Afitarctic convergence at the surface. Middle temperature and range of the gradient u Middle temperature u n 0 Pi u .2 XI Middle temperature Pi t-t :2^ Middle temperature 1) a 0 Pi Middle temperature CQ 0 Pi 1 Serial number Middle temperature (3° I 5-25 2-1 43 2-00 3-2 67 4-00 1-8 91 4-40 — 121 4-00 i-o 3 2-20 — 44 3-80 2-5 68 2-55 1-3 92 4-00 2-6 122 4-50 2-8 4 3-80 0-4 45 4-30 17 69 3 -20 3-3 94 4-50 2-0 123 4-20 1-4 6 4-50 — 47 4-10 2-2 70 4-50 07 95 3-40 o-S 124 4-15 2-5 7 4-90 2-0 48 5-25 2-5 71 4-00 i-i 96 3-40 II 125 4-40 1-5 12 2-00 — 49 — 2-0 72 3-50 1-5 99 5-50 — 126 3-50 2-2 13 3-20 — 50 3-10 1-0 73 4-10 3-4 100 5-80 — 127 4-05 1-3 15 5-89 — 51 4-45 1-9 74 4-45 2-3 102 5-3° — 128 3-4° 1-4 17 3-8o — 52 3-60 2-2 75 4-50 2-0 103 2-8o — 129 2-85 17 22 2-6i — 53 3-45 2-9 76 2-8o 2-0 104 2-95 — 130 2-50 1-0 24 — 0-3 54 4-00 4-0 77 2-50 1-0 105 2-00 1-2 131 170 1-0 27 4-08 — 55 2-00 1-4 78 3-20 2-4 1 06 3-15 2-1 132 2-40 1-0 29 3-40 2-8 56 375 3-5 79 — 2-4 107 3-50 3-0 133 2-85 0-9 30 3-50 3-0 57 4-25* 1-5 80 1-50 1-5 109 2-15 3-2 134 3-50 i-o 32 5-85 27 58 2-00 2-0 81 270 2-2 III 4-20 2-0 135 4-50 1-0 33 4-00 I-O 59 1-35 3-5 82 — 1-9 1 12 5-50 2-0 136 — 0-5 35 — 0-8 60 3-10 3-8 83 1-65 I "3 113 5-50 1-0 137 4-90 0-5 36 — 07 61 3-63 i-o 84 175 II 115 4-50 0-6 138 5-30 0-3 37 475 2-5 62 3-85 17 85 2-00 2-0 116 3-25 2-3 139 5-15 3-3 38 4-40 1-2 63 3-30 2-4 86 1-30 2-0 117 4-10 1-8 40 4-25 2-5 64 4-00 1-8 87 2-15 2-1 118 3-50 1-0 41 2-00 2-0 65 4-50 10 88 3-00 0-8 119 4-00 — 42 2-50 2-0 66 4-90 I-O 90 3-10 37 120 4-50 1-5 * Doubtful. Analysis of this table indicates that the middle temperature varies mainly according to the time of year, and to some extent according to the latitude of the convergence, while the range varies in different longitudes and appears to depend on other factors. Taking the middle temperature first, Table 4 gives the average and maximum and minimum values for each month in arbitrary ranges of latitude. Divided in this way the quantity of data is rather small, and it is evident that the middle temperature can vary considerably in any one month and latitude, but it is obvious from the average figures that it generally falls to a minimum about September and October, and rises to a maximum about February. This is naturally to be expected since the sea tem- perature varies with the time of year while the latitude of the convergence is little affected. The figures also leave little doubt that where the convergence is in a high latitude the middle temperature tends to be lower than where it lies in a low latitude, though the difference is not very great. This implies that surface isotherms near the convergence do not always run quite parallel to it. THE ANTARCTIC CONVERGENCE 189 Table 4. The observed middle temperatures of the co?wergence gradient for month and latitude North of 50° S 5o°-54° S S4°-S8° S South of 58° S i i — Mean No. 2 Max. Min. Mean No. Max. Min. Mean No. Max. Min. Mean No. Max. Min. September 2-25 2-50 2-00 — — — i-6s 3-15 2-82 3 375 2-00 175 1-32 2-38 2 2 2-00 i'15 1-50 1-30 October November 2-63 3-63 3 5 4-00 4-10 1-70 2-8s 2-21 3-43 3 3-25 3-63 2-59 4 3-10 2-15 3 3-00 2-00 December 4-04 4 4-.;o 3-5° 3-57 6 4-20 3-20 2-7b 3 3-20 2-00 2-55 January February S-iS 4 .;-30 4-90 4-20 4-37 2 3 4-40 4-50 4-00 4-10 4-50 5-17 I 3 5-51 4-50 4'5° 4-47 4 5-89 4-00 March April 4-64 4-6i 4 4 5-15 5-30 4-00 4-25 5-23 4-25 3 3 5-80 4-50 4-40 3-8o 4-30 379 3-60 372 4 2 475 4-08 3-5° 3'50 4-15 2 4-90 3-40 May June 4-33 2-95 2 I 5-85 2-8i 3-45 3 4-45 2-8o 2 4-00 3-45 — — — — July 4-05 I — — 3-50 I ~~ _ — August 2-99 4 3-40 2-5° 2-00 I 1 It is evident' that the average figures can be improved by some method of smoothmg, even though the data are insufficient for an accurate estimate of the true averages for each month and latitude. I have therefore adopted the rough and ready method of plotting the averages and drawmg curves by eye to make as good a fit as possible. Figures showing the average for all latitudes in different months, and for all months in different latitudes might be misleading, for there are insufficient winter observations in the higher latitudes. In Fig. 6 A-D therefore the monthly averages are plotted separately for each range of latitude and the curves are drawn both to represent the monthly trend of the points, and to resemble each other, for it is assumed that the rate of change does not differ very much in different latitudes. In Fig. 6 E the four curves are shown on the same scale and are marked with the middle latitude of each range. Thus the second curve applies to all records between 50 and 54 S and can be taken to represent 52° S. It will not be far wrong to say that those north of 50 represent 48 b and those south of 58° represent 60° S. Table 5. Provisional estimate of the average middle temperature of the convergence gradient in different months and latitudes {figures smoothed) 46° S 48° S 50° S 52° S 54° S 56° S 58° S 60° S 62° S Average September October November December January February March April May June July August 2-6 2-6 3-9 4-3 5-0 5-4 5-1 4-6 4-2 3-9 3-6 3-2 2-5 2-5 3-6 4-1 4-9 5-2 5-0 4-5 4-1 3-8 3-5 3-1 2-4 2-3 3-4 3-9 47 5-0 4-8 4-4 4-0 37 3-4 3-0 2-3 2-2 3-2 37 4-6 4-9 47 4-3 3-9 3-6 3-3 2-9 2-1 2-1 3-0 3-4 4-5 4-8 4-6 4-2 37 3-5 3-2 2-8 2-0 2-0 2-8 3-2 4-4 47 4-5 41 3-6 3-4 3-1 27 1-9 1-8 2-5 3-0 4-3 4-5 4-4 4-0 3-5 3-3 3-0 2-6 1-8 17 2-3 2-8 4-2 4-4 4-3 3-9 3-4 3-2 2-9 2-5 17 1-6 2-1 2-5 4-1 4-3 4-2 3-8 3-3 3-1 2-8 2-4 2-1 2-1 3-0 3-4 4-5 4-8 4-6 4-2 37 3-5 3-2 2-8 Average 4-0 3-9 3-8 37 3-5 3-4 3-2 3-1 3-0 3-5 Table r is derived from Fig. 6, and it includes figures for interpolated latitudes and a slight extra- poltion to 46 Id 6." S. The'steps leading up to this table have been explained ■" i> bi) 0 rt ^ 1m U > rt D 0 ^ ■ nj ■ 1- 0 < ■S . — uJ a w-a CO 0 .5-3 • D h 2'3 0 "J o-c ■ ^ "ttj u D _ D 0 0 0 ca f- u •5' J3 tn 0 i3 4-* -oz 0 u 60 0 C -ru rt Lri _ (U 60 0 "5 S u > rt u 0 j= -tD U.4 0 CO 0 .0 .2 *n c^ 0 > r^ bib •••4 fe •3 533^930 Nl 39NVy "94 DISCOVERY REPORTS PART II. THE DISTRIBUTION OF SURFACE TEMPERATURE IN ANTARCTIC WATERS INTRODUCTION In 1940 I published, in collaboration with Mr H. F. P. Herdman, a paper on the distribution of the pack-ice, which consisted essentially of a series of charts showing the observed or estimated latitude of the ice-edge in each month in all parts of the Southern Ocean. The present paper is the next step in mapping the surface conditions in Antarctic waters, and its object is to estimate, as nearly as the data allow, the mean positions in each month of the surface isotherms between the ice-edge and the Antarctic convergence. During the cruises of the ' Discovery ', ' Discovery II ' and ' William Scoresby ' between 1926 and 1939 a very large quantity of data on surface temperatures has been collected. These data of course include far more than the surface temperatures published in the Station Lists {Discovery Reports vols. I, III, IV, XXI, XXII, XXIV). It is the thermograph carried in these ships which furnishes the mosi important body of data, for this instrument was in continuous use (except for one short period) throughout the voyages of the 'Discovery' and ' Discovery II', and in the last commission (1937-8) of the William Scoresby '. v vj / y The data are unfortunately distributed very unevenly in time and space, for the movements of the ships were generally dictated by other considerations than covering the surface as equally as possible. u ', r ^"' ^'"""^^ ^^^'^ observations were obtained in each month, and it will be seen that although the Atlantic sector is well covered at nearly all times of year there are large gaps in other sectors, especially in the winter months. Some data on surface temperatures south of the Antarctic convergence are of course available from other sources. These are the observations made by other Antarctic expeditions and some published observations made by whaling factories. I have not, however, obtained very much assistance from this material. A large number of the observations were made within the pack-ice or near its fringe and It IS not certain whether some of the others are sufficiently accurate (see pp. 197-8). The number of reliable records from all other sources of temperature in the open regions of the Antarctic surface water are very small compared with the Discovery Committee's material, and they cover few of the gaps mentioned above. ^ ^ Since the surface temperature in any given locality depends primarily on the time of year, I have set out to plot all appropriate records on separate charts for each month, irrespective of the year in which they were made. In some years of course the temperature at a given place in any one month is above or below the average, but these irregular annual variations seem to have only a local and tem- porary significance, and their range is much less than that of the regular seasonal rise and fall of temperature. The charts used were semicircular overlapping charts identical with those used for plotting the ice-edge (Mackintosh and Herdman, 1940). The isotherms were then drawn to fit the actual observations as nearly as possible. Some adjustments were made after a comparison of the monthly position of each isotherm, and in areas where no observations were available in certain months entative isotherms were filled in by interpolation as described below. Finally, the isotherms were ransferred to the circumpolar charts reproduced here in Plates II-XII. The various steps in the treatment of the data are described below in some detail because they raise certain points in technique which seem worth recording. ^ DISTRIBUTION OF SURFACE TEMPERATURE iQS TREATMENT OF THE DATA As mentioned above, the continuous thermograph provides the most important body of data. In this instrument the thermometer bulb Hes in a pocket in the ship's hull about 14 ft. below the surface, and the temperature at that depth is recorded on a chart (see Fig. 5, p. 187) attached to a clockwork drum in the laboratory on deck. Each chart covers a week. The instrument has given very satisfactory results and, provided certain checks are made, the temperature can be read correctly to within less than 0-2" C, and the time to within less than half an hour. Certain possible sources of error have to be considered. The clock has generally been found to keep excellent time, but allowance had occasionally to be made if the drum had not been accurately set when the new chart was attached. The timing could generally be checked at points where the ship was stopped on station, for here the temperature is often traced as a steady horizontal line which is usually distinguishable from the numerous small fluctua- tions of temperature which appear while the ship is moving (for example, see Fig. 5 A, p. 187, St. 451). The time when the ship is stopped can of course be checked from the logs. Correction of the tempera- ture is more important as there is frequently a slight error which is possibly due to a small amount ot play on the arm of the pen. This error seldom exceeds 0-5^ C and is usually constant until the chart is replaced unless substantial changes of temperature occur. The thermograph record should, however, always be checked against direct thermometer readings, preferably with the Nansen-Pettersson water- bottle Since the ' Discovery ' and ' Discovery II ' normally worked full stations at frequent intervals, there were few thermograph charts which could not be checked in this way at several points, but it is verv evident that the method of reading the temperature in a bucketful of water drawn over the ship s side is wholly unreliable as a check, however carefully the thermometer is read. Such readings seem nearly always to be too high and the error may amount to i-o° C or more. This method has been criticized elsewhere (see Lumby, 1928). , , r r. 1 A further question which arises is whether the thermograph, recordmg at a depth of 14 ft-, Proper y shows the 'surface' temperature, and whether thermometer readings at o m. constitute a valid check. In other words, what diflFerence may there be in Antarctic waters between the temperature at o m and about s m.? At full stations the vertical series of temperatures, etc., included readings at o and 10 m and I have worked through the station lists and assembled the data on this point m Table 8. This includes 987 vertical stations at which the temperature at o m. did not exceed 5-0 C (i.e. the vast majority of stations south of the Antarctic convergence) and it will be seen that, except in the summer months (December-February) about 90% showed less than o-z^ C difference between o and 10 m Even in the summer months when occasional superficial sun-warmed patches may be expected, any significant difference is rare, and the difference between o and 5 m. must be still less. There have beL one or two occasions when a vertical station has been worked in the vicinity of an iceberg, and where a cold, less saline, surface stratum has depressed the temperature at o m. below that at 10 m In any case where there is a significant diflference between o and 10 m. it is preferable to check the thermograph chart from other stations. , , 1 . f*^„ Since ship's time changes with changing longitude the clocks are frequently changed at sea, often several times in a week. The thermograph cannot conveniently be adjusted to ship s time and is there- fore set at G.M.T. But the temperature readings must be linked with the ship s position the stations, nd her observations, all of which are recorded by ship's time. It is therefore desirable to mark a "ak of ship's time on the thermograph charts. I found it best to trace the corrected temperature on to new charts, marked with the ship's time scale, periods on station, and other -notations. The next step was to draw track charts, marked with ship's time, for all voyages south of 45 S except such complex local movements as were undertaken in the oceanographic surveys around South 196 DISCOVERY REPORTS Georgia and in the Bransfield Strait. The tracks were plotted on the usual semicircular charts. The scale was 6 mm. to a degree of latitude, and it was thus possible to mark the ship's hourly positions by dots at intervals of a millimetre or less (i mm. = 10 miles). The ship's positions recorded in the logs, usually at 4-8 hr. intervals, were plotted first and the hourly positions interpolated. Such plotting of the ships' tracks over many thousands of miles was naturally very laborious, but the resuhing charts should in due course facilitate the plotting of various observations in addition to surface temperatures. A small part of one of these charts is reproduced on the original scale in Fig. 8 A. These track charts were then compared with the thermograph charts and the ship's track was marked in pencil at each point at which the temperature line crossed the line of a whole or half degree C (-0-5, o, 0-5, i-o°, etc.). The hues of the ships' tracks, together with the temperatures were then traced on to monthly tem- Table 8. Frequency of differences between the temperature at o m. and at 10 m. {including all records where the temperature at o m. does not exceed 5° C) Range of difference rc) September, October, November December, January, February March, April, May June, July, August No. 0/ /o No. /o No. /o No. % 0-00-0-09 0-10-0-19 0-20-0-29 0-30-0-39 0-40-0-49 0-50-0-59 0-60-0-69 0-70-0-79 0-80-0-89 0-90-0-99 I -00-1-09 1-10-1-19 1-20-1-29 1-30-1-39 I -40- 1 -49 239 22 3 I I 89-8 8-3 i-i 0-4 0-4 368 46 26 12 6 8 3 I 2 I I 2 77-3 9-7 5-5 2-6 1-2 1-7 0-6 0-2 0-4 0-2 0-2 0-4 196 II 5 2 I 91-2 5-1 2-3 0-9 0-5 27 2 I 90-0 6-7 3-3 266 100 476 100 215 100 30 100 perature charts, all data for any one month being included on one chart or set of overlapping charts, irrespective of the year. The year was, however, marked against each track. A comparison of Figs. 5 and 8 will^show the method of plotting. For example, in Fig. 5 A (p. 187) the thermograph record crosses the 3° level at noon on 14 October 1930. The position of this point is found in Fig. 8 A just north of 50° S, and hence 3° is marked in the corresponding position on the ship's track in Fig. 8 B. A large number of temperature records at single positions were then added to the temperature charts. These included the surface temperatures read at stations made by the Discovery Commhtee's ships when no thermograph chart was available (mainly derived from the 'William Scoresby'), and published records from other sources. Part of a chart of October temperatures is reproduced as an example in Fig. 8 B. The mean isotherms indicated here are in their final position, as shown in Plate II, after comparison with the corresponding isotherms of other months. They represent the temperatures for both September and October, which are very similar; but there is not room to show the observed temperatures for September on the same chart, for in places the ship's tracks for the two months almost coincide. It should be mentioned that DISTRIBUTION OF SURFACE TEMPERATURE i97 the mean isotherms cannot always be made to correspond with the obser^'ed temperatures so well as they do in this chart. • . tv/t * Of the published records from other sources, perhaps the most important are those m the Meteor Reports (Bohnecke, 1936). Here a very large number of surface temperature readings are tabulated, but they are from all parts of the Atlantic Ocean and only a small percentage he south of the Antarctic convergence Bohnecke gives two principal tables. One shows the temperatures (also S°/oo, etc.) recorded by the 'Meteor', with date and position (pp. 41-64). Only a small number of these are m Fig 8 A. Part of a track chart for October 1930. showing hourly positions of the ^Discovery II ' during a voyage from Cape Town to South Georgia. See Fig. 5 A for the thermograph record of part of this track. 12 n X- Ship's position at 1200 hours on 13th October. 20^00 (451): Stopped on Station 451 from c. 2000 to 0000 hours. B. P^Jof^rCperature chart for October, including ^Je t.J in A w^^^^^^^^ rrtte-^s^trorie^^^^^^^^^ -P-- -' -'«- --''-' from Plate^I, and the 2° isotherm lies on the mean pos.t.on of the convergence^^^ A.C. (No. 33): Observed position of the Antarctic convergence (No. 33 m lable 9, p. 20OJ. ('25): Meteor observations, 1925. ('26): 'Wi.lliam Scoresby' station, 1926. Antarctic water, but they arc of some assistarrce although they fill no important gaps in *e Discovery material For tkc other tabic (pp. 65-186) the Atlantic Ocean ts div.dcd into areas enclosed by ro Tnati udc and longitude, and these are subdivided into 1° areas. Data from many sources are mcluded aid o far as the data permit, the average temperature is given for each f area for each momh. Some of A e a e n Antarctk waters, bu, it seems that such averages, compiled from m.seellancous sources, 1 u d be t Led with caution. B6hnecke drew charts of the Atlantic Ocean showmg the surface '9^ DISCOVERY REPORTS isotherms for each month (bound separately in an ' Atlas '), and although these no doubt give a correct picture of the general distribution of Atlantic surface temperature, some features of the isotherms south of about 50 S (e.g. o and ^ 1° in March) are certainly not compatible with the Discovery data In his later paper (1938) Bohnecke gives a table of surface temperatures in 2° squares The Norwegian Meteorological Institute published a series of records (1935) collected by whaling factories for the International Polar Year, 1932^3. These include readings of sea temperature which are no doubt accurate, but the great majority were taken near the ice-edge and for our present purpose do httle more than confirm that temperatures here are about or below -- iC. Some temperatures recorded in the first part of the homeward voyages, however, are helpful, though these again do not fill any important gaps. A limited number of observations between the ice-edge and the convergence are to be found also in results of the ' Norvegia ' expeditions (see Mosby, 1933 1934) The Australasian expeditions under Sir Douglas Mawson provide material in the Australian Sector in some months in which the Discovery Committee's ships collected no data in that region Some of the temperatures recorded in the earlier expedition (1911^:4) seem very high, and here again it is TaZVI^T 1"!''^'''^'^ -^ f""y ^-1-ble; but those recorded in the Station List of the B.A.N.Z.A.R. Expedition of 1929-31 are undoubtedly reliable (see Mawson, 1940; and Johnson, 1937). METHOD OF DRAWING THE ISOTHERMS The distribution of surface temperature is of course subject not only to annual variations, but also to innumerable local and temporary variations and complexities. The object here is to draw isotherms to show as nearly as possible the average distribution of temperature for each month. Thus Plates II-XII show the temperatures which .^ould probably be found in certain positions at different times of year. For example, in January the 2° isotherm crosses longitude 30° W in c^° 30' S This means that according to past experience a ship arriving at that position about the middle of January IS more hkely to find that sea temperature than any other. If a higher or lower temperature is found It IS probable but not certain, that in that region the surface water is warmer or colder than usual, and the degree of probability will depend mainly on the number of different years from which data near to that position are available. Data from the preceding and succeeding months will be relevant as well as from the month in question. Since in some regions data are lacking, and in some other regions insufficient to indicate the extent . of annual variations. Plates II-XII can only be a first approximation to the average distribution of temperature The data are, however, very extensive, and the charts should give a more correct view of the distribution ot temperature in these waters than any that have been published before It IS obvious that the basic pattern of the isotherms must be a system of concentric rings, generally representing a falling gradient from north to south. The rings, however, are distorted in various ways by water movements and the configuration of the land and the sea floor. The gradient is not usually a steady one, but apart from small fluctuations it is not often inverted on a substantial scale. Inspection of the temperature charts ^ indicates that the highest temperatures generally prevail in February and suggests that the lowest occur about September and October (see also Fig. 6, p. 190). In drawing the isotherm charts I have assumed that any isotherm should be placed progressively farther south in each month from October to February, and similarly farther north from February to September A break in the average seasonal rise and fall of temperature is not perhaps impossible, but the assump- tion that there is no such break seems justified in the absence of evidence to the contrary. There are are rlftVero'l^Ts^thr "chti " ""' '°^ '""^ '^"^^'"^' ^'^"^^ °" ^^^'^'^ '^^ ^^^^ temperatures were plotted. Plates II-XII DISTRIBUTION OF SURFACE TEMPERATURE i99 comparatively few records for July and August, but these months are evidently not much warmer than September. Before any isotherms were drawn the Antarctic convergence was drawn in its mean position on each of the temperature charts, and the mean position of the pack-ice edge for the month in question was added. The latter was traced from the original charts reproduced in the report on the pack-ice (Mackintosh and Herdman, 1940, plates Ixxi-xcv), and no alterations were made except for one small adjustment. Inspection of the sea temperatures in the Falkland sector in November and re-examination of the ice records suggested that the mean position of the ice-edge here should be drawn a little farther south. Accordingly the line of the ice-edge in November between 10 and 70° W has been moved about 20-40 miles farther south. The lines of the ice-edge and the convergence form very useful starting points in the drawing of the isotherms. The line of the ice-edge can be taken as coinciding approximately with the isotherm of - 1-5° C. (It does not of course always coincide for the sea temperature at the ice-edge is sometimes above -1° in summer or below -1-5° in winter, but some allowance can be made for this.) From Tables 5 and 7 (pp. 189 and 192) we can calculate which isotherms should lie on the line of the con- vergence. If these steps are accepted then the intermediate isotherms must lie at intervals withm a limited belt. The isotherms are drawn at intervals of 1° C, and in the first place were sketched in pencil on the temperature charts. Those falling on the convergence were entered first, and they were derived from Tables 5 and 7. For example, in April in 60° W (see Plate VIII) the middle temperature by Table 5 is 4-0^^ and the range by Table 7 is 2-6°. This gives a gradient from 27 to 5-3° which includes the isotherms of 3, 4 and 5°. In 50" W, however, the gradient is from 3-2 to 5-0° which includes only the 4° and 5° isotherms. The f isotherm must therefore diverge away from the line of the convergence between 60 and 50'' W. In this way the isotherms were drawn along the whole length of the con- vergence for each month. 1 At some points (e.g. 10° E) the average range of the gradient is less than 1° C and the line of the convergence cannot accommodate more than one isotherm. Elsewhere the convergence can be distinguished by the concentration of two or more isotherms. The isotherms between the ice-edge and the convergence, and one or two also in sub-Antarctic water north of the convergence, were then sketched in on the temperature charts. With data from several different years included in a chart for one month, it was at once obvious that any attempt to join all points of a given observed temperature would lead to absurdities. However, such points of equal temperature were connected as often as possible provided this did not involve any improbable deviation from the general trend of the adjacent isotherms and of the lines of the ice-edge and convergence. If in any locality observations were available from only one year it was assumed that they represented average conditions, unless observations from that and other years in adjacent regions indicated any departure from the average. Where observations from different years suggested different latitudes for the isotherm the line was drawn as nearly as possible in the mean position. At many points of course it was difficult to decide where to place an isotherm, and even where observations were plentiful the isotherms could be drawn with much more confidence at some points than at others. However, they were drawn so as to make as good a fit as possible with the observed temperatures. Fig. 8 B (p. 197) is an example of a fairly good fit. In this area the October temperatures for 1930 and 1938 corresponded almost exactly In 1925 and 1926 in October and in 1936 in September the water was a little colder, at least in the northern part of the area. September and October temperatures in 1938 were very ''whereas the majority of isotherms were in this way drawn by eye on the charts of observed 1 But see below with regard to special adjustments at one or two points. 200 DISCOVERY REPORTS temperatures, the isotherms on the Hne of the convergence were derived from Tables 5 and 7. Ahhough both are dependent for the most part on the same body of data this might have resuhed in some inconsistencies, for Table 5 is no more than approximately correct, and as noted on p. 191 it is not certain that it is equally applicable to all longitudes. Furthermore, the position of the isotherms not on the convergence are in places influenced by extraneous data. However, the relative positions of isotherms on the convergence and of those to the south of it were generally found to be quite con- sistent, and only at one or two points was a compromise necessary. Thus in 150-170° E the 2° isotherm in December and the 3° isotherm in January are placed a little south of the convergence, although they should just be on it according to Tables 5 and 7, and a similar adjustment has been made in 60° W in January. Tentative isotherms were thus drawn for all months and for all regions except for the major areas in which observations were lacking, and the next step was to compare single isotherms month by month. For this purpose the isotherms of one temperature, for example all the 0° isotherms, for each of the months October to February (i.e. coldest to warmest months) were traced on to one chart, as in Plate XIII. It is to be supposed that during the whole of this period the sea temperature is on the average rising, and that therefore the 0° isotherm (or any other) will lie farther south in each successive month. This was found to be so, the 0° isotherms forming a succession of concentric rings which did not touch or intersect each other except at one or two points. Such instances were assumed to be due to abnormal conditions or annual variations, and the positions of the isotherms were adjusted accordingly. For example, it was found at one point that the November isotherm, depending mainly on temperatures in 1926, ran for a short distance a little north of the October isotherm which depended on temperatures in 1936. Further inspection of the temperature charts indicated that in an adjacent region the spring temperatures in 1926 were rather lower than in several other years. Therefore where the overlap occurred the November isotherm was adjusted so as to lie a little south of the October isotherm. The fact that such adjustments were few, and involved only small changes in the position of the isotherms, suggested that they had been drawn not far from the correct positions. Some other adjustments were made so that isotherms of adjacent months should be approximately parallel and show similar features ; but this was only done where it involved small alterations which were not too inconsistent with the observed temperatures. Sometimes the form of an isotherm changes considerably from one month to the next (cf. the 0° isotherms for December and January between 30° W and 30° E, Plate XIII). One isotherm having been adjusted in this way, it was traced back in its new form on to the tem- perature charts and the process was repeated for the other isotherms for September-February, and then for all isotherms for February-September (the period of falling temperatures). It amounts of course to a means of supplementing the data for one month with the data for other months on the assumption that there is an unbroken rise of temperature from October to February and an unbroken fall from February to September (see p. 198). The revised isotherms were checked against the temperature charts in case of any unjustified departures from the observed temperatures, and were then transferred from the original semicircular charts to the circumpolar charts of a rather smaller scale which are reproduced in Plates II-XII. It remained to fill in the isotherms in the regions in which no observations had been made (mainly in the winter months). These were sketched on the circumpolar charts in what appeared to be the probable positions by analogy with the isotherms already drawn. Each isotherm was then traced as before on to a separate chart, adjusted with the same isotherms for other months, and traced back again. These isotherms are shown as pecked lines in Plates II-XII. Often it was a matter of simple interpolation. For example, between 80 and 110° W the 0° isotherms for November and January are based on DISTRIBUTION OF SURFACE TEMPERATURE 201 observed temperatures, but there were no data here for December. The December isotherm, however, must he between those for November and January, and since the observed temperatures nearly everywhere indicate that it Hes nearer to the November than to the January isotherm there was no difficuhy in deciding where to draw it here (see Plate IV). A good interpolation of this kind may some- times give a more reliable position for the isotherm than a limited set of direct observations. Farther west the November and December isotherms are both interpolated between those for October and January, and in the winter months in the Pacific sector the 0° isotherms for five or six consecutive months have to be interpolated (see Plate XIV). Here of course there is less certainty, especially as the September isotherm rests on observations only from one year (1932). Even so there is not very much choice in the positions in which they might be drawn. NOTES ON THE DISTRIBUTION OF TEMPERATURE It will be realized that Plates II-XII are semidiagrammatic charts which represent only the major features of temperature distribution. Surface temperatures are naturally subject to some appreciable widespread variations and to more numerous local variations ; and if the isotherms could be drawn in full detail for any area at a given time they would usually reveal elaborate complexities, and bear little resemblance to the smooth lines in Plates II-XII. Fig. 9 shows the apparent actual distribution of temperature between South Georgia and the Falkland Islands in September and October 1934, and may be compared with Fig. i (p. 182) and with the corresponding region in Plate II. The isotherms are correctly placed where they cross the ship's tracks, and between these tracks they are drawn in what seem to be the most probable positions ; but there is not much doubt that, if additional lines of observations could have been made at the same time, a still more complex picture would be presented. co- Fig. 9. Observed temperatures in the Scotia Sea, 1st September to 14th October, 1934, showing the apparent positions of the isotherms. Half-degree iso- Most charts of temperature distribution as observed therms are shown as pecked lines, and the ship's tracks at a particular time would be likely to show some features, such as tongues of warm or cold water and other bends in the isotherms, which are temporary, and which become smoothed out when the isotherms are drawn in their mean positions. Plates II-XII are intended to show only those features which seem to be constant, or which at least are found in most of the years for which observations are available. Even such constant features have been drawn with caution, and it may be that some of them are not sufficiently emphasized. The following points may be noted. South of the Falkland Islands in about 55^ W there is some evidence of a northward thrust of cold water. This is noticeable in the chart for September and October (Plate II) where the isotherms are crowded up to the convergence. For this chart there are data only for one year (see Fig. 9), but the bend in the convergence here and the similar shape of the isotherms in other months suggest that this is a constant feature. It is perhaps more noticeable in spring than m summer. A more conspicuous tongue of cold water projects northwards around the South Sandwich Islands (near 30° W). This is often very pronounced, and it is probably connected with a comparatively narrow thrust of cold water off the north-east side of South Georgia, which was found during intensive observations around the island. Warmer water is found to the east of the South Sandwich Islands, and it is possible that the southward bend of the isotherms about 10-20° W is not sufficiently emphasized in the charts. It is best seen in the charts for March and April, but it is not certain that it is really 202 DISCOVERY REPORTS more conspicuous in these months than at other times of year. It seems probable that further examina- tion of such features as these will show that they are correlated with the effects of bottom topography. Between about io° W and 30° E the positions of the isotherms are very much influenced by the eastward flow of cold water from the Weddell Sea, which also has an important effect here on the distribution of pack-ice in the early summer (see Mackintosh and Herdman, 1940, p. 293). In winter and spring the isotherms tend to be pressed up towards the convergence which here lies about its lowest latitude, but in December, while pack-ice and cold water persist in a relatively low latitude between o and 30° E, east of 30° both ice-edge and isotherms bend far to the south, and the ice belt begins to break up internally, leaving an outer zone which contracts towards the South Sandwich Islands (compare the ice-edge in Plates IV and V, December and January). As the outer zone of ice 60 ■ S-6 - 40- 3-0- 2-D- 1-0- 00- -l-O- 1 1 1 52 53 54 1 55 1 55 1 57 1 53 1 59 1 60 1 61 1 62 I 63 1 64 I 65 t 6S \ 67 1 6B DcGRCES SOUTH Fig. 10. Comparison of average monthly temperatures, based on monthly isotherms. melts away it tends to leave a long tongue of cold water which is seen as a conspicuous S-shaped turn in the isotherms in Plates V-VII. The shape of the isotherms here is subject to a good deal of variation. The data indicate that a line of observations running south from the convergence will usually, but not always, reveal a slight rise in temperature about 60-65° S, and if there is no actual rise in temperature, there will at least be a long expanse of ocean in which the temperature at the surface will change very little. The existence of a belt of relatively warm water here, which melts the ice in a high latitude while a zone of pack-ice still persists to the north of it, is noted by Deacon (1937, pp. 18, 28, etc., and Fig. 8), who describes it as an ill-defined divergence region with an upwelling of warm water between the Weddell Sea current moving eastwards in a lower latitude, and the current moving westward in a higher latitude near the continental coast. Deacon also refers (p. 30) to the outer and inner belts of pack-ice, and adds : ' It is just possible that even in winter there may be open water between the two ice-streams in the eastern part of the Atlantic Ocean.' 203 DISTRIBUTION OF SURFACE TEMPERATURE Meridional lines of observations in this region in April are only available for one year (1938). No warm belt was then distinguishable, and the isotherms in Plate VIII are drawn accordingly, t is no perhaps certain that the S-shaped bend in the isotherms is normally smoothed out in Apri bu it may be that the advance of winter conditions has by then obliterated any signs of a warm belt at the The loop in the convergence in 30° E is referred to on p. 184. It is possible that more of the adjacent isotherms are involved in this disturbance than the charts suggest, but more data would be needed to decide this point. . , ^^^^^„t The course of the isotherms in the Indian Ocean and Australian regions call for no special comment. The charts are self-explanatory, and it will be seen that more data are needed in some months. The Ross Sea is another locality in which open water is found in a high latitude while pack-ice still lies to the north. The isotherm for - 1° should perhaps reappear here in January, February and March, but there is little material to show in what shape it should be drawn, and it is therefore omitted. 4 0- 3-0^ u 2-0- o kJ a: D 5G S I 0- 0. 5 u 1-0-0- -I 0- , I I I I I I I I I I I I I ' ' SEPT OCT NOV DEC JAM FEB MAR APR MAY JU^.E JULY AUG SEPT OCT NOV DEC Fig. II. Seasonal rise and fall of temperature, derived from Fig. 10. In the Pacific sector the northward extension of cold water between 120 and 160° W, and the com- pression of S^ isotherms in the winter and spring months in the Bellingshausen Sea and Drake Strait ^"XTs^e^^e^^^^^^^ of the temperature m different months and diff nf iXdes. For Fig. :o the latitude of each isotherm in each month ^^^--f-]^ II-XII at every 30^^ of longitude, and the average was plotted. For example, in Plate VI (Februa^^ tL avLag lal^ of the 0° isotherm was found to be 65° 40' S. Thus m Fig xo the curve f r the average latitu isotherms are correct the diagram should '^eTrect ii; esl ^Z^^. i. ave.ge .e.pe..u.e between .he differen. .on.hs and o ,he Ze of the temperature gradient south of the convergence. The slope of the gradient on the cllrgence and to the north of it, and differences in the shapes of the monthly curves are perhaps notve^ rel able The figure shows that surface temperatures are much the same m January and March, : il Nov mber and j!ly, that south of the convergence the water is warmer „t December than m T ,! b,„ if wou d OTobably not be justifiable to infer much more than this, '"pt ;'is de W dtom' FTg .0, the teripera.ure for each month being plotted at arbitrary m.ervals of Lfitude Ttflattening of the curves at the foot of the diagram indicates that the water ,s (on the 204 DISCOVERY REPORTS average) covered by pack-ice in the months and latitudes in question. The figure suggests that when the sea becomes free of ice in the early summer, the rise in temperature (December-January) is more rapid in high than in low latitudes. Here again, however, the curves can only be taken as a rough indication of the rise and fall of temperature. REFERENCES BoHNECKE, G., 1936. Temperatur, Salzgehalt wid Dichte an der Oberfldche des Alla7itiscben Ozeans. Erste Liefcrimg. Wissen- schaftliche Ergebnisse der Deutschen Atlantischen Expedition auf dem Forschungs- und Vermessungsschiff 'Meteor', 1925-7, V, pp. 1-186, and Atlas. BoHNECKE, G., 1938. Tetnperatur, Sahgehall und Diclite an der Oherfldche des AtlanHschen Ozeans. Ziveite Lieferung. Wissenschaftliche Ergebnisse der Deutschen Atlantischen Expedition auf dem Forschungs- und Vermessungsschiff 'Meteor', 1925-7, v, pp. 187-250. Deacon, G. E. R., 1933. A general account of the hydrology of the South Atlantic Ocean. Discovery Reports, vii, pp. 171-238, pis. viii-x. Deacon, G. E. R., 1937. The hydrology of the Southern Ocean. Discovery Reports, xv, pp. 1-124, pis. i-xliv. Hart, T. J., 1934. On the phytoplankton of the South-West Atlantic and the Bellingshausen Sea, 1929-31. Discovery Reports, VIII, pp. 1-268. Hart, T. J., 1937. Rhizosolenia curvata Zacharias, an indicator species in the Southern Ocean. Discover}' Reports, XVI, pp. 413-46, pi. xiv. . „ r r Hart, T. J., 1942. Phytoplankton periodicity in Antarctic surface waters. Discovery Reports, xxi, pp. 261-356. Hastings, .'\. B., 1943. Polyzoa {Bryozoa). Discovery Reports, xxii, pp. 301-510, pis. v-xiii. John, D. D., 1936. The southern species of the genus Euphausia. Discovery Reports, xiv, pp. 193-324. Johnson, T. H., 1937. Biological organisation and Station List. B.A.N.Z. Antarctic Research Expedition, 1929-31. Reports, Ser. B, I, pt. 1, pp. 1-48. LuMBY, J. R., 1928. Modification of the surface sampler with a vino to the improvement of temperature observation. Journal du Conseii, iii, 3, pp. 340-50. Mackintosh, N. A., 1934. Distribution of the macroplankton in the Atlantic sector of the Antarctic. Discovery Reports, ix, pp. 65-160. Mackintosh, N. A., 1937. The seasonal circulation of the Antarctic tnacroplankton. Discovery Reports, xvi, pp. 365-412. Mackintosh, N. A. and Herdman, H. F. P., 1940. Distribution of the pack-ice in the Southern Ocean. Discovery Reports, XIX, pp. 285-96, pis. Ixix-xcv. Mawson, D., 1940. Ilydrological Observations of the Australasian Antarctic Expedition, 1911-14. Sci. Reps. Ser. A, I, pt. 4, pp. 103-25. Meinardus, W., 1923. Meteorologische Ergebnisse der Deutsche Siidpolar-Expedition, 1901-1903. Deutsche Siidpolar- Expedition. HE Meteorologie, Bd. I, Heft 1. Meteorological observations made on 9 Norivegian ■whaling floating factories during the International Polar Year 1932-1933. Publications from the International Polar Year 1932-33, No. 1. Det Norske Meteorologiske Institut. Oslo, 1935, pp. 1-53- MosBY, H., 1933. The sea-surface and the air. Scientific Results of the Norwegian Antarctic Expeditions, 1927-1928 et sqq.. No. 10. Det Norske Videnskaps-Akademi i Oslo, pp. 1-140. MosBY, H., 1934. The waters of the Atlantic Antarctic Ocean. Results of the Norwegian Antarctic Expeditions, 1927-1928 et sqq.. No. 11. Det Norske Videnskaps-Akademi i Oslo, pp. 1-131. Neaverson, E., 1934. The sea-floor deposits: 1. General characters and distribution. Discovery Reports, ix, pp. 295-350, pis. xvii-xxii. Norman, J. R., 1938. Coast Fishes. Part III. The Antarctic zofie. Discovery Reports, xviii, pp. 1-104, P'- i- SvERDRUP, H. v., Johnson, M. W. and Fleming, R. U., 1942. The Oceans. Their physics, chemistiy and general biology, pp. 1-1087. New York. • 205 A.PPENDIX Table 9. r/(e Antarctic convergence. List of all occasions oft which the convergence has been crossed or located by ships of the Discovery Committee Abbreviations : adj. adjacent ship s T.s Routine 4-hourIy temperatures (Fahrenheit) entered in ship's log approx. approximate St. Station betw. between surf. surface D.R. dead reckon! ng temp. temperature hr. hours W.S 'Will lam Scoresby' min. minimum Note. Where the hour is noted in the ' Remarks ' column it refers to ship's time on the date in the 'Date' column, and is given mainly to facihtate reference back to the original data . Degree No. Ship Voyage Date Estimated position of Remarks accuracy I ' Discovery ' Tristan da Cunha to 14- ii. 26 47° IS' S, 25° 20' W Probable By min. temp. betw. Sts. 10 and 11. By S. Georgia 14. ii. 26 48° 05' S, 27° 05' w ») thermograph crossing loops of the con- IS- ii. 26 49° 10' S, 28°35'W »j vergence at 0800 and 2000 hr. on 14th, IS- ii. 26 49° 20' S, 28°55'W »» 1500 and 1800 hr. on 15th, and 0100, 16. ii. 26 49° S3' S, 29° 40' W »» 0400 and 1700 hr. on i6th, the last 16. ii. 26 50° 10' S, 30° 00' W ' crossing being the most definite. 16. ii. 26 5°°3S'S, 30^ 45' W Good 2 »> S. Georgia to Falkland Is. 20. iv. 26 52° 10' S, 43° 00' W Uncertain No adj. Sts. Thermograph not clear; probably 0200 hr. 3 'W. Scoresby' Cape Town to S. Georgia 28. X. 26 47° 00' S, 06° 45' w Probable Betw. Sts. WS II and 12. Ship's T.s indicate about 1800 hr. 4 ' Discovery ' >> 6. xi. 26 47° 10' S, 14° 10' E Uncertain i No adj. Sts. Thermograph suggests 1800 1 \ ' hr., but doubtful. 5 n S. Georgia to S. Orkneys 8. ii. 27 S4° 35' S. 42° 50' w Approx. Thermograph rises to over 4° C at 1600 hr., indicating proximity of a loop of the convergence. 6 'W. Scoresby* S. Georgia to Falkland Is. 20. ii. 27 S3°i9'S, 45^ 40' w Good By min. temp. betw. Sts. WS 67 and 68. By ship's T.s about 1500 hr. 7 * Discovery * S. Shetlands to C. Horn 18. iv. 27 58° 28' S, 67°SS'W Approx. By min. temp, about St. 216. By thermo- graph about St. 217. 8 *W. Scoresby' Falkland Is. to S. Georgia 13- v. 27 52° 30' s, 48°i5'W Approx. No. adj. Sts. By ship's T.s about 1800 hr. Position by D.R. 9 »» S. Georgia to 31- v. 27 49° 12' S, 35°5o'W Probable No temps, at Sts. Ship's T.s show marked Gough I. I. vi. 27 48° 40' S, 34° 00' W Uncertain rise about 1800 hr. on 31. v. 27, and I. vi. 27 47° 47' S, 3i°34'W )) indications of loops at about 0200 and 1500 hr. on I. vi. 27. 10 )» Falkland Is. to S. Georgia II. ii. 28 S3° 17' S, 47° 02' W Approx. By min. temp, about St. WS 140. St. showed surf, cooled by nearby ice island and brash ice. II )» S. Orkneys to 26. iv. 28 S6° 27' S, 54° 22' W Good By min. temp, about St. WS 204. Ship's Falkland Is. 27. iv. 28 56° 24' s, S4° 30' W Uncertain T.s show rise at St. WS 204 and indicate 27. iv. 28 56° 12' S, 55" 00' w )» loops about 0800 and 1400 hr. on 27. iv. 28. 12 )) Falkland Is. to 22. viii. 28 52° so' S, 47° 40' W Good By min. temp. betw. Sts. WS 253 and 254. S. Georgia By ship's T.s about 0600 hr. 13 )» S. Georgia to Falkland Is. 4- xii. 28 52°4S'S, 48° 25' W »> By min. temp. betw. Sts. WS 316 and 317. By ship's T.s about 0200 hr. 14 J) Falkland Is. to S. Georgia 14. xii. 28 S2° 30' S, 49° 00' W ** No. adj. Sts. Ship's T.s show clear con- vergence at 1600 hr. 15 t» S. Shetlands to C. Horn 23- ii. 29 58°2i'S, 65°37'W »» By min. temp. betw. Sts. WS 404 and 405. No clear indication by ship's T.s. 16 1) Falkland Is. to S. Georgia iS/i 5. iii. 29 .? — No temps, at Sts. Ship's T.s indeter- minate. 17 )> S. Georgia to Falkland Is. 30- iv. 29 52° 40' S, 48° 09' W Probable By min. temp, near St. WS 429, but ship's T.s indicate on St. WS 430. 18 j» Falkland Is. to S. Georgia 7- V. 29 52°4S'S, 46° 20' W Approx. No. adj. Sts. Ship's T.s show conver- gence, but D.R. positions doubtful. 19 )) S. Georgia to 19. v. 29 50° 55' s, 29° 53' w Uncertain By min. temp. betw. Sts. WS 435 and 437. Cape Town 19. V. 29 So° 13' S, 27' 40' W jj Ship's T.s suggest running parallel to convergence between positions shown (1200-0000 hr.). 206 DISCOVERY REPORTS Table 9 [cont.) ■ • Degree No. Ship Voyage Date Estimated position of Remarks accuracy 20 'W. Scoresby' Cape Town to 19/27. X. 29 47° 12' S, 07° 50' w Uncertain No informative Sts. Ship's T.s showing S. Georgia 52° 15' s, 28° 20' w >» fluctuating temps, betw. positions shown. 1800 hr. on 19. x. 29 to 0200 hr. on 27. X. 29. 21 »» S. Georgia to Falkland Is. 2. xi. 29 52^34'S, 48° 48' w ») No temps, at Sts. Ship's T.s suggest at or near St. WS 466. 22 » Falkland Is. to S. Shetlands II. xi. 29 57° 16' S, 57° 14' W Approx. By min. temp. betw. Sts. WS 469 and 470. Ship's T.s fluctuating. 23 >> S. Shetlands to Falkland Is. 9- i.30 57° 50' S, 59° 40' W )» No adj. Sts. Ship's T.s clearly indicate 0600 hr., but latitude doubtful. 24 ' Discovery II ' Montevideo to i5(?)-i-30 ? — No adj. Sts. No indication by thermo- S. Georgia graph. 25 'W. Scoresby' Falkland Is. to S. Georgia 16. i. 30 52°45'S, 44°35'W Probable No adj. Sts. Ship's T.s suggest 0200 hr. 26 i» S. Shetlands to Falkland Is. 20. ii. 30 56°47'S, 58°i5'W Approx. No adj. Sts. Ship's T.s indicate 0200 hr., but positions not certain. 27 >> NW'ward in 55° W 3- iv. 30 55°45'S, 54° 50' W Good By min. temp. betw. Sts. WS 529 and 530. Ship's T.s indicate 0200 hr. 28 >» E'ward in Scotia Sea 6. iv. 30 55°28'S, 54° 30' W Approx. Noadj.Sts. Byship'sT.sat iSoohr. Sub- sequently probably running parallel to convergence. 29 'Discovery II' S. Shetlands to C. Horn 14. iv. 30 59° 15' S, 64° 10' W V. good By min. temp, near St. 384. By thermo- graph at 2330 hr. 3° >i C. Horn to 18. iv. 30 55°47'S, 52° 50' W Good Surf, temps, only at Sts. By thermograph S. Georgia 18. iv. 30 55°4o'S, 5i°oo'W J» clearly running parallel to convergence 19. iv. 30 55° 10' S, 47° 20' W )» and crossing loops at 0830 and 1900 hr. on i8th, and 0530 on 19th. Some addi- tional fluctuations of temp. E of 47° W. 31 'W. Scoresby' S. Georgia to Rio. ?. V. 30 ? — No adj. Sts. Ship's T.s indeterminate. 32 'Discovery II' S. Georgia to Cape Town 13- v. 30 48° 21' S, 2i°55'W V. good Surf, temps, only at Sts. Sharp rise in temp, at 1800 hr. Temp, rather high for May. 33 >» Cape Town to Bouvet I. 14. X. 30 48° 20' S, 10° 00' E Good By min. temp. betw. Sts. 451 and 452. Thermograph shows slight but distinct fall at 0900 hr. 34 'W. Scoresby' Montevideo to S. Georgia 9- j-3i 50° 26' s, 4i°24'W Approx. No adj. Sts. Ship's T.s show changes at 0600 and 1800 hr. 35 'Discovery II' W'ward from S. Georgia 4- iii. 31 54° 05' S, 46 ' 00' W V. uncer- tain Surf, temps, only at Sts. Thermograph indeterminate but probably crossed 36 convergence at 0430 hr. yt S'ward in Scotia Sea 6. iii. 31 55° 40' S, 5i°3o'W >j By min. temp, near St. 633. Thermograph indeterminate. Crossing possibly at 0500 hr. 37 »» S. Shetlands to 12. iii. 31 58" 00' S, 60° 10' W V. good By min. temp. betw. Sts. 648 and 649. By 38 Staten I. thermograph at 1930 hr. »> Falkland Is. to 24. iii. 31 53° 17' S, 47°35'W Probable By min. temp. betw. Sts. 656 and 657. By S. Georgia thermograph at 1200 hr. 39 'W. Scoresby' S. Georgia to Falkland Is. 13- iv. 31 52° 30' s, 47°45'W Uncertain No. adj. Sts. By ship's T.s about 0200 hr. 40 'Discovery 11' N'ward in 30° W 18. iv. 31 49° 50' S, 29°55'W V. good By min. temp. N of St. 666. By thermo- 41 1} S'ward in 75° W 20. xi. 31 59°47'S, 75°o5'W Approx. graph at 0900 hr. By min. temp. betw. Sts. 730 and 731. 21. xi. 31 62° 05' S, 75°oi'W )» Thermograph indicates confused loops 21. xi. 31 62° 53' s, 75° 02' W )» with sharper fluctuations farther S. Times adopted: 1400 hr. on 20th, 1330 and 1930 hr. on 21st. 42 >) N'ward in 56° W 28. xi. 31 56° 50' s, 55° 50' W V. good 7 J By min. temp. betw. Sts. 745 and 746. By 43 »» S'ward in 49° W 3- xii. 31 55° 50' S, 49° 02' W »» thermograph at 1700 hr. By min. temp. betw. Sts. 754 and 755. 4- xii. 31 56° 10' S, 48° 58' w >» Sharp convergence at 1845 hr. on 3rd, 4- xii. 31 56° 30' s, 48° 56' w »> and well-marked loop or patch of sub- Antarctic water at 0100 and 0300 hr. on 4th. THE ANTARCTIC CONVERGENCE Table 9 (cont.) 207 Ship 44 45 46 47 'Discovery II' 48 49 5° SI 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 Voyage North of S. Georgia »> )» Date S. Georgia to Falkland Is. Falkland Is. to S. Georgia S. Georgia to Cape Town SE'ward from Cape Town NE'ward to Fremantle SE'ward from Fremantle NE'ward to Melbourne S'ward from Melbourne N'ward to N. Zealand SE'ward from N. Zealand NE'ward betw. 150° and 140° W SE'ward in ii3°W NE'ward to Magellan Str. S'ward in 80° W S. Shetlands to Falkland Is. S'ward in 45° W North of S. Georgia S. Shetlands to Falkland Is. Falkland Is. to S. Georgia N'ward in 13° E 16. xii. 31 17. xii. 31 10. ii. 32 20. ii. 32 Estimated position Degree of accuracy Remarks 25. n. 32 25. ii. 32 25. ii. 32 25. ii. 32 26. ii. 32 14. iv. 32 50^48'S, 37^22'W 49"4o'S, 37'15'W 52^40'S, 43'30'W 52'=So'S, 43^15'W 49°45'S, 24°03'W 49° 25' S, 23° 00' W 48°S5'S, 2i°35'W 48°35'S, 2o°35'W 48° GO'S, i8''5o'W 5o°25'S, 3i°3o'E Approx. Good I. V. 32 23- V. 32 31- V. 32 20. vi. 32 27. vi. 32 7. ix. 32 13. ix. 32 14. Lx. 32 14. ix. 32 26. ix. 32 30. ix. 32 28. x. 32 9. xi. 32 18. xi. 32 4. xii. 32 4. xii. 32 7- ii- 33 21. ii. 33 20. iii. 33 Probable Good By min. temp, near St. 776. By thermo- graph near St. 775. By min. temp, near St. 776. By thermo- graph probably about 1 800 hr. , but some indications of loops in the convergence. No. adj. Sts. By ship's T.s about 1 100 hr. By min. temp. betw. Sts. 829 and 830. By thermograph at 0600 hr. betw. Sts. 830 and 831, showing surface stratum of sub-Antarctic water drifted east of nor- mal position. Surface position adopted. Cf. No. 46. Surf, temps, only at Sts. By thermograph crossing loops near St. 835 at 0300, 0730, 1500, and 1900 hr. on 25th, and 0200 hr. on 26th. 52° 00' S, 95° 50' E Probable 52° 07' S, 121° 35' E 54° 25' S, 136° 00' E 53° 45' S, 151° 33' E 54° 15' S, 151° 55' E 55° 00' S, 152° 00' E 54° 30' S, 162° 35' E 57° 50' S, 165° 35' w 55° 30' S, 144'^ 55' W 53° 45' S, 141° 45' w 53° 05' S, 140° 50' W 56° 55, S, 113° 25' W 61° ID'S, 60° 20' s, 60° 10' s, 62° 45' s, 55° 17' S, 5i°2o'S, 49° 35' S, 50° 28' S, 93° 20' W 9i°i5'W 90° 45' W 80° 00' w 56° 00' w 44° 41' W 36°35'W 36° 57' W 58°2o'S, 6o°35'W 52°i3'S, 47°25'W 48°3o'S, i3°3o'E By min. temp. betw. Sts. 849 and 850. By thermograph at 2130 hr. Indication of ■ long northward bend in convergence approaching ship's track betw. Sts. 848 and 849. By min. temp. S of St. 866. By thermo- graph at 1500 hr., but loop comes near ship's track at 1800 hr. on 2nd. Approx. By min. temp, near St. 883. By thermo- graph about 1500 hr. Good By min. temp. betw. Sts. 891 and 892. By thermograph at 1000 hr. By min. temp, near St. 903. By thermo- graph at 0200 hr. with loop at 0630 and 1200 hr. V. good By min. temp, near St. 920. By thermo- graph at 0130 hr. Probable By min. temp, near St. 949. By thermo- graph probably at 1000 hr. Good By min. temp. betw. Sts. 961 and 962. By thermograph at about 0700 hr. Well- defined loop of Antarctic water 0530 to HOC hr. on 14th betw. Sts 962 and 963- Probable By min. temp. betw. Sts. 970 and 972. By thermograph at oioo hr. By min. temp. betw. Sts. 975 and 976. By thermograph at 0330 hr., with loop of Antarctic water at 1 130 and 1300 hr. V. good By min. temp. betw. Sts. 990 and 991. By thermograph at 0500 hr. By min. temp. betw. Sts. 1017 and 1018. By thermograph at 0500 hr. Uncertain By min. temp, near St. 1027. By thermo- graph indeterminate. V. good By min. temp. betw. Sts. 1054 and 1055. By thermograph at 0400 hr. By min. temp. betw. Sts. 1055 and 1056. By thermograph at 2300 hr. just S of 1056. Approx. By min. temp. S of St. 1 1 17. By thermo- graph probably at 1600 hr. Uncertain By min. temp. E of St. 1 123. By thermo- graph at 0400 hr. By min. temp. S of St. 1162. Thermo- graph suggests 0900 hr. 208 DISCOVERY REPORTS Table 9 (cont.) No. Ship 67 Voyage Date 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 'Discovery II' Tristan to S. Georgia N'ward in 78° W S'ward from Falkland Is. N'ward to Auckland N. Zealand to Ross Sea N'ward in 79" W Falkland Is. to Elephant I. NW of S. Georgia Enderby Land to Durban Cape Town to S. Georgia SW'ward to S. Georgia S. Georgia to S. Shetlands N'ward in 78' W Falkland Is. to Elephant I. N'ward in 46° W NW of S. Georgia S. Georgia to Falkland Is. S'ward in 79° W E'ward betw. 105° and 96° W 23- 24. 24. 25- 25- i6. •^'- 33 -^i- 33 xi-33 xi- 33 xi-33 -'^''- 33 Estimated position 47° 56' S, 23° 10' W 48° 30' S, 25° 15' W 49° 04' S, 26° id' W 49°48'S, 28"o3'W 1 49° 55' S, 28°34'W! 61° 40' S, 78" 00' vv Degree of accuracy Probable 29. 29. ^11- 33 ■^>'- 33 29. xii. 33 29. xii. 33 29. xii. 33 29. xii. 23 29. xii. 33 3°- xii. 33 30- ^ii- 33 22. i. 34 19. n. 34 12. iii. 34 29. iii. 34 8. iv. 34 8. iv. 34 15- V. 34 II. viii. 34 21. viii. 34 22. viii. 34 23. viii. 34 23. viii. 34 4. ix. 34 5. ix. 34 13. ix. 34 26. ix. 34 55" 24' s, 55° 30' S, 55°48'S, 56° 25' S, 57° 20' S, 57'^ 40' S, 57° 58' S, 58° 20' S, 58° 45' S, 57°58'S, 60° 10' W 6o°i5' W 60" 28' W 60° 4s' W 6i°3o'W 6i°48'W 62° 05' W 62° 25' W 62° 50' W 170° 45' W 60° 47' S, 174" 50' w 63°3o'S, 79°o8'W 56° 35' S, 55° 40' W 5i°2o'S, 44" 09' W 5i°2o'S, 42°4o'W 50°20'S, 44°54'E 48'55'S, 04^15'W 47 4°' 48° 57' 49° 45' 49° 55' 55° 40' 56° 03' 61° 25' 56° 20' I. x-34 56° 06' 2. x-34 55° 12' 3- ^•34 51° 55' 5- x-34 51° 20' 12. •^•34 52° 48' 29. ^•34 62" 10' 9- xi. 34 61° 02' 10. XI. 34 60° 15' 10. XI. 34 60° 03' II. XI. 34 59" 35' 11. XI. 34 59° 33' II. XI. 34 59" 40' S, 23°56'W S, 25°4o'W S, 26°5o'W S, 27'05'W S, 44°2o'W S, 45"2o'W S, 78° 30 W S, 55^40'W S, 45°47'W S, 45°45'W S, 45°2i'W S, 42°25'W S, 47°57'W S, 79'2S'W S, 105" 05' w S, 100° 45' w S, 99°3o'W S, 97°i5'W S, 96°28'W S, 96° 10' W V. good Uncertain J) Probable V. good Good V. uncer- tain Good Uncertain Probable Gt)od )) Approx. V. good Good Remarks By min. temp. betw. Sts. 1190 and 1198. Conditions confused at intermediate stations. Loops probably crossed at 1500 hr. on 23rd, 0830 and 1330 hr. on 24th, and 0300 and 0530 hr. on 25th. Course of convergence very confused. By min. temp, near St. 1224. By thermo- graph at 0400 hr. By min. temp. betw. Sts. 1233 and 1234. By thermograph a series of loops or patches at 0030, 0200, 0330, 0730, 1400, 1600, 1800, 0100 and 0400 hr. By min. temp. S of St. 1276. Thermograph indeterminate, but 0800 hr. adopted for position. No. adj. Sts. Thermograph suggests 1200 hr. By min. temp, near St. 13 15. By thermo- graph at 0600 hr., but ship near con- vergence until 0400 on 13th. By min. temp. betw. Sts. 1325 and 1326. By thermograph at 0400 hr. By min. temp. betw. Sts. 1337 and 1338. By thermograph at 0500 hr. By min. temp. betw. Sts. 1338 and 1339. By thermograph at 1700 hr. By min. temp. betw. Sts. 1366 and 1367. By thermograph at 0200 hr. Surf, temps, only at Sts. Thermograph indeterminate, but crossing perhaps at 0730 hr. By min. temp, near St. 1391. By thermo- graph sub-Antarctic water SW of 1391 but not SE of it. Crossings at 1400 hr. on 2ist, about 2000 hr. on 22nd, and 0400 and 0530 hr. on 23rd. Surf, temps, only at Sts. By thermograph ship's track cuts loop of convergence SW of S. Georgia at 1900 and 0300 hr. By min. temp. betw. Sts. 1416 and 1417. By thermograph about 1800 hr. By min. temp. betw. Sts. 1424 and 1425. By thermograph at 0700 hr. Surf. temp, only at St. By thermograph ship's track cuts loop of convergence SW of S. Georgia at 1700 and 0500 hr. By min. temp. betw. Sts. 1434 and 1435. By thermograph at 2100 hr. By min. temp. S of St. 1435. By thermo- graph at 0200 hr. By min. temp. betw. Sts. 1439 and 1440. By thermograph at 0300 hr. By min. temp. betw. Sts. 1446 and 1447. By thermograph at 0500 hr. By min. temp. St. 1467 very close to con- vergence. By thermograph crossing loops at 2100 hr. on 9th, 1230 and 1700 hr. on loth, and 0430, 0830 and 1330 hr. on nth. THE ANTARCTIC CONVERGENCE 209 Table 9 (cont.) Degree No. Ship Voyage Date Estimated position of Remarks accuracy 88 'Discovery II' N'ward in 80= W 17- xi-34 60' 20' S, 79° 54' W Uncertain By min. temp, close to St. 1476. By thermograph doubtful. 89 'W. Scoresby' SE'ward from Cape Town xi./ xii. 34 ? No adj. Sts. Convergence doubtful by ship's T.s. 30. xi. or i. xii. 34. 90 ' Discovery II ' Falkland Is. to 4- xii. 34 56" 05' S, 59° 00' W V. good No adj. Sts. By thermograph at 0030, S. Shetlands 4- xii. 34 56^ 55' S, 59° 20' w }) 0700 and 1430 hr. on 4th. Rise of temp. 4- xii. 34 57' 54' S, 59° 45' W )» to 2-4° centred at 0800 hr. on 5th 5- xii. 34 59° 47' S, 60" 20' W Good indicates proximity of a loop of the convergence. 91 >> N'ward in 44^ W 25- i-35 52° 06' S, 44° 05' W >» By min. temp, at St. 1495. By thermo- graph at St. 1495. 92 )» NW'ward to Cape Town 7- "•• 35 49° 50' S, 3i°05'E >» By min. temp. betw. Sts. 1552 and 1553. By thermograph at 0400 hr. 93 'W. Scoresby' 1 NE'ward to Cape Town 24. "i- 35 50" 57' s, 05°i5'E Uncertain No adj. Sts. By ship's T.s probably about 1000 hr. 94 ' Discovery II ' E'ward betw. 20° 2. iv.3S 48° 50' s, 20= 45' E Good By min. temp. St. 1561 in Antarctic water. and 35 E 5- iv.3S 49° 35' S, 29° 07' E j» By thermograph crossings at 11 00 hr. 5- iv-35 49° 30' S, 3i°oo'E )» on 2nd, 0500 and 1600 hr. on 5th, and 6. iv.35 48° 40' S, 34° 10' E Approx. about 0900 hr. on 6th. Pronounced bend in convergence about 30' E. Cf. No. 49, 14. iv. 32. 95 >> S'ward in 56' E 20. xi. 35 48° 08' S, 56" 30' E Uncertain By min. temp, near St. 1618. By thermo- graph probably 0800 hr. on 20th. 96 jj NE'ward to Fremantle 7- xii- 35 52° 30' s, 117' 05' E Probable No adj. Sts. By thermograph at 1700 hr. 97 *W. Scoresby* S'ward from Cape Town 9- xii- 35 49° 20' S, 18° 48' E )) No. adj. Sts. By ship's T.s about 1600 hr. 98 * Discovery II ' N. Zealand to Ross Sea Balleny Is. to 6. i.36 60" 22' s, 178° 15' E )» No adj. Sts. By ship's T.s about 2100 hr. 99 )> 8. ii. 36 57° 20' S, 163° 00' E Approx. By min. temp. betw. Sts. 1679 and 1680. Melbourne By ship's T.s about 0600 hr. 100 )» S'ward from Melbourne 9- iii. 36 53° 20' S, 146° 44' E Good By min. temp. betw. Sts. 1690 and 1691. By ship's T.s near 1200 hr. lOI 'W. Scoresby' N'ward to Cape 29/30. iii. 36 ? — No adj. Sts. Indeterminate by ship's T.s. Town 102 'Discovery II* N'ward to Fremantle 2. iv. 36 48° 05' S, 109° 47' E Good By min. temp. N of St. 1730. By ship's T.s near 1730 hr. 103 )> S'ward in 0° 29. V. 36 49° 37' S, 00' 06' E Probable By min. temp, near St. 1777. By ship's T.s about 1400 hr. 104 )j N'ward to Cape Town 12. vi. 36 48° 40' S, 18" 50' E Approx. By min. temp. S of St. 1798. By ship's T.s about 1200 hr. 105 )) S'ward in 0" 26. ix. 36 49° 54' S. 00^ 06' E Uncertain Bv min. temp. betw. Sts. 1810 and 1812. By thermograph about 1200 hr. 106 >» NW of S. Georgia 5- xi. 36 5i°io'S, 41° 40' W Good By min. temp, near St. 1857. By thermo- graph at 0400 hr. 107 ») It yy »» 6. xi. 36 52° 04' S, 43' 00' W Approx. By min. temp. N of St. i860. By thermo- graph about 1600 hr. 108 )» S'ward in 43° W 7- xi. 36 54° 10' S, 42° 50' w j> Surf. temp, only at St. By thermograph rise of temp, to 3-0° centred about 1200 hr. indicates proximity of loop of convergence. 109 )) Elephant I. to Falkland Is. IS- xi. 36 57° 08' S, 55°i5'W V. good By min. temp. N of St. 1877. By thermo- graph at 0300 hr. no 'W. Scoresby* S'ward from Cape Town I. xii. 36 47° 58' S, i3°03'E Good No adj. Sts. By ship's T.s about 1800 hr. III 'Discovery IF E'ward to S. Georgia 3- xii. 36 53°35'S, 48^ 30' W Uncertain By min. temp. E of St. 1916. By thermo- graph about 0300 hr. 112 i> S. Orkneys to Falkland Is. 17- ii-37 56°45'S, 5i°4o'W V. good By min. temp. S of St. 1969. By thermo- graph at 1600 hr. "3 )j Falkland Is. to S. Georgia 2. iii- 37 53°25'S, 45° 3°' W Uncertain By min. temp, near St. 1975. By thermo- graph about 0230 hr. 114 'W. Scoresby' N'ward to Cape Town 19. iii- 37 48° 20' S, ii°25'E n No adj. Sts. By ship's T.s perhaps at 2200 hr. "5 'Discovery II' N'ward in 0° 28. iii. 37 49° 5°' S, 00° 20' E Approx. By min. temp. betw. Sts. 2022 and 2023. By thermograph probably at 2300 hr. 2IO DISCOVERY REPORTS Table 9 (cont.) No. Ship Voyage Date Estimated position Degree of accuracy Remarks 116 'W. Scoresby' 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 'Discovery II' 'W. Scoresby' ' Discovery II ' Whale marking betw. Montevideo and S. Georgia SE'ward from Cape Town NE'ward to Fremantle S'ward from Fremantle N'ward to N. Zealand SE'ward from N. Zealand S. Shetlands to Falkland Is. NE'ward to Falkland Is. Falkland Is. to S. Orkneys N'ward to Cape Town S'ward in o N'ward in 20' E S'ward in o^ N'ward in 20^ E S'ward in o" N'ward in 20^ E S'ward in 0° N'ward in 20' E S'ward in o" N'ward in 20^ E S'ward in o N'ward in 20^ E S'ward in o N'ward in 20^ E 27- X. 37 47° 32' S, 33° 12' W 28. X. 37 ] 49° 00' S, 30° 32' W 30. X. 37 j 50° 04' S, 32° 10' W 5. xi. 37 j 48° 20' S, 33° 12' W 19. .XI. 37 10. xu. 37 5- i- 38 5- i- 38 6. i. 38 24. i. 38 14. ii. 38 24. ii. 38 47" 05' S, 2S° 00' E 52° 12' S, 98° 25' E 49° 47' S, 1 15° 47' E 50° go's, 115° 47' E 51° 12' S, 115° 50' E 59° 30' S, 165° 15' E 59' 52' S, 168^ 45' W 56°47'S, 58"oo'W Probable )? Good Probable Uncertain Probable 13. iii. 38 57° 57' S, 62" 15' W 22. iii. 38 29. iv. 38 29. iv. 38 30. iv. 38 10. vii. 38 23. vii. 38 14. viii. 38 27. viii. 38 23. ix. 38 3- X. 38 26. X. 38 5. xi. 38 2. xii. 38 14. xii. 38 16. i. 39 3- "• 39 23- ii- 39 15. iii. 39 56°3o'S, 52°i5'W 49° 45' S, 47° 50' S, 46° 28' S, 50° 12' S, 48° 20' S, 49° 28' S, 48° 30' S, 50° 00' S, 48° 55' s, 50° 06' S, 48" 10' S, 49° 45' S, 47° 55' S, 49° 45' S, 48" 20' S, 49° 33' S, 48° 20' S, 19° 22' E 19° 20' E 19° 18' E 00° 32' E 21° 00' E 00° 30' E 20' 30' E 00^"' 40' E 20° 00' E 02° 05' E 20" 15' E 01° 03' E 19° 55" E 02° 48' E 19° 45' E oi°o6'E i9°38'E Uncertain Probable V. good Probable Good Approx. )» Good »j Uncertain Good Uncertain V. uncer- tain Uncertain Good Probable V. uncer- tain Probable Approx. Good Complex movements by ship and v. sharp fluctuations of temp, by thermograph. Loop of convergence approached at 0015 hr. on 27th. Crossings at 0600 hr. on 28th, 1300 hr. on 30th and 0500 hr. on 5th. Loops possibly shifted betw. 27. X. and 5. xi. 37. By min. temp. N of St. 2088. By thermo- graph at 0100 hr., S of St. Latter position adopted. Obscure by min. temp. By thermograph probably at 11 00 hr. By min. temp, near St. 2158. By thermo- graph at 1500 and 1700 hr. on 5th and 0800 hr. on 6th. Track close to con- vergence from 1700 to 0800 hr. By min. temp, near St. 2206. Thermo- graph shows rise in temp, before and after St. 2206. By min. temp. betw. Sts. 2220 and 2221. By thermograph at 1330 hr. No adj. Sts. By thermograph at 1000 hr. By min. temp. betw. Sts. 2288 and 2289. By thermograph at 0200 hr. Probably oblique crossing. By min. temp. betw. Sts. 2292 and 2293. By thermograph at 0900 hr. By min. temp. betw. Sts. 2347 and 2348. By thermograph at 0300 and 1900 hr. on 29th and 0800 hr. on 30th. By min. temp. betw. Sts. 2358 and 2359. By thermograph at 1330 hr. By min. temp. betw. Sts. 2377 and 2378. By thermograph at 0630 hr. By min. temp. S of St. 2387. By thermo- graph probably at 1500 hr. By min. temp. betw. Sts. 2416 and 2417. By thermograph at 1400 hr. By min. temp. betw. Sts. 2426 and 2428. By thermograph at 0100 hr. (doubtful). By min. temp. betw. Sts. 2449 and 2450. By thermograph at 1300 hr. (doubtful). By min. temp. N of St. 2461. By thermo- graph S of St. at 0300 hr. By min. temp. betw. Sts. 2481 and 2482. By thermograph at 0600 hr. By min. temp. betw. Sts. 2495 and 2496. By thermograph at 1600 hr. By min. temp. betw. Sts. 2521 and 2522. By thermograph at 0130 hr. By min. temp. S of St. 2533. By thermo- graph possibly at 0030 hr. By min. temp. betw. Sts. 2573 and 2574. By thermograph at 1030 hr.- By min. temp, near St. 2585. By thermo- graph indeterminate. By min. temp. betw. Sts. 2622 and 2623. By thermograph at 1400 hr. 211 NOTES ON THE PLATES Place-names are shown in Plate I. Plates II-XII show the mean surface isotherms for each month in Antarctic waters and for a short distance north of the Antarctic convergence. The pack-ice edge, shown by a hatched line, is in the mean position for each month, and is reproduced from Mackintosh and Herdman (1940) without alteration except at one point in Plate III between 10° and 100° W (see text, p. 199). The positions of thermograph records, and single observations included in the Discovery Committee's data, are shown in yellow; pecked lines in yellow represent ship's routine temperature readings. DISCOVERY REPORTS, VOL. XXIII PLATE I GO. ,,,,.,'.,,/>•:./> . J ,'!'■• ,'r, I fr^'i-l ri I irfc •■ ' i ','.■ I ■ 60 30 120 150° V/ 180°E Estimated mean position of the Antarctic convergence. The mean northern limit of the pack-ice is the average Lstimated mean po^s^i^^^ ice-edge for September reproduced from Mackintosh and Herdman (1940)- DISCOVERY REPORTS, VOL. XXIII PLATE II Surface temperatures, SEPTEMBER and OCTOBER. DISCOVERY REPORTS, VOL. XXIII PLATE III Surface temperatures, NOVEMBER. DISCOVERY REPORTS, VOL. XXIII PLATE IV oO 120 120 150° W I80°E 150° Surface temperatures, DECEMBER. DISCOVERY REPORTS, VOL. XXIII PLATE V 120 150° W 180-E 150° Surface temperatures, JANUARY. DISCOVERY REPORTS, VOL. XXIII PLATE VI 30° 60 120 ISO- W 180"E ISO- Surface temperatures, FEBRUARY. DISCOVERY REPORTS, VOL. XXIII PLATE VII Surface temperatures, MARCH. DISCOVERY REPORTS, VOL. XXIII PLATE VIII Surface temperatures, APRIL. DISCOVERY REPORTS, VOL. XXIII PLATE IX Surface temperatures, MAY. DISCOVERY REPORTS, VOL. XXIII PLATE X Gtf 150° W 180"E 130° Surface temperatures, JUNE. DISCOVERY REPORTS, VOL. XXIII PLATE XI GO; 90 120 iso° W 180°E 150° 60 30 120° Surface temperatures, JULY. DISCOVERY REPORTS, VOL. XXIII PLATE XII Surface temperatures, AUGUST. DISCOVERY REPORTS, VOL. XXIII PLATE XIII 0° ISOTHERM, summer months. DISCOVERY REPORTS, VOL. XXIII PLATE XIV 30° go: 90 120 ...,..■.,..>"■/-"■-■■. /.,.mmm/u riwriTi I I'rrrrI' r .■Ml.f.. I m\m I'l I < ' IMM'l'V \'y'\'\'\' ^■\' \'"> 150° W 180-E ISO" 0° ISOTHERM, winter months. [Discovery Reports. Vol. XXIII, pp. 213-222, Plate XV, August 1946.] NEBALIOPSIS TYPICA By H. GRAHAM CANNON, Sc.D., F.R.S. NEBALIOPSIS TYPICA By H. Graham Cannon, Sc.D., f.r.s. Beyer Professor of Zoology in the Victoria University of Manchester (Plate XV) THE FORM OF THE BODY IN a recent report which has just reached this country, Linder (1943) has called in doubt the shape and configuration of Nebaliopsis typica which I described in an earlier Discovery Report (193 1) and illustrated by photographs of a specimen which I stated (loc. cit. p. 201) was complete and undamaged. The only specimen which, previous to my report, had reached the surface intact, I stated on the authority of Dr Odhner of the Riksmuseum, Stockholm, was most probably lost. Dr Linder has now found the missing specimen, and it is this that he maintains represents the normal appearance of this rare deep-sea crustacean. My specimen was labelled F 2, and was illustrated by three untouched photo- graphs on plate xxxii, and text-figs, i and 2, based on these photographs. Linder's chief criticism is that the cephalothorax is distended, more especially in the posterior part, while the carapace has remained unaffected (loc. cit. p. 5). As a result, part of the thorax is not covered by the lateral carapace, and the length of the carapace relative to the rest of the body is abnormally small. The difference is very marked by measuring Linder's figure and my own. A comparison shows that the length of the cephalo- thorax in his specimen is about 54 % that of the carapace, while in my specimen it is 96 %. Obviously such a large difference calls for further investigation. I pointed out (loc. cit. p. 202) that measurements of all the specimens I had, together with that of the specimen we may now call Linder's, established the existence of a considerable variation in the length of the carapace and left the matter at that. I took the view that there was no doubt about it as, while Linder's specimen and mine were both presumably as perfect as could be, his had a large carapace and mine had a small one. I was relying for details of Linder's specimen on a sketch which had been published by Ohlin (1904, fig. i). From this, however unsatisfactory it might have been in other respects, there was no doubt about the length of the carapace— it was relatively long. The question to be settled is whether the relative shortness of the carapace of my specimen F2 is abnormal. Linder's explanation (loc. cit. p. 6) is that at the great depths at which Nebaliopsis lives, there must be inside the body an enormous pressure. This is a point which no one will dispute. However, he then says that as the specimen is brought to the surface in the collecting gear this pressure inside, acting in all directions, enlarges the body so that by the time atmospheric pressure is reached at the surface the body is completely distended. This distension will occur to different degrees in different parts, and Linder assumes that the carapace is not affected by the pressure. Hence it retains its normal size while the body becomes bloated, and thus the carapace appears abnormally short. Now this argument represents a widespread fallacy, a fallacy which has arisen first from the known fact that specimens of deep-sea fish are occasionally completely distended when they reach the surface, and secondly from the persistent and erroneous belief that sea water is much more dense at great depths than at the surface. Actually Linder's argument is valid, and then to a very uncertain degree, only if the body of Nebaliopsis contained gas. As long as the body contains no bubble of gas, no such distension as Linder describes can take place. Now there is no reason to suppose that the body of Nebaliopsis contains gas any more than one may expect to encounter gas on opening up a lobster. Its body can be looked upon very largely as a mass of aqueous liquid and, moreover, a liquid closely similar in its physical properties to sea water. Now sea 2j^ DISCOVERY REPORTS water, like other liquids, is only slightly compressible, or, more emphatically, is practically incom- pressible. The coefficient of compressibility of sea water is roughly 44 x io-«, which means that at the depth of a mile— the depth at which Nebaliopsis is known to occur— the density of sea water would only be 1/130 the greater. From this figure a rough computation shows that a specimen such as Nebaliopsis, if lowered to a depth of a mile, would diminish in length by approximately 1/400. Clearly, then, a specimen raised up from i mile would expand by 1/400 of its length, that is, by an imperceptible amount that would be extremely difficult to measure. Obviously such an expansion would lead to nothing like the distension of my specimen F2. I have stated that the presence of gas in a specimen, such, for instance, as occurs in the air bladder of a bony fish, would make a considerable difference to the above argument, for a gas, unlike a liquid, is extremely compressible. Now the pressure at a depth of 2000 m. is approximately 200 atm. Hence, if there were a bubble of gas in such a specimen at such a depth it would enlarge to 200 times its original volume by the time it was drawn up to the surface, or roughly a bubble the size of a pin's head at the depth of i^ miles would expand to the size of a pea at the surface. This is a considerable expansion and is of the order that might produce the enlargement which Linder maintains has occurred in the Discovery specimen F2. Is there any reason, therefore, to suppose that gas is produced in the body of Nebaliopsis as it is brought to the surface? It might be argued that as it passes upwards the resulting reduced pressure would cause the dissociation of the oxyhaemocyanin that is presumably present in its blood. But the solubility of a gas is proportional to its pressure, and hence oxygen at 2000 m. is 200 times as soluble at the surface. Hence any oxygen which might be set free from the dissociation of the oxyhaemocyanin would immediately go into simple solution. For the sake of argument, however, let us suppose that the dissociation took place so quickly that there was not sufficient time for the oxygen to become dissolved before the animal reached the surface. What would be the result? Most probably the animal would burst under the strain of the sudden expansion inside it and the gas would escape. But if it were able to stand the strain without bursting one thing that is certain is that the oxygen would remain in it as a bubble. Anyone who has handled preserved specimens of Crustacea in spirit knows the difficulty of getting rid of a bubble of air in a specimen once it has got in. In my specimen F 2 there was no bubble of air, as I think my photographs (loc. cit. plate xxxii) established without a doubt. As far as I have argued, therefore, I have shown that my specimen of Nebaliopsis could not have expanded as suggested by Linder merely by being relieved of the enormous pressure under which it lived, simply because liquids are almost incompressible, which is the same thing as saying unexpansible. Neither could it have become extended with gas, because there was in fact no gas in the specimen when it reached the surface. My chief argument, however, in refuting Dr Linder's suggestion is the obvious undistorted condition of specimen F2. Naturally, when I first received the specimen I was struck with its bloated appearance, and this caused me to wonder whether it could be abnormal. Directly I examined it ventrally and dorsally, however, I had no hesitation in deciding that it must be undistorted, for it showed the complete ventral chain, of nerve ganglia, and, more striking, the complete tubular heart. I am afraid I assumed that this perfection of the inner organs was so obviously a testimony to the condition of the animal that I did not comment on it in my paper ; I merely relied on the photographs. Now, surely it is practically impossible for the hind part of the cephalothorax to be enlarged without producing an obvious distortion of either the nerve chain or the heart. Certainly if the supposed expansion were caused by an expanding gas bubble (and I have shown that this, however remote, is the only possibility) the expansion would not be bilaterally symmetrical, for the bubble would of necessity lodge to one side of the gut. The final demonstration, however, that the Discovery specimen is normal comes from the words of Dr Linder himself. He states (loc. cit. p. 7) that 'a study of sections provides a certain proof. . . NEBALIOPSIS TYPICA 2i7 specimens that are only slightly swollen show the muscles of the thorax torn away from the integument '. Now, specimen F2 was sectioned down to the sagittal plain as I stated in my report (loc. cit. p. 204). The sections are quite normal and are typical of deep-sea material. There is no such distortion of the musculature as Linder predicts, or, in fact, of any other organ. The proof of this is that Miss Rowett (1943) has used the series of sections to work out with conspicuous success the anatomy of the gut. Miss Rowett 's work was published a few months after the appearance of Dr Linder's paper, and throws a completely new light on Nebaliopsis. Moreover, it supplies an obvious explanation of the differences between our two specimens. She showed that the large opaque mass in my specimen F2, which I tentatively suggested was the ovary (loc. cit. p. 203), was in fact an enormous sac-like diver- ticulum of the mid-gut. It was filled with a homogeneous coagulum ; that is, with a mass of food material in which there was no structure (Rowett, 1943, p. 15). From this, with admirable argument, she puts forward the view that Nebaliopsis is an egg-sucker. She then shows that, quite apart from the apparatus for fiker-feeding, this remarkable form is fully adapted in a variety of ways to this peculiar diet. In a paper that is now in the press she has gone further and has pointed out the extraordinary correspon- dence that occurs between the adaptation of Nebaliopsis and of the nudibranch Calma glaucoides which is known to feed exclusively on a diet of eggs. The adaptation which is of importance in the present discussion is the enormous mid-gut sac. Miss Rowett points out (loc. cit. p. 8) that this sac is 'without any convolution and with only a few septa arising from its walls'. Therefore while it does not provide much extra surface for its digestion it is admirably suited for a storage organ. Now eggs will certainly not always be present at the great depths at which Nebaliopsis lives. During the breeding periods of neighbouring animals there will be abundance, but in between whiles scarcity. However, it must be remembered that the neighbours of Nebaliopsis are few and far between, so that even when eggs are present they will be patchy in their distribution. A large storage organ is clearly an adaptation to this. It enables the animal to take a con- siderable quantity of food on the infrequent occasions when it happens to encounter a patch. My specimen F2 is clearly a specimen that had just had a meal and it became distended in the same natural way in which, for example, a blood-sucking tick becomes bloated after a meal. Another parallel is to be found in a deep-sea fish such as Chaismodiis niger (Murray and Hjort, 1912, p. 721, fig. 515) which has an abdomen so distensible that it can accommodate a larger specimen of the same species. The musculature of the body appears to be arranged so as to allow the body to expand in the hinder trunk region. The abdominal region is a packed mass of muscles. The cephalic region contains all the musculature of the antennae and mouth parts and the muscles extend ventrally in association with the trunk limbs. They become less marked posteriorly in relation to the simple condition of the eighth trunk limbs. The dorsal and ventral longitudinal muscles are practically non-existent in the thoracic region. There is thus a large region of the body, the posterior and dorsal thoracic region, which is almost devoid of muscles. The integument over this region is very thin and flexible and it is here that, as my photographs show, the expansion takes place. u • u ^ To summarize, the enlarged appearance of the Discovery specimen F2 is due to the fact that it had just taken a meal. It is not in any way an unnatural distension. A comparable specimen in which the mid-gut digestive sac is empty is shown in Plate XV. This beautiful example occurs in a second small Discovery collection. I consider this specimen much more perfect than Linder's, for it shows the rostrum, eyes, and antennae in a normal position for one of the Nebaliacea. . , ^ , . The photograph which Linder published (loc. cit. Taf. I, fig. i) agrees fairly closely with the sketch published by Ohlin (loc. cit. fig. i) ; enough, in fact, to make it fairly certain that the sketch was made after fixation and not while the animal was swimming round. How the animal was preserved we do not know There is nothing to indicate the use of any special fixative, and so most likely it was placed in 2,8 DISCOVERY REPORTS spirit or formalin in sea water. Also it was almost certainly moribund, when it was pickled, for as Linder (loc. cit. p. 7) points out after quoting Ohlin, it could only swim on its side, a quite unnatural position. Moreover, having obtained for the first time a living specimen of a deep-sea crustacean it is only natural that the scientist would keep it alive as long as possible. Whatever happened, it was during the fixation that the first damage to the specimen occurred. Now spirit and formalin are both slow fixatives and take a considerable time to kill an organism. I have even seen a specimen of the fairy shrimp Chirocephalus placed in the relatively fast fixative Bouin, and after one minute it was still wriggling. During this time the outermost musculature becomes fixed, while the inner muscles still contract. Naturally the result is a distorted specimen. It looks to me as though Linder's specimen, during fixation, became distorted in this way, for in the head region the rostrum, together with the antennules and antennae, appear to have been pulled inwards so as to become completely covered by the carapace. This does not occur in any other nebaliacean. Even in badly damaged specimens of Nebaliopsis, in which the body has become wrenched away from the carapace, the head region appears to remain intact with the antennae and rostrum projecting beyond the edge of the carapace. This is very clearly shown in one of Linder's own specimens (loc. cit. Taf. i , fig. 3), which shows just the same arrangement of these parts as in the photograph I am publishing in this paper as well as in that of my original F2 specimen. In addition to this fixation damage I think there must have been slow shrinkage of the whole body relative to the carapace after fixation, for in Ohlin 's sketch the hinder margin of the carapace reached only to the front end of the seventh abdominal segment, while in Linder's photograph taken years later it stretches to the hind end. Some deep-sea Crustacea are able to stand a journey up to the top and appear quite normal when caught ; thus Gigantocypris is often collected in numbers from great depths and will swim about actively in bowls of sea water. My own studies of this form (Cannon, 1943) have, I think, demonstrated that these forms are quite unaffected by the enormous reduction of pressure which they undergo in their passage upwards of maybe more than a mile. Nebaliopsis, on the other hand, is now known from a considerable number of specimens, and yet only three are anywhere near perfect. The greater number are completely disrupted. Now Nebaliopsis is a much more delicate form structurally than the robust Gigaiitocypris, but I feel certain that this does not explain the difference between them. Gigantocypris has a firm and substantial outer shell, while that of Nebaliopsis is extremely delicate. But, on the other hand, both have bodies constituted of cells containing living protoplasm and, quite irrespective of their exoskeletons, that protoplasm is in equilibrium under enormous pressure. When they are brought to the surface that pressure is relatively quickly diminished. Now, there can surely be no doubt that this change will produce an immediate disturbance of the equilibrium of the protoplasm. In the majority of cases it is probable that the protoplasm would be unable to readjust itself and so would be precipi- tated, with inevitable death resulting. This is what I consider happens in Nebaliopsis. On the other hand, if the protoplasm can so adjust itself as to remain in equilibrium, then the specimen will appear living and normal at the surface as in the case of Gigantocypris. If my argument is correct, it means that Nebaliopsis collected in deep-sea trawling is normally fixed (but not preserved) soon after leaving the depth at which it occurs. It is fixed, that is, its protoplasm is precipitated, by suddenly reduced pressure. Now this will act on all parts of the body quickly at the same time. Thus, there will be no question of penetration of fixative. The protoplasm of all cells of the body, whether they are deep or superficial makes no difference, will suddenly precipitate. There will be no distorting death struggles- one part of the body still alive pulling against another part already fixed. There will be a sort of shock suddenly immobilizing all parts of the body at the same instant, and the result will be a fixed prepara- tion as nearly like the living form as possible. If, now, such a specimen can finish the journey to the top NEBALIOPSIS TYPICA 219 and, more important, on to the boat, without being crushed, and further, if it is subsequently handled and preserved by an expert, such specimens as my F 2 and the one I illustrate in this paper will be obtained. THE FILTRATORY FEEDING MECHANISM In Nebaliopsis I described in my earlier report (1931, p. 210) a fihratory mechanism unique amongst the Crustacea, for it was based on the joint action of the maxilla and the first trunk limb. I gave evidence that this type of filter-feeding must have evolved from that of Nebalia, which I had previously shown (1927) resulted from the combined action of all the trunk limbs, the maxilla being minute and taking no part in the filtering process. I suggested (1931, p. 216) that along the lines leading to Nebaliopsis the ancestral form ' developed the maxillary-first trunk limb filter mechanism, at first, to aid its more posterior trunk limb filter. . . . Then, when this became sufficiently advanced, it opened up the carapace — maybe to allow a greater inflow of water on to the maxillary region. . . . The maxillary filter now became the chief feeding mechanism. . . . As it developed so the carapace widened out and water came to be sucked in from all directions. The trunk limb filter was then almost abandoned, the trunk limbs swinging forwards to act as a subsidiary mechanism supplying water to the . . . maxillary-first trunk limb filter.' Linder (loc. cit. p. 30) accepts my description of the new filter apparatus, but considers that in addition the trunk limbs together still form an efficient filtering mechanism. He even goes further (loc. cit. p. 31) and suggests that the filtering power of Nebaliopsis is more effective than that of Nebalia, a point which I most strenuously deny. The filter process of Nebalia is, as far as I can judge, the most efiicient of all those crustacean feeding mechanisms that I have studied, for water carrying suspended food particles is sucked into a filter chamber just as if it were being sucked into a cylinder by a movable piston, and after being filtered is passed to the exterior through a valve as efficient as a rubber gas valve. The filter chamber of Nebalia into which the water is sucked is the median ventral space between the trunk limbs. The entrance lies anteriorly in the mouth region. Posteriorly the chamber is closed by the eighth trunk limbs uniting medially to form a wall. Laterally the trunk limbs form its walls, the spaces between the limbs being spanned by continuous sheets of filter setae. Dorsally it is roofed by the mid- ventral body wall, while ventrally there is a complete and thick floor formed by the endopodites of the trunk limbs which recurve sharply backwards and slightly inwards so that their tops touch in the middle line. Thus the filter chamber is a simple laterally compressed space with but one entrance, and that is relatively small. It is a slit extending from the lower edge of the labrum to the 'elbow' of the first trunk limb. Now the first trunk limb is itself small — it is only about two-thirds the length of the middle trunk limbs which are the main limbs acting as pistons sucking water into the filter chamber. Thus we have a relatively large filter chamber with a small opening into it. Clearly suction will be very powerful, so that once water has been sucked in it will not escape out again through the same opening — it must remain to be filtered. Obviously then any enlargement of this entrance will lead to a diminution in the force with which the water is sucked in. In Nebaliopsis the whole floor of the filter chamber has opened up by the disappearance of the long recurved posteriorly projecting endopodites of Nebalia. Hence, even if the trunk limbs were acting as in the latter form, the suction into the filter chamber would be relatively weak. However, it is wrong to call this median space between the trunk limbs of Nebaliopsis the filter chamber. Apart from the fact that I showed that another filter chamber had developed between the maxilla and the first trunk limb, this space, now that it is completely open ventrally, is directly comparable with the mid-ventral space of a branchiopod. Thus, again supposing that the trunk limbs were still filters, as they are in Nebalia, the efficiency of their combined efforts in 220 DISCOVERY REPORTS Nebaliopsis would be comparable with that of the feeding mechanism of the fairy shrimp Chirocephalus. Now this type of filtratory mechanism is not nearly so powerful as that of Nebolia. A floating particle, once it has been sucked in between the limbs of Nebalia, cannot, as far as I can see, escape unless it is forcibly ejected. On the other hand, a particle, on being sucked into the mid-ventral space of Chiro- cephalus, as often as not is blown out again, and it would be the same with Nebaliopsis. However, this is all on the supposition that the trunk limbs of Nebaliopsis are still efficient filters. But I pointed out (1931, p. 21 1) that this cannot be so because, among other things, their lateral parts do not form a valvular system as they do, in fact, in all other Nebaliacea. In Nebalia and Paranebalia it is the epipodites together with the exopodites, and in Nebaliella, where the epipodites are absent, the exopodites alone, which project backwards and completely and accurately span the gaps between successive limbs, thus preventing any lateral entry of water into the inter-limb spaces during the suction phase. In Nebaliopsis, on the contrary, there are wide lateral gaps both proximal and distal to the epipodite through which water could pass unhindered. Linder (loc. cit. p. 30) denies this and says that the gaps are simply due to the distension of the limbs. He implies, although he does not say so in as many words, that the main axis of the limb has elongated under pressure, while the epipodite has remained the same size, and hence a gap has occurred between the epipodite and the body wall. The fallacy of this is obvious. There is no reason why the epipodite should not enlarge along with the rest of the limb. In fact from the anatomy of the limb it is fairly certain that if this hypothetical swelling did take place it would be the epipodite which would enlarge most and not the main axis of the limb, for the latter is skeletally relatively rigid, while the epipodite is extremely delicate. One further argument : if the lateral gaps which are so clearly seen in the upper photograph of my plate xxxii (193 1) are due to artefacts, how is it that the lateral space between the first and second trunk limbs is so accurately covered by the epipodite of the first trunk limb (loc. cit. p. 206, fig. 3)? Is this also due to distension? Surely not — it is simply a manifestation of suction between the first and second trunk limbs which is absent, or practically so, between the other limbs. Having explained away the lateral gaps between the trunk limbs, Linder still has to deal with the distal gaps that occur between the tips of the limbs. These gaps result from the absence of the pos- teriorly curved endopodites. The typical trunk limbs ' are unsegmented, the exopodites being repre- sented by a slight protuberance. . . .The endopodite must be considered as the tip of the limb distal to this exopodite lobe' (Cannon, 193 1, p. 207). They are, as Linder correctly states (loc. cit. p. 10), turgor limbs, that is, limbs that depend for their rigidity on an internal blood pressure. Moreover, ' the distal two-thirds of each limb is devoid of musculature ' (Cannon, loc. cit. p. 210), so that the limb cannot be bent by internal muscles — it can only be moved as a whole by the muscles at its base after being made rigid by being pumped full of blood. Now Linder explains the closure of the gaps between the tips of the limbs during the 'Abduktions- phase ' (suction phase) as due to a bending over. of one limb on to the limb in front, as a result of suction between the limbs (loc. cit. p. 30). This would imply that directly any two successive limbs commence their suction stroke, the suction between them becomes so great that it causes the hind limb to bend in the middle so that its tip comes down on the limb in front and closes the distal gap. But what causes this original powerful suction? Unless there are efficient valves, the suction between the limbs will be negligible. The gap is there to start with and hence the small suction caused by the inter-limb space enlarging will simply draw in water through the gap which will at once neutralise the suction. Linder says that similar conditions obtain in the Anostraca (loc. cit. p. 30) and he sees no reason why the same thing should not happen in Nebaliopsis. But things are quite different in the Anostraca. There the limb is jointed and the distal endopodite ffap is provided with a complete set of muscles by which it can be pulled down and held, if necessary, on to the next limb. NEBALIOPSIS TYPICA 221 Linder goes further and, as I have mentioned above, states (loc. cit. p. 31) that the filtering mechanism of Nebaliopsls is an even more efficient system than that of Nebalia. He points out that it differs in two respects — first that the eighth trunk Hmb is not fihratory, and secondly that the maxilla, together with the first and second trunk limbs, form a separate filter system. It can only be deduced from this that Linder considers that the Nebaliopsis mechanism evolved directly from that of Nebalia, a fact with which all will agree, and further, that it only shows the two differences which he enumerates. There is, however, another fundamental difference which he has overlooked. In Nebalia the trunk limbs curve backwards so that one limb overlaps the limb behind. In Nebaliopsis, according to Linder's hypothesis, the limbs bend forwards so that one limb overlaps the limb in front. This being so it is clear that Linder's Nebaliopsis mechanism could not have evolved from that of Nebalia, for such an evolutionary process would involve an intermediate stage when the limbs curved neither backwards nor forwards and so would be unable to carry out any filtration. If, therefore, the trunk limbs of Nebaliopsis do filter according to Linder's mechanism, then this would be an entirely new development and this is not what Linder has been arguing. There would be the same difficulty here as occurs when it is attempted to derive the malacostracan maxillary filter from the anostracan trunk limb filter (Cannon, 1928, p. 820). Both these are based on typical phyllopodia, but in the former the limb is concave anteriorly and in the latter posteriorly. But what is more important, the functional activities of the limbs are dependent absolutely on these arrangements of the limbs. The same applies to Nebaliopsis and Nebalia. The activities of their trunk limbs, whether actually filtratory or only supposedly so, depend on their arrangement and since they are arranged as mirror images, one pointing forwards and the other backwards, it is not possible to derive one mechanism from another, without postulating an intermediate stage during which the limbs could not function. BIBLIOGRAPHY Cannon, H. G., 1927. On the feeding mechanism 0/ Nebalia bipes. Trans. R. Soc. Edinburgh, lv, pp. 355-70. Cannon, H. G., 1928. On the feeding mechanism of the fairy shrimp, Chirocephalus diaphanus Privost. Trans. R. Soc. Edinburgh, LV, pp. 805-22. Cannon, H. G., 1931. Nebaliacea. Discovery Reports, Cambridge, hi, pp. 199-222, pi. xxxii. Cannon, H. G., 1943. On the anatomy of Gigantocypris muUeri. Discovery Reports, Cambridge, xix, pp. 185-244, pis. xxxix-xlii. Linder, Folke, 1943. t)ber Nebaliopsis typica G. O. Sars, nebst einigen allgemeinen Bemerkungen iiber die Leptostraken. Dana-Report, No. 25. Murray and Hjort, 191 2. The Depths of the Ocean. London. Ohlin, a., 1904. tjber eine neue hathypelagisch lebende Phyllocaride. Zool. Anz., Leipzig, xxvii, pp. 59-61. RowETT, Helen G. Q., 1943. The gut of Nebaliacea. Discovery Reports, Cambridge, xxiii, pp. 1-18. PLATE XV Nebaliopsis typica x 9. St. 1636. Date 30. xi-i. xii. 35. Net N 100 B. Depth 380-150 m. DISCOVERY REPORTS, VOL. XXIII PLATE XV NEBALIOPSIS TYPICA [Discovery Reports. Vol. XXIII, pp. 223-408, Plate XVI, December 1946] REPORT ON TRAWLING SURVEYS ON THE PATAGONIAN CONTINENTAL SHELF Compiled mainly from manuscripts left by the late E. R. Gunther, M.A. By T. JOHN HART, D.Sc. CONTENTS Foreword, by N. A. Mackintosh page 226 Introduction 227 Field methods and preliminary observations 228 The first survey 230 The second survey 234 The third survey 236 Topography of the shelf 238 Hydrology 242 Plankton 246 Methods of presentation 250 General account of the fish fauna 251 Distribution and general notes on the species . ..... 259 Petromyzonidae ........... 259 Myxinidae 259 Lamnidae 259 Scyliorhinidae ........... 260 Squalidae 260 Squatinidae ............ 260 Torpedinidac 260 Rajidae ............. 260 Chimaeridae ............ 272 Summary of observations on Elasmobranchii ...... 274 Clupeidae ............ 275 Galaxiidae ............ 278 Aplochitonidae ........... 279 Syngnathidae 279 Macruridae 279 Merlucciidae . 280 Merluccius hubbsi . 280 Introduction: economic importance of allied species . . . 280 The sizes of Patagonian and European hake compared . . . 284 The distribution and relative abundance of Merluccius hubbsi within the area surveyed, and the effect of latitude on numbers, size and sex ratio . . : 289 The relation between length and weight of Merluccius hubbsi, and its value as an indicator of the spawning season, and for other purposes 291 Migrations ^00 Conclusions on migration 106 The food and feeding of Merluccius hubbsi 308 Parasites ,jq Macruronus magellanicus ' . .312 Comparison and contrast of the main features in the bionomics of Merluccius hubbsi and Macruronus magellanicus . . . . ^20 Gadidae ^20 Micromezistius australis ^20 Salilota australis ~~. Physiculus marginatus ^24 Muraenolepidae Carangidae Bovichthyidae ^ CONTENTS Nototheniidae P^S^ 32? Notothenia ramsayi 33 Harpagiferidae . ....••■•■• 34 Chaenichthyidae 34' Summary of observations on Patagonian Nototheniiformes . . -342 Gempylidae ....•••••••" 347 Thyrsiies atun 347 Scombridae 35 Zoarcidae 35 Summary of observations on Zoarcidae 353 Lycodapodidae . . • • ■ ■ • ■ • ■ '354 Ophidiidae 355 Genypterus blacodes 355 Brotulidae 35^ Centrolophidae 35 Stromateidae 359 Stromateus maculatus 359 Atherinidae 374 Scorpaenidae 374 Congiopodidae ...•••••••" 375 Psychrolutidae 37 Agonidae 37 Liparidae •' ' ' Bothidae 377 Thysanopsetta naresi . ■ ■• 377 Other species 3 Features of general biological interest 3 2 Prospects of commercial development 3 7 The weight of catches 3 7 Conclusions 39 References 39 Appendix I. Particulars of trawling stations 39^ Appendix II a. Hake data: first survey 4°° Appendix IIb. Hake data: second survey 4oi Appendix lie. Hake data: third survey 402 Appendix III. Approximate positions of localities mentioned in the text but not charted in text-figures '^° pjgjg ' following page 408 FOREWORD By N. a. Mackintosh The position, extent, and physical features of the Patagonian Continental Shelf are such as to prompt comparison with the major fishing grounds of the northern hemisphere ; but prior to the trawling surveys of the R.R.S. 'William Scoresby ' the nature and magnitude of the population of demersal fishes on the shelf had not been explored. Three surveys were carried out, for the most part at different times of the year, in 1927, 1928 and 1931-2; and the principal results are set forth in Dr Hart's report. The report is based to a considerable extent on the unfinished work of the late Mr E. R. Gunther, and the circumstances in which it was prepared require a word of explanation. Several members of the Committee's staff took part in the work at sea, but the largest part was played by Mr Gunther, and he was in charge during the third and most extensive survey. It is very largely owing to his untiring energy and enthusiasm, and his broad conception of the problems involved, that a most comprehensive and thorough investigation was carried out. The surveys have resulted in a very great mass of data, and the preparation of a general report on the results was undertaken by Mr Gunther. Some delay in the completion of this report was inevitable. A taxonomic account of the fish fauna, and a description of the marine deposits of the shelf, were needed first, and these have already been published in the Discovery Reports, together with systematic papers on certain invertebrate groups. In the years before the war, however, Mr Gunther had made good progress with the general analysis of the material, though his work was unavoidably interrupted from time to time by other responsibilities. Since he held a commission in the Territorials he was called upon for military service just before the outbreak of war, and further progress was therefore suspended. His death on active service in 1940 was a severe loss to the Discovery Investigations. Owing to the dispersal of the Committee's staff in wartime no more could be done until 1943, when Dr Hart was able to take over the work. The task of picking up the threads and collating the data naturally involved considerable difficulties; and for many aspects of the subject it was necessary to start again from the original data. The substance of the report can be regarded as the combined work of Mr Gunther and Dr Hart, but the latter, as he explains below, has written the entire text in its final form. It will be realized that this report deals mainly with the general biology and ecology of the demersal fish, and with the prospects of commercial trawling. The surveys were planned for this purpose and did not include an investigation of the pelagic fish such as the Falkland herring. Various references to these fish are included in the report, but there is still little information on the prospects of com- mercial fishing by other means than trawling. The principal conclusion of the report is that hake, and some other edible species, are obtainable in moderate numbers by trawling. Although the shelf has been found to be less rich in trawlable fish than might have been expected, it is possible that enough could be taken to support an industry if markets could be found, and problems of preservation and delivery could be overcome. The report may be regarded as a contribution to our knowledge of the fish faunas of the world, and it is hoped that it will be of assistance in any consideration of the future economic development of the Falkland Islands. REPORT ON TRAWLING SURVEYS ON THE PATAGONIAN CONTINENTAL SHELF Compiled mainly from manuscripts left by the late E. R. Gunther, M.A. By T. John Hart, D.Sc. (Text-figs. 1-53, Plate XVI) INTRODUCTION THE Patagonian Continental Shelf extends from the River Plate in the north to Staten Island in the south, and from the South American coast to an average distance of some 250 miles offshore to the eastward.' Outlying areas with depths of less than 200 m. extend round the Falkland Islands and on the Burdwood Bank to the south. Beyond the 200 m. contour, which may be taken as the edge of the shelf, the descent to oceanic depths is more or less abrupt. Faunistic writers referring to the Patagonian region commonly include the Magellan Channels and the coast of southern Chile m their geographical unit, but there is no need to qualify our definition of the Patagonian Continental Shelf so as to exclude the west coast, for there the descent to oceanic depths is so immediate that a shelf (in the accepted meaning of the term) can scarcely be said to exist. The investigations to be described here covered the whole of the shelf south of lat. 42° S., an area of some 150,000 square miles which is larger than the entire North Sea. Except for descriptions of small collections made in coastal waters the marine fauna of the region was almost unknown when the Discovery Investigations began, although it constitutes the largest expanse of shoal water (accessible to trawling) in the 'cold tem- perate' or sub-Antarctic Zone of the southern hemisphere. The need for a fisheries survey of the shelf, to gain information on the prospects of developing a commercial fishery from the Falkland Islands, was recognized from the outset of the Discovery Committee's work (Kemp, in Kemp, Hardy and Mackintosh, 1929, P- 148)- The greater urgency of problems relating to whaling and sealing in the more southerly (Antarctic) waters of the Falkland Islands Dependencies limited the scope of the trawling surveys however, and combined, with Mr E. R Gunther's untimely death in 1940, to prevent publication of results until now. Vast collections of benthos with lesser but probably representative collections of plankton and hydrological data, were obtained Gunther had hoped to use the information gained from these, as they were worked up by various specialists, in presenting the ecological study of the fish fauna in much greater detail than can now be attempted. Continued work on many of the groups may not be possible for years, but by great good fortune the taxonomic revision of the fishes had been completed by the late Mr J. R. Norman in IQ.7 As the need for more knowledge of the bionomics of the fish fauna became urgent, Dr Mackintosh asked me to prepare this report, working from Gunther's manuscripts. I found this unusually difficult because for the first time in my experience it involved work upon data which I had not helped to collect Moreover, Gunther had planned the production of five separate papers, and the manuscripts were in widely different stages of incompleteness. As a single report was called for I have entirely re-written the text myself, retaining Gunther's leading ideas and indicating our indebtedness to him so far as I am able. I found it necessary to recalculate all numerical data, using the origmal log books, except where the reasons for alterations made by Gunther himself could be traced. Any mistakes in this part of the work are my own responsibility. Both Mr Gunther and I have gained much from discussion of hydrological results with Dr G. E. R. Deacon, F.R.S., and the brief notes on the hydrology of the region presented here owe much to him. 22g DISCOVERY REPORTS My own experience of trawling in the southern hemisphere has been hmited to a few experimental hauls off the Falkland Islands and off the south-west coast of Africa, but I have seen enough to recognize the tremendous amount of hard work in the field that these trawling surveys must have entailed. This is especially true of the last and most comprehensive of the surveys, when Mr Gunther was in charge of the scientific work and had as his assistant Mr (now Comdr) G. W. Rayner. I think this survey was one of the most arduous pieces of field-work ever completed by the Discovery Investigations ; and, as my two colleagues were always anxious to point out, its success was largely due to the able and willing co-operation of the Captain, net-man and ship's company. The work of writing up the report has been carried out in the Laboratory of the Marine Biological Association at Plymouth, by courtesy of the Director, the late Dr S. W. Kemp, F.R.S., who, while Director of Research to the Discovery Investigations, planned much of the work here described. Dr Kemp's personal kindness and encouragement have helped me throughout my working life, and I am sure all members of the Discovery scientific staff would wish to say the same. I have gained much from the advice and encouragement of the staff of the Laboratory and more especially from hints on the handling of numerical data by Mr E. Ford and Mr G. M. Spooner. Had these two gentlemen not been away on war service during most of the period my task would have been lighter. The rapidity with which the librarian, Miss M. Sexton, procured obscure references under all the difficulties of wartime conditions was a great help. I have gained much from an all-too-brief interview with Mr C. F. Hickling of the Ministry of Agriculture and Fisheries (now Fisheries Advisor to the Colonial Office), who also helped with the loan of some of his important papers on Hake. This brings me to a point that needs emphasis if the work described here is to be justly appreciated — it was planned and carried out before the results of the last decade of fishery research, prior to the war, were known. If, for example, the final results of Hickling's prolonged work on European Hake had been available there is no doubt that our programme could have been modified with advantage, but our data were collected before Hickling's work was complete. FIELD METHODS AND PRELIMINARY OBSERVATIONS The three trawling surveys were carried out by the R.R.S. 'William Scoresby' in autumn (March- April) 1927, winter (June-July) 1928, and throughout the whole of the warmer half of the year 1931-2 (October-April). A few additional observations were made within the area by the R.R.S. ' Discovery' and by the R.R.S. ' Discovery II '. A description of the ship and of the gear will be found in Kemp, Hardy and Mackintosh (1929). In the more detailed parts of this report abbreviated descriptions of the gear have been used, as standardized throughout the station lists in Discovery Reports. Meanings of the abbreviations relevant to the present work are repeated here for the convenience of the reader. For this study of the bionomics of the fish fauna, the gear may be grouped under two main headings : ' Trawl -|- accessory nets' and 'Other gear'. ' Trawl + accessory nets ' comprises : OTC Commercial otter trawl, 80 ft. headline, 3 in. cod-end mesh. N7-T Net 7 mm. meshl Fine-meshed nets attached to the back of N4-T Net 4 mm. mesh ,- the trawl as described in the work men- NCS-T Coarse silk net J tioned above. ' Other gear' includes a motley collection of apparatus that helped to extend our knowledge of the distribution of the fish fauna in lesser ways. (The finer plankton nets, which scarcely ever catch fish except in their larval stages, and hydrological apparatus are not considered here.) INTRODUCTION 229 OTL Large otter trawl, 40 ft. headline, 1 1 in. cod-end mesh. BTS Small beam trawl. NRL Large rectangular net. BNR Russell's bottom net. DC Conical dredge. N450H Large plankton net 4I m. diameter at mouth. TYF Young fish trawl, stramin, 2 m. diameter ring^ Fished either horizontally N looH (or B) Plankton net i m. diameter at mouth / (H) or obliquely (B). NH Hand net. LH Hand line. 65" •i O O c^.«& ■ WSI09 'WS79 ""^3° •WS76_./VS75 • WS73 .ws72 WS92 wsea • WS9l' VVS7I 5d •WSS4 WS87 • ^i<^/ •■•••■■ 200M-" ■1^ h"'-^'f f- 65° 55 60° Fig. I a. First trawling survey, station positions in March and April 1927. This report is primarily based upon the hauls of the trawl + accessory nets, and full details of date, time, position etc. of all such hauls (apart from total failures) are tabulated in Appendix I. Similar details regarding the working of all the other gear will be found in our Station Lists when completed. For the present purpose reference to other gear has only been made where it directly affects our knowledge of fish distribution. 230 DISCOVERY REPORTS THE FIRST SURVEY The first survey was aimed at investigating the area immediately surrounding the Falkland Islands, with some observations on the shelf between them and the mainland, and on the Burdwood Bank. Dr N. A. Mackintosh, now director of research to the Discovery Investigations, was in charge of the scientific work. With him were Mr E. R. Gunther specializing on the fishes, and Mr D. Dilwyn John +5 •WSEIB WS237 O ■ wsz3a W5 WS2I4. ZI5 ai6 • VVS2Z5 '1Z33 210 'WS (230 '1231 ■.VV5229 '. •WS228 •WS243 .WS239 5rf ■•WS246^ • WS227 WS250 ' WS245 ..I • WS248 55 65° Fig. \h. Second trawling survey; station positions in June and July 1928. working on the invertebrates and the food and parasites of the fish. The normal routine at the trawling stations was: first, a sounding; then a haul with the conical dredge; then, unless the bottom had proved impossibly rough, an hour's haul with the ' trawl + accessory nets'. The trawling was followed by a second sounding, and the collection of water samples from surface and bottom for the deter- mination of temperature and salinity. A line of full hydrological and plankton stations between the Jason Islands and Port Desire was also worked, so that with incidental stations elsewhere and minor digressions to land the fur seal guard on Elephant Jason, and to mend nets, much work was accom- plished in the period of just over two months devoted to the survey. INTRODUCTION 231 The species of fish were counted, and the important ones measured and sexed, and data on the stages of maturity at different lengths collected. Occasionally observations on scales, gonads, and pathological growths, and sexing of the less important species were possible. Provisional identifi- cations proved adequate for the subsequent accurate determination of almost all the specimens taken, in spite of the facts that it was only possible to preserve a small proportion of the catch, and that some Fig. ic. Third trawling survey; station positions October to April 193 1-2. Nototheniidae and Rajidae presented special difficulties which were only cleared up in the course of the preparation of Norman's report years afterwards. The invertebrates were more difficult owing to the great bulk of many of the catches. Quantities of previously preserved species were estimated and noted when possible, and all or a noted pro- portion of the others preserved; but the bulk of the catch and masses of broken sponges, coralline polyzoa, large Scyphomedusae, etc., frequently made it impossible to sort the catch adequately m the limited time available. 232 DISCOVERY REPORTS Some forty species of fish were taken, and it was reckoned that the most important were : Squalus lebruni Psafnmobatis spp. Rajidae (several spp.) Clupea fuegensis Macrurotius magellanicus Merluccius hiibhsi Salilota australis Cottoperca gobio Notothenia ramsayi N. gunt fieri Champsocephalus esox Stromateus maculatus^ The existence of overlapping characters in the ' tesselata group ' of Notothenia was noted in the field, also the profuse variation among the Rajidae of the locality, which necessitated a large collection Fig. id. Plankton and hydrological observations; station positions 193 1-2. for further revision. Five teleosts new to science were collected, and specimens of Cottoperca gobio and Dissostichus eleginoides much larger than any previously recorded were obtained. The most important result of the survey was to demonstrate that hake {Merluccius hiibbsi) were to be found in moderate quantity on the shelf to the north and west of the Falkland Islands. This fish was for long confused with Merluccius gayi, the west coast species, and was the object of several small ^ The names in this list are those determined after Norman's revision. INTRODUCTION 233 trawling ventures from the great ports of Montevideo and Buenos Aires. These all worked very much farther to the northward and close in to the land ; it is safe to say that no considerable haul of the species had been taken south of lat. 42° S before. On this autumnal survey the hake were distributed over the shelf to the north of a line from Cape San Sebastian to the northernmost of the Falkland Islands (Figs, i a, 2). The largest catches were obtained north-west of the Falkland Islands, but there was little evidence of any special concentration there. A moderately rich haul was obtained so far away as St. WS90 near Magellan Straits. Females were commoner than males, males were com- monest in the shallower water along with the smaller females, and the catches with the largest females were almost devoid of males. Notothenia spp. were the most widely distributed and most numerous fishes, but were obviously less important than hake owing to their small size. A very heavy catch of Notothenia was obtained at one station north-west of the Falkland Islands, and another on the Burdwood Bank. Macruronus magellanicus was most frequent at hauls made in the centre of the plain of the shelf. Its excellent edible qualities and freedom from superfluity of small bones were gratefully recorded. Stromateus macidatus was commonest near the mainland, but was also found at two ofltshore stations towards the end of the survey. The flesh of this species was described as resemblmg that of the herring, but not quite so good. It was remarked of both these species that though not very abundant they occurred along with hake and might serve to supplement catches of the latter. John's observations on the food of fishes showed that squids, Clupea spp., Thysanopsetta naresi, hyperid amphipods and small Euphausiidae formed the bulk of the hake food. Each predommated m different hauls to the exclusion of some or all of the others. The fact that hake snatch up food while in the trawl was noted. This habit is also common among European hake, a point that became well known with the publication of Hickling's work later. The stomach contents of small numbers of other less important species of fish were recorded, and John also collected large numbers of fish parasites. Invertebrates formed the bulk of the nine catches made within 20 miles of the Falkland Islands in depths of 80-130 m., where fish were scarce. Sponges, actinians, Alcyonaria, coralline Polyzoa, spider crabs echinoderms and ascidians were extraordinarily abundant. These invertebrate hauls were heaviest to the south-east, south and west of the Falkland Islands. Smaller invertebrates reached their greatest abundance where coralline Polyzoa or Alcyonaria predominated. This ground was too rough for profitable trawling, owing chiefly to the corallines, large catches of 'rubbish and few fish^ This dis- appointing negative result must be regarded as the second important point proved during the first survey. Invertebrates were somewhat less numerous between the Falkland Islands and the mainland, but some very heavy catches were obtained. In general the bulk of invertebrates decreased to the west- ward and was least near the Argentine coast. Of nektonic forms squids were taken at nineteen out of twenty-nine stations, sometimes in considerable numbers. They form an important part of the diet of the larger fishes, birds and seals, of the area. Large Scyphomedusae were taken at seventeen stations, and sometimes completely smothered the rest of the catch. .^, ^ ^. A broad tentative correlation between the nature of the bottom and possible fishing prospects was ventured upon at the close of the survey, and may be summarized as follows. On the shelf and to the north of the Falkland Islands a dark greenish brown sand predominated on fairly clean ground. This was the best ground for hake also and is therefore by far the most P^^^ WSsV a"mt east south and west of the Falkland Islands, on the Burdwood Bank and at St. WS 88, a similar sand mixed with a high proportion of shell fragments occurred. This was associated with a rich nvertebrate fauna and few fish, except perhaps Notothenia spp. The Burdwood Bank presents very poor prospects owing to foul ground and prevalent gales with a steep breaking sea. Subsequent work substantiates these pronouncements in greater detail. 234 DISCOVERY REPORTS THE SECOND SURVEY The second trawling survey was carried out in the winter (June and July, 1928). Mr D. Dilwyn John was in charge of the scientific work and was assisted by Mr J. W. S. Marr. This survey was planned as a continuation of the first, with additional observations designed to discover the conditions to be found along the edge of the shelf in depths below 200 m. Routine methods at the trawling stations were the same as those followed during the first survey, with the addition of extra water sampling before as well as after the trawling. At this time the ship was (unavoidably) without an experienced trawler hand, and Mr John remarked that this led to much difficulty until the necessary experience was acquired. Much foul ground was encountered so that much time was spent in mending nets. The weather was often atrocious, and some minor breakdowns hindered operations still further, so that the ship. did well to complete a rather larger programme than before (Fig. i b) in about the same time. Material was dealt with as in 1927, and a large representative collection of bottom samples and of the fauna was made. On this survey the numbers of hake taken were small (1071 in all) and good hauls few, but it is probable that they were present nearby in greater abundance, as the following considerations show. The best catches were obtained on the edge of the shelf in depths of 200-300 m., on a line running north to a point some 300 miles north of the Falkland Islands. Later the same area was disappointing. The bottom was of clean, fine, dark green sand. On the coarse brown sand, pebbles, and shells, of the shallower waters of the shelf, very few hake were taken in the north, and fewer or none to the south. In the trough of relatively deep water to the west of the Falkland Islands, two very moderate catches were obtained. No hake were taken on the shelly bottom with heavy invertebrate fauna to the south and south-west of the Falkland Islands. Comparing these results with those of the previous autumnal survey, when most hake had been captured on the shelf to the north and west of the Falkland Islands, it was seen that the different distribution observed in winter would agree with an offshore migration. Such a seasonal movement, connected with the sexual rhythm, was already known to occur in the closely allied European species, and by analogy John concluded that the Patagonian hake were summer spawners also. The data on condition of the gonads, though not entirely satisfactory, were compatible with such a view. The comparatively good catches in shallower water of March and April 1927 would thus be accounted for by closer proximity to the spawning season, and the poor hauls of June and July 1928 mainly by fishing ' out of season '. It was known that European hake were caught most readily when concentrated for spawning in relatively shallow water, and that deep-sea trawlers from British ports followed the ' seasons ' south- wards, sometimes as far as the Moroccan coast. Also some of the British boats were already working ' over the edge ', in far deeper water than any in which large-scale trawling had previously been carried out, in order to keep the market going during the off-season. Hence John's decision to work extra stations along the edge of the shelf. The fact that these were only moderately successful was almost certainly due to the steepness of the slope (far greater than that off the west coast of the British Isles) which left a very small area of moderately deep water in which fishing was possible. John concluded that a commercial fishery would have to follow the spawning fish throughout the year, and that that would mean going farther afield to the northward than the first two surveys had proceeded. Sub- sequent work fully substantiates this view. The most important result of the second survey lay in this recognition of the fact that Patagonian hake would be found to move with the seasons in much the same way as their better known European relatives (allowing for the reversal of the seasons in the southern hemisphere). INTRODUCTION 235 Apart from the direct distributional study, the sex, lengths and maturity stages of all hake captured were recorded. The great difficulty of recognizing clear-cut stages in the development of the ovaries of these fish first became apparent at this point. The difficulty of standardizing such observations on this subject as are possible by direct inspection remained a handicap throughout. Probably it can only be tackled by large collections for subsequent microscopic examination such as Hickling (1930ft, 1935ft) developed during his prolonged work on the European species. Hickling found it possible to distinguish major stages by naked-eye appearance, but our observers agreed that this was never satisfactory with females of the Patagonian species. These are only slightly smaller than the European fish, but they often mature when considerably smaller, with a consequent increase in the amount of overlapping of ovarian developmental stages among fish of the same length class. Hickling was working upon a single species, but a limited staff investigating virgin ground could not make a large enough systematic collection of ovaries without neglecting other essential work. On the winter survey female hake were markedly more numerous than male. As in the autumn the males were associated with the smaller females, and very few males were taken where large females preponderated. The majority of the smaller females were immature. Very few of the large females were ripening, and the majority seemed to be 'spent'. Among niales the proportion of immature to mature fish was roughly 3:2. Very few were 'ripe and running'. The hake were found to be feeding very largely upon Euphausiidae, Clupeafuegensis, Notothema spp. and squids. Apart from hake, no fish were taken in such quantity as to be considered of possible commercial importance. Notothenia ramsayi were fewer than on the autumnal survey, and the best catches were taken on the same grounds as the hake. All were measured and sexed. There was a marked preponderance of females. Cottoperca gobio seemed to favour the rich invertebrate area to the south of the Falkland Islands, but was found elsewhere as well. Some seven or eight species of Rajidae were observed. Raja brachyurops being the commonest. Very full notes on the large range of variation to be found among the members of this group were made These were of great value later, when our data were brought into line with Norman's revised taxonomy of the Patagonian species. Salilota australis was taken quite frequently (fifteen stations), mostly in the more northerly part of the area, but in small numbers. Specimens of three species of rays and two species of Zoarcidae, all new to science, were obtamed during the second survey. ,,-11 c Three well-defined associations of invertebrates were observed, correlated with three types ot bottom deposit. Along the edge of the shelf north of the Falkland Islands to 45° S a fine green sand giving a clean bottom for trawling was prevalent in depths of 150-300 m. This ground was very rich in the smaller invertebrates-small ophiuroids, echinoids, other echinoderms, Serohs, amphipods, cumaceans and ostracods. Small quantities of Cephalodiscus occurred at seven of these stations. The Falkland trough yielded similar results but without the vast numbers of small ophiuroids, in slightly greater depths of water. Both these grounds yielded bigger quantities of hake and of rays than the other areas worked during the winter. . . . , , . , On the continental shelf to the north, west and south-west of the Falkland Islands there was darker, coarser sand, often brown, with pebbles and shells. Some patches were heavy with large invertebrates such as sponges and molgulids, and others comparatively clean. The invertebrates were heaviest close in to the Falkland Islands, and especially to the west of the Jason Islands. . ^ ^ The coastal waters to the south-west, south and south-east of the Falkland Islands again showed a strikingly rich invertebrate fauna of sponges, hydroids, coralline hydroids, Alcyonana and Polyzoa. This sheltered a rich fauna of smaller invertebrates: holothurians (especially Synapta), polychaetes and small Crustacea. Here the bottom was of light green sand with pebbles, shells (many of pectens), 236 DISCOVERY REPORTS large stones and boulders. It was found that this area extended over the 200 m. line for some distance beyond the immediate coastal shelf investigated during the previous survey. Mr John concluded his preliminary report (which has not been published) with the statement that Mr Marr emphatically agreed with his opinion that ' any decision on a commercial fishery in and near the waters of the Falkland Islands must depend on a greater knowledge of the seasonal variations of the hake of the locality, and of waters farther north, and a full knowledge of what is known concerning hake in other waters '. The preliminary results of the first two surveys have been treated at some length to show how great was their value when the third survey was planned. Several of the pioneers (notably Mr John) were engaged upon other studies after finishing their share of the field work, and the value of their efforts should be recognized. The third, most extensive and important survey, was carried out by Gunther himself with the assistance of Mr G. W. Rayner. The results form a large part of the main substance of this report. In this section, therefore, I have referred to the field observations only where they modified the general picture previously gained of the conditions. The longer period available, and increased experience, permitted improvements in working methods which must first be made clear. THE THIRD SURVEY At Dr Kemp's suggestion the plan of the third survey was designed to include five lines of stations spaced at regular intervals over the shelf between 44 and 54° S. Rough ground was to be avoided when possible, but eflForts made to keep the stations uniformly spaced. Each line was planned on a course of 111° — roughly normal to the coast. The isotherms here run almost parallel to the coast, and the value of observations upon temperature and salinity is much increased if they can be made as nearly as possible at right angles to the isotherms, as our hydrologists have frequently pointed out. After the two northernmost lines of stations had been completed, permission was obtained still further to extend the scope of the survey. Gunther did this by interpolating four lines of more closely spaced trawling stations, and by making many additional observations, including a north-south line that pro- vided valuable evidence of the effect of latitude. This tends to be masked (in the absence of such evidence) by seasonal movements, the effects of increased depth, and so forth. As finally carried out the programme included vertical hauls of the Gran international net (N50V) with water samples from surface and bottom at all stations. The vertical nets provide evidence on the distribution of fish eggs and phytoplankton, the water samples gave temperatures and salinity determinations. The trawling stations were of three types : (i) On the lines A, B, C, D and E (those of the original plan) the stations were 60 miles apart and the procedure the same as on previous surveys, with the addition of the N50V. (2) On the intermediate lines W, X, Y and Z, the stations were 30 miles apart, and in addition to the normal routine the trawl was shot for a further period of four hours whenever fish seemed plentiful. (3) On the additional lines the trawl was shot for one hour at stations 60 miles apart, the conical dredge was not used and no repeat hauls were made (see Figs, i c, d). Russell's bottom-net, with plankton nets towed obliquely on the same wire, was fished on six rather irregular lines during passages to and from the trawling stations, and on two more lines of additional observations where a small beam trawl was also used. The observations made included some important additions to the standard practice of the first two surveys. Among the essential routine observations the sorted fish were weighed in addition to being sexed and measured, and the larger species of invertebrates were also weighed in addition to being counted and listed. At selected stations whenever possible the important fishes, after being sexed and INTRODUCTION 237 measured, were sorted into length groups for weighing. Notes on maturity stages of ovaries, stomach contents and parasites of fishes were made as before. In addition, an attempt was made to assemble collections of scales and otoliths from prescribed length groups of Merluccius and Macruronus. Unfortunately, it has not yet been possible to work them out. It is known that they were too small to provide conclusive evidence as to growth rate, but they will give a clue which may help to clarify conclusions drawn from other lines of inquiry that must still be regarded as tentative. A series of carapace-breadth and weight records of the centolla crab, Lithodes antarctiais, were kept. These also await examination. Some biometric data on squids (body lengths and weights) were recorded, and a lot of data on the numbers, size, sex and incidence of bopyrid parasitism of Miinida spp. These have been utilized by Rayner (1935) in an important study of the growth of these Crustacea. In their pelagic stage they are predominant among macroplankton animals of the more coastal waters of the area (in due season), a fact observed by the earliest navigators of these waters (see also Matthews, 1932). Munida spp. are important as food for fishes, birds and whales. The much greater scope of the third survey, especially in more northerly parts of the area, led naturally to more and different kinds of fishes being met with. To the north such forms as Callo- rhyncJnis caUorhynchus, Seriolella porosa and two soles were unfamiliar, and there also the centolla crab was to some extent supplanted by the large red oxyrhynch Libidoclea sp., though at some stations both were found. To the south our fauna list expanded as a result of more extensive observations in the deeper waters of the Falkland trough, and in shoal water close to the mainland. Two macrurids, a ray, Parana signata and Sebastodes oculatus were the principal additions here. Only two species new to science were recorded from the trawling stations, although the third survey was more ex- tensive than the first two combined, and carried out under better conditions with all the added advan- tage of previous experience. By contrast, ten new species were found during the first two surveys. This is good evidence that the gear and methods used were adequate to provide a general picture of that portion of the fish fauna that can be sampled by trawling. The two new species discovered during the third trawling survey were Raja midtispinis and Notothenia macrocephala (Norman, 1937, pp. 20, 68). A most important addition to ideas gained from previous surveys was the discovery of rough ground at the edge of the shelf in several places, whereas Mr John had been fortunate enough to find fine, clean sand there. This rough ground produced striking examples of specialized distribution of fishes found elsewhere in shallow rocky waters but not on the shelf between. The weight records constituted a big advance, and permit of a much better general idea of fishery prospects, the relative importance of potentially useful species, and probable breeding seasons. The observations made by Gunther and Rayner with spring balances seem amazingly accurate, from the consistency of ponderal indices {K) calculated from them. Hickling (1930Z), pp. 7-8) has also shown that very good results can be obtained with weighings made on small ships in rough seas. With less detailed studies in view, Gunther and Rayner were weighing several fish at a time, in length groups ; not individual fishes and organs as weighed by Hickling. Since the individual lengths were almost always accurately known, it is possible that for the calculation of broad mean values, such weighings are better than individual ones. A second innovation during the third survey, that of repeat or control hauls, gave valuable evidence on the shoaling of fishes. They also showed that the nature of the bottom changes so gradually on the shelf that little would be gained (from the viewpoint of such studies as these) by a closer spacing of the stations. During this survey Gunther recorded his opinion that ' the extreme scarcity of fish in the immediate vicinity of the Falkland Islands is unquestionably due to seals. Seals were frequently met far from 23g DISCOVERY REPORTS the coast and seem likely to affect the fishing over a much wider area. This raises the question whether fish or sea-lions are the greater asset. While trawling on rough ground near the Falkland Islands would be difficult, so that here the sea-lion would be the more remunerative, it is possible that if their numbers were reduced fish farther afield might be turned to still greater account.' It is now known that sea- lions feed largely upon cephalopods (Hamilton, 1934), but there is little doubt that they eat con- siderable quantities of fish also. The cephalopods are themselves among the most important fish foods, and the way in which sea-lions frequented areas in the open sea where fish were shoaling was remarked by all who took part in the trawling surveys. The Falkland Islands are also the breeding grounds of vast numbers of oceanic sea birds, many of which eat fish whenever they can get them. These, too, probably play their part in keeping down the local fish population. Gunther's notes also include some pertinent statements about the invertebrate fauna of the region : 'Lithodes and squids are of direct economic importance to the South American market; macro- planktonic forms like Parathemisto and Munida spp. are among the important constituents of fish- food ; the rich benthos appears to reflect the character of the sea floor, and is thus bound up with the distribution of fish. Very full notes upon the quantitative distribution of invertebrates were therefore made, but it will be impossible to make full use of them until systematists have revised the taxonomy of the several groups.' 'The plankton of the area appeared to be poorer than that of corresponding latitudes off the west coast of South America, and on the sub-Antarctic whaling grounds.' Unfortunately, one has to add that, owing to the war, most of the collections, both of plankton and of benthos, still remain to be worked up. The magnitude of this task may best be judged from a quotation from the summary of Gunther's unpublished report on the work of the third survey: '. . .the OTC was shot at 80 out of 131 Sts. and repeat hauls made at 24 Sts. BTS was shot at II Sts., BNR at 46, oblique plankton nets at 50, and N50V at all Sts. Roughly 150,000 square miles between the Patagonian Coast and the Falkland Islands, and from 44° to 54° S were examined. The trawling stations were arranged in 11 lines and at a rough average were 50 mi. apart.' This report of Gunther's includes memoranda upon visits to the fish markets of Montevideo and Buenos Aires, and upon the history of the trawling industry off the River Plate, based upon the account of Devincenzi (1926). Both have been of great help to me in writing this work. Throughout the period of these investigations and for some years afterwards, our collections were supplemented by specimens of littoral fishes obtained at the Falkland Islands by the government naturalists Mr A. G. Bennett and Dr J. E. Hamilton. Both these officers worked in close collaboration with the Discovery Investigations; indeed, Dr Hamilton was seconded for service with us for many years, his chief work being on the bionomics of the sea-lion, which happily gave opportunity for many incidental observations on shore fishes. The value of their work, especially Mr Bennett's, can best be appreciated from the frequent references to it in the main part of this paper and in Norman's (1937) systematic report on the coast fishes. Norman's work gives us the taxonomic foundation without which this could not have been written. In addition to the Discovery material he was aided by various small collections made on the mainland coast by workers to whom he has already made acknowledge- ment. Norman also had access to all the material from the region already preserved in the national collection. TOPOGRAPHY OF THE SHELF The locality and extent of the region surveyed have already been described. The names of the features most useful for general descriptive purposes are shown in the general chart of the area (Fig. 2). This chart also shows the arbitrary division into northern, southern and intermediate regions best suited to the handling of the data available. Most of the names are those of salient features of the coast and INTRODUCTION 239 I I 63° GO" 45 ■ SO Rierto Mddrun \ \^ T Golfo NuevD Northern Region RiaCoig -^l Grande Bai:^ t=^ ^^^^^-APuertx) G alleg-os S^ — w)) C Virgins ^ Sir. Sebastian. 9t' FALKLAND Is/-. Southern Kegi.DtS ■^ ..• C.Horn °Diego Ramirez 1; ^-^ Statenl. N I I ' W"' J> BURBY^OOD BANK I ...- ■•• 200m"" II I f* '-n-'-l i-r-r-r- 65° 60° -I " ' ' ' t^' ■ ' ' I Fig. 2. General chart of the area surveyed. 45 50 55 240 DISCOVERY REPORTS require no comment. I have preferred the name ' Falkland trough ' for the broad tongue of relatively deep water running north between West Falkland and the mainland, to the ' Falkland channel ' of some earlier writers. In describing faunal distribution in this part of the world it is particularly important to avoid such misunderstandings as could arise by indiscriminate comparison of this natural feature with the numerous and straiter 'Magellan Channels' to the westward. The broad term 'plain of the shelf has been used to describe the central portion of the area where the gradient of the sea floor is exceptionally slight. The topography of the shelf has been admirably described by L. Harrison Matthews (1934, pp. 177-9) ii^ ^^is valuable account of the bottom deposits sampled by the conical dredge during the trawling surveys. The sections shown by him (loc. cit., Fig. i) give ample illustration of the features that exert most influence in a study of the distribution of the fish. These are: (i) The extremely slight gradient from 80 m. right out to the shelf edge at the 200 m. contour (the distance exceeds 200 miles in places). This is most marked in the north; farther south the sea floor slopes a little more steeply, but it is generally true to say that most of the shelf lies below water of more uniform depth than is to be found over comparable areas elsewhere. The great importance of this fact in studying fish distribution lies in the difficulties that result in interpreting depth relations, especially of migratory species. Elsewhere hake movements lead to clear correlation between size of fish and depth of water at appropriate seasons. Here the depth gradient is so slight that evidence of similar movements can only be detected when distance offshore is substituted for depth. The diffi- culty was augmented by areas of very slightly shallower water offshore, notably off the Golfo San Jorge and south-east of Puerto Deseado. These very slight elevations can hardly be termed 'banks'. (2) The slope from the shelf edge to oceanic depths is very steep. In the north it is almost pre- cipitous, so that trawling in water below 200 m. was limited to within a very few miles of the shelf edge. Farther south trawling between 200 and 450 m. was possible over a wider area, but was difficult owing to rough ground. (3) The Falkland trough and the area of deeper water separating the Falkland Islands from the Burdwood Bank are also well illustrated by Matthews's sections. It is possible (for our present purpose) to make a slight improvement on Matthews's general chart of the bottom topography (1934, pi. iii), by including several more recent soundings and plotting the 80 m. contour. This is shown in Fig. 3, which should be compared with Fig. 2 and with the separate distribution charts when depth relations are under consideration. , The main body of Matthews's work dealing with the grading and distribution of the deposits themselves is very instructive. The distribution of the coarser grades (mostly in the south of the area and along the landward and seaward margins of the shelf) has obvious practical significance : coarse grades, especially large fragments (loc. cit., pi. iv), usually coincide with foul ground for the trawler. Matthews's main conclusion was that deposits became finer as one proceeded northwards (fine sand and silt clearly predominate to the north), and that this is due to elutriation by the prevailing north- ward flowing Falkland current, which may be said to act as a natural levigator. This is one reason for the better trawling conditions found to the northward. To extract the full benefit from Matthews's work one would need to study first the relations between the distribution of the deposits and the sessile benthic fauna which is more directly affected by the nature of the bottom than are most of the fishes. Unfortunately, as already stated, this colossal undertaking can only be begun when taxonomic revision of the main invertebrate groups is achieved. However, one very striking correlation between fish distribution and bottom deposits can already be demonstrated. The abundant occurrence of the flatfish Thysanopsetta naresi was almost entirely restricted to a central area of brown sand delineated by Matthews (loc. cit., pi. xii, area 'C'). INTRODUCTION 241 Fig. 3. Topography of the sea floor. (Depths in metres.) 3-2 242 DISCOVERY REPORTS Thysanopsetta is one of the important forage species for larger fish, so that the restriction to a particular type of bottom deposit, later described in detail, affords a good illustration of the potential value of Matthews 's work. HYDROLOGY The water movements over the area covered by the trawling surveys are comparatively simple. Over the whole of the plain of the shelf and eastwards beyond the shelf edge, relatively cold sub-Antarctic water flows northwards in what is known as the Falkland current. To the east and north of the area the warmer subtropical water of the Brazil current flows southwards, and in the region of the con- vergence between these two (corresponding to the subtropical convergence in the open ocean still farther to the east) hydrological conditions are more complicated. Here streams of sub-Antarctic and subtropical surface waters may alternate, giving rise to large differences in salinity and tem- perature within a few miles. Klaehn (191 1) was able to trace the southward movement of the Brazil current as far as 49° S to the north-east of the Falkland Islands, but Deacon (1937, pp. 5S-9) has shown that south of about 431° S, the subtropical water is becoming more and more mixed with sub- Antarctic water. The influence of the Brazil current is strongest in summer, when relatively unmixed subtropical water may extend some 4° farther south than in winter. The complicated conditions around the southern extremity of the Brazil current rarely impinge on the trawling area, but this is the probable route by which fishes that normally live in warmer seas occasionally reach the north-east coasts of the Falkland Islands. The main facts concerning the southward limits of the Brazil current have been ably summarized by Deacon (1937) and an attempt has been made to depict them in Fig. 4, which should be regarded as a pictorial representation of the current system, and not an exact hydro- logical study. The Falkland current, which bathes most of the shelf, is composed of sub-Antarctic surface water. Deacon (loc. cit., p. 51) has described how the main west-wind drift of the south Pacific is com- pressed while passing through Drake passage. This augments its speed, and the sub-Antarctic portion of It IS swollen by a relatively small amount of warmer, poorly saline coastal water flowing southwards down the south-west coast of Chile. The resultant of these forces is the so-called Cape Horn current, which IS really a local intensification of the west wind drift. It sometimes reaches a speed of as much as 40 miles per day. To the east of Staten Island the Cape Horn current divides in the form of the greek letter y the esser branch swinging north round the Falkland Islands, but mainly between them and the main- land, to form the Falkland current; and the main branch proceeding north-east and then east until it merges into the main easterly (i.e. 'west wind') drift of the open ocean south of the Atlantic The Falkland current itself flows most rapidly on its right flank, well oflFshore and beyond the shelf edge, outside our immediate area. It is here that the coldest water is found, but the lower tem- peratures are not caused solely by the greater speed of flow than that obtaining over the shelf. Both Krummel (191 1) and Klaehn (191 1) postulate upwelling as an additional source of the cold water. Where the Brazil and Falkland currents are flowing in opposite directions alongside each other it is natural to suppose, as did Klaehn, that the dynamic disturbances so set up favour the creation of whirls with consequent upwelling. Deacon (1937) considers that Klaehn 's demonstration (191 1, pi. 35 hg. 4) of isolated patches of relatively cold water towards the northern end of the Falkland current is proof of upwelling, but our own data did not then provide any fresh evidence of its mechanism.^ ^^\^^^^^:^^X^£^^^ -' ^- -"-d by the Dis^very Cc.iUee's INTRODUCTION 243 Fig. 4. Current system of the area surveyed. Black stipple indicates cold currents, and red indicates relatively warm currents. 244 DISCOVERY REPORTS The Steep slope of the shelf edge is almost certainly one of the factors involved. These complications in the coldest part of the Falkland current occur mainly to the eastward of the trawling area, and therefore need not concern us further here. Along the mainland coast the speed of the Falkland current is greatly reduced and close inshore southerly movements of surface water may predominate. Consequently the water close to the main- land is warmer than that over the plain of the shelf. A definite counter-current close to the beach is set up in the summertime, which may flow (more intermittently) at other seasons also. These inshore conditions are clearly indicated by the direction of the isotherms in Klaehn's charts (191 1, pi. 35). The warmer inshore water does not result from any southward translocation of subtropical water, like the Brazil current offshore. The latter begins to swing away from the land well to the north of our area, usually in about 30° S. The warmer inshore water is formed by 'warming-up' of sub- Antarctic water, owing to the slower rate of advance on the left flank of the Falkland current. It may be described as ' old shelf-water '. In the area of the trawling surveys there are only these two main hydrological features to be visualized, the northward flow of the Falkland current over the plain of the shelf, colder and faster^ offshore along the shelf edge ; and the old warmer water close inshore with a more or less definite southerly trend. It is helpful in considering the distribution of fishes to gain some idea of the annual cycle of tem- perature of the water. This enables some direct comparison to be made with conditions on better- known fishing grounds elsewhere. Here we are handicapped by the fact that our three surveys were made at different seasons in different years. Klaehn's very thorough averaging of results from many years' observations, at a time when the region was much more important for traffic,^ provides the general picture we need, though it deals with surface waters only. From a careful check of Klaehn's charts against our own results it appears that the three years in which our own surveys took place must have conformed closely to the 'average' conditions depicted by him. Comparing observed tem- peratures obtained with the best modern apparatus on our surveys with Klaehn's monthly means, we find a resultant mean error of -0-26° C. in autumn, -0-40° C. in winter and -0-07° C. in summer. The range of error introduced by 'assuming' Klaehn's mean values instead of our own would be -i-oi to +0-49° C, -1-35 to +o-8i° C, and -3-21 to +1-95° C. respectively. Some error must be introduced by the time interval between our individual observations and the middle of the month, the rest may be confidently attributed to diurnal variations, as indicated by the dis- crepancy being greatest in summer. Klaehn worked on ten years' records from sailing ships (doubtless mainly nitrate clippers) and four years' records of the German Admiralty. We know that two years (1896-7 and 1906-7) out of the total studied by him were exceptionally cold, with icebergs drifting far north in the Falkland current (Krummel, 191 1, p. 606, fig. 172). Such conditions are rare, and this fact alone is probably sufficient to account for the slight tendency of Klaehn's values to be lower than ours, especially in winter. More important still, a comparison between the observed differences m surface temperature between successive pairs of our own observations, and the differences that would be expected from Klaehn's results revealed a close correlation {r= +0-85, with P much less than o-oi). It seems clear, therefore, that Klaehn's results give a very adequate picture of the ' average ' conditions, which will give a more satisfactory idea of the annual cycle of surface temperature than any scheme of plotting our less numerous and more scattered observations with interpolations for the gaps. 1 Some 13 sea miles per day according to Klaehn (191 1). sequently""^"^ °^^" ^^"' °^ '^^ southern ocean, this region was far better known during 'the last days of sail' than sub- INTRODUCTION 245 Graphs for the annual cycle have been plotted from Klaehn's charts in Fig. 5. The three positions have been chosen arbitrarily on account of their approximation to the centres of the most numerous of our observations on the fishes of the shelf, in each of the three main latitudinal regions into which our stations may be grouped. The curves show that the temperatures in our 'Northern region are very slightly colder than those found (say) on the hake grounds south of Ireland. The range is similar, but the winter temperatures nearly i^ C. lower. The temperature cycles plotted for our ' mtermediate and 'southern' regions show a decrease with increasing latitude, as is to be expected; while the three curves together show the diminution in annual range as one proceeds southwards. This is also well illustrated by Klaehn (1911, pi. 34, ' Jahresamplitude der Wassertemperatur')- 15-1 14- 13- u o uj 12- D ^ 1|. u, 9- u ■ c Northern Region --X-- Intermediate Region — Jk— Southern Region < in 8- 6- JULY 1 AUG I SEPT | OCT | NOV | DEC | JAN | FEB | MAR \ APR ] MAY | JUNE | Fig. 5. Representative annual surface-temperature cycles for the plain of the shelf, after Klaehn (191 1). For consideration of the bottom temperatures, which are more significant in relation to our trawled samples of fishes, we have only our own scattered observations to go upon, ^he actual %^^^^^^^^ eiven in Appendix I. It is obviously impossible to show typical seasonal curves, like those derived from Klaehn's monthly charts of surface temperatures, owing to the small proportion of these obser- vatLs that fall within any one depth range over a reasonably restricted area. A single, partly hypo- Ih Ucal, curve for the intermediate region is shown m Tig. 6. With this as a rough guide it is possibl to perc ive some important features by simple inspection of the figures quoted in the Appendix. The seasonal cycle appears to be centred later in the year m the bottom water, with maximum tempera- Tre^i: Mich. I autumn and the first half of the winter there is much less dif^^^^^^^^^ between surface and bottom water than at other seasons; it is rarely more than 2 C. whereas in summer it usually amounts to some 4° C. This is due to intense mixing over the shelf m winter, aradrdescribed'by Deacon (.937, P- SD- ^ spring and early summer even in the -rthern -^^^^^^ some temperatures below 5° C. were recorded (lower than the winter values), but most of these were in fairly deep water well ofTshore. 246 DISCOVERY REPORTS These bottom temperatures are particularly important in their relation to the most promising fish of the region, the Patagonian hake. It is quite clear, from the general run of the figures, that these are plentiful at some seasons in waters where the bottom temperatures are around 6° C. Their lower hmit seems to be some 3° below the figure 87° C. which Hickling (1927, p. 67; 1928, pp. 81-2, 88) found to be the normal cold limit of penetration of European hake. The ' stumpies ' of the Norwegian deep are commonly found at lower temperatures, it is true; but Hickling 's later work provides ample evidence that they are a local race, unrepresentative of the main stock, quite apart from the tem- perature of the waters in which they are found (Hickling, 1930ft, pp. 50-1). The bottom temperatures north and south of that part of the shelf represented by our 'partly hypothetical cycle' seem to vary with the seasons in similar fashion. To the north they probably 'average' 1° C. warmer, and to the south perhaps 1° C. colder (certainly so in summer). The main JULY I AUG I SEPT OCT NOV I DEC I JAN I FEB I MAR I APR I MAY I JUNE Fig. 6. Bottom temperature on the shelf; partly hypothetical annual cycle for the intermediate region, corrected to an assumed mean depth of 100 m., with surface cycle for comparison. feature is the lag of at least a month after the surface maximum, and the relatively high autumn and winter values due to mixing. For this reason the seasonal graph does not fall anything like so steeply as does that showing the surface temperatures, from the maximum to midwinter. Observations of the nutrient sah content of the waters of the Falkland current are not yet available for this area. PLANKTON During the trawling surveys some phytoplankton samples were collected from the waters of the shelf with the Gran net. They have not been fully analysed because of pressure of other work. Most were obtained during the third survey. Mr Rayner made some preliminary observations on them (un- published) and has recorded his impression that they indicate a poorer phytoplankton than that found off the west coast of South America. Some of our early phytoplankton work in the ' Discovery H ' extended northwards, to the extreme south of the trawling survey area, and there we found some INTRODUCTION ^47 evidence of a secondary autumnal increase in March 1930, with Rhizosolenia spp. dominant (Hart, '^OurTatlr wo/k from 1933 onwards was based on results obtained by Harvey's (1934) method. I have not previously dealt with our results from the sub-Antarctic Zone, because they were relatively few and widely scattered, not lending themselves to the methods of presentation adopted in dealing with the Antarctic material, which was the main object of study (Hart, 1942). Although most of these observations are from oceanic waters, I think it profitable to discuss them briefly here since they seem to me to give a good indication of the type of seasonal cycle (of standing crop) that may be expected on the shelf. In November 1936 several observations were made on the shelf and in the oceanic waters just beyond the shelf edge. These showed very little phytoplankton on the shelf but some rich hauls offshore. . ..^u^f^.oc The widely scattered observations available throughout the sub-Antarctic Zone, north of 54 b are shown as a scatter plot in Fig. 7. No grouping of these observations by time intervals would permit calculation of means that could be plotted to show the seasonal variation effectively^ Owing o the'mall numbers of observations and their wide dispersion, the differences between the monthly rneans for example, would not be statistically significant. Wide dispersion is a common source of rifficuity in all 'samples' of quantitative estimations of phytoplankton, whatever method is adopted^ I Teerns certam that it is due to the extremely patchy distribution of these orgamsms m nature, though 1 meThods have limitations that may exaggerate this feature. Plotting the logarithms of the individual observations instead of the actual numbers of units per m.^ overcomes another great difficulty due to t^^^^^ of finding a scale on which all the observations can be shown. The observations Iwn we not all obtained in the same year and, as already stated, they were scattered throughout Llblntarctic Zone, but apart from the logarithmic plotting they have not been otherwise mam- pulated fnty way. Moreover, nearly all the species involved have a completely circumpolar distn- Kiitinn as in the Antarctic Zone to the southward. tZ2Z shows clearly that the only months in which hauls exceeding looo umts per m.» occurred weTe Nov:ltr Dec "bl and March' Further, .he proportion of very small hauls, be ow too unrts were NovemDer ^ ^^^^^^^^ ^._^.|^_^ ^^ ^^^^ postulated for he rnhtroctreTa^r: IfThe Ant^arctie Zone (Hart, .,4. PP- 307-8). but centred slightly earher " wLr 1: Ts Manation of the low values recorded over the shelf in Novenrber .,36? The 'Tild not fake place un^ December in our southern locality. But cond.tions in the two areas are not "' I " n the Antar ti 'the time of the main increase in oceanic sub-Antarct,c waters ts remark- s'^(^hirrths' of the year, as compared 7;;-^-:^ rTm r^!::: L:^^SytSe:rr:eitres:^^^^^^^^^ 248 DISCOVERY REPORTS 5-R CO CO CO n • • V ^ rvj to -•-•— A- lb CO CD CD • •'• • Sxg' CO 3^ CO is t ; s CO CO CM CO O O O O ID O O o • 9. • CD O o o o [3ll^Hliyv90l]rrj y3d SUNG A3AaVH c o c o 3 I 3 rr- c li 1 o m ^ ca If 1 1-4 ^ > Z nt C o CO bo INTRODUCTION ^49 shelf waters sampled in November 1936. The fact that the temperatures diminish as one proceeds offshore supports this view. There should be an earlier tendency towards establishment of a thermo- cline, favouring earlier onset of the main diatom increase, under the warmer conditions prevailing '"'itTs^thus probable that over the area of the trawling surveys the times of maximal production of phytoplankton are only slightly later than in corresponding latitudes in European waters, and not so much later as in the oceanic sub-Antarctic waters beyond the edge of the shelf. , , , . The effect upon fishes of a slightly later timing of this basic element in the 'plankton-calendar of the locality is likely to be a corresponding shift in their own dispositions, in so far as these are affected by the plankton, for it is generally true that zooplankton maxima follow the maximal ' standing crop of phytoplankton. Some of the evidence on this point has recently been well summarized by Bogorov ^' One of the most striking features of the zooplankton of the region is the swarming of lobster-kriU (the Grimothea post-larval pelagic stages of Munida gregaria\ which occurs most frequently during the summer months. The aduhs of this anomuran also swarm at the surface occasionally. The swarms of post-larvae are often thick enough to impart a reddish appearance to the surface of the sea, observed by many of the earliest navigators of these waters (Matthews, 1932, P?" 479-8;, several references) The later phases of this species are limited to coastal areas round the Falkland Islands and along the mainland coast, within our area (Rayner, X935, %• 18). A closely allied iorr.M. suhrugosa, is even more widespre;d and abundant on the shelf (Rayner, i935, P- ^38, fig- ^D- The later stages of this species are more strictly bottom dwellers. Munida spp. are of great importance in the ecology of the region, as a staple food of whales (Matthews, 1932, PP- 481-3)- seals (Hamilton, 1934. p. 295), bird and fihes M. gregarra is also found fulfilling a similar role in New Zealand waters (numerous eferences quotef by Matthews). Other allied species of Anomura are equally -P-^-^/^^^^^^ Jf^^ Pacific off Mexico and southern California (information supplied by Capt. Fagerh and by Dr Waldo L Schmidt of the Smithsonian Institute, quoted by Matthews, 1932, p. ^']^). Rayner's detailed work on the growth of the Falkland species of M...W. shows that they are c-- paratLly long-lived (5 years or more of post-larval life in M. subrugosa) and are sexually matu e rom the'end of the first year of post-larval life. The Antarctic krUl, EupH..^a ^^P^rt^;^^^^^^ a less specialized group, and pelagic throughout its life history, lives for a much shorter time^ The di St X differences in pelagic life of the two regions may be affected by this difference in the If histo" of their respective key-industry animals, as well as by the more obviously important physical ^' Euphausians are quite important constituents of the macroplankton over the shelf, but are obviously les! important than'm the Antarctic. None have been observed to form ^^nse -rms^ ^^^^^^^^^^^ the surface of the sea, as E. superha commonly does in the Antarctic, and as Meganychphanes tZtallT^ inermis) more rarely do in north European waters. Two or more Southern p ies of Thysanoessa are to be found over the shelf, and of these T. gregana is the mo rindant From Tohn^s (1936) work we know that of the genus Euphausia, E. vallenUm is the most wilpread Z.^My the Ust numerous species around the Falkland Islands and in our southern lion £ ZL E. Lus, E. longrrostris and E. triacantha also occur. Judging by its general dis Sion tough^ut the sub- Antarctic Zone, E. luce, may be the most important species on the shelf farther north but quantitative data are not yet available. , . if ^ fi^I. wprP The ub nu ,ous hyperid amphipod Para,hen.hio gaudichaudii is abundant on .he shelf, and fi h were faqundy found .o'have been'feeding upon h. Numerous -'anoid copepods oecur and „^a - dance of large Scyphomedusae is frequendy referred to m the log books. The general 250 DISCOVERY REPORTS zooplankton is indeed not unlike that of north European waters, but important differences may be expected to appear, especially in the times of maximum abundance of the different groups, when the material is fully worked up. As already explained, however, this must wait upon the completion of taxonomic revision of the groups. Moreover, the environmental features of an area of this size, both physical and biological, present almost unlimited scope for further investigation. In general we may say that most of the more important macroplanktonic forms have a wide, often a circumpolar distri- bution within the sub-Antarctic Zone. They are noticeably less restricted than the benthos in their latitudinal range also, including a higher proportion of species that may extend into the distinct water- masses both to the south and to the north of the sub-Antarctic Zone. Among microplankton forms this wide tolerance is even more marked ; many of them are cosmopoHtan. METHODS OF PRESENTATION It has not been considered advisable to attempt to publish the biometric data on fishes in full, for the raw data alone in manuscript form weigh over i cwt. These papers will be stored by the Discovery Committee, and it is hoped that much information may still be gleaned from them. Full details of the station positions are appended, with some of the abstracted data on the most important.fishes. The numbers of fishes recorded are shown with station numbers only, in the general distributional accounts of each species. Wherever it has been found profitable to consider length and weight measurements these are given in the form of summarized tables and diagrams with legends which It is hoped contain sufficient explanation. The data relating to different species varies so much that It IS obviously impossible to adopt uniform treatment throughout. Such simple statistics as have been ventured upon relate mainly to length-frequency distributions and mean lengths. Where mean lengths have been used to compare or contrast catches from difl^erent stations or groups of stations I have included sufficient information as to dispersion to enable the significance of the differences to be determined. Simpson and Rowe's book (1939) has been a great help m dealing with the numerical data. The systematic arrangement and taxonomy of the fishes follow Norman's report (1937) throughout In discussing bionomics of the fishes I have stressed differences and resemblances between Patagonian species and those most nearly allied to them that have already been studied on better-known fishing grounds elsewhere. I believe that it is easier for readers without experience of this particular area to visualize the conditions if they are presented in this way. Since hake are the most important species on the shelf, I have tested out the theories developed by Hickling during his prolonged work on the iiuropean species, m so far as our scantier data on the Patagonian species permit. With the possible exceptions of cod, haddock, herring and plaice, Hickling's work on hake probably constitutes the most complete picture we have of the life history of any marine fish. Consequently all members of the Discovery staff who have been concerned in this work have studied Hickling's work intensively Iwo conventions have been used to facilitate general descriptions offish distribution, and to ensure that seasonal comparisons should be kept roughly comparable: . First, the whole area has been arbitrarily subdivided into ' northern ', ' intermediate ' and ' southern ' regions, according to latitude. The interval chosen was 4° of latitude, so that the northern region includes all trawling survey stations north of 46° S, the intermediate region all those between 46 and 50 b, and the southern region all those south of 50° S. This subdivision has already been indicated mtig. 2. It IS an arbitrary division introduced solely for the purpose of reducing the data to manage- abe proportions, but it approaches a natural division inasmuch as the physical conditions that change with latitude afl^ect the flora and fauna. Provided that large-scale migrations in a north and south GENERAL ACCOUNT OF THE FISH FAUNA 251 direction are not involved, observations at different seasons within any one of these regions are obviously more fairly comparable than those over the whole area. In the specific distribution lists, the regions to which individual stations belong are indicated by printing the numbers of northern stations in heavy type, intermediate stations in ordinary type, and southern stations in italics. Secondly, the reversal of the seasons in the southern hemisphere is indicated by beginning all time scales on i July, comparable to i January in the northern hemisphere. It should also be remembered that when the looser seasonal expressions 'spring', 'summer', etc., are used, a similar reversal is 'Twing to the shght gradient on the shelf, distance from the mainland coast is more significant in relation to seasonal changes in distribution than changes in depth of water inhabited by migratory species This figure has been calculated (in sea miles) for all the stations and is tabulated m Appendix I Conditions round the Falkland Islands themselves are peculiar, and they do not appear to be regarded as a coast at all by the main migratory species, which are almost absent f»-om their immediate vicinity. I have therefore calculated the distance from the nearest point of the Falkland Islands also, for those stations that fall within 100 sea miles of them. Beyond that distance migratory fishes seem to proceed towards the mainland (in due season) as though the Falkland Islands did not exist even though the mainland may be twice as far distant. The exact distances from the Falkland Islands of all other stations has not, therefore, been calculated; they are tabulated merely as more than 100 miles'. • ,• ■ r>- Abbreviated descriptions of gear are the same as those used throughout the station lists m Discovery Reports. The meanings of those used in this work have already been given m the section on field methods. GENERAL ACCOUNT OF THE FISH FAUNA A list of all the species of fishes recorded on the Patagonian Continental Shelf and immediately adjacent coasts is given in Table i. This also shows which species were obtained during the surveys in ' Trawl + accessory nets ', in ' Other gear ' and by shore parties. Most of the specimens recorded in the last category were obtained by Mr A. G. Bennett and Dr J. E. Hamihon. I have not included some recent records of Pozzi and Bordale (1935) which are mentioned by Norman (1937, PP- HS-^)- Some of these require confirmation. The list follows that of Norman (loc. cit., pp. 143-5) but omits species recorded only from the west coast (columns A and B of Norman's list). The historical aspect of the growth of our knowledge of the fish fauna is admirably dealt with by Norman (1937, PP- i37-4^). and from his account the debt we owe to earlier expeditions-British, French, Swedish and American — can be assessed. . . r From Table i it can be seen that ninety-five species are recorded from the region, and specimens ot seventy-eight of these were obtained by ships or shore parties in the course of the Discovery investi- gations. Also an undoubted basking-shark was observed. The records of two of the species that we did not capture seem somewhat doubtful: Bunocottus apus Kner is based on a single specimen said to have come from the Burdwood Bank, no other Cottidae are known from the region and Kner s description does not agree with his figure (cf. Norman, 1937. P- HS)- Alphestes afer (B och), a small serranid common in the West Indies, has a normal range extending from Cuba to Brazil (Jordan and Eigenmann, 1890, p. 35°). It is clearly a tropical species, so that its occurrence south of 42 ^ would be most extraordinary, and I am unable to trace the authority upon which Norman placed it in the Patagonian list. Among the other species not taken on the surveys, Notothenia tngramma Regan and Crossostomm fasdat^^s (Lonnberg) are known only from their unique holotypes, and it is possible that 252 DISCOVERY REPORTS Table i . Fish fauna of the Patagonian Continental Shelf Taken in ' Trawl + accessory nets' Taken in Taken by Common names Family Species 'Other shore adopted in gear' parties this report Petromyzonidae Geotria australis Gray — — X — Myxine australis Jenyns X X X Hagfish *M. affinis Guntherf — — — — Lamnidae *Cetorhinus maximtis (Gunner)| — — J.E.H. obs Basking shark Scyliorhinidae * Scyliorhinus (Halaelurus) bivius (Smith)f — — — Carcharinidae *Mustelus cams (Mitchill) — — — ■ Dogfish *Centroscy Ilium granulatum (Giinther) — — — Squalidae Squaliis lebruni (Vaillant) X — — ^ Torpedinidae Discopyge tschudii Heckel X — — V Rajidae Raja flavirostris Philippi X — — R. doello-juradoi Pozzi X — — R. macloviana Norman X — — R. magellanica Steindachner X X — R. multispinis Norman X — — Rays R. scaphiops Norman X R. brachyurops Fowler X X — R. griseocauda Norman X — — Psammobatis extenta (Garman) X P. scobina (Philippi) X X — Chimaeridae Callorhynchus callorhynchus (Linnaeus) X — — — Clupeidae Clupea fuegensis Jenyns X X X 'Herring' C. arcuata Jenyns X — — ' Sprat ' Galaxiidae Galaxias attenuatus (Jenyns) X *G. maculatus (Jenyns) — — *G. smitliii Regan — — Aplochitonidae Aplochiton zebra Jenyns — — X *A. taeniatus Jenyns — — — Syngnathidae Leptonotus blainvilleanus (Eydoux and Ger\'ais) X — — Pipefish *Entelurus aequoreus (Linnaeus) — — — — Macruridae Coryphaenoides holotrachys (Giinther) X — — Coelorhynchus fasciatus (Giinther) X — — Merlucciidae Merlucciiis hubbsi Marini X X Hake Macruronus magellanicus Lonnberg X 'Long-tailed hake' Gadidae Micromezistius australis Norman X Salilota australis (Giinther) X Physiculus marginatus (Giinther) X — Muraenolepidae Muraenolepis microps Lonnberg X — — M. orangiensis Vaillant X Lamprididae Lampris regius (Bonnaterre) X Serranidae *Alphestes afer (Bloch) — Carangidae Parana signata (Jenyns) X Cheilodactylidae Cheilodactylus bergi Norman — X Bovichthyidae Cottoperca gobio (Giinther) X X Bovichtus argentinus MacDonagh§ X Nototheniidae Notothenia macrophthalma Norman X *A^. trigramma Regan — — — — N. canina Smitt X X A^. jordani Thompson X X A^. tessellata Richardson X X X A^. brevicauda Lonnberg X X A^. guntheri Normann X X , A'^. ramsayi Regan X X N. wiltoni Regan — X X , A'^. squamiceps Peters — X — — N. sima Richardson X X N. cornucola Richardson " X X A'^. elegans Gunther X X — GENERAL ACCOUNT OF THE FISH FAUNA 253 Table i {continued) Taken in ' Travel + accessory nets' Taken in Taken by Common names Family Species ' Other shore adopted in gear' parties this report Nototheniidae Notoihenia macrocephala Giinther — X X — *N. microlepidota Button — — — — Dissostichus eleginoides Smitt X — — — Eleginops maclovinus (Cuvier and Valenciennes) — X X 'Mullet' Harpagiferidae Harpagifer bispinis (Schneider) X X X — Chaenichthyidae Champsocephalus esox (Giinther) X X X — Gempylidae Thyrsites atun (Euphrasen) X X X — Scombridae Gasterochisma melampus Richardson — — X — Zoarcidae Iluocetes fimbriatus Jenyns X X — — /. elongatus (Smitt)|| X X — — Austrolycus depressiceps Regan — — X — *A. laticinctus (Berg)t — — — — Phucocoetes latitans Jenyns X X X — *Crossostomus fasciatus (Lonnberg) — — — — Pogonolycus elegans Norman X X — • — Platea insignis Steindachner — X — — *Maynea patagonica Cunningham — — — M. brevis Norman|| X — — — Melanostigma microphthalmus Norman X — — — Ophidiidae Genypterus blacodes (Schneider) X X — — Brotulidae Cataetyx messieri (Giinther) X — — — Centrolophidae Seriolella porosa Guichenot X ■ — — — Palinurichthys caeruleus (Guichenot) X — — — P. griseolineatus Norman X — — — Stromateidae Stromateus maculatus (Cuvier and Valenciennes) X • — — ' Spotted pomfret ' Atherinidae Austromenidia smitti (Lahille) — X X } 'Smelt' A. nigricans (Richardson) — — X Scorpaenidae Sebastodes oculatus Cuvier and Valenciennes X — — — Congiopodidae Congiopodus peruvianus (Cuvier and Valenciennes) X — — — Cottidae [?!] *Bunocottus apus Kner [?!] — — — — Psychrolutidae Neophrynichthys marmoratus Gill X X — ■ — Agonidae Agonopsis chiloensis (Jenyns) X X ■ — — Liparidae Careproctus falklandicus (Lonnberg) X — — — *Liparis antarctica Putnam — — — — Bothiidae Thysanopsetta naresi Giinther X X — 'Scald fish' Paralichthys isosceles Jordan X X — Xystreurys rasile (Jordan) X — — Mancopsetta maculata (Giinther) X • — — ' Achiropsetta tricholepis Norman X — X * Species not taken during the surveys. f Species not taken in this area, but specimens obtained by the expedition elsewhere. \ Observed, but not taken. § Taken by Mr MacDonagh at Puerto Madryn. || Species taken only in 'Accessory nets' when trawled. the latter may prove to be a young example of Austrolycus depressiceps Regan (Norman, 1937, p. 106). Specimens of three species, recorded from the shelf but not taken there by us, were obtained on other occasions among the channels to the westward. Of our seventy-nine 'shelf species, sixty-one were obtained in the ' Trawl + accessory nets', thirty- five in ' Other gear' (including ten that were not trawled), and twenty-three by shore parties (including seven not taken by other means), the total being made up by the basking shark seen by Dr Hamihon. Thus we see that of the ninety odd species recorded from the region, two-thirds were obtained by trawling. The remainder are mainly littoral fishes; for example, eight of the notothenias and two zoarcids have never been taken in water deeper than 46 m. (25 fm.). The difficulties of obtaining littoral fishes in this area are great : the tidal range is small and the surf often heavy, while the great 254 DISCOVERY REPORTS Table 2. Fish fauna of the Patagonian Continental Shelf (trawled fish only). Total numbers taken and frequency of occurrence I Numerical Total nos. of Frequency of Frequency of class Species fish taken occurrence in 178 hauls occurrence as percentage > 1000 Notothenia ramsayi 9599 130 73-0 Merlitccius hiihhsi 5748 109 6l-2 Macruronus magellanicus 5336 63 35-4 Thysanopsetta naresi 1916 33 i8-5 Stromateus maculatus 1044 51 287 100-999 Clupea fuegensis 694 28 157 Micromezistius australis 557 21 II-8 Salilota australis 485 65 36-5 Cottoperca gobio 414 50 28-1 Raja brachyurops 274 48 27-0 Notothenia guntheri 267 11 6-2 Psammobatis scobina 153 45 25-3 Coetorhynchus fasciatus 140 12 6-7 Champsocephalus esox 124 15 8-4 Notothenia tessellata 102 9 5-1 50-99 Coryphaenoides holotrachys 95 8 4-5 Genypterus blacodes 74 33 i8-s Raja doello-juradoi 68 12 6-7 R. magellanica 54 21 II-8 Congiopodus peruvianus 5° 16 9-0 25-49 Agonopsis chiloensis 47 IS 8-4 Thyrsites atiin 45 6 3-4 Raja flavirostris 41 24 13-5 Clupea arcuata 39 I 0-6 Iluocetes jimbriatiis 38 22 12-4 Parana signata 35 2 II Notothenia canina 30 5 2-8 Callorhynchus callorhynchus 28 6 3-4 Notothenia jordani 25 3 1-7 Neophrynichthys marmoratiis 25 7 3-9 1-24 Myxine australis 24 7 3-9 Raja macloviatia 23 12 6-7 Physicidus marginatus 22 4 2-2 Dissostichus eleginoides 9 5 2-8 Squalus lebruni 8 6 3-4 Seriolella porosa 8 I 0-6 Raja griseocauda 8 6 3-4 Paralichthys isosceles 8 5 2-8 Raja scaphiops 7 5 2-8 R. albomaculata 7 6 3 "4 Notothenia elegans 7 S 2-8 Iluocetes elongatus 6 0-6 Careproctus falklandica 6 06 Disco pyge tschudii 4 0-6 Leptonotus blainvilleanus 4 0-6 Phucocoetes latitans 4 0-6 Maynea brevis 4 4 2-2 Xystreurys rasile 4 3 1-7 Harpagifer bispinis 3 I 0-6 Palinurichthys griseolineatus 3 3 1-7 Sebastodes oculatus 3 3 17 Pogonolycus elegans 2 2 i-i Melanostigma microphthalma 2 ' 2 i-i Cataetyx messieri 2 2 i-i Palinurichthys caeruleus 2 2 I-I Raja multispinis 0-6 Psammobatis extenta 06 Muraenolepis microps 0-6 M. orangiensis 0-6 Notothenia macrophthalma 0-6 Mancopsetta maculata 0-6 Achiropsetta tricholepis 0-6 255 GENERAL ACCOUNT OF THE FISH FAUNA beds of kelp {Macrocystis, Durvillea, etc.) may extend out to 30 fm. and render any form of fishing difficult, though they certainly harbour many species of fish. The trawl seems an adequate sampling instrument within the limits set by the mesh used. Hicklmg (i933' PP- I ^-^9) ^^^ g^'^^^ adequate demonstration of this, and our results with a closely allied species of hake (pp. 284-9) conform with his as closely as could reasonably be expected. Faulty hauls cannot prove absence or relative abundance, but have helped to prove presence of certain species on some occasions. The trawl cannot provide adequate data for pelagic species such as herring, or very small species such as some Zoarcidae. Immature specimens of some of the larger, more important species also escape through the meshes. This loss is most serious with the more slender forms, such as very young hake (especially males), immature Macrurontis, Mkromezistim and Genypterus. The fine- meshed nets attached to the back of the trawl caught enough of these to enable us to outline their probable distribution, but do not provide comparable quantitative data. Provided that these limita- tions are borne in mind. Table 2, which gives the total numbers of the species in ' Trawl + accessory nets', and their frequency of occurrence, helps to extend the outline of our general picture of the fish fauna begun by the first table. The fish fauna of the Patagonian shelf is not rich in species, as the full list in Table i shows. In a preliminary account of the trawling surveys,^ Gunther pointed out that it is less than one-third of the strength of the British list, and nearly twice as many species occur in the Gulf of Maine (Bigelow and Welsh 1925) Apart from the numbers of species, there are big qualitative differences from the types of fish' faunas known from other parts of the world. The marked predominance of the percoid group Nototheniiformes, with four families, seven genera and twenty-one species from the area of the surveys, and relatively large number of Zoarcidae with eleven species representing eight genera, are a most peculiar feature. Among elasmobranchs Rajidae show many species and remarkable diversity for an area where nearly all the known changes in environmental conditions are gradual. Dogfish are not common on the shelf, and the numbers of species of true codfishes (Gadidae) is small; in these features we see a great contrast to the fish faunas of northern Europe and of the New England states. Table 3 showing the relative strengths of some of the important groups in British seas, on the Pata- gonian shelf and in the Gulf of Maine, summarizes these points. The comparison between British and Patagonian fish faunas was first made by Gunther in the preliminary account mentioned above, and I have abstracted figures from Bigelow and Welsh (1925) for the Gulf of Maine to make the comparison wider. Table 3. Relative proportions of certain taxonomic groups in British seas, on the Patagonian shelf and in the Gulf of Maine No. of spp. Rajidae Percomorphi Gadidae Heterosomata British seas 350 5% .. 31%: No Nototheniiformes Zoarcidae i sp. 8% Patagonian shelf 95 10% 48%: Nototheniiformes 21 spp. Zoarcidae 11 spp. 3% 5% Gulf of Maine 173 3*°' 27%: No Nototheniiformes Zoarcidae 3 spp. 7% -710/ 72 /o The Nototheniiformes are, of course, an essentially southern group, but it has further to be noted that most of the Patagonian species are distinct from the Antarctic ones; and the number of species comln to other sub^Antarcdc localities, such as the Antipodes, is small. The relations between the 1 'A Fishery Survey of the Patagonian Continental Shelf read before Section D of the British Association, July 1938- 256 DISCOVERY REPORTS Patagonian fish fauna and that of Kerguelen and Heard Island were dealt with by Regan (1914, p. 36), who treated the latter as a peripheral district of the Antarctic Zone. Norman (1937, p. 148), reviewing the question with more recent evidence, reached the conclusion that the dissimilarity between the fish faunas of the Kerguelen and Patagonian regions was not so great as Regan had supposed, and pointed to several pairs of species of the closest phylogenetic relationship from the respective regions. However, it is easy to demonstrate similar close relationship between Patagonian species and others from very widely remote regions, as we shall presently show. Norman himself, in a later work (1938, pp. 100 et seq.), summed up the present position in regard to this question with the statement: 'it is clear that, although the coastal fish fauna of the Kerguelen district shows certain features of resem- blance to that of the Patagonian region and the Antipodes, its affinities are mainly with that of Antarctica.' This is precisely what one would expect from our latest knowledge of the hydrology of the regions concerned (Deacon, 1937). In strong contrast to the marked differences between the Patagonian fish fauna and that of other regions, some of which have been shown in Table 3, several important and familiar species from better-known grounds can be 'paired-off' with Patagonian species closely allied to them. This feature was first made clear in Gunther's unpublished work, and was subsequently expanded by Norman (i937» P- 146), and by Gunther himself in his address to Section D of the British Association during the following year. These workers were concerned to show the parallels between allied British and Patagonian species, as an aid to general description of the fauna. I have attempted to widen the basis of this comparison with further parallels from the Gulf of Maine and from South Africa. Naturally there are fewer closely allied species from such widely diverse regions. South African waters are subtropical, though with low temperatures in the Benguela current, and the Gulf of Maine is a very specialized ' cold-temperate ' area with exceptionally high summer temperatures due to the influence of the Gulf Stream. Nevertheless, I find these broad comparisons helpful in gaining an idea of the character of the Patagonian fish fauna, especially in conjunction with the roughly quantitative work to be described later. I therefore include them here in the hope that those who read this report may similarly be aided. Table 4. Closely allied species from the Patagonian shelf, British seas, the Gulf of Maifie and South African seas Patagonian shelf British seas Gulf of Maine South African seas Myxine australis Squabis lebruni Raja flavirostris Clupea fuegensis C. arcuata Merluccius hubbsi Micromezistius australis Salilota australis Sebastodes oculatus Stromateus maculatus M. glutinosa S. acanthias R. batis C. harengus C. sprattus M. merluccius M. poutassou Urophycis blennioides Sebastes marinus M. glutinosa S. acanthias R. stabuliforis C. harengus M. bilinearis Microgadus tomcod U. tenuis et spp. Sebastes marinus Poronotus triacanthus M. capensis S. acanthias R. batis M. capensis Sebastichthys capensis Stromateus fiatola It IS clear from this table that quite a number of Patagonian species have close relatives elsewhere, but when the quantitative aspect is taken into account we find that with the exceptions of the hake,' Micromezistius and Stromateus,^ the Patagonian species listed are relatively far less numerous than are their nearest counterparts on the better-known grounds we have considered. Thus the differences 1 Also the Falkland herring, but this could not be adequately sampled by the trawl. GENERAL ACCOUNT OF THE FISH FAUNA 257 between the Patagonian fish fauna and that of these better-known fishing grounds outweigh the resemblances, and .f we are to succeed in descnbing the general facies of fePatagoman fauna by analogy with that of other regions it will be necessary to cast our mmds still farther afield. The differ- ences have already been summarized (in part) in Table 3. Four notable ones are: predominance of Nototheniiformes the relative unimportance of Gadidae and of flatfish, and an absence of Salmomdae of useful size (which was not previously mentioned). In all these respects the Patagonian fauna differs markedly from that of better-known fishing grounds in the northern hemisphere. It is also most Unfortunately true that relative scarcity and small size of flatfishes and Salmomdae are features common to all the other fishing grounds of the southern hemisphere. Consider, then, an area in the northern hemisphere which we have not yet taken into account-the North Pacific. If we allow ourselves to imagine this fauna without its two best fishes, salmon and halibut, what would be its emating characteristics? Predominance of Scorpaemdae and allied families, especially Hexagram- mle (with Ophiodon elongatus, the cultus cod) and the Cottidae or sculpins. Herring would be Cort nt, but though Gadidae would be fewer than in other parts of the northern hemisphere they n s'ili rank high The relative importance of Merluccius productus, a true hake hitherto despised on that favoured coast, would be much enhanced. • , . . Now Jordan (,,05, vol. „. pp. 50. e. seq.) has pointed to the c ose analogy wh.eh ex.sts between the exdusnely southern Notothemiformes and certain families of the great group of ma.l-cheeked fishes whch L calls Pare.oplitae (including the Hexagrautmidae, etc.). Of the Notothenndae he retarklAeir general resemblances to small Hexagrammidae •, and a little later he speaks of them as ■ he n ipod s of the Cottidae and Hexagramm.dae ; although lacking the bony stay of the latter, they show severa analogical resemblances and have very similar habits', and aga.n ' . . ^Harpagrtendae resemb e u pins even more closely '. To th,s may be added the close superfiaal resemblances between o hewia nt Antarcttc) Nototheniiform fishes (Ba.hydraconidae and Chaemch.hydae) and ye other ?r,;ilsofth'emai.,cheked«..,.^^^^^^^ Z ^:: ^fi^tt: ■rhrN^JSrl^alue biol„g,ca1 coLerpart of the Notothen.iform draw between uie r g a r f^.^oP^.n Micropadus znA Micromezistius, Merluccius productus callorhynchus, Clupea pallasei and C. fuegensis, Microgaaus an Sebcutodes is of Uckthy. spp.) m S™* Am ™^7*J.i, „„ait ons, and in colder waters the scarcity of flatfishes :rj:f:Ui:ri:Tn-Uhrrngtn"he southern hem.sphere. Sizeable Salmonidae are also unknown 258 DISCOVERY REPORTS in the southern hemisphere (apart from introduced trout) so that these two deficiencies are by no means peculiar to the Patagonian shelf. The most encouraging features, more apparent when the quantitative data are considered, are that a true hake is common on the Patagonian grounds, that minor quantities of another merlucciid, a butterfish or pomfret, and lesser numbers of other species are all good eating, and that there is a Falkland herring which may prove to be of real value. Though our gear was admittedly unsuitable for sampling this last species, its abundance is placed beyond doubt by the frequency with which it was observed in the stomachs of larger piscivorous species. The Patagonian fish fauna has certain species in common with other southern hemisphere localities, and it may be thought that in attempting to draw descriptive parallels it would have been better to turn to these rather than to the northern hemisphere. I did not attempt this because the essentially subtropical conditions of the other southern localities leads to the prevalence of various percoid groups very different from the Nototheniiformes, so that although New Zealand for example has some species m common with Patagonia, and a few others closely allied, the general character of the vastly richer fish fauna is ahogether different from that of the Patagonian one. This remark applies with even greater force to the fish faunas of Southern Australia and South Africa. It is mainly in the scarcity or absence of certain groups, as Salmonidae and Gadidae, that the fish faunas of these areas can be said to show any resemblance to that of Patagonia. The extent to which we are forced back to the northern hemisphere for closer parallels is a measure of the extent to which meteorological factors, chiefly temperature, determine the conditions of life in the sea. It IS a striking fact that a true hake is found in each of the regions we have discussed : Merluccius merluccius to the west of Great Britain, M. bilinearis in the Gulf of Maine, M. productiis on the Pacific coast of North America, M. capensis off South Africa, and M. hubbsi on the Patagonian shelf. Three of these five species are already heavily exploited: prior to the war Hickling's work had shown (1935a) that the European stock was being overfished, while in the last twenty years M. bilinearis has risen from the status of 'rubbish' to the New England fishermen, to the highest place among the frozen fish products of the eastern states. Moreover, only half the catch is frozen. M. capensis forms a third of the catch of the trawling industry at the Cape, where large quantities of it are salted and dried. M. productus is not yet sought after. In a region where better fish are still abundant the softness of Its flesh makes it unpopular. In less favoured localities the other species of hake, probably little better m this respect, are valued and are successfully marketed, owing to improved methods of pre- servation and storage. This is especially true of M. bilinearis, half of the catch being sold inland in the form of frozen fillets, etc. This species is the one most nearly allied to the Patagonian M. hiibbsi. Such small quantities of the last named as are caught by the small trawlers which operate from the mouth of the River Plate, well to the north of our area, fetch prices well up to the average in the Buenos Aires market. In summarizing the points dealt with in this section of the report, we may say that the fish fauna of the Patagonian shelf is peculiar in quality and in the small number of species to be found there The number of potentially valuable forms is small too, and their quantity not encouraging, but hake are moderately abundant. The nearest parallel among fish faunas of better-known regions is probably that with the North Pacific, but with important reservations detailed above. There is also some resemblance to the conditions found on the hake grounds to the west of the British Isles. 259 DISTRIBUTION AND GENERAL NOTES ON THE SPECIES PETROMYZONIDAE A specimen of Geotria australis Gray was obtained by Dr Hamilton in the Falkland Islands, but none was taken during the trawling surveys. Possibly the marine phase of the life history does not extend far beyond the littoral zone, but the general distribution of the species appears to be circumpolar in sub-Antarctic waters, and even extends to subtropical waters in AustraUa. MYXINIDAE Three 'good' species of the genus Myxine were recognized by Norman (1937, pp. 4-7) from the Patagonian region. One of these, M. tridentigera Garman, was not obtained by the Discovery Com- mittee's ships. It is known only from the unique holotype from Magellanes. Four specimens of M affinis Gunther were taken at St. WS582 in a dip-net, but none from the area of the trawlmg surveys The remaining species, M. australis Jenyns, seems to have a much wider distribution. It is common at the Falkland Islands, where specimens have been collected by Bennett and HamiUon, on both coasts of South America and particularly in the Magellan channels. A specimen has also been collected by Hamilton at the South Shetlands, a fact of exceptional interest, as I am not aware of any other record of a cyclostome being taken south of the Antarctic convergence. Details of our obser- vations on the distribution of this species are : I : in ' Trawl + accessory nets ' Mean depth Station WS763I WS789 WS792 A WS797C Myxine australis Jenyns. Mean depth m. 84 94 104 112 Numbers I 7 I 10 Station WS812 I WS833 WS834 Numbers Station Port Stanley (J. E. H.) Salvador waters (A. G. B.) Port Stanley (A. G. B.) WS835 WS836 II: in 'Other gear' Mean depth Numbers m. ? ? ? 15 64 etc. I I I 6 in BTS I in BTS Station WS856 WS871 WS873 WS878 S3 34 32 Mean depth m. 104 338 93 (-°) 121 (-0) Numbers etc. I in BTS I in BTS I inNR 5 inNR Thus although the majority of specimens were captured in shallow inshore waters the species has been taken below 300 m., and at four of the remaining fifteen stations the depths slightly exceeded 100 m It also extended much farther north on the east coast than the other species, a considerable concentration being found off the Golfo san Jorge. The other records are mostly grouped much farther south, off the mouth of Magellan Straits and the north-east coast of Tierra del Fuego There is no obvious reason for such a discontinuous distribution, but the gear used is obviously far from efficient for the capture of such slender organisms, and as it was not obtained in l^'-f ^.^^ers any- where, insufficient sampling is quite possibly the sole cause. Apart from the probability that M australis will be found to attack useful fish, especially those that come well inshore, it is of no potential economic significance. LAMNIDAE Cetorhinus maxirmu (Gunner). No basking sharks were seen or captured by our ships within the trawling area. Norman (1937, P- V) gives the details known concerning one washed up in East Falkland and quotes HamiUon's observations on another seen off Cape Dolfin m 1936. He notes that it may prove to be distinct from the common species of the northern hemisphere. 1 See pp. 250-1 for distinction between ' northern', ' intermediate', and ' southern ' regions. 26o DISCOVERY REPORTS SCYLIORHINIDAE Scyliorhinus {Halaelurus) bivius (Smith) was not obtained at the trawUng stations. Five specimens were obtained with ' Other gear ' at three stations in the western channels. SQUALIDAE Squalus lebruni (Vaillant) is very close to the familiar S. acanthias of the northern hemisphere. From the material at Norman's disposal up to 1937 he judged that 'there appears to be only one species of Spotted Spiny Dogfish in the southern hemisphere' (Norman, 1937, pp. 9-10). Dogfish are rare on the Patagonian grounds, in strong contrast to their abundance in European seas. This would be a helpful feature if long-lining for any of the more useful species were found practicable in the future. They could of course be utilized in the same way as the northern species, but our records suggest that they are far too scarce ever to form an appreciable part of the catch of a commercial fishery. S. lebruni was not taken in ' Other gear ' ; records of its occurrence in the trawl are : WS90 I of 67 cm. length. WS791 B i of 66 cm., 1 150 g. WSg4 3 of 61, 62 and 64 cm. WS797C i of 66 cm., 1300 g. WS218 I of 64 cm. WS853 I of 65 cm., 1200 g. These scanty records suggest that the species probably ranges over the whole of the shelf in summer, and that it commonly attains a length of 2 ft. and a weight of some 2\ lb. SQUATINIDAE Squatina armata Philippi was not taken by the expedition, but Norman (1937, pp. lo-i i) provisionally identified two Argentine specimens with this Chilean species, so that it may possibly occur within the area of the trawling surveys. TORPEDINIDAE Discopyge tschudu Heckel. This species is known to range far north on both sides of South America, but was only once taken by us : WS776 103 m. 4 $? specimens RAJIDAE The general distribution and depth relations of the species of this family taken by us have been summarized, after the notes on individual species, in the form of tables and a figure dealing with all the Elasmobranchs taken in the trawl (Fig. 18, p. 276). Measurements of disk width and many weights were recorded, but owing to the scarcity of most of the species they are not sufficiently numerous to repay statistical treatment. The individual records therefore refer only to numbers of specimens taken. It is noteworthy that disk-width measurements made by Gunther and his colleagues in the field show discrepancy with some of Norman's (1937) records, owing no doubt to shrinkage of pre- served material. Rajaflavirostris Philippi. This species is morphologically very closely related to R. batis of British seas and R. stabalifons of New England. It was widely distributed over the shelf in small numbers rather more frequently in the northern region than elsewhere. In common with R. magellanica \t favoured shallower waters than other members of the genus (Fig. 18), in spite of our difficulties in sampling such a large area we can claim to have revisited most parts of the shelf at least three times so that It is a remarkable fact that this skate was never found in the same place twice. Evidently it must roam widely over the plain of the shelf (Fig. 8), and such behaviour is precisely the opposite of that which Steven (1936) found in R. clavata. The young and adolescent stages of the thornback seem nearly stationary, while even the migration of adults (Steven, 1932) is probably on a small scale DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 261 R. flavirostris was taken in the ' Trawl + accessory nets' at twenty-four stations as shown below. None was captured with ' Other gear ' : WS77 WSyg WS9S WS214 WS217 WS233 WS236 WS245 WS763 WS76S WS789 WS790A WS790B WS791B WS792B WS793 WS796B WS797C WS810 WS815 WS816 WS817B WS834 WS857 I 8 I 3 I I 2 I Of the twenty-eight individuals which were sexed, there were fifteen males and thirteen females; sex ratio 53 "5% /W« which show ^ le in the aduh stages) a movement into the North Sea in summer, but are found chiefly - deeper wat r to the north and west at other seasons (Meek, 1916, pp. 41-3)- More recently Steven (1932, P- ^o) ha shown that R. clavata appears to hatch out in shallow water, that the dispersal of young and adolescent fishi greater depths is'Lw (Steven, 1936), but that in the English Channel adults are rarely found inshore except in spring just prior to the deposition of eggs. , , . If an an logous movement takes place in R. brackyurop. we should expect our deep hauls to contain a hig'r proportion of large ind.viduals than the shallow hauls, taking the year as a whole, though the re^d might be masked by a few good catches of adults m shallow water durmg then m.grat.on. 266 DISCOVERY REPORTS Disk-width frequencies of all records from the two depth categories, shown as percentages in Fig. 13, clearly support the possibility of such a migration : individuals of the greater widths were more frequent in the deep hauls and the smaller individuals in the shallower hauls. About the main mode, however, the deep hauls show considerably greater frequencies, so that the difference is less well defined than could be wished, while the adverse factor of some migrating adults being taken on the shelf is unavoidable by this method of array. Fig. 12. Distribution of Raja brachyurops. Spring: diamonds; summer: circles; autumn: triangles- winter- squares- negative observations left blank. Note. Symbols here refer to seasons, and not fo s^™" ^ Fig. 13 gives just a hint of possible submodes at roughly 4-5 cm. intervals that may indicate year classes. Irial freehand curves drawn from grouped data by Buchanan Wollaston's (1929) method show some support for this view, which seems reasonable from the annual increment of 6 cm. established by Steven (1936, p. 614) for R. clavata, which is a larger species. Our data are notnumerous enough for us to regard the 4-5 cm. increment for R. brachyurops as more than a possibility. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES ^67 , ;^ r»„r nrpfl there were only 14^ measured individuals (of "" Iirer r sul The w d di persion. leading ,o large values for c. renders .he difference between g,ve abetter resull he WP ^^^ ,^^ ^^^^^ility of migrat.on was estab, shed, means of some 2 5 ^■''"If "" J ^*^„,di ,„ reason could be undertaken without fear of waste ^ff:": rr: sTetttmejrLSlish b^eyond al, reasonable doubt the fact of migratron on to the shelf in summer and over the edge into deep water m wmter. II I > A. HAULS OF LESS THAN ZOOM MEAN DEPTH B. HAULS OF MORE THAN 200 M. MEAN DEPTH DISC WIDTHS IN CMS Fie X. Size-frequency distnbution (percentage at each en. of disk-width) of Raja brachyurops Fig. H shows that whether we consider total abundance of ^^^^^^ ^^^"t.^ depth categories, or frequency of occurrence, the l^'^^^^:^ i:i'J':i''^,^:%,,,.^ appears to more abundant on the shelf in summer and -'^T'^-T'^^nXt^i conformity with our thesis ; 268 DISCOVERY REPORTS 80- > liJ 60 < t; 40 20 Fig. 14. A. (<2O0 Seasonal variation in total abundance of Raja brachyiirops 'over the edge' (> 200 m ^ m.) white; also seasonal variation in its frequency of occurrence B on the shelf, and black, and on the shelf C 'over the edge'. Season Spring Summer Autumn Winter Spring Summer Autumn Winter Table 5. Full data relating to Raja brachyurops /)/o«g^ in Fig. 14 On the shelf, depths less than 200 m. No. of individuals 6 62 74 17 % of total caught in each season 60 95 97 14 Over the edge, depths more than 200 m. No. of individuals No. of hauls 14 65 42 17 No. positive 3 17 18 3 positive 21-4 26-2 42-9 17-6 4 3 2 106 % of total caught in each season 40 5 3 86 No. of hauls 2 10 8 13 No. positive positive SO 10 12-5 38-5 Separate depth-frequency polygons of the type used for summarizing the main features of depth dj nbutton of all elasmobranchs may also be plotted for R. brachyr.ops at each season, as m Fig '5 This presents the evidence of seasonal migration even more clearly ^' Very good evidence that the young of R. brachyurops are hatched on the shelf can be obtained bv considering our records of captures of minute post-embryos and yearlings less than zo cm wide' The r seasonal distribution also is so circumscribed as to establish the approximate hatching peT od and they are accordingly tabulated in full (Table 6) naicning period, 3oMaThaLTMlv'lltt' 't '"" V"' T'"''''^'' ^^^ ^"^^" ^^^^^^^^ -- -"g^^ between xcfp ons I'o cm Id f T'n . ' '""'"' observations were made at other seasons. Of the exceptions, the 9 cm. individual in October was presumably a late-hatched survivor of the previous DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 269 in UJ an U) Q- Fig :5. Seasonal variation .n depth distribution of Raja Irachyurops. The widths oj the Polygons a^e P-portional to the ^ ^ percentage depth-frequency distribution when the observations are graphed m 50 m. classes. Table 6. Captures of Raja brachyurops less than 10 cm. wide Station WS87 WS90 WS94 WS9S WS109 WS76S WS797 48 51 51 WS867 Date 3. iv. 27 7. iv. 27 16. iv. 27 17. iv. 27 26. iv. 27 17. X. 31 20. xii. 31 3. v. 26 4. V. 26 4. V. 26 30. iii. 32 Numbers, size I of 4 cm. I of 5 cm. I of ;"=""=: J™ .probably the specimen 274 DISCOVERY REPORTS of C. smythii and C. callorhynchtis . The particulars for the specimens not seen by Norman are given separately : WS96 WS762B WS763 WS788 16 2 WS847B WS853 Specimens not seen by Norman : WS762B i male of 48 cm. ; WS763 unsexed specimens of 25, 30, 31 and 33 cm. and one unmeasured; WS788 unsexed specimens of 32 and 40 cm.; WS84JB one female of 90+ cm. (tail damaged), weight 4750 g. SUMMARY OF OBSERVATIONS ON ELASMOBRANCHII The main features of the distributional trends of the various species- of elasmobranchs found on the shelf are summarized in Tables 7 and 8. These show frequency of occurrence relative to the total number of hauls made in each region, and abundance of individuals in each region relative to the total for each species for all regions. With these and the data given in the Appendix any desired com- putation as to fish per hour's trawling, or per hour's positive hauls, could also be made ; but as the group is not sufficiently abundant to encourage commercial exploitation, I have not done so here. Table 7. Distribution of Elasmobranchii: occurrence in total of roughly comparable hauls in each region Species Northern region Intermediate region Southern region No. % of hauls + No. % of hauls + No. % of hauls + Squalus lebruni Raja flavirostris R. doello-juradoi R. macloviana R. magellanica R. scaphiops R. albotnaculata R. brachyurops R. griseocauda Psammobatis scobina Callorhynchus callorhynchus 3 8 2 I I I 0 6 I 4 4 ii-i 29-63 7-41 370 370 37° 0 22-22 370 14-81 14-81 I 9 4 3 7 0 0 17 I 17 I 1-89 16-98 7-55 5-66 13-21 0 0 32-08 1-89 32-08 1-89 2 7 6 8 13 4 6 26 4 24 I 2-o8 7-29 6-25 8-33 13-54 4-17 6-25 27-08 4-17 25-00 1-04 Number of roughly com- parable hauls 27 53 96 Table 8. Distribution of Elasmobranchii: regional abundance of individuals of the trawled species Species Squalus lebruni Raja flavirostris R. doello-juradoi R. macloviana R. magellanica R. scaphiops R. alburnaculata R. brachyurops R. griseocauda Psammobatis scobina Callorhynchus callorhynchus Total Elasmobranchii Total no. of individuals 39 68 23 54 7 7 292* 8 iS7t 28 697II Northern region No. 3 10 5 4 I 2 o 43 2 5 26 77f /o 37-5 25-6 7-4' 17-4 1-9 28-6 o 14-7 25-0 3-2 92-8 ii-o Intermediate region No. I 18 5 7 27 o o 53 2 93 1 /o 211' * Includes eighteen southern specimens taken in X Includes two taken in 'Other gear'. II Includes six rare specimens not tabulated. ** Includes four Discopyge tschudii not tabulated. 12-5 46-2 7-4 30-4 50-0 o o i8-2 25-0 59-2 3-6 Southern region No. 30-3 Other 4 II 58 12 26 5 7 196* 4 59§ /o 409tt 50-0 28-2 85-2 52-2 48-1 71-4 1 00-0 67-1 50-0 37-6 3-6 587 gear f Includes three taken in 'Other gear'. § Includes one taken in 'Other gear', f Includes one Psatnmobatis extenta not tabulated. ft Includes one Raja multispinis not tabulated. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 275 Fie i8 to which frequent reference has been made in the notes on individual species, summarizes the main 'features of their depth distribution. The statistical significance of the diiferences between the effective mean depths observed for the several species, which are plotted m the figure, are given '"^ iT tm be seen that the species may be grouped into three classes according to their depth relations, '^l' Shallow, species of the shelf: Squalus lebruni, Raja flavirostris, R. magellanica, Psammobatis '''".^t^.: Raja maclo.i^a and R. brachyurops. The peculiar depth distribution of the last named has been shown to be due to a seasonal migration over the shelf edge. The less common R. macloviana shows a general depth relation of the same type (an hour-glass or dumbbell-shaped polygon), and there are other grounds for supposing that a similar migration in tha species is po sible m. Deep-water species found almost exclusively over the shelf edge: R. doello-juradot, with the rarer forms R. scaphiops, R. alhomaculata and R. griseocauda. CLUPEIDAE Clupea fueoensis (Jenyns). This is one of the most numerous and important species of our area, but ^C^l main y pela ic habits and small (' Pochard ') size, it could not be adequately sampled wi h ol ge For this reason I give only the list of stations where it was captured. The numbers mean 3i tie. The Falkland herring is a most important forage fish for l-g-/P--;. ^^f^^^^.^tr' anZi also extensively eaten by birds and seals. It is possible that it could be utilized directly for human ood-it is most excellent eating, as the Falkland islanders well know-but prospects of regTr a^- m numbers sufficient to support, say, a small canning factory are not good. Any fofm of gm net would be continually threatened by seals and penguins anywhere wi hin reasonable SllLce of the Falkland Islands, while the herring trawl would almost certainly fail on the rough bottom Some specialized form of purse seining might possibly provide an answer. Some other clupeoids recorded from the Patagonian region may be mentioned here, though we d.d not cTotu e any of them and they are probably not normal inhabitants of the area surveyed. The ; c es rr^ownTchU "; ' sardina' was Lnd by Norman (:937, P- 38) to be quite di^mct from the vTa ly ealrn C.fuegensis. Norman has named it C. berrtincki, after Mr Cavendish Bentinck, who sIntHm a tne series Lm Talcahuano harbour. From Col. Tenison's drawings it can at once be Te^n Z I is much deeper in profile than the species so common on the continental shelf to the east. It resembles typical northern hemisphere herrings much more closely. Gunther has a note that Clupea sp. (? maderen.s) was on sale at the -/^^ ^ ^^^ t"; f Itfs u f oAor^T^rnv This soecics he savs, s represented farther south by C. /z^c^ewm. n is smaller than C. fuegensis, of ' sprat ' as compared with ' pilchard size. ^ We obtained specimens of C.friegensts at the following stations, mostly in the accessory S wis WS^So A TS8%B WS^3S ^fi4 ^St Ws4b WSS^S S^f/,(N:ooB) Also: Stanley Harbour, 2. .i. zj and Chartres River, West Falkland Islands, 13. iii- 33- 276 DISCOVERY REPORTS fe 'So t^ DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 277 S •5^ s a c o <^ 'S E .£ S 3 -4- c O + 1 +-c« Cfi +CAJ + Ci5 ^C« +Cfi +CJ5 ^.Cfi +'cfi 2 a oi- e«)' 0) ^ e V + o ■^f. ^ 03 . O tjo ;«5 rr, 6J) "^i . 2. • o f^ 60 "^ o« - "~>.oo ^ 7^ +z '^ 1 ^ + 00 ») 1^ ro 6a fv o<| 7^5 7<^ A CO -a «J ^ e (U ^4-H C a ?s + 1 1^ tQ.a^ t^-^ f^.^ 1 Co I CO +cfi 1^ +z +1 '^ +z I CO 7^ .'^ s "fe se s a -S ■^^ a « 3 ■^ O^ CO ^ Qi Q> 278 DISCOVERY REPORTS Cliipea arcuata Jenyns. Our gear was even less suitable for sampling this small coastal but mainly pelagic species, which is very closely related to the European C. sprattus, and which is known to be abundant at times within the area surveyed. In life the muscles show up yellow through the skin, permitting rapid distinction from small fry of C. fuegensis of similar size. Doubtless a frequent food of larger fishes, we only happened on it in numbers at St. WS8g, where thirty-nine were taken in the N7-T attached to the back of the trawl. This station was worked in very shoal water close in to the north-east coast of Tierra del Fuego. Quite possibly the normal habitat of the species is too coastal for it to have been taken by the fine ' accessory nets ' on other occasions. GALAXIIDAE Galaxios attemiatiis (Jenyns). This fish was not trawled by us— it is improbable that they ever depart far enough from the coast to be sampled by a trawl— but a Falkland specimen was readily obtained for Norman's report of 1937 by Mr Bennett. The majority of the galaxiids are fresh-water species, though an increasing number have been found in the sea. Most of them can probably be regarded as anadromous. G. attenuatus is catadromous, descending to brackish water or to the sea itself to spawn. It is possible that much of the spawning takes place in the lower reaches of estuaries rather than in the sea itself (Phillips, 1924), but it is certain that the larvae must be widely distributed in the sea, or the species could never have populated the smaller brooks in which it is found, for many of these have no estuarine transition area at their mouths. Moreover, there is an interval between the time of known spawnings and the records of upstream movements of larvae. A species of Galaxias unknown m fresh waters, G. bollansi Hutton, was described in 1899. Very recently, Scott (1941) has shown that apart from G. attenuatus (which he studied in very great detail in an earlier paper there quoted) the type species of the genus, G. truttaceous Cuvier, is euryhaline, and that 'it is not improbably facultatively catadromous, and that when not confined in land-locked waters it may retain the pre- sumably primitive spawning habit of the family'. Since Scott had already shown (1938, pp. 125-6) that young G. truttaceous may arrive with the upstream spring immigration of G. attenuatus, that adults descend (1941, pp. 57 et seq.) to almost completely saline sea water, and that when ripe they are found at or near the coast (p. 68), his conclusion seems most cautiously worded. I cannot understand why the catadromous migration of G. attenuatus was ever doubted Beginning with Hutton in 1872I the facts had been described by several observers in New Zealand, Tasmania and Australia, who were familiar with the fish in fife. It is true that McCulloch (1915) had had to improvise a very primitive experiment in the endeavour to prove his point, but there is nothing to show that It was not effective ; and Meek's assertion (1916, p. 147) that the 3 J-4 cm. larvae are denatant IS at variance with McCulloch's direct observations. It is, indeed, probable that the catadromous migration of G. attenuatus was well known and understood by the Maori— who still eat them— before white men ever came to New Zealand. Phillipps (1919, quoting Best, 1903) was doubtful of one tradi- tional Maori account that described spawning at the mouths of rivers. He then thought that the spawning was entirely marine, but within five years his own observations (Phillipps, 1924) had shown that the Maori account was substantially correct. It may seem unnecessary to labour the point now but G attenuatus is an important fish in New Zealand, where the ascending fry are captured and canned as whitebait', at one time the only fishery product exported from the Dominion Moreover visiting the Dommions twenty years afterwards in a research ship, one found that biologists 'from home still tended to be weighed on their willingness to accept enlightenment on the subject The general distribution of G. attenuatus is of particular interest; it is known from southern Australia, Tasmania, New Zealand, both coasts of Patagonia and the Falkland Islands. The family ^ I have not seen this paper. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 279 Galaxiidae has a marked ' circumpolar sub -Antarctic ' distribution, but the other, mostly more truly lacustrine, species have localized distributions, and only G. attenuatus, known to be catadromous, extends across the great ocean barriers. r ^ .. , .« According to PhiUipps and Hodgkinson (1922) the older lacustrme stages of G. attenuatus are generally known as 'minnows' in New Zealand. The young fry eaten as whitebait are very good. All the whitebait examined from Auckland market were found to belong to this species.^ They were on sale in August, September and October. In the Falkland Islands G. attenuatus is one of several fishes locally termed ' smelt '. They sometm.es shoal in littoral waters, probably when spawning, or later when re-entering the rivers. Their excellence as food is already appreciated. Adults are said to attain a length of 6 in. which is consistent with Scott's voluminous biometric data from Tasmania. This is undoubtedly one of the most promising species for small-scale local exploitation. Owing to the small size of the rivers m the Falkland Island he scale of operations would have to be kept down, or rapid depletion would result. The stock cannot be a large one for there is not enough fresh water to maintain large quantities of the maturing adults. ' 1 :r .^c./... (Jenyns), which occurs m the Falkland Islands and on both coasts of Patagonia, was not obtained by us. It is known chiefly in streams and brooks, but also occurs in the sea (Norman, ",7 p 40 In the Falklands it is known as ' trout', but this local name is also given to AplochUon II; 'it' noteworthy that the larger galaxiids of southern Australia, Tasmania and New Zealand were al o called 'trout' by the settlers until the artificial introduction of true salmonids from he were also calleat y observation in the application of the ra^::t ' 'o3' nle^oTui:: rtTtely related fishes m some instances ; while m others, also dating from the great period of human expansion, no discretion at all has been used. APLOCHITONIDAE AMuton ^ebra Jenyns. No specimens of this fish were obtained by the expedition, but Bennett Aplocfuton .ebra jenyns p forwarded it to Norman when the latter s found a specimen which had been '^^ ^^^ ^^/^^ ^9 ^^^^^.^^ maculatus, report was being written, ^^^^^l^^ .^J^^^l ^^^^^ might indicate. Indeed. but there is no ^^^^^ ^^l ^^^^^^^^ ,he F^land Islands, Norman (x937, P- ^37) It was probably one of the first fishes ever o ^^^il Darwin brought Jenyns in 1842. SYNGNATHIDAE Upto^,us min.:iUanu. (Eydoux a„a Gerva,s). Th,s pipefish -Z:^^;:':^l\tr t:^::::^Z l^^^XXX^^^ 'ee„ Unown for J. cen.u^, .he • WUia™ Scoresby • also ob.a.ned a specimen in a tow-net, at St. WSS93. MACRURIDAE "'" TT'ZZZ'C^Z^^^^^ in co„siae«b,e depths is possible. I. ™y ™:fd fl:: nonh, bu. *::: *e s.eep„ess of .he slope p.ecluaes aae,ua.e exp,o.a.,on a. su,.ab,e . T™.,, wh„*.i, c„„ls, ,.gd, of Voc.i.o„ld.e wi,K G. .,-™-.. .nd »,n...™. O. ..-« (ScC, ,,36). 28o DISCOVERY REPORTS depths. Out of ninety specimens, two only could have been taken in less than 150 fm., and none in less than 272 m. The only large catch showed a high proportion of males, but at WS821A (ten individuals) the entire catch consisted of females, which generally preponderated in the small catches: 2?? WS82O I? 2$? WS821A 10 ?? 3?? WS839 69: 54cJiJ, IS?? 2?? WS84O i^ Coelorhxtichus fasciatiis (Giinther). All our specimens of this species were obtained in summer and autumn (third survey) in deep water over the shelf edge, mainly in the Falkland trough. Out of 140 specimens captured one only could have been taken in less than 200 m. of water. A notable prepon- derance of females was observed in three out of four catches of fourteen or more individuals. This may be due to the higher escape ratio of the smaller males, but abnormal sex ratios were also observed in Coryphaenoides, and it may be that there is a tendency towards sexual segregation among shoals of fishes belonging to this family at certain seasons. It is possible that the geographical range of Coelo- rhyfichus fasciatus extends considerably farther north. Owing to the steepness of the descent from the shelf there our chances of trawling in suitable depths were extremely limited. It is noteworthy that this smaller species, though found exclusively over the edge like members of the family everywhere, favoured slightly shallower depths than did Coryphaenoides. From Phillipps (1921) we learn that Coelorhynchiis australis, called 'javelin-fish' in New Zealand, is occasionally taken by trawlers in Golden Bay, and highly esteemed as a food fish. It may therefore be that the usual dumping of macrurids, as ' rattails ' among the rubbish of trawl catches, as is general off the British Isles and off" South Africa, is a needless waste. If Coryphaenoides is similarly edible it would be the more valuable of the two Patagonian species on account of its larger size — up to 87 cm. as against 38 cm. in our catches: WS817A I WS819A 3 (2 3o) WS829 7 (all 66) IVS817B 57 (4 o^.^) WS819B 5 (4 o'^) WS840 I IVS818A 15(20^0^) WS820 14(11 'SS) WS870 8 IVS818B 8 (6 ,^0") WS821 A H (I S) WS875 7 MERLUCCIIDAE Merluccius hubbsi Marini INTRODUCTION: ECONOMIC IMPORTANCE OF ALLIED SPECIES Merluccius hubbsi Marini, the Patagonian hake, is the most important and one of the most numerous fishes of the region. In our trawling catches it ranked third in numbers to Notothenia ramsayi and Macruronus magellaniciis, but the weight of Merhicciiis captured exceeded that of the other two species combined, and formed 47-3 % of the weight of all the fishes caught at eighty-three stations worked during the third survey for which weight data are available. All the true hakes are very closely related ; there is, indeed, still room for doubt as to whether some are specifically distinct. Their distribution raises many problems of great general biological interest which I hope some day to discuss at length elsewhere. For our present purpose it is sufficient to note that the Patagonian species is known to range from the southern coast of Brazil (probably from about the point where the Brazil current begins to swing off'shore) to the neighbourhood of the eastern entrance to Magellan Straits. From Norman's (1937, p. 46) diagnoses there appears to be no doubt that it is specifically distinct from M. gayi (Guichenot), the common species of the west coast of South America, with which it was for long confused. Indeed, M. hubbsi in its general bodily proportions resembles the silver hake or 'whiting' of New England and the north-west Atlantic, M. bilinearis DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 281 (Mitchell), more closely than it does any other member of the genus; but here again Norman's scale counts point to the specific distmction being justified. (Compare Col. Venison s figures of M^^a>. and M hMsi (Norman, 1937, P- 46), with the figures of M. productus and M. bthneam by H. L. Todd in Goode's Atlas, 1884, pi. 65.) ,,,.,. • r 1 ^u All the true hakes are edible : M. merluccius of Europe and North Africa, M. bibneans of the north- west Atlantic and M. capensis (Castelnau) of South Africa are already the staple of important fisheries^ M productus (Ayres) of the north-east Pacific has been little utilized, and the same may be said of M australis (Hutton) of New Zealand; but so long ago as 1907 the British Columbia Fisheries Com- mission reported that M. productus was not inferior to the Atlantic species, and it would seem that Therneglecr^ due merely to the plenitude of better food fishes in these favoured regions. Mgayr of Chile is captured and eaten locally, but its exploitation by modern large-scale trawling methods is rendered impossible by the absence of any continental shelf on the west coast of South America FinaUy the Pa^agoman species M. hubbsr is captured by the small trawling mdustry operating from the mouth of the River Plate. This originated with Don Pedro Galceran in Montevideo, but the U urayan enterprise failed and the small-scale industry was then carried out from Buenos A. re (DevTncenzi 1926). In 1932, when Gunther visited the fish market at Buenos Aires, he found tha '^eriuv^'ere selling at o 50 pesos per kg. The Buenos Aires trawlers are not known to have operated Zt^Z lile so far afi'eld as th'e area we surveyed. Up to the time of our last survey i^,^^)r. was said that thev rarely proceeded out of sight of the land at the mouth of the Plate. In orL to IsLs the potential value of Patagonian hake we may briefly consider the history of th expbitatn of the three'species of Merlucaus that already provide the raw material for considerable The European hake, M. nrerlucaus (Linnaeus), ranges from the Norwegian Rinne southwards along the edge ofThe continental shelf as far as Dakar on the Mauntanian coast, and perhaps even farthe south I local race has been reported off Cape Verde (Belloc, 1937)- Belloc's observations on vertebral rlrt !how that, while the previously known stocks of European hal^e/-^"-^ -- m the west of Ireland, had 'vertebral numbers' regularly increasing from 50-48 (±029) o 51 ^5 lo6rthe number or the large sample he obtained off Cape Verde was 54-09 (±o-3) Ooc cit., fi^^TWsT of exceptional interest because increased numbers towards the northern end of the fig. 3)- ^his IS ot exceptioi .0 . ^;j, .o^^a and herring (numerous workers, quoted by 1 r.;^c Tf Belloc is right n maintaining that the Cape Verde haKe are uui d , , • • '^^ spe -t-ana L . e.pHa.c *a. a,, ^ ^^^^^^^^Z:;^!::::^ -"^ '^^::X::t:z:^^t!:^"^^-^ a.f ... a„a *. ' T ' L r o b Mlowed by the other stoeks of European hake for which data are ava.lable. r rrrknutth detailed observations on vertebral numbers of Merluum wuh the pre- LZs aXfinle^'that Ford (,,38) has shown to be desirable, cou.d not fa„ .0 be of exceptronal '"xhThake is contnron in the Mediterranean especUl, on t^.enorth«n shore whe.^^^^ . nrentioned ,n the literature' from the t.nre »' Ar.s.o.^ Cou h .86 . v ^ ■ ■ • P^ J ;__^ l^ ^^^ ^^^ rpo;tL':tcid;:—i^^^ 282 DISCOVERY REPORTS principal fish captured at Fiume in 1879-80, ibid. p. 167), also mentions that the superiority of hooked over net-caught hake was so marked that they could command an appreciably higher price. In the north records of M. merliicciiis from Iceland have been confirmed, but some from south- west Greenland are more doubtful. Even at Iceland such stragglers are rare and represent the extreme range of the species. Hake are not common in the North Sea, though there is a regular small-scale immigration into the north-eastern portion of it. In fact the fish, which lives for a good part of the year over deep water oflF the edge of the continental shelf, is essentially an inhabitant of warm tem- perate waters from the west of Scotland southwards. Its southern limits in all probability are normally defined by the limits of influence of the Canary current, which is cold relative to the tropical surface waters south and west of it, and relatively rich in nutrient salts and all the larger forms of life (cf. Hentschel, 1936, p. 243 and Beilage ix). M. merlucciiis, the merluce of heraldry, has been an important constituent of the fish food of the western European nations throughout historic times. In Britain it has been the subject of various commercial treaties from the time of King John to Queen Mary. Much was eaten during Lent. Latterly it fell into disfavour, partly perhaps because the disestablishment of the Church Ted to a less rigorous insistence upon traditional Lenten fare, but chiefly because improved boats and gear enabled fishermen to catch far greater quantities of the choicer fish than before. In France, Spain and Portugal It must always have been relatively important, owing to the lack of the colder-water gadoids near at hand; but as recently as the second half of the last century we find leading ichthyologists in this country dismissing it as a poor, coarse fish, of inferior table qualities. It is said, indeed, that the German and Dutch names for hake, Stockfische or Stokvische (there are others), derive from the habit of lettmg smack's boys keep them for ' stocker ', which makes it certain that they were practically unsaleable. The rapid development of otter trawling, especially the introduction of steam trawling, at the turn of the century saw a rapid decline in the proportion of prime fish landed (though of course the actual quantity was at first increased), and the development of fried-fish shops greatly stimulated hake trawling. Indeed, Hickling (1935a, pp. 70-1) has been able to show that there was serious depletion of the stock through overfishing before the war of 1914-18. During that war all fish stocks recovered to some extent, and with improved boats and gear (the V.D. trawl) and the development of deep-sea traw mg on distant grounds throughout the whole latitudinal range of the species, British hake trawling showed increasingly heavy catches until well into the 1920's. But the catch was only main- tained by increased fishing efl^ort, as Hickling (1935., pp. 74-5) has so clearly demonstrated. Since 1925 catches declined and depletion continued until 1939. The importance of hake to the modern British trawling industry is very great. It is the staple catch of our great west coast trawling ports, Fleetwood, Milford, Cardifl^ and Swansea. Even in 19^0 when decline of the stock was beginning to show itself, it ranked third of all our trawled fish whether reckoned by quantity or by value. Only cod and haddock were more important. During 1920-5 the British landings averaged 38,500 tons, and some of the fleet were working as far south as North Africa sZ^r' ^ V r' r'""^" P'"''^ '^' ''''''''^^ °f fi^h^^g by French and (latterly) modern Spanish vessels, in the southern part of the range of the species, also greatly increased. Though the c^;ra"fi::::::rtrpeX^^^^^ rp, ., ,,,<,. y : 920 33 was ^1,532,000 per annum (Hickhng, 193 5 «, p. 74, fig 7) to No^th cllt T 7' ' t ""T ^r^^^'^'^' ""■ ''''''^'' ("^^^^^''^)' -"g- f-- Newfoundland to No th Ca olina. To the south it has been recorded (from deep water only) as far as the Bahamas and Florida (Longley and Hildebrand, 1941, p. 38). Panamas The history of the commercial exploitation of M. bilinearis aflfords a close parallel to that of the ■ DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 283 European hake, but with a characteristic acceleration due to the American fishermen having a totally undepleted stock to deal with when modern methods of capture and treatment of the catch were introduced. In dealing with this species it is important to realize that much of the earlier American and Canadian literature is vitiated by confusion between it and several species of Urophysis (Gadidae which are called ' hake ' with or without distinguishing prefixes along those coasts. The once important bv-product trade in ' hake-sounds ' for isinglass was based on these, and not on Merluccius. In the early days silver hake were regarded as rubbish by the New England fishermen, and also as a great nuisance when large numbers were caught in mackerel nets (Goode et ai, 1884 pp. 240-3) and had to be discarded. Their inshore migration in summer appears to be even more marked than that ot the European species, and they frequently become stranded in pursuit of prey. At times they were used for manure (Bigelow and Welsh, ig-S, PP- 386-96), and as recently as 1895 only 37,000 lb. were marketed from Massachusetts and Maine. By 1919 more than 14 million lb. were sold, and even a that da e most were still caught in traps and weirs, the price being too low for the offshore fishermen to bring m those they captured. Since then the demand has increased enormously with the development of large cold stores, especially in the mid-western states. From the Statistical Digest No. 4 of the U.S. Bureau of Fisheries (i94o)-a mine of fascinating information-it can be seen that nearly 50 million lb. of M.^SLn. were marketed m that year, more than 80O/. of the total coming from the New England States. Nearly half this catch was frozen (forming 11% of the total frozen fishery produce of the country). It has, indeed, become the most important single species of the frozen fish trade. This rapid develo ment ha; synchronized with a big change in fishing methods, more being ^^^^^^^ZZ^r than in pound-nets now, except in New Jersey and Rhode Island. One wonders whether the over fishing problem may not soon become evident with the New England whiting ; M capensis (Castelnau), which is known in South Africa as stockfish or stokvisch has been the stapk of the modern trawling industry developed in that country. In the absence of adequate data ^ets stm some room for doubt as to whether the distinction between this and the Eu-p^anha -s sufficient for M. capensis to be regarded as a separate species (Barnard, 1925, P- 3^0). The bouth Afric n form ranges from Angola to Natal, in deeper waters towards the equatona limits of its range on dthe side of L continent. It may be noted that this seems to be a general rule for all species o rltU and for very many other fish besides. There are no statistics ^J^^^/ ^f ^^^ ^^^^ l^nrlinps to enable one to trace the growth of the modern fishery at the Cape, but the figures tor Tat; uH^he^ by von Bonde (^934) leave one in no doubt as to the i-P^^-;^^^^^^^^^^^^^^ 5n that period the catch was more than one-third of the total weight of fish landed: some 7850 tons nf ctnrkfish worth ovcr / 1 ^0,000 per annum. r ^^ 7 Sorrjes relating .0 the siz^ and value of the fisheries for these three speces of S^^-^"-"-" sumrrized n Table ,0 below. They reveal truly astonishing difTerences tn the value of the fi h ,n hecTu 1 concerned. The British catch used to be nearly double that 0^*= An^encan wh,,.„g fishery but its value at first sale was nearly eight times as great, weight for we.ght. The South A rtcan ctchw- about 40% of the weight of the American catch, with the value at first ^k micrn,.d,.v. bettrthTprici obtained in the other two countries: about 3! times the value °f *e A^- and tr,hrhai?the value of hake landed in Britain. While the difference between Bnt.sh and South A i an prict m yTn part be explained by the employment of cheap coloured labour tn South Afnca fhr"ch chlper Amertcan prices point to profound differences in eeonomtc cond.t.ons wh.ch cannot be understood without first-hand knowledge of the ' whnmg trade. Twould seem therefore, that while it is obviously desirable to exam.ne our b.ologtcal dat con- It would seem, me Knowledge accumulated concernmg the speces already 284 DISCOVERY REPORTS by direct biological comparison alone. Peculiar economic and even political considerations — in fact, the bionomics of the prospective human producers and consumers — play at least an equal part in the determination of industrial possibilities. At the same time, some knowledge of the natural history of the Patagonian species is obviously a basic necessity in any attempt to deal with the problem. Our data are therefore considered here from the biological point of view, while suggestions as to com- mercial prospects are deferred to a later section of this report where the summarized weight data allow one to take into account the possible value of other less important species. Table lo. Figures illustrating the size and relative values of the principal fisheries for Merluccius spp. Country, period, species, source of information Approximate average annual catch in tons, cwt. or lb. Average price at first sale per unit of weight, in sterling or in % Great Britain 1920-33 Merluccius merluccius Hickling (1935) United States 1939 and 1940 Merluccius bilinear is Statistical Digests, i and 4 Union of South Africa 1929-32 Merluccius capensis von Bonde (1934) 36,904 tons or 738,088 cwt. or 82,666,000 lb. 19,767 tons or 395,335 cwt. or 44,277,500 lb. 7845 tons or 156,900 cwt. or 17,572,800 lb. £4,1. 6s. i^d. or I171.49; £2. IS. 6d. or I8.62 ; 4|^. or nearly 8 c. ^4. i()s. 2ld. or I20.62; 45. \i\d. or $1.03; \d. or just under ic. l\(). gs. 2,d. or $79.15; igs. o\d. or $3.95; zd. or 3|c. THE SIZES OF PATAGONIAN AND EUROPEAN HAKE COMPARED First, it is desirable to establish the sizes of the hake we captured and to see how they compare with the sizes of better-known species captured with similar gear. If it were possible to obtain ' ideal samples ' of a slow-growing fish like hake, whose length increases almost as a Imear function of age, the length frequencies, when plotted graphically, would approximate to the Ix curve of a life table (a curve like a left-handed ogive, but with the curvature reversed and enormously produced at the very beginning, because of the high infant mortality rate). An imaginary curve of this nature is shown by the solid Ime in Fig. 19. It would begin with astronomical numbers of newly-hatched larvae, and end with the largest hake caught. Departures from this curve would, in part, be due to slight changes in growth rate. Such fish grow rather faster than usual early in life and slower near the end of their lives. The straight line of the age-length relationship becomes slightly bent at the ends, approaching a parabola. But the ideal length-frequency distribution would show other more important deviations about the smoothed curve if considered with small class intervals (say i-ocm.): there would be modes or shoulders whose magnitude would reflect the difl^erential survival of hake hatched in successive years the resultant of all the environmental factors, animate and inanimate, that influenced their lives An imaginary distribution of this type is shown by the pecked line in Fig. 19. Hake have a prolonged spawning season, and the changing environmental conditions would not have favoured early- and late- hatched fry equally in successive years. This would be a further source of variation in the modes or shoulders in our imaginary curve, for it is their dispersion that helps in age determination, and this even more than their magnitude, may thus be distorted. In actual hake samples this factor so com- plicates the length-frequency distribution that Pettersen's method of age determination is rendered unsatisfactory except for the younger (smaller) fish. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 28s 50:- 40^ 202- \oZ- 0 r 0 10 In practice samples of a hake population are modified by the catching power of the trawl used. An ' idealized ' trawl, that captured all of the fish above a certain size in its path and allowed all the others to escape, can be imagined to catch samples showing a length-frequency curve of the same pattern as our ' ideal ' sample, but with the point of origin at (say) 40-0 cm. instead of the length of the newly-hatched larvae. Unfortunately such ideal conditions do not exist, and the selective action of a sampling instrument such as a commercial otter trawl of normal mesh retains a varying proportion of the smaller fish below the size at which all are captured. The proportions of these smaller fish retained varies in regular fashion according to their size, and follows yet another ogive curve, the ' selection ogive ' of the trawl used. Selection ogives for trawls of diflFerent mesh taking European hake are given by Hickling (1933, p. 71, fig. 38). It will be seen that as the size of the mesh is increased and the selection ogives become centred higher in the length scale the curves slope less steeply. Theoretically it would be possible to obtain length-frequency samples approximating to a normal distribution by increasing the size of the mesh. By covering an ordinary trawl with shrimp netting, Hickling has been able to obtain samples which, when treated cumulatively over five annual surveys of the grounds to the south and west of Ireland, show length frequencies approxi- mating to the smoothed curve one would expect in ' ideal ' samples for hake upwards of two years old. No practical gear could sample the younger length classes simultaneously, for the minute fry are pelagic, and the yearlings do not inhabit the same grounds as the older fish. Curves of percentage length fre- quency of European hake, derived from Hickling (1933, table IXa), are shown in Fig. 20, where they may be compared with the age distribution (a partial 'life curve') upon which that particular series of mean lengths was based. These data of Hickling's also give us a valuable demonstration that the numbers of the two sexes of European hake are nearly equal, the actual ratio from that set of figures showing a slight preponderance of males. If samples such as can be obtained with a commercial trawl are considered, it is found that females preponderate to a considerable extent, owing to the higher escape ratio of the smaller males. Unfortunately, such data as we have for Patagonian hake are not directly comparable with those for the European hake set out in Hickling's table IXa, but they do compare very well with mean values of the figures set out in his series of measurements on commercial trawlers, of European hake from areas 1-5 (Hickling, 1933, tables XI a, XI b). These areas cover that part of the European hake's geographical range most nearly comparable to the areas best sampled by the ' William. Scoresby ' when \\. ^^ 80 90 100 110 Fig. 19. Imaginary length frequency curves for an 'ideal sample' of hake. Solid line: smoothed curve; pecked line: with small class intervals. 286 DISCOVERY REPORTS catching hake off Patagonia, and direct comparison of the two sets of length frequencies seems highly instructive. These data of Hickling's are given in the form of percentage length frequencies for each month, irrespective of sex, over a period of more than two years. To facilitate direct comparison with our southern data the results have been recalculated to exclude the very small number of fish less than 20-0 cm. long, and then meaned. For the southern species we have 4704 measurements of hake caught with a commercial trawl at all seasons, between 42 and 53° S, the latter probably being the normal southern limit of the species. In addition to the calculation of the mean, standard deviation, etc. , by the long method, they have been secondarily grouped into 5-0 cm. length classes, and the percentage length frequencies computed, so that curves could be drawn for direct comparison with Hickling's data. -1 30 so 60 70 LENGTH _ CMS u In IbtItI ytIyhImiI kIx Ixr Ixnl YEAR CLASSES Fig. 20. Percentage length frequencies, and a partial life curve, for Merluccius merhicciiis. From Hickling, 1933, table IX a. Before this comparison is made, it is necessary to make clear one most important difference between the two stocks : the males of the Patagonian hake are relatively much smaller than those of the European hake. It will be shown that Patagonian hake are some 5-0 cm. smaller than European hake, on the average, when the sexes are lumped together; but although directly comparable data for the separate sexes are not available, it can be shown that males of the Patagonian stock are relatively much smaller than this difference would indicate. Means for each sex of European hake may be obtained from Hickling's table IX a. While the means for the Patagonian stock are not directly comparable with these, the difference between the mean lengths for each sex, in each set of data, may readily be com- puted; and the difference between these differences is probably significant, though this cannot be established statistically. For Patagonian hake the difference is 49-9 cm. for females less 36-4 cm. for males, i.e. 13-5 cm. For European hake, with data including a much higher proportion of the smaller individuals of both sexes, it was 34-9 cm. for females less 31-0 cm. for males, i.e. 3-9 cm. Moreover, if we take the largest decile of the 1396 Patagonian males, we find that they show a mean length of 46-1 cm., the largest individual was 64 cm. long, and only i % of the total of 50-0 cm. and over. The largest decile of the European males has a mean length of 56-1 cm., while the largest individual age group had a mean length of 767 cm. ; and some 7% of the total were 50-0 cm. long and over, in spite of the much greater proportion of small individuals in this series of measurements. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 28? From these considerations it would seem that if we assume the disparity in size beUveen the sexes to be at least twice as great in Patagonian hake as it is in European hake, we shall not be far wrong. Turning back to direct comparison of the lengths of the two stocks irrespective of sex, the figures for European hake being derived from Hickling's (1933) tables XIa and Xlb as previously stated, we obtain the curves of percentage length frequency shown in Fig. 21 a. This shows higher frequency ot the European species in the higher length classes; but since all frequencies m these higher length classes are low this difference is much better illustrated if the results are plotted on an anthlog scale, as in Fig 2ib\t must also be remembered that Hickling has shown that the figures for the European species were derived from a heavily overfished stock, in which the proportion of large fish had been seriously depleted, so that the smaller size of the virgin Patagonian stock is even more marked than can be shown from these figures. • r ^u Proceeding to direct comparison of mean lengths it was found that the secondary grouping ot the southern data leads to an error of +0-5 cm. Further, the secondary grouping m itself reduces the significance of the difference between the means, for using percentage length frequency as the basis, iV becomes 100, and the value of cr, is grossly exaggerated. It is not surprising, therefore, that on comparing the means of the two sets of percentage length frequencies and applymg the test rf/a,>3, their difference cannot be shown to be significant. Thus : European: Mi = 5i-4, ffi=i5-475> (?f = 239-476, Ni=ioo; Patagonian: M2= 46-4, (72=12-535, al=i57-i26, A^2=ioo; J=Mi-M2=5i-4-46-4 = 5-o. d 5-0 . — = ^ — = 250. Orf 1-99 If, however, we take the true values for the southern species, calculated by the long method from all the individual observations, we have : European (M. merluccius) values as before ; Patagonian (M./?M6fc«):M2= 45-9. 0^=12-^2, al= 154-33. iV2 = 4704; J=Mi-M2=5i-4-45-9 = 5-5. ^ __ /(4+4U /(219j476i54l3_3\ ^2-42757= 1-56, °^~VliVi N2' V\ 100 ^4704/ Oa 1-56 ^^^ Here the difference is clearly significant. Moreover, if the number of individual observations upon wHchtrln of Hickling's means depends were known. N, would be much larger, and the signifi- rnnre of the difference even more marked as Oa diminished. i„„„tl. nf Even ift assume that the unknown grouping error introduced by ^^'^^'^'^f^^l^TJ^^^^^ the European species from the secondarily grouped percentage frequencies was of h s^me magm^^^^^^^^ and of opposite sign, from that which was found for the Patagonian data so treated, we Mi=5o-9, Patagonian values as before, ^=5-° and ad=i-63, 1=. 1^ = 3.07. Oa 1-63 -* j88 DISCOVERY REPORTS 50 50 70 ao LENGTHS _ CMS 90 100 110 'i';:cSrc£Sx-ijs='t^t3J:sr,iSi^^^^^ 60 70 80 LENGTH5_CM5 Fig. 21 J. The same as Fig. 2i«, on an arithlog scale. 90 100 DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 289 It therefore seems quite clear that Patagonian hake really are some 5-0 cm. shorter than European hake, and that the disparity would be even more marked if it were possible to compare the males separately. Another interesting point is brought out by comparison of the length-frequency curves: Patagonian hake, despite their smaller ' average' size, are relatively more numerous in the 50-60 cm. length classes (to the right of the mode) than are European hake. The sharp inflexion above the mode is doubtless due to scarcity of males over that length; but it can be seen that the succeeding flattened portion of the curve is on a distinctly higher level than the corresponding part of the curve for European hake of about 5 cm. greater length. The 50-60 cm. hubbsi—^hont 'ordinary chat' size by Fleetwood market standards of the period-probably correspond physiologically, if not in actual age, to European hake one category larger; so that it seems to me that we have here a striking indirect confirmation of Hickling's proof that the European stock was heavily overfished. THE DISTRIBUTION AND RELATIVE ABUNDANCE OF MERLUCCIUS HUBBSI WITHIN THE AREA SURVEYED, AND THE EFFECT OF LATITUDE ON NUMBERS, SIZE AND SEX RATIO The general distribution and abundance of this species cannot satisfactorily be shown on charts of such scale as could be reproduced here, because of its widespread occurrence and pronounced migratory movements. These are essentially similar to the migrations of the European and north-west Atlantic species, as will be shown in a later section. The occurrences of the species at our trawling stations have been tabulated hi exienso in Appendices II A, B and c. From these it can be seen that the most important feature of the general distribution is a marked decrease in relative abundance from north to south. Local concentrations were encountered inshore in autumn (St. WS853, Wb855), offshore in winter (St. WS216, WS217), and at intermediate distances from the coast m early summer (St WS7Q0 WS791). There were, of course, numerous less pronounced concentrations, most ot which tended to conform to the general pattern of migration suggested by the extreme examples quoted. . , .... Now since hake tend to be more closely congregated when on their inshore spawning migration, and the larger fish tend to move inshore first (cf. Hickling, 1927, P- 59. on the European species), a series of observations in early summer might give an erroneous impression of the effect of atitude on the size of the fish. Spawning takes place earlier in the year in the more equatorial part of the range of the species, and so considerable concentrations of individuals larger than the average for their latitude may be sampled when fish in higher latitudes are not so concentrated. At the same time, moderate numbers of the smaller hake, which do not seem to migrate so far or so fast, can nearly always be found in relatively shallow water throughout the year. In Merlucaus Imbbstj. found that the resultant of these factors completely masked the effect of latitude upon size of fish in December, but this effect was quite clear when the data for all seasons were considered together. A small series of observations in a single longitude taken over a narrow time interval later in the year, when the sma ler individuals were at the peak of their inshore movement, also showed the effect of latitude q-^e clearly^ The general decrease in abundance of hake from north to south of our area is deinonstrated by the figures in Table 11, which are taken from eighty-three hauls spread over the whole of the third survey. Earlier resvdts are m agreement with these, but are not considered here because comparable weight data are lacking : ^ 290 DISCOVERY REPORTS Table ii. Decrease in relative abundance o/ Merluccius hubbsi with increasing latitude Northern region 42-46° S Intermediate region 46-50° S Southern region 50-54° s Number of hauls Hours trawling Mean number of hake per hour's trawling Mean weight of hake per hour's trawling 14 23 100-48* 54-817 kg. 29 44 20-59 19-399 kg. 40 60 6-67 7-464 kg. * Obviously fractional hake could not long exist in nature, and it is mathematically indefensible to treat fish as indiscrete objects— but I feel that the fractional expression is less misleading than giving results as hake per 100 hr. trawling in order to get whole numbers, because these results are based on less than 100 hr. of comparable hauls-nn each region. It will be seen that these figures provide evidence of the second feature of the influence of latitude already mentioned, namely, the increase in size (and weight) of the individual fish as one proceeds southwards. This can better be demonstrated by considering the mean lengths and length frequencies of the fish caught in the three regions, including data from the earlier surveys (Table 12). Table 12. Variation in size 0/ Merluccius hubbsi, as shown by the differences in mean lengths for all comparable hauls in the three regions here surveyed, regardless of season. The sexes considered separately Note. The numbers of individuals do not indicate the relative abundance in the respective regions because of the very difl^erent number of hauls made m each of them. They show merely the number of individual measurements upon which these mean lengths are based. The sex ratios shown at the head of the table are based on a different array of the data in- cludmg some specimens sexed but not measured. ' Northern region Intermediate region Southern region Sex ratio, % males 54-1 24-0 19-3 Males: No. measured Mean length oMl ^ 943 g- = 304-89129 X 943 g. = 287,512-48647 g., say 287-5 kg. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 299 21 1 females had a mean length of 41-4 cm. K for March females of 4i-i-42-o cm. (Table 17) is 0-695. Hence the theoretical weight of the females is given by W- 41-4' X 0-695 X 211 g- Table 17. 100 = 709-56x0-695 X 211 g. = 493-14420x211 g. = 104,053-4262 g., say 104-1 kg. Table of K for female Merluccius hubbsi throughout the year, computed as described in the text. Interpolated values in italics Length in cm. 34-I-35-0 35-1-36-0 36- 1-37-0 37-1-38-0 38-1-39-0 39-1-40-0 40-1-41-0 41-1-42-0 42-1-43-0 43- 1-44-0 44-1-45-0 45-1-46-0 46-1-47-0 47-1-48-0 48-1-49-0 49-1-50-0 50- 1-51-0 51-1-52-0 52-1-53-0 53-I-54-0 54-i-55-° 55-1-56-0 56-1-57-0 57-1-58-0 58-1-59-0 59-1-60-0 60-1-61-0 61-1-62-0 62-1-63-0 63-1-64-0 64-1-65-0 July 0-759 o-y6i o-y62 0-763 0-763 0-762 0-762 0-760 0-759 0-757 0-755 0-753 0-750 0-749 0-746 0-743 0-740 0-737 0-734 0-731 0-728 0-725 0-723 0-720 0-718 0-715 0-712 o-7og 0-707 0-706 0-704 Aug. 0-777 0-778 0-780 0-781 0-781 0-780 0-779 0-777 0-776 0-774 0-772 0-770 0-768 o-j66 0-763 0-760 0-757 0-754 0-751 0-748 0-745 0-742 0-739 0-736 0-734 0-731 0-72^ 0-726 0-723 0-723 0-720 Sept. Oct. 0-796 0-799 0-798 0-801 0-799 0-802 0-800 0-803 0-800 0-803 0-799 0-802 0-798 o-8oi 0-797 0-800 0-795 0-798 0-793 0-796 0-791 0-794 0-789 0-792 0-787 0-790 0-786 0-789 0-782 0-785 0-779 0-782 0-776 0-779 0-773 0-776 0-769 0-772 0-766 0-769 0-763 0-766 0-760 0-763 0-758 0-761 0-755 0-758 0-752 0-755 0-749 0-752 0-747 0-750 0-744 0-747 0-741 0-745 0-741 0-744 0-737 0-741 Nov. Dec. 0-789 0-750 0-791 0-752 0-792 0-753 0-793 0-754 0-793 0-754 0-792 0-753 0-791 0-752 0-790 0-751 0-788 0-750 0-786 0-748 0-784 0-746 0-782 0-744 0-780 0-741 0-778 0-740 0-775 0-737 0-772 0-734 0-769 0-731 0-766 0-728 0-763 0-725 0-760 0-722 0-756 0-719 0-753 0-716 0-751 0-714 0-748 0-71 1 0-746 0-709 0-742 0-706 0-740 0-704 0-737 0-701 0-735 0-699 0-734 0-698 0-731 0-695 Jan. Feb. 0-720 0-700 0-721 0-702 0-723 0-703 0-723 0-704 0-723 0-704 0-723 0-703 0-722 0-702 0-720 0-701 0-719 0-700 0-717 0-698 0-715 0-696 0-713 0-694 0-7II 0-692 0-710 0-691 0-707 0-688 0-704 0-685 0-701 0-682 0-699 0-680 0-696 0-677 0-693 0-674 0-690 0-671 0-687 0-669 0-685 0-667 0-682 0-664 0-680 0-662 0-677 0-659 0-675 0-657 0-672 0-654 0-670 0-652 0-670 0-652 0-667 0-649 Mar. Apr. 0-694 0-695 0-696 0-700 0-697 0-701 0-698 0-702 0-698 0-702 0-697 0-701 0-696 0-700 0-695 0-699 0-694 0-698 0-692 0-696 0-690 0-694 0-688 0-692 0-686 0-690 0-685 0-689 0-682 0-686 0-679 0-683 0-677 0-681 0-674 0-678 0-671 0-675 0-668 0-672 0-666 0-670 0-663 0-667 0-661 0-665 0-658 0-662 0-656 0-660 0-653 0-659 0-651 0-655 0-649 0-652 0-647 0-650 0-646 0-650 0-643 0-647 May 0-714 0-715 0-716 0-717 0-717 0-716 0-716 0-714 0-713 0-711 0-709 0-707 0-705 0-704 0-701 0-698 0-695 0-693 0-690 0-687 0-684 0-681 0-679 0-676 0-674 0-672 0-669 0-667 0-665 0-664 0-661 June 0-735 0-736 0-738 0-739 0-739 0-738 0-737 0-736 0-734 0-732 0-731 0-728 0-726 0-725 0-722 0-719 0-716 0-713 0-710 0-708 0-705 0-702 0-700 0-697 0-695 0-692 0-690 0-687 0-684 0-684 0-681 Thus the gross weight of the whole sample should ' theoretieally ' be 39i« kg or, fi'^f''^% '» % tolet our ■ probable minimum- weight, we should sa, that at the very least 3« kg. of hake had been The actual weights recorded at St. WS853 were: 414 males (weighed in 4 length groups) totalled 121-350 kg. 529 males unmeasured, weighed in bulk , iSS'OOO kg. 211 females (weighed in 6 length groups) totalled 103-290 kg. Grand total 379-640 kg- and the total for males alone, for comparison with the first calculation above, was 276-350 kg. 300 DISCOVERY REPORTS From these it can be seen that the theoretical weights, using mean monthly figures for K from the tables, gave errors of: + ii'i5/276*35 kg., or +4-07% for males, + o-8i/i03-29 kg., or +078% for females, + ii-96/379-64kg., or +3-15% for the total sample, but consideration of the ' probable minimum ' weight would have saved us from any over-optimism. Converting to units used among British practical fishermen, our theoretical 'guess' at the catch would have been less than 2 stone too big in a catch of 7I cwt., and our cautious ' probable minimum weight ' would have involved the statement that the catch was at least 708 lb. or 6| cwt., instead of the 7I cwt. that we happen to know that it was. Another way in which the ponderal index K can assist us is well shown by the figures for this station: the actual weights of the 414 measured males being known, 'true' mean K for them can be computed for this individual catch thus : j^ w 203-116 , K = j^ X 100 = -^^ — X 100 = 0-603. Now we have the bulk weight of the 529 unmeasured males, and if we calculate their theoretical weight from this figure, assuming the same mean length as the measured subsample, and find good agreement between theoretical and actual weights, we have a good argument that our assumption as to length is justified, thus : P ^"^ loo '^ ^ ^ 529 = 486-27 X 0-603 X 529 g- = IS5-II4 kg., in fact we know that they weighed 155 kg., and the agreement is so close that there is little doubt the unmeasured subsample really had a mean length almost identical with that of the measured one. MIGRATIONS We have seen that the seasonal changes in condition of M. hiibbsl correspond to those that occur in European hake. Variation in condition is primarily connected with the sexual cycle, and since the mam inshore movement of European hake is a spawning migration, it was expected that the move- ments of Patagonian hake would also correspond with those of the better-known species. It was by no means easy to demonstrate this from our scattered data. The necessity for investigating the whole area as fully as possible prevented us from repeating observations within more localized portions to the extent desirable when trying to follow the seasonal movements of a single species. One of the most valuable features discovered by Hickling in his work on European hake was a direct relation between increase in depth and size of fish, which may be locally reversed at the time when the larger fish are moving inshore. By watching the seasonal variation in size over limited ranges of depth one may thus obtain valuable clues as to seasonal movements. But the peculiar topography of the Patagonian Continental Shelf, with its uniform depths prevailing over vast distances, was found to defeat this method of attack. Although correlations were found between depth and size of the fish, and these showed the change of sign with the seasons that one would expect if they moved in a fashion analogous to that of the European species, the correlations were not large enough to be considered significant. The slight gradient of the Patagonian shelf thus masks a feature that is beauti- fully clear off western Europe, where the shelf slopes more steeply and the edge is less abruptly defined, giving a more continuous depth gradient in most localities. 301 DISTRIBUTION AND GENERAL NOTES ON THE SPECIES However, the distribution of exceptionally rich hauls of M. hubbsi clearly indicated the possibility of seasonal migrations similar to those of M. merluccius, and eventually it was found that this could be shown with some certainty when relative abundance and size differences were considered in relation to distance from the mainland coast. It was necessary further to restrict the data to be con- sidered, so as to follow the changes within limited areas (the northern and intermediate regions con- sidered separately), for it is well known that such movements may take place at different times in different parts of the range of a widely distributed species. In the northern region, three series of observations at different seasons seem together to give reasonably good evidence of shoreward movement in summer, and will now be described in some detail. In October-November there are data from nine stations that can be arranged according to their distance from the coast, and which were completed within a reasonably narrow interval of time (see Table i8). Table i8. Merluccius hubbsi captures betzveen 15 October and 3 November 1931 in the northern region, with their distance from the coast Station Depth m. Distance offshore sea miles Numbers of hake Total hake, remarks Males Females WS762A WS762B WS777 WS763 WS771 WS764A WS764B WS765 WS772 66 66 98 85 90 108 107 116 236 H 10 43 75 130 145 1 140/ 212 236 0 0 0 ? 41 12 II 0 0 0 0 ? 50 15 16 9 0 3 juv. in accessory net 0 (torn) 1 5 not sexed, some juv. 91 27 27 9 Consideration of the lengths of the females, together with the distance intervals, seemed to justify a further lumping of these data, to the form shown in Table 19. Table 19. Summary of October-November data bearing on hake movements in the northern region Distance offshore sea miles Total hake Hours trawling (no. of hauls) Hake per hour Males Females Sex ratio No. Mean length No. Mean length Less than 100 miles 100-200 miles Over 200 miles 18 117 36 3(3) 3(3) Ii(2) 6 39 28 53 II 35-8 32-4 65 25 43'5 50-1 45% ^cJ 30% cJc? From these two tables it seems fairly clear that at this time of the year the hake were mamly con- centrated more than 100 miles from the coast. A shoreward movement may have begun. (St. WS771, the richest of this series, was the innermost of the three between 100 and 200 miles offshore.) The length data do not help here; one would not expect the smaller females to head the advance, and m fact the size distribution might well be indicative of a prolongation of winter conditions. The offshore males were too few for the size difference to be considered significant for that sex. However, the sex ratios suggest the beginnings of concentration shoreward, and it is possible that bigger females catch up and pass the smaller ones later. (There is considerable diffuse evidence that the speed and extent of migratory movement is a function of size of fish.) 302 DISCOVERY REPORTS In December there were only six stations suitable for this comparison, but at three of them additional 4-hr. hauls were made, so that there is a considerable body of data over the most critical part of the distance range. The whole series of observations was completed between 13 and 16 December 1931, so that there is little chance of the comparison being vitiated by the time factor. Table 20. December data bearing on hake movements in the northern region Station Depth m. Distance offshore sea miles Hake nos. Hours trawling (no. of hauls) Hake per hour Mean length Sex ratio % males Males Females Total and remarks Males Females WS788 WS789 WS790A + B WS791A + B WS792A + B WS793 8S 94 100 97 107 no 27 53 85 121 147 17s 4 5 58 53 13 0 2 6 319 15° 74 7 6( + 24Juv.) II 377( + SJuv.) 203 87 7 1(1) 5(2) 5(2) 5(2) 1(1) 6 ( + 24) II 75( + 0 40 17 7 37-8 30-6 31-8 31-9 36-8 6o-8 617 51-2 43-3 54-6 54-0 66 45 18 26 15 0 I The results are summarized in Table 20, and show that the greatest concentration was at St. WS790, 85 miles offshore. The length data for males are again unsatisfactory because of the wide dispersion and smallness of the samples containing the larger fish. I suspect that the males were still widely dispersed, as is suggested by the low sex ratios at the richer stations. Here the female length data are very interesting, the mean length at the peak station being considerably higher than at the richest station of the earlier series, although 45 miles farther inshore. Thus Difference ^=7-2 cm Mean length ?, WS790 A + 6 = 51-2, Mean length ?, WS771 =44-0, \/i-^i()= 1-19 6-05. Strongly significant. 0-4672, N' ^=0-9488. od=\/{o-i\.6']2 + 0-9488) -- d 7-2 Od 1-19 This agrees well with the view that larger females may catch up and pass the smaller ones on their way inshore. But, considering the December figures alone, we find that at St. WS791 A + B the females were significantly smaller than those on either side of them ; the relation of length to distance offshore is discontinuous. The detailed length-frequency distributions showed that this was due to a much higher proportion of immature fish, especially of about 30-32 cm. length, at St. WS791. It would seem that large females are definitely heading the shoreward movement at this time, a few having penetrated right inshore among the juveniles, which are perhaps almost non-migratory. Com- parison with the October-November results suggests that in the interim (about 7 weeks, taking mean dates) the large fish have travelled shorewards some 100 miles, while the smaller fish advanced some 15 miles only. In March also there were only six stations providing comparable data from the northern region, but fortunately five of these yielded rich hauls. The remaining station, WS860, presents some anomalies that spoil an agreement with our theory that is otherwise complete ; for a small number of large hake were found less than 100 miles from the mainland, where a good haul of small hake should have been taken. It is probable, however, that the net did not fish properly at this station. A note in Gunther's hand in the original rough log reads : ' Haul disappointingly small. A few hake escaped. Those present of large size— majority caught in after wing which opens suspicion that net may have DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 303 been fishing foul.' Possibly, therefore, we should be justified in disregarding the results from this anomalous station as 'not strictly comparable'. The data from the series as a whole, however, appear to fall quite naturally into inshore and offshore groups, and when so lumped the doubtful result from St. WS860 is completely swamped by the more abundant data from the other two inshore stations. Table 21. March data bearing on hake movements in the northern region (21-25 March 1932) Station Depth m. Distance ofFshore sea miles Hake nos. Hake per hour Mean lengths Sex ratio % males Males Females Total, and remarks Males Females WS853 WS860 WS855 WS859B ■ WS859A WS8s8 90 102 112 108 108 127 44 77 80 142 146 212 943 5 130 64 113 28 211 16 120 42 82 62 1154 (only 414 a measured) 21 (? net foul) 250 106 19s 90 ^"54 21 250 106 195 90 36-4 38-0 33-8 38-5 38-9 40-9 41-5 54-6 37-8 45-9 45-3 50-3 8i-7 29-4 52-0 60-4 S7-9 31-1 The data from individual stations are shown in Table 21. The very large catch at St. WS853, the station nearest the land, was the best we obtained at any time. The hake here were of small size, but some were ripe, and the high proportion of males was also noteworthy. It would seem clear that the smallest mature females are the last to spawn in any given season. Table 22. The March data combined into inshore and offshore groupings Distance ofFshore sea miles Less than 100 More than 140 Total hake 1425 391 Hours trawling (no. of hauls) 3(3) 3(3) Hake per hour 475 130 Males Nos. 549' 205 Mean length 35-8 39-0 Females Nos. 349 186 Mean length 407 47-1 Sex ratio 75 52 * =29 males were not measured at St. WS853, the number given refers only to the measured specimens upon which the mean length given is based ; hence the discrepancy with the ' total ' column. The combined data (Table 22) show very clearly that smaller fish predominated in the inshore catches and that the proportion of males was higher inshore. The differences m mean lengths are strongly significant by the usual statistical tests. Comparison with the December figures (Table 20) indicates that a complete change in the hake population had taken place; in March at the inshore stations, the females were much smaller and the males larger than in December. Another striking feature of the March results is the high sex ratios-even at the offshore stations the proportion of males was much higher than earlier in the year. This is probably due to the main concentration for spawning being later in the summer than exact correspondence with the habits of the European species would demand. Thus the larger oflfshore fishes were probably only just beginning to disperse after spawning. In this respect M. hubbsi may come closer to M biMnearrs (where the correlation between shoreward concentration and the peak of annual temperature is very strong) than to M merUcciMs, where the first wave of larger spawners moves inshore at least two months b fore^he maximum temperatures are reached, and the second wave contains ";7;'^77^^,^^^^^^^^^^^ be borne in mind that the range of temperature is much greater in the habitat of M hdrneans than m ha o^rJ./.^.) The timing of the cycle of movement of M. hubbsi ^^Pf ^^^^ --^J^;^^^^^ that lateness of the whole 'plankton calendar' of these southern waters, which was described m the 304 DISCOVERY REPORTS introductory section of this paper. It is not likely to be a direct effect of temperature (as seems possible with M. bilinearis), for the range of temperature here is small, approaching British conditions more closely than New England conditions. The broad fact that smaller fish are still closely congregated inshore, while large fish are beginning to disperse offshore, seems sufficiently clear for us to state that the March results are consistent with the view that the seasonal movement of M. hiibbsi is essentially similar to that of other species of hake elsewhere. Ponderal indices also support this view: the small inshore population, with a probable majority of late spawners and immatures, showed average K 20% higher than that of the larger offshore population sampled at the same time. The latter must have completed spawning, for they showed the lowest average K values (around 0-550) recorded at any season. In the Intermediate Region data covering more of the annual cycle are available, if we may con- sider results obtained in different years as roughly comparable for our present purpose, but they are scantier and less satisfactory than those from the northern region. We have already seen that the hake diminish in numbers towards the south, so that this difficulty was only to be expected. In October and November no satisfactory series of observations was obtained here, but the size and abundance of the hake at two stations, WS773 and 775, 206 and 82 miles off the land respectively, were consistent with the view that a winter type of distribution still prevailed. Hake were nearly twice as numerous at the offshore station where they were very much larger than those found farther in. The proportion of males was greater inshore. Twenty-two juvenile hake of indeterminate sex, less than 20 cm. long, were also captured during this period. This was at St. WS776, 60 miles from the land. In December numerous observations were obtained during the third survey, but they were too scattered, and the samples too small, to warrant individual treatment of the results. When the stations are grouped according to their distance from the coast, the picture of frequency distribution obtained IS m accordance with more conclusive results from farther north at the same season. It therefore seemed legitimate to use similarly grouped data in studying the size distribution, etc., since although either chain of evidence may appear bald and unconvincing by itself, they corroborate each other. Table 23. December data bearing on hake movements in the intermediate region, 5 to 22 December 1931 Distance grouping sea miles from mainland coast Actual mean distance Hours trawling Hake nos. Hake per hour Mean lengths Sex ratio % males Males Females Total Males Females I, 0-49 II, 50-99 III, 100-149 IV, 150-199 V, over 200 35 75 129 161 219 6 6 12 3 2 34 17 20 I 0 77 134 219 25 17 III 151 239 26 17 18 25 20 9 8 33-1 31-9 ?* 42-0 41-9 457 56-8 58-6 55-5 30-6 II-3 8-4 3-8 0-0 * The mean would be misleading here (see text). These data are summarized in Table 23. It will be seen that the relative abundance was greatest between 50 and 99 miles from the coast, and that the proportion of males was highest close in to the land. The length data for males are unsatisfactory owing to wide dispersion in the small samples. In distance grouping III a mean length for males would be meaningless, for the sample was composed of some very small and a few large fish with intermediate lengths entirely unrepresented. The more abundant length data for females tell a consistent story: the differences in mean lengths of the two inshore groups from all the offshore groups are significant. The inshore fish were smaller, and the offshore population apparently very homogeneous at this time. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 305 Here, then, we have the suggestion that shoreward concentration is beginning, but is not yet so well defined as in the northern region at the same period. A few January observations, quite insufficient by themselves, fit in so well with the general theory that I give them here in full (Table 24) and have included them in the general diagrammatic summary of the observations on hake movement in Fig. 25. Following on the December observations these few January stations clearly suggest an increase in the tendency to shoreward concentration. This seems mainly due to an incursion of larger females (just as we should expect), for the difference in mean length of over 5 cm. between the peak station WS809 and the inshore figures for distance grouping II of the previous month is strongly significant. Table 24. Observations from -] to () January 1932, to be considered with the December data bearing on hake movements in the intermediate region Station Distance offshore sea miles Hours trawling (no. of hauls) Hake nos. Hake per hour Mean lengths Sex ratio Males Females Total Males Females WS810 WS809A + B WS808 WS807 20 52 77 104 1(1) 5(2) 1(1) 1(1) 0 7 0 0 2 69 3 0 2 76 3 0 2 15 3 0 30-7 37-0 51-4 58-0 0 9-2 0 Some observations made in April during the first survey are also in very good agreement with the theory Catches were more uniform than they were during the peak of shoreward movement (m March only northern region data available), and the distribution in relation to distance from the land was bi-modal The hake were most abundant at two distinct points over the range observed, at 94 and 238 miles from the land, the former being the richer haul and composed of significantly smaller fishes A distribution of this type is just what one would expect if there had been a double wave of shoreward movement, and if in April (autumn) the fish were again moving offshore towards their winter quarters. We have already seen that the larger fish get farther out than the small ones, except at midsummer, and they must therefore lead in the offshore movement, just as they seem to catch up and pass the smaller fishes during the season of shoreward movement. The April data are shown in detail in Table 25. April data bearing on hake movements in the intermediate region, 17 to 25 April 1927 Table 25. Station WS96 WS95 WS108 WS97 WS98 WS99 Distance offshore sea miles 34 68 94 1 65 238 285 Depth m. 96 104 119 146 172 237 Hake nos. Males 14 10 62 14 II o Females 16 29 66 20 60 19 Total per hour 30 39 128 34 71 19 Mean lengths Males 30-0 39-4 35-6 38-1 39-4 Females 33-7 43-4 43-6 57-8 53-6 63-3 Sex ratio % males 46-7 25-6 48-4 41-2 15-5 0-0 Another point brought ou, by this table is the relatively greater abundance of ma es than at any other season except March, and the.r tendency not to go so far offshore as the larger fcn>ales^ tL dXenees in mean length for either sex at the two peak stations are sigmficant by the usual rf, Jtsf wh eh has been appM throughout this section wherever differences » nn,ean length have b'een consTdeL specifically I have not tabulated a or 0I.IN io. all the nreans, because many of them did not individually assist in the building up of the general picture. 3o6 DISCOVERY REPORTS Finally, during the second survey, a series of observations in June (midwinter) yielded two rich hauls at the greatest distance offshore, while farther in there were few hake, and only a singleton within loo miles of the land. These results also showed that, as at all times except midsummer, the mean lengths of the more adequate samples increased with their distance from the coast, and at the same time the proportion of males diminished. Table 26. June data bearing on hake movements in the intermediate region, i to 8 June 1928 Station Distance offshore sea miles Depth m. Hake nos. Mean lengths Sex ratio % males Males Females Total per hour Males Females WS222 WS223 WS220 WS219 WS216 WS2I7 29 95 126 134 197 221 103 114 106 115 176 146 0 0 3 16 87 118 I 0 3 26 181 294 I 0 6 42 268 412 44-7* 36-8 40-1 39-4 59-0* 43-0* 407 47-5 46-5 0 50 38 32 29 * Too few for means to have any significance. CONCLUSIONS ON MIGRATION The relative abundance of the hake caught in relation to distance from the coast is the best means of studying the probable seasonal movements of M. hubbsi, from our unavoidably limited data. For comparative purposes the catches per hour for each series of observations may be summed, and the catch at each distance category expressed as a percentage of the figure so obtained. A diagrammatic summary of the seasonal observations, obtained in this way, is shown in Fig. 25. We have seen also that comparisons of mean lengths and sex ratios corroborate the general picture so obtained, wherever the data are adequate. M. hubbsi seems to migrate towards the coast in summer, and offshore in winter, in much the same way as do better known species of hake elsewhere. It would seem that as in M. merlucciiis, the larger females move inshore first, passing their smaller sisters who may, however, begin to shoal somewhat earlier. There is a strong suggestion that the smaller fish are rarely abundant at the greater distances from the coast ; probably they do not migrate so far or so fast as the bigger ones. This may be a function of size, and not only due to the greater proportion of immature fish among the smaller individuals, for the proportion of males always diminished as one proceeded seawards, and in this species, where' the disparity in size between the sexes is much more marked than in M. merhiccius, it is certain that many of the males of even the smallest length class are mature. Probably the movements of males show some marked differences from those of the females: the proportion of males was noticeably high in March, at the time of greatest shoreward concentration of the smaller females, still fairly high when seaward movement had begun in April, and low at all other seasons. This suggests that, except at the height of the breeding season and for a short period afterwards, the males are more widely dis- persed, and less inclined to shoal, than are the females. Although M. hubbsi seems to move shorewards in a double wave rather like M. merhiccius, this movement begms later m the year, and in this respect perhaps resembles more closely the movement of M. bibneans. It is thought that this is probably connected with the general lateness of the biological seasons in this part of the southern hemisphere, which has been described in the introduction to this report. In M. bihnearis the later timing of the cycle may well be a direct effect of temperature, for off New England the annual range is great. It is unlikely to be so off Patagonia, where the annual cycle DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 3°7 of temperature, though on a rather low level, shows small annual range. Thus the temperature con- ditions in the habitat of M. hiibbsi resemble more closely those found on the European seaboard of the North Atlantic (see also the Hydrological Notes in the Introduction). > o INTERMEDIATE REGION 50 IQO 150 200 250 50 100 150 200 250 MILE5 FROM THE COAST Fie 2. Diagram showing seasonal variation m relative abundance of Merluccius hMsi at ^' ^ different distances from the coast. 3o8 DISCOVERY REPORTS THE FOOD AND FEEDING OF MERLUCCIUS HUBBSI Owing to the wide scope of the survey programme it was not possible to carry out such detailed observations on this subject as have been found desirable by European specialists working on single species of fish (e.g. Hardy (1924) on herring, Hickling (1927) on hake). On our first survey 191 stomachs were examined and on the second 186. On the third survey forty-six notes on stomach contents were made in the field, but owing to pressure of other work the numbers of fish examined in this way were not always recorded. It was evident, however, that the proportion of fish containing food was least in summer, which accords with the known habits of European hake (Hickling, 1927, p. 49). A general loss of appetite just prior to spawning is known among many diverse species of fishes. ^ Our first survey (autumn) and second survey (winter) results show that a slightly higher proportion of Patagonian hake were found to contain food in winter, but I do not believe this indicates more extensive feeding during that season. From notes on the size and numbers of food organisms in individual stomachs it is evident that feeding was heaviest in autumn. The apparent anomaly is due partly to a seasonal change in diet, partly, no doubt, to the limitations of the data. The number of times that food of recognizable categories was recorded, and the percentage occur- rence of each category, during each survey are set out in Table 27. From this it is at once apparent that the feeding habits of M. hubbsi art essentially similar to those of M. merluccius, and perhaps even closer to those of M. bilinearis (Bigelow and Welsh, 1925, pp. 389-90). It feeds chiefly upon other fishes, mcluding even its own species, squids and more or less planktonic Crustacea. This last constituent was chiefly found in the stomachs of the smaller and younger fish. The list of fishes eaten includes the commonest species of the area. Falkland herring {Cliipea fiiegensis) was by far the most important forage species, especially in winter, but other species are preyed upon with equal voracity when readily available. Thus as many as sixty-seven individuals of the small scald-fish Thysanopsetta naresi have been taken from the stomach of a hake less than 56 cm. long. Merluccius hiibbsi, like other species of hake, are known to devour individuals of their own kind more than half their own length. Had we been able to obtain more food records during the offshore phase of the seasonal migration there is little doubt that such instances would have been commoner. There is one peculiar difference from the feeding habits of European hake that makes the paucity of deep-water observations during winter the more regrettable.^ Although the distribution of M. hubbsi overlaps that of Micromezistiiis australis to the southward, the latter was not recorded as a constituent of the hake food. This is extraordinary because M. australis is very closely related to our own blue whiting M.poutassou (Norman, 1937, p. 51), which Hickhng (1927, p. 42) had shown to be such an important constituent of the food of European hake. It seems that M. australis keeps more exclusively to deeper water and higher latitudes than its European counterpart, and therefore its habitat— during the warmer months of the year— does not coincide with that of the hake to anything like the same extent. In winter the southern blue whiting may move northwards as well as offshore, as will be shown in the section dealing with that species. Possibly the hake then feed upon them as we should have expected ; but the fact remains that there was no evidence of this at the few stations where we did locate the two species together. It will be noted that the list of food organisms includes several bottom-living fishes and a little benthos. The slightly more benthic tendency in choice of food is doubtless occasioned by the uniformly " hv^H^t^'^T t° this phenomenon in Quinat salmon (Jordan), silver eels (Petersen) and pleuronectids (Todd), are quoted by H.ckhng (1927), and he was further able 'to suggest that the blue whiting {Micromezistius poutassou) also feeds much le7s ? olmrilerv.Thr'^?" -h'™'" behaviour among Labrador cod ^o^uld seem to belplied ly Harold TZmpson seasoA^' ^ ^ ^^ "P°" ''P'" ^°' '^^ P'™'^ °^ S'"'' ^""'^'"g ^"'^ recuperation which succeeds the spawning season . 2 I must reiterate that this was due to the precipitous slope preventing trawling, and not to any lack of endeavour. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 309 moderate depths of the Patagonian shelf, which gives a much sUghter depth gradient than the habitat of the better known species, until the edge is reached. The few echinoderms, etc., were, moreover, recorded in autumn, when the fish feed ravenously upon whatever comes their way. Table 27. Feeding of Merluccius hubbsi Food categories Times recorded and percentage occurrence (italics) Third survey (summer) First survey (autumn) Second survey (winter) Clupea fuegensis Merluccius hubbsi Macruronus magellanicus Salilota australis Notothenia ramsayi Notothenia spp. Stromateus maculatus Thysanopsetta naresi Unidentified fish Post-larval fish 12 26-1 2 4-3 1 2-2 7 15-2 2 4-3 I 2-2 19 i6-5 1 o-g 3 2-6 2 1-7 2 1-7 7 6-1 8 7-0 I o-g 34 28-1 I 0-8 6 5-0 5 V-! 9 7-<^ Total fish 25 54-2 43 37-4 55 45-4 Large squid Squid 10 21-y 7 6-1 14 12-2 4 JJ Total squid 10 21-J 21 i8-3 4 3-3 Decapoda Munida subrugosa M. gregaria Munida unidentified Euphausians Parathemisto gaudichaudii Hyperid amphipods* Amphipods Serolis sp. 6 13-0 I 2-2 3 6-5 I 2-2 7 6-1 13 11-3 1 o-g 18 15-7 4 3-5 2 1-7 2 i7 56 46-j 2 i7 Total Crustacea II 23-g 45 39-2 60 .^97 Holothurians Asteroids Ophiuroids Sponge fragments — — I o-g 1 o-g 2 1-7 2 1-7 2 1-7 Total echinoderms, etc. — — 6 5-2 2 i7 Total separate records 46 gg-8 115 100-1 121 lOO-i * Doubtless mainly Parathemisto. Hints of some interesting seasonal changes in dietary may be gathered from Table 27, although the data are not quantitative. Squids were obviously an important food in summer and especially in autumn, but were rarely recorded in winter-caught hake. Of the crustacean constituents, chiefly important to the smaller hake, Munida were eaten largely in summer, when there were no records of euphausians in the stomachs. In autumn there was still some Munida, considerable quantities of euphausians and many hyperid amphipods. (There is little doubt that nearly all of these were Para- themisto gaudichaudii, a very common species over most of the southern ocean, but I give the categories as stated in the original records.) In winter euphausians were the most frequently recorded food, and few other Crustacea were present. It is probably significant that the widest variety of food was re- corded in autumn, when the hake were feeding intensively after spawning. The relative importance of the different constituents of the food of Merluccius hubbsi cannot be 3IO DISCOVERY REPORTS accurately gauged without weight or volume records and altogether more extensive data, but a system of arbitrary' weighting (kept well on the ' safe side ') permits a diagrammatic presentation of seasonal changes in the major (lumped) categories, that gives a useful general picture. It may be regarded as a cautious understatement of the predominance of fishes and cephalopods in the diet, and of the obviously great importance of Chipea, especially in winter. The 'weighting' employed, after due con- sideration of known weights of most of the food organisms was : regarding Crustacea and benthos as unity, cephalopod records were multiplied by four and fish records by five. The results are shown in Fig. 26. Doubtless this picture would be altered by more detailed results, especially if the size of the fish could be taken into account, but such work will only be possible when a naturalist can devote his whole time to the one problem. A B C Fig. 26. Diagrams showing crude relative proportions (arbitrarily weighted) of the main food categories of Merlucdus huhbsi at different seasons. Weighting: fish x 5, squids x 4, Crustacea x i, and echinoderms, etc., x i. Echinoderms, etc., which are rarely eaten are left white in the diagrams. A, third survey, summer. B, first survey, autumn. C, second survey, winter. PARASITES Like most other sea fishes M. hubbsi were observed to be very commonly infested with nematode, cestode and trematode worms. On the first survey these three classes of parasites were observed to be present in (roughly) the order of frequency in which they are named above, and it was noted that they seemed particularly abundant in the larger (older) fish. Almost all the specimens examined had nematodes in some part of the digestive tract or in the body cavity. Copepodan parasites were evidently less frequent but not uncommon. Chondracanthidae were more than once recorded as numerous in the mouth, and Miss N. G. Sproston informs me that members of this family frequently infest European hake also. There aje numerous references to Lernea and Lerneidae in the log-books, which introduce an un- fortunate element of doubt into some carefully collected statistics of the incidence of this form of parasitism during the third survey, when the hake were sorted into length classes for weighing. I believe that these records all refer to a lernaeocerid either identical with our own Lernaeocera branchialis or very closely allied to it, but earlier references to ' Lernea ' on external situations (L. branchialis is strictly a blood-vascular parasite, and has been recorded from European hake) leaves some element of doubt. The situation is clouded by the unfortunate change of status of the genus Lernea so justly deplored by Gurney (1933, p. 336). To be quite safe, these parasites may all be referred to the family Lernaeoceridae as proposed by him. The bulk of the figures were obtained in December, in the northern and intermediate regions, that is, in the most favourable part of our area for hake, at a time when the seasonal shoreward migration DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 3ii had begun. There is thus a good chance that the figures may be fairly representative, though the chances of infection by these parasites might vary with the seasons. A few data obtained farther south, at other times, do not suggest this, and in the main confirm the picture given by the good December figures. The latter only are considered here. The results are given in summarized form m Table 28, and the increase in percentage of female fish infected with increasing length is shown graphically in Fig. 27. Table 28. Incidence of infection by Lernaeoceridoe, Merluccius hubbsi, north of 50° S, December 1931 Length classes cm. No. females examined No. females infected Females 0/ /o infection No. males examined No. males infected Males /o infection 21-30 31-40 41-50 51-60 61-70 71-80 81-90 63 107 216 334 164 37 I I 4 13 24 15 8 0 1-59 374 6-02 7-19 9-15 21-62 64 70 29 I 4 0 I 0 6-25 0 3-44 S922 26s M7-05 5:164 S5 M3-05 It will be seen that female Merluccius hubbsi showed a steady increase of ' percentage mfection' with these parasites as they grew longer (and older), and marked increase in the highest length group for which adequate figures are available. Females appear to be more than twice as often mfected as are males. Zl-30| 31-40 I 4-1 -50 I 51-60 | 61 "70 1 71-80 LENGTH CLASSES _ CMS Fig. 27. Increase in infection by Lernaeoceridae with increase in length (and age) of female Merluccius hubbsi. 312 DISCOVERY REPORTS The males do not show any correlation between increase in length and increase in liability to infection. The smallest length class showed the highest percentage infection, which was also con- siderably higher than that of females of the same length. I think this discrepancy is in some way bound up with the marked differences between the sexes that became apparent in the study of the general bionomics of this fish. Apart from the size difference, and almost certainly slower later growth rate, male M. hiibbsi reach maturity at a much smaller length than females, they seem less given to shoaling during the ' off' season, and do not migrate so far as the older females. All these factors may influence their chances of infection by Lernaeoceridae. Macruronus magellanicus Lonnberg This is a long, slender fish, with close superficial resemblances to the Macruridae, with which it was formerly classified. Norman (1937, p. 49) has shown, on osteological grounds, that it should be placed in the family Merlucciidae. The tail, tapering to a point and without a separate caudal fin, is the most noticeable point of similarity to the macrurids, but Macruronus lacks their projecting snout, and its distribution and habits are markedly different. Macruronus is found in numbers only on the shelf, in relatively shallow water during at least nine months of the year, whereas the macrurids are essentially a deep-water group inhabiting the slope beyond the shelf edge and even greater depths. The coloration of Macruronus reflects this difference in habitat. In Plate XVI a water-colour sketch of a living specimen taken at St. WS99, by E. R. Gunther, is reproduced. At an earlier station he had described its coloration thus : ' laterally a pale lustrous blue, becoming more intense dorsally into tones of sapphire and turquoise, ventrally losing colour becoming silvery white.' A colour pattern such as this is normally characteristic of mid-water fishes inhabiting moderate depths over a well- illuminated sandy bottom, and this would apply fairly to that part of the plain of the shelf where we found the smaller individuals most abundant. It IS interesting to note that the correct taxonomic position of Macruronus is reflected in the local Spanish-American name ' Merluza de cola ' which Norman (loc. cit.) states is applied to it. This might be freely translated into 'long-tailed hake' with advantage, for English-speaking fishermen tend to apply the names ' rat-fish ' or ' rat-tail ' to anything remotely resembling a macrurid, consequently confusmg them with fishes as gsnetically remote as chimaerids in some parts of the world. M. magellanicus is the second most important fish of the Patagonian Continental Shelf. In our catches it was outnumbered only by Notothenia ramsayi, but although slightly more numerous than hake in the aggregate it was less widely distributed, a few exceptionally rich hauls augmenting the total unduly. A much more slender fish than the hake, it is about half as heavy at a given length, and a larger proportion of the smaller individuals escaped through our normal cod-end mesh. There are numerous references in our rough logs to 'Macruronus seen escaping'. In the eighty-six summer hauls of the third survey, for which roughly comparable weight data are available, Macruronus yielded 29-5% by weight of the total fish taken (rubbish excluded) as against the 47-3% of hake. The relation to other less important categories can be seen from the tables in the concluding section of this report. Macruronus was the most important species in the southern region, where, as we have already seen, the hake diminished greatly in numbers. Of the weight of fish caught here, Macruronus provided 4^2 % • M. magellanicus is most excellent eating. In the third (unpublished) scientific report' on the work of the ' William Scoresby ' Dr Mackintosh wrote : ' It is generally agreed that Macruronus is superior to any of the other common fishes. The flesh is reasonably firm and free from too many small bones.' DISTRIBUTION AND GENERAL NOTES ON THE SPECIES Our records of the total catch of this species are as follows : 313 WS77 WS79 WS91 WS92 WS94 WS95 WS99 WS108 WS216 WS218 WS762A WS762B WS763 WS764B WS765 WS771 WS774 WS775 WS776 WS786 WS788 WS789 12. 111. 27 13. iii. 27 7. iv. 27 8. iv. 27 16. iv. 27 17. iv. 27 19. iv. 27 25. iv. 27 1. vi. 28 2. vi. 28 16. X. 31 16. X. 31 16. X. 31 17. X. 31 17. X. 31 29.x. 31 i.xi. 31 2. xi. 31 3.xi. 31 7. xii. 31 I3.xii. 31 13. xii. 31 I 4 3 24 49 2 4 32 I 3 I 123 6 4 25 12 368 43 162 34 I I WS790A WS790B WS791B WS792A WS792B WS793 WS796A WS796B WS797B WS797C WS799A WS799B WS800B WS805 WS806 WS807 WS810 WS811I WS811II WS812I W 881211 WS813 14. xii. 31 14. xii. 31 14. xii. 31 15. xii. 31 15. xii. 31 15/16. xii. 31 19. xii. 31 21. xii. 31 20. xii. 31 20. xii. 31 21. xii. 31 21. xii. 31 22. xii. 31 6. i. 32 7- •• 32 7- i- 32 9. i. 32 10. i. 32 12. i. 32 10. i. 32 12. i. 32 13. i. 32 107 23 30 9 932 5 56 3 26 37 37 12 I I I 12 3 162 75 33 5 80 WS814 WS815 WS816 WS817A WS817B WS818A WS818B WS838 WS848 WS8S3 WS8s5 WS857 WS858 WS859A WS859B WS864 WS866 WS868 WS870 WS874 WS875 13.1.32 13.1.32 14. 1. 32 14. i. 32 14. i. 32 17. 1. 32 17.1.32 5. ii. 32 10. ii. 32 21. iii. 32 22. iii. 32 23. iii. 32 24. in. 32 25. iii. 32 25. iii. 32 28. iii. 32 29. iii. 32 30. iii. 32 31. iii. 32 3- iv. 32 3. iv. 32 981 2 56 5 2 2 I 33 3 26 I 48 5 12 II I 1180 98 227 II 78 Two main features of the distribution already mentioned become quite clear from these figures: the greater relative abundance of the species in the southern region, and the tendency to form dense local shoals, so that a small minority of the catches are vastly bigger than the others This latter feature is the probable reason for lack of a clear north to south gradient in abundance of Macruronus m our records A single extra large catch (of small immature individuals) was made in the northern region, but no corresponding shoal happened to be encountered while we were sampling the intermediate region. It is clearly necessary to consider other lines of evidence, bearing on the probable spawning time and movements of the fish, that may help to explain the observed distribution. The most fruitfu studies possible from existing data appeared to be considerations of seasonal variations in ponderal index regional variation in mean length, and relation of relative abundance and mean length with depth • but before we pass on to these one most important feature, quite clear from the catch records, must be emphasized: Macruronus was present at only two stations worked in winter, and these were among the most northerly of thirty more or less comparable hauls worked along the shelf edge, and over the shelf, in the main summer haunts of the species. This strongly suggests that the fish move north in winter. The repeated attempts made by Mr John to trawl in deep water over the shelf edge during the winter survey sufficed to show the offshore movement of hake at that season and should also have revealed the presence of Macrr^ronus offshore if there were not also a considerable mendiona component in the direction of movement of the latter species. We have no proof, of course, tha Macruronus does not move eastwards offshore-it probably does, though not to the same ex^nt as th hake-but it seems certain that it moves north as well, whereas any meridional component in the direction of movement of the hake would seem to be too small to be demonstrable from existing "^Tn Table 29 the data have been grouped at mean dates, from stations selected according to their tine distribution, so as to show the seasonal variation in 'average' ponderal index of Macruronus in each o the three regions. The chief stress was laid upon inclusion of stations within a narrow interval ot: about the m'ean date and therefore regional differences in the other futures shown-reative abundance mean lengths and sex ratios from the same groups of data-are not fully illustrated by this array The' Tata suffice to indicate three main points, however: greater abundance and size of in- d "duals in the south, and the constancy of the sex ratios, showing a slight preponderance of females 314 DISCOVERY REPORTS in nearly all groupings. The seasonal variation in ponderal index shows a rise in all three regions from spring and summer to autumn. These results are also shown graphically in Fig. 28. Table 29. Data selected over short-time intervals in each region, to show the increase in ponderal index of Macruronus magellanicus during the season, with corresponding figures showing relative abundance, mean lengths and sex ratios, and the mean sex ratio for each region i Mean date Total Macruronus Hours positive hauls Fish per hour's + haul Sex ratio i, males Mean length cm. Average K Sex ratio mean for region 23. X. 31 15. xii. 31 23. iii. 32 37 1089 45 Northern region 2 IS 3 18 73 15 41-7 40-0 6o-o 417 37-6 33-8 0-290 0-312 °-354 40-8% 2. XI. 31 20. xii. 31 23. iii. 32 570 153 48 3 15 I Intermediate region 39-9 190 10 48 49-0 43-8 38-9 37-6 40-6 0-270 0-345 0-358 41-9% 45o Spring Hours trawling 0 6 8 0 1 4 1 0 0 0 0 N. ramsayi per hour 0 39 14 0 8 8 0 0 0 0 Summer Hours trawling 2 32 57 7 5 5 5 i 2 i TV. ramsayi per hour 0 8 yi 16 5 0 1 5 0 1 0 Autumn Hours trawling \ 8 25 7 6 1 0 0 0 0 A'^. ramsayi per hour 0 II 157 40 238 8 0 0 0 0 Winter Hours trawling 0 1 9 6 7 •/ i 0 0 0 N. ramsayi per hour 0 0 49 29 20 67 0 0 0 0 An investigation of the mean lengths within depth categories at each season showed that there was a significant increase of size with depth. The fish caught between 201 and 250 m., and 251 and 300 m., in winter, do not show this ; but all the other observations, taken in pairs successively (twelve pairs in all) show the deeper sample of the pair to contain significantly longer fish (Table 32). It is therefore clear that A^. ramsayi conforms to the general rule 'larger fish in deeper water'. Table 32. Variation in mean length of Notothenia ramsayi at different depths during each of the four seasons of the year Depth range m. Spring Summer Autumn Winter Mean length cm. oUN Mean length cm. ol^lN Mean length cm. <^l,IN Mean length cm. o^N 1-50 51-100 101-150 151-200 201-250 251-300 None 8-9 II-8 None 20-5 26-9 0-0270 o-i8gi 0-1250 2-0763 None 13-0 13-9 17-8 26-9 None 0-2107 o-oigo 0-5113 1-3578 None i6-7 219 22-3 27-5 31-0 0-2221 0-0045 o-og28 o-oo8g 1-1250 None None 23-4 24-8 24-2 26-1 0-0348 o-og27 0-1386 0-0513 These results also indicate that the vast majority of that part of the N. ramsayi population found beyond the shelf edge (i.e. in depths greater than 200 m.) are more than 20 cm. long. Reference to the actual frequencies shows that only sixty-three out of 1894, or 3-3 % , of the fish captured over the edge were less than 20 cm. long. The winter migration to deeper water is thus almost entirely confined to the larger fish that, as will presently be shown, are almost certainly more than two years old. The numerous length-frequency data available for N. ramsayi can only be used to test the probable ages of the younger fish, by Pettersen's method. It was rarely possible to sex the individuals, owing to pressure of work upon the more obviously useful species, and, when sexing was achieved, it was found that a very large majority of the larger fish (of more than 25 cm.) were females. It is therefore almost certain that after maturity is reached the male and female growth rates diverge, the females growing the faster. Consequently one cannot base any conclusions as to age groups or growth rate upon the length frequencies of the larger fish in unsexed data. If the mature males are considerably smaller than the females, the apparent excess of the latter will in the main be accounted for by the selective action of the net. For the younger stages, however, the rich autumn hauls (many of them taken in the ' accessory nets') furnish length-frequency data that seem to show age groups with some certainty, judging by the consistency with which modes recurred at the same lengths. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 333 Fig. 37. Depth relations of Notothenia ramsayi. Above: relative abundance of fish caught within each depth-grouping at each season (circles), and the relative amount of time spent trawling within the same limits (crosses). Below: widths of polygons are proportional to the relative abundance of fish within the limits shown. (Data from Table 31 transposed to a percentage basis for comparison.) 334 DISCOVERY REPORTS 30- ;20- 10 B Fig. 38. Percentage length frequencies of A'^. ramsayi taken (A) between 51 and 100 m., and (B) between loi and 150 m., in spring. We have no extensive autumn data for the smallest Notothenia ramsayi, but the pooled length frequencies of spring samples taken in 51-100 and 101-150 m. show such strong modes about 8-10 cm. (with the larger fish in the deeper water) that it is reasonable to assume that they represent a year class. These fish are thought to have been almost one year old — o-group becoming I-group (Fig. 38). In autumn we obtained eleven rich hauls of between eighty and 161 6 individuals of this species, and on plotting the percentage length frequencies (Fig. 39) it appeared that modes at around 14-16 and 22-23 cm. recurred with such consistency that there can be little doubt that they represented year classes. They are thought to indicate I-group and I I-group fish respectively. The scale of the figure is necessarily much reduced, in order to permit comparison of all the samples on one page. Table 33, summarizing the important points arising from Fig. 39, and giving relevant geographical data, has been prepared to cover any loss of information due to the unavoidably small scale of the figure. It will be seen that clear evidence of either or both of the two year classes mentioned is provided by all but three of these samples notwithstanding their diverse locations. At St. WS83 there was a strong mode at 17 cm. — considerably higher than the modal values for most of the presumed I-group fish captured around that time, which were at 14 and 15 cm. It is believed that this is explained by the geographical position of St. WS83 — close in to, but on the southern side of, the Falkland Islands, considerably farther south than any other station at which such small A^. ramsayi have been taken in quantity. From the general distribution of the species as already described it seems at least highly probable that only the largest members of the I-group would be likely to penetrate so far south. This notion is perhaps strengthened by the strong 14 cm. mode shown by the sample from St. WS73, almost equally close in to the islands, but to the north of them. The very rich sample from St. WS97 yielded length frequencies which are not incompatible with the idea that the Il-group predominated, but the mode is ill-defined, as stated in Table 33. The suggestion that this may be due to the slower growth rate of mature males as compared with females is strongly supported by the fact that such sexually differentiated grovi^h is known to take place in various other fishes (e.g. hake). Finally, at the only deep station at which a large haul of N. ramsayi was secured in autumn, St. WS214, it seemed that we were dealing with an altogether larger age group (? Ill-group) with the mode at 27 cm. This would be in full accordance with our findings as to the general relation between depth and size of fish where N. ramsayi (and many other species) are concerned. Considering these results in conjunction with the general distributional data, it seems probable that early growth of N. ramsayi takes place somewhat as follows. The fish probably hatch in early summer and grow rather more than 10 cm. in their first year of life, at least 8 cm. in their second year and 6 cm. in their third. At this point maturity is probably reached and no reliable conclusions can be drawn from unsexed data, for it is probable that mature females grow considerably faster than males, but there is some evidence suggestive of a 4-5 cm. increment during the fourth year. I have included this brief and admittedly speculative suggestion as to growth rate in N. ramsayi because it will at worst provide a working hypothesis if any future work on the shelf is possible, and we know so little of the growth of any of the fishes in southern temperate waters. Several records of the stomach contents of A^. ramsayi were made in autumn and winter, when some 75 % of the fish contained recognizable food. The number of times food of each category was recorded, 10 10 lo- ws 73 10- 10- 10- 10- 10- 10- 10- 20- 10 WSS3 WS97 30- 20- 10- WS2I7 W5859B Fig. 39. Percentage length frequencies of TV. ramsayi at autumn stations with more than eighty individuals, show- ing probable year classes. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 335 0 and an arbitrary weighting system (based on some contemporary volumetric records) which provides some approximation to relative values of major food categories, are shown in Table 34. Clearly N. ramsayi is much more of a bottom-feeder than most of the other Patagonian species for which we have any data. Benthic Crustacea were the largest item in the dietary observed, and a considerable proportion of poly- chaetes were eaten. A'^. ramsayi, however, feeds heavily upon Falkland herring whenever opportunity offers, though the same might be said of almost every animal in the area capable of swallowing them, including such other typical bottom-feeders as Cottoperca and some of the rays. In winter some plank- tonic food was taken by some of the larger A^. ramsayi far offshore, and this brings us to a point that the existing data are not adequate to solve, namely, that it is highly probable that in this species, as in so many other larger demersal fishes (cf. hake, cod, Macruroniis, etc.), increased size and mobility is accompanied by a change-over from a carcinophagous to a fish diet. It is therefore probable that our approximate diagram really represents a sort of summation effect of the dietary of Notoihenia ramsayi over the whole of its life, and that had it been possible to examine stomach contents of samples of small fish and larger fish separately, we should find the organisms occupy- ing (roughly) the upper and lower halves of the diagram in very different proportions. I should expect the benthic invertebrates to be relatively more important in the smaller Notothenia, and fishes to predominate in the larger. Another notable feature is that no Miiiiida were recorded from the stomachs of Notothenia ramsayi, but it is highly probable that they would have been had it been possible to make summer observations on this point. Enough has been said to show that, if only on account of its small size, A'^. ramsayi is not likely to be of much value as human food in spite of its abundance within our area. The largest individuals of 35 cm. and upwards may attain a weight of a pound, and are not less palatable than the (rather insipid) larger Antarctic members of the genus. The average weight is little more than J lb. however, and the majority of the trawled fish are just too large to be fried and eaten as ' sprats ', a process that the crew of the ' William Scoresby' found most efficacious with the young fry some 10-12 cm. long. These were said to be '. . .exceedingly good, though the bones are rather hard. The flesh resembles that of whiting'. A^. ramsayi is of the first importance as a forage fish for larger and more useful species on the Patagonian Continental Shelf, being one of the main sources of food supply for hake and Macruronus, as we have already seen. Notothenia wiltoni Regan. This species seems to be one of the more extremely littoral Nototheniiformes, though it may depart from the shallowest waters in winter. In our collections with ' Other gear' it was taken between 2 and 35 m. of water at the Falkland Islands and in the Magellan region, the mean depth being but 5 m.: Port Stanley 45 (by A. G. B.) 56 2 (in BTS) Field Anchorage i (on LH) 222 i (m TNL) 55 I (in BTS) None was taken at the regular trawling stations. 15 336 DISCOVERY REPORTS Table 33. Summary of the observatiom on length frequencies of Notothenia ramsayi at autumn stations shown in Fig. 39, with relevant geographical data Station WS73 WS79 WS83 WS86 WS92 WS94 WS95 WS97 WS108 WS214 WS859B Date 6. iii. 27 51° 02' Position Lat. S 13. ni. 27 24. iii. 27 3. iv. 27 8. iv. 27 16. iv. 27 17. IV. 27 18. iv. 27 25. IV. 27 31.V. 28 25. iii. 31 oi|' 52 29 53° 53^ 51° 58i' 50° ooi' 48° 58' 49° ooV 48-31' 48° 25' 45° 14' Long. W 58° 55' 64° 59r 60° 07I' 60° 34I' 65° 01' 64° 57l' 64° 45' 61° 58' 63° 34' 60° 40' 61° 56' Distance in sq. m. From Main- land 226 126 232 176 125 107 68 From Falk- lands 13 78 > 100 o> 10 > 100 165 > 100 94 199 Depth m. 142 132 133 149 144 118 109 146 119 214 108 No. of N. ramsayi measured 82 124 131 lOIO 183 269 92 1616 166 1362 81 Remarks Very strong mode about 14 cm., hint of a submode at 22 cm. Fairly strong submode about 14 cm., mode at 22 cm. Strong mode at 17 cm. Strong mode about 23 cm. Strong mode about 22-23 cm. Hint of a submode at 16 cm., mode at 23 cm. Higher values tail off rather gradually, owing perhaps to larger males growing more slowly than females of similar age Submode at 14 cm., mode at 22 cm. Ill-defined mode — the values at 19-22 and at 24 cm. all high. This is readily understandable if the hypothesis as to slower growth of mature males is ac- cepted Two well-defined modes at 14-15 and at 22-23 ^^■ Strong mode at 27 cm. Here we seem to have a later year class (? Hl-group) dominant than at any of the other stations. Note that this is the only rich autumnal haul in deeper water Very strong mode at 15 cm. Table 34. Observations of stomach contents of Notothenia ramsayi in autumn and winter, and an approximate evaluation of the relative importance of the main food categories by arbitrary zveighting {data of Fig. 40) Food category Times recorded Food category Times recorded Food category Times recorded Sagitta sp. Nemertinea Nereidae Sabellidae Terebellidae Other Polychaeta I I 2 8 4 5 Mysidacea Serolis sp. Other Isopoda Parathcmisto Hyperiidae Other Amphipoda Euphausiidae Paralomis granulosa Eurypodius latreilli Other Brachyura 2 II 3 I I 15 I I 7 I Cephalopoda Ophiuroidea Holothuria Clupea Notothenia sp. Other fish I 2 I 13 2 2 Major (grouped) Times Weighting Relative Relative impor- food categories recorded factor importance tance as % Polychaeta 20 X I 20 6-7 Benthic Crustacea 41 x3 123 41-1 Other benthic invertebrates 4 X2 8 2-7 Planktonic invertebrates 4 X I 4 1-3 Cephalopoda I x8 8 27 Clupea 13 x8 . 104 34-8 Other fish 4 x8 32 107 DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 337 N. wiltotti is somewhat larger than the shallow-water nototheniids discussed previously, attaining a length of 34 cm. with individuals of more than 20 cm. fairly common. It is very easily confused with small specimens of N. ramsayi, and resembles A^. longipes Steindachner so closely that Norman (1937. PP- 8°~2) found it impossible to give complete synonymies for the two species, and believed that they might yet prove to be identical. However, Norman was able to give characters that should prevent confusion with the deeper-water A^. ramsayi in the future. Bennett's interesting notes on the occurrence of this species in shallow harbour waters at the Falkland Islands are also quoted by Norman. It is common inshore during summer, becoming scarce in autumn, when the gonads are enlarged. Hence Bennett concludes that its departure may be for Fig. 40. Approximate relative importance of the main food categories of N. ramsayi, the data being arbitrarily weighted as shown in Table 34. breeding purposes. N. wiltoni is known as ' rock-cod ' locally, but this vernacular name, perhaps the most promiscuously applied of all fish names m English-speaking countries throughout the world, is also eiven to several other Notothenia spp. in the Falkland Islands. Notothenia longipes Steindachner. We have just mentioned the possibility of confusion between this snecies and N. wiltoni. Such scanty distributional evidence as we possess favours the view that they really are distinct. All the specimens referred with any confidence to A^. lorigipes seem to come from the west coast of southern Chile or the western end of the Magellan channels, where they show a depth distribution rather similar to that of A^. canina at the eastern end of the straits (Fig. 42) and very different from that of the more exclusively littoral distribution shown by A^. wrltont and various other Notothenia spp. We took no N. longipes at regular trawling stations, but specimens obtained as shown 338 DISCOVERY REPORTS below gave an 'effective mean depth' of 43 m. with extreme range 12-78 m. Our largest specimen was 17 cm. long, but most of them were very small: WS582 I (on LH) WS583 20 (in BTS) Ringdove Inlet i (on LH) Notothenia sqiiamiceps Peters. The 'William Scoresby' did not capture any specimens of this small species, which seems to have an extremely littoral distribution at the Falkland Islands and in the Magellan region. The 'Discovery', however, obtained a few in the autumn of 1926, all from East Falkland and some with ripe eggs. None of these can have come from depths greater than 16 m. The largest specimen was but 12 cm. long: 53 3 (in RM) 55 I (in BTS) 56 3 (in BTS) Notothenia sima Richardson. This is yet another exclusively littoral species, taken plentifully by Mr A. G. Bennett with shore seines and traps in Port Stanley, and (rarely) by the ' Discovery' using the small beam trawl in very shallow water (10-16 m.); but it was never encountered at the regular trawling stations. Norman (1937, p. 85) quotes interesting observations by R. Vallentin that seem to show that this species spawns in littoral waters in spring or early summer, i.e. at quite a different time of year from that at which A^. squamiceps was taken with eggs. A^. sima is another small species. Bennett's largest specimen was 14 cm. long: Port Stanley (Nov. and Feb.) 24 (in seine and trap, A. G. B. co//.) cc 2 (in BTS) 56 I (in BTS) ^:> K I Notothenia cormicola Richardson. This species has been recorded at depths down to 35 m., but in the main it seems almost as exclusively littoral as the last two. It was never taken in the trawl, and with one exception (from Cape Horn) our few specimens were all obtained at the Falkland Islands. A^. cormicola resembles A^. sima very closely, but may be distinguished by the absence of scales on the lower part of the operculum (Norman, 1937, p. 85). Norman (p. 87) also quotes references from the literature that point to the possibility of an extended breeding season in A^. cornucola. The record of this species from New Zealand is extremely doubtful, as Norman has shown, but it is known from southern Chile (northwards to Chiloe) and the Magellan channels in addition to the localities where we obtained specimens. The largest A^. cornucola obtained by us was only 13 cm. long: 52 I (on LH) 222 I (in NRL) 53 I (in RM) Port Stanley Several (A. G. B. coll.) 55 I (in BTS) New Island (West Falkland) 6 (J. E. H. coll.) 56 I (in BTS) ' Notothenia elegans Giinther. This little species with its slender body and proportionately large fins, so well figured by Col. Tenison (in Norman, 1937, fig. 42), cannot easily be confused with any of the other Patagonian Nototheniidae. It was the smallest nototheniid that seemed regularly to inhabit moderately deep water on the plain of the shelf and, rarely, beyond. This is shown by the depth- frequency distribution (Fig. 42). N. elegans seems, moreover, to have a more northerly regional dis- tribution than most of the other species, having been recorded twice in the northern region, and more frequently in the intermediate than in the southern region in our catches. Too much stress should not be laid on this point, however, because on account of its shape and small size (we took none more than 12 cm. long) it is certain that our gear could not sample this species adequately. It was captured chiefly in ' Other gear ' or in the accessory nets attached to the back of the trawl. A^. elegafis is doubtless eaten by larger fishes, but is probably not sufficiently common to rank high in importance as a forage species : WS83 I WS808 3 WS861 I (in BTS) WS878 ii(inNR) |f^9J I 51 4(inOTL) WS863 26 (in BTS) ' WS237 I WS767 5(inNR) WS867 I (in BTS) WS795 1 WS836 6 (in BTS) WS873 I (in NR) DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 339 Notothenia macrocephala Gunther. Norman (1937, pp. 89-90) has shown that there seems to be no reason to doubt the identity of this species with examples from Kerguelen and New Zealand. It is one of two Nototheniids that seem to have a truly circumpolar «/6-Antarctic distribution. The silvery young seem to live pelagically for a much longer period than those of most other members of the group, and this may have favoured wide dispersal of the species.^ A striking illustration of this was afforded by the capture of nine specimens (up to 9 cm. long) at St. 63 in 48° 50' S : 53° 56' W, right out in the open South Atlantic some 300 miles NE x N of the Falkland Islands. These specimens were caught at the surface with a dip-net, and not by hand-lines as erroneously stated in Norman's report. We got no A'^. macrocephala in the trawl^the sizeable adults seem to be mainly littoral in their habits — but some were captured with 'Other gear', and Mr Bennett secured numerous specimens with seine and hand-lines in Port Stanley. Bennett's notes, quoted by Norman (loc. cit.), show that the larger individuals are common inshore at the Falkland Islands, where they are known as 'yellow- bellies ', and stay close inshore later in the year than most other Nototheniidae. He found them good eating though they are rarely used as food. At the Falklands they attain a length of over a foot and I lb. weight. The depth distribution shown in Fig. 42 refers to the larger individuals, taken only in littoral waters, and excludes the pelagic young which (as we have seen) may at times be met with far out at sea over oceanic depths. We obtained specimens of Nototheftia macrocephala as follows : 63 9 (in dip-net) 222 i (in TNL) Port Stanley 8 ( + several not preserved, A. G. B. fo//.) 229 i (in NiooH) From Phillipps (1921, p. 123) we learn that in New Zealand this fish goes by the name of 'Maori chief, but in Wellington, where a few were marketed in autumn, the fishermen know it as 'More- pork'. The fish should not be confused with the hairy owl {Ninox novae-selandiae) or the Tasmanian night-jar [Podargus ciivieri) that go by the same vernacular name. Notothenia microlepidota Hutton. We did not obtain any specimens of this species, which is, how- ever, of special interest because Norman (1937, pp. 90-1) believed that some Patagonian specimens, variously described, and authentic New Zealand ones, were identical. ' There is, thus, a second species common to the Patagonian and Antipodes regions.' Dissostichtis eleginoides Smitt. This is the largest of the Patagonian Nototheniidae. Superficially it bears a very strong resemblance to a hake, but Norman (1937, p. 92) found the skeletal relationship to Notothenia very close. It seems to be a rare fish in the Patagonian region, and we captured nine specimens only. Dissostichtis was always taken in the trawl, never in 'Other gear'. It occurred at such widely divergent depths that the depth relations could not be expressed by the methods used for other Nototheniiformes, and individual occurrences are plotted against depth in Fig. 42. The largest specimen, which was 90 cm. long and weighed 7710 g. (nearly 17 lb.), was taken in 297 m. At shallower depths the specimens showed regular increase in size with depth, ranging from a specimen of 13 cm. in 84 m. to one of 33 cm. in 172 m., but two five-pounders (about 64 cm.) were taken m one of the deepest hauls made (418 m.). It would seem that at most seasons only immature individuals of this species are to be found on the shelf, the larger fish ranging the deep water beyond the edge. They may migrate to shallower water to spawn, but we lack any direct evidence on the matter. Our larger specimens showed a steady increase in ponderal index with increasing length, from about 07 at 47 cm. to a value exceeding unity for the largest fish caught: WS75 I WS97 I WS98 3 WS245 2 WSSjg 2 1 It is quite probable that adequate search would reveal the presence of iV. macrocephala at Gough Island, the Crozets, Marion and Prince Edward Islands, and perhaps even at St Paul's Island and New Amsterdam. Our knowledge of the fish faunas of the isolated sub-Antarctic islands is deplorably fragmentary, especially in the Indian Ocean sector. 340 DISCOVERY REPORTS D. eleginoides is one of the few Patagonian fishes known also from the Antarctic Zone. It has been taken among the islands off Graham Land (Vaillant, 1906, pp. 36-9). Eleginops maclovinus (Cuvier and Valenciennes). This genus may at once be distinguished from others of the family Nototheniidae by the entire absence of a lower lateral line. E. maclovinus is a sizable fish, and appears to be exclusively littoral in its habits. We never captured any in the trawl. Specimens were secured with ' Other gear ' as follows : 724 10 (in seine) Connor Inlet 2 or 3 (on LH) WS586 i (on LH) In addition to these, Mr A. G. Bennett collected six specimens for Norman's report by seine-netting in Port Stanley harbour, mostly at Weir Creek. Bennett also provided valuable notes on the habits of the fish which Norman (1937, pp. 93, 94) quotes at length. The species is known locally as 'mullet', and this vernacular name is justified not only by its strong superficial resemblance to true mullets {Mtigil spp.), which are absent from the region, but also by its habits. Notable among these is its tendency to run right up into the mouths of fresh-water streams on the last of the rising tide. Bennett tells us that it may grow up to 2 ft. long, but unless this length is at times considerably exceeded, his figure of 15 lb. for the maximum weight, quoted by Norman, must be a mistake. A fish 2 ft. long and 15 lb. in weight would have a ponderal index just over 3-0 (calculated from K=w (g.)// (cm.)=^x 100),^ and it is clear that no fish approaching the proportions of Eleginops could give even half this value. Authentic weight records of smaller specimens give ponderal indices from 0-69 to 0-85, and if we assumed an index of i-o for a 2 ft. specimen its weight would be just 5 lb. Conversely, even if we assumed an index as high as 1-25, a 15 lb. fish would be no less than 32 in. long. I believe that in all probability Bennett actually wrote 5 lb. and that some error crept in subsequently. Eleginops is eaten quite frequently at the Falkland Islands, but often has a muddy taste. Otherwise it would be a promising subject for small-scale local exploitation by seine-netting. It extends round both coasts of the mainland of South America from the River Plate in the east to northern Chile in the west — much farther towards the equator than most other Nototheniidae. HARPAGIFERIDAE In his later report on the coast fishes of the Antarctic Zone Norman (1938, p. 43) places Harpogifer, with four other (exclusively Antarctic) genera in this separate family and not, as heretofore, in the Nototheniidae. The chief characteristic of the family is absence of scales on the body. Harpagifer bispinis (Schneider). This is the only member of the family found in the sub-Antarctic Zone. To the southward it has a wide distribution in the northern part of the Antarctic Zone, having been recorded from Graham Land and almost all the isolated island groups (Norman, 1938, pp. 52-3). Norman's description of it as mainly littoral (and frequently intertidal) in habit, applies accurately enough in the Patagonian region, though even there we have taken it down to 95 m. ; and the depth relations, shown in Fig. 42, show it closer to the 'first-slope' dwellers than to the exclusively littoral species of Nototheniidae. Farther south, however, where the intertidal zone is usually small and subject to ice action, Harpagifer usually occurs at greater depths, although it is true enough that it ' The high 'condition factors' quoted in some salmon literature (of the order 36-40 or more) are obtained from the formula w (lb. & fractions)// (inches)^ x 10,000. Menzies' Scottish 'coefficients' make use of the same heterogeneous British units, but get rid of the unwieldy decimal ciphers resulting from the first term of the formula by dividing by 0-00036, a figure just below the mean for 'normal' east of Scotland salmon. This has the effect of bringing all the values close to unity (and close to those obtained by direct use of the metric formula). His system is perhaps ideal so long as we wish to consider salmon only (and only east of Scotland salmon !) but the principle of dividing by the mean value implies that the factors for any given species (or local race) of fish will be grouped close around unity. If we wish to visualize the difference in the ratio weight to cube of length between fish of diverse form it is not possible to use his method, whereas direct application of the metric formula permits this, and with less heavy arithmetic. A propos the immediate problem above, Menzies' figures (1925, p. 190) show that a 24 in. salmon should weigh about 5 lb. 2 oz. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 34i keeps close in to the land. In the Patagonian region we captured specimens of this very small species at one trawling station only: WS89, in shallow water, where three were taken in one of the ' accessory nets'. Other specimens were obtained for Norman's 1937 report as follows: Port Stanley 24 (under stones, A. G. B. coll.) New Island (West Falkland) 3 (J. E. H. coll.) WS749 3 (in NR) WS832 I (in NR) None of these exceeded a length of 10 cm. If readily available between tidemarks at the Falkland Islands, Harpagifer might be useful as bait. CHAENICHTHYIDAE Champsocephalus esox (Gunther). This species occurred with some frequency in trawl catches on the rough ground near the Falkland Islands in the southern region. A single large specimen was taken farther north, at St. WS97 in 49° S: WS71 30 WS73 6 WS75 29 WS76 I IVS81 I WS83 41 WS84 2 WSgs I WSgj 2 WS97 I WS756B 2 WS802B I WS823 5 WS834 WS837 51 724 Port Stanley 22 (in OTL) 2 (in seine) 3 (A. G. B. coll.) Most were taken in autumn, and none during the winter survey when a large proportion of the stations were worked in deep water beyond the shelf edge. The species is, indeed, very much an inhabitant of the plain of the shelf, judging by the depth distribution shown in Fig. 42. From this it can be seen that while a few have been taken in shallow LENGTH. CMS littoral waters there was only a single specimen from beyond the shelf edge. From Bennett's notes, quoted by Norman (1937, p- 96), it would seem that if there is any inshore migration of Champsocephalus, it will take place in late summer or autumn, for his records of the infrequent capture of the species in littoral waters at the Falkland Islands all date from that season. The length frequencies of our autumn-caught speci- mens (Fig. 41) show two very strong modes at 15 and at 28 cm. If these indicate year classes it follows either that this species is of extremely rapid growth,^ or that ''"■"■ r;£r^;l*?;iXTi.°r""'"- a„ intermediate year class ts entirely lacking in our samples. The presence of such an mtermediate year class would indeed, bring the growth rate into line with that observed for Notothenia ramsayi, a related fish of much the same size, but it is extremely difficult to see why such an intermediate age group should be absent from the grounds frequented by both younger and older fish of the same species. Moreover in view of our extensive observations, coupled with those of Hamilton and Bennett in littoral waters, we should certainly expect to have found evidence of such a group somewhere if it actually exists. . „ • u Champsocephalus esox, which has near relatives in the Antarctic Zone, is excellent eating: much firmer and of better flavour than most Nototheniiformes. Unfortunately, we rarely captured it in any great quantity, and our largest specimen was only 36 cm. long, so that it is unlikely that the 'pike', as it is called in the Falkland Islands, could be exploited profitably. 1 14 cm. in its second year. Among better-known fishes of similar size such a rate is approached by estuarine pollack and whiting, which later descend to the sea (Hartley, 1940, pp. 47-5°)- 342 DISCOVERY REPORTS SUMMARY OF OBSERVATIONS ON PATAGONIAN NOTOTHENIIFORMESi The survey of our observations on the distribution and bionomics of the individual species of Pata- gonian Nototheniiformes enables us to present some of the main features regarding the group as a whole in more concentrated form. Table 35 gives a list of all the species known from the area up to the time of the publication of Norman's report, and shows which were captured in our trawls or with 'Other gear', with brief notes on records of occurrence outside the area investigated. Table 35. List of the Patagonian Nototheniiformes, their occurrence in our material and their distribution outside the area surveyed Page references Family Species in Norman (1937) A B C Distribution outside the area Bovichthyidae Cottoperca gohio (Giinther) 63-65 + + — Southern Chile Bovichtus argentinus MacDonagh 65 — — + Northwards to La Plata Nototheniidae Notothenia macrophthalma Norman 68-69 + — — Holotype only, not yet known else- where N. trigramma Regan 69 — — + Holotype only, not yet known else- where Magellan Straits A'^. canina Smitt 69-70 + + — A^. jordani Thompson 71-72 + + — Magellan Straits A^. tessellata Richardson 72-73 + + — Magellan Straits, and southern Chile north to Chiloe N. brevicaiida Lonnberg 74-75 — + — Magellan Channels N. giintheri Norman 75-76 + + — Not yet known elsewhere A'^. ramsayi Regan 76-80 + + — Not yet known elsewhere A'^. wiltoni Regan 80-81 — + — Magellan Straits A^. longipes Steindachner 81-82 — + — Magellan Straits, southern Chile A^. squamiceps Peters 82-83 — + — Magellan Straits N. sima Richardson 84-85 — + — Magellan Straits A'^. cormicola Richardson 85-87 — + — Magellan Straits, and southern Chile north to Chiloe, ? N.Z.* N. elegans Giinther 87-88 + + — Magellan Straits N. tnacrocephala Giinther 88-90 — + — Circumpolar sub-Antarctic N. microlepidola Hutton 90-91 — — + Circumpolar, but not at Kerguelen Dissostichus eleginoides Smitt 91-92 + — — Magellan Straits, Graham Land Eleginops maclovitius (C. and V.) 92-94 — + — Up to River Plate on the east, and to northern Chile on west coast Harpagiferidae Harpagifer bispinis (Schneider) 94-95 + + — Circumpolar northern Antarctic Chaenichthyidae Champsocephahis esox (Giinther) 95-96 + + " Magellan Straits (near relatives Antarctic) A = taken in ' Trawl + accessory nets'; B = taken in 'Other gear' or by shore parties; C = not taken by the Expedition. * See Norman (loc. cit.) for the doubtful records of this species in New Zealand. In their regional distribution within our area, the group as a whole is a southern one. Of the deeper water species only the dominant Notothenia ramsayi, with Cottoperca gohio and Notothenia elegans were recorded in the northern region ; and Cottoperca was much more plentiful in the southern region, while Notothenia elegans seemed to find its optimum in the intermediate region. N. guntheri appeared twice in the intermediate region, but 49° S was its northern limit among our observations, which was true also of Dissostichus and Champsocephahis, characteristically southern genera. Little can be said concerning the probable northern limits of the littoral species, owing to unavoidable lack of observations in Argentine territorial waters. Eleginops and Bovichtus are known to range far to 1 This supra-family grouping is convenient for purposes of this summary. It includes the families shown in the Table with other exclusively Antarctic ones. Elsewhere in this report I have avoided using such groupings, as many of them involve questions of classification that are still unsettled. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 343 the north, and W. F. Thompson took Notothenia jordani in Grande Bay. We took young A^. macro- cephala in 48° 50' S in the open ocean, but large individuals were obtained by us only in southern coastal waters. In New Zealand, however, adults of this species are known to range as far north as the southern coasts of North Island. All the other shallow-water Nototheniiformes were taken (by us) only in the southern region. The Nototheniiformes are preeminently an Antarctic group, completely dominant over all other fishes in such small areas of the vast oceans south of the Antarctic convergence as are sufficiently shallow to support any coastal fish fauna. Their abundance and variety in the Patagonian region and their tendency to be distributed mainly near its southern limits, are consistent with the view that they have spread northwards, but we have seen that only Harpagifer and Dissostichus have species common to both zones. It thus seems probable that the Patagonian fish fauna has been insulated against invasion from the south for a very long time, and that the hydrological barrier of the Antarctic convergence provides too big a contrast in environment for the majority of such fishes to overcome. Quite early in these investigations, when attention was focused on large species (notably hake) that migrate over long distances in a comparatively short space of time, we found it impossible to gain much by the study of depth distribution because of the very slight gradient on the plain of the shelf. This necessitated laborious calculations of the distance of each observation from the coast before we could attempt to follow the movements of such fishes. With the Nototheniiformes it is quite otherwise. Here we have a group consisting for the most part of fairly small bottom-living fish with limited powers of movement, and the study of depth distribution has helped a great deal in our attempts to gain some insight into their probable way of life. The depth relations of the Patagonian species are summarized in Fig. 42, and the differences m the mean depths recorded are given in Table 36, with data sufficient to determine their statistical signffi- cance. The figure shows the effective mean depths and extreme ranges observed for all the species except Dissostichus eleginoides, our few specimens of which were so widely dispersed as to demand individual plotting. The distribution of all the species descending below the 50 m. level is also in- dicated by the black polygons. The widths of these are proportional to the relative abundance of each species within each 50 m. depth grouping. The littoral species are further indicated by stippling of the rectangle covering the whole of the observed depth range of each. It will be seen that the species can be divided into three main groups according to their depth distributions : I Deep-water species found mainly on the plain of the shelf and rarely beyond the shelf edge: Cottoperca gobio, Notothenia guntheri, N. ramsayi, N. elegans, Dissostichus eleginoides and Champso- cephaliis esox. , II ' First-slope' dwellers, rarely descending below 100 m., where the plain of the shelf may be said to begin: Notothenia canina, N. jordani, N. tessellata, N. hngipes and perhaps Harpagifer hsprnis. III Exclusively littoral species: Notothenia brevicauda, N. zviltoni, N. squamtceps, N. stma, N. cornucola, N. macrocephala (adults) and Eleginops maclovinus {Bovichtus argentinus may fit m here). Within the first two groupings it proved possible to recognize further distributional trends, either regional or bathymetric, which serve to differentiate the species still further. As a result of this (always excepting the dominant and ubiquitous Notothenia ramsayi) most of the species show a distinctive distributional pattern that tends to minimize territorial overlapping between them. The possible ecological ' advantage ' of this in lessening competition between species of similar size, may be one of the factors that has led to the slight modifications in food requirements and general habits that it must have entailed. 344 DISCOVERY REPORTS 1= 9 L 3 IN t p II " • • ii • • « • • i 5 i == Pi ^ * ^ m - 1^ V NOTO- THENIA L0N6IPE5 > NOTO- THENIA WILTONI NOTO- THENIA RAM SAY 1 w K- r= ( > 1^1 i 1 ^ % H "^ ^ rfi s9 il^ H ft- m 9 W DEPTH GROUPS IN METRES s o o o 1 O LH (\i 1 O o o in o in hi ) o o o 1 in o f o o o o o Ml 4-1 4".2° ?.2° ?^.y= I CO I CO I CO I CO CO :2.2> + CO ^2" -0.2° _|.CO +CO 00 p) . "^ bVi "^ bO -CO bo + 'co ~t-° S.2> 1 z +^ Z fZ vO 2 6b Z +«^ I o z C;.23 ^.25 °.2f CO I CO .:? tb ^ M ^.2) S .2> ^.25 ^.2' S .2) ^.2' ^.2° - y" + CO I I CO +CO +C0 I CO CO vo'.S" vd -XT' vd .s^ ■*." "o ._ I CO I CO I CO +CO I CO luvpxoC -jV vumvo viu3i^tofopj OtqoS V3X3 "^ 2° CO S.£P "^.2° {0.2' ~ I CO I CO I S.2° CO + 03 ^, bo + ^ I ^ <^.2° CO I CO I cio I CO I CO Af/1^ vO OS + 00 00 0 Ov 00 Ov VO Ov 00 VO vO b o 00 ov 00 o -*- o m vo ov m ov 34S 1.2= S.2> I CO I CO oJ.y + CO -I 2.2= + Z +CO •+K -CO VO o ::;.2' S.sp 2.£P I CO I CO I CO >^^ s;.M Z +CO Tz + Z +CO I + Z +CO ^.2= ^.2= vS.23 I CO I CO 4- CO Z 4-Z +CO CO I CO I CO 00 I CO I CO I CO Ov.ir + CO vo'.a' V0.2" ^.2= I CO I CO +CO ?,.23 ^.2° S>.£f I CO I CO +CO J] bb J\ 60 lz j / f A very closely allied form, G. capensis, may yet prove specifically identical with G. blacodes (ct. Norman, 1937, p. 113). G. capensis, known as 'king-klip', is an important fish m South Africa Most of them are obtained by trawling in rather deep water, and aUhough they form only about 2 /o by weight of the total landings, they fetch prices considerably above the average. Thus they repre- sented 2-o-4-5% of the value of the fish landed during three years 1929-32. and the sums realized at first sale ranged from ^C 11,000 to just over £20,000 (S.A.). Evidently the edible qualities of the genus are appreciated wherever they are to be found, but they cannot be a cheap fish to catch, for they are not much given to shoaling. Numerous trawling records off South Africa (von Bonde, 1934, pp. 42- 63) show 'king-klip' in consistently small numbers where they occurred at all, though large catches of shoaling species like Merluccius capensis were being made. In the area of the trawling surveys Genypterns blacodes was taken in very small numbers throughout the year but the data are far too scanty to permit of any detailed consideration of the bionomics of the species The distributional data are, however, very interesting and sufficient to suggest two main trends of movement, supporting the view that most of the G. blacodes taken in the area are seasonal southern stragglers from a stock inhabiting warmer waters, farther to the north. The records are: WS78 T, WS783 I WS811II 3 WS7Q T. WS785A I WS812II I WSos I WS789 3 WS816 I WSq8 I WS792A I WS817A I WS108 I WS792B I WS819B I WS214 2 WS794 3 WS^4S I WS216 S WS795 6 seen to escape WS217 4 WS797B I ^SSjo 1 WS218 13 WS809A I WS875 2 WS772 I WS8IO 2 , TU\ WS773 4 WS8nI I WS586 3(onLH) •^G,!jj(, I Connor inlet i (on LH) It can be seen that apart from four specimens caught on hand-lines in the western channels, outside the area of the trawling survey, all the specimens were taken in the trawl. . r ,t. In Fig 43 the records are shown on three seasonal charts with a schematic representation of the trends of movement they are believed to indicate. In winter and spring (a) the species seerned to be confined to a small area of deep water near the edge of the shelf, and near the northern boundary assigned to our intermediate region. In summer (b) the records suggest that these fish tended to move into slightly shallower water, and also southwards. At the same time it seemed that other Genypterus blacodes invaded the area from the north, working south along the 80 m. 1^-e m^he relaUvely warm inshore counter-current. Where this current peters out, in the northern part of our 3S6 DISCOVERY REPORTS r '10 O €%""-■■■■ s c o 3 Q M E O O u O) c ■t.r^. ^ c in (J QJ (- L. IT -? bo c 3 bo DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 357 c o CL. a. < I ^ O 3 .0 U' ,v>Q°? O.'^/^'^'^- if-p-t^i^- o ■'"6 O./ ;0\ a B 3 *^ 3 •« p I 3 CiO 358 DISCOVERY REPORTS southern region, there was in late summer a strong indication of offshore movement, that continued into autumn (c) as one would expect if Genypterus moves offshore in winter like so many other demersal fishes. The measurements — too few for detailed consideration — show extremes of size in the north, while individuals caught in the south were more uniform. Possibly the southward movement down the coast and out again is mainly confined to adolescent individuals. Really small young individuals were taken at one northern station only. The interesting point that seems fairly certain from the distribution observed is that Genypterus seemed to avoid the central plain of the shelf, being found mainly along the lines of steepest gradient, i.e. the ' first slope ' inshore, and the edge of the shelf to the north. Our only captures of the species on the central plain were near its southern limits in autumn, when those that have worked down the coast are believed to be seeking deeper water. This they seemed to succeed in reaching in the deeper southern portion of the Falkland trough, and since they have not been seen at any of the numerous stations worked in the immediate neighbourhood of the Falkland Islands I conclude that if they further succeed in regaining a more normal, northern habitat, it is by way of the Falkland trough that they do so, as indicated by the dotted portion of the line on the diagram (Fig. 43 d). BROTULIDAE Cataetyx messieri (Giinther). Two small specimens of this rare fish were taken in deep water beyond the edge of the shelf, one in the intermediate region and one farther south. It has also been recorded from Messier Channel, Chile, and South Africa. A specimen 63 cm. long has been taken in very deep water off Cape Point. E. R. Gunther considered that the normal habitat of the species lies below the depth hmit of ordinary trawling : WS248 I WS773 I CENTROLOPHIDAE Seriolella porosa Guichenot. This species was only captured at a single northern inshore station, WS853 (8), worked early in autumn. The French found it common in Orange Bay, near the eastern end of the Magellan Strait in 1883 (Vaillant, 1888, p. C30), and it might therefore be expected to occur throughout the length of the shelf. Endeavouring to explain our lack of evidence of such extended distribution, E. R. Gunther noted that its shape and colouring suggest fast swimming near the surface, which might help to account for its being missed by the trawl. Alternatively, it may be a strictly coastal species rarely moving offshore so far as the main trawling grounds. The species is also known from New Zealand, and from Tasmania, where Neptonemus dobiila Gunther was the synonym used by Johnston (1891), and Seriolella douhla^ Gunther the synonym used by Lord (1923, p. 66). These writers tell us that it is known as 'trevally' or ' mackerel-trevally ' in Tasmania. The closely allied Seriolella brama Gunther, the 'snotgall' or ' snotgall-trevally ' of Tasmania, is also sold under the name of 'trevally' in Christchurch and Dunedin, New Zealand, but in that country the name is more usually (and correctly) applied to Caranx platessa Cuvier and Valenciennes. It is also noteworthy that in Tasmania a true carangid (C. georgianus, Cuvier and Valenciennes) is called ' silver-trevally ' (Lord, 1923, p. 67), while the usual New Zealand name for Seriolella is 'silverfish', which can of course lead to immediate confusion with atherines in that and most other countries, 1 A misprint. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 359 The etymology of the name ' trevally ' is exceedingly interesting. It appears to be derived from the Portuguese cavalha, a horse or horse-mackerel (spanish caballa), names which are still applied to Caranx in their countries of origin. In the Gulf of Mexico and eastern Florida this name applies to Caranx hippos (Linnaeus) but has become corrupted to 'cavally', and even to 'Horse crevalle'^ in South Carolina. As a further complication on the ichthyological side, ' crevalle ' is sometimes applied to Trachynotus carolimis (Linnaeus), although this species is widely known as the ' common pompano' in the eastern states. I suggest that 'crevalle' became readily corrupted to 'trevally' and, as applied somewhat loosely to carangid or closely allied fishes, reached the southern hemisphere with the New England whalers and sealers early in the last century. Whether the final corruption took place en route or after arrival must remain a matter for speculation. Palinurichthys coeruleus (Guichenot). Norman (1937, pp. 116-18) tentatively refers both this and the next species to the genus Palinurichthys, remarking that the genera of Centrolophidae need revisioi]L. It was with some doubt that Norman identified our two specimens from the centre of the shelf with Guichenot 's Seriolella coenilea from Juan Fernandez off the west coast, owing to vagueness of the original description : ^^g^^ ^ ^^^^6 ^ Palinurichthys griseolineatus Norman. Our specimens of this new species, which would seem to be rare, all came from the centre of the shelf. I have found a note of E. R. Gunther's which states that: ' Its colour in life is of a delicate blueish and silvery gray : the gray running along the sides in horizontal undulating bands which divide and merge. The reference by Norman (1937, p. 117) to brownish and yellow has doubtless resulted from staining by teak and alcohol used in storage and transport.' WS75 I WS97 I WS108 I STROMATEIDAE Stromateiis maculatus Cuvier and Valenciennes. This fish is more likely to become of value as food for man than most of the others found within the area of the trawling surveys. Its flesh is rich and well flavoured (other members of the family being also noted for their pleasant taste, and the high fat and mineral content of their flesh) and free from too many small bones. In South America, where Norman (1937, p. 118) tells us that it is called 'pampanito', the species seems to have acquired an undeserved reputation for causing gastric disorders when eaten, and some hint of this prejudice may be found in the Falkland Islands.'^ The crew of the 'William Scoresby' ate Stromateus with relish and without any ill effects, and from what is known of closely allied species that have long been exploited on a large scale in the eastern U.S.A. and in China, I have no doubt that its ill repute is due merely to more rapid decomposition (when stored too long under primitive conditions) than species of lower food value. There seems to be no English vernacular name for this fish, and I would venture to suggest 'spotted pomfret' as most suitable; alternatives based on other common names applied to members of the family elsewhere would probably leave more scope for confusion, as will be shown later. Stromateus ranked third in total weight of the fishes captured during the third survey, and fifth in total numbers during all three surveys. Although of uniformly small size they are nearly all potentially saleable, owing to their compact bream-like shape. They commonly range from 10 to 13 in. (25-34 cm.) in length, and from 7 oz. to i lb. (200-470 g.) in weight. A very high proportion of the fish (almost certainly over 60%) is edible. Analysis of the closely allied Chinese species Parnpus (Stromateoides) argetiteus (Euphrasen) shows 64% edible (Read, 1939, p. 45). Our largest specimens of Stromateus maculatus measured 38 cm. 1 Truly a work of supererogation ! . , ^ „ , , 1 u . c 2 Stromateus, however, seems rarely to penetrate to the immediate vicinity of the Falklands: we took them at tour only of the many stations worked within 100 miles of the islands, and the nearest was 75 miles away. 18 360 DISCOVERY REPORTS Although we took only ahout a thousand of this species during the three surveys, we have strong evidence that many more large hauls of it could have been obtained in summer (when it tends to shoal inshore), had we been concentrating upon it as a commercial fishing vessel would have been able to do, instead of sampling the whole area. During the three trawling surveys, S. maciilatus was captured in the ' Trawl + accessory nets' at fifty-two stations, as shown below. It was never taken by us with ' Other gear ' :^ WS78 20 WS792B 13 WS817A 4 WS97 8 WS794 2 WS817B 2 WSI08 22 WS797B 2 WS833 203 WS2I4 I WS797C 8 WS838 2 WS2I7 4 WS798 10 WS847A 146 WS762A 9 WS799B 6 WS847B 100 WS763 I WS800A 32 WS848 27 WS764B 4 WS800B 7 WS84g 26 WS77I 2 WS806 I WS850 2 WS784 4 WS809B 92 WS8S3 3 WS785A 6 WS810 3 WS8s8 I WS785B 7 WS811I 21 WS8S9B I WS786 3 WS811II 58 WS862 I WS788B I WS812I 19 WS864 5 WS790A 14 WS812II 40 WS866 18 WS790B 19 WS813 52 WS868 I WS79IA I WS814 I WS79IB 7 WS815 2 The distribution of these catches at different seasons of the year is charted in Fig. 44. From this it can be seen that in spring (a) a few Stromateus were taken on the shelf in the northern region, and that there a few were already right inshore. The species was not observed on the single line of stations worked in the intermediate region, nor at a couple of odd stations worked farther south at that season. From this it may be concluded that in spring some shoreward movement has begun, but probably little southward movement. In summer {b) the species was much more frequently taken and in very much larger numbers, especially to the southward. Several really good hauls of Stromateus were made inshore, from Cape Virgins northwards to a point off Puerto san Julian. Moderate numbers were taken at several stations on the plain of the shelf, but the species was not taken in the deeper water to the southward. It was also absent from all but one of the catches at the outermost stations (those worked nearest to the shelf- edge) on the northern and intermediate lines, and from the immediate vicinity of the Falkland Islands. It would seem clear that the species works southwards inshore during summer, and reaches maximum concentration (shoaling) at that season. In autumn {c) some Stromateus were caught still inshore in Grande Bay in the southern region, but this was in only one of the two years for which we have autumn data, and the location of most of the catch was farther offshore (though still on the plain of the shelf) and to the north. The distribution IS clearly compatible with the hypothesis of dispersal of the southern summer inshore shoals in that direction. Stromateus was captured at two stations only during the winter survey {d). These were right offshore on the shelf edge in the intermediate region, and the numbers of individuals were but one and four respectively. None was taken at any of the numerous stations worked on the plain of the shelf and to the southward at this season. These observations clearly point to continued dispersal offshore and to the north. Perhaps the ' spotted pomfret ' should be regarded as a semi-pelagic wanderer in winter. Its seasonal 1 Cf. Goode's remark of Poronotiis triacantlius: 'it has never been known to take a hook.' Quite recently, however, some Ime-caught Poronotus were reported from Florida. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 361 B B S e &0 3 Xi c o en a CO $ 'id CO 3 XI « c o S3 v en a $ l8-2 362 DISCOVERY REPORTS DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 363 movements find a close parallel in those of its near relative Poronotus (Stromateus) triacanthus (Peck) on the north-eastern seaboard of the United States. This 'butterfish' finds its optimum from New Jersey to Chesapeake, and its shoreward congregations in summer may be even more pronounced than those of the Patagonian species, for pound nets are the main means of capture. Near the northern limit of the butterfish, considerable quantities are trawled in Massachusetts ; and in more southerly waters off North Carolina, where it was previously only known from the summer inshore fishery, •400H •350- 1-300- ^ l'250- i-zoo- M50 DECEMBER j JANUARY | FEBRUARY | MARCH Fig. 45. Stromateus maculatus: seasonal variation in ponderal index, and corresponding mean lengths. small quantities have been trawled offshore in winter (Pearson, 1932, Table 2, p. 18). Thus the butterfish compares with our spotted pomfret not only in shoaling shorewards in summer, as so many useful fishes do, but also in being trawlable near the colder limits of its geographical range, and in wide dispersal to warmer and deeper waters in winter. A first step in further efforts to elucidate the bionomics of Stromateus maculatus was to test the seasonal variation in ponderal indices for indications of the spawning season. Weight data were available only from a limited period during the third survey, and average K for either sex, plotted at mean dates, yielded the values joined by the heavy lines in Fig. 45. 364 DISCOVERY REPORTS This shows, first, a 27% rise in average K values between December and the end of March, with httle discrepancy between the sexes. Secondly, an apparent drop in K values in February, to a level which was, however, still well above the December figures. This was found to be associated with a drop in mean lengths (thin lines in Fig. 45), i.e. with an increased proportion of smaller fish in the population sampled in February over that of the population sampled in January. It seems fair to conclude that most spawning takes place in early summer, and further (though at this stage only as a working hypothesis) that larger fish spawn first and therefore recover condition earlier than smaller ones. The smaller fish shoal inshore later than the larger ones, as will presently be shown ; and any mature individuals among them are presumably correspondingly later in spawning. But it is also quite possible that a proportion of the later, smaller shoalers are immature. If so they may still tend to decrease the mean February K if, like adolescent hake, they show seasonal harmonic variation in condition like mature fish, but on a lower level. 1350- 2 hi 1-300- 2 1-250 ■ --0 [8:37] \^:46o] — \ — 25 50 75 iOO MILES FROM THE COAST 125 150 Fig. 46. Stromateus maculattis: variation in average K with distance from the coast; southern region, summer. Figures in brackets denote numbers of weighings, with the numbers of fishes measured in iUilics. The further testing of these hypotheses as to the movements of S. maculattis, from the available data, depends mainly on considerations of size distribution (length frequencies) in relation to distance from the coast and at diflFerent periods; but one further application of the use of K seems to help, and may be described before we proceed to the other evidence. In Fig. 46 the mean K values of Stromateus taken at different distances from the coast in the southern region, in summer, are shown joined by a line (merely to guide the eye). It will be seen that the values increased sharply with increasing distance from the land. This shows that the ofl'shore fish had had longer to recover after spawning, for it is reasonably certain that the converse explanation (inshore population spawning while that off'shore had yet to do so) could not hold, for we have already seen that the general rapid increase of K values throughout the summer suggests that most spawning takes place early. The point becomes even clearer when the length-frequency data are considered. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 365 In Fig. 47 the length-frequency distributions of Stromateus taken inshore in the southern region, in successive summer months January and February, are shown as histograms. Here we see that in January the distributions were strongly unimodal, with the mode at 31 cm. (males) or slightly higher (females). In February the distributions were clearly bimodal for both sexes, with the modes at about LENGTH5_CM5 Fig. 47. Stromateus maculatus: seasonal variation in percentage length frequency at inshore stations in the southern region. 25 and 30 cm. for males, 26 and 31 cm. for females, the smaller fish being more numerous than the larger. This proves the point made above, that the smaller fish moved inshore later than larger ones. It will also be shown later (Fig. 50) that in the intermediate region the length frequency of inshore Stromateus around midsummer resembled the February distribution in the southern region. The smaller fish in particular are thus mainly confined to the warmer inshore counter-current in their movement down the coast, but also the whole of the seasonal cycle is probably centred earlier m the 366 DISCOVERY REPORTS year farther north. Some of the northern and intermediate fish probably do not proceed so far south before they begin to move offshore again ; and indeed it can be shown that the mean lengths of southern samples were significantly greater than those of fish from the intermediate region. We cannot deter- mine the relative extent of meridional and on- and offshore trends of movement, owing mainly to dearth of inshore data from the intermediate and northern regions early in summer, particularly from the Golfo san Jorge; but it seems certain from the general distribution observed that both trends exist together. The disposition of our catches of S. maculatiis in the southern region in summer, the time of maximum concentration, is shown diagrammatically in Fig. 48, in relation to distance from the main- land coast. This arrangement of the data gives some idea of what a practical fisherman in quest of the species could expect, in addition to amplifying our ideas as to its general bionomics. It is very evident that the species was most plentiful close inshore, but well distributed up to 100 miles from the land. It was rare farther offshore at this season. The sex ratio was highest inshore ('normal') and the proportion of males diminished as one proceeded seawards. The males of this species are slightly (but in large samples significantly) smaller than the females, and it is just possible that this difference has some small effect on their mobility, for we know that in other fishes where the difference in size between the sexes is more pronounced, the larger females travel farther and faster than the males. Here, however, the sexual dimorphism is so slight that one would expect that some factor not yet determined, such as more rapid dispersal after spawning among males, must be mainly responsible for the diminished sex ratios in samples of the offshore population.^ In Fig. 49 the percentage length frequencies of both sexes of Stromateus for the same area and period are considered in relation to distance from the coast, the observations being grouped according to the same distance intervals as were used in constructing Fig. 48. The inshore grouping showed bimodal distribution, with modes at about 25 and 29 cm. for males, 26 and 3 1 cm. for females. These probably indicate year-classes. There was an increased tendency towards suppression of the smaller mode as one proceeded off- shore, until, at distances of more than 100 miles from the land, such few fish as were caught belonged almost exclusively to the older (larger) year class. Within a brief period around midsummer 193 1-2, a series of observations were obtained that fall naturally into two groups, serving to show the strong contrast in the population of Stromateus inshore and that found offshore in the intermediate region, in respect of length-frequency distribution (Fig. 50). The inshore grouping showed bimodality with the smaller length-class dominant; the off- shore grouping was unimodal, with almost complete suppression of the smaller class. There are still some big gaps in our knowledge of S. maculatiis, for pressure of work in the field upon other, more important species, made it impossible to collect routine data on the condition of the gonads, for example. We have very little idea of its diet, though some were found to have fed upon Parathemisto. Hake have twice been found to have preyed upon Stromateus, but we can only guess at its other natural enemies in this region. The main features in the bionomics of S. maculatus that seem clear from our data are : a double trend of seasonal movement, inshore in spring and summer with maximum concentration, and offshore in autumn and winter with maximum dispersal ; and, superimposed upon this, a meridional move- ment southwards in summer and northwards in winter. A schematic representation of these move- ments is given in Fig. 51. 1 Probably females have somewhat greater food requirements than males during the recovery period, with a resultant tendency towards greater local concentration where food is plentiful (though such concentrations would not be comparable with the spawning shoals). This would lead to the same result. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 367 LESS THAN 25miles FROM LAND 26 -50mile5 FROM LAND 51-IOOmiles FROM LAND MORE THAN IOOmiles FROM LAND Fig. 48. Stromateus maculatus: relation to distance from coast in the soutiiern region in summer. A. Total fish taken m each distance grouping. B. Number of hauls (positive black, negative white). C. Number of fish per positive haul, showing also the sex ratio. White segments indicate the percentage of males. D. Hours trawling within each distance groupmg. E. Number of fish per hour's trawling. D XXIII 19 368 DISCOVERY REPORTS LENGTH5 CMS 5 O en o o o H Fig. 49. Stromateus maculatus: percentage length frequencies at different distances from the land, southern region, summer. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 369 Fig. 50. Stromateus maculattts: variation in length frequency with distance from the coast, intermediate region, 17. xii. 31 to 9. i. 32. 19-2 370 DISCOVERY REPORTS Stromateus probably spawns early in summer, but we do not yet know how prolonged the spawning season may be. It seems clear, however, that the larger, older fish lead in the migratory movements, and spawn earlier than any mature individuals among the smaller ones. The summer shoreward movement of the spotted pomfret finds a close parallel in the behaviour of the most valuable of its northern relatives, Poronotus triacanthus (Peck), off the southern New England, middle Atlantic and Chesapeake Bay States of America. With Stromateus the superimposition of a marked meridional trend of movement is probably occasioned by the necessity for maintaining a species with a prolonged pelagic post-larval phase within its ecological norm (cf. E. S. Russell, 1937, p. 321) and directly correlated with the current system of our area (Fig. 4).^ Any similar tendency in Poronotus will be less obvious because of the more complex hydrological conditions of its habitat. Only the most northern parts of its range show resemblance to conditions off Patagonia. The main locus of Poronotus is influenced by the warm gulf-stream flowing towards higher latitudes, which is the converse of the Patagonian conditions. Moreover, the coastline in the habitat of Poronotus runs more east and west than the coasts frequented by Stromateus, and therefore any meridional component of movement will be acting more nearly in the same direction as that of the primary on- and offshore movements of Poronotus. Stromateus maculatus has an extensive distribution up the west coast of South America as well as off eastern Patagonia, and Norman (1937, p. 1 19) has pointed out that large series of specimens from the more distant localities might reveal the existence of two or more races or subspecies. He also stated that there is some doubt as to the specific identity of the specimens reported as far north as Peru by Valenciennes. However, the narrower Peru coastal current brings relatively temperate con- ditions very much farther north up the west coast, than does the Falkland current off the south-east coast of South America; and since these Peruvian specimens were recorded as most abundant in the Lima market in winter, they might have resulted from seasonal meridional movement of the most northerly of the west coast stocks, in view of the known behaviour of the species elsewhere. It is well known that the extreme northward extent of the Peru coastal current leads to a more northerly dis- tribution of other temperate types, including such fish as hake {Merlucchis gayi), than is to be found anywhere else south of the equator. In order fully to appreciate the potential value of the spotted pomfret in our area, it is instructive to consider the Stromateidae already exploited elsewhere. The family has been much subject to taxonomic changes, and further systematic revision — impossible without further widespread col- lecting— is still needed. It is hoped that Table 37 gives sufficient synonymy to leave no doubt of the identity of the species referred to. There remains the bewildering tangle of common names, many raising most interesting etymological problems, which I give with their localities in the second column. To quote individual authorities for all these would take up too much space. The object of the table is to give some idea of the range and relative importance of the Stromateidae already exploited, in compact form, before proceeding to some further consideration of their common names. The etymology of the name ' pampano ' or ' pompano ' is extremely interesting. Its literal meaning in Spanish is 'a young vine branch or tendril', but the ichthyological connotation is very old, and may possibly have first applied to Sparus salpa. Of the ultimate origin of the word from the root 'pomum', a fruit, there can be little doubt. The Dutch word ' pampelmoes ', applied to Stromateus fiatola in South Africa, derives from the French ' pampelmousse ', the grape-fruit or shaddock (compare 'pomelo' and 'pompoleon', names occasionally applied to this fruit formerly in England). Perhaps the likeness of a deeply compressed fish seen in profile to a grape-fruit viewed in the same way accounts ^ Such very few Stromateus juv. as we captured were found only in the north. DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 371 ' ' I I J'-l ' ' ■ ■ ■ ' ' ' '~^~' Spring" • — Summer ■ Autumn ...—..-- Winter 1 1 1 1 1 1 ' I I h-^ I I I I 1 I I I I 1 I I I I I i=c=\ 45 50" 55 Fig. 51. Schematic representation of the probable main trends of movement of Stromateus maculatus. 372 DISCOVERY REPORTS r o .8 '9 o o e o <3 S S t; o a, e o c o o W j2 ^o't; O T) 9 c T)- Q. 9" o S o ^ c ^ a^rS U-; 60 ■as °0 « I ^ XI o d o -r;^.3 *~» *-> C r3 ■*-' (X o ■S-a « rt CD 3 ^ M -^■2 ^ ^ f2 S U-. r2 T3 ~ -.o O S " . h °~- M O ^ >, fi , , C u > rt ° 3T3 •£ ^ g m 2 E t*- M C il ;5— O « « C -a d « ex -5 ' M .™_>-. to « .S -S ■" o D '-' 3 ^ 3 — ^ C — u ^ 4J CO I -.£; 6 O 73 uS « C U I) ■ — •s-s &■ >r^ C/? lU rt ' o c P M rt 5 -P -C 60 O 60 2 .2 CI X -3 Mug pa 1-3 -a c; a a u M o U J, '^ '^ 1i. < ^ s ^ -^S ^ t3 &■ . « CQ ' — > 5 c/3 ^"^ • ^ to « -. Z O rt rt ^ h C '^ 1) u o « J3 ^ 2 <^ *^.S r^ O P « 6 C ° !3 60 c 3 C Q c8 ^ > "^ ^ c 9 rt cc;^ o u. o Id a a a C IS U — Ph S2 ^^ Crt 6 O fct, o PJ o W M G C o 6 e o U o V Oh is -H .. o a, O o U d 1— . 1) c « s c o « V 2 TJ C c =3 O a CO Vi e E OJ -r) 6 n a W , •^ , -^ "a « ^ •^ M It^ iS^t' ■« -« Jt c>3CoaH « 6 >--s o O rt ^ en n g o ■5 c ex o -a P e -P ex g >. h o 60 8 X W a.^ „ 3 « -S ^-o u So .- -IN < :s-£ o •S 2 p -* ,p 3 <(< « . ex _p — -in manize y have Philo ( ,^ -« S ■« o « pane erfis fish pelrr iLa e (r is m that i: C ^ S ^^ j= _ « a 0,05 O ^ T3 rp i> ■ — •-^ P rxi 60 P o be ' P o X « J! S:^-^t3 «^ «J N « O r^ =^ P . U -a; -a; ■ct, o s ^ J3 «3 60 P a j= P ■Ci, S Co a s CO 2-^ 1^ 1-2 < U W - I t <^ CT] }-( a lu o S >. ii b 3 '5-« «a -a G C/3 J3 <3 3 )^ P ~ p "« p "J p P VI 3 u P P « ft; < XI -^ Sl CO a ^ S c -s: G G ft5 J2 ^ i7 WS583 I in BTS WS97 I WS847A I WS832 I in NR WS244 2 WS851 I WS877 I juv. in NR WS825 I AGONIDAE Agonopsis chiloensis (Jenyns). To the south Agonopsis was found chiefly on the shallower coastal grounds. Farther north it penetrated to the centre of the shelf also, and there some 30% of our specimens were secured. Its distribution is of great interest, for it seems to show a strong tendency to remain in localized areas, roughly intermediate between the contrasted types of distribution shown by the two species last mentioned, and in conformity with what one might infer of its habits. The fish is small, seldom exceeding a length of 15 cm., and the body is, in Boulenger's words, 'com- pletely cuirassed with osseous scutes', giving it a lizard-like appearance. The body is marked by a number of brown cross-bars on a sandy buff background. Beyond a very evident lack of agility there is little direct evidence as to its habits. The colour of Agonopsis was often matched by hauls of buff-coloured invertebrates such as the hydrocaulus of tubularians and the reef-like polyzoan Smittina among which it was found ; or the sand, shell and coral fragments of the bottom deposit. Agonopsis was often caught in the Russell bottom-net. Unless these were individuals behaving in an exceptional manner, it shows that the fish rise when disturbed, or alternatively are in the habit of swimming several inches above the sea-floor. The small flatfish Thysanopsetta lives right on the bottom and was very commonly caught in the fine nets attached to the back of the trawl, but was taken in the Russell net only once. As can be seen from Fig. 52, Agonopsis was caught in five separate localities. The conclusion that this indicates a stationary habit is borne out by the fact that within each group catches were secured in different months. In the most northern of these localities the fish were met with in October, December and March; in the most eastern (Falklands area) in February, May and September; while in two others the fish were met with in different years : WS7I 20 WS243 I WS583 5 (in BTS) WS81 2 WS787 4 WS754 3 (in NR) WS83 4 WS791B I WS767 I (in NR) WSgj I WS799A I WS832 3 (in NiooB) WS95 s WS800B I 6 (in NR) WS216 I WS847A I WS836 2 (in BTS) WS219 I WS860 I WS852 I (in BTS) WS221 I WS873 Post-larvae present (in NR) DISTRIBUTION AND GENERAL NOTES ON THE SPECIES 377 Fie ,2 Distribution of Agonopsis chiloensis: positive records only, roughly contoured to show their localization in differerit seLns'. oLonds, springfcircles, summer ; triangles, autumn ; squares, winter. Solid symbols represent captures m Trawl + accessory nets'; cross symbols with 'Other gear'. LIPARIDAE Careproctus falklandica (Lonnberg). Six specimens of this fish were trawled at St. WS89 off Cape Virgins in April 1927. It was previously known only from the Falkland Islands and the Burdwood Bank. This suggests a distribution similar to that of Neophrynkhthys, which it resembles m havmg a soft and gelatinous body. Its colour in life was pale orange. Unidentified liparids were obtained at two stations in the Magellan Strait, outside the area of the trawling surveys. BOTHIDAE Thysanopsetta naresi Gunther. This small flatfish, resembling the 'scald-fish', Arnoglossus laterna (Walbaum), of British seas, was the only member of the family at all numerous m catches obtamed during the trawling surveys. Its small size renders it unsuitable for human consumption, the largest 378 DISCOVERY REPORTS ... >i C O 3 05 O O J3 rt - >, 6* •^ u< ■? -^ 'A O 0^ - c U= J3 11 .t; •o ^ -« M C U rt u. M 3 C D. ^ n O J3 -^ »-»-. • - O tfl 1) M > OJ -jS b ca t« bO V "3 c 1-1 c-S S-^ « . ' -C M OJ o c c fc^ o o ca i2 + "« E-1 - )-« >-. u C 3 o ^ 6 6 3 05 ^-R C/3 n .3 aj •C-fl o ^ ^ o CO — a. rt 3 03 ^ '^^-a 05 . c C en o o e M =" rt ta "• Goode, G. Brown and Associates, 1884. The Fisheries and Fishing Industries of the United States. Section I: Natural History of useful Aquatic Animals. Text. [No. 72. The Silver Hake and the Merluccio, pp. 240-3.] Atlas of 272 plates published as a separate volume in the same year. Washington. Gran, H. H., 1929. Investigations of the production of plankton outside the Romsdalsfjord, 1926-7. Rapports et Proces-Verbaux des Reunions, Cons. Perm. Internat. Explor. Mer, lvi, 112 pp. GiJNTHER, Albert, 1887. Report on the deep-sea fishes collected by H.M.S. 'Challenger' during the years 1873-6. Rep. Sci. Res. Challenger, Zool, Vol. xxii. GURNEY, R., 1933. British Freshwater Copepoda, ill. Ray Soc. Monograph 120. London. Hamilton, J. E., 1934. The southern sea lion, Otaria byronia [De Blainville). Discovery Reports, viii, pp. 271-318, pis. i-xiii. Hardy, A. C, 1924. The herring in relation to its animate environment. Part I. The food and feeding of the herring. Min. Agric. Fish., Fish. Invest. Sen, 11, Vol. vii. No. 3. Hardy, A. C. See Kemp, Hardy and Mackintosh (1929). Hart, T. John, 1934. On the phytoplankion of the south-west Atlantic and the Bellingshausen Sea, 1929-31. Discovery Reports, VIII, pp. 1-268. REFERENCES 393 Hart T Tohn 1042. Phytoplankton periodicity in Antarctic surface waters. Discovery Reports, xxi, pp. 261-356. Hartley, PH.' T 1940. t£ saltash tuck-net fishery and the ecology of some estuarine fishes. Journ. Mar. Biol. Assoc, xx.v, HartTey!'r H.' Tt^!945'^rSfooi of Coarse Fish, being the Interim Rep. on the coarse fish investigation. Freshwater Biological Association, Sci. Publ. No. 3, pp. 5-33. 5 Ags- xvii tables. Harvey H W., 1934. Measurement of phytoplankton population. Journ. Mar. Biol. Assoc, N.S., xix, PP- 7f>i-73- Hentsc^el Ernst 1936. Allgemane B^olog^e des Sildatlantnchen O.eans. Wiss. Ergeb. Deutsch. Atlantischen Exped. 'Meteor', 1925-7, Band xi. [Beilage ix-xlii published under separate cover, same year.] p , ,• u ^ K r A P Herklots, G. a C and Lin, S. Y., 1938. Common Marine Food Fishes of Hong-Kong. 75 PP- 4i fig^- Published by G. A. C. Herklots at the University, Hong-Kong. ■ „io ^ fi„c HiCKLiNG C F., 1925. A new type of luminescence in fishes. Journ. Mar. Biol. Assoc, N.S., xiii pp. 914-3?. iv pis., 7 hgs. HiCKUNG, C. F., 1927. The Naturai History of the Hake. Parts I and H. Min. Agric. Fish.. Fish. Invest. Ser., 11, Vol. x, HiCKUNG,' C. F., 1928. The Fleetwood exploratory voyages for hake. Journ. du Conseil Internat. Explor. Mer, iii. No. i, HiCKUNrc%., i930«. A contribution to the life-history of the spur-dog. Journ. Mar. Biol. Assoc N.S., xvi. pp. 529-76. HICKUNG, C. F., 19306. The Natural History of the Hake. Part III. Min. Agric. Fish., Fish. Invest. Ser.. 11, Vol. xii, Hicku°nJ,'c. F., 1931. A new type of luminescence in fishes. III. Journ. Mar. Biol. Assoc, N.S., xvii, pp. 853-67, iv pis., HickLn'g'c F..1933. The Natural Hrstory of the Hake. Part IV. Min. Agric Fish., Fish. Invest. Ser.. 11, Vol xni. No. 2^ HICKUNG, C. F., 1935 «. The Hake and the Hake F.hery, being the Buckland lectures for 1934. London: Edwin Arnold HICKUNG%. F., 19356. Seasonal changes in the ovary of the ^mmature hake, Merluccius merluccius L. Journ. Mar. Biol. Assoc, N.S., XX, pp. 443-62. Hildebrand, Samuel F. See Longley and Hildebrand (1941). HoDGKiNSON, E. R. See PhiUipps and Hodgkinson (1922). Hodgson W. C. 5ce Wollaston and Hodgson (1929). „ m r ., JIrvis. SoRMAN D., 193- The Fisheries of Puerto Ruo. U.S. Dept. Comm., Bur. Fish., Invest. Rep. No. 13. pp. t-41, loHN d'dilwyn, 1936. The southern species of the genus Euphausia. Discovery Reports, xiv. pp 193-324- JOHNSTON, R. M., 1891. The Fishes and FtsMng Industries of Tasmania, together wUh a revised hst of tndigenous speaes. Paps.'Proc Roy. Soc Tasmania for 1890, pp. 22-46. , u-u ,^ n . ki ,.^ Po Jordan David Starr. 1905. A gmde to the study of fishes. 2 Vols. Westminster: Archibald Constable and Co. ORDAN DAVID Starr and EiGENMANN. Carl H.. 1890. A review of the Genera and Species of Serramdae found rn the waters oMmmV« ««<^ £«ro/... Bull. U.S. Fish. Comm. VIII (fori888), No. 9, pp. 329-441. pls- Ix-lxix. Kemp, S. W.. Hardy, AC. and Mackintosh, N. A.. 1929. Disc^ery investigations: objects, equipment and methods. Dis- K.Zl.^ZTm''le ZZr^mungen .wischen Kap Horn und der La Plata-MUndung. Ann. Hydrogr. Mar. Meteorol.. K.Z:.ZoSo%t H^^^^^^ Band II. Die Bewegungsformen des Meeres. Pp. i-xvi, 1-766, 3 tables. Kyle'hIrry M.; f926!'?A. Biology of Fishes, xvi + 396 PP- 77 %-. xvii pis. London: Sidgwick and Jackson. L^ngL' W^Tan^d^'ni^^^^^^ Samuel F.. 194^. Systematic catalogue of the Fishes ofTortugas, Florida. Being Vol XXXIV oJ Papers from the Tortugas Laboratory. Carnegie Inst. Washington Publ. 535. -» + 33- PP-, 34 pls- Wash- L0Ri;!'?u;rf923. A List of the Fishes of Tasmania (Read Sept., 1922). Paps. Proc Roy. Soc. Tasmania for 1922, MCCULLO^! Alan R., 1915- The migration of the jolly-tail or eel-gudgeon, Galaxias attenuatus, from the sea to fresh water. Australian Zoologist. I (2), pp. 47-9- Mapkintosh N a See Kemp, Hardy and Mackintosh (1929). , t^. d ..» „ Matthews LhIkison. 1932. Lobster-Krill: Anomuran Crustacea that are the Food of Whales. Discovery Reports, v. MATTHEfs!L. HARRiloN, 1934- The Marine Deposits of the Patagonian Continental Shelf Discovery Reports, ix. pp. 175-206. Meek,'1™r, 1916. The Migrations of Fish. Pp. vii-xviii. 1-427; -8 figs., xi pis. London: Edwin Arnold. ^^^ 394 DISCOVERY REPORTS Menzies, W. J. M., 1925. The Salmon, its Life Story, vi + 211 pp., 5 diags., 36 figs., photographs. Edinburgh and London: William Blackwood. MuNRO, I. S. R., 1943. Revision of Australian species of Scomberomorus. Mem. Queensland Mus., xii, 2, pp. 65-95, pis. vi-viii. Navarro et ah, 1943. La Pesca de Arrastre en las fondos del Cabo Blanco y del Banco Arguin {Africa Sahariana). [F. de P. Navarro, de F. Lozano, J. M. Navaz, E. Otero, J. Sainz Pardo and others.] Trabajos No. 18, pp. 1-225, Lam. i-xxxviii. Madrid: Minis. Mar. Inst. Espagnol de Oceangr. Norman, J. R., 1937. Coast fishes. Part II. The Patagonian region. Discovery Reports, xvi, pp. i- 150, pis. i-v. Norman, J. R., 1938. Coast fishes. Part III. The Antarctica Zone. Discovery Reports, xviii, pp. 1-105, pi. i. Pearson, John C, 1932. Winter Trawl Fishery off the Virginia and North Carolina Coasts. U.S. Dept. Comm., Bur. Fish., Invest. Rep., No. 10, 31 pp. Phillipps, W. J., 1919. Life history of the fish Galaxias attenuatus. Australian Zoologist, I (7), pp. 211-13. Phillipps, W. J., 192 1. Notes on the edible fishes of New Zealand. N.Z. Journ. Sci. Tech., iv, pp. 114-25. Phillipps, W. J., 1924. Notes on the life history of the New Zealand minnow {the adult whitebait) Galaxias attenuatus. N.Z. Journ. Sci. Tech., vii, pp. 1 17-19. Phillipps, W. J. and Hodgkinson, E. R., 1922. Further notes on the edible fishes of New Zealand. N.Z. Journ. Sci. Tech., v, pp. 91-7. Pozzi, A. J. and Bordale, L. F., 1935. Cuadro sistemdtico de los peces fnarinos de la Repiiblica Argentina. Anal. See. cient. Argent., cxx (4), pp. 145-89, i map. Radcliffe, Lewis. See Evermann and Radcliffe (1917). Rayner, G. W., 1935. The Falkland species of the crustacean genus Munida. Discovery Reports, x, pp. 209-45. Read, Bernard E., 1939. Common Food Fishes of Shanghai. 52 pp., 32 figs. North China Branch, Royal Asiatic Society. Shanghai: Mercury Press. Regan, C. T., 1914. British Antarctic {' Terra Nova') Expedition, 1910. Fishes. Terra Nova Reports, Zoology, I (i), pp. 1-54, 13 pis., etc. RowE, Anne. See Simpson and Rowe (1939). Russell, E. S., 1937. Fish migrations. Biol. Rev. Vol. xii. No. 3, pp. 320-37, 4 figs. Cambridge. Schmidt, J., 1930. The Atlantic cod (Gadus callarias L.) and local races of the same. Compt. Rend. Lab. Carlsberg, xviii, p. 6. Copenhagen. Scott, E. O. G., 1936. Observations on some Tasmanian fishes. Part III. (Read Sept. 1935.) Paps. Proc. Roy. Soc. Tasmania for 1935, pp. 113-29. Scott, E. O. G., 1938. Observations on fishes of the family Galaxiidae. Part II. (Read Nov. 1937.) Paps. Proc. Roy. Soc. Tasmania for 1937, pp. 1 11-43, P's- xvii-xxvii. Scott, E. O. G., 1941. Observations on fishes of the family Galaxiidae. Part III. (Read Nov. 1940.) Paps. Proc. Roy. Soc. Tasmania for 1940, pp. 55-69, pi. ix. Simpson, George Gaylord and Rowe, Anne, 1939. Quantitative Zoology, xv + 414 pp., 52 figs. New York and London- Magraw Hill Book Co. Inc. Steven, G. A., 1932. Rays and skates of Devon and Cornwall. II. A study of the fishery ; with notes on the occurrence, migra- tions and habits of the species. Journ. Mar. Biol. Assoc, N.S., xviii. No. i, pp. 1-33. Steven, G. A., 1933. Rays and skates of Devon and Cornwall. III. The proportions of the sexes in nature and in commercial landings, and their significance to the fishery. Journ. Mar. Biol. Assoc, N.S., xviii. No. 2, pp. 611-25. Steven, G. A., 1936. Migrations and growth of the thornback ray (Raia clavata L.). Journ. Mar. Biol. Assoc, N.S., xx No 3 pp. 605-14. Taning, a. Vedel, 1937. Some features in the migration of cod. Journ. du Conseil Internat. Explor. Mer xii No i ic no 8 figs. ' ' • . Ji FF-. Thompson, Sir D'Arcy Wentworth, 1942. Growth and Form. Revised ed. Camb. Univ. Press. Thompson, Harold, 1943. A Biological and Economic study of Cod (Gadus callarias L.) in the Newfoundland area including Labrador. Newfld. Govt. Dept. Nat. Resources. Res. Bull. No. 14, 160 pp., viii charts, 12 figs., Appendices a b c and A-Z. > > > Vaillant, Leon, 1888. Poissons. Mission Scientifique du Cap Horn, 1882-83, T. vi, Zoologie, pp. Q-C35, pis. i-iV Paris Vaillant, LfoN, 1906. Poissons. Exped. Antarct. Fran, commandee par le Dr Jean Charcot, 52 pp., 4 figs Paris- Masson Waite, E. R., 1899. Fishes. Sci. Res. Trawling Exped. H.M.C.S. 'Thetis'. Part I. With addendum. Mem. iv Vol \ PP- 27-132. Sydney: Australia Museum. Waite, E. R., 191 i. Pisces, Part II. Sci. Res. N.Z. Govt. Trawling Exped., 1907. Rec Canterbury Mus. i, No 3 pp 1,7- 272, pis. xxiv-lvii. J' ff 3/ Waite, E. R., 192 1. Catalogue of the Fishes of South Australia. Recs. South Australian Mus., 11, i, pp 1-208 ^^2 fies Welsh, William W. See Bigelow and Welsh (1925). Wollaston, H. J. Buchanan and Hodgson, W. C, 1929. A new method of treating frequency curves in fishery statistics with some results. Journ. Cons. Internat. Explor. Mer, iv. No. 2, pp. 207-25. APPENDICES 396 DISCOVERY REPORTS o < o I— I o o Pi < Ph I— ( Q iz; w PL, Ph o a, o c o c O H O 2 2; i>l u ^ a. o c S a, J3 Cao bjo c , 3 J3 1 1 o -4-* O XI y3 c 3 » o -o CJ H (D 3 O 2 fc u s o 4-* O pq 3 V , 12 li o (- "^ - ^'S,;;'^ —O"^ r^r^N N O O O N I^OO t^^o CO ^6 t^ob CO vo v6 CO ^^ob Kob r^ t^cb 6 >o io iovo r^ t^so CO ioob f^ NCO r<^HH r*-)ONroO '^foa^O^ CO '^ C^ "^ 0^ O^ O^ OnO i-h n N ro wico li^ u-,co J^ ro C^ tnco N M M iy-> u-i Lo io to io^b Tf- f^ f^ob ob f^ N lo ro\0 Tj- t^ ro N O -^ ro rOCO ON J-^ M C>vD ^ oj r^ O NO CO OnOO O On On On O On OnOO nO CO CO OD f^NO CO ONf^N OnOOO O " O Onu^On N O rooo i-H y^^^-^ONO N N On ONCO CO 00 CO ON 6 00 ONCO On IT) lo Ln -ho NO NO O -i-NO ui W o N NO NO NO nd so NO NO f^NO t^ r^ob cb r^ CO o o o O OO 1-1 M r<-)NO N CO ro O O l-^ « On M i-< I^OO O O O O O -i- O O OONf^Ot^NNOOOOOOOOOOOO OnOOOOio"-iij-iOOOOOOOOOOOO AAA AAAAA AAAA AAAAAAAAAAAA O ON 'i-co NO ^ lO N 00 N ro fo ro N N - - -_ ON^r^*- o u-i N N M r- o l-< 0^ rrj looo t-< 11 i-« -J On t^ 1^00 '^-N '^ON-i-N w H l-l ►-'t-'i^r^'-HMMojoo I I I I I I I I I O t^ioONi-HNOCO roN ^f^f^t^voONt-i MOO lo ro O NO CO O -^ r«^ (N) O NO M W M hH I I I I M LO fO O C7N T^ ro tH l>^ M lO O ■^ t^ (N) N I-, On 11 M N M I I I I I I OnnO NO r^ 1-1 OO o ON -^ r^ to w 1-1 w « N " ON ON OnnO conO t^ -^ '»^ no "^ I^ -^ fO 1-1 1-1 -^ N N N I r^ I I I " " M (Nj OO NO •* ^ O " N N N '^ r^ ^ rh )_, o 1— 1 o HH I—I 1— ( ^ N "-I X K- 1 H 0^ OO ^n on n -i- 1— 1 ^ 1— 1 o r^ o N N ro i-< l-H O O ^roM loLoij-, N o >oio looo O u-.«rpu^ro;+i-. O O loN w « io« -i O u-i" « u-i^M ""iroioi •^ >-o o oo •I "-^ N r^iONvO r^N r^O r^O u-^O Om .-hhhhhwi-.O'-hOO'hOO'-h O O »J-)N ^^-u-iu-iioi ON On r^ 1-1 r^ t^ 1^00 NO ^ - - Tt-moNt^^oco o 1-1 M M n o t^oo o LO N -^ O lO 11 LO r^ OnnO lt-^nO nO no NO t^ i-i«„Oi-iOi-iOiH" tor^ino o o o too r^oo u-i lo w ^ HH LT-jroi-i O roroN i-i O fOio-^ O N O 'I- OnnO COnOOOCOnOOO t^u-j >-iihi-cmO«Oi-iOO'-iOOi-i NTt-lOW roi-i ^ONt^N t^r^ONM -+NO i^nO 03 Aw roroiocoo roO lorO" ^O O" ro'S !o, O ^ P^ O °r^°i^?^ °/^ °». °;^ 2, °_._° ooooooooooo to f^ lOOO CO to rj- Tj-00 nOnOOOOOOCOK-iNOOOOOO tOtOtOOO^^^NO^C^Ng S^^ ^^S nS-N^N^vS N^^nJnS ^ S^^ N8Ngv§N8N§N8v§N8S;NgN2 N^N^V^ !^^ ■'r^ "*. "^ ^^ ■" *■ '~'''^'"- ^ ^ ». «]rl.^|c.lr-*|CJH«^^ (-OOOOOOOOu-iroON c^ O K O § O Ov "tI-oo 00 O CO t-H O "-< loioioo LOO Hlei^ ^ ^ ^ O tJ- N M 00 O u-i '^ m M r^ !>. N N lO r^ r^oo loo ^O m ro r^ LOlOVOU-llOLOLOlOlj-ilOlO '^-N M N ro-rf-ThfON N tototo.o.o^^to^to.0 S>=§-<=^ ???=§- aS>&>??=^q:*J^¥!J^5:^? >N M C n u .>.>.> .> > > > fOrO-^-ONl-i ■n^-'^tocO fONO wiHwwNNNN l^ r^OO 00 ONNO r^ t^oo 00 ON ionO OOCOCOOOQOOO°°°°°0°0°°000000 COCOCOOOOQCOmnnnNNNN > > > > > ro ro -^-oo 00 N N N N N N M > > > > > > >■? "> ON N ON N O O ro ro ro « w N N N m »^ eg 0-; l^NO t\Oo 0\ O >N (Ni i-r^ ^ lo NTN l\ Or\ rKf>*^fN, r^v4..^«^ COO\0>HNrO^ ionO t^OO On O PARTICULARS OF TRAWLING STATIONS 397 otrope ed well o ■5 -s 1 1 1 1 1 1 1 1 1 1 1 E 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ^111 Sill Z o ^ bjo •T3T313 ^T3-a "^-^En 333 ih33 C jsS-^^m Z Z Z Z ZZ Z JfiZ hS 0 s i-irt-wvO-^-i-i">- o rt tfl ? o U -1 N-!hOOOOiH>00000000 m OOOOTt-u-it^vO^OOOOOOOOOOOOOOi-OO 0^\0 un u-1 o Tj- Th u-iO vOOOOOOO 1 N i i OOioO"NN ""OOOOOOOOOOOOO-^-^ -^vO .2i; = A AAAAAA AAA AAAAAAAAAAAAA Q P V u •a o C '— S £ -2 2J •^1- m u-i M ov -hvO vO t^ t^ LOOO t^I^ivOi |vO«T|-i-<0 irivO 0000 ^O inu-iO N O >-^OCO N O mt^ro f^OO 2 o ON t^ u^0O 00 00 LO lO -h ^ " t^^O " N OOir,"^"00"-it^t^«i-.t^^-t-«ror^O ^CO vO -1- O " O r^ .2i NN"r^<^r^roroNNNP-i«Ol'N' •-•N"NNr<-)NN m«NmNNi-i Nr^rOr^, Q 6 SO M N loO Mt^OvQiot^u-tOO (-^"-"r^ONt-ir^OOONf^Tl- O ^00 N 00 ^ O O sD OS c N O »o O^ rn t^vO u-iONI^Oir^O»ri on^^'^*-''^^<^0'^'-h 00»^i^Ni-tO'-'OsOO>'^'-tOsa^ rh-^ui HHh-tl-lMNN'-'l-HI-HWN>-'rON 1 MM^t-HWMNNNrOtH OSVO \D00MWI-lONt-lN'-''-tON ON*^' hH »-< 1-1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 ) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ! 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -^ N r(- O CTs 0 0> r^OO toioinN O ' sO -hroON^ro-^-r^O "-i 00 -^t^r^ t^sO O CO O ON f' ON lo 1^00 CO '-' M U-) C 3 NnO -ThN N n-isOnO OnoO ON^Ot^i'^ On'^'^-i-' -^loOvO ^ lo *-< OnDnOOO O "-• '-' OnO ONro"-" O O^ ^ -^ -^ ^^ MMi-nroNNi-'MiHi-.i-iMNi-. «««iH«Nr^MNNMi-c mm,-, c^nwmm wmmm c I^nO " OnoO O O " m 't- O uiOO Lr)N"'1-^'-'N0r^0"->0r^0"-i0000'J-)00tnro00io00u-,0 0 o o roiON ION O iOfO»or^io»oO iom»-< "^N ■Tj-'-i HH >H fi lo-^- m-^r^'-' ror<^tH wir'-j'^-M roN N CON N N *-* •^ H o N CNOO On N Tj- ONOO N CO CO r^ " o 00 onno 00 r^ onoo ■-■ oc o " loco o o CO o onoo no loco r- r^ ON onco ■-■ on CJNOOO'-i'-HOO»-t'-HO»-'»Hi-Hi-<0»-iO'-'0'-''-tON hH>-iO'-'NO'-H'-i'-<'-i'-<'-i>-«0'-i'-iO'-'0 e 2 OOO Tt-00 O O NO lO tJ- O >000 >ON00NN0N0r0toio0^00i000io0ioio00ro00io00>O0>o E H fe tHNOOOt-iMOO'-'»-<0'-iMO'-<0'-*0»-tO'-''-'0'-i t--'w»-iwi-(MO»-'i-00 rorOO ^ O rOO N u-ii-i i-i O c ooooooooooooooooooooooooo ooooooooooooooooooo o O^ On N NO NO r^OO OOOnOnQ'-'OOi-'Nio tovO ''t'NrO'-toOt^ ONONtotoro'-'NONOONrO ionO CO 00 GO O c: J lO ionC lo lo lo lo to »o tONO vOnOOOnOnOnOnOnOnOnOnO lOiO to tONO vOnO^nOnOnOnOnOnOnDOnO toio u^nO C/2 OOOOOOOioOOiOiONNOiOOOOio lONO NO OnO >oO ionOnO "00 "I-nO O no O oioOrO"ioioOOOroN-*-^ioioio^"tOTt-0-*-i-iNO'i-"'ON-j>.,m-m""" u NNNNN rorotororoeotorocororo'^'^'^'^'^. . . Q ■>'>'>'>'>'>'>'>'>'>'>">>'>'>'>>>>>>>>>> X X X y. '^ X y, y, X y, y, y. X X X ■■< •f< >; y. onono n d " coTi-toio toNO NO t^oo iono no r^ r^oo oocnoo ooNONONor^r^i^ONOt-tMNro cond -^ ^- to ""„„Ji5 ^ „„„„„i_„„NN«M""<""«MNroro §^ '^o; i<-)nO t^oo On O -^ N ro ^- tONO r^OO gN O 3 J> £2 3: IS'S.'S. ° "R"^ N N r'J-r-^iOM N p;* tONO t^ -H ^ ^ r:^ Pi 0] rl fN) (N) !Ni fn^oro^o^orooc«coc'}'>3'>3^cncfic«Mcr!c«a)fococotocojoto^coai cococ«c«iy3c«aic«c«c«cncADcflMaijojococo C/3 ^ 2^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ 398 DISCOVERY REPORTS o Q w PL, < o Trawl fouled u o 3 2 e e o o %^\ irKS^Kf^l^lll^^^Kl JCI IS^I i 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 1 1 1 1) o 3 C/2 moo 00 moo 1- M 0 LoOOO" t^ lot-^o n or-, ir'^TfT-ri'^irir^r'T^ir^l |9?^|9^ir^| 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ; U -1 o _ .22 i = S ^ "- '- 2 c " c S^-S o ca M ^ o V u-iooooooooooooooooooooooooooo 'i-o o^t>oogoooogggg t-OOOOOOOOOOOOOOOOOOOOOOOOOOO"" roO vOO^OOOOOOOOOO AAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAA 2 2 .22 i Q III oo w mo\o or— r^io»'^»-' lo«oa^ Looo '-' o o oo a^f— c^»^o^n t^oo n o^r^c>i^»^-+-r-N i^io t— o^o^r» O u-> ^ -1- ON r^ N •i->0O 00 M N ^ -)- r-CO "OOOOr-oiONNMi-cro -t-vO -I- -h On r-vO -<-NOt^in-t-Nror^ " NWIHl-l „„„„„„« „„M„„p)N«««w«« Q .^r.^r^j-^ONO HH^ MsDoJ^sO^-ifON '-'■^ t— O^t^ONtoN O^O roOOO NvO f— -hO „„„««m00ON"'-'On""""'-''-'"'-' I mi-.-0«"mi-cm'-'«i-'"'-'""'-i""'-'<:7>ONOn "-)C0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 O O O O r^.vO N i/-> ON ON t-^C N nO oo r^i t-vO CO ' u-i«ONi-00 ""^^i-^r^ (^nO N r^t^"-)T^-T^^^ N •-■ O O O^ONO^i^Ti- o u-iOOOOmirimOOOOOOOOMOO'^.N"-lNr-«/^OOK-i:^Ou-ir-u->mioO''^"->mLr>i-iOOin CO" u-ir^mM u-iTt-m'^O K-)iriLou-i«0" O t^ 1 " •^rj-O ^m u-iu-i-|-rO'*--1--i-^" -^mu-ir^ino O N " GO i-( N '1-nO r— 00 ro r— M t— r^ fOOO roONN r^t— 1 00 mN rot— M NnO -^nO r^NO OO I^I^ror-inO NGO O •<*-r— „w„«O""NO«"M0000m>-'0 «Ni-imwNOOm"NOO"0>-0000»J^"^r— *^*^00 0000'-'OOOrlONt^u-^0 0»'^i'^0»ot— iou-)»00 00»00»-'OOLn fO^-^ li-iroN ►-! i^N M roO t/^i^iou^iO"-* O u^O « O -1-0 -+-•-< uiiy-, roro'^'^'^'^'-' ThroiorDO O O N « I— O " rM "^nO t-^NvOOOvO OnN '^NOO m n rOOO t— ©"MrO-Nuirou-iNu-, I— vO nO N nD tJ-\D N I^ O r^ <••-) „w«i-iOi-'>-'NOO«i--< r^<:^'H O I>mgo mt^'+O O roro^»ON ON Thrl-u-iThN m -tJ-Q N-^rJ-O N O O ^ON w LomN NO '^'loN m O-^N-^OO O^OtO oooooooooooooooooooooooooooooooooooooooooooo •-H NNNrniOU-^T^-r^r^NN NNhh h-i O roroiO'+»OLONNrl hh •-( i-(i-h NNN ro*^»^ iri\0 O f^OO OO OO OO 0'0'^^0'00'OsOO'O^OOsDOOO'0'00'0'OOO^OvO'sD\DO'0'0'00'OOOsD\00^0^0'>0'0'0 h-1 ■^NNNO^O'-^OIN^n'^T^u-llJ-^l-HwlOli-lr^T^ror*^00•-H^-^Nr^■^-^^NN'-lOlO'^NNl-HNN'-«l-H oooooooooooooooooooooooooooooooooooooooooooo ON c^ c^ o^ ONOo u-iu-^voii-iiou-iiovn u-i\o ^o r^ t^ i> i^ r^ r^oo ooooooooo o o o o o oo^c^o^c^c^'-' m »-« ^ Q ^ m r-, ro r-. r-. r^ C-. r-. CO r^ r-. c-. r^ m ro CO r-. CO ro r^. r-. r-. r^ ro ro r--._^ ro ^ ^ ^ ^ ^ ^ „ ^ ^ ^ „ ^^ ^ ^ ^ ^ .^ .« .^ .^ .-; .^ .-; .-: .-; .— .-^ .— ..-' •-; --4 .-h ._; .-: .^' ...;._; ..4 .^ .« -^ .-; 'p ._; m ro m ro ro ro ro rn r^. f^- ro ro ro ro fo ro '>< '>< 'S '>< 'S '?< '>< 'i< '>< 'x '?. 'x "P, 'S 'P 'x 'x 'x 'x 'x 'x 'x 'x 'x 'x 'x '. 'x .-; .« .« .« .- .-: .- .- •-; •-: .« ■- .- •-: .- •-; N lONO nOvC I— t— f^c^Tj-Tl-.^.rJ-iy^ u-iO r— CC ONtHONOOOt-'HHMNMLrj*^ u-inO nO nO t— t— 00 00 00 On O N O N N o c 1] 1- Ti- u-i u-i u-iND r-00 a^oOMMNNf<^'t• >^vo NO r- r- t^oo ONONOOi-co<5 r-oo on on o -i •-< c^i (m oocooooooocoooeo o^o^ONONONONa^ONa^ONa^ONONONONONONO o o o o o o o o o o o o o « "-< "-i •-( >-( t-i^i^t— t— r-t^t^i^t^t^c^t^r>.t^t-~t^t^r-~t— i^r-r-:^ t— oo oooooocooooooooocooooooooooooooooooo PARTICULARS OF TRAWLING STATIONS 399 J3 perly «c o 1-1 a. •a s , . . . "> s -o ^ j: o 1 1 1 1 u- 1 1 1 1 1 •? M 1 1 1 1 1 1 SI -5) 1 1 1 1 1 1 i 1 ll 1 1 1 1 1 1 1 1 jz; u '-• B — T3 C D. ^ n o JS ~™ o c >, a g- '^ ^ S 2 2 o P3 Z Hta ZZ u o 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 *-* £2 u a. 6 o 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 o v2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 M 1 1 1 1 1 1 1 1 ! 1 1 1 1 1 1 1 1 1 1 M 1 1 H 3 CO tf) li- o o o o o o N.-.OCOOO^COOwOOOOOOOOOOOO r^*ooooooooooooooo o o o o o o ON ON r^^ ^ ^ ri'i-ooooooooooooo NOOOOOOOOOOOOOOO SJS-g o )-||-«l-HI-ll-H>-> l-ll-ll-IHH .2i = A A A A A A AAAAAA AAAA AAAAAAAAAAAAAAA Q t V (U -« ii != ';:r ^sii C> M o OO O N O N ■**-0 sO -I H- lo N u^O O NO NGO rou^'-' »>^On O " (^ "-> t^ O N u-i U-, N ThOO »J^ w 'fi- tJ- r^O O m lO N \0 '-' fO .2i 'S'^e I-. I-. *-. HI WHHW.-.HH^Nrl'-^, H-« _►-, M (VJMW MWWW WW Q 6 N 00 N O N CO OO -^ ON N r^OO t^ ^ N ON -1- r^i O r- t^O t^ O -t- <^- O O -t- u-iO rl- N O w M -^ O^O'-'NN'-'OO'-'N '+0 t~^ ^ m rn O " O "-1 o m t^OO N -^O sO iJ-^ro-^i-*©© O "1 r^O N ^C CO '-' r^O 3 M 1- w « N N (N| N for^mTj-ON'-H i-i rr.ro'-' •-< '^'i-'-i >j-i»h m ^ i- hhONwwwwwwwwwwN'-'i-'N 1 1 1 1 1 1 \0 " N O « N 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 i 1 I 1 1 NOONON-^OO tnOO t^OO GO coco ON « ^C' u^ I> t^ 1 T 1 1 1 1 1 1 1 1 1 i 1 1 1 1 '-' OiONNO O '-' NOO r^O ON ON u-1 N C O M CO "-) 0^ O t^ r^ HH o» i^\0 00 '^rOfON ON'^J-OnO O u-jiO'-i ro^^ NONwNmOwQwM rovO -O ro ro *o ! w l-l M li M N NNf^>fOrOTl- WW ►-.Tj-row www M _wwwwwwwwwNMwN| c O O »^ O Ln i^i O O LT-i^OO O O tou->o u~)ir^t^r^LriO toO "J^ lOOO LOO O »OioO NQiO lOOO O O Lo O ui O K-i Tl- o o -t- 'i-'^Tl-r*-lU-)T^^^, O r^iO N N O O --< -h- N M ^-r^ w N 'i-roroiororoio»oto"rt-0 O >-• *o t^ CO '^- Th -fOO O O -1-N rO'^i^*l^0\0 ^ H lo Ti- i-( rn Ov r^ O w ON -^w r^w w ONt^r^r^.r^w Q ON'^r'^wOO 2 ? 3 g '^ o o -+ '^'i f- r*^ ^ r^ mo rr.O N N O O •-' rn >-> N N '+r^, M N H-roroio*rororoio»oH-0 O '-' lO u ■^ m O N CO ov u-i r^O COfO'-'^O O OOOO^O N NO ONOO O N O r^\0 N r^i f^> r^i r>. ON ON ro "-I N c*-i •^\0 On u-j H ta O « N " " - OOwwQwONOwOwwrJwwOwwNO H-wwwwOOwwwwwwwOW| t ^ ^HM- ^ ^ ^ ^ CO CO vO 00 t^ C^ r^N HH r^r^r^N oncoo r^o n w w coon ^i-^o o >^ NtoOtOOOwOlOtO lOO Lo LO t^ G tuO ^ r^ -^ lo ^ O N w -rhNir^toONi-H O O r^O roinNu-ji^N O rr> Qwwwr^lOOrOi-iwwwwwwlo o o o o o o ooooooooooooooooooooo oooooooooooooooo t^vO lO ^ Tt- -)- f^cDNN'-.OOOO r^OO OO O u^ r^, w Q f^ t^sC tn r^ N -^-^-^O ^ N fO'^'^'^'^'^Thu-iro _o hJ s^ \0 ^ *0 vO 'O sOOOsOOsO^D u-j LOO sOOnOsOnOsOOO-OOO ^O O ^O ^O O ^O O O ^O "O ^O ^O "O ^O 'O "O CO O CD H«rH(ci^ i-4c]v Hs^i ■-.ici ---i-t ^ ■-'Ic-)^ Hci^ ^ ^ ^ wlcJ^ Hci.-<[(M^ *. ^ ^ r-i?]^ r^ TJ- 1/-1 O N N O T^-w IONIAN w GOO r^OO w ON '-' w u-^oO r^-O 00 ON00\O <-> N Tt-w rorO'-'\0 u-jOnOnioio •!-• m rr, "^ -rf- u-^ tn >-> rrju-iN vtitJ-w n u-jw w w r^tow rorolOw -+W w u-)N roro^'^'^cOfO hJ o o o o o o ooooooooooooooooooooo oooooooooooooooo 1-c w i-c N N N NNNNNNNNONNNf^, rocO't-OOGO^-' w tJ- lo I> u-1 lo u-> t(-00 on G '-' N CO N N lOTt--^rt-Ti-T^-J-^Tt--^U1lOU->lOlOlO ir^ lo u^ u-j to lo iy^u-^iy^Lnu^u^triioiJ^iO»^LniOLOiOU-iLOLoiO»OLri u N N N N N N r^ ro en ro m ro N nnnnnnnn'^*^'^^*^^*^^^^'^^''^ mNNNNNNNNNNNNNNN Q /4v*-- ••-^ "^ ••H •« '-^ '-^ --4 -^ --^ --^ '-J .-; :s ::: := :3 :s :=! :S :=! :S :3 :S :s > > > N .... ro CO fn -^ "^ *i- t^ t^ t^ l^oo oo o o^Jg- « pj m lo loo \c o o o o " w w M, ro-^iOU-)ior^OO OnQ '-' '-1 coco """""" i-HNNNNNNNNNNCOro c c o o •^oq •^aq-^oq ^ oj < W 7 M — r-i -^h 'o^ f^ t^ OoOo 0\0\0 -^ ^*J->0'-T-1- t^Co On C -^ 1^ t^^o O, 2 r^ r^ u^ r~~oo On a^ O N ^O CO O N rt- to ii S y •-i •-< ^ f-H '-^ -^ ►s •>! -< -1 <^l 0) 0| 0) M '^ Po '^ '^ '^ -^ -^ -+- ~h ■> -h lo. 'j~; lo lo ^^' lo »o uosc ooomd r^r^i>r^ 2 Si 1- 00 Co Co Co Oo oo OoCoOoCoOoOoCoOoOoCoCoaoOoOoOoOoOoCoOo>DOo ^■OOOOCOOOOOOOOOOOOOOOOOCOOOOOOO Co C CO Q Pi X < Q W < X u Q w -a c 11 60 J3 .60 'S a. s 60 X ^ B o o c 60 C s n J2 3 1^ No hake No hake 27 + 2 juv. 26 14-500 90 47-800 7-500 68 102-650 No hake 37 22 juv. ig-g I -000 No hake No hake No hake No hake 8 1 1 -000 8 8 8. 1 ^§ 1 1 1 1 ltnlir5llo^lMO^IKvol"lo'6lll 1 1 1 looob" wM^ro >OOf>N low H B 0 0 1 1 1 1 1 ^ 1 „ 1 1 « 1 ^: 1 1 1 ,, 1 ^ 1 r. 1 - 1 1 1 1 1 1 i 111 MM ■^ HH a. 3 2 60 ^ 60 C o 1— ( .3 Of Of 1) o o 1 1 1 M II 1 1 1 1 II II 1 1 1 1 1 1 1 1 II II M 1 1 1 1 o o ft 1 1 1 1 1 II 1 II 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 1 1 1 1 o r CO 1 1 1 1 1 1 1 1 1 II 1 1 1 II 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 o oo 1 N „? 1 1 ^^1 1 *.?! 1 1 1 1 1 1 i 1 1 1 o 000 0 llll 0 f-^i >o^ oSioo8| '^"^lllll '^1 1 000 0 ll||| 0 r^ \ r^^ -mSi -,81 «^^||lllll ° \ to <>-> 'Tvo ro'orotoin^vovovo r^-^ob >o v6 u->w-)K-l>ON'ON<0 U-) o 1 5- 000 0 Mill "-^111 ^p ??- ^?^l ?p|lllllllll o 1 CO 1 1 1 1 1 «^ mt 1 m?:l III III 1 ^^1 1 1 1 1 1 1 1 1 II 0^ r^ M OTi (Vi o 1 N 1 1 1 1 1 m^ .§.1 1 1 i 1 i 1 1 1 1 1 0.?.^ 1 1 1 i 1 1 1 1 1 1 C\l 0.) M Cvj D. 3 2 60 a 3 6 o o c 'o 'o j a _ < m <; PQ "^ cq JOvOvOvOvOvO vO t^ !>. t^ t^t^ t^ t^OQOoCooo ^ tn in vi Ui Ui tn xn vi tn cncn ixi mcootjcooq HAKE DATA: THIRD SURVEY 403 O rt o o I ° so I O o o I ° O 3 + > > 0 3 0 3 U-, •— > 00 t^ ro ON N ib + vO + t^ -1- 0 so H o o 00 N O N O O O ON CM O o \0 ■tt "I o C o o o o o o ■^ ^ o On CO 00 •^ m 00 ON to o o 0 O rt I o o o I ON o ON I 3 2 bo e>o c s o o o 00 o I o NO I o o 3 o a 6 u o o o NO I o I o r o o o o o o ON r^ nO o Vi N IN) O 9 o o 9 T' n <>-) t^ o o o m ON --t- io r»5No NO o o 00 vD c povb NO On »1- 10 NO >-i 0-1 N NO 10 fn iNa o o o o o o o ^1 « o o 1- o ^ N^OO On irjOO ir-i -^ 10 "t^nO O ^ O 00 10 O f 1-1 10 1-1 N "O rn 10 N NO N 00 o o o •^NO 00 000 000 o NO ON onno "^ ^No r^ o o r-- < ob 6 o o NO f^ "ONO "1 N. CO o 0 o ° rr i\ « ON ij^ b M (Nj ^l o o o o o O NO < o ON en n o ON en On en M (Nl M ON en < N ON en 23-3 404 DISCOVERY REPORTS s I— I Q w 0 0 m 0 0 ,2 "-I Ou*+ "1 0 0 0 |t^ I "^ \ ^ I'"- l"^ 1^ 1" I'^W" rl 1'^ |0 1*^ 1^ CO 00^0000*^0 U-1U1000 OONOOK1MO MNunioO .r^ |W-j i^tl- it-^ lO |a^ |io rroi |CT> i'^' iN ir^ lOO «/il«volv6 olvo Ml6>clNoolr^vol6ool'o 1 lol^fl^o nIoo wlr^wlN vOt^ \Oio •^'*-r>.u-iO\t^ mOMroNNMi- r\.. M tH Total males J- 0 0000^_i_00 "T 0 LoroiOLn'^OO Os|o.l 1 1 nIm 1 1 1 Mlbf^l^-^lavr^iN 1 r^l^f^l^nii 1 1 1 1 1 1 g a Ji B o o M a o^ Of u 0 T 0 III III III III III III III III 1 III III III III III 0 0 M 1 HI 0 III III III III III III III III 1 III III III III III 0 1 00 III III III III III III III III 1 „ ? t 1 M III III III 00 0 00 1 M 9?^ 9^ ?=?llllllllllll 9?^! ?r^°??>lllllllll 0 0 00 000 0000 10 0 0 0 0, 0 0 0 "O,,, C^^. Oro Oom t^u-1 CMu-> 00 ~^0 *l-K-i tvjio MM M'O 0 M'O" M^Oi-iVO MVON ^"NVO-4-^ VO 0 -0 1 IH >0 000 000 00000 0 0 Oiiiii, 0 0 0, 0 0 0 0 >o 0"-> U-joO '^"^ "^lio ^tn w~)0 >~iO >-iLri 0\0 i~iw-i VOO 0 iri 1 'I- 000 000 00000 0 w-1 0,., 0 0 0, 0 0 lo 0 10 ^rt. OMD t-ii-i ot^ o^" "oo 1-0 oio 'n "^ t^ t^ cjvo oi^ NJs Mt-^oroK^MC^b 0 t>vob ooioNOO^N ot^t^ a^'o " r^6\N m>-o6 moo oo»o 0000 10 lOiot^ 0^000 N 0100 Ofo ^111 ""1° '^°° -^o '0"->| '^'-^111 'II III III 0 1 0 ir> 0 ir> ^ ^ 1 1 1 1 1 1 1 1 1 1 II 1 1 1 '-^ v'^ 'r* ?^ 1 1 1 1 ?? 1 1 1 1 1 1 1 1 1 woo 0''' ''' '' ''vo'-^ooo'oo' '''mi^o''' ''' ''' a. i BO •£ 60 C •u o _c »o »o u .8 III III III III Ml III III III 1 III III III III III 0 1 M 0 00 000 cv.. "-• 1 1 1 \n 0 1 1 1 III 0 w-1 iN,vo 1 1 1 1 1^^ i^<>o 0 00 1 1 1 III °r ^^? 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ^?;|^>f 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 N 0 0000 ■^^111 III III ??^ 9?=P^ ^"^1 III Ml III III III a o ■M CO N ro'i-io^^o i^t^ooc^<^o 0 ^ C^C^C?^C7^0^a^a^O^O^O^CT^0 O O t>l>« t^ r^ t^ :^ i^ t^ i^t^ r-»oo 00 00 ai «3 CO CO CO CO CO-CO coco co co co co u j: .3P -33 & ■M O G e u N HAKE DATA: THIRD SURVEY 40s 4>W4i 0 0 0 Oq> 0 0^0 0 0 0 0 MMj^ 0 0 0 ^^t *A "■) OS^ 0 »o 0 »o ««« ico 1^- |t^ ir'« 1°° |<=p— »ir^ ir- 1^ ic- i=p 000 0 «+t-. NNWco OOOO ooooooo OOOw-j u-»uiNO»oO»o III |°P 1:^ |f^ 1^1 1=? i:*-os|? ir- 1^ 1?^ |°P 1 1 1 MloNlNC»5lro^lcnl wlco©lt-«ol"Nlor»jlcoolo^OOlo M t-- N N N CO 0-. ^1 0 > 0 0 1 1 1 Ml III III III 1 1 1 1 N 1 ?■; 1 r 1 1 1 III II 1 ,, 1 2 S M B u 0 _c ot Of 1 0 1 0 1 1 1 1 1 1 1 1 1 1 1 1 II 1 M 1 1 1 1 1 1 1 1 1 II II 1 1 1 1 1 1 1 0 0 0^ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II II 1 1 1 1 r 00 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M 1 1 1 II 1 1 1 1 1 1 1 1 1 1 1 0 1 1 M 1 1 1 1 1 1 1 1 1 II II 1 1 1 1 1 9^ III 1 1 1 ?| 1 1 1 f*i ^00 M C^ CO 0 0 00 000 llllll 91 ^^lllllll 00 v^S^iii 0^ 00 o)^ 0 1 00 000 000 III III III Ot*- >no| 00? oSv ^o||| oJo <^ S-, chS^ mo-n-^n' r»job co^<>)co'«->oco ' MO" 100b 00 00 in N 0 T 0 00 0.0 0 0 III 000 o| or^ "^"^llll °^ "^olll ^"^lll °°^ HpOMOv' MC^6N^v" ' ' ' ' MrvON^oob ' ' ' wobb ' ' '00^6^0 •O ^ ^ ^ ^ H '^t*^. "^ ^ ?■ M 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ^^L^L^I 1 1 1 i 1 1 «§.! (V) 0-) (V-) CV) 0 CO 1 1 1 III III III III 1 III 1 1 1 ^?| 1 1 1 III III III Oh 1 1 0 M _s u 0 vO 1 M to 1 1 1 III III III III 1 III III III III III III Ml 0 5- 1 1 1 III III III III 1 III II 1 ? 1 III III II 1 ?| M CO 1 1 1 II 1 II 1 1 II 1 II 1 1 II II 1 ^^l II 1 1 II M 1 „?| 0 T M N 1 1 1 III III III III 1 M 1 „^^ II 1 III III III III ■^1 c 0 ■*-» CO WS802A WS802B WS803 WS804A WS804B WS805 WS806 WS807 WS808 WS809A WS809B WS810 WS811I WS811II WS812I 4o6 DISCOVERY REPORTS Q w Ph Cl, 97 I No hake No hake 1-700 6-100 14-050 10 5-000 12-400 6-650 No hake No hake No hake No hake No hake 000000 0 0 "^ 0 0 >/i nI mI hI 1 1 HlMMlvor^l^ol hI^nInhIso 1 1 1 1 1 ^ IH M IH O rt ^1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M Oh 3 s 00 1 i o G. Of o CD o 1 o 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1? 1 M 1 1 1 1 1 1 1 1 M 1 o o 1 00 000 lllllllllll 9r 9^*^ 9? 9?" 9^11111 On 0 0\ 0\ 0\ 0 o 1 00 M II II 1 1 III 1 1 1 H?,!^^ i 1 1 H§l 1 1 1 1 1 1 1 1 00 00 Os o ,? „9 1 i 1 1 1 1 1 1 1 1 „?L? 1 1 1 1 1 1 1 1 1 1 1 1 1 1 In. t^, tx «>, o 1 vO ^5. 1 1 II 1 1 ??■ 1 1 1 III II III III III 1 1 1 1 1 1 ^^ 1 1 ■ ? 1 1 III III III II III III III 1 1 1 1 1 o 1 IH ^ ?; 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 1 1 M \h 1 IH ^^11 II 1 1 III III III II III III III 1 1 1 1 1 o i 1 1 II 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 II M M 1 1 1 1 M to Q. 3 e 1 6 tj o IH O ^? 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 1 1 1 1 1 o 1 IH ^^: 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 o „? 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i M 1 1 1 1 1 1 Co O M N 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 a .2 S CO WS812II* WS813* WS814 WS815 WS816 WS817A WS817B WS818A WS818B WS819A WS819B WS820 WS821 WS823 WS824 WS825 WS833 Xi J3 'S HAKE DATA: THIRD SURVEY 407 -a ■4-J e2 I) a> a>Q>VQ> 0 0 0 Guv 0 0 »o 0 »^ M »^_id_;d_^ 0 u-i u-^ o_M_»^ 0 0 r^ 0 r^ « |«««M ir* 1^ IP |r<«•«T^|^ ir> 1° 1 1^ ir Ji nI J3 Ji J3 Ji oviN mIm ovIm oInj5J3 lnl0^l-llo^Nl-^-O^I lolOvfil'l- _ __«_H MN N MM M__M r^moofiMOO o^oo »« 0 0000 OOMmN MMM en 0 S H g 0000 OO"^ 0«rt 0"-)"-)0 inON Ot^ 1 nI i 1 1 1 f^loMlNOvlMolf) 1 1 HlrnMlro>olo>iHl Nlc^Nl^^ N MM --- CO 0 CJ * 0 ^000 00 1 II 1 1 1 1 ^, 1 r 1 M III III 1 1 + if 0 il^ 1 ?« 1 r^il ^li '«-i^ror--)M N M'l->on H N M M CO CI. 3 2 — 2 CH- u 0 1 HI 0 II II 1 1 1 1 1 1 II 1 1 1 1 II 1 1 1 1 1 1 1 II 1 II II 1 1 II 1 1 8 »-i 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 II 1 1 II 1 1 1 0 T 00 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 00 1 0 0 ■ 00 lllllll ??'lll ?°lllll '.^? 9^111 °P|||||| 0 0 0000 000 00 00"^0,, 000 QUI 0 T' 9 T^ 9 97*" 9"^ T' 9 9? T' 9 T ";'P V^r^ \0 <5yD^O vO^mvo^vO^ 0 1 OOOO GOO GO , , , , , , GG»00,, OOG GO r^ r^ (MO 7100 "^ M T- G tV) "1 P-lOO 'i~i >0 O '^ in ' ' ' o'oNwSN>ou-if-- oo\t-avt«^>>5N^'^rof^u-)m n b H in MU-JiH U-l U-) mU-JMHW-jm U-) MU-JMU-JM Irj" 0 5- GOGO GOO GO m m m O , , GOO OG O >Orn t^r^ OvO 0"-> ~t-0 T-tn OG ^ n-"-i l-G ' Mt^ N0bMr5fs,MMVo6N6\M lO-^MlO^MNirjMOOioMOrjOl^rxN 0 1 IH * G G G "^ G G ||||||||l|l|| o8||||| -oSi Qo8 oSl O ^o!? voS ' NO"''' ' 'oovovoo-^f-Mobb ONob « >o v6 wvo ;^ 0 0 1 N lllllllllllllllllllll 9^ ^? ^?^|| ^r^ 9? a. 3 2 J5 1° 6 0 M ?■ u 0 u-1 1 II 1 1 1 i III III III 1 II 1 1 „?| 1 II 1 M „? 1 1 1 1 1 1 M ».i tj M ">! 'n "-^ 0 i o o o o o o lllllll 9?^||lllllllll ^? 9? ^r^ ^ 9?> ?^?' ^ t^^l-rON-^-M •^M^Tf-q-MN'tt- CO O O G O O lllllllllllllllllllll 9? oS>op2 o opS, ^&. 0 1 O G O GO lllllllllllllllllllll ^v^ '^'^ r'?^ 1 1 v^"? r^^ ' ' ' ''' ''' ''' ' 'l«t^n-N t^O rOOo O ' ' N oo O -^00 G c 0 o^jo-jnorj-j.-^-.). i^ ■^ >). in 'n in m to mm lo OOOoOoOOOoOoOO 00 Oo 00 cooooo 00 00 OOOO 00 CoOnOriCntOCOC/3 JO fr) Cd CrjCOcfl c« CC MM M ^s ^fefe^fe ^ ^ ^ ^S^ ^ ^ ^ ^ ^ o c 3 J3 bJO N 4o8 DISCOVERY REPORTS Si X! Q w < O ra o o o o o o ^ o o 00 O vnu-iO Om ^5 .gv ,00 ,u-» |0p IN ly-'-S IP 1? M loo fi ' •'iv© 1 r--loloO Osl " N 1 ""••C ">! t^'^lvD N"f»j"Trror^mMN ^ !z; ^1 OOOOOO oo 0u-jir-»000 tno 1? |0 IN |t-. |- IV^I 1? 1? '>olr^irjlooool^ool"->t-.iONl"-il lAlt^^lvO M MM^. wmN fOM N e21 o o o o o ^ifoo 1 £00 1 ?;,^i £n 1 ^ 1 1 1 1 III III H 1 M C 6 u o CH- o T o II II II 1 1 1 1 1 1 1 1 II II 1 1 1 1 1 1 1 o o T 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 1 II 1 1 o f M CO III III III III III III 1 III III o 00 1 p. 0000 0 III III ON 00 000 Om| OvO||| t^ «^ 0 iosb rovo rn m ■^ rn 0 \h N r^^o 00 W o-i en N K N o 1 *-* 00000 00000 Com M« Mm '^ in ' III III Ngl^|f 1 1 1 III 1 1 1 Mi 1 1 O 1 M N 1 1 1 ^^1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 1 1 1 a, 3 2 60 S o o M O vO 1 0 1 1 1 ??^ 1 1 1 III III ill 1 III III o r 00000 00 ui 0 0 0 O'*- -(t^ ON OOTj- 0"|||||||||| H g_0 O^;*- -i-^N .O^Sn f,^:, 1 1 1 III III 0000 opS, .08 -p2 o^Milllllllill ■0 txN -"TOO "NOqII' III 1 III III O 1 N 1 1 1 H^S 1 1 1 III III ill 1 III III c WS860 WS862 WS864 WS866 WS868 WS870 WS872 WS874 WS875 C/2 W I— I J Q < W OH 0 ^ J 1-1 m 0 m m m -^ tJ- -^ m w^ '^ Date ""«« ■-■■-.■-■I...'-' 'mm r^ c*^ m f^ >> > > > > > > -d t~- 6 pi 00 r~- 6^^ N M Mm M *^ c .0 CS «5 WS576 WS582 WS583 WS586 Ringdove Inlet Puerto Acero Puerto Bueno Bahia san Nicholas Connor Inlet ni mail Ito aaiziij'^ nioi'i) .ybt PLATE XVI Macruronus magellanicus ( x \). Sketch by Mr E. R. Gunther from a damaged specimen in which the blues were less vivid than in individuals from previous stations. Note that the extension of the fins alters the apparent proportions of the body. (From St. WS99. OTC.) DISCOVERY REPORTS, VOL. XXIII PLATE XVI W:^ E^. X m 'P a I— f •< W O < m Z O OS P Pi U DISCOVERY REPORTS Vol. XXIII, pp. 1-18 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands THE GUT OF NEBALIACEA by Helen G. Q. Rowett CAMBRIDGE AT THE UNIVERSITY PRESS 1943 Price Three shillings net CAMBRIDGE UNIFERSITT PRESS LONDON: BENTLET HOUSE NEW rORK, TORONTO, BOMBAT CALCUTTA, MADRAS: MACMILLAN All rights reserved PRINTED IN GREAT BRITAIN BY WALTER LEWIS MA AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XXIII, pp. 19-36 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands ON A SPECIMEN OF THE SOUTHERN BOTTLE- NOSED WHALE, HYPEROODON PLANIFRONS by F. C. Fraser, D.Sc CAMBRIDGE AT THE UNIVERSITY PRESS 1945 Price three shillings net I CAMBRIDGE UNIVERSITT PRESS LONDON: BENTLEY HOUSE NEJV rORK, TORONTO, BOMBAY CALCUTTA, MADRAS: MACMILLAN All rights reser'ved PRINTED IN GREAT BRITAI N BY b^ LEWIS M A AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XXIII, pp. 37-102 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands 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. CAMBRIDGE AT THE UNIVERSITY PRESS 1945 Price eleven shillings net CAMBRIDGE UNIFERSITT PRESS LONDON: BENTLEY HOUSE NEW rORK, TORONTO, BOMBAT CALCUTTA, MADRAS: MACMILLAN All rights reserved PRINTED IN GREAT BRITAIN BY WALTER LEWIS M A AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XXIII, pp. 103-176 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands THE DEVELOPMENT AND LIFE-HISTORY OF ADOLESCENT AND ADULT KRILL, EVPHAVSIA SUPERBA HELENE E BARGMANN, Ph.D. i -.r CAMBRIDGE AT THE UNIVERSITY PRESS 1945 Price fourteen shillings net CAMBRIDGE UNIFERSITr PRESS LONDON: BENTLEY HOUSE NEfV rORK, TORONTO, BOMBAT CALCUTTA, MADRAS: MACMILLAN AH rights reterved PRINTED IN GREAT BRITAIN BY WALTER LEWIS M A AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XXIII, pp. 177-212 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands THE ANTARCTIC CONVERGENCE AND THE DISTRIBUTION OF SURFACE TEMPERATURES IN ANTARCTIC WATERS N. A. Mackintosh, D.Sc. \ CAMBRIDGE AT THE UNIVERSITY PRESS 1946 Vrice ten shillings and sixpence net CAMBRIDGE UNIFER^SITT PRESS LONDON: BENTLEY HOUSE NEW YORK, TORONTO, BOMBAY CALCUTTA, MADRAS: MACMILLAN All rights reserved PRINTED IN GREAT BRITAIN AT THE CAMBRIDGE UNIVERSITY PRESS DISCOVERY REPORTS Vol. XXIII, pp. 213-222 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands CAMBRIDGE AT THE UNIVERSITY PRESS 1946 Price two shillings and sixpence f->^ hlEBALlOPSlS TYPICA \L , H. Graham Cannon, Sc.D., F.R.S. Printed in Great Britain at the University Press, Cambridge (Brooke Cnitchley, University Printer) and published by the Cambridge University Press Cambridge, and Bentley House, London Agents for U.S.A., Canada, and India: Macmillan DISCOVERY REPORTS Vol. XXIII, pp. 223-408 Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovernment of the Dependencies of the Falkland Islands REPORT ON TRAWLING SURVEYS ON THE PATAGONIAN CONTINENTAL SHELF Compiled mainly from manuscripts left by the late E. R. Qunther, M .A. T. John Hart, D.Sc. CAMBRIDGE AT THE UNIVERSITY PRESS 1946 Price thirty'five shillings net 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 DISCOVERY REPORTS Issued by the Discovery Committee, Colonial Office, London on behalf of the Qovemment of the Dependencies of the Falkland Islands Vol. XXIII, pp. i-vi TITLE-PAGE AND LIST OF CONTENTS \' L CAMBRIDGE AT THE UNIVERSITY PRESS 1947 Price one shilling net 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 iiiiip iiiiiii ' :^r liil ill iiiiiiiii