Given in Loving Memory of' <£>anidM,c Crew Member on the maiden voyage of the R /V Atlantis « Corporation Member, 194^-79 Trustee, IQ^H- -6^ Honorary Trustee e^ Corporation Member, 1979- Oceanog-ra-pher, Writer, Editor, Fisherman, Educator, Mentor Woods Hole Oceanographic Institution So o O m O MARINE BIOLOGICAL LABORATORY LIBRARY WOODS HOLE, MASS. W. H. 0. !. THE LIFE-HISTOKIES OF THE BRITISH MARINE FOOD-FISHES. Honfcon: C. J. CLAY AND SONS, CAMBKIDGE UNIVEESITY PEESS WAREHOUSE, AVE MARIA LANE. ©Insgoto: 263, ARGYLE STREET. llfip?ig: P. A. BROOKHAUS. THE MAOMILLAN COMPANY. WEST OF FIFTEEN - SPINED STICKLEBACK. Castle Rocks, S* Andrews. 5th June, 1885. (Reduced) EE Prince, del F, Huth Littf Edmr ,/z. THE LIFE-HISTORIES ft OF THE BRITISH MARINE FOOD-FISHES BY WILLIAM CARMICHAEL Mf'INTOSH, M.D., LL.D., F.R.SS. L. AND E., F.L.S., C.M.Z.S., ETC. PEOFESSOB OF NATURAL HISTORY IN THE UNIVERSITY OF ST ANDREWS, DIRECTOR OF THE UNIVERSITY MUSEUM AND OF THE GATTY MARINE LABORATORY, AND LATELY MEMBER OF THE FISHERY BOARD FOR SCOTLAND, AND ARTHUR THOMAS MASTERMAN, B.A. (CANTAB.), B.Sc. (LoNix), FORMERLY SCHOLAR AND DARWIN PRIZEMAN OF CHRIST'S COLLEGE, CAMBRIDGE, ASSISTANT-PROFESSOR AND LECTURER ON NATURAL HISTORY IN THE UNIVERSITY OF ST ANDREWS. LONDON: C. J. CLAY AND SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE, AVE MARIA LANE. 1897 [All Rights reserved.] CTambrttigf : PRINTED BY J. AND C. F. CLAY, AT THE UNIVERSITY PRESS. TO THE MEMOKY OF GEORGE WILLIAM, 13TH EARL OF DALHOUSIE, CHAIRMAN OF THE ROYAL COMMISSION ON TRAWLING, 18&3— 85, AND SUBSEQUENTLY SECRETARY FOR SCOTLAND, A NOBLEMAN DISTINGUISHED NO LESS FOR HIS INTEREST IN THE WELFARE OF THE FISHING POPULATION THAN FOR HIS SYMPATHY WITH THE SCIENTIFIC INVESTIGATION OF THE FISHERIES, THIS WORK IS DEDICATED BY THE AUTHORS. PREFACE. TN presenting this little volume to our readers we do not primarily lay claim to any special elucidation of new facts, but it has rather been our intention that it should constitute a popularised epitome of the results achieved by British and foreign scientific workers at the St Andrews Marine Laboratory and elsewhere. These results, hitherto, have been from their strictly technical nature and method of publication inaccessible to many to whom the facts themselves would be fraught with interest. At the same time we may remark that, in dealing with each species of fish we have in many cases been enabled to add, by a careful examination of the type -collection and the fresh forms at St Andrews, important links in their life- histories which till now have escaped observation. We hope that this may enhance the value of the publication, especially to scientific workers in this field of research. As far as has been possible we have, by means of footnotes, acknowledged the authorities from whom we have quoted or have derived our information, and we have not hesitated, both in the text and figures, to avail ourselves of the published contributions of our fellow- workers at home and abroad. From no work have we quoted more largely than from Mclntosh and Prince's Researches, which may be said to have attempted for Teleosteans what the lamented Frank Maitland Balfour did for Elasmobranchs. Vlll PREFACE. As regards our respective shares in the preparation of this book, we may state that the Introduction and Chap. I are the result of our conjoint labours, and that, whilst the senior author is responsible for the Chapters dealing with the Life-history and development of a fish from a pelagic egg (Chaps. II and IV), upon Pelagic Fauna (Chap. Ill), and the typical development of a Teleostean Fish (Chap. V), — the Chapter on the Rate of Growth of Fishes (Chap. VI) is the work of the junior. The composition of the special life-histories of each species was originally divided between us, the latter taking in hand those of the grey gurnard, dragonet, angler-fish, wolf-fish, viviparous blenny, mackerel, fifteen-spined stickleback, cod, haddock, whiting, poor cod, green cod, the rocklings, ling, hake, the sand-eels, plaice, dab, flounder, herring, sprat, pilchard, eel, and conger, the former of us undertaking the remainder. To this we should add that certain remarks were later made to such an extent in the parts dealing with e.g. the dragonet, viviparous blenny, cod, haddock, whiting and rocklings, that they must be regarded as a conjoint production. While this work is only a step in a department to which future investigators will continually add, it is not without some satisfaction that one of us contrasts the condition of to-day with what it was when he undertook the Trawling work towards the end of 1883. Then the life-history of not a single British marine food-fish was known, at least from observations in our country. In the present work between 80 and 90 species are dealt with, the majority of the important forms more or less exhaustively. The larger share of this work has fallen to the St Andrews (now the Gatty) Marine Laboratory, and amongst those who have given their energies to this task E. E. Prince, R. Scharff, J. Cleland, J. Burdon Sanderson, F. Gotch, Marcus Gunn, E. W. L. Holt, W. L. Calderwood, W. E. Collinge, G. Sandeman, H. C. Williamson, J. R. Tosh, J. H. Fullarton, H. M. Kyle, W. Wallace, G. Lawrence, and J. L. Steven deserve acknowledgment. PREFACE. IX For many interesting papers connected with the Fisheries we have to thank our fellow-workers in the maritime states of the Continent, in America, and the British Colonies. Such have been and will be of great value in these investigations. In conclusion, we must express our obligations to the Council of the Royal Institution for the use of the woodcuts illustrating the life-history of a Marine Fish ; to Prof. E. E. Prince, Dominion Commissioner in Fisheries, Canada, for drawings illustrative of the development of a fish in Chap. V : to the Fishery Board for Scotland for many acts of courtesy in these investigations ; to Dr Fulton, their Scientific Super- intendent, for his uniform zeal and energy in promoting the increase of knowledge in his department, and to Mr H. C. Williamson for much information from Italy and Germany. W. C. M. A. T. M. GATTY MARINE LABORATORY, April, 1897. CONTENTS. PART I. GENERAL REMARKS CONNECTED WITH THE LIFE- HISTORIES OF MARINE FOOD-FISHES. PAGE INTRODUCTION . . .1 CHAPTER I. GENERAL REMARKS ON THE EGGS OF MARINE FISHES . . . 11 CHAPTER II. LlFE-HlSTORY AND DEVELOPMENT OF A FlSH FROM A PELAGIC EGG . 28 CHAPTER III. PELAGIC FAUNA .36 CHAPTER IV. LlFE-HlSTORY AND DEVELOPMENT OF A FlSH FROM A PELAGIC EGG — CONTINUED 57 CHAPTER V. GENERAL SKETCH OF MARINE TELEOSTEAN DEVELOPMENT . . 67 CHAPTER VI. RATE OF GROWTH OF FOOD-FISHES . . 97 Xli CONTENTS. PART II. LIFE-HISTORIES OF THE SPECIES. CHAPTER VII. PAGE ORDER I. ACANTHOPTERI . . . .116 The Perch Family 117 The Sea-Perch 117 The Comber 118 The Red Mullet Family 119 The Red Mullet 119 The Bergylt Family 120 The Norway Haddock 120 The Sea-Scorpion Family . 122 The Short-spined Sea-Scorpion 122 The Long-spined Sea-Scorpion 129 The Four-Horned Cottus . . .... 132 The Red Gurnard 134 The Sapphirine Gurnard 135 The Grey Gurnard 135 The Armed Bull-head Family 143 The Armed Bull-head 143 The Angler Family 149 The Angler-Fish . 149 The Weever Family 156 The Lesser Wecvcr 156 The Mackerel Family . .160 The Mackerel 160 The Scad Family . .165 The Horse-Mackerel 165 The Boar-Fish 166 The Dory Family 167 The Dory .... . . 167 The Goby Family 167 The Freckled Goby . . . . 167 The Black Goby 170 The Doubly-Spotted Goby 172 Nilssou's Goby 173 CONTENTS. Xlll PAGE The Dragonet Family 175 The Dragonet 175 The Sucker Family 181 The Lump-sucker 181 The Common Sea- Snail 190 Montagu's Sucker 191 The Lepadogaster Family ... 195 The Bi maculated Sucker .... . 195 The Blenny Family . 200 The Wolf-Fish ... . .200 The Butterfly Blenny .... .205 Yarrell's Blenny ... . .206 The Shanny 206 The Gunnel 210 The Viviparous Blenny .217 The Sharp-tailed Lumpenus . . 223 The Grey Mullet Family ... .223 The Grey or Thin-lipped Mullet 223 The Lesser Grey or Thick-lipped Mullet . . .224 The Stickleback Family . . 224 The Fifteen- spined Stickleback 224 The Wrasse Family .... 229 The Ballan Wrasse ... . 229 The Conner . . 232 Jago's Goldsinny The Rainbow Wrasse .... . 233 ORDER II. ANACANTHINI. A. ACANTHINI GADOIDEI . . 235 CHAPTER VIII. The Cod Family .... 236 The Cod • 236 The Haddock • 245 The Bib .253 The Poor-Cod .... .254 The Whiting 257 XIV CONTENTS. PAGE The Poutassou . 265 The Green Cod 266 The Pollack ... .269 The Norway Pout .... . . .273 The Hake 274 The Greater Fork-beard . 277 The Ling .... . . .277 The Rocklings generally ... .... 284 The Five-bearded Rockling ... . 288 The Four-bearded Rockling . . 294 The Three-bearded Rockling . . . 295 The Lesser Fork-beard . .297 The Torsk ... 299 The Sand-Eel Family . . . .303 The Greater Sand-Eel . 304 The Lesser Sand-Eel 305 CHAPTER IX. ACANTHINI PLEURONECTOIDEI . . . 315 The Flounder Family 315 The Halibut 315 The Long-Rough Dab ... .... 319 The Turbot 328 The Brill 337 Miiller's Topknot 345 The Norwegian Topknot 349 Remarks on Young Topknots 350 The Sail-Fluke . 352 The Scald-Fish 355 The Plaice 356 The Lemon-Dab 366 The Witch . 372 The Dab 374 The Flounder 380 The Sole ... .... 387 The Variegated Sole . .395 The Little Sole 395 CONTENTS. XV CHAPTER X. PHYSOSTOMI. The Argentine Family . The Argentine ...... The Garfish Family The Garfish The Skipper The Herring Family The Anchovy The Herring ...... The Pilchard The Sprat ... The Allis Shad The Twaite-Shad ... The Eel Family The Eel . The Conger Concluding Remarks upon the Mursenoids i'AO E 400 400 400 400 403 404 404 405 422 429 432 433 434 434 450 458 SYNOPTICAL TABLE OF THE EGGS OP MARINE FISHES . APPENDIX. Instructions for the Transmission of Living pelagic and other Eggs of bony fishes to the Laboratory . 465 PLATES I. to XX., with explanations, INDEX . to follow 4C8 509 FRONTISPIECE . To face the Titl<'-)-> ERRATA. Pp. 30 and 168. Hoffman should be Hoffmann. P. 119. Mullus surmuletus, L. should be Mullus barbatus, Cicer. Pp. 121 and 318. Collet should be Collett. P. 191. Hydrallmannia should be Hydrallmania. INTRODUCTION. WHEN we take into consideration the very important place which the marine food-fishes occupy in the general food-supply of the country it is remarkable to find how very little is really known, even by scientific men, of the life-history of the various species so familiar to every housewife. Few there are who could not readily distinguish such common forms as the cod, the haddock, the plaice or the sole when either or all of these were before them, but how many know anything whatever of the habits or the past history of these denizens of the deep ? True it is that a cod is a cod, and a knowledge of its young stages and their food does not in any measure add to the intrinsic value of the fish as an article of diet ; in fact we may go further and say that perhaps — in the case of some fishes- ignorance regarding their habits and food were better, lest their ready sale be affected thereby. We shall see, later, instances of fishes which are of so grotesque or repulsive an appearance that they must often of necessity have their form disguised or their heads and skins removed before being sent into the market lest their mere ugliness should seriously militate against their sale to an unreasoning and easily influenced public. Be that as it may, we find in every other case in which man makes use of the lower animals for his own sustenance a wonderful amount of intelligent observation and study of the laws governing the existence and prosperity of these animals has been brought to bear upon them from the very earliest times of man's existence. M. F. 2 INTRODUCTION. We may say with perfect justice that the fashing industry has been conducted by man, until recent times, in a manner which is at the same stage of evolution as that reached in the adaptation of cattle to his use when the primitive man hunted and slew in the chase just the few oxen he might from time to time require, quite regardless of the laws governing the existence and increase of the victims. Now however we find that cattle and sheep are carefully bred, fed and tended with one sole object in view, namely, the production of a maxi- mum in quantity and quality of nourishing food, with a mini- mum expenditure. To attain this end a large proportion of the human race spend their ceaseless energies upon the breeding, rearing and fattening of the animals which serve for the food of themselves and of their fellow-creatures. The same remarks apply to the other large food industries, always excepting that of marine fishes. In this case we have only the mere capture, and neither time nor thought is ex- pended upon the breeding or rearing of the finny tribe. There are obvious reasons for this state of things — when we come to consider the vast extent of the watery area dealt with and also the inaccessibility of the ocean-beds to the would-be investi- gator; in fact so great have these impediments been to the seekers after knowledge in all generations that the majority of the common inhabitants of the sea have clinging round the story of their birth and habits a long train of strange fables and weird anecdotes. What more extraordinary and grotesque inventions could one desire than the common but now historical notions regarding the life-history of the barnacle, or the habits of the pearly nautilus ? Similar remarks apply to the quaint fables which are prevalent amongst those who are continually in contact with, and derive their livelihood from the capture of, the food-fishes surrounding our shores. So great have these obstacles been that only in the last few decades have there been any really successful attempts to elucidate the history of the egg of the fish after its deposition by the female. It is evident that before attempting to bring the fishing INTRODUCTION. 3 industry in any way into line with other industries, a thorough knowledge of the life-history, food, habits and diseases of the food-fishes is essential. With this aim in view scientists con- tinue to accumulate facts relating to these, and although to the ordinary observer some of these facts may appear to be of no economic importance, as not directly bearing upon the points at issue, those more familiar with the progress of knowledge are aware that at any time some discovery seemingly insignificant in itself may throw a flood of light upon certain phenomena in a way which may directly benefit the progress of the wholr industry. Historical Remarks. It is, as already indicated, little more than a decade since the eggs and larval stages of almost all our British food-fishes were unknown — at least so far as regards their study by men of science in this country. For_ the discovery of the fact that the eggs of the cod, haddock and gurnard are pelagic, that is, float freely in the ocean, we are consequently indebted to Prof. G. 0. Sars of Christiania, an able naturalist- trained from boyhood under a distinguished father, and who by a fortunate appointment to a fishery post in Norway was enabled to make these and other important observations from 1864 onwards. Prof. Sars had gone in January of the year just mentioned to examine the cod-fisheries of the Lofoten Islands, and had watched enormous numbers of the fishes which come in to spawn. Toward the end of February he observed the earliest spawning. " By fishing with a fine net1 on the surface of the sea I caught some small, completely transparent globules, floating on the water, which I at first took for some very low species of aquatic animals, as I was entirely ignorant of the peculiar spawning-process of the codfish, to which I shall now refer."" I had in former times heard fishermen say that the roe of the codfish could be seen floating in the water, and that at certain seasons it filled the sea to such an extent as to make the water appear quite thick : but as this was in direct opposi- tion to anything I had hitherto known of the spawning of fish, I could not but suppose that what had been taken for spawn 1 Quoted from U.K. Fish Coininis. Kept, (for 1877), 1879. ] 2 4 INTRODUCTION. was in reality nothing but those lower aquatic animals which (as is well known) often fill the sea... Gradually these eggs, floating about freely, became more numerous, until, about the end of March, they filled the sea... I now succeeded in following their development step by step until the tender little fish slipped out of the shell and swam about in the water." On a calm day on one occasion he found the surface of the sea covered with a dense layer of floating spawn, so that with a sufficiently large net he could have taken tons of it. This occurred over a celebrated fishing-ground, on which the cod were present in enormous numbers, so as to form what the fishermen called a "fish-mountain," the sounding lead or sinker in going down striking them as it passed. He also noticed that the spawning of the cod did not take place all at once, but thought it lasted several days. He also observed that the female fishes were nearer the surface than the males — both eggs and milt rising towards the surface. The haddock was noticed to have similar eggs. In 1866 and 1867 he followed the further growth of the larval cod to the post-larval stage of 7 — 8 mm. with their large broad/ heads and protruding eyes, while their tails vibrated like a fine thread in the surface- water as they fed on the very minute crustaceans. In June again, in the sounds and inlets inshore, he found a larger stage up to 24 mm. in incredible numbers snapping at the myriads of small crustaceans just mentioned, and which are generally known under the name of herring-food. He failed for some time to get the later stages, but at last he met them and the young haddock sheltering under the discs of the jelly- fishes, for the purpose of feeding on the small crustaceans that are parasitic on them or that become entangled in their envenomed tentacles. His largest form had now reached the length of 40 mm., a little more than 1^ in. Next season he procured in August older forms of 50 to 60 mm., near lines of floating sea-weed and debris, and pointed out how the dark cross-streaks he noticed in the younger forms of the previous year had now dissolved into three or four parallel lines of square spots of a more or less bright reddish-brown colour, which contrasted beautifully with the light hue of the body, resembling, INTRODUCTION. indeed, a chess-board in the regularity of their arrangement. The sides and the head had an alternating silvery and golden gloss. The thread-like "beard" under the chin was now present, and the aspect was truly cod-like. They also occurred abun- dantly at the margin of the rocky shores and quiet bays, where they were better protected from the voracious pollack which decimated them in the open water. Their average size inshore remained till far into September from 60 to 70 mm., and this apparently stationary condition was explained by Prof. Sars as due to constant emigration and immigration of the swarms of young cod. The larger forms sought the deeper water, the smaller the inshore water. Nor was the backward condition of British information on this head to be wondered at. The authorities entrusted with the patronage of posts in which marine zoology could be studied as a rule and with a singular impartiality filled them with those accustomed to other departments of the subject, while men imbued with enthusiasm for marine zoology were stationed far inland. Marine investigations, therefore, were often made under disadvantageous circumstances, or altogether ignored. More- over, it is only within comparatively recent years that young observers could be trained in marine zoology with any prospect of future advancement, and even now the condition in this respect is far from satisfactory. It is no marvel therefore that progress in this country in regard to the life-histories of the food-fishes was more or less in abeyance. Trawling Commission of 1883-84. When, in 1883, the Government of this country, in consequence of urgent com- plaints made by the line fishermen against the practice of beam- trawling, which, long pursued in English waters, had now made considerable progress in Scottish seas, appointed a Royal Commission to inquire into the subject, it was soon found that there were no well-ascertained facts with which to meet the great variety of statements brought forward on all sides. It . became necessary, therefore, to undertake a series of scientific investigations in connection with the marine fisheries, and, amongst other points, the eggs, larval stages, young and adolescent conditions of the food-fishes received special attention. G INTRODUCTION. At this time, indeed, the observations of Prof. G. O. Sars were little known in this country, the floating or sinking of the eggs of such fishes being vaguely associated with the temperature of the water. Yet Sars had shown at the International Fisheries Exhibition in London this year (1883) a series of drawings illustrative of the development of the cod, and the Americans had for some years been experimenting in the artificial hatching of the same species. Fishermen's views. At this time the almost unanimous opinion of British fishermen, and not a few others, was that the common food-fishes sought the shallow water of the bays and inshore grounds for the purpose of depositing their eggs on the bottom, for demersal eggs were alone known to them. There is no doubt that the masses of the eggs of the lump- sucker and sea'scorpion (Oottus) had been in some instances mistaken for those of the haddock, and, so great was the interest, even ingeniously hatched to prove it. Nor were British fishermen alone in this idea. The late Prof. Spencer Baird. the originator and first Chief of the American Fish- Commission, found the same condition on the other side of the Atlantic. The fishermen had not the slightest idea of this floating property, but thought that the female food-fishes deposited their eggs on the rocks, where they were fertilised by the males. They had indeed noticed the little transparent bodies in the water, but it never occurred to them that they were the eggs of any fish. In this respect, therefore, as already stated, they were considerably behind the Norwegian fishermen, who pointed out the pelagic eggs to Prof. G. O. Sars. Yet it was only necessary to examine the more or less ripe " roes " of such fishes as the cod as they were thrown, for instance, amongst the offal on the pier at Anstruther to satisfy even the most cautious on the subject, and more especially by placing the mature eggs in a pail of sea-water. Literally this was the mode by which the late Lord Dalhousie, chairman of the Royal Commission above alluded to, first became acquainted with such pelagic eggs — while waiting at Anstruther for the ship which then carried him sea- wards— where multitudes of similar eggs, mingled with the larval fishes from a few hours to a few days old, filled the tow-nets. INTRODUCTION. 7 Yet though this was the condition of the information on the subject in our country, the labours of the American Fish Commission and Prof. Alex. Agassiz had shown that besides the cod, haddock and gurnard the majority of the American flounders, certain kinds of wrasses, a species of sparling, shad, mackerel, Spanish mackerel, a kind of dory and the frog-fish were amongst those which had floating eggs. The late Dr Malm of Gothenburg further increased the list by discovering that the eggs of the plaice were similarly buoyant ; and the late able worker, G. Brook, added to this category the eggs of the lesser weever and the rockling. When the Trawling Report was issued considerable dis- appointment was evinced by those who firmly believed in the supposition that the eggs of all the food-fishes were deposited on the bottom, and thus were destroyed by the trawl. Much of this, however, was due to other influences than those which arose on the part of the fishermen themselves, a feature as prominent and far-reaching to-day as in 1884. Those fishermen, however, in contact with the Marine Laboratory at St Andrews, and especially those who saw the work amongst the pelagic eggs then being pursued by the earnest and genial Prof. E. E. Prince1, and observed the manipulations of the attendant who went amongst them and removed from the dead fishes the ripe eggs and milt, placed them in sea-water and showed them how the former floated as minute spheres of glassy transparency, soon altered their opinions. Many of them were by and by provided with earthenware jars which they took to sea, and in some instances were successful in bringing to the laboratory fertilized and floating eggs of forms not yet examined. Thus a change was speedily brought about without much notice, so that in July of the same year (1885) when, with Prof. Ray Laukester, at Oxford, urging the University to support the scheme for the proposed Marine Biological Laboratory at Plymouth, one of us was able to adduce this fact as an instance of the effect of such institutions even on the opinions of the fishing population. 1 Now Commissioner of Fisheries in Canada. S INTRODUCTION. Floating or Pelagic Eggs. During the investigations con- nected with the Royal Commission on Trawling ample oppor- tunities were afforded for becoming acquainted with the pelagic eggs of sea-fishes. They were found to frequent the surface, mid-water and bottom, and their abundance in a given area was generally diagnostic of the prevalence of the parent fishes. The latter feature was boldly illustrated by contrasting in April the area at Smith Bank, where the eggs were in enormous numbers, with that off the Forth, where they were fewer. In the former area the vast multitudes caused the sea to resemble a great hatching-pond in which eggs and larval fishes occasion- ally were driven by surface-currents into long lines marked at a distance by numerous ducks and other sea-birds that greedily fed on them. It often happens that when one region of the sea is almost devoid of pelagic eggs, for example the surface, they may be found in mid-water or near the bottom. Thus the investigations with the trawl-like tow-net show that a vast number of pelagic eggs, such as those of the cod, whiting, rockling, sole, flounder, gurnard, sprat and other forms, are to be found near the bottom, when the surface and mid-water nets are devoid of them. The cause of this feature is not fully known, but it may, as in the case of pelagic invertebrates, be due either to temperature or to currents. The areas usually called fishing-banks especially abound in these pelagic eggs during the spawning season, and thus in a sense the examination of the tow-nets gives certain data in regard to the prevalence of the food- fishes. Pelagic ova, however, may be borne long distances before hatching takes place, and the larva? may subsequently be still further carried from the spawning ground, so that their distribution is amply provided for. Thus it happens that the inshore grounds receive supplies of eggs from the offshore, where many of the breeding fishes are, and on the other hand the young and adolescent fishes leave the inshore and seek the deeper waters beyond. To close the inshore grounds therefore (except to the liners) and leave the offshore free to all may not have the result of increasing the food-fishes of the area to a noteworthy extent, since the spawning fishes which supply a constant increment INTRODUCTION. 9 from the offshore are decimated and scattered, and the re- cuperation of the inshore is interfered with. One of us1 has pointed out the necessity for caution in drawing deductions from the experiments hitherto conducted by the Fishery Board in Scottish waters. The great increase in the number of fishes caught by the ' Garland ' in St Andrews Bay and elsewhere in the second year of the work, caused the Board to extend the closed areas, and further to close a large portion of the Moray Firth. Yet next year the numbers captured in St Andrews Bay by the 'Garland' fell greatly, and still more so in 1889, again rising in 1890, but falling in 1891 only a little above what they were at the commencement of the experiments in 1886. No safe deduction therefore was possible under the circumstances, and since the number of the larger fishes seemed in no way to increase otherwise than the special opportunities for capture explained, such action was probably the result of other data than those supplied by science. The whole subject will shortly receive treatment by one of us, so that it is un- necessary to go into detail at present. Examination of Areas. Since the period of the Trawling Commission more or less systematic examination of the various areas closed to trawlers has been carried out by the Fishery Board for Scotland by aid of the steam-ship ' Garland,' and constant examination of St Andrews Bay by the boats connected with the Marine Laboratory has given us more precise know- ledge concerning pelagic eggs, larval and young fishes and their surroundings. In many instances the eggs and larvae have been examined in the living condition, while in the others careful preparation has enabled us to obtain a fair knowledge of species, and also to estimate their numbers. As a rule the eggs of each species have certain characters peculiar to them, whether in regard to size or structure, and thus their prevalence in a given area and at a given time may approximately be ascertained. The characteristic appearance of the egg of the sole, for instance, at once indicates the presence of the adult in the neighbourhood, and the introduction of a larger number of 1 A Brief Sketch of the Scottish Fisheries chiefly in their Scientific Aspects. By Prof. Mclntosh, 1892. 10 INTRODUCTION. adults into a particular area is followed by an increase in the numbers of such eggs, as illustrated recently in St Andrews Bay, where about five hundred soles of various sizes were transferred from English waters1. In the subsequent pages are set forth the leading facts concerning the life-history of our more common food-fishes, the more exclusively scientific terms being as far as possible avoided in their narration. 1 The kind co-operation of J. W. Woodall, Esq., M.A., of St Nicholas House, Scarborough, deserves cordial acknowledgment. CHAPTER I. GENERAL REMARKS ON THE EGGS OF MARINE FISHES. Groups of Food-jis/ies. The British food-fishes fall natu- rally under two groups (excluding the lamprey and the hagfish, not treated of in this work), the first, that of the cartilaginous fishes, the sharks, skates and dog-fishes, and the second, that of the bony fishes, comprising all the remainder, and whose life-histories form the special subject of this work. The former differ mostly from the latter in having a soft cartilaginous skeleton throughout life, in having the gill-slits separate from one another and not covered by a shield or operculum, and in other points of anatomy, into which we need not enter here ; but the great difference which concerns us is that of development. The cartilaginous fishes have a very few, very large eggs, often protected in a hard chitinous skin which is popularly known as a " sea-purse " or " mermaid's purse." This is usually laid by the parent either on sand or in the vicinity of sea-weeds, amongst which the dark olive colour of the " purse " serves to hide it, or again in other cases the long filaments at the end are attached to various foreign bodies. Most cartilaginous fishes carry the system of " pro- tection " still further and the whole egg is carried about inside the parent's body till the embryo is at a sufficiently advanced stage to take care of itself, when it emerges from its parental covering and commences its individual existence ; this is termed a viviparous form of development. On turning to the bony fishes, such as the cod or the herring, 12 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. we are at once struck by a great contrast between their development and that of the cartilaginous fishes. The egg of the cod is very small, and is surrounded by a delicate capsule or membrane. The main part of its contents consists of a mass of nutritious yolk, ready-prepared by the parent for the use of the little embryo, and as the egg is entirely left to shift for itself, the numbers destroyed by voracious enemies, and still more by adverse physical condi- tions1, must be very great indeed. Hence, in order to maintain the normal number of the species, the cod lays an enormous number of eggs, one female laying several millions. The same carelessness as regards its young is exemplified by these fishes beyond the egg-stage. The young cod escapes from its egg at a much earlier stage of development than the young skate, and is hence in a much more helpless condition, larval cod being at first unable to look after themselves, and at this stage also the mortality of the little fishes must be very high. In this respect the cod is an extreme type of the bony fishes, and in such a very large group of animals there are endless variations both in the character of eggs and in the mode of development. General Remarks un Eggs. In general form the eggs of the ordinary food-fishes2 are circular. On deposition they are usually invested by a single layer (capsule or zona radiata). But some are ovoid or fusiform, as in the case of the anchovy and goby, and others have long filaments attached to the capsule, as in the gar-pike, saury -pike, flying-fish and sparling, these filaments occasionally fixing them to foreign structures. Amongst other interesting types are the large eggs of the stickle- back, the wolf-fish and the salmon-tribe. These eggs, however, are surpassed in size by those of the Siluroid genus Arius, found both in the Old World and the New (Ceylon and Guiana), the eggs of which are somewhat larger than a pea (viz. 5 to 10 mm.). This is not the only remarkable feature in these fishes, for as Drs Giinther and Wyman and Prof. Sir William 1 "Effect of pelagic spawning habit upon life-histories of food-fishes." Report Brit. Assoc. 1896. A. T. Masterman. * Vide Prof. Mclntosh, Nature, April 9, 1885, Vol. 31, pp. 534 and 555. GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 13 Turner have shown, the large eggs are carried without injury by the male in his mouth and gill-chamber until hatched, the small and almost granular palatine teeth making this possible. He thus acts the part of a dry nurse, as also does the male pipe-fish and the sea-horse, the eggs being borne by the male in a pouch on the under surface, the young fishes after hatch- ing even returning when alarmed to the pouch for safety. In another Siluroid (Aspredo) from Guiana the remarkable exception occurs of a female fish interesting herself in the care of the young. At the breeding season the skin on the under surface becomes soft and spongy, and the eggs, which are deposited on the ground, adhere by simple pressure of the body over them, somewhat after the manner observed in the Surinam toad. An Indian Lophobranch (Solenostoma) shares the distinction just mentioned, in which the pelvic (ventral) fins — free in the males — coalesce to form, with the integuments, a pouch for the reception and hatching of the eggs. These and other examples, such as the stickleback and the viviparous blenny, form a series of types in which the amount of protection given to the future generation by the parent approaches that shown in the viviparous cartilaginous fishes. Viviparous Fishes. Too little is known of the life-histories of the viviparous fishes to enable us to generalise with safety, but this much is evident that the young fishes on extrusion are in some cases of a very large size proportionally, as in the viviparous blenny, where they measure 2 inches, but they are comparatively few in number. This widely distributed species, however, has held its own in the struggle for existence, but it nowhere occurs in profusion. The same remarks apply to the viviparous " Norway haddock " of our northern waters. Nowhere are these viviparous fishes so abundant as on the western coast of America, where Prof. Carl Eigenmann, who has recently made a careful study of their development1, found them forming no less than 30 per cent, of the bony fishes at San Diego. They belong to two families (Embiutocidae and Scorpcenidae), the former including fishes frequenting the in- 1 " On the Viviparous Fishes of the Pacific Coast of North America," Bull, U. S. F. C. 1892. 14 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. shore waters, the latter the deeper water, as in our own country (viviparous " Norway haddock "). The large number of species in each family showing this condition on the west coast of America would point to it as the most favourable for their survival. It is probable that this habit was slowly acquired, those having the tendency best marked being enabled to leave a larger number of young, so that their chances of continuing the race were greater. Divisions of Eggs of Food-fishes. The eggs of the bony food-fishes (Teleosteans) may conveniently be divided into two great divisions in accordance with their structure and environment, viz. (1) those which are deposited on the ground, usually called demersal, and (2) those which float about in the water, and are termed pelagic. In the former group, viz. that of the demersal ova, only the herring, the sand-eel, and the wolf-fish (catfish) are conspicuous, while the large majority of the food-fishes have pelagic eggs ; such as all the cod-tribe, flounders and gurnards. No very evident connection exists between the habits of fishes and the condition of the eggs. Thus the pelagic herring — one of the most prolific of marine fishes, and one still as abundant as ever, notwithstanding the enormous efforts of man to compass its destruction — has a demersal egg, deposited in masses on the bottom, while its congeners the sprat and the pilchard have pelagic eggs. The ground- loving frog-fish has pelagic eggs, but the wolf-fish which likewise haunts the bottom has large demersal eggs. The wandering cod-tribe and the race of sand and mud-loving flounders have each buoyant pelagic eggs, which rise to the surface of still sea-water in a vessel like globules of oil. It would indeed be a difficult task to predicate from the habits of a fish the nature of its eggs, since forms frequenting the same region, such as the sand-eel, armed bullhead, weever and dragonet, have eggs totally different in nature, the eggs of the two former being demersal, while those of the latter are pelagic. Pelagic and Demersal Eggs. But whatever may be the cause of this essential difference in the eggs of the marine fishes, and more especially the food-fishes, there can be little question that the pelagic character leads to the dispersion of the GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 15 species throughout the ocean, tends to minimise the destruction of the eggs by any special agency, and appears to have played an important part in the preservation of the various food-fishes. Little explanation of the first proposition is necessary, for it can readily be understood how oceanic currents and tidal changes carry the buoyant ova far and near, peopling regions unknown to the adults and spreading such species as the cod throughout the North Sea and both sides of the Atlantic, and this altogether irrespective of the swimming powers of the larval, post-larval, adolescent and adult fishes. In the case of those fishes which inhabit the bottom and the range of which in many cases is therefore restricted, the pelagic eggs and young carry the species widely throughout the ocean — a provision so conspicuously observed in fixed invertebrates, the larval forms of which often have remarkable swimming powers. The second proposition, that the pelagic eggs escape wholesale destruction by special agencies, will be understood if the almost invisible glassy spheres drifting in every direction are contrasted with the masses of the ova of the herring deposited on the bottom, and which form the food of hordes of haddocks, cod and green cod (saithe) following in the wake of the spawning fishes, the cod for instance greedily gulping the gravel to which the eggs adhere in its eagerness to secure them. As many as eighty boxes of large haddocks have been trawled1, every one of which had its stomach distended with the eggs of the herring, and this is but an indication of the enormous loss sustained twice a year in the life-history of this species. A similar con- trast may be made between the eggs of two rock-loving fishes, viz. the lump-sucker and the rockling. The former deposits its eggs in masses attached to ledges of rocks and stones, and they are often laid bare by the tide, so that besides the depredations of whelks and other forms in the water they are devoured by crows, starlings, and in some cases even by rats when exposed at ebb-tide, notwithstanding the faithful guardianship of the male. The latter has pelagic eggs that are widely disseminated through the water. It is true that occasionally pelagic eggs when occurring in 1 By Messrs Joseph Johnston and Sons, Moutrose. 16 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. enormous quantities, as G. 0. Sars, for instance, describes off Lofoten, may be beached by the tide, but such has never been observed in our country, probably because the fishes are fewer and the spawning-areas at a greater distance from the shore. Again, these small glassy spheres are often engulphed by such pelagic fishes as the herring, when swallowing their crustacean prey, as many as twenty having been observed in good condition in the stomach of a single example. They may also be destroyed by the small crustaceans (shrimp- like forms) amongst which they float, as well as by other invertebrates which form the pelagic fauna usually so abundant in our waters. The third and last proposition, viz. that the pelagic con- dition of the eggs appears to have played an important part in the preservation of the various food-fishes, offers many points for remark. In the first place, fishes which are endowed with this property do not shed all the eggs at once, but only a portion of the roe ripens at a given time and the eggs pass externally, and so at intervals until all the mature eggs are discharged. Several weeks would thus appear to elapse in certain cases before an individual ceases to spawn. The effect of this condition is twofold, viz. to give a much wider area for distribution and a series of gradations in the growth of the young of the same fish. In this way a succession of larval fishes is liberated, and time is afforded for those of one stage to disappear from the surface before those of the succeeding stage take their places. Moreover, it is evident that even if circumstances were unfavourable for the vitality of one series of such ova, they would rarely be unfavourable for all. Such eggs are therefore in a different category from those of the salmon and wolf-fish in which the contents of the ovaries ripen simultaneously and are discharged about the same time. On the whole, it would seem that the advantage is on the side of the minute translucent pelagic eggs, which under the varied circumstances of their periodic discharge seem to be placed in favourable conditions. Again, pelagic eggs as a rule are small, so that instead of the 28,000 or so of ova of the salmon, there are 6 — 9 million eggs in the cod, and a still larger number in the turbot. GENERAL REMARKS ON THE BUGS OF MARINE FISHES. 17 An ample margin is thus provided for losses during the development of the embryo and the growth of the young. The pelagic condition of the eggs likewise gives the species a double means of dispersion, which is important when we consider the great areas of the North Sea and the Atlantic that are yet unexplored as regards their food-fishes. The capture of the adults on a large scale can only be carried on— as a rule — in comparatively moderate depths, the deeper waters thus forming a sanctuary, from which may issue eggs and young fishes for the recuperation of the areas which have been thinned. One of us1 has recently argued for the probability that the pelagic spawning condition is the more primitive, and that the demersal habit has been independently acquired in the later history of the various species which now spawn in that way. The following considerations amongst others were cited in favour of this view : — The pelagic-spawning fishes exhibit a greater fecundity, they have a more extended period of ripening and deposition of eggs; they show no secondary sexual characters, and have promiscuous fertilization. The young in these fishes make the least demand upon their parents (i.e. if we may judge of demand by supply) for nutrition or protection, and are hatched at a much earlier stage of development. Lastly, it is amongst the species with pelagic spawning habit (e.g. the plaice) that the larval and post-larval migration is most marked, the demersal types forming a series (from herrings to small littoral forms) in which the early stages are as it were telescoped up and the migration thereby eliminated from the life-history. The pelagic eggs are deposited by the parents in the water and there fertilized. They are very little lighter than sea water and hence rise slowly towards the surface of the ocean, when their development proceeds till the capsule is burst by the little larva, which then emerges from its confinement. These eggs are all beautifully translucent, a fact which combined with their small size no doubt saves an immense number of them from destruction, as their presence even in a small 1 A. T. Masterman. Op. cit. M. F. 2 18 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. confined volume of water (e.g. in a small jar) is easily over- looked by the casual observer. The appearance presented by a number of pelagic eggs floating in sea water, like tiny pellucid bubbles, is not easily forgotten, especially if they should happen to belong to one of the species (e.g. Gurnard) which have inside the egg a glistening little oil-globule which moves freely in the yolk. Condition of pelagic eggs. Pelagic eggs usually float singly in the water, except in certain notable instances, e.g. the frog-fish, in which the eggs are surrounded by a transparent gelatinous substance having the form of a long riband, as first described by Prof. Alex Agassiz. Prof. E. van Beneden also found some minute isolated and agglutinated floating eggs which he was not able to determine, and Prof. Haeckel procured similar eggs off the coast of Corsica1. Such eggs differ from the pelagic nests of the Sargasso sea, with their masses of fimbriated eggs as described by Louis and Alex. Agassiz, J. T. Cunningham, Mb'bius and others, and which belong to such fishes as Antennarius aod Pterophyrnoides, near allies of the frog-fish. A marked translucency, in many cases almost perfect trans- parency, of both capsule and contents usually indicates the healthy pelagic egg. When developing in the ovary such eggs are quite opaque, as observed in the " roe " on the fishmonger's slab, but towards maturity the granules of the yolk disappear and the ovarian egg becomes quite translucent. From observations made by Mr Harold Dannevig at the Dunbar Fish Hatchery of the Fishery Board in 1884-5 the pelagic eggs of the plaice, lemon-dab and other forms would seem to be shed for the most part at night. When placed in a vessel of sea water the eggs persistently float on its surface, descending but a very little when the jar is rudely shaken. Even after a protracted journey only the dead ones roll on the bottom of the vessel. All the floating eggs are living. For instance, eggs removed from the cod and fertilized on Smith Bank off Caithness, and even at 'a much greater dis- tance, were carried to the Marine Laboratory at St Andrews, still 1 Fierasfer? GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 19 floating, on arrival, at the surface of the water. On transferring the eggs of the cod to a larger jar and turning on a tap of sea water they in a few minutes went to the bottom, the impure sea water on these occasions proving speedily fatal. The addi- tion of a small quantity of methylated spirit in the same way sends all the eggs and embryos to the bottom. Sars, indeed, mentions that if the eggs of the cod are placed in fresh water they sink and never again rise. They are killed — just as a newly hatched salmon is killed, though somewhat more slowly, by immersion in sea water. Sars thinks that even a fall of rain might affect the floating of the eggs in the sea, but this is unlikely, since as a rule the eggs are not, in our seas at least, found quite at the surface. Occasionally the diminished density of shore water suffices to send the eggs captured in the open sea to the bottom of a vessel, but they are by no means killed. The attachment of fine particles of mud and sand in the same way carries the eggs downward, and often proves disastrous to them. More than once the eggs of the haddock and other fishes have been brought under notice as lying on the bottom of a vessel, and therefore held as proof that they did not float. But in every such case of pelagic eggs they were either dead or dying, unripe, and often unfertilized. If in removing eggs from a fish too much pressure is applied, unripe eggs escape ; and they either sink or float ambiguously according to the stage of development. Unless this fact is borne in mind disappoint- ment naturally ensues, especially in the case of those who have brought them from deep-sea fishing to vindicate their state- ments. No one ever asserted that dead eggs floated. It is the ripe and living eggs that are so buoyant. While thus, if care be exercised, there is no great difficulty in transmitting the pelagic eggs1 of marine fishes great dis- tances immediately after fertilization, it is, in critical cases, better to wait till the embryo is outlined before subjecting them to such vicissitudes. The mortality is by this method 1 In stoneware jars tied over with cheese-cloth, and comparatively few eggs in each vessel. 2—2 20 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. greatly reduced, and eggs can be transmitted, for instance, from Shetland to St Andrews with comparatively slight loss. Even though kept for ten days without renewal of the sea water, and the eggs hatch on the way, the lively little cod, with their characteristic black pigment-bars, swim actively near the surface of the water, darting hither and thither when interfered with, while a stratum of the dead lies on the bottom. The water may even be somewhat milky and the odour perceptible, and yet the larvoe survive — until, as Sars also found, the yolk-sac, which supplies them with nourishment, is absorbed. Throughout the spring and summer countless hosts of these little eggs belonging to many species, varying slightly in size and structure, float about near the surface and in mid-water, tossed here and there by the waves and at the mercy of every tide or current. At certain spots off the coasts, at the meeting place of currents, these eggs accumulate in millions and may then be taken in any numbers by dragging fine tow-nets through the water. During the wanderings of these eggs the development of the embryo inside progresses slowly but surely, until through the transparent tissues the principal organs and shape of the future little fish can be clearly discerned, and soon after the lively movements of the same cause a rupture of the egg- membrane, which, no longer required, sinks to the bottom of the sea, leaving the helpless larval fish to toss about in the surface- water until it gradually gains size and strength to regulate its own movements and to hunt for the minute crustacean organisms upon which it feeds. Number of floating eggs diagnostic of number of breeding fishes. The condition of the fish-fauna of the various grounds may to some extent be estimated by the number of the eggs floating near the surface. It has been seen that Sars found the water crowded with a multitude of eggs off the Loffoden Islands, where enormous numbers of cod are captured. In our seas no fishing bank is so prolific, the greatest number of eggs occurring on Smith Bank, off Caithness, and the next 011 the rich grounds off the Island of May — both of which present a great contrast with the meagre supply of eggs of round fishes floating in St Andrews Bay. The proportional numbers in each GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 21 case accorded fairly well with the captures of adult cod at that time (1884) in the several areas. No sight can be more interesting to the naturalist than the surface of the sea, in the condition just mentioned, about the beginning of April. The rough water of the great fishing grounds, such as off Smith Bank, and somewhat further from land, is enlivened by large groups of gulls, guillemots and the ubiquitous gannets, apparently feeding on the smaller fishes which have been attracted to the surface by the wealth of food. At short intervals the long dorsal fin of a huge "killer" appears above the surface and the water behind it is churned into foam by the powerful strokes of its screw-like tail ; while a small group of bottle-noses (another kind of toothed whale — the ca'ing whale of the Shetland Islands) is recognized by the noise and spray, as one or more leap like dolphins from the sides of a huge wave. The tow-nets collect large quantities of eggs and larval fishes which have just been hatched. They further show that innumerable minute crustaceans (Copepods &c.), multitudes of the young or nauplius-stage of sea-acorns and other forms, Sagittce (arrow-like worms), and many peculiar annelids are present. It is evident, therefore, that the young fishes are placed in the midst of a rich surface-fauna, the more minute forms of which would readily serve as food as soon as the supply of yolk disappears and the mouth opens. In the study of nature and nature's ways the naturalist is often brought face to face with truths bewildering in their remorseless fatality, but few chapters of animal development present one with such a history of ruthless destruction of the many and survival of the few as that presented here. A ripe female cod may at a low estimate be reckoned to give rise to 5,000,000 eggs (though in nature it is doubtful if all these are available), and as we have no reason to believe that the number of cod is appreciably increasing we may say with certainty that 4,999,998 of these, at some stage of their career from the egg to the mature adult, will meet with an untimely death at the hands of countless enemies. Facts such as these bring home the intensity of the struggle for existence going on around us even in so large and open an arena as that 22 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. presented by the ocean. The struggle can be none the less keen though in this case the agencies of destruction are probably more of a physical than an organic nature. Before leaving this part of the subject it is well to note that the demand of the succeeding generation upon its parents is the same in each case : — for the cod, in return for the entire neglect of its young from the moment of oviposition, has to meet the enormous demands upon its organisation Avhich are caused by the production and ripening of the huge number of eggs which are necessary to maintain the balance of nature, demands so great as to render the parent-fish almost unfit for food when in the spent condition. This utilization of the tissues of the parent-fish during the growth of the roe is a subject that has recently attracted much attention in connection with the salmon, since it has been found that peculiar chemical changes take place in the muscles under these conditions. Dr Noel Paton, with the aid of Dr Gulland and others in Edinburgh, is at present engaged in an extensive inquiry of this nature on the salmon, a fish in which the changes have important legal as well as physio- logical bearings. We look forward to the publication of these observations and also those of Dr Alex. Brown, on the salmon of the Aberdeenshire rivers, with much interest. We also find that the fishes with pelagic eggs have a larger proportion of the female sex, whilst the reverse is the case with fishes laying ' demersal ' eggs. Thus female gurnards are found in the proportion of four to every single member of the opposite sex, whilst the male gurnard appears to be slightly smaller than the female. In the case of the haddock and whiting the proportion is that of one male to two females ; while in the cod the number of each sex is much more equal, about four males occurring to every five females. As regards the relative size of the different sexes, the whiting, like the gurnard and all flat fishes, has a larger female, whereas the male haddock and cod are slightly larger than the females. In the ' pelagic ' fishes not only are the females more abundant but the ovaries are much greater in bulk when GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 23 compared to the milt than is the case in ' demersal ' fishes. In the cases where there is not a great disproportion between the relative number of the sexes, as, for example, the cod, the disparity in the relative sizes of the organs is more marked, always to the disadvantage of the male sex. Demersal eggs. In the fishes of ' demersal ' habit the males are usually in excess of the females, which also probably results from the fact that the process of fertilization in these fishes is not assisted by physical factors, as in the ' pelagic ' fishes. In the latter the micropyle or minute aperture of the egg is downwards, and the ' milt ' being specifically lighter than the water passes upwards and meets the egg as it slowly ascends through the mid-water. In the 'demersal' species the batches of eggs are close together and are not so readily accessible to the male element. In this category are most fresh water fishes (except the shad), the eggs being deposited on the bottom like those of the salmon, attached to foreign bodies like those of the sparling, or to water-plants as in the carp and the pike. These fresh water fishes are held to be descended from marine ancestors which had already acquired a demersal spawning habit. There are obviously many physical difficulties in the way of a fresh water fish reproducing itself by buoyant eggs. (See ' Flounder ' and ' Eel.') Amongst the marine fishes demersal eggs occur — in the herring, wolf fish, shanny, various suckers, gobies, armed bull-head, sea-scorpions (Cotti), ball an wrasse, fifteen-spined stickleback, gunnel and others. These 'demersal' eggs are so called because they are de- posited by the mother upon the bed of the ocean, and in many cases some amount of regard for concealment is shown in the se- lection of a site for the deposition of them. As we might expect, therefore, these eggs are not usually by any means so numerous and are not so translucent, but often have a typical colour and are opaque. They are usually deposited in masses of scores or hundreds attached to weeds, shells, or rocks, where the development takes place up to the hatching stage, often under the direct protection of the male, as in the case of the stickle- back and lumpsucker. If, as already indicated, we separate 24 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. the bony fishes into two divisions according to the character of their eggs (either pelagic or demersal) we at once note that this character by no means agrees with the characters of the adult fishes which determine their relationships ; thus we note for example that the sprat has a typical pelagic egg, and its close ally the herring, so close that their young forms are hardly distinguishable, has a demersal egg, laid in clusters upon gravel, seaweeds and zoophytes. This and other instances point to the fact that the pelagic or demersal habit is not determined by the general structure of the adult but by other factors. Quite the same absence of connection occurs between the size of pelagic eggs and that of the adult, the turbot for instance having a smaller egg than the plaice, and the cod than the whiting, although the greater size of the turbot and the cod when compared with the plaice and whiting must be familiar to all. On the other hand the size of the eggs seems in some measure to vary inversely as the number, as will be seen by comparing the cod and the whiting. Too much reliance however must not be placed upon this comparison. Lastly, the oil-globule does not by any means characterise one particular group of fishes, the closest allies differing in this respect ; the single instance of the turbot's egg with its oil- globule and that of the plaice with none must here suffice. The study of the development of food-fishes furnishes one with beautiful examples of what is known as convergent evolution. The most important demersal egg amongst the food-fishes is certainly that of the herring. This is attached so securely to the structures on the bottom or to the bottom itself that so far as experience goes it is never tossed on shore or brought up by the trawl in the ordinary course of work. The eggs have a firm capsule coated with an adhesive substance which fixes them firmly to each other and to foreign surfaces. The adults, which show a remarkable persistence in attaining their spawning grounds, take, like the salmon, no further heed of them after deposition and fertilization. In some cases, however, the eggs are tended by both or by one of the parents. The well-known GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 25 case of the stickleback and its nest is an example in fresh water. The fifteen-spined stickleback, gunnel, bimaculated sucker, pipe-fishes, and lumpsucker are instances in the sea, the male in the majority acting as the guardian. Whether the eggs of the wolf-fish (cat-fish) are at first fixed to the bottom is unknown, but they form bulky masses occasionally about a foot square, which are carried by the tides into neighbouring areas. These masses are only a little heavier than the water, and hence are readily wafted about by currents. In other groups of animals, for instance the cuttle-fishes, the same divisions occur in the eggs, some forming demersal masses of various shapes, or being fixed to foreign bodies, while others are pelagic. The majority of them, however, are demersal. Summing up then the general facts concerning the early development of the food-fishes we find three leading types :— A. That of the cartilaginous fishes with a minimum number of large eggs, with a maximum amount of ' protection ' (in the widest sense) in the way of thick capsule, and great quantity of yolk, and a lengthened period of incubation so that the young is considerably developed before hatching. B. That of the bony fishes with demersal eggs. An abundance of small eggs, with a varying amount of protection, either by a tough capsule, concealment, colour and position or by direct contiguity of the parent, more or less close, and a fairly long period of incubation. C. That of the bony fishes with pelagic eggs. An immense number of very small eggs, with small quantity of yolk, an objective protection by translucent appearance but no parental protection whatever, a rapid period of incubation and more or less helpless larval form. Stages after hatching. The later stages of development are usually classified for convenience into Larval, Post-larval and Adolescent. The larval stage extends from the period of hatching till the complete absorption of the yolk-sac, which in larval forms hangs like a large ' hernia ' from the ventral surface of the newly hatched fish. During this period its movements are greatly hampered by this large sac of mi- 26 GENERAL REMARKS ON THE EGGS OF MARINE FISHES. triment, which must be a very costly necessity to the young fish, as a source of attraction to many a hungry foe (see Herring). The post-larval period extends from the end of the larval period till the time when the adult characters are assumed. This period is marked by numerous 'larval organs,' e.g. spines and pigmentations, which disappear before the adolescent stage, and must be regarded as secondary adaptations to the sur- roundings of its youth, affording, in the wide sense, means of defence or safety. At the adolescent stage the young fish has arrived at an age when it is to all appearances an exact miniature of the adult, and further changes take place mainly in growth and maturation of the reproductive organs, though there may be important changes in the relative size of various organs and in pigmentation. In the progress of the evolution of a fish from the moment of its individual existence to the adult stage we have care- fully to distinguish the processes which take place side by side but which are quite different in their nature, namely development and growth. In the former we see the evolution of new organs and the progress of each to greater complexity, form and function rapidly differentiating till the adult stage of organism is reached. In the latter we have cell-division, resulting in increase of bulk, which continues long after the development, in the strict sense, has come to an end, namely at the end of the post-larval period, and which in fishes there is reason to believe only terminates with the death of the animal. In more highly organised animals, as birds and mammals, there is a fixed limit of growth when we may say that the organism has reached maturity and when the re- productive organs come into functional activity, and beyond which there is no further normal increase of bulk, but in the case of fish the size attained seems to depend largely upon the amount of nutriment consumed and upon the time over which the active consumption extends. Thus the size of a fish except in the widest limits is no criterion of the stage of its sexual maturity nor even of its age. It is to be noted GENERAL REMARKS ON THE EGGS OF MARINE FISHES. 27 that a curious parallel to this is found in the group of jointed animals; in these we find that the more lowly organised aquatic shelled crabs and lobsters grow throughout life and have continually and periodically to throw off their hard shells because they become too small for them, whereas the more highly organised land insects and spiders have no such troubles to contend with, but accomplish all their growth in an ado- lescent period before the sexual organs become mature. CHAPTER II. LIFE-HISTORY AND DEVELOPMENT OF A FISH FROM A PELAGIC EGG. IN the roe or ovary the egg of the fish is developed from a minute cell — quite invisible to the unaided eye — to a size that can readily be observed. During this process many important changes go on both with regard to the egg and its surroundings, and amongst others is the change from an opaque mass to a transparent glassy egg at maturity. This is due to the dis- appearance of the granules of the yolk which fills the egg, and of the contained bodies (nucleus and nuclcolus), together with the formation of the female pronucleus. Moreover, at this time the egg increases considerably in size, especially in the pelagic forms, and hence the advantage of the provision whereby only a certain number assume this condition at a given period. If the whole of the eggs of such a fish, for instance as the cod, ripened simultaneously the body of the fish would be ruptured. The condition therefore is different from that in such fishes as the herring, wolf-fish, lump-sucker, sand-eel and salmon, in which the majority of the eggs ripen simultaneously and are shed about the same time. The translucent eggs float about in the water, rising near the surface wherever the adults have shed them, and many are remarkably hardy. Thus the late Prof. Huxley having suggested in 1884 that perhaps the floating or sinking of the eggs was a question of temperature, a series of eggs of the flounder which had been fertilized on the 2nd May were placed in a test-tube LIFE-HISTORY AND DEVELOPMENT OF A FISH. 29 on the 5th1. After standing an hour the majority were floating on the surface, one or two lay on the bottom, while others rested in mid-water. Placed in a vessel of water at 98° F. the eggs exhibited lively movements for several minutes, being carried up and down by the currents, but never remaining at the bottom. The test-tube felt quite warm, yet the eggs floated, and remained floating, as buoyantly in the warm water as in the cold, so that their floating in the sea is not a question of temperature. An interesting sequel, moreover, remains to be told in connection with this experiment. The test-tube had been placed aside and forgotten, but on May 10th while we were explaining the matter to Prof. Ewart of Edinburgh motion was noticed in the dusty test-tube, and it was found that the eggs which had been raised to a temperature of about 98° F. had given birth to little fishes, which thus survived the exigencies of their surroundings — both as regards temperature and water. On the other hand severe frosts are fatal to eggs crowded in shallow vessels, in many cases actual rupture taking place, and the same occurs with large eggs, such as those of the wolf-fish, deposited on the bottom of the vessel. Thus to take the example of the eggs of the haddock during a severe frost in February. The earlier stages had been successfully passed, when the water at the surface of the vessels was frozen into soft flakes of ice — on which many of the eggs were elevated. No sooner was this ice broken than all the eggs were observed to present a whitish opacity and to sink to the bottom. Some of those which had floated in mid-water, or under the trickle from the supply-pipe, escaped destruction, but in a few days they also succumbed during a night of unusual severity, and after the embryos had been outlined. These circumstances, however, are for the most part artificial, for in the sea the danger from such extreme cold would be minimised, since the eggs in winter and spring are generally some distance from the surface2. Out of the little glassy sphere, after a longer or shorter interval (varying from a few days to a few weeks, according 1 Nature, vol. 31, p. 555. 2 Ibid., vol. 34, p. 148. 1886. 30 LIFE-HISTORY AND DEVELOPMENT OF to temperature), comes a minute and nearly transparent fish (Fig. 1), which at first is often as passive in the currents as the eggs themselves1. The tiny young in their helpless state are carried along with multitudes of eggs by every tide into sheltered creeks and bays, peopling waters in which, perhaps, no adults are, and finding both safety and food in the shallower water. These little larvae, each furnished with a yolk-sac, are so fragile that they would seem to fall an easy prey to hosts of swimming companions, on which in the adult state they FIG. 1. Larval Ling, immediately after hatching. would not even deign to feed. The larva soon, however, uses its tail for swimming and its breast-fins for balancing. Its shape is somewhat like that of a tadpole, partly from the large head, but mainly from the great size of the yolk-sac, which contains a store of nourishment on which the little mouthless creature, about 3 mm. long, sustains itself for a week or ten days. In this respect it somewhat resembles the young salmon, in which a much larger collection of the same food supports it about six weeks amongst the gravel in the spawning-bed of the river, though a closer scrutiny reveals certain essential differences. Thus the store of nourishment in the yolk-sac of the salmon is taken up by the blood-vessels which branch in a complex manner over the whole yolk ; whereas in the young cod, though the heart is present and pulsating, not a blood-vessel at first is seen, and none ever enters the yolk-sac. The absorption of this nourishment there- fore must take place by aid of the cells and tissues themselves, 1 For some years the development of fishes has been studied by able workers, amongst others on the Continent, by Gotte, Kupffer, Hoffman, Henneguy, E. Van Beneden, Osjannikov and Raffaele; in America, by Alex. Agassiz, Ryder, Whitman and Bashford Dean ; while in our own country, Ransom, Klein, Cunningham, Prince, Brook and Holt have carried out similar researches. A FISH FROM A PELAGIC EGG. 31 and there is nothing specially wonderful in this, when the conditions in the endoderm of Hydra, and other instances of intracellular digestion are considered. It has been mentioned that these minute and delicate little fishes are nearly transparent, and this is more or less the case throughout, though in the majority — even before they leave the egg — points of pigment ap- pear here and there in the skin, so as to give them a distinctive character (Fig. 2). After hatching, these pigment- spots branch out in a stellate manner, thus becoming more evident, and it is found that in most cases each little food-fish has colours of its own. Thus the cod (Fig. 3) is known by its four somewhat regular black bands, the pig- ment on the haddock being less defined, and chiefly aggregated behind the head, the whiting by its canary-yellowish hue, the gurnard by its chrome-yellow, the ling by its gamboge-yellow, the flounder by its yellow and black, the turbot by its ruby-red, the sole by its stone-colour and so on. All these hues, however, become greatly modified during subsequent development, indeed the pigment in no group of vertebrates shows more remarkable changes between the young and adult states than certain of our food-fishes. Thus, for instance, the cod is characteristically FIG. 2. Flounder, show- ing pigment in the egg- FIG. 3. Larval Cod with black spots or bands, slightly enlarged. FIG. 4. Aggregation of pig- ment in Post-larval Cod. speckled in its tiny youth (Fig. 3), next it becomes more or less uniformly tinted, then the pigment groups itself somewhat irregularly on the sides (Fig. 4) ; thereafter it is boldly tes- sellated (Fig. 5), subsequently blotched with reddish brown, and finally in its adult condition it again puts on more or 32 LIFE-HISTORY AND DEVELOPMENT OF less uniform tints. The ling shows a similar series of trans- formations, the colours, however, differing in their arrange- ment, being marked with gamboge-yellow in its larval, slightly barred in its early post-larval stage, then the body becomes FIG. 5. Tessellated condition of young Cod (spirit -preparation). more or less uniformly tinted in its post-larval phase, and the little fish is furnished with a pair of enormously developed and bright yellow ventral fins (Fig. 6) — so different from the short ones of the adult. It is next striped longitudinally when FIG. 6. Long-finned Post-larval Ling, enlarged. about three inches long (Fig. 7), thus affording a great contrast to the tessellated condition of the young cod. In this stage an olive-brown band passes from the tip of the snout in a line with the middle of the eye, straight backward to the base of the caudal fin-rays. The pale ventral surface bounds it inferiorly, while a dorsal stripe with a beautiful opaline lustre FIG. 7. Young Ling, about 3 inches long (in spirit). runs from the tip of the snout, over the upper part of each eye to the tail, on which it is opaque white, thus giving the fish a characteristic appearance. The dorsal line from the A FISH FROM A PELAGIC EGG. 33 brain backward is distinguished by a narrow edge of dull orange or pale olive, which relieves the colours formerly men- tioned, and the general effect is varied by two black specks in the dorsals. When it is double the length (i. e. 6 or 7 inches) a complete change has taken place in its coloration (Fig. 8). Instead of being striped the fish is now boldly and irregularly blotched — both dorsally and laterally, the region of the white stripe being indicated by the pale and somewhat scalloped area dividing the dorsal from the lateral blotches. Fourteen or fifteen brownish blotches occur between the breast-fins and the base of the tail, and they are separated by the whitish areas, which thus assume a reticulated appearance, and both kinds of pigment invade the dorsal fins. Other touches of FIG. 8. Young Ling in the barred stage. About half natural size. dark pigment on the fins and tail increase the complexity of the coloration at this stage. Again, some species, like the gurnard, have pigment over the yolk-sac before they are hatched, others have not. The dragonet in its post-larval (and pelagic) stage has its ventral surface deeply tinted with black pigment, while in the adult, a ground-loving fish, it is white, and the same occurs in the armed bullhead. The St Andrews cross in the eye of the post-larval stage in a kind of sea-scorpion is another remarkable feature (Fig. 9). No more interesting or more novel field indeed than this exists in the whole range of zoology; but the investigations need ships and boats, with expensive appliances, as well as persevering work for several FIG. 9. Head of young Cottoid seasons. We have only been able to open the field at St Andrews with St Andrews cross in eyes. M. F. 34 LIFE-HISTORY AND DEVELOPMENT OF by the help of the Trawling Commission under Lord Dalhousie, and subsequently by the aid of the Fishery Board. It may be asked, Why is all this remarkable variation in colour ? Just for the same reason that the young tapirs and wild pigs are striped, or the young red deer spotted — the adults in each case being uniformly tinted. Such features indicate their genetic relation with ancestral forms having these marks ; and, moreover, in the struggle for existence such variations in tint conduce to the safety of the young. The view of Eimer that the markings in animals are primitively longitudinal would not suit for many fishes, notably for the young cod, ling, and Pleuronectids, and, indeed, Haacke has already pointed this out from a study of the Australian fish, Helotes scotus1, the adult of which is marked by eight longitudinal bands, while young specimens present in addition a row of clear transverse bands which subsequently disappear. The larval salmon enters the world of a size — though small— that is readily recognisable, viz., about three- fourths of an inch in length, but the marine forms under consideration, from their minute size and glassy translucency, are almost invisible to the naked eye — just a gleam of light broken by the passage of a different medium, or a tinge of pigment, arresting attention. Only in the wolf-fish or cat-fish (which is not much — though it ought to be more — of a food-fish) with its large egg, have we a size nearly reaching that of the salmon at birth. We had left the larval fish tossed about by the currents and unable to struggle against them, now floating with its yolk-sac uppermost, or hanging in the water with its head downward, and again making spasmodic darts hither and thither. Soon, how- ever, it gathers strength, arid at the end of a week or ten days it glides actively through the water, and avoids , , * . FIG. 10. Anterior (ventral) both obstacles and enemies, the young region of larval Cod with cod nimbly escaping the forceps, poising great^ breast-fins. Mag- 1 One of the Pristipomatidre. A FISH FROM A PELAGIC EGG. 35 itself iii the water with its large breast-fins (Fig. 10), and evincing both intelligence and dexterity. Nothing is more interesting than to observe the young fishes in the hatching boxes at Dunbar keeping their heads to the current as soon as strength permits. They form dense parallel rows like a vast regiment. Moreover, this activity greatly promotes re- spiration in those like the gurnard with a motionless mandible, the water being thus sent through the mouth and over the branchial region. Its mouth has now opened and the yolk-sac has been absorbed, while it feeds on the more minute of the little Copepods, especially those almost microscopic in size, that swarm in the surrounding water, and the consider- ation of which, before proceeding further, will form the theme of the next chapter. 3—2 CHAPTER III. PELAGIC FAUNA. THE provision whereby such little fishes find in the ocean food suited to their capacities is one of the most striking features in nature, but it has only recently been carefully investigated1. It is a notion no longer tenable that during the winter and spring the sea, to a large extent, is devoid of the wealth of pelagic life — so characteristic of the summer months just as it is of the genial waters of the tropics. For many years, however, it has been known that a vast abundance of minute life of all kinds is present throughout the entire year- arid from the surface to the bottom. Moreover, during the warmer months a constant succession of you-ng forms rises from the eggs both of the sedentary and creeping animals on the bottom to the surface — where they sport in the summer sun, undergo certain changes, and again descend as they assume the form of the adult. The pelagic young food-fishes — swimming freely in the ocean — thus have a double chance at them — first in their very early stage as they rise, and again in their larger 1 Vide "La faune pelagique du Golfe de Marseille," par Gourret, Ann. du Musee d'ltist. nut. de Marseille, n., 188-4. The pelagic fauna of our shores in relation to the nourishment of the young food-fishes, Ann. Nat. Hist., Feb. 1887. Also Hensen and Mobius in Fiinfter Bericht der Kommission zur wiss. (&c., der deutschen Mee.re, Berlin, 1887, pp. 1 and 109. Recently various observations have been made on the Pelagic Fauna, perhaps the most im- portant being those of Prof. Hensen in the Plankton Expedition of 1889, of Prof. Haeckel in his Plankton- Studicn, of Mr Garstang on the Pelagic Fauna of Plymouth; and "The Sources of Marine Food," James J. Peck, U.S. Fish Com. 1895. PELAGIC FAUNA. 37 and later condition as they descend. The enormous numbers, countless variety, and ever-changing nature of the small animals, either directly or indirectly constituting the food of these little fishes, form an important feature in the economy of the sea. Now it is the arrow-like Sagitta that fills the tow-nets to bursting, now Appendicularians, again the Cydippe-group or Crustaceans (Thysanopods). Such animal forms comprise those long known in the British seas, besides others more familiar to arctic voyagers, or to the sunny waters of the Mediterranean, for with modern apparatus and persistent effort (thanks hitherto, at any rate, to the enlightened views of the Government acting through the Fishery Board) our knowledge is always extending. It is a remarkable fact that it is primarily to plants in inshore waters that the abundance and variety of animals are in many respects due, especially if estuaries also debouch in the neighbourhood. Thus nowhere are the swarms of Sagittce, Appendicularians, Crustaceans, and other forms of fish-food more conspicuous than in the midst of a sea teeming with diatoms, Rhizosolenise, and other algoid structures l. These nourish many of the lower forms upoii which the crustaceans, annelids and other higher types feed, the latter again falling a prey to the fishes. Moreover, while the larger forms of the Copepods and other crustaceans, for example, afford suitable nourishment for the more advanced post-larval fishes, the multitudes of larval crustaceans (Nauplii) are adapted to the needs of the smallest larval food-fishes. Now this plant-life is specially rich in April and May, just when the larval and very young post-larval fishes appear more abundantly in the inshore waters, so that the cycle is nearly complete, viz. from the inorganic medium — through microscopic plant and larval crustacean — to the post-larval fish. We have mentioned the neighbourhood of an estuary as a prolific source of food for young fishes, and we need only explain further by instancing the case of mussel-beds, which for months pour countless myriads of larval mussels into the adjoining sea, far beyond the needs of the area as regards mussel-culture, arid which form a favourite 1 The fact that certain fishes feed on Infusoria has not been overlooked. 38 PELAGIC FAUNA. food of the little fishes at all stages, but especially from an inch and a half to three inches in length. These fishes feed on the young mussels as they settle down on the sea-weeds, rocks, and zoophytes in August, after a free-swimming larval existence. Like some of the forms indicated above, mussels live to a considerable extent on microscopic plants and various minute organisms contained in the mud of the estuaries and other sites, so that a rich and favourite food, universally liked by fishes, is the product of these uninviting flats. Moreover, in passing, it may be remarked that, while everywhere preyed on by the food-fishes, it occasionally happens that in turn the mussel proves a source of inconvenience to them, for, settling on the gill-arches of haddocks, the mussels flourish on a site so suitable for aeration and food that they by-and-by press out the gill-cover and impede respiration, just as the shore-crab (which is also fond of mussels) has its eye-stalks wrenched out by the slow but sure growth of the young mussels which have fixed themselves, wedge-like, in their sockets. Nemesis thus, by a chance of anchorage, converts a favourite food into a permanent inconvenience. Again, in connection with the pelagic food of fishes, it is a well-known fact that adult cod are extremely fond of sea- anemones 1, and some of the rarest species may be procured in their stomachs, a feature by no means surprising when we remember that Abbe Dicquemare cooked and ate his sea- anemones with great relish, and wrote in their favour, as also did Mr Gosse in our country. Now, the pelagic young fishes, instead of roaming near the bottom in proximity to the anemones fixed on the rocks, and running the risk of being themselves captured for food, find in the inshore waters in summer the larval Peachice, worm-like anemones, in great numbers conveniently attached by the mouth to the little jelly-fishes (Thaumantias hemisphcerica, and T. melanops) which occur in swarms in mid-water. Moreover, the somewhat larger young food-fishes (cod and green cod of 2 — 8 inches) 1 A favourite bait for cod in some parts (e.g. Aberdeen and St Andrews), and from the fact, amongst others, that star-fishes do not molest them on the hooks, no bait is more successful. PELAGIC FAUNA. oO show the same liking for the coelenterate group, by browsing on the zoophytes (Obelia geniculata) which cover the stones and rocks with feathery tufts, yet the zoophytes are not much the worse for this treatment, for they by-and-by shoot afresh, and clothe the area once more with a minute forest. The rapidity with which such zoophytes grow is remarkable, though we must remember that in some cases, as in those (Gonothyrcea) clothing the mussels in the Eden, the old stock naturally dies off after having produced swarms of pelagic young. General Remarks on Pelagic Life. In the year 1888 one of us prepared a special report1 on the pelagic animals of the Bay of St Andrews, a region which may be taken as more or less typical for the east coast of Scotland if not for a wider area. The investigations were carried out for a year and summarised for each month. From the results of previous observations made during the trawling expeditions, the pelagic forms may be classified into two great divisions, viz. the Temporarily Pelagic and the Permanently Pelagic. In making these observations a different method from that of the Germans, especially Dr Hensen, was adopted. Prof. Hensen sank nets of a special construction, and calculated the amount of water passing through them in a given time, then having counted the various eggs and animals he apportioned them to the cubic foot of water. The method followed at St Andrews consisted of the use of tow-nets at the surface and the bottom, as well as in mid- water, the latter being a large net — 24 feet in length — attached to a triangle of wood or bamboo — 10 feet each way and hinged at two of the angles, the third being secured with spun yarn. The mid-water net has been of great service in regard to the post-larval and young fishes, indeed no other known form of net has been so successful in capturing these active forms2. If worked from a large steam-vessel this net, which is steadied at the required depth by a heavy bar of lead and floated by a galvanised iron rloat, requires much care in manipulation. 1 Seventh Ann. Rept. Fishery Board, 1889. - Dr Nansen amongst others carried one of these St Andrews nets with him in his expedition to the North Polar regions. 40 PELAGIC FAUNA. Another net has been extremely successful at St Andrews in examining the pelagic fauna near the bottom, viz., one constructed after the manner of a small trawl. The trawl- heads are of very light iron, the beam a slender bar of elm, 8 feet 6 inches long, and the net (18 feet long) is in the form of a trawl-net, but composed of cheese-cloth, with a terminal region 3 feet in length of fine cotton cloth. The mouth of the net is elevated about 9 inches from the ground by being drawn " taut " between the trawl-heads, and has an oblique aperture of 3 feet from beam to foot-rope. This is a fatal net for larval fishes and the multitudes of invertebrates at various stages that haunt the bottom-water in early spring, and should be used only for a short time. The mathematical apportionment of the animals composing the pelagic fauna therefore falls short of the German method (Prof. Hensen's), but for all practical purposes connected with the fisheries the plan here adopted is fairly satisfactory. It gives at a glance the vast resources to be found in the sea for the nourishment of the food-fishes — resources ranging over the vegetable as well as the animal kingdom, the latter compre- hending representatives of every class from the fishes downward. Moreover, an intimate acquaintance with this pelagic fauna alone, and leaving out of view for the moment all reference to the multitudes of animals in sand and mud, under stones and elsewhere, and which are all beyond interference by man, demonstrates the unsatisfactory position of those who labour, either through misapprehension or simply ad captandum— to prove that a beam-trawl deprives the food-fishes of nourish- ment by rendering the sea-bottom barren (sic) — just as a roving enemy might starve the flocks of a population by burning the grass. The importance of the pelagic fauna in supplying nourish- ment for fishes cannot be over estimated, for while the adults of many food-fishes might obtain support from the bottom- fauna alone, it is certain their post-larval and young stages could not. Further, the remarkable adaptation whereby the most minute post-larval forms, such as the very young cod, find in the pelagic organisms every want supplied is a striking PELAGIC FAUNA. 41 feature. Agaiu, the terminal portion of the intestine in the larval herrings captured in March shows a deep greenish coloration, which may be connected with chlorophyll, the green colouring matter of plants. Moreover, even so minute a pelagic form as an Infusorian is occasionally, as for instance by the sardines, swallowed in masses as food. There is no more interesting part of the inquiry than the gradual advent in the early part of the year of the larval fishes, and their great abundance in the spring months, such as March, April and May. Then, while the larval stages of a few still appear in the warmer months, viz. June, July, and August, the predominance of post-larval and young stages are the main features. These become rarer in the pelagic fauna as winter approaches, and finally almost cease to occur in the nets — pro- bably in many cases as much from their increased power and activity as from their scarcity. In the same way the larval stages of mussels and other shell-fishes of importance make their appearance at a stated period, continue in great profusion for some time, and then gradually diminish and disappear. In the case of the mussel the advent of the larval forms in swarms is in touch with the previously ripe condition of the adults on the beds, the long- continued presence of certain of these Iarva3 being connected with the later maturation of the reproductive organs in the littoral and often stunted examples so abundant in many parts, especially on the margins of rocks. Besides the special interest of the pelagic fauna in con- nection with the fisheries, the bearing of many of the facts, e.g. the appearance and disappearance of multitudes of swimming jelly-fishes, is of a novel kind, since even the most recent and most authoritative investigators of the subject, such as Prof. Haeckel, do not exhaust the question. Little indeed has been done in this respect since the days of Edward Forbes, with the exception of the Notes made at the Naples Zoological station, and those at the Plymouth Laboratory. The pelagic fauna round the British shores seems to have many features in common, but the presence of the phos- phorescent Noctiluca in the south, of the large Arachnactis in 42 PELAGIC FAUNA. the north, of the purple lanthina, of Velella and of Salps on the west, and of Cydippe, Lesueuria and Appendicularians, on the east, indicates certain regional distinctions. Though not strictly belonging to any special group or to any special season, the pelagic mud carried by the bottom-water plays an important part in feeding the sedentary mollusks, asciclians, cirripedes and other fixed types. It is only necessary to examine the stomachs of asciclians and of edible mollusks, such as the oyster and the mussel, to observe the large number of Infusorians. Diatoms and other Algse, the abundance of sponge-spicules and organic matter of various kinds, which have thus been swept by currents in their neighbourhood amidst a plentiful supply of mud and sand. Pelagic Life during the various Months. January. It is unnecessary in the present work to go fully into the nature of the pelagic fauna during each month : the salient features alone will be alluded to. Thus in January the floating eggs of food-fishes were for the most part absent, as were also the larvae, though young sprats were occasionally found near shore. Pro- minent amongst the mollusks was Spirialis. Crustaceans were in great force, such as Thysanopods, sessile-eyed forms (e.g. Parathemisto, which attained its maximum this month), with swarms of Copepods. The arrow-like worms (Sagittce) were often in great profusion, indeed it was in January that the tide stranded them on the west sands many years ago, when first found in Scotland by the lady to whom the "St Andrews Fauna" is dedicated. They were observed to sparkle along the line of the retiring tide in the setting sun like needles of glass. The equally translucent Tomopteris, an annelid of great beauty, was also very common. Numerous jelly-fishes likewise thronged the sea, such as Tima, Stomobrachium and Aglantha, with the Ctenophores — Cydippe, Lesueuria. and Beroe. The mitre-like Aglantha was one of the most characteristic jelly-fishes of the winter months from January to April, and was frequently in great numbers. Of Infusorians, Ceratium tripos, a phospho- rescent form, was in abundance, and Radiolarians occurred occasionally. Amongst plants, diatoms and the lower algoids were frequent — with a few Rhizosoleniee. The sea is thus very PELAGIC FAUNA. 4H different from the land, since even in mid-winter a profusion of animals — both minute and delicate, together with many microscopic plants — are carried about by its currents or swim actively in its midst. February. During February the pelagic eggs of the food- fishes were more noticeable, and included those of the plaice, long-rough dab, dab, haddock, and green cod — along with various larval and post-larval stages — amongst others those of the large form resembling a goby (Grystallogdbius) and young wolf-fishes. Numerous young Appendicularians repre- sented the allies of the vertebrates. Of shell-fishes, minute bivalves, numerous young examples of the pteropod Spirialis and a few of another example of the same group, viz. Clione, were the most noteworthy. No form is more beautiful than the latter (Clione), which was first observed by Prof. Prince, the gaily coloured little sea-butterfly, as it has been aptly termed, dancing to and fro in the water with its extended arms, like a deft human swimmer. The larval form (Cyphonautes} of an encrusting species allied to the sea-mats was the only example of the Polyzoa. Crustaceans were in great numbers, such as forms mentioned under last month, and especially Parathemisto, besides many additions. Of Annelids the females of a species (Autolytus) carrying eggs on the under surface of the body, and the male with bifid palpi, were not infrequent, while the sexual forms of another (Nereis), which undergoes strange modifications of structure and habit at the reproductive season, occasionally occurred. Various larval forms of the Nerine group also abounded. Sagittcv were in swarms throughout the water. The jelly-fishes in addition to Tima and the prevalent Aglantha, included an occasional Thaumantias, besides the Ctenophores formerly mentioned, some, such as Cydippe, in the young condition. Towards the end of the month also the saucer-like divisions of the hydroid stage in the develop- ment of the abundant yet pretty jelly-fish Aurelia became frequent. They often indeed appeared in swarms in the tanks, having been pumped up from the sea. Of the elementary types various species of Infusoria and Radiolarians were captured. Plant-life was present in the shape of 44 PELAGIC FAUNA. multitudes of spores, minute algae, a peculiar algoid body like a minute radiolarian, and numerous diatoms. In southern waters, e.g. the estuary of the Thames, the brilliantly phos- phorescent Noctiluca was conspicuous, and occasionally at Plymouth the Siphonophore Diphyes. March. A great increase took place in the number of pelagic eggs of fishes in March, those of the haddock, ling, rockling, gurnard, plaice, dab, long-rough dab and flounder having been met with ; while towards the end of the month larval cod, plaice, rockling, and the demersal eggs of the gunnel and other forms occurred. The most conspicuous larval fish, however, was the herring, which was found in vast swarms in the bottom-nets in such bays as that of St Andrews, where the adults are seldom seen in numbers. Young appeiidicu- larians were also frequent and in considerable profusion, yet at Plymouth ripe adults were common. The pelagic shell-fishes consisted of a few examples of Spiricdis, other larval univalves, a few minute bivalves, including mussels. The cuttle-fishes were represented by a young squid an inch long. The larvse (Cyphonautes) of an encrusting Polyzoan were plentiful. The crustaceans again comprised many adult opossum- shrimps, the larvae (Zoece) of the shore-crab, the larva? of shrimps, and myriads of larva? of sea-acorns. The Copepods were in swarms, Tomopteris and the larval annelids were very abundant, while Sagittw, which form a favourite food-supply of fishes, were in myriads. No larval star-fishes were present, but minute brittle-stars occurred the first week. Larval sea-anemones, the young of true jelly-fishes, swarms of developing small jelly-fishes (Hydromedusce) of various species were frequent. The cteno- phores (Lesueuria and Pleurobrachia) were common, and larva? of the former were also captured. April. No month presented a more conspicuous collec- tion of pelagic eggs than April, the surface of the sea in many parts abounding with vast multitudes of them. They consisted chiefly of eggs of cod, haddock, whiting, poor cod, rockling, gurnard and sprat amongst round fishes, and of plaice, dab, long-rough dab, brill and flounder amongst flat PELAGIC FAUNA. 45 fishes. The larval and post-larval forms comprised herrings, sprats, gadoids of various kinds, gurnards, swarms of sand-eels, besides suckers, long-spined Gotti and others. The appearance of multitudes of sand-eels in the bottom-nets, as well as by and bv at the surface, was a characteristic feature. v While the Appendicularians occasionally showed themselves in February, yet their enormous numbers during April and at the beginning of May far surpassed the earlier period. The huge mid-water net was filled like a balloon with them and their gelatinous " houses," so that the patience of the boatman was well-nigh exhausted by the constant and heavy strain, as well as the frequent ruptures of the net. It was a relief when they disappeared. There could be little doubt that like other ascidians they were eaten by fishes, and from their prodigious numbers they were thus important. They have long been known to occur in Scottish waters, for Edward Forbes in 1845 found that the cloudy patches of red colouring matter in the sea off the north of Scotland consisted almost entirely of them. They were also frequently met with all along the eastern shores during the work for the Trawling Commission, but their prodigious numbers were only clearly made out at St Andrews. Their food apparently consisted of the peculiar gela- tinous algoid masses and similar structures, and many were ripe. Of pelagic shell-fishes the most remarkable was the graceful pteropod, Glione, hitherto considered one of the rarest British marine animals, indeed the late Dr Gwyn Jeffreys, long the authority on the group, could only quote Dr Leach, who found in 1811 several mutilated specimens on the rocks of the west coast, and a single living one off the coast of Mull. It forms a prominent part of the food of the right whale in the Arctic Sea, and is a prize for any food-fish, though the size fell considerably short of the northern examples. Crustaceans were represented by vast multitudes of Schi- zopods, which were occasionally stranded like lines of chaff along the beach, or made the littoral pools semi-solid. They were eagerly eaten by most food — and other fishes. Larval cirripedes in various stages and allied forms — with myriads of Copepods and other minute crustaceans besides larval stages (Zoew) of the higher forms, still further augmented the list. 46 PELAGIC FAUNA. An increase in the numbers of the larval annelids also took place in April. The sexual forms of Nereids and the tinted Syllids (Autolytus) bearing eggs were also common, along with the translucent Tomopteris and Sagitta, though the latter were sometimes less conspicuous than during the previous month. The jelly-fishes, especially the smaller forms ' (Hydromedusce), were plentiful, some bearing buds from the central stem (manubrium) or at the margin, as in Lizzia and Hybocodon, others issuing in swarms from the hydroid stocks, as in Clytia. The pelagic stages (Arachnactis-st&ge) of Peachia were also present. The ctenophores were abundant, and young stages were occasionally met with. Minute forms of plant-life crowded the water, and their great profusion had a close relation to the abundance of various pelagic animals. May. Amongst the pelagic eggs of the food-fishes that of the gurnard was prominent. It was accompanied by those of the whiting, pollack, poor cod, rockling, green cod, sprat, haddock, plaice, dab, lemon-dab, brill, topknot and others. As a rule, however, during easterly gales most of the eggs were found in the lower parts of the water. Post- larval gadoids were abundant, and during the month it was easy to follow some of these to unmistakeable young cod. There were also numerous clupeoids, myriads of sand-eels and young flat fishes, besides the larval and post-larval stages of many inshore fishes, such as the lump-sucker and armed bullhead. On the south coast pelagic brill were common in the inshore waters. The Appendicularians occurred generally throughout the month in all the nets (surface, mid-water and bottom). The largest were frequently found in mid-water, and they were quite as fine as any procured during the Challenger expe- dition. It sometimes happened that when certain algoids (Rhizosolenise) occupied the upper regions of the water the Appendicularians held the lower regions. Young mollusks were represented by minute univalves like Velutina and many others like Natica. Crustaceans from Copepods upwards were plentiful, and PELAGIC FAUNA. 47 larvae (Naiiplii and Zoeat) were especially prominent, some of the latter being in the large-eyed stage towards the end of the month, while the larval sea-acorns (Cirripedes) had also fixed themselves as Cypris-larvse at the same period. The pelagic larval annelids had increased in number and variety, and the larger adults, such as the sexual forms of Nereids, Tomopteris and the Cheetognath, Sagitta, were likewise frequent. Amongst Echinoderms the reddish eggs of sea-cucumbers were often captured. These issue as long strings but after- wards break up into isolated eggs. As a rule the ' painter's- easel ' larvae were later in making their appearance, but occasionally some were procured as early as the 14th May in the northern waters. The various forms of jelly-fishes were largely augmented, especially the Hydromedusse — amongst which the somewhat rare Hybocodon was numbered. It was interesting to watch the growth of the true jelly-fishes as the month advanced. The Ctenophores were abundant, viz. small examples of Cydippe from 1 to -| an inch in diameter, and the brilliantly phosphorescent Beroe and Lesueuria, the latter often in swarms; yet hitherto it had been unknown in Britain, though discovered in the Mediterranean in 1841 by the distinguished French zoologist, Milne-Edwards. Minute alga3 were plentiful, such as the curious gelatinous masses alluded to in April, Rhizosolenise, diatoms and myriads of spores. Rhizosolenise were sometimes so abundant as to interfere with the working of the nets, the pores of which were plugged — thus retaining the water and masses of Appen- dicularians with their gelatinous "houses," so that the boat was anchored or the ropes of the sails broken. June. In June the pelagic eggs became less conspicuous than the post-larval and young stages of fishes ; yet eggs of the gurnard, rockling, sprat, dragonet, sole, solenette, dab and flounder were still present. The young gadoids for instance were prominent pelagic forms during the month — ranging from 7 mm. with an entire marginal fin, to 24 mm. They fed on larval cirripedes and other forms (NaupUi 48 PELAGIC FAUNA. and Copepods). Some of the smaller of these were shorter and stouter — with a shorter snout and heavier head, while the jaws were less lanky and the permanent rays of the fins seemed to be far advanced for their length. They appeared to be mere varieties of the cod, and it has since been found that the young haddocks seek the offshore water along with many of the whiting, while the young cod pass inshore to the laminarian region. Young cod, again, of an inch and a six- teenth in length and having tessellated pigment on the sides, appeared in the bottom-nets towards the middle of the month, and older forms were captured in the trawl along with young clupeoids an inch and seven-eighths. Young sand-eels were still abundant, along with young gobies from 3 to 8 mm., and other forms already mentioned. Appendicularians were somewhat less common than in May, only a few small examples now and then appearing. Rarely were they numerous. At Plymouth salps sometimes abounded. Pelagic young mussels now appeared in considerable numbers in the tow-nets, their size varying from '0055 to '014 inch. It is during this month that they fix themselves on the various ropes and other parts of the salmon-stake-nets. The pteropod, Spirialis, with other minute univalves were abundant. The larval form (C yphonautes) of an encrusting Polyzoan was common. Crustaceans everywhere abounded. In the bottom-nets Copepods were especially prominent, and stragglers from their dense ranks often sought the upper regions of the water. They were accompanied by swarms of the viviparous Evadne. The later stages (Cypris-stage) of Cirripedes of many larvas of the shore- and porcelain- crabs were frequent, while the young of other crustaceans also occurred. Numerous examples of pelagic young annelids of many species were procured, all of which form a favourite food of the young fishes. Tomopteris, in the adult state, with the fully developed reproductive organs, was not uncommon. The rare Mitraria was also occasionally found. Sagittce were few and small. The 'painter's-easel' larvse (Plutei) of brittle and other star- PELAGIC FAUNA. 49 fishes and of the sea-urchin were procured in considerable numbers towards the end of the month — sometimes sooner. On one of the small jelly-fishes, the larval stage of an anemone (Peachia), frequenting the sand, was often found attached to the disc by the widely open mouth and tentacles. The young anemones are thus carried about — without effort on their part — by the beautifully transparent and festooned coach. But while this is so, they are also placed within the reach of the active young gadoids and flat-fishes — both of which probably diminish their numbers at this stage as well as subsequently when they are settling in the sand. No food is more tempting. The number and variety of these little jelly-fishes were remarkable, indeed on many occasions they thronged the water. Moreover, the enormous quantity of eggs given off by them must largely increase the food of other invertebrates and even of the larval and early post-larval fishes. A consideration of their life-history, further, is a curious comment on the views of those who imagine that the bottom of the sea can to a serious extent be rendered barren by the use of the trawl. While the Hydromedusa? were thus generally plentiful, the true jelly- fishes (e.g. Aurelia and Cyanea\ on the other hand, were vari- able in their appearance, sometimes abounding in June while in cold seasons they were few. Infusoria (Tintinnus, Ceratium and Peridiniuni) were especially numerous in the bottom-water, and though not previously mentioned, occurred every month of the year. The phosphorescence of Ceratium caused the interior of the fine tow-net when suddenly jerked at night to gleam like a fiery funnel. Amongst algoid forms Rhizosoleniae were common at the beginning of the month, but gradually diminished. Diatoms, the gelatinous algoid masses, and spores frequently occurred. July. The pelagic eggs of July were not numerous, though the variety was considerable, viz. gurnard, sprat, rockling, dragonet, weever, and the eggs of the frog- fish in long gelatinous ribands, besides eggs of the dab, topknot, lemon-dab, turbot, sole and little sole. Amongst the young fishes M. F. 4 50 PELAGIC FAUNA. were gadoids, clupeoids (15 — 17 mm.), sand-eels (7 — 12 mm.), rocklings, gunnels, gobies, lumpsuckers, dragonets, pipe-fishes, brill (20 mm.), topknots and the flat fishes. Moreover, abundant food for the adult fishes was present in the shape of larger forms, such as sprats of 2^ inches, gurnards 3 to 4 inches, and whitings of a smaller size. Throughout the month, Appendicularians were seldom absent from the mid-water and bottom-nets, but if the weather was cool they did not appear on the surface till the latter half. As a rule — whether numerous or few — they were small. Young mussels and other bivalves, with the little pteropod, Spirialis, had now reached the surface, but they were most plentiful near the bottom. The larval forms of Polyzoa were abundant, e.g. Cyphonautes, which agreed with the able description of Prof. Allman, who found it in the Moray Frith. Schneider, again, traced it to the adult condition (Membranipora). The wonderful larval form of Phoronis, viz. Actinotrocha, was common this month. As in June, crustacean life swarmed from the surface to the bottom, though now and again the one or the other species would be more abundantly captured. On certain grounds some species (e.g. Boreophausid) rendered the nets semi-solid, and attracted many fishes. Moreover the pinkish oil of these crustaceans would certainly suffice to colour the muscles of the fishes which devoured them — if coloured they could be by such food1. A great number of pelagic annelids in their larval stages were present, most of the littoral forms being represented. After this nomad existence they settle down in chinks of rocks and under stones, in sand or in mud, or as borers in stone. The wealth of life, in this group alone, was great. Besides the annelids, Sagittce occasionally appeared in masses. ' Painter's easel ' larvas of Echinoderms (sand-stars etc.) thronged the water, along with larval forms of the common cross-fishes, of the holothurians and Synaptce (Bipinnarians, Brachiolarians and Auriculariaus). The various kinds of jelly-fishes were in great profusion and 1 See Marion's Pelagic Faima. PELAGIC FAUNA. 51 many of them ripe. So plentiful were many of the smaller Hydromedusae that occasionally they were beached on the sands. Cyanea and other types reached the surface. The Infusorians were especially numerous at the surface on fine days, and during the still warm evenings, forming a thick phosphorescent coating (Ceratium) to the tow-nets, and spark- ling with every stroke of the oars. It is well known that even these (e.g. Peridinium) are found in masses in the intestine of the sardine. The algoid contents of the nets sensibly increased in July, probably from the profusion of spores and rapid general growth, and they were abundant offshore as well as inshore. Altogether the enormous variety in pelagic life — both plant and animal — constituted a conspicuous feature in July. August. During August the number of floating eggs had considerably diminished, those of the gurnard, rockling and sprat being most conspicuous — along with a few of the sole, turbot, and lemon-dab. The midwater-net inshore also showed a diminution of the post-larval round fishes, though taking the offshore and inshore together, a considerable number of small forms were still present. Two sizes of clupeoids, viz. 5 '5 to 8 mm. and 16 to 18 mm. were yet obtained. There were also families of gurnards ranging from 5 to 10 mm., sand-eels at 9 mm., rocklings in great variety at the surface from 4 mm. to silvery mackerel-midges of an inch or more in length : besides young whitings at li in., Montagu's suckers, gunnels, skulpins, lumpsuckers from 6 mm., gobies, pipe-fishes and pleuronectids (flounders, etc.) from 7 to 11 '5mm. Perhaps the most frequent round fishes in the nets were gobies. Small appendicularians were not unfrequent during the whole month; while the larval form of Phoronis (Actinotrocha) was a special feature, the transformation of the one into the other being followed with ease. Swarms of young mussels and other bivalves were present in both surface- and bottom-nets ; while Spirialis, Natica and others in the veliger-stage represented the univalves. Occa- sionally a young cuttlefish was likewise captured. Molluscan life indeed was plentiful. 4—2 52 PELAGIC FAUNA. Polyzoa were still represented by various larval forms. No diminution in the wealth of crustacean life took place in August, but it rather seemed more plentiful. While many of the higher kinds were still in their larval forms, such as porcelain crabs, others were in the large-eyed stage or beyond it. A marked feature off the Forth was the abundance of the post- larval stages of the Norway Lobster, which, mingled with the long trailing tentacles of jelly-fishes (Gyanea), formed an inextricable rope or chain. The larval stages (Nauplii) of the lower crustaceans likewise were still common. Thus the waters abounded in this rich food, the swarms of the post-larval stages being especially suitable for the growing fishes. From the beginning to the end of the month the various remarkable larvre of Echinoderrns (starfishes, sea- and heart- urchins) in various stages thronged the water — both surface and bottom ; moreover, many minute sea-urchins, heart-urchins, sand-stars and brittle-stars showed that considerable advance- ment had been made in growth. The long spines of some of these larvas may be protective, but — notwithstanding — it is probable that they are largely diminished by the young food- fishes. The jelly-fishes were in considerable profusion during the month, and their distribution seemed to be more general throughout the water. Occasionally it happened that a square mile of sea was densely covered with the common jelly-fish with the lilac bands (Aurelia). The uncertainties in connection with pelagic life were seen in the case of Beroe, which was plentiful the first week and afterwards disappeared. The Infusorians were abundant, especially at the bottom, and those at the surface were occasionally accompanied by Radiolarians. The minute alga?, such as Diatoms, spores, Rhizosolenias and others, were as plentiful as in July, and now and then were stranded as a greenish scum on the sands. September. The study of the pelagic eggs of fishes in September presented a contrast to that in the preceding months, for they had disappeared. A great diminution like- wise had taken place in the post-larval food-fishes. Young PELAGIC FAUNA. 53 gurnards, however, were not uncommon, varying in size from 5 to 12'5 mm. — thus showing an increase on the previous month. Moreover young clupeoids — from 7 to 16 mm. — of the autumn brood had made their appearance, the larger forms only after the middle of the month. A single gadoid of 5'5 mm., rocklings from 6 to 24 mm., larval sand-eels and pleuronectids from 5 to 11 mm. completed the list of the food- fishes. Gobies were still frequent, the extremes in length being 6 and 21 mm., while the majority were about 7 mm. Skulpins from 3'5 to 10 mm. were often met with, the former appearing at the beginning of the month, and having the tapering larval tail with permanent rays below. Young sea-scorpions (Cotti) of 9 mm., bimaculated suckers of 7'5 mm., young whiting from 1 to 3 inches, young gadoids of similar size, and young pipe- fishes were likewise obtained. Appeiidicularians were fairly numerous throughout the mouth, as also was Actinotrocha, with young Phorotiis. Young mussels were still plentiful, though not in profusion, and they were accompanied by other bivalves and many univalves. The larval Polyzoan, Cyphonautes, was still common. The crustaceans were apparently as numerous and varied as in August. Larval stages also (Nauplii and Zoew) were prevalent. They formed an almost inexhaustible store of nourishment for the smaller fishes. The bottom-nets especially abounded in the larval and post- larval annelids, their profusion in all the nets being remarkable. Young examples of most of the littoral annelids also made their appearance. Sagittoe from 12 to 16 mm. were frequent. The larval Echinoderms were less numerous than in August, but yet they occurred in almost every haul of the surface- and bottom-nets — together with a considerable number of minute star-fishes. In the rnid- water-nets jelly-fishes (Hydromedusoe) were frequent at the beginning of the month, but were rare in the surface-net. The ranks diminished as the month advanced, probably by the death of the jelly-fishes after the discharge of the eggs. 54 PELAGIC FAUNA. Infusorians in considerable numbers occurred in almost every net, especially in mid-water. Diatoms and other algoids were as abundant as in August. October. No pelagic eggs of fishes were found in October, and the pelagic post-larval and very young fishes were few in number. They consisted of clupeoids of the autumn series, from 17 to 18 mm., with permanent rays appearing in the dorsal fin and in the tail ; rocklings ; an occasional gurnard of 16 mm., various gobies and a stray pipe-fish or two. Young appeudicularians were met with several times. The pelagic shell-fishes were still procured in considerable numbers, the majority of the bivalves being apparently young mussels, though other forms were likewise present. Larval univalves were frequent. Cyphonautes (the larval stage of an encrusting Polyzoan) continued throughout the mouth. The bottom-nets teemed with minute crustaceans from the beginning to the end of the month. Larval types (Nauplii and the Cypris-stage) were plentiful. The paucity of the larval phases of the higher crustaceans was a marked feature, but very young forms were often met with. The larval annelids showed little diminution in their numbers in the bottom-net, and their variety was great — showing how ample the food-supply of the smaller fishes is from this source. Arrow-worms occurred sparingly at first, but were numerous at the end of the month. The diminution of the larval Echinoderms was most marked. Only a few ' painter's easel ' larvae (Plutei) appeared at the surface about the middle of the month. The small jelly-fishes (Hydromedusae) were still occasionally numerous, though the species were limited. They were, however, comparatively rare in the surface-nets. The cteno- phores were represented by myriads of Cydippe, while Beroe was generally present in most hauls of the nets. Infusorians occurred in considerable numbers, and one form (Ceratium tripos) was sometimes in vast profusion. The pelagic diatoms and algse were in great variety, and quite as numerous in the bottom- and surface-nets as formerly. PELAGIC FAUNA. 55 November. The post-larval fishes had almost but not quite disappeared in November. Clupeoids of 14'5 to 16'5, bima- culated suckers, an occasional pleuronectid with the eye on the ridge (11 '5 mm.), and gobies from 18 to 23 mm. were obtained. The contrast to the preceding months was thus marked. The shell- fishes were represented by a few bivalves — apparently mussels, and a few minute univalves — both groups occurring only in the bottom tow-nets. Cyphonautes was the only larval polyzoan. Minute crustaceans (Copepods) were numerous, with occa- sional swarms of the large-eyed stages (Megalops) of crabs. Amongst annelids the larval forms of certain kinds (Nerine) and small examples of others (Tomopteris) occurred. Arrow- worms on the other hand often were in profusion and of large size. They took the place of the crowds of small jelly-fishes of the earlier months. The appearance of a mitre-shaped jelly-fish (Aglantha) indicated the presence of the fauna of winter, the others being Tima and the ctenophores — Cydippe, Lesueuria and Beroe in considerable numbers. The majority were beneath the surface. The Infusorians and alga? were as numerous as in the previous months, arid the former occasionally appeared in the surface-net. December. In December the post-larval fishes appeared to be extremely rare, though the older stages of various forms were met with. Larval univalves in limited numbers represented the shell- fishes. The larval Polyzoan, Cyphonautes, was still common. Crustaceans were often in multitudes, the sessile-eyed forms also being present (e.g. Parathemisto), and a few larval forms (Nauplii) were still found. There was no lack of nourishment for fishes in this group. The larval annelids were inconspicuous, but a few adults still occurred. Arrow-worms, on the other hand, were remark- ably prevalent, and of large size. They formed a considerable element of the food of fishes at this season. 56 PELAGIC FAUNA. Contrary to what might have been anticipated one or two ' painter's easel ' larvae (Plutei) of Echinoderms still appeared. The same jelly-fishes mentioned in November were present, with the addition of one or two others. The numbers of the Ctenophores were frequently large. Infusorians of the same species as formerly were often in considerable profusion ; while plant-life — such as Diatoms, spores of alga? and minute algae — was plentiful throughout the month. In connection with a general collection of pelagic larval and post-larval fishes — chiefly the latter — it is interesting to note that if they are arranged according to the months a spindle is formed, with the thick central mass in May, April being next, and followed by a nearly equal series in June, July and August, while the tapering at each end, viz. the beginning and end of the year, is marked. When the same collection is grouped according to species, the first place is held by the pleuronectids, then follow in order the gobies, gadoids, sand- eels and cottoids, after which come Montagu's suckers, the clupeoids, dragonets, armed bull-heads, rocklings and gunnels. No particular weight need be put on this remark, but it gives an idea of the comparative abundance of young fishes usually captured in the tow-nets. FIG. 11. Jelly-fish, with disc everted, partly engulphiiig a post-larval flounder. A. T. M. CHAPTER IV. LIFE-HISTORY AND DEVELOPMENT OF A FISH FROM A PELAGIC EGG. THE perusal of the foregoing chapter on the pelagic animals will demonstrate the inexhaustible nature of the food provided for fishes at every season of the year, and, further, will indicate how little reliance is to be placed on the opinions of those who imagine that man can to any serious extent, by any method of fishing at present practised, denude the bottom of the ocean of its inhabitants. It has already been shown by one of us, that even though it were granted that a particular area of the sea were by the agency of man rendered " barren " the next tide would bring in a sufficient stock of temporarily pelagic forms to settle on the bottom and re-people the hypothetical waste, and that with rapidity. Under this rich food, the young fishes grow apace — head and eyes, mouth and accessory organs, body and fins — all rapidly increase, and the little fish, hatched in the spring, say from March to May, is soon in what is known as the post-larval stage, that is, has a mouth, has lost its yolk-sac, has assumed a more or less uniform tint, and has gill-fringes and teeth. It is about a quarter of an inch long, and is both active and in- telligent, the large head and large eyes of the young food-fishes being at this stage specially conspicuous, and in marked contrast with such as the sea-scorpion (Coitus). The marginal fin is continuous at a quarter of an inch, and the lancet-like termina- tion of the caudal end of the body is noteworthy. 58 LIFE-HISTORY AND DEVELOPMENT OF About this time the ventral fins of the pelagic fishes first make their appearance, for hitherto they have managed to do without them. Moreover, these fins in some, such as the rockling and ling, undergo remarkable development, forming in the latter (Fig. 6) a pair of great ventral wings conspicuously coloured yellow ; yet in the adult (a ground-fish) they attain no greater dimensions than in the cod, both having at a certain stage soft, free filaments or tactile processes at the tip. The ventral fins in the post-larval rockling are equally large, the distal half being black, so that at first sight the little fish when captured seems to possess a great ventral spine on each side FIG. 12. Post-larval liockling, enlarged. (Fig. 12). In the post-larval gurnard again, the huge breast- fins form a drapery for the entire body when folded back, only the tip of the tail extending beyond them (Fig. 13). They are indeed proportionally as large as in the southern flying gurnards, but in these the fins reach full development only in adult life, while in the young stages they are comparatively small — exactly the reverse happening in the grey gurnard of FIG. 13. Post-larval Gurnard, enlarged. our seas. The presence of the broad arches of pigment on the breast-fins of several forms, such as the present species, the green cod, and the armed bullhead, is also an interesting feature. We have not yet read the riddle of all these changes, but in the ling the great ventral fins are probably connected with its roaming or pelagic life, and this explanation would also suit in the cases of the rockling and the armed bullhead, both, in their mature state, seeking their food on the ground. A FISH FROM A PELAGIC EGG. 59 The little fishes at this stage are still more or less trans- lucent, except in the region of the eyes, which are silvery, and on the parts where the pigment occurs. Moreover, their fond- ness for a minute reddish Copepod (Calanus finmarchicus), which occurs in myriads around them, gives the region of the stomach a faint pinkish hue from the translucency of the tissues. Soon, however, pigment appears — at first chiefly along the dorsal and ventral margins of the body — and, by- and-by, foreshadowing in the cod those peculiar squares which give the sides, at a somewhat later stage, their tessellated or tartan-like aspect. Besides, they are found nearer the bottom of the water, so that they can be captured in a naturalist's trawl with a fine gauze-bag at the end. There is, therefore, a downward tendency as the little fishes get older and stronger, and thus in many cases a parallelism exists between them and the minute forms on which they prey, for on deposition the eggs rise towards the surface, where the helpless larvae (or newly hatched young fishes) also often occur, and then they seek the lower regions of the water as their size increases. There is much that is wonderful in such a life-history, especially in the metamorphoses or changes of form undergone by many of our best fishes such as the flat fishes (Pleuronectidre), which come out of the egg just like a haddock or a cod, with an eye on each side, as in Fig. 1, yet in after life have both eyes on the same side. ^IG 14 Young Lemon.Dab Nothing like this occurs in any of in the third stage, enlarged, the higher vertebrates. Gradually during growth the body of the fish increases in depth (Fig. 14), then the right or left eye passes over (Fig. 15) the ridge of the head to the opposite side, while the creature, hitherto pelagic, sinks deeper in the water and exhibits a tendency to lie on the side from which the eye has passed, and which gradually loses its dark pigment so as to become white1. It finally reaches the bottom, taking up its residence amongst the sand or sandy mud, and lying with the two eyes and the coloured side up, the 1 The tardy disappearance of the pigment in some forms is interesting. 60 LIFE-HISTORY AND DEVELOPMENT OF white underneath. At the same time they push inshore, and multitudes are found on muddy flats between tide-marks, and in shallow rock-pools, their bodies being immersed in mud, but the two active eyes raised above it. When disturbed, a little streak of muddy water alone tells their course. The mode by which the eye travels round has been a fruitful source of discussion with scientific men, and amongst these the names of Steenstrup, Malm, Schiodte, and Alex. Agassiz abroad, FIG. 15. Young Lemon-Dab at a later stage, the left eye just appearing on the ridge of the head : enlarged. with Wyville Thomson and Traquair in our own country, are well known. The fact is — two methods exist in nature ; in the one the eye travels over the ridge of the head, as just described in the flounder; in the other it traverses the soft and yielding tissues of the young fish, and so gains the other side. In Plagusia, the species in which the latter remarkable change occurs in the post-larval stage, the general tissues are so transparent that the creature in a glass vessel can only be noticed by the two apparently disembodied eyes, or by the gleam of light caused by its movements ; and before the change ensues in its eyes it can look obliquely through its own body and see what passes on the other side1. Up to this stage in the life-history of both round and flat fishes it will have been apparent that the efforts of man can have little effect on the vast multitudes of the eggs and minute fishes. His trawl sweeps beneath them, or they are carried harmlessly through its meshes. Not even in a trawl blocked by a fish-basket and several large skate are any likely to 1 Alex. As;assiz, Proceed. Americ. Acad. Arts and 8c. vol. xiv. p. 8, 1878. A FISH FROM A PELAGIC EGG. 61 occur. No example indeed was procured in the trawling expeditions for the Commission under Lord Dalhousie. The hooks of the liners are too large for the mouth at this stage, and hence they escape capture. Their small size and translncency also seem to afford protection in the case of predatory fishes of their own or other kinds, for they are rare, so far as present observation goes, in the stomach of any fish. Their great numbers are doubtless kept in check by some means, and we know that even jelly-fishes (e.g. Pleurobrachice, Fig. 11) and sparrows are very fond of post-larval fishes. It is only when they become somewhat larger that they are preyed on by their own and other species, and are swept up in thousands by the destructive shrimp-nets on our sandy shores. While the little food-fishes are assuming the change of hue indicated in the preceding pages, they in many cases seek the inshore waters ; at least, systematic use of the mid-water and other nets prove that at certain seasons they are met with in large numbers at the entrance to bays or off-shore, and that a little later, in the case of the cod — from the first of June onwards, they are visible from the rocky margins. The colora- tion in this species (cod) is now beautifully tessellated (Fig. 5), and they swim in groups, often in company with the young green cod, at the margin of the rocks at low water, and in the little tidal bays connected with rock-pools. The latter are often richly clothed with tangles, bladder-weed, red and green seaweeds, and the green Ulva — amidst the mazes of which the young fishes find both food and shelter, capturing the little crustaceans (Copepods, Ostracods, and others) swimming there, and snatching the young mussels and minute univalve mollusks from the blades of the seaweeds. To the zoologist few sights are more interesting than to watch the little cod in these fairy lakes, as they swim in shoals against the current, balancing themselves gracefully in the various eddies by aid of their breast-fins. In a mixed company, the young cod are easily recognised by their coloration, and the reddish hue of the occiput, for the blood-vessels there shine through the tissues, which generally are more translucent than in the green cod. Prof. G. O, Sars considered that about this stage there was 62 LIFE-HISTORY AND DEVELOPMENT OF an intimate connection between them and the hordes of jelly- fishes (Aurelia and Cyanea) which abound in the inshore waters towards the end of summer. He thought the young cod approached the jelly-fish for the sake of the minute pelagic animals stupefied by its poisonous threads, and that the fish repaid this favour by picking off a parasitic crustacean (Hyperia medusarum) which clings to the jelly-fish. Observations, con- tinued for a long period in this country, however, show that this connection is only casual and of very little importance, and that certain Hyperice are occasionally found in vast numbers in a free condition. As the season advances, the young cod are joined off the rocky ledges by a few pollack and whiting, but not by the haddock, which has certain social views of its own — keeping to the deep water farther out. The size of these cod late in autumn, as in October, varies, some reaching from 4 to 5 inches in length. Their food ranges from zoophytes to crustaceans, mollusks, and small fishes, and in confinement the larger are voracious, an example in the laboratory about 5 inches readily attacking a smaller (3 inches) and swallowing it as far as possible, though for some time a considerable portion of the body and tail of the prey projected from the mouth. Moreover, the tessellated condition becomes less marked, and as they approach 8 inches in length a tendency in some to uniformity of tint is noticeable. Many of those, however, that continue to haunt the rocky shores and the tangle-forests beyond low water still retain for some time mottled sides, and they are known by the name of rock-cod. Further, while their growth in the earlier stages is less marked, it is now very rapid — even in confinement. The exact rate of growth in the free condition in the sea is difficult to estimate, but the little cod of an inch and a half to an inch and three-quarters in June reach lengths varying from 3 to 5 inches in autumn, and in the tanks of the laboratory, specimens 5 inches in August attain 8 inches the following March. At Arendal, in Norway, where opportunities for watching the growth of cod in confinement have been supplied with a liberality not excelled in our country, Capt. Dannevig found that the cod of 3 mm. in April reached only A FISH FROM A PELAGIC EGG. 63 15 mm. in June, a length somewhat at variance with the condition, as above stated, on our shores. In July they mea- sured 2 inches, in September 3 inches and a half, and in October about 44 inches. The second year they attained 14 to 16 inches in length. In artificial circumstances, as well as in nature, it is found that great variation exists in the sizes of the young fishes of the same age, and this variation would not seem to be related to temperature. At the stages just mentioned they now come under the notice of both liner and trawler, for young cod 5 or 6 inches in length occasionally take a haddock-hook, and those somewhat larger (9 to 18 inches) occur in certain hauls of the trawl, especially off a rocky coast like that of Aberdeenshire, south of Girdleness, as well as on the hooks of the liners on rough ground. Special trips, indeed, were, and perhaps are, made by the liners for the capture of these young cod (termed codling), and thus their numbers are kept in check. So far as present observations go, therefore, the young cod in a free condition reach the length of from 4 to 10 inches the first year, while in the second they attain from 10 to 20 inches or more. It probably takes 3 or 4 years (and this is the original opinion of Sars) or more, to reach full maturity and a length of 3 feet or upwards ; though he mentions having seen young cod a foot in length, with mature roe and milt in the fish-market of Christiania. These, however, were probably examples of the small race of cod characteristic of the fjords of Norway. The young round fishes, such as cod, haddock, and whiting, of similar or nearly similar size, seem respectively to herd together. Thus it happens that in certain hauls of both liners and trawlers the majority agree in size. This is well known to the liners, who in former days specially sought out the young cod as already indicated. The same feature is observed in many other fishes, and probably conduces to their safety. So far as known, the adult fishes of the three kinds specially alluded to in the preceding paragraph (viz., cod, haddock, and whiting) follow no very definite law in regard to migrations, if we except the apparent congregation in 64 LIFE-HISTORY AND DEVELOPMENT OF certain regions during the spawning season, as pointed out for instance by Sars, off Lofoten, where they occur in vast numbers from January to March. In our own country, again, the appearance of shoals of haddocks and whiting in certain localities is another example. How far such multitudes, however, are influenced by the abundance of food is still an open question. In British seas the herring is the main cause of these congregations of cod and haddock ; the former chiefly pursuing the fishes, the latter their eggs. In the same way, the abundance of Norway lobsters and similar food on the grounds called banks exercises considerable influence on the presence of cod. It has already been pointed out, however, that in their young stages certain migrations do occur. Thus the post-larval cod by-and-by seeks the laminarian region, while the older forms for the most part tend to go seaward. The haddock keeps to the deeper water in its post-larval and very young state. The same occurs even in a more pronounced manner with the ling, the adults of which as a rule are found in deep water. The pelagic post-larval ling seeks downwards as it grows, and is seldom found near the shore till it attains the length of six or seven inches, in short, until it is barred with pigment. As it increases in size it migrates seaward. Similar features are noticed in the plaice. As observed in the trawling expeditions of 1884, only large plaice as a rule are procured in deep water off the east coast, while the sandy bays abound with those ranging from 11 inches downward, and none of the females of which appear to be mature. Multitudes of little plaice haunt the margins of these sandy beaches, but it cannot be said that forms which have the length, for example, of 3 inches, are confined to any particular line drawn across a bay, for small forms (2-4 in.) occur in hauls all over such a bay as that of St Andrews. Small turbot and halibut in the same way are often found in the shallow bays, while the large adults are inhabitants of the deeper water. Such would not, however, seem to be the case with certain skate, very large adults of which occur in the shallow water of the sandy voes in Shetland. A FISH FROM A PELAGIC EGG. 65 On the other hand, the witch (Pleuronectes cy no gloss -us} keeps to its special areas, both as regards the young and the adult condition, so that the movements of eggs, larval and post- larval forms are circumscribed ; and the same, to some extent, would seem to be the case with the topknot (Zeugo- pterus) and sail-fluke (Arnoglossus). The dab (Pleuronectes limanda), again, is found in all stages both in comparatively deep and in comparatively shallow water. Almost all our valuable food-fishes, therefore, are produced from minute pelagic eggs, the enormous numbers of which provide for a vast increase and wide distribution of the species ; yet it cannot be said that this habit alone provides for their multiplication when the case of the herring with its demersal eggs, fixed firmly to the bottom, is considered. It has to be borne in mind, however, that the larval herring mounts upward toward the surface as soon as its strength suffices. Many striking changes occur during growth, both in external form and coloration, but it is difficult at present to lay down any general law that would apply to all cases, though those in which certain migrations take place during growth show such changes very prominently. The young round fishes by-and-by roam about the sea in shoals, led hither and thither mainly by the presence of food; yet in the case of the larger and adult forms, safety or freedom from molestation may have some influence. Though so minute on escaping from the egg, their growth is, by-and-by, rapid, and the duration of life in such as the cod is considerable. Abundance of food, more than any special instinct, would appear to be the main cause of their migrations in the adult or adolescent state, and that food is as varied as their haunts ; in short it embraces every sub- kingdom up to their own, for fishes and their eggs form a large share of their diet. There would be little difficulty in adding to the sea great numbers of larval forms of any species of which eggs can be procured : yet if a few adults can be obtained in such waters at the proper season it is still an open question whether the natural process with its surroundings would not be more successful. M. F. 5 66 LIFE-HISTORY OF A FISH FROM A PELAGIC EGG. In the foregoing remarks a few of the leading features of the life-history of a food-fish have been but touched on, for the subject is one of vast extent, and some of the points embraced in it are by no means easily solved. We have only earnestly entered on the study of the subject in this country within the last few years, and much yet remains to be done, even in some of the most common marine fishes. However, the zoological investigator is here stimulated by the fact that all his labours directly bear on the public welfare, for it need hardly be pointed out that a thorough knowledge of the development and life- histories of our food-fishes is the first step to sound legislation and effective administration. The State has in past years spent princely sums on more or less pure science, as in the memorable voyage of the Challenger. There can be no doubt, notwith- standing the recent opposition, that at the present moment the public interests demand a searching and long-continued inquiry nearer home, viz. the exhaustive investigation of all that pertains to the food-fishes of our shores, since the problems connected therewith affect the prosperity of so large a portion of the population. The difference between the larval cod and the young salmon just hatched is striking. The former, that is, the young cod, is in a very rudimentary condition, not only in size but in structure. For instance, it is devoid of mouth and vent, and though the heart pulsates, it, as our colleague, Prof. Pettigrew, observed, is devoid of blood, and there are no blood-vessels. Those, therefore, who thought that the heart in animals contracted from the stimulus of its living blood, would have here found little support. The tiny larval cod of about 4 mm. is so fragile that it can be handled with difficulty, and it would seem as if a breath would destroy it. On the other hand the newly-hatched salmon has attained great complexity, and is about three-fourths of an inch in length, while several days may be spent in delineating its elaborately branched blood-vessels — through which the blood-discs may be followed as they swiftly circulate in the transparent tissues. CHAPTER V. GENERAL SKETCH OF MARINE TELEOSTEAN DEVELOPMENT. IN a work like the present it is unnecessary even were it practicable to give in detail all the authors who have laboured in the field, or to enter into narrative in regard to the type taken to represent the development of the food-fishes. Lists of those who have made advances in the development of these fishes will be found in the various treatises and mono- graphs on the subject. In the list appended to the ResearcJies, by one of us and Professor Prince, no less than 160 authors of importance are cited in 1889, and the list has considerably increased since. It will suffice if a brief and somewhat popular resume be given, firstly, of the changes which are externally visible in the egg up to the period of hatching, and secondly, of some of the leading structural changes at various stages, mainly followed out by the systematic study of sections. For illustrating the changes undergone by such an egg floating about in the sea, we shall take the translucent pelagic egg of the whiting (Fig. 16) in the month of April, and it may at once be stated that it forms a very good type for almost all the bony fishes. As discharged by the adult on the spawning-grounds the translucent sphere consists externally of a hyaline capsule, which under a high power is observed to be minutely punctured, and under certain conditions has a tendency to break into film-like scales of membrane. It really consists, however, of a single layer and is not compound as some authors have supposed. At one part of the capsule a minute indentation (micro- pyle) is seen leading to the interior, and this aperture has 5—2 68 GENERAL SKETCH OF various shapes in different fishes (Fig. 19). Fluid is readily transmitted through this entrance, and also by the pores — as may be seen by partially drying the egg, and then immersing it in sea water — a feature connected with the aeration of the contents. Within the capsule we have the globe of more or less fluid yolk or nutritive material — surrounded by a delicate layer of protoplasm, between which and the capsule is a space termed the perivitelline space. The entire egg of the cod or whiting (Fig. 16) is a transparent spherical floating mass FIG. 16. Egg of Whiting, with cap of protoplasm. E. E. P. delicately sensitive to every movement in the water, even a breath in a still vessel setting the almost invisible spheres in commotion; so that under the microscope even in a quiet room, and still more in a wooden building, we may understand it is not always easy to keep them in the field of vision. In the spawning-areas, at the period of the shedding of these eggs, the water swarms with the minute sperms of the male, one of which, in the course of events, finds entrance by the aperture (micropyle Fig. 19 a) in the wall of the egg, and then begin those changes which mark fertilisation, such as the fusion of the male with the female pronucleus, the protrusion of the polar globules and the formation of the first segmentation- nucleus. Shortly after fertilisation the egg in many cases becomes more tense and the space within the capsule, the "breathing chamber" of Needham, more evident. In the cod and others, moreover, a streaming movement of the jelly-like protoplasmic covering of the yolk downwards, MARINE TELEOSTEAN DEVELOPMENT. 69 towards the animal pole, takes place. At this pole, always on the lower side of the floating egg, the protoplasm collects, forming a faintly straw-coloured cap in which granules are scattered (Fig. 16). At this part the first segmentation- nucleus may be seen. Soon after the appearance of this cap, e.g. in 40 minutes in the whiting, longer in the cod, an indentation begins in the middle which gradually deepens until the cap is bisected (Fig. 17). Thus, instead of one, two segments are now present. FIG. 17. Egg of Whiting. Commencing segmentation. E. E. P. In a similar manner the splitting proceeds in each of these, resulting in the formation of four segments, and these again are subdivided by bisecting lines, so that eight segments result. This process is continued with considerable, but by no means absolute, regularity. In the colder months during which these changes take place in the cod it is a much slower process than for instance in the whiting, which at a later season develops more rapidly. In other words, the rapidity of segmentation is directly dependent on the temperature. Thus the egg of a whiting, observed by Prof. Prince, which showed at 6.40 the first furrow, presented at 9 p.m. the accompanying changes (Fig. 18, a, b, c, d, e) — the last stage (e) having the specks or nuclei in the segments. Thus the whole disc, passing from the 8 to the 16, 32 and higher stages, at last forms a finely divided cap (often termed the blastoderm), which is now said to be in the many-celled stage ; this is reached on the second day. The whole process of segmentation thus consists of the dividing up of the protoplasmic 70 GENERAL SKETCH OF cap into a number of cells, each containing its nucleus and enveloped in its cell-membrane (Figs. 19 and 20, surface and 9. P.M. 9.10 9.30 9.40 2 c d, e FIG. 18. Egg of Whiting, a. View of segmenting blastoderm, 9 p.m. b. View of segmenting blastoderm, 9.10p.m. c. View of segmenting blasto- derm, 9.25 p.m. d. View of segmenting blastoderm, 9.30 p.m. e. View of segmenting blastoderm, 9.40 p.m. E. E. P. lateral views). Moreover, surrounding the disc, which in lateral view soon becomes lenticular or lens-shaped (Fig. 20), is a ring of small pinkish dots or nuclei, in the protoplasmic belt or FIG. 19. Egg of Cod. View of many-celled blastoderm from above. E. E. P. FIG. 20. The same. Egg of Cod. View of blastoderm from the side. E. E. P. periblast, forming the so-called nuclear zone. This is more clearly seen in the highly magnified portion of the edge of the disc in Fig. 20 a. The nuclei have a central speck or nucleolus. IQth April, 2nd day. Next day it was found that at 9.30 a.m. the disc of the egg presented little change, except that the nuclei in the zone surrounding it were more numerous. This nuclear zone became less distinct at noon and about 1 p.m. it had all but disappeared. MARINE TELEOSTEAN DEVELOPMENT. 71 and 18th April, 3rd and 4ith days. When segmen- tation is complete the blastoderm undergoes a change of a striking character. It becomes raised and separated from the yolk so that a chamber, not coincident with the centre of the disc, is formed between its under surface and the yolk (Figs. 21 and 22, surface and lateral views). This chamber, FIG. 20 a. Edge of blastoderm with nuclear zone (periblast). Highly magnified. E. E. P. FIG. 21. Egg of Whiting with germinal cavity. Dorsal view. E. E. P. usually termed the segmentation-cavity, has also been styled the 'germinal cavity' by Prof. Prince, to emphasise the fact that it differs in some respects from the segmentation-cavity of other types of vertebrate animals. The blastoderm at this stage shows no clear differentiation into layers, though the upper stratum is usually distinguished as a layer of ectoderm or epiblast-cells, and that below, forming the main mass, as the lower-layer cells. A new series of changes now ensues. Formerly, in the FIG. 22. Egg of Whiting with many-celled blastoderm and germinal cavity. Lateral view. E. E. P. 72 GENERAL SKETCH OF early ovum, the protoplasmic covering of the yolk tended to stream downwards and augment the disc, but now the blasto- derm commences to grow up and over and to envelop the yolk (epiboly, Fig. 22). With the commencement of this process the blastoderm flattens, and the vertical height of the germinal cavity is reduced to form a mere fissure, whilst the germ becomes generally thinner, yet along one radius there is a distinct cellular thickening, viz. the embryonic radius (Fig. 21), and this is noticed many hours before the inflection of the inner layer or hypoblast, which will be mentioned later. When the blastoderm covers about a quadrant of the globe of the yolk, the rim or border is visibly thickened, forming, as some authors name it, 'a true pad around the egg.' This curious condition is usually explained thus : — the amount of yolk prevents the formation of an ordinary ' embolic gastrula '- stage in this case, hence the yolk has to be enveloped by growth of the blastoderm over it. At this stage we may glance at the formation of the yolk (deutoplasm) in such eggs. As a rule in pelagic eggs, the yolk is a transparent homogeneous liquid through which a body like an oil-globule, such as occurs for instance in the gurnard, freely passes. In the earlier stages a different condition prevails : thus in the roe or ovary the developing yolk is quite opaque, composed of a number of spherules with intermediate proto- plasm, as seen for example in the early ovum of the turbot (Plate III., fig. 22). In such species as the cod, the yolk- mass clears up before ripening, and the protoplasm collects mostly on the surface, but that all the protoplasm is thus removed is not yet proved. For instance, in some cases collec- tions of protoplasm take place on the yolk after the closure of the blastopore. In some eggs, as in that of the sole, a vesicular condition of the surface of the yolk is caused by evident strands of protoplasm cutting off portions of yolk, and in the herring- tribe and mursenoids the whole yolk is thus divided. The transparent condition of the yolk is not found in the demersal eggs and it is probably to be regarded as a special protective adaptation to a pelagic habitat ; the pelagic fauna, consisting of examples from most of the leading groups of MARINE TELEOSTEAN DEVELOPMENT. 73 animals, are as a general rule characterised by transparency of the tissues. The yolk forms a store of nutriment which is little used in the earlier stages of development, though there is no reason to suppose that it is altogether inert, since the neighbourhood of the layer of protoplasm separating it from the blastoderm is a scene of considerable activity. In the later stages the rapid shrinking of the yolk before the embryo leaves the egg- shows how important it then is in affording nutriment for the development of the embryo. All the changes are not readily seen without making sections of the prepared eggs, especially as regards the origin of the second primary germ-layer (hypoblast). Its exact derivation has given rise to much speculation, but it would appear mainly to arise as an infolding and ingrowth of the outer- layer cells (epiblast), supplemented by periblast-cells. This folding is seen at a very early stage, and after the disc has flattened out it can be followed to the central region of the animal pole, a region which corresponds with the embryonic thickening formerly mentioned. The effect of this ingrowth of cells is that the originally very definite outline of the embryonic shield becomes irregular and finally disappears or passes im- perceptibly away on all sides, except posteriorly. A typical section of a teleostean egg (e.g. that of cod) when the yolk is about half-covered by the blastoderm shows a single layered corneous outer stratum and beneath it a thicker mass of cells, the former being the outer and the latter the inner layer of the epiblast. Another layer of cells, the hypoblast, borders the embryo towards the yolk. The third primary layer, or meso- blast, has yet to be described. It arises in great part from the lower-layer cells, that is, those seen very early under the epiblast, and probably also in great measure from the hypoblast, and the cells so derived soon become divided into longitu- dinal sheets or masses. These three primary layers are clearly distinguished in the early embryo, except at the extreme hind region, in which they are all confluent in a mass of in- different cells. At the extreme head-end, on the other hand, there is still only epiblast and hypoblast, as the middle layer does not extend into this region till later. 74 GENERAL SKETCH OF As the rim extends over the yolk the embryonic shield lengthens, the keel thickens, and soon a spathulate flatten- ing of the central or nervous layer is noticed anteriorly, that is, at the opposite end from the aperture of the blastoderm (the blastopore), the spathulate process trending posteriorly to the margin of the aperture. Then differentiation of the spa- thulate region (Fig. 23) occurs by the definition of the optic vesicles from the solid cellular margin (Fig. 24). Very soon after the formation of these organs traces of the ears appear FIG. 23. Egg of Whiting with spathulate neurochord. FIG. 24. Egg of Cod with embryo. Lateral view. more posteriorly, as elongated or elliptical thickenings of the sensory layer of the epiblast, and a large fissure or chink develops in the centre. To return to the growing blastoderm as it covers the yolk, it is found that the thickened margin or rim bounds an aperture called the blastopore by embryologists, this aperture attaining its maximum when the margin of the blastoderm has reached the equator, and thereafter decreasing in calibre as the rim advances towards the vegetal pole of the yolk (Fig. 23) ; the latter indeed in certain views bulges from the rim like an india-rubber ball pressed in the fingers. The embryonic rim is used up in the formation of the embryo. The aperture or blastopore then becomes contracted, marked with radial lines or furrows, and finally closes. In eggs with oil-globules it is at this time that the globule becomes surrounded by the protoplasmic periblast-layer which envelopes the yolk. Only in one instance, and that probably abnormal, has the globule been found freely moveable in the MARINE TELEOSTEAN DEVELOPMENT. 75 larval yolk-sac after hatching, viz. in the torsk or tusk1. More- over, black pigment at a somewhat later stage develops in the periblast bordering the yolk. We may further note that in demersal eggs with vitelline circulation, the germinal cavity disappears, after closure of the blastopore, partly by the pushing in of mesoblast. When the blastopore closes, or a few hours before, a vesicle (Kupffer's) appears on the ventral aspect of the embryo slightly anterior to the tip of the tail (Fig. 24). Occasionally other vesicles occur along the under surface of the embryo, in front of the foregoing. At the same time, a canal (neurenteric canal) is seen passing, from the transient homologue of the medullary groove of the dorsum, to the under surface of the embryo. In most other vertebrates, the groove in the dorsum, called the medullary groove, which closes later to form the canal of the spinal cord, is a very early feature in development, whereas in the cod and other Teleosteans the solid mass forming the spinal cord has a fissure formed in its axis only at a later date. In the same way the neurenteric canal and blastopore are much more con- spicuous features of the larval stages of other vertebrates. I9tk April, 5th day. With the closure of the blastopore, the definition of the embryonic fish as an elongated rod pressing into the surface of the yolk, becomes marked. The middle layer or mesoblast has by this time segmented into five or six transverse divisions called muscle-plates or protovertebne. The primary axis (notochord) can be traced some distance forward, but only as a broad translucent streak. This organ arises from the hypoblast as a median axial rod of cells growing upwards and separating the two rows of mesoblastic lateral protovertebrae. These increase in number rapidly, so that with the extension of the mesoblast forward, the trunk is wholly segmented, and they give the embryo a transversely banded appearance. On either side of the head an invagination of a small area of epiblast gives rise to the lens of the eye, and an expansion of the lateral walls of the embryo on each side forms what is called the alar membrane. 1 W. C. M. Wth Ann. S.B.F. Rep., p. 290. 76 GENERAL SKETCH OF The heart is formed of a solid and rounded projection of the middle layer bulging out towards the subjacent periblast. Wth April, 6th day. The brain is now differentiated into fore-brain, mid- and hind-brain (Figs. 25 and 26). The lens FIG. 25. Egg of Cod with embryo on 6th day. Lateral view. FIG. 26. Egg of Cod with more advanced embryo on 6th day. Ventral view. of each eye is complete, and otoliths have appeared in the otocysts. The heart assumes a conical tubular appearance, and the inner layer (hypoblast) has formed the gut, in which a cavity is now apparent — large in the median region, and tapering to the solid anterior part. The liver appears as a blind diverticulum of the gut, on the right side in most views. The otocysts are more ovoid in shape, and the notochord reaches forward to them. It has a downward flexure at its anterior end and shows signs of vacuolation. The third pair of sense organs, or nasal pits, is now indicated. Three diverging lines on the ventral surface of the brain indicate the lateral ventricles. The pectoral fins are outlined, and delicate protoplasmic FIG. 27. Egg of Cod with embryo on 9th day. Nearly dorsal view. FIG. 28. Egg of Cod with embryo on 9th day. Lateral view. MARINE TELEOSTEAN DEVELOPMENT. 77 expansions stretch out from the body in front of them on the yolk. 2lst April, 7th day. The body is now jerked from side to side by the contraction of the trunk-muscles, and has lengthened ; the caudal extremity is flexed (Figs. 27 and 28). The ear-capsules or otocysts, although they arose in a position behind the head, have now moved forward towards the eyes. The heart forms a long cone with the open end to the left, and debouching into the germinal cavity, while the apex passes to an expansion in the mid-region. The heart beats languidly and irregularly, the pulsations sometimes ceasing for 15 or 20 seconds. The gut is hollow as far forward as the heart. The primitive axis, or notochord, is now completely crossed by intermingling arcs. FIG. 2!). Egg of Cod with advanced embryo (9th day). FIG. 30. Egg of Cod with advanced embryo. 22nrf April, 8th day. The posterior regions of the trunk and tail are flexed, and the yolk has considerably decreased. The continuous median marginal fin is now in evidence, and the FIG. 31. Egg of Cod with more ad- vanced embryo. Lateral view. FIG. 32. Egg of Cod with more ad- vanced embryo. 78 GENERAL SKETCH OP pectoral fins are well defined, rounded anteriorly and pointed posteriorly. The liver (arising from the ventral region of the gut) forms a rounded process. The heart has a trumpet-shaped venous end and a boldly flexed arterial portion. The pulsations are now more regular — about 25 per minute. Rounded black chrornatophores, or pigment-cells, have appeared on the head and dorso-lateral region of the trunk, but they have no regular linear disposition (Figs. 2.9 — 33). FIG. 33. View of the head of the same embryo from the front. 23rc£ April, 9th day. By this time the eyes become pigmented with rounded black chrornatophores and the pig- ment-spots of the body and tail are more numerous. Three branchial clefts have made their appearance behind the ears. The olfactory bulbs are connected with the nasal pits. The liver has acquired an irregular lobulated appearance, and the gut is a prominent vermiform structure (Fig. 34). FIG. 34. Nearly lateral view of an embryo of the Cod on the 9th day. Almost ready to hatch. The violent wriggling of the embryos indicates their advancement, and a few issue from the eggs this day. The empty capsules retain their spherical shape though a lenticular rent passes two-thirds across their diameter (Plate TIL, fig. 2). April, 10th day. Five-sixths of the embryos are still MARINE TELEOSTEAN DEVELOPMENT. 79 in the eggs. They present a similar appearance to those of the previous day, though the increasing complexity of the gill- region is evident, and four clefts are visible, extending outwards to the otocysts. The heart is larger and the two chambers are more distinct. 25£/i April, Iltk day. Black pigment-spots are sparsely scattered over the dorsal and lateral regions and a few on the tail. The eye has a bright bronze hue. The remainder of the embryos emerged on this day. The subsequent changes are best described under the special head of each organ, so that we have now to retrace our steps and observe what is the history of the various organs we have mentioned. These organs may be classified according to their origin from each of the three primary layers of the germ, namely, the outer layer or epiblast, the middle layer or mesoblast, and the inner layer or hypoblast. They are most conveniently described under these heads. ORGANS DERIVED FROM THE EPIBLA.ST OR OUTER LAYER. Nervous system. We have seen that the outer layer or epiblast was esta- blished as a single layer of cells in the developing disc or blastoderm. These cells early become flattened and in section spindle-shaped. The outer or corneous layer at first covers the trunk of the embryo and the arc of the blastoderm beyond, but later, cellular thickening occurs beneath — forming the nervous layer of the epiblast. We have, indeed, seen that this thickening forms the keel (carina) that presses on the yolk. It becomes the rudiment of the spinal cord (neurochord). Its con- stituent cells are full and rounded but, as downward proliferation proceeds, those forming its lateral borders become columnar and unmistakeably mark off the neurochord from the adjacent cells, especially anteriorly. In the latter region they proliferate so rapidly, ventrally and laterally, that they come into contact with the inner layer or hypoblast and thus leave little room for mesoblast. Both layers of epiblast seem to extend over the 80 GENERAL SKETCH OF yolk-sac. It is below the second layer that the pigment- corpuscles occur. The enlarged anterior part of the neurochord forms the brain. It is rounded above, deeply carinate below, arched over by the epiblastic integument and limited ventrally by the hypoblast. From the anterior region of the brain, or fore-brain, grow out the paired optic vesicles, and the mid-brain is easily distinguished by its greater breadth and volume. Immediately posterior to it is the hind-brain which passes into the neuro- chord. It is remarkable that the brain forms nearly ^ of the total length of the embryo in its early condition. When a quarter of the yolk is enveloped, a cleft appears on each side of the post-optic region-. One anterior portion, the united mid- and fore-brain, can now be distinguished from the hind-brain, which is soon separated by a similar slight fold from the spinal cord behind. The mid-brain, lastly, is constricted off by an interorbital fold and the three regions become clearly defined. Instead of the medullary folds in the other vertebrates, which close together and form the spinal cord with its median cleft (neural canal), we have in these bony fishes a solid mass or rod of cells forming the neurochord, which later acquires a cavity by a separation of the median cells. In some Teleo- steans, for example the gurnard, a remnant of the medullary groove is seen in the deep temporary groove in the dorsal surface of the neurochord about the fourth day — whilst still in the egg. In the cod, soon after the closure of the blastopore, a fine cleft appears by separation of the median cells in the brain along a vertical longitudinal plane. It commences in the mid- brain and thence extends into the fore-brain almost to the anterior limit of the latter. This is the first indication of the true neural canal. Then two lateral continuations form a cruciform fissure, marking off the fore-brain ; whilst a T-shaped lumen is found in the mid-brain, the roof being thinner than the walls or floor. The canal rapidly extends into the neural cord in the trunk, giving off a pair of (vertical) lateral cavities, forming the optic ventricles and iter a tertio ad quartiim MARINE TELEOSTEAN DEVELOPMENT. 81 ventriculum, and a median pair constituting the fourth ventricle. Several days before escaping from the egg, great extension of the mid-brain occurs. On hatching, the mesencephalon (optic lobes) embraces the largest extent of the brain, and gives it the bulbous form so characteristic of young Teleosteans. The cerebellum is almost entirely covered by the posterior enlargements of the optic lobes, a thickened ridge only projecting. The fore-brain forms a narrow laterally compressed mass with a small dorsal cham- ber. On the second or third day after hatching, a deep fold divides this into anterior fore-brain (or cerebrum) and posterior fore-brain (or thalamencephalon), in addition to the longitudinal fold formerly present. Very early a posterior part of the floor of the thalamen- cephalon is directed backwards as a prolongation beneath the elevated medulla oblongata, and at the same time obliquely downwards. The cells surround a cavity continuous above with the third ventricle, or cavity of the thalamencephalon. This is known as the infundibulum, which abuts on the roof of the oral cavity. A loose mesh-work of cells lies behind the infundibulum and into this the front end of the notochord pushes. On the summit of the arch caused by the elevation of the oral roof a mass of cells appears, proliferating from the oral roof-cells. This is the pituitary body (hypophysis) and it lies in front of the infundibulum. A small median swelling, not unlike the hypophysis, lies in front of the other, i.e. behind the point where the optic nerves cross. Such appears to form the inferior lobes (hypoaria), so well developed in Percoids. The small extent of the roof of the thalamencephalon ex- posed becomes very thin and much folded. In sections there is seen a central aggregation of cells, which later push out as a papil- liform process, containing a cavity which is continuous with the third ventricle below. The cells become columnar and it forms the pineal 'gland.' It becomes truncated, more or less plicated and pressed against the arachnoid membrane. Later still its lumen is obliterated and it forms a deeply folded mass of cells. The spinal cord, when fairly advanced, proceeds quite to the M. F. G 82 GENERAL SKETCH OF termination of the notochord posteriorly. Usually the terminal part is attenuated, but occasionally it is slightly enlarged. Eyes. (Optic nerves and vesicles.) One of the earliest features in the embryo is the great development of the cephalic region, which is chiefly due to the protrusion of two rounded lateral masses (optic vesicles) from the sides of the narrow fore-brain. This cellular proliferation later shews a median slit, the only approach to the hollow brain-vesicle in the primitive forms. This slit passes along the stalk into the fore -brain ; and subsequently it is seen that the optic nerves are at a different level. Very soon the vesicles become cup-shaped by the pressure of an almond-shaped mass of thickened epiblast — the future lens, and the thinner margin creeps over the intruding mass of epiblast so as to form a circular lip around it, except at the lower side. The gap thus formed is called the choroid fissure and persists for some time. Mesoblastic cells, which were included as a thin plate between the optic nerve and the brain, have spread over the front as an outer layer (sclerotico-choroidal sheath) and pushed their way through the choroidal fissure into the internal chamber of the eye, and probably break down to form the vitreous humour and other structures. Just before hatching, pigment occurs outside the optic vesicle and in the external investment. Moreover, no less than six layers of the retina are present before extrusion occurs1. After hatching, the complexity of the eye is greatly in- creased, one of the most marked features being the develop- ment of well-defined rods and cones. The brilliantly opalescent iris not only adds to the remarkable appearance of the trans- parent larva but, with the pigment of the choroid, generally enables it to be discerned in the tanks. Auditory organs. The ears (otocysts) are formed, soon after the differentiation of the optic vesicles, as solid proliferations of the sensory layer 1 Those specially interested should refer to Dr Marcus Gunn's account of the development of the Teleostean eye, Ann. Nat. Hist. Sept. 1888, pp. 263—268. MARINE TKLEOSTEAX DEVELOPMENT. 83 of the epiblast, considerably behind the latter organs. Then, in the radial arrangement of cells, a cavity appears ; and this at first, in most bony fishes, has the form of an elongated slit, then it becomes broader and finally rounded. The walls are originally very dense, but become thinner subsequently. Otoliths occur about the sixth day, small at first, afterwards as two, occasionally three, distinct rounded calcareous masses. The otocysts change shape — becoming like an oyster-shell. Sensory cushions and semicircular canals are later developed, so that in the post-larval fish the auditory apparatus becomes complex. Olfactory pits and nerves. The olfactory pits are distinguishable on the sixth day or thereabout, as paired thickenings of the sensory epiblast in front of the upper part of the hemispheres. Each soon forms a flat oval sac of slightly elongated cells. The nerves are minute proliferations of the wall of the anterior fore-brain, which coalesce with the proximal surface of the nasal pit. Each pit has at first a single opening, but later a slight promontory appears in the middle of each lip and in a few days a junction occurs. The bridge becomes broad and each aperture is sur- rounded by an elevated rim. Cranial nerves. The embryos are unfavourable for tracing the development of the third, fourth and sixth, but the fifth (Trigeminal) is a large nerve which springs from the upper lateral margin of the hind-brain at a late embryonic stage. Just as it emerges, it separates into several branches, the maxillo-palatine and man- dibular, each having a large ganglion from which other twigs pass. Behind the foregoing nerve, the seventh and eighth arise in close proximity, the auditory being posterior and exhibiting a large ganglion. Fibres from the former can be traced to the base of the third ventricle or more correctly, above the pyramids, whilst the eighth or auditory consists of thin fibres 6—2 84 GENERAL SKETCH OF which emerge close to the surface of the medulla, just below the overlapping posterior part of the optic lobe. The vagus or tenth arises by two roots, the first, which probably includes the fibres of the ninth, issuing from a point near the lateral summit of the medulla oblongata. It per- forates the auditory cartilage and sends branches to the gill- arches and pharynx. The second part describes a curve and can be traced to the median region of the medulla below the floor of the fourth ventricle, and above the pyramids, part of its fibres having a more superficial origin. They form in front of the pectoral girdle a large double ganglion below and, to some extent, internal to the ear. Fibres pass, further, to the pharynx and branchial arches, and, from a smaller ganglion, pharyngeal and cardiac branches are given off. Lateral sense-organs. Generally one of these occurs on the top of the head just behind the eye, a second behind the pectoral fin, and one or two along the caudal region. Each consists of a somewhat elliptical aggregation of granular columnar cells from which a number of very fine palpocils project. A delicate nerve-filament passes from the apparatus to the muscular plates, and this shows a slight enlargement at its proximal end, and another as it approaches the sensory organ. In some instances, the latter is absent whilst the nerve is present. In the haddock, the facial region has numerous papillary sensory bodies, which have a similar structure, viz. lengthened spindle-shaped cells. In the gurnard also, on the snout, tubes with cushions of columnar epithelium and sensory hairs occur, and they communicate with the cushions. Skin. It has already been mentioned that in early embryonic life there is externally a flattened layer or corneous stratum of epiblast, and beneath it is the nervous layer. Soon after the notochord is defined, these layers extend as a distinct integu- ment all over the embryo and its yolk-sac. The inner stratum at a late stage consists of several layers of rounded cells, the MAR1NK TKLKOSTKAX DKYKU )I'M KXT. 85 innermost part probably constituting a Malpighian layer. A sub-epidermal space is often present, especially in those reared in confinement. No true skin at first occurs beneath the latter, but pigment-corpuscles appear in it, the hues in the different species being described elsewhere. Beneath the former, and constituting the true skin, the mesoblast extends later ; in this the scales are subsequently developed and burst through the epiblast. When pigment occurs over the yolk it develops in the periblastic covering. In the development of a pigment-spot, as a rule, a colourless corpuscle, often branched, precedes it. At a late stage iridescent plates occur in this layer. The Fins. Median Fins. The embryonic median fin arises as a minute fold of the outer layer (epiblast) of the embryo within a day or two of the closure of the blastopore. This becomes a broad membrane, which increases after hatching. Its thinness and transparency are remarkable. In section it consists of the two layers of the epiblast and a central fissure continuous with the sub-epidermic space, which is filled with a jelly-like lymph and gives passage to delicate nerve-strands to the sensory papillae of the skin. In this continuous fold, as F. M. Balfour said, by local hyper- trophy the permanent unpaired fins arise, but atrophy of the interlying membrane takes place during development. Certain parts also in early stages are characterised by remarkable increase of pigment. Shortly after hatching very fine fibrillar lines appear in this continuous fin, commencing generally in the tail. They form at first granular tracts indefinite in outline, and un- connected with the axial skeleton. They are usually termed, after Ryder, embryonic fin-rays. The mesoblast at a later stage extends between the layers of the fin, and thus it is this layer which gives rise to the true rays, which appear in the cod as three dorsal and two anal fins. Each ray at first forms a slight opacity extending towards the free margin, and the intermediate membrane disappears by absorption. The de- velopment of a mesoblastic granular thickening, a short distance 86 GENERAL SKETCH OF from the tip of the notochord of the larval tail, is the first sign of the permanent organ. The true rays form in the lower region of the tail, the whole generally pushing the notochord with the larval tail upwards, and giving it a characteristic bend or slope, the boundary between the permanent and the larval tail being marked by a notch. In some fishes, as in the cod and the haddock, though the ventral thickening is most distinct, the tip of the notochord remains median, and rays develop on both sides of it, making a feather-tip. Paired fins. When the embryo is first outlined, an alar expansion, con- sisting of epiblast and hypoblast, resting on periblast, stretches away on each side along the whole trunk. No mesoblast apparently extends into it. Soon two flattened oval pads, consisting of a double fold of epiblast, are differentiated from the rest of the expansion. Then the outer border of the cell- mass (mesoblast) near the Wolffian ducts sends a process between the layer and spreads out radially, but does not quite reach the distal margin. The fin gradually becomes discon- nected from the covering of the yolk-sac, the central meso- blast assumes a columnar character, and later towards the trunk a stout peduncle is formed, cartilage-cells are developed, radial structures appear and lastly pigment may be seen on its surface. The fin, moreover, leaving its primitive horizontal position, becomes more or less vertical. During the third week after hatching, the rotation of the fin has made it obliquely vertical, and then the basal attachment is placed almost dorso- ventral. Embryonic and permanent rays develop as in the dorsal fin. A pectoral bar appears on each side, the first part of the girdle, and then the various elements of the latter, as shown in a special research of Prof. Prince1, are outlined. The ventral fins of the lesser weever appear in the egg shortly after the pectoral, and from the same alar expansion, but as a rule, in the forms with pelagic eggs, the ventral fins 1 " On the Development and Morphology of Limbs of Teleosteaus." Glasgow, 1891. MARINE TELEOSTKAN DEVELOPMENT. 87 only appear in the later post-larval condition as shown under the various species. ORGANS FORMED FROM THE MESOBLAST. Heart and circulatory system. The heart in the Teleosteans here described develops at a very early stage — before the oesophagus is formed — as a cylindrical cellular process of splanchnic mesoblast in front of the pectoral region, that is, between the otocysts and the optic vesicles. After the alimentary canal is defined and when about 24 muscle plates are marked off, the heart has a vermiform shape and is still solid. A fissure soon appears in the centre, and the heart becomes a simple tube, then muscular twitchings and finally pulsations occur (48 to 60 per minute). The cavity is lined by a single layer of cells, but no fluid is yet present. Then the cardiac tube becomes conical, and a pericardial chamber is apparent, also formed from mesoblast. The heart next becomes L-shaped, the arterial end being median, the auricular passing off at right angles. The cardiac lining becomes papillose, and granules may occasionally be seen at the open end of the auricle which is in communication with a space — the germinal cavity. This cavity becomes indistinct after the closure of the blastopore but again becomes evident between the periblast and epiblast of the yolk-sac (periblastic blood-sinus of Cunningham). The latter author is of opinion that this cavity is theoretically no longer the segmentation- cavity, since it is on its inner or periblastic side partially lined by mesoblast-cells, viz. chromatophores from the periblast. Morphologically, it is homologous with the vitelline vessels of the salmon, and, similarly, is continuous with the auricle of the heart and shut off from the pericardium by a definite meso- blastic membrane. Meanwhile vascular canals are in course of formation, the dorsal aorta being hollowed out of the mesoblastic cells above the gut and under the uotochord, and two lateral venous trunks 88 GENERAL SKETCH OF in the connective tissue external to each head-kidney. In the larval ling the large channel from the liver to the tail, con- taining numerous large round corpuscles, can be traced on the fourth day. The venous trunks in front of the pectoral fins send prolongations downward and communicate with the venous end of the heart, which at this time is directed upward and backward as a sinus venosus which receives the venous trunks (Ductus Cuvieri) on each side, the latter trunk receiving not only the cardinal vein but the jugular. Around the cardinal veins the cell-tissue of the head-kidney or pronephros grows. Before the end of the first week after hatching a simple circulation can be detected. The anterior bulbus and ventricle of the heart drives the blood upwards behind the eyes, pro- bably by the artery of the hyoid arch, and along the dorsal aorta to the root of the tail, there it forms a loop and returns by a large venous trunk which anteriorly splits into the two cardinals. A day or two later a venous branch, the sub- intestinal, leaves the sub-vertebral vein and passes to the lower side of the gut along the margin of the liver to the sinus venosus. The intestinal artery (coeliaco-mesenteric) has a much longer course, for it leaves the dorsal aorta in the posterior region, traverses the mid-gut in descending, then courses beneath the rectum, ascends before reaching the vent, and passes along the anterior margin of the urinary vesicle to join the cardinal vein. In a cod of the seventh day the aorta and vein reached nearly a quarter the length of the cardinal trunk, while on the fourteenth day they extended almost to the tip of the tail. As was observed in the salmon many years ago by one of us the force of the current seems to hollow out the yielding channel. During the second week there are two gill-arches and sometimes three, which have developed arterial channels, and another vessel, the hyo-opercular artery. The mandibular courses along the ventral margin of the mandible, the vessels of opposite sides meeting and returning by a single vein. In the cod, when the caudal artery extends f the length of the tail, four branchial arteries are visible, and a sub-maxillary artery beneath the eyes, whilst a return current is directed over the eyes by the supra-ocular vein. The vessels of the abdomen — MARINE TELEOSTEAN DEVELOPMENT. 89 especially the veins — have largely increased, and the portal system is indicated. The proximity of the yolk to the great vessels returning to the heart makes it probable that these are connected with its absorption. As the absorption of the yolk takes place, however, before any circulation occurs, it is possible that the open end of the heart takes in the yolk-materials, as Mr H. C. Williamson has recently more clearly pointed out. The origin of the blood-corpuscles is still open to discussion, more than one mode of origin being suggested. Briefly the formative elements of the blood seem for the most part to be derived from the periblast, the primary corpuscles being moulded from the detached cells of the sub-notochordal trunks. Renal organs. Soon after the splitting of the mesoblast into an outer layer (somatopleure) and an inner layer (splanchnopleure) a rod of cells is budded off from the latter. In the transparent pelagic forms, e.g. the cod, we find that a solid cylinder appears on the outer margin of the intermediate cell-mass, and it rapidly develops forward to the pectoral fins, but advances posteriorly more slowly. A lumen is formed by a radial arrangement of its cells. Thus a pair of simple ducts, whose walls consist of a single layer of columnar cells, originate from the pectoral region to the root of the tail. Anteriorly each tube, widely separated from its neighbour, is folded on itself inwards towards the notochord, and ends in a trumpet-shaped infundibular opening. Posteriorly the tubes approach each other and co- alesce into a single tube, which at first is of small capacity with thick walls. Somewhat superficial at first, the ducts lie later ventro-laterally to the notochord and ultimately protrude into the peritoneal cavity. A connection is established pos- teriorly, at the urinary vesicle, with the cavity of the hind-gut. The crozier-shaped loop at the anterior end has in front of it a mass of trabecular tissue in which tubules appear to some extent to enter, but it is also penetrated by the basilar plate of the skull. Such is the condition of the pronephros or head- kidney at the time of hatching. When the young fish emerges, the folds of the loop in front 90 GENERAL SKETCH OF increase in complexity and a vascular glomerulus is developed near each trumpet-shaped opening (nephrostome). A little later a capsule encloses the opening of the head-kidney and the glomerulus — shutting off both from the body-cavity. Originally the waste-products passed directly from the body-cavity, but they are later conveyed from the special capsules to the urinary vesicle behind. As development proceeds the opening between the latter and the gut closes and the fluid has a separate passage posterior to the anus (ureter). In the later stages the lymphatic trabecular tissue in front is traversed by blood- vessels and probably in some by tubules. In the last stages of larval existence mesoblastic cells become aggregated along the dorsal region of the segmental ducts, especially in the fore and hind region, and minute sinuous tubules appear in their midst, which pass down and open into the segmental ducts. In such Teleosteans as the Wolf-fish, in which the later stages of development are well shown, the segmental tubes are absent at first in front, but develop largely when the ducts reach the middle line beneath the aorta, and the pigment having also greatly increased, the structure of the organs thus gradually assumes the adult character. Generative organs. As soon as the segmental organs have reached their final position on each side of the dorsal aorta, a straud of cells of the inner fold of mesoblast (splanchnopleure) passes below them. They form the generative epithelium though at first not a genital ridge. On the genital ridge which afterwards appears on either side of the mesentery, a groove develops, according to Jungersen, which subsequently closes to form a canal. In others (Physostomata), the lower edge of the genital ridge coalesces with another ridge projecting from the peritoneum — to the outer side. MARINE TELEOSTEAN DEVELOPMENT. 91 ORGANS OF THE HYPOBLAST OR INNER LAYER. Notochord or central axis. In the earliest section of the trunk no trace of the notochord is visible, a single layer of hypoblast limiting the neurochord below with the plates of the mesoblast above at the sides. About the time when the lip of the blastopore has reached the equator, a median mass of rounded cells is observed between the neurochord upon which they press and the thin median stratum of hypoblast below — in the middle region of the trunk ; it gradually extends forward, ending above the cardiac rudiment on the second day after the blastopore closes. In the early condition of the notochord the cells forming it are more numerous ; thus from six to eight cells extend across the diameter of the chord, due to irregular vertical septa (vacuolation), after the breaking down of cells. Later they form a meshwork, the outermost portions of which constitute a limiting membrane with occasional nuclei. Intruding meso- blast limits the chord below and later still surrounds it with a very thin layer. This is the simple condition of the parts about the 20th day after hatching — the sheath proper being very thin, and the mesoblastic sheath (perichordal) but little increased in thickness. From the latter the future vertebrae are formed, and an external limiting layer can be made out. This structure (external elastic membrane) gives origin to the arches, the upper (neural) preceding the lower (haemal) in development. In some forms (e.g. in the snake) cartilage develops preco- ciously in the vertebral column, but in many cases the arches and outer laminae of the vertebral bodies are not preceded by preformed cartilage but by ossific matter of a clear, homo- geneous, and brittle nature. Indeed in more advanced stages a spicular sheath has been found in the connective tissue out- side the external limiting membrane. 92 GENERAL SKETCH OF Branchial system. In the early embryo the hypo-blast forms the roof of the sub-germinal cavity, which has inferiorly the periblast envelop- ing the yolk. Behind and below the ears a fold grows up on each side of the middle line, meeting with the epiblast on each side of the ears. The lower edges of the lateral folds approach each other in the mid-line, fuse and close in the branchial region of the gut (Wilson). The same takes place with the trunk behind, but Cunningham denies this. Behind and below the ears a large oval area is apparently pushed in, resulting in the perforation of the epiblast on each side. The formation of the primitive operculum is not readily understood, for the oesophageal lumen is not yet formed. The opercular flap is a much later outgrowth from the tympanic region — which grows backwards over the gill-slits. Below the hind-brain and ears the hypoblast shows a great increase of cells, and beneath them mesoblastic cells make their way, bounded below by periblast, and these become columnar, forming paired rod-like masses. Some clays before emerging from the egg three branchial bars are usually visible, a long thread-like lacunar vessel is found along the posterior margin of each bar, and later a septum forms — constituting an upper arterial and a lower venous trunk. Five transverse bands, sometimes indications of six, extend across the floor and under the flattened O3sophagus, forming a series of mesoblastic cross ridges, and between these the hypoblast is pushed in as septa. The hypoblastic diverticula indicate the future position of the gill-openings, while the mesoblastic margin forms the gill-arches and branchial skeleton. The fifth branchial arch remains rudimentary in the forms dealt with here. In front of the four branchial arches are the stout bars of the hyoid arch, and again in front of these is the mandibular arch with cartilage-cells which are readily dis- tinguished by their height and stoutness. The elements connected with the suspensory apparatus of the lower jaw and the skeleton of the jaws gradually develop. In the young cod MARINE TELEOSTEAN DEVELOPMENT. 93 three weeks after hatching the branchial system is wholly converted into cartilage. In general the arches have a meso- blastic core and a hypoblastic epithelial covering. Skull. At first the brain is covered by a thin epiblastic layer com- posed of the flattened corneous stratum, and the thick sensory remnant beneath. A sub-epidermal enlargement often occurs beneath this in larvae in confinement, giving a peculiar outline to the animal. Then the mesoblast is aggregated between the eye and the neurochord, and, passing upward as a thin membrane, invests the brain. Ultimately the three brain-mem- branes are developed from it, and pigment is also formed on its inner surface. At the time of hatching the base of the brain is strengthened by two masses (parachordals), and in front of these are two cylinders (trabeculoe), which unite posteriorly. Within a week or 10 days after hatching these are converted into cartilage-cells. About the middle of the third week various other cartilages of the skull are formed, e.g. the optic and auditory, but the translucent opercular plates on the surface of the more exposed skeletogenous elements of the face occur later (post-larval stage). While cartilage thus early develops in the skull, that of the axial skeleton is later. Alimentary canal. In its earliest condition the alimentary tract consists merely of a thickened layer of hypoblast, intervening between the neurochord above and the periblastic covering of the yolk below. When little more than one-third of the yolk is covered by the blastoderm the hypoblastic cells assume a columnar character, arching over a lumen — beneath which is a floor of periblast. This is the first indication of a gut — the continuity of which with the neurenteric canal and medullary groove has been described. From the arched roof the notochord is diffe- rentiated. The lumen extends only a short distance forward, and is lost in hypoblastic cells which reach as far as the cardiac region, where they spread out and form a delicate limiting mem- brane below the head. The ventral wall of the canal is formed 94 GENERAL SKETCH OF of hypoblastic cells pushed in from the side, or budded off from the nucleated periblast below, probably both1. The mid-gut soon, when the invagination of the lenses becomes complete, forms a massive cylinder, and the oral tract a wide sheet of hypoblast. By the time the ear-capsules have thinned out and the otoliths have been developed, a fine fissure traverses the pharynx, and the lumen continues to the blind end of the canal. The cells become cubical and of many layers, and the inner lining has a granular or mucoid appearance. The inner mesoblastic sheet (splanchnopleure) e.g. gives a coating to the canal — forming the muscles and connective tissue, while ex- ternally it gives rise to the epithelial peritoneal layer. The oesophageal tract at least appears to be covered with cilia. At first straight and smooth — the walls of the canal in the later stages become folded and wrinkled — especially posteriorly. Indeed, about 10 days or a fortnight after hatching some forms show a capacious though flattened oral chamber and oesophagus, the latter giving origin to the duct of the swim-bladder. The enlarged stomach follows, the liver lying beneath, with the dense pyloric section posteriorly — the pyloric casca subsequently springing from it as evaginations, as observed in young cod ^ to 1J inch in length. From this position also passes the bile-duct. The intestinal walls are likewise dense, and rapidly develop ruga3 and a glandular character. Posteriorly the rectal region is marked by a cincture or valve, followed by an enlarge- ment ; it then bends downward and narrows to form the small vent — opening upon the muscular papilla. Mouth. An involution of the outer layer (epiblast) to form the mouth does not occur in these forms. The oral cavity is capacious and the branchial framework supporting its floor and sides well advanced when a fissure from the chamber bursts through in front. The jaw-cartilages grow rapidly forward, and the mouth, at first ventral and shark-like, opens from above — from the rapid forward growth of the lower jaw. J Cunningham says periblast only. Q. J. Mic. Sc., vol. 26, 1885. MARINE TELEOSTEAN DEVELOPMENT. 95 Vent. In the same way the posterior opening of the gut is not produced by an inpushing of the outer layer (epiblast), but is formed five or six days after hatching by the extension of the internal lumen. At first the aperture is above the ventral margin of the fin, but later the membrane is absorbed, and the condition is almost that of the adult. Liver. Soon after the ear-chambers (otocysts) are formed the ventral wall of the mid-gut shows an enlargement anteriorly, and from this the liver arises as a solid proliferation. Into the early organ a delicate canal passes from the fissure of the gut— the common bile-duct. By-and-by the liver becomes bifid — a right and a left lobe being distinguishable — and in the midst a spacious gall-bladder develops. The importance of the liver in the larval fish is shown by its rapid growth into the yolk-sac of the gunnel, and Wilson is of opinion that it is the medium for the absorption of the yolk and periblast. Swim-bladder. From the dorsal wall of the alimentary tract the swim- bladder is given off as a very thick-walled diverticulum which presses upwards against the notochord, and remains for some time connected by a fine canal with the gut. Before the embryonic period ends, however, the duct atrophies in the forms under consideration. In recent years considerable attention has been devoted to the artificial hatching of sea-fishes, especially since their life- histories have been more accurately known. In the Trawling Report of 1884 one of us suggested that efforts might be made in our country to augment the " soles, turbot, brill, and other flat fishes," and perhaps also " the cod and other round fishes." " Such an experiment, scientifically carried out, would give a valuable basis for future legislation, tend to increase our knowledge of the food-fishes in a remarkable degree, and would 96 GENERAL SKETCH OF MARINE TELEOSTEAN DEVELOPMENT. be worthy of the interests which this country has in the department of sea-fisheries1." In a later publication-, when dealing with the same subject, it was distinctly stated that, " It has yet to be proved that the artificial hatching of sea-fishes, even on a large scale, will be beneficial to the fisheries generally ; yet the importance of the issue demands that a thorough trial be made." In view of the latter opinion, and when opportunity offered, cordial support was given to the experiment begun by the Fishery Board (previous to 1892) at Dunbar. Until, indeed, the experiment is performed on a scale sufficiently large to give reliable results, criticism on the general question is unnecessary, though in some cases it may be valuable in the suggestion of methods. 1 W. C. M. Trawling Report, 1884-85, p. 379. - Ibid. A Brief Sketch of the Scottish Fisheries, 1892, p. 15. CHAPTER VI. THE RATE OF GROWTH OF FOOD-FISHES. THE importance of this subject from an economic point of view would be difficult to exaggerate, and at the same time it is remarkable how very little is definitely known in connection with it. It has already been pointed out in the Introduction that the various changes undergone by a fish in its transition from the egg to the adult condition fall under two heads, — firstly, quantitative changes, which involve increase in bulk, and secondly, qualitative changes, or those which are concerned with the differentiation of organs. It is well to understand clearly that these two series of changes have not necessarily any direct connection, although the quantitative changes of growth appear to precede to a great extent those of development or differentiation. These quantitative changes which cause the phenomena of growth are held by most zoologists to be either the direct effect of, or intimately connected with, cell-division. Thus if cell-division takes place at a more rapid rate than the loss of cellular tissue to which the organism is constantly subjected, the result is an increase in the number of cells and of general bulk, or, in other words, growth, and so far as is known there is no inherent reason why this growth should not proceed throughout the life of the organism. We do not mean by this that there is no limit whatever to the size attainable by any given species; thus, to quote Weismann, "although many fishes, reptiles and lower animals are said to grow during M. F. 7 98 THE RATE OF GROWTH OF FOOD-FISHES. the whole of their life, we do not mean by this that they possess the power of unlimited growth any more than that of unlimited life. There is everywhere a maximum size which, as far as our experience goes, is never surpassed. The mosquito never reaches the size of an elephant, nor the elephant that of a whale." Limit of Growth. The problem of the growth of fishes is thus very different in many ways from that of the growth of quadrupeds and birds. As one would expect, from theoretical considerations, the fish as an organism shows itself to be more directly susceptible to the influence of its environment. Thus, as mentioned below, the period of incubation can be altered at will between very wide limits by simply varying the tempe- rature, whereas one would hardly expect to alter the period of mammalian gestation, or even of avian incubation, except within extremely narrow limits. Other instances could easily be given to show that the direct effect of the environment upon the piscine organism is much greater than that upon land organisms. Without considering the effect of environment it is probable that the life-cycle of the individual fish also differs in very important respects from that of the higher animals. ' In the mammalia there is a definite duration of growth in bulk and of life quite apart from the environment, and either period can only be altered by continued action of the environmental factors through many generations. There are some experiments which point to the conclusion that the duration of life of the individual fish is only limited in the widest sense (with immunity from environment), but leaving this out of the question there is no proof of the hypo- thesis that the individual fish ceases to grow at any period of its life ; on the contrary, there are considerations which point to the other view, i.e., that a fish continues to grow throughout its life. 'Thus, if a mammalian or avian species be subjected to the destroying agency of man, there follows a diminution in numbers; whereas, if a fish be subjected to like conditions a reduction in the size of the individuals is the immediate result. These facts can be explained as follows : — In the former case THE RATE OF GROWTH OF FOOD-FISHES. 99 very little, if any, growth in size takes place after the attain- ment of sexual maturity, whereas, in the latter case, growth continues indefinitely after that event1. We may take an example to make this clear. Let us suppose that a certain fish of species A first comes to sexual maturity at a size of 6 inches, and that it grows 1 inch a year, and continues to reproduce its kind annually up to a length of 20 inches. In this case there is an active reproductive period of 14 years. It is a well-known physiological fact that the period of reproduction is not constant in intensity from its inception to its end, but both in quantity of eggs laid and vigour of offspring it varies, attaining a maximum in the middle and minima at each end. In this way the fish although attaining maturity and producing sexual products at a length of 6 inches will then only produce smaller and weaker offspring, whereas by the time of reaching, let us assume, some 19 inches, offspring of greatest vigour will be the result. In a state of nature a balance is struck and both kinds are found living together, or rather, to be more accurate, every gradation, from those maturing from the first time at 6 inches and those maturing at 10 inches. Now let us consider the effect of over- fishing. Assume that fishing has been carried to such an ex- treme that an inappreciable quantity of fishes reach the length of 10 inches, so that only those are left which commenced breeding at 6 inches and somewhat upwards, and their offspring- will have the same tendency to commence breeding at the earlier age and shorter size. The first effect will therefore be that of reducing the average size of the fish. It is a well-known fact that a far greater number of fry are annually produced than will ever attain sexual maturity, and the numbers surviving in any given district till sexual maturity will bear a constant relation to the amount of nutriment supplied by that area. Returning to our example let us suppose that in a given district at a given time there are of species A (1) 50,000,000 adults past their maximum reproduc- tive capacity and of 12 to 24 inches in length, (2) 100,000,000 1 A. T. M. 'Rate of Growth of Marine Food-Fishes.' 13th Ann. S.F.B. Report, p. 289. 7—2 100 THE RATE OF GROWTH OF FOOD-FISHES. at a length of 8 to 12 inches at the maximum age and length for reproductive purposes and (3) 120,000,000 at a length of 6 to 8 inches which have reached the inceptive stages of repro- duction, and (4) 300,000,000 immature individuals. For the present purpose we may neglect the (1) and (4) and consider (2) and (3). The effect of over-fishing is to remove the second series, for apart from other reasons the members of (2) have been subjected to the risks of being caught at stage (3) before attaining to stage (2). The removal of (2) and with it (1) means that, the sup- porting powers of the district remaining the same, a much greater quantity of (3) and their offspring included in (4) will be able to find subsistence and will survive. Hence the second effect of over-fishing a district will be that of multiplying the numbers. The results at which we have arrived here are found in nature. One of the most obvious effects of over-fishing a district is that there results a great number of small-sized individuals of the particular species affected. Looked at from a general aspect we may express this result thus: The species is subjected to risks of extinction and the counteracting effect is a reactive production of a greater number of smaller indi- viduals, so that extinction may be prevented by multiplying the chances of a sufficient number reaching sexual maturity. The want of this ready re-activity to changed surroundings in mammals helps to render the whales and seals liable to extinction at the hands of man. In the case of birds and quadrupeds these laws apply, as already stated, only to the extremely limited extent to which the attainment of sexual maturity and the cessation of growth do not quite correspond in point of time. Hence protective reduction in size of a bird or mammal can only take place at an extremely slow rate. The amount of growth of a fish per annum may be expressed as a certain proportion of the whole, and must form an ever- decreasing geometrical series, so that the total amount of growth is only limited in the same sense asl+| + i + ^, etc., is limited THE RATE OF GROWTH OF FOOD-FISHES. 101 in a finite series to less than 2. A study of the average sizes of fishes shows that the annual increase is practically distinctly appreciable. There is no reason, therefore, to believe that there is any definite limit to the size of a fish, as used in the sense in which we speak of a definite size attained by mammals and birds, beyond which growth in bulk does not proceed. Exceptions to this rule have been put forward from time to time ; for example, some observers hold that the herring forms one of these, so that after attaining a definite size it entirely ceases to grow. It seems, however, that these state- ments, though handed on from one authority to another, do not rest upon sufficient basis of fact. We may next enquire consecutively into the various factors of environment which are known to affect the rate of growth of fishes. Nutrition. The direct effect of nutrition upon growth is well known, even in the case of the higher animals, so that the result is probably intensified in the case of fishes. This factor obviously cannot come into force till the post-larval stage is reached, for before this, the larva is supplied with a copious quantity of yolk-material, which though doubtless varying in slight degree may be reckoned as constant for all practical consequences. Dr Meyer in 1878 made experiments with young herrings, and found that those reared artificially grew at a slower rate than those living under natural conditions, but that later they, so to speak, made up the leeway. This is shown as follows :— Length in millimetres attained by fish under Artificial Natural Age from impregnation. conditions. conditions. One month 10—11 17—18 Two months 17—19 24—36 Three months 30—35 45—50 Four months 48 — 54 55 — 61 Five months 65—70 65—72 From this table it can be seen that the two months' herring brought up in captivity are some of them less than half the 102 THE RATE OF GROWTH <>F FOOD-FISHES. size of some of the same age reared under natural conditions, but that at 5 months the average sizes are closely similar. Dr Meyer was inclined to ascribe the small size of the former to the insufficient or unsuitable food. Detailed experiments have yet to be made with regard to the direct effect of varying the supply of food upon the growth of fishes of a more advanced age than these herrings. Temperature. One of the commonest observations in re- spect to the development of fishes is the great variation in the length of the period of incubation according to the temperature of the surrounding medium ; indeed it is now usual for natu- ralists in quoting the duration of incubation, in any given instance, to mention the temperature of the water in which the eggs were hatched. Dr Meyer, in the same series of observations as above re- ferred to, experimented with the eggs of the herring, and came to some important conclusions. He found that at a temperature of 51°8F. to 53°6F. they hatched in 10 to 11 days, but at a temperature of 35°6 F. the embryos emerged on the 29th to 33rd day, and even later. Again at 32° F. the hatching period was postponed till the 47th day, and at 30°56 F. normal development was no longer possible. At this low temperature the yolk became opaque and burst the egg-capsule, thus destroying the embryo. He also found that the influence of cold in retarding growth is more marked upon the later stages than upon the earlier1. It might at first be supposed that we have here a pheno- menon of retardation of development or differentiation and not of growth, but it must be remembered that in Vertebrate embryos all the earliest stages are almost entirely taken up with processes of growth or cell-division, and that differen- tiation of organs only takes place later, so that it is not a great assumption to suppose that we have to deal mainly with the retardation of growth. Recently Mr Harald Dannevig'2 has attempted with some success to determine in an exact manner the connection 1 Vide also W. C. M. Nature, Vol. 34. 2 Superintendent of the Hatchery at Duubar. THE RATE OF GROWTH OF FOOD-FISHES. between the duration of the incubatory period and the tempera- ture. Thus by hatching eggs of some of the commoner species at successive constant temperatures and comparing the results, he has been enabled to make curves dependent on geometrical principles, which show the exact length of the period of incubation at any given temperature. The subjoined table is a reprint from his work1. He was led, from his observations, to the following conclusions :— ' 1. That the period of incubation is various under the same circumstances for different species. ' 2. That this difference is in relation to the size of the eggs in this way, that the large eggs take a longer time than the smaller ones. ' 3. That the time of incubation for the same species varies according to the temperature. ' 4. That this variation for each degree is comparatively larger in a low temperature than in a high. ' 5. That development of fish-eggs takes place also at temperatures below zero (Centigrade), when the specific gravity of the water is sufficiently high to prevent it from freezing.' Dannevig obtained almost the same results with the eggs of the cod and those of the haddock, but it is well to remember that although the average size of the former is slightly smaller than that of the latter yet the two overlap each other considerably. Thus Mr Williamson'2 recently found that out of a considerable number of examples, the smallest cod's egg was 1'35 mm. in diameter and the smallest haddock's egg was 1*38 mm., whilst the largest haddock's was 1*458 mm. and the largest cod's was 1'467 mm. in diameter. The mean 1 Fahrenheit i 28°'4 30°-2 32° 33°'S ' 35°-6 37° -4 39° '2 41 i 42°-8 46°'4 50° s.r-6 57-2 temp. } Centigrade 1 -2° -1° 0° 1° 2° 3° 4° 5° 6° 8° 10° 12° 14° temp. J 1 Cod Whiting Haddock 42 42 23 23 201 131 154 8* 13 104 lOf 9| 64 84 5| 8f Time of incuba- Plaice Flounder 64 4 12 44 104 tion in days. 8 13?/; Ann. S.F.B. Eeport. 104 THE RATE OF GROWTH OF FOOD-FISHES. size for cod's eggs was T386 mm. and that for haddock's was 1'458 mm. These figures show that the two species have eggs so near in average bulk that the difference is probably in- appreciable by the apparatus used by Dannevig. This third conclusion is suggestive. Would it not be possible, with exact physical apparatus to show that at a constant temperature the time of incubation is inversely as the mass, or inversely as the relationship of the bulk to superficial area, or in some way to reduce the specific variation in time of incubation to a function of the mere bulk contained within each egg ? Passing to the further effect of temperature-variations upon the growth of the older fishes, there are a number of obser- vations which tend to show that fishes as a rule grow very little during the coldest months of the year. PatJtological. Fishes appear to be very susceptible — as regards their growth — to slight injuries. Small abrasions of the surface of the body seem to have a marked retarding effect upon the growth of the individual. The importance of this will be seen later. Lastly, we have to deal with what we may term the ' indi- vidual tendency.' It is an accepted axiom that individuals vary not only in size but in other respects, and there is no reason to hold that the rate of growth forms an exception to this statement. In the same brood marked differences occur in the size both of the eggs and embryos and there is reason to believe that, cceteris paribus, these differences become ac- centuated as time goes on. These factors, briefly touched upon, form such a formidable array of tendencies, all active under natural conditions and all working to the same end, namely, diversity of growth, that they might well appear to make the rate of growth of fishes an insoluble problem, but we do not think matters will be so hopeless as this if carefully considered. Under natural conditions, all the fishes of a species in a given limited area have the same habits and all are subjected to approximately the same conditions of temperature, the pathological element may be eliminated by selections of THE RATE OF GROWTH OF FOOD-FISHES. 105 specimens, and the 'individual tendency' has a correcting factor in the action of natural selection. Thus the mean average size of the species at every stage has been determined under natural selection, as that which has the greatest im- munity from destructive tendencies of every sort, and greatest chances of surviving ; and the same destructive tendencies still act upon every generation to weed out the maxima and the minima to the preservation of the mean, so that after the early stages the largest or smallest are continually removed, the mean only surviving. This fact is important, for it follows from the same, that if one takes a number of fish-eggs or young Iarva3, already varying in slight degree as regards size, and places them in tanks in artificial conditions — such as easy access to food, and immunity from actual foes ; then, all the slight varia- tions in size tend to be emphasised, and in course of time fishes of very various sizes will result1. This does not come under the head of a theoretical hypothesis except to those, a few of whom may yet be found, who will not acknowledge the principles of natural selection, the struggle for existence, and the survival of the fittest. In natural conditions, especially in gregarious fishes, the environmental factors of temperature, nutrition, &c., must be very closely similar, if not identical, for all the individuals of one brood, so that, taking this into account, and also what has been stated about the variation, there are good grounds for supposing that the rate of growth is closely similar for each individual, and that a mean average length for any given species at a given age is a fixed and determinable quantity. The greater the difference in environment the greater the dif- ference in the mean average, till ' forms ' may become marked. Thus, in estimating the rate of growth, the great effect of marks or abrasions upon the fish precludes the employment of this method of procedure, although apart from this drawback it enables us to study the growth -rate of individuals under their actual conditions. The method of rearing young fish in confinement and measuring them periodically must be condemned once for all 1 A. T. M., loc. cit. 106 THE RATE OF GROWTH OF FOOD-FISHES. as absolutely useless ; for, although the temperature may be registered, the conditions of nutrition and the exclusion of the check caused by natural selection upon ' individual tendencies ' make all results obtained in this way abnormal and useless. This method has been resorted to by Mr Cunningham at Plymouth and by some other workers. Although we have under the heading of each species quoted some of his obser- vations by this method, yet from what we have remarked above, it will be manifest that there are strong objections (which cannot be lightly dismissed as being 'theoretical') to conclusions deduced from such experiments. This being the case the growth of fishes under natural conditions must be studied in another way. If a great number of fishes be caught and measured and the date of capture registered, and again, if the spawning-period of the species be known, then the age of each fish may be estimated with a fair degree of accuracy. A number, the more the better, may be caught and measured at one elate and the mean size determined ; we may then assume that the indi- viduals of mean size at any date correspond in age with the mean spawning period, or the period in the middle of the spawning-time — at which there are always the greatest number of eggs laid. In doing this we assume that the greater part of the diversity in size of one ' haul ' of fishes, if they are within a year of each other in age, is due to the extended spawning- period ; in fact, is due to a greater or lesser age. A great number of facts and observations tend to bear out this hypo- thesis. The method is probably capable of great accuracy, provided a sufficient number of specimens are caught and measured and the true mean for the district thereby deter- mined, but approximate results may be obtained from even a few specimens, by gauging the age of each individual, within the limits allowed by the spawning-period, as being that which best corresponds with that of a different age. If the age of each fish be expressed by a point upon a curve then each is so estimated as to make the curve as regular as possible. Here we assume that the individual fishes are derived from eggs spawned at different times. The method is capable of a fair THE RATE OF GROWTH OF FOOD-FISHES. 107 degree of accuracy but one or two extreme cases may con- siderably upset the results. A comparison of the curves derived by this method with the table of maxima and minima of sexual maturity given in this chapter shows that the majority of the food-fishes attain sexual maturity for the first time during their third year, and in the case of the males it is very likely that most come to maturity during their second year. As in other remarks concerning fishes in this department we must be understood to refer only to the average, and indeed it is likely that the average attainment of maturity may be as late as the third and fourth years for males and females respectively. Further we may notice that the larger species of fishes appear to differ from the smaller in their size mainly on account of a greater rate of growth from the outset, by which is meant, from the hatching epoch, although the size of the egg and newly hatched larva only corresponds in a very rough way to that of the adult. Thus the ling shows a more rapid rate of growth during the whole period of maturity than the gurnard or smaller species, and the same applies to the plaice amongst flat fishes compared to the dab and others. We have, up till now, left out of consideration all points in connection with differentiation of organs. This development, proceeding contemporaneously with growth, reaches a culmi- nating point in the maturation of the sexual organs. These are the last organs of the body to become functional, and a fish which has reached this stage may be known as ' mature ' or ' sexually mature,' while one which has not yet become sexually functional is 'immature.' Some observers have attempted to apply these terms as if they were synonymous with ' fully- grown,' but from what has been said above, the latter term is meaningless and even misleading as applied, at any rate, to the vast majority of fishes. If we wish to speak of young and adult fishes we can only do so by selecting an anatomical or physiological feature as a line of demarcation between the two, and the feature which is eminently suitable for this purpose is that of functional sexuality. The terms ' immature,' ' mature,' ' young ' and ' adult ' can not be scientifically defined in any 108 THE RATE OF GROWTH OF FOOD-FISHES. other way. Dr Fulton insisted upon the same interpretation of the terms 'immature' and 'mature,' in 1889 1> and determined the average size limit of immature individuals in the case of several of the commoner species. His work led him to conclude that the following sizes repre- sent the lowest length, without reference to sex, at which maturity is nearly reached in the following species: Plaice 12 inches Turbot 18 inches Lemon-dab 8 „ Brill 16 „ Dab 6 „ Sail-fluke 9 „ Flounder 7 „ Haddock 10 „ Witch-sole 12 „ Cod 20 „ Long-rough dab 6 „ Whiting 8 „ Little sole 3J „ Gurnard 8 „ Cat-fish 20 (?) inches. The practical application of our knowledge of growth-rates depends largely upon the determination for each species of the average size and age at which sexual maturity is reached. A large number of fishes may be caught and measured and their sexual organs examined, and we may thus determine roughly a mean minimum size for the attainment of maturity ; and this has been done in a great number of cases. The results of examination of a number of fishes caught by the ' Garland ' show the following maximum and minimum sizes which were sexually mature2. Whilst those below the minima were not sexually mature, it does not follow that individuals above the maxima no longer spawn, in fact, a cod of 46 inches is, on the same Table, recorded as in the ' spent ' condition. Male. Female. Inches. Inches. Cod 23 — 37 36 (only one example). Haddock 12—16 14—24 Whiting 9—15 9— 18-J- Bib 8—8 7|— 10 1 T. W. Fulton, 1th Scot. Fish. Board Report, p. 161. - T. W. Fulton, Wth Scot. Fish. Board Report, p. 239 a. THE RATE OF GROWTH OF FOOD-FISHES. 109 Male. Female. Inches. Inches. Gurnard 11—16 8|— 18 Long-rough dab 5 — 8 5 — Plaice 13—22 20—28 Lemon-dab 8^—16 11—19 Dab 6—12 51—17 Flounder 6—14 7— 17£ By this method one important fact has been brought to light, namely, the difference in size between the sexes. It appears to be the rule that the attainment of maturity takes place at a different size in the two sexes, the male pretty generally being the smaller among Teleosteans. Occasionally, as in the long-rough dab and the salmon, this is very marked. Here again we have the same two alternatives by which to explain the facts. Does the male grow more slowly than the female, both maturing at the same time ? or does the male mature at an earlier age, the rate of growth for the two sexes being the same ? In other words, which varies — the rate of growth, or the period of development ? We have data which distinctly point to a variation in the duration of the development, the male sexual organs maturing at an earlier date than the female. This is undeniably the case in Myxine, and also in the salmon. Thus we may assume that the rate of growth in the two sexes is closely similar. Both sexes are placed under the same environmental conditions, and hence a different rate of growth would be at least im- probable, whereas the more rapid maturation of the male elements may be traced to a deep-seated origin which may be best explained by the following remarks by one of us1. " There are numerous facts in the ontogeny and structure of the Vertebrata which point to a hermaphrodite chordate ancestor. " Amongst these we may cite Nansen's observation of the protandric hermaphrodite condition of Myxine* and the 1 A. T. M. ' Hermaphroditism in the Cod,' 13th Scot. Fish. Board Report, p. 298. • Aarsber. Beryens Mus. 1887, op. vii. 110 THE RATE OF GROWTH OF FOOD-FISHES. important fact that amongst Teleosteans, as a rule, the males are considerably smaller than the females, and there is in some cases (salmon, etc.) certain proof, and in others a great pro- bability, that the males are mature at a much earlier date than the females. Above the fishes also, many of the Amniota show a tendency for an earlier maturation of the male products. The bearing of this may be seen as follows : — We may suppose the primitive hermaphrodite chordate ancestor to have a continual rhythmical and perhaps seasonal predominance of one sex more or less directly dependent upon the environment (cf. Yung's1 observations upon Tadpoles). The tendency for a seasonal repetition of the same environment would then have the effect of causing a perfectly rhythmic sexual cycle, from male to female, in each individual. This would be a case of poly cyclic hermaphroditism. " This condition appears to be exemplified in the abnormally hermaphrodite Teleosteans. The lengthening of the cycles would result in a monocyclic hermaphrodite condition, as still persistent in Myxine, and lastly, the reduction of the stage at the commencement of the cycle to vanishing point in some individuals (females), and hypertrophy of the former half in others (males), would cause a hemi-cyclic dioecious species, as exemplified by all other vertebrata. If we may regard this as, in a general way, the line upon which the dioecious condition has been evolved phylogenetically in the chordata, we have an explanation of the facts of the earlier maturation of the males." If the maturation of the sexual organs takes place in a certain district at a smaller size than is the case with the average, then a " race " or variety of smaller fishes ensues. Thus the plaice in the south of England, as far as statistics go, appear to be smaller on an average at the attainment of maturity than those, e.g., on the east coast of Scotland ; and Petersen deduces facts to show that a smaller 'form' is present in the Baltic than in other Danish waters ; while in Iceland and certain other regions a ' giant race ' predominates. If we assume that in all these ' forms ' the sexual organs mature at 1 Arch. Zool. Expr., vii. Arch. Sci. Phys., Nat. xiv. THE RATE OF GROWTH OF FOOD-FISHES. Ill the same lapse of time from the hatching date, then the different rate of growth, due to the different environment, would account for the smaller size. The other alternative is to assume an approximately con- stant rate of growth in each case, and an earlier (in time) maturation of the sexual organs. Further investigation alone can show whether one has to deal with a case of a hastening of sexual development (paedo- genesis) or a retardation of growth. Petersen is inclined to think that the period of attainment of maturity (roughly three years in the Danish waters) corresponds with that of our east coast ' form ' of plaice, which rather indicates the latter. We may perhaps appeal to a similar series of phenomena to account for these ' races ' or ' forms ' as in the case of the effects of over-fishing (see above). If in a given district the increase of natural enemies to the plaice is so great that their influence may have the same injurious effect as that of over-fishing by man, then a small form would result, and this would possibly be assisted in warmer regions by earlier maturity, whereas in districts where the species is comparatively unmolested, and especially when in the colder climes, with, for this reason, a retarded sexuality, a larger race or ' form ' would be evolved. This phenomenon of ' forms ' practically resolves itself into a special case of the general law stated in the early part of this chapter, namely, that the mean average size of the species at every stage has been determined under natural selection, and is obviously mutable under changed conditions. It is convenient to express the facts of growth-rate by growth-curves. These curves are intended to express the nor- mal rate of increase in length of the species under natural conditions, and we should add, in the particular environment in which the specimens were caught. It is preferable to gauge the size of a fish by its length because this dimension is most readily determined, and in consequence all ' prohibitive ' legis- lation must adopt a standard of length. From a scientific point of view the method, as expressing comparative increase in bulk, is inaccurate, for in comparing the growth- curves we must assume that the length is in all cases an equivalent 112 THE RATE OF GROWTH OF FOOD-FISHES. proportion of the whole bulk, and this is not so. The curves, for instance, of the gadoids may be compared amongst them- selves with a fair degree of accuracy, but a glance at the shape of a pleuronectid will at once remind the reader that the length is a less predominant factor of the total bulk than is the case with a gadoid. It is well also to bear in mind that the length is not a constant factor in the bulk of a single fish throughout its life ; and it bears a greater proportion to the other dimensions as a rule, in the younger stages, so that the curves cannot in any way be said to represent accurately the increase in bulk or total growth of the species in question. Any rhythmical disturbing factor in the growth under natural conditions should appear in these growth-curves as a constantly repeated deflection in the curve, so that the seasonal variation in temperature should, if the curves were quite accurate, cause a recurrent series of secondary curves. The method of investigating rates of growth, as described and advocated here, depends, as already stated, upon a know- ledge of the spawning-period, and the inverse to this holds. Thus, given a sufficient number of specimens of a certain species and their date of capture, the curves of maximum and minimum size at every stage may be defined, and these, produced back to the base line, will give the duration of the spawning-period. This method has been pursued recently for the herring1 and the sand-eel2 and shows that both these forms have two spawning-periods in a year, besides indicating other important facts, such as the difference in size and rate of growth of the spring and autumn ' races ' of herring, and the habitat of the young fishes at each stage, as mentioned under the sections devoted to the herring and the sand-eel respec- tively. Table A will give a good idea of this graphic method of representing the rate of growth and life-history of the herring; the gradual progression of the young forms as they 1 A. T. M. ' Rate of growth of the Herring.' l±th Scot. Fish. Board Report. 2 A. T. M. 'Life-history and rate of growth of the Lesser Sand-eel. Annals and Magazine of Natural History, Sept. 1895. THE HATE OF (ilU)WTH OF FOOD-FISHES. 113 grow, through the in id- water and surface and thence to the shore, is indicated by special markings at each stage. It may be noticed that the close approximation to the same parallel of the limiting curves is an additional justification of the view that the greater part of the disparity in size of young fishes at any date is due to a disparity in age, which is itself made possible by the prolonged spawning-period. Further remarks upon this table will be found in the part dealing with the herring. It is patent to all that, notwithstanding what has been done in this branch of ichthyology, the results are at present very meagre, and although the experiments upon fish-eggs may lead us in the right direction for elucidating the laws which govern the growth of fishes and may enable us eventually to reduce them to concrete terms, yet this end is at present distant. Every step in advance has however its peculiar interest from a general point of view, for the fishes, with their simpler organisation and their less fixed and definite pre- determining influences, lend themselves the more readily to experimental investigation, and form a vantage point from which to make a flank attack upon the great problems of growth and reproduction — as exemplified in their greatest complexity by the higher Vertebrates, including man himself. M. P. 8 PAET II. LIFE-HISTORIES OF THE SPECIES. CLASS PISCES (FISHES). Aquatic vertebrates breathing by gills. Heart, as a rule, of two chambers ; blood cold. Two pairs of limbs in the form of paired fins supported by fin -rays are usually present, and are not divided externally into arm, fore-arm and hand, or thigh, leg and foot. They are the only animals possessing median fins supported by fin-rays. A row of sense-organs along each side of the body. Skin generally covered with scales. No amnion and no allantois. SUB-CLASS I. TELEOSTEI (BONY FISHES). Skeleton bony, with well-developed skull. Tail externally symmetrical (though it may not be so structurally). Brain with large optic lobes ; optic nerves simply cross • olfactory organ double. Gills free, in one chamber, and with one aperture of exit behind the ^'ill-cover. Heart with a non-contractile bulb at the origin of the o ° aorta. No spiral valve in the intestine, which opens independently of the urinary and genital apertures. Air-bladder usually present. Eggs numerous, generally small, almost always fertilized in the water. ORDER I. ACANTHOPTERI. Fishes generally having ctenoid scales. Ventral fins thoracic or jugular. Some of the fin- rays spinous, that is, unarticulated. No duct to the air-bladder (physociistous). CHAPTER VII. LIFE-HISTORIES OF THE SPECIES. The Perch-Family. (Percidae.) THE SEA-PERCH. (Boccus labrax, L.) THE eggs of this species have not been procured in a ripe condition, and indeed, on the east coast, the fish rarely comes under notice. On the west coast, however, as at Southport, observations should be comparatively easy. Day considered that they deposited their spawn about the mouths of rivers during the summer months ; while Couch states that July and August are the breeding-months in Cornwall. Thompson found the ovaries enlarged about the end of March, and the eggs about the size of millet seeds. Raffaele found the ripe pelagic eggs of the sea-perch in the Mediterranean from January to the beginning of March, and from what has been observed in a large female at St Andrews in May the spawning- period would seem to be late, for the ova were comparatively small1. The eggs, which, according to Raffaele, range from 1*155 mm. to 1*2 mm., have the usual pelagic characters, viz. a transparent capsule, homogeneous yolk and an oil-globule with a diameter of 0-333 mm. to 0*366 mm. (Plate I. fig. 2). Occasionally two or three oil-globules may be present, as in the gurnard and other forms. During the development of the egg blackish pigment occurs along the dorsum, and yellowish pigment on the sides. The latter colour also appears under the oil-globule. The abundance of the pigment and the large size of the chromato- 1 W. C. M. fith Ann. S.F.R. Report, 1888, p. 276. 118 THE PERCH-FAMILY. phores, indeed, form a feature of the egg. Hatching occurred in three or four days. The larva (Plate V, fig. 1) measures on extrusion 2'5 mm. and is thus about the size of the American species (L. lineatns). It is distinguished by the presence of the large oil-globule with its yellow pigment at the posterior and inferior border of the ellipsoidal yolk, by the presence of a pre-anal marginal fin of considerable length, for the vent is far back, and by the yellowish pigment of the body and anterior part of the yolk-sac. No pigment is present in the eye. On the 6th day (Plate V, fig. 2) the larva measured 4'7 mm. to 4'8 mm., and the pigment had grouped itself in two massive touches on the body, besides patches on the head, tail and yolk-sac, but instead of the yellowish hue it was now greyish-brown, with black corpuscles along the abdominal roof. Only at the tip of the tail did the pigment pass into the fin-membrane. The yolk soon disap- peared, and by the 12th to the 15th day the post-larval sea- perch had the vent about the middle of the body, with a long pre-anal marginal fin, bluish eyes, very large ear-capsules, erect breast-fins and a deep brown pigment-band along the lower border of the muscle-plates from the swim-bladder nearly to the tip of the tail, only a short line of pigment occurring on the opposite border of the muscle-plates in the tail. The head is somewhat large and the mandible massive. Closely allied young percoids 11 mm. in length were sent from Naples by Mr H. C. Williamson in June. THE COMBER OR SMOOTH SERRANUS. (Serranus cabrilla, L.) This species is common off the Channel Islands and the southern coast, and occasionally is procured in the Moray Frith. All that Day observes with regard to reproduction is that the ripe fishes occurred at the end of summer or in August and September. Raffaele studied the egg at Naples, where he found it in spring and the beginning of summer. The egg (Plate I, fig. 3) is pelagic, as first stated by Hoffman, compara- tively small, viz. 0'90 mm., with a small oil-globule measuring 0'15 mm. in diameter. The pigment in the developing embryo THE RED MULLET-FAMILY. 119 is slower in making its appearance than in the sea-bass, but the tints are similar, viz. yellowish and black. On its escape the small larva (Plate V, fig. 3) is characteristic, the large ellipsoidal yolk projecting in front of the depressed snout, while the oil-globule is ventral and median. A pre-anal fin is present. The pigment-corpuscles are few, but pronounced above and below the muscle-plates, and occur toward the posterior third of the tail. By the fourth or fifth day (Plate V, fig. 4) the eyes are pigmented, the yolk has disappeared, the mouth is widely open, the pre-anal marginal fin is long — with a pigment- spot in front of the vent, and some chromatophores have appeared in the marginal fin dorsal ly and ventrally. The Red Mullet-Family. Mullidae. THE RED MULLET. (Mullus surmuletus, L.) To those accustomed to the trammel-net on the shores of Guernsey and the Channel Islands generally, not even the numerous blue sharks and huge spiny lobsters that meet one under these circumstances are more interesting than the bright orange-red mullets that entangle themselves in the meshes. Their striking coloration and the eagerness with which their large lateral scales are torn off by the fishermen to impart to their prizes that reddish tint so popular in the market, impress themselves on one's memory. This fish has pelagic eggs (Plate I, fig. 1) having a diameter of O93 mm., a single large oil-globule with a diameter of 0'23 mm., and which in the tanks at Naples, Raffaele tells us, were shed in great quantity in early spring. The egg, further, is characterised by the very evident pores of its capsule, by the large vesicles of its yolk — most distinct at the surface, but which change position during the growth of the embryo, and by the proportionally large oil-globule, under which pigment is developed after fixation. Black pigment only, and that sparingly distributed, appears on the body of the embryo. The eggs hatch in three or four days, and the larva (Plate V, fig. 5) has a remarkable aspect 120 THE BERGYLT-FAMILY. for the yolk projects in front like a prow with the oil-globule at the tip, and the vent is close to its posterior border, while sparsely distributed black stellate pigment occurs along the dorsum and sides. By the second day the prow of yolk has been considerably abbreviated, and a larger space exists between the vent and the receding yolk, while the /"-shaped curvature which characterised the newly-hatched fish is now scarcely noticeable. The usual extension of the marginal fin forwards to the mid-brain and the appearance of larval sense-organs have also taken place. Between this and the 7th or 8th day the yolk has entirely disappeared (Plate V, fig. 6), and the black pigment has grouped itself chiefly along the ventral border of the muscle-plates with a very little on the side above, in the middle of the body, and a speck above and below the plates about a sixth from the tip of the tail, and also along the abdominal roof. The eyes are richly pigmented and re- splendent. In the middle of June, and in July and August, Raffaele also found young fishes (2 — 3 cm.) with a barbel and silvery aspect, for the roseate colour had not yet appeared. In the young examples described by Malm1 the profile of the head is almost like that of the cod, but as it increases in size the interorbital space rises so as to give it the charac- teristic form of the adult. Cunningham mentions that young examples, 3 inches long, occur in Plymouth Sound in summer. He considers them a year old. The Bergylt-Family. Scorpaenidse. THE NORWAY HADDOCK. (Sebastes marinus, L.) As one of the two British viviparous forms amongst bony fishes this species is of special interest, though few opportunities have been had of examining it, the only specimen in the collec- tion at St Andrews having been obtained from Mr Sim, of Aberdeen. It was brought from the Moray Frith, where it occasionally occurs. It is common off the Norwegian shores. 1 Goteborijs och Bohusl. F., p. 383, and Scanflin. Fishes, p. 63. THE BERGYLT-FAMILY. 121 Ekstrom and Smitt observe1 that the male is generally rarer than the female. The eggs are fertilized internally and developed on the walls of the ovaries. Collet considered the number of eggs in a female 550 mm. long (21'5 inches) to be about 148,000. Ryder, however, estimates the number of embryos in each ovary to be about 1000, and thus it is pro- bable that only a portion of the eggs arrives at maturity at the same time2. He also believes he had found an abundant covering of flat, fleshy and highly vascular processes which to some extent corresponds to the maternal placenta of verte- brates. Ekstrom frequently had one of both sexes forwarded to him, and therefore he was inclined to think the fish monogamous. He found the embryos far advanced at the end of May, while Kroyer stated they left their parent in July when they are from 3 to 5 mm. in length, and swam near the surface. In the same month (July) they reach from 9 to 19 mm. in length, and are captured in the surface-nets. They soon, however, seek the lower regions of the water. In an example procured from the Moray Frith the surface of the ovary was furnished with a vast number of villous processes, smaller and finer than those in the viviparous blenny. They were more or less digitate, lobed processes, which presented numerous small blood-vessels and various minute ova — ranging from '0762 mm., or under, to about '2286 mm. The larger forms of these projected from the surface either as sessile or pedicled processes, and their sub- sequent history is probably closely analogous with that of the viviparous blenny, the numerous long vascular processes, amidst which they were, having formerly carried developing ova from which the embryos had now escaped. The free embryos in the ovarian chamber were all of the same size, viz. 4'9 mm., and generally had the body bent round on the yolk, some of which still remained. The eyes were deeply pigmented, and black chromatophores occurred along the roof of the abdomen, dorsally along the edge of the muscle-plates, and ventrally to a less extent in the caudal region, the pigment of the dorsal line 1 Scand. Fishex, p. 152. 2 See Ryder, Bull. U.S. Fish Com. Vol. vi. (1866), p. 92. 122 THE SEA-SCORPION-FAMILV. almost coming up to that on the abdomen, while the ventral series is confined to the tip. The specimens were very in- differently preserved, but the marginal fin showed no fin-rays except in the caudal region, the tip of which had well-marked embryonic rays. The pectorals seemed to be little developed. The hyoidean and mandibular cartilages were fairly developed and the mouth open. So far as could be ascertained from this example there were no signs of a rapid succession of series of eggs, but, on the contrary, a wide interval evidently existed between the free embryos and the largest of the developing eggs, an interval probably extending over an entire season. The appearances of the villi pointed to their having similar functions to those in the viviparous blenny, viz. to supply amongst other things a nutritive pabulum for the embryos, and this before any special development of the intestinal canal of the embryo could take place. The larvas of Scorpcvna abound at Naples in June, but Mr Williamson did not make out the species. The Sea-Scorpion-Family. Cottidae. THE SHORT-SPINED SEA-SCORPION. (Coitus scorpius, L.) The short-spined sea-scorpion was one of the fishes very early examined in regard to spawning during the trawling work of 1884 and the following year. Comparatively little was known previously in regard to it. Thus Day, in his British Fishes1, states that, 'In Greenland it has been observed to deposit its eggs on the seaweed in December and January. Its eggs are very small, and in this country are extruded during the spring in the sand or pools in the rocks. The male is said to make a nest of seaweeds and pebbles for the re- ception of the spawn ; while he is believed to watch over, as well as protect, the young when hatched.' On the other hand, Prof. Alex. Agassiz records the ova of certain American Cotti 1 Page 186. THE SEA-SCORPION-FAMILY. 123 as pelagic, a feature very different from those of our country, and probably requiring corroboration. Whether he had the condition of Scorpcena with its ovoid floating mass of mucus and eggs in view is unknown. On the east coast, as at St Andrews, the eggs of this fish occur abundantly in March attached to stones, tangle-roots, old shoes, tin vessels, and, indeed, almost anything convenient. They are found again somewhat earlier (February) at Gairloch and other parts on the west coast. The authors of the Scandinavian Fishes broach the idea that the roe of this fish may be fertilised before de- position, and suggest that the serrations on the inside of the breast-fins may be useful to the males for this purpose. There is no reason to suppose that in Britain the eggs are so fertilised ; on the contrary, it is evident that they are not fertilised before deposition. As an example we may take a female observed shortly after the opening of the St Andrews Marine Laboratory1. This specimen, whose abdomen was distended, had been isolated in a glass vessel, so that its movements were somewhat limited ; and it is probable, there- fore, that the deposition may have been hastened. It had been observed to be somewhat restless the previous day ; and on the 1st March it rested quietly on the bottom of the vessel, and in a few seconds deposited a mass (as large as a duck's egg) of faintly pinkish eggs, keeping its breast-fins in active motion during the process, and then it dashed through the water, sending some of the eggs over the edge of the vessel. The mass of eggs, at first quite soft, though cohering to- gether by a secretion, soon harden, the capsules adhering by facets to each other as in the lump-sucker, so that the egg-mass (Plate I, fig. 6) resembles a spongy structure into which water freely enters, and is retained in considerable quantity even though the eggs are uncovered by the tide, a provision of some importance. They vary in colour from that first mentioned to roseate, orange, straw-colour, and deep red, and have a diameter of about 1'5 mm. to 2 mm. (Holt). The capsule is thick, tough, and resistant, and shows the facets or processes by which it adheres to surrounding eggs (Plate I, 1 W. C. M. 3rd Ann. S.F.B. Report, p. 59, and Uth Ann. Report, p. 181. 124 THE SEA-SCORPION-FAMILY. fig. 5). It appears minutely punctured, under a high power, the punctures having, as a rule, a more regular (linear) arrange- ment than in the lump-sucker. Moreover, larger dots occur at intervals all over the surface, resembling those seen in the lump-suckers' eggs removed from the stomachs of young cod. The yolk internally has several colourless oil-globules, from three to nine, as mentioned by Mr Holt, and they vary in size from '015 mm. downward. The yolk itself is tinted pale brownish or faintly reddish-brown. Mr Holt, who carried on special observations on the eggs of this species at St Andrews, could not make out the passage of the oil-globules through the yolk, as had been described by one of us in the gurnard ; but, so far as observed in 1884, there was no reason to doubt that the oil-globules followed the same movements as in other forms. In the developing embryo the oil-globules coalesce, so that but a single large globule remains. In the tanks of the laboratory the eggs are readily eaten by other specimens of the same species. The development of this form is somewhat slow, especially in cold seasons, so that masses of eggs with advanced embryos are often found in April and even in May. Mr Holt found the larval fishes, on emerging from the egg, 7'5 mm. in length, and this accords with our own experience. The yolk forms a com- paratively small prominence ventrally, and the large oil-globule lies at its front inferiorly. ' The head is large and broad ; the profile of the snout abrupt ; the eyes large and fully pigmented ; and the ear-capsules, of about the same size as the eyes, lie close behind them. The mouth is open, but the lower jaw is at first immovable.' The internal organs are well developed (Plate V, fig. 7). The tail in May shows only embryonic rays. The breast-fins are large and fan-shaped. The heart and blood-vessels are in full activity, the returning blood streaming over the yolk, and finally entering the heart. The oil-globule is often on the right of the middle line. The coloration consists generally of an olive-green hue, a series of distinct black chromatophores over the head, and a few about the base of the breast-fin. They form a broad band on each side of the abdomen over the yolk, and extend THE SEA-SCORPION-FAMILY. 125 from the breast-fins to the vent. ' Pigment of a bright yellow colour by reflected, and orange by transmitted, light occurs also at the base of the breast-fins, on the top of the head, and on the abdominal roof.' The eyes are black with a greenish iridescence. ' In the post-anal region the only pigment is a ventral line of black chromatophores, sometimes very small or absent in the anterior region, and ceasing before reaching the tail.' The vitality of the larval sea-scorpions is remarkable. They will survive for a fortnight in March, in a small quantity of water, in a glass vessel 2 inches across and 1 inch deep. On the 10th October, on one occasion, a larval form resembling a sea-scorpion was captured in the tow-net. In general outline it resembled that figured in the 'Researches,' with the vessels coursing over the yolk-sac. The oil-globule remained at the anterior part of the yolk-sac. Small specks of black pigment occurred along the sides of the body, one set forming a row near the upper-lateral region. No distinct coloration was visible on the pectorals. The eyes were iri- descent-greenish, like the inner surface of the shell of Haliotis. Hitherto it has been unusual to get larvae at this season of the year, so that the deposition of such eggs must have been antedated by some months on this occasion — if the inter- pretation of the nature of the larvae be correct. Swarms of the early post-larval sea-scorpions, indeed, just after the absorption of the yolk, about 7 to 7*5 mm. (in spirit), are occasionally captured in the surface tow-nets, as in the Forth, e.g. in March and April. Such pelagic forms have only embryonic rays in the tail-fin. The body and tail are trans- lucent, whilst the head and abdomen have a pale greenish hue with black chromatophores, and the eyes have a silvery lustre. A line of black pigment-specks runs along the ventral edge of the muscle-plates behind the vent almost to the tail. Mr Holt1 observed that the lower jaw is movable two days after hatching, and that the vent is open. In our examples, a thickening below the axis of the tail occurred. When six days 1 E. W. L. Holt, S'c. Tranx. Roy. Dab. Soc., v. 2, p. 21, &c. 126 THE SEA-SCORPION-FAMILY. old the length is 8*4 mm., and the yolk has diminished, while the oil-globule has been elevated to the gullet. The young forms were kept in the tanks till the 17th day, but development proceeded slowly under the somewhat un- favourable circumstances. On the 7th day they measured 7 '7 mm., and during the two or three subsequent days the pigment made great progress, extending behind the vent, and passing from the dorsum down the sides. The chief changes are the straightening of the ventral line from the head to the vent, from the diminution of the yolk, the disappearance of the oil-globule, the increase in the length (forward) of the lower jaw, and the presence of a distinct and broad pigment- band on the side of the body a little behind the vent. The notochord is still perfectly straight at the tip of the tail, and the circulation is much as before. On the 10th day the yolk had disappeared, but the embryonic fin-rays were present only in the tail. The absence of food would, as Mr Holt suggests, suffice to explain the slow progress, but not altogether, since, in the open sea, specimens of 9'5 mm. are occasionally procured in a similar condition, viz., having a membranous dorsal and anal, and only embryonic rays in the tail. Those in confine- ment differed in having no trace of the ventral fins, and the thickening beneath the tail was better marked in the free forms ; spines on the gill-cover were also present. In specimens of 7 mm. (in spirit) captured in the bottom- net in the bay, the head has much increased in size, the fish is thick-set, and the gill-cover has minute spines. Embryonic rays are well developed in the tail, and a thickening occurs beneath the central axis. The black pigment has largely extended along the dorsum to a line behind the vent, and it is more abundant on the head. Such a form contrasts with the slender and ill-nourished specimens reared in the tanks, in which the abdomen was shrunken and the end of the gut distended, as if the vent were closed. The breast-fins, however, were large. At the end of April and beginning of May, pelagic forms of from 9'5 to 10 mm. are not uncommon at the surface, e.g., off the Isle of May. The body has now considerably increased in THE SEA-SCORPION-FAMILY. 127 bulk, and a series of sharp spines project from the gill-cover, and two on each side of the occiput. The larval tail is present, but it is bent upward by the development of the true rays inferiorly. The ventral fins appear as minute processes. In contrasting such a stage (e.g., one 9 mm. in length) with a gadoid, it is easily distinguished from the latter by its shorter snout, smaller mouth, and smaller eye, as well as by the deeper greenish pigment, with a trace of yellow on the head and abdomen. Moreover, the latter is much more densely and somewhat regularly spotted with blackish pigment, the whole having a tessellated aspect. Further, from the greater tenacity of life in the sea-scorpion, the body does not so soon assume the whitish opacity so characteristic of the gadoids ; indeed, though perfectly motionless, the heart may be pulsating. The blackish pigment again is confined to the ventral edge instead of passing along both dorsal and ventral edges, as well as some distance up the sides, as in the gadoids. Except in the tail, the young gadoid of the same size has only embryonic rays in the continuous marginal fin, while in the sea-scorpion a considerable number of rudimentary true rays occur both dorsally and ventrally (10 or 11 dorsally and 6 ventrally). In the anal fin the true rays commence anteriorly. Those in the dorsal begin just above the latter (that is, a little behind a perpendicular line from the vent). No permanent rays appear in the dorsal and the ventral marginal fin of the gadoid, even though the example exceeds in length the cottoid of this stage. A very evident difference exists in the tail of those of equal length. Thus, the upper and lower elements (epiural and hypural) are more or less equally developed in the gadoids ; the ventral series, however, terminating in one or two larger cartilages. The tapering notochord (elementary back-bone) is straight, and extends considerably beyond both series. True caudal fin-rays, moreover, are developing both dorsally and ventrally, giving the tail a peculiarly symmetrical or ' feathered ' appearance. On the other hand, the notochord in the sea-scorpion is somewhat less finely tapered, has a thicker sheath, and the 128 THE SEA-SCORPION-FAMILY. inferior (hypural) elements alone are conspicuous in the form of a large inferior and two upper cartilages. The permanent caudal rays are developed only inferiorly, while the whole dorsal half, and the region extending to the last ray of the dorsal fin, have embryonic rays. At 11 mm. many are still pelagic (17th May), and show the three anterior gill-spines, — the occipital, superciliary, and nasal spines (Plate V, fig. 10). The ventral fins are minute. The pig- ment approaches that of the adult stage, only it is not so largely developed. The larval tail is at the upper edge of the organ, and the marginal fin is continuous and has only embryonic rays. When from 14 to 18 mm. they .still occur as pelagic fishes in the tow-nets. The head and body are now larger and more deeply pigmented, the former being entirely covered and con- tinuous with the dorsal pigment, which passes downwards to the cheeks and chin. A bold bar in many exists at the base of the breast-fins, another across the region of the first dorsal fin, and one at the second dorsal, the latter, moreover, extending down- ward on each side to the ventral edge. The pigment of the two latter bars in some is specially dense, though in others the tint is more uniform dorsally. The head is cottoid in appear- ance, the superciliary ridge and the occipital tubercles with their spines being conspicuous. Three of the spines on the gill-cover are large, the fourth at the inferior edge being small. The boldness in the demarcation of the pigment gives the fishes a piebald aspect in spirit (Plate V, fig. 8). The larval tail is represented in the smaller forms by the upturned central axis (notochord). True rays now occur in all the fins. The voracity of these young forms is remarkable. One of 16 mm., for instance, in captivity swallowed a young flounder not much shorter than itself, just as the larger examples cleared the young gunnels out of the tanks. At 22 mm. (27th May, estuary of the Eden), the bar of pigment behind the vent has sent a process backward to the tail, but it goes no further than the basal region. Symmetrical white spots — one dorsal and two ventral — occur in this pro- longation. The occipital and supra-orbital tubercles are less prominent, but the supra-nasal are distinct. A large spine THE SEA-SCORPION FAMILY. 129 occurs on the gill-cover (operculum). The upper spine on the gill-cover (pre -operculum) is largest. The first dorsal fin has 9 rays, the second 15, the variegated pectoral 16, the ventral 3, and the anal fin 13. The caudal has 12 long rays besides four or five shorter at each edge. The chief difference, there- fore, between this and the adult is the increase in the caudal rays, but the short basal ones probably disappear during growth. One specimen in July reached 2 inches. In June they reach 23 to 24 mm., and in July 38 mm., (1^ inch) with adult characters, the first dorsal having 9 rays, the second 16 rays, and the anal still constant at 13. In the same month one of 88 mm., also in its second year, was captured in Guernsey. In September specimens 54, 65, and 85 mm. (nearly 3| in.) occur, the first being considered by Mr Tosh as the young of the season at 5i months. It would be difficult to separate that at 65 mm. from the same series, but that at 85 is considerably older, probably by a year (or, as Mr Tosh puts it, 1 year and 3 months). The usual arrangement of the spines on the gill-cover in such forms is as follows: — A spine points down- wards at the ventral edge of the pre-operculum, two short spines occur above, then the upper long spine, above which is the opercular spine. Two short spines appear on the sub-operculum. Those of 57 (2^ in.) and 74 mm. (3 in.) in February repre- sent specimens about a year old, while those of 98 mm. are in their second year or approaching it. One of these (98 mm.) had almost ripe eggs, so that the remark in the new edition of the Scandinavian Fishes, that it does not propagate its species until about 150 mm. (nearly 6 in.) long, is not applicable to our country. LONG-SPINED SEA-SCORPION. (Coitus bubalis, Euphras.) The eggs of this form (Plate I, fig. 4) were alluded to and figured in the Researches, and they have been frequently observed since. Moreover, Mr Holt had an opportunity subse- quently of studying them at St Andrews Marine Laboratory, and his careful observations and those of Mr Cunningham have M. F. 9 130 THE SEA-SCORPION FAMILY. been drawn upon for the present purpose. The eggs are found between tide-marks in April amongst tufts of sea-weed, or on ledges of rocks, so as to be partially exposed at low water. The masses are usually flattened, and of a size less than the palm of the hand. The diameter of the egg is 1*70 to T88 mm. The capsule is peculiarly corrugated or minutely nodulated, and the yolk has a golden yellow or straw-colour. The oil-globules are numerous, as shown in the figure. The newly hatched larva (Plate V, fig. 9) is smaller than that of the short-spined bull-head, measuring about 57 mm. The choroidal pigment of the eye has a deep blue colour (Cunningham). Yellow pigment occurs on the head and the pectoral region. No black pigment exists on the top of the head, or along the ventral edge behind the vent. The pigment along the roof of the abdomen is peculiarly dark, and has quite a different aspect from that of the species above mentioned. Mr Holt has likewise pointed out that the circulation of the yolk-sac is more complex than in the other species, the blood leaving the liver in at least two and generally in three vessels. The oil-globule lies behind the heart in front and has a diameter of -22 mm. In two days the length in confinement has increased to 6 mm., and the yolk is reduced. Black chromatophores occur along the edge behind the vent, which is open, and canary-yellow pigment amongst the black on the roof of the belly. On the fourth day embryonic rays appear in the tail-fin ; the length is 0'28 mm., and the oil-globule is reduced. On the sixth day the yolk is still further diminished. Black corpuscles have appeared in the covering of the mid-brain, internal to the ear-capsules and round the vent. The gills show papilla?. At the tenth day the length reaches 6'42 mm., chiefly from increase in the tail, the jaws move freely, an opercular fold is present, and the gills are pectinate. The liver is much larger. Black chromato- phores form a conspicuous median dorsal line over the anterior part of the abdomen, and similar pigment appears on the pre-anal fin. At the 20th day in confinement, increase in length was found by Mr Holt to be ml, probably from want of food. All THE SEA-SCORPION FAMILY. lol traces of the yolk and oil-globule have vanished. TLe breast- fins are very large fan-shaped organs. A thickening (hypural) beneath the notochord causes a slight upward bend of the tail. A single process in the upper opercular region (hyomandibnlar) indicates the pre-opercnlar armature of the adult. A specimen of 7'3 mm. procured south-west of the Bell Rock on the 8th August, 1888, may be a post-larval example of this species. The body is short and thick, the two dorsal fins are out- lined, but still connected with each other and with the tail by the larval fin. The caudal fin is well formed, but a trace of the larval tail exists superiorly where the tip of the notochord also appears. The breast-fins are large. The ventrals form minute processes. The front gill-cover (pre-operculum) has about four spines. The posterior occipital tubercles have spines, the pair in front do not show them. No turbinal spines are yet visible. The specimen is much shorter than the short-spined Cottus of the same age, and the body is deep and somewhat flattened. The identity of this form is still doubtful. The subsequent stages of this species are still in need of observation. Mr Holt found a form of 10'5 mm. which he doubtfully refers to it. Though it has two pairs of small tubercles connected by longitudinal ridges on the top of the head, it does not appear to be a young Cottus quadricornis, since both turbinal and supra-orbital spines are absent, and therefore Mr Holt's suggestion is probably correct, viz. that it is a stage in the development of the long-spined Cottus. In the Irish expeditions of 1890-1891 Mr Holt took a young example of 21 mm. in Jeelin harbour on the 19th May, a fact which shows that the spawning-period is considerably earlier on the West coast of Ireland. On the East coast of Scotland this species spawns somewhat later than the short- spined bull-head — if the interpretation in regard to the eggs and larvae, by Mr Holt and ourselves, is correct. In the recent Scandinavian Fishes the spawn ing- season is stated to be the end of November and in December ; Mr Cunningham again at Plymouth found the period of deposition to be from January to March. In August a specimen 44 in. long occurred at St Andrews. 9—2 132 THE SEA-SCORPION FAMILY. FOUR-HORNED COTTUS. (Coitus quadricornis, L.) In the recent edition of the Scandinavian Fishes1 the spawning season of this species is given as November, December and January in the Baltic. " The roe is deposited, says Sunde- vall, like that of the Perch, in one single mass ; but this is attached to the bottom in water of some depth, possibly even several fathoms. From a piece of roe which Baron Cedestrom found in a seine, the young were hatched during the latter half of April. They were then about 11 mm. in length, and their external and internal organs were far more developed than is generally the case in the fry of other fishes. They swam about freely, but soon sought shelter in the roe from which they had emerged." This account is somewhat remarkable — especially in regard to the size of the larval fishes on emergence, and fresh investigations are required. Post-larval stages of this species have occasionally been obtained in the mid- water- and other nets in and off St Andrews Bay during the last ten or twelve years, but the eggs have never been accurately differentiated if they have ever been obtained. When engaged in investigations at this laboratory in March 1890 Mr Holt concluded that a cottoid larval form which was familiar to us from the mouth of the Forth and 8t Andrews Bay might belong to this species, since the larvae of the short- and the long-spiued were already known. A certain amount of doubt remains, especially as the eggs have not yet been identified, but as the larvae have not been linked to other forms, and as no centra-indication is apparent, the suggestion may be allowed to stand. Mr Holt's youngest examples, which were 3 mm. in length and had considerable oil- globules, were dead. A somewhat older form from St Andrews Bay measured 4'08 mm. and much resembled the father-lasher. " The globular urocyst, and the separation of the coalesced ureters from the immediate neighbourhood of the gut appear to be features essentially characteristic of the two forms in their late larval stages." A brownish-yellow mass in the anterior 1 p. 179. THE SEA-SCORPION FAMILY. 133 region of the abdomen represents the yolk with its single oil-globule (now minute). A feature which from the first was most diagnostic was the dense black pigment of the lining membrane of the abdominal roof, and it is outlined by a fringe of yellow. A single ventral line of black chromatophores occurs behind the vent. Forms somewhat older were caught in the bottom nets at the beginning of May measuring 4'8 mm. (in spirit), and with well marked embryonic rays in the tail. In these as well as in the younger forms a curious separation of the black chromatophores of the abdominal roof takes place. Mr Holt procured specimens of 6'25 and 7 mm. off the West coast of Ireland ; while one of 8 mm. was formerly described in the Researches1. The deep black patches of abdominal pigment were very conspicuous. The head is greenish and the body comparatively pale. The eyes, as in the common form, are bluish, with a remarkable St Andrews cross radiating from the pupil, the long axis being horizontal (Plate V, fig. 11). Stellate black chromatophores occur on the under surface of the abdomen, a touch of the same pigment at the anterior region of the branchiostegal rays, and a row runs along the ventral edge of the body above the anal fin. One or two specks are also present on the cheeks, and a considerable number over the brain, the latter being bounded by a curved line which joins a median black band in front of the dorsal fin. The four tubercles on the head are distinct, the posterior pair being the larger. The turbinal spines are not visible ; but the four on the anterior gill-cover (pre-operculum) are well-marked, the superior being especially evident. The first dorsal fin is only slightly arched, the second is continuous posteriorly with the larval tail-fin, which now lies at the upper angle. The permanent rays give a somewhat conical shape to the tail ventrally. The anal is likewise joined to the caudal. The breast-fins form fan-shaped organs, the rays passing close to the surface of the body. The rays are massive though soft, and, as in the adult, present considerable free portions at the tip. The ventrals are small, and arise somewhat behind the bases of the pectorals. 1 Trans. E. S. E. p. 862. 134 THE SEA-SCORPION FAMILY. Another example measuring about 11 '5 mm. was procured on the 20th June. A considerable amount of black pigment occurs on the head and cheeks, and a dark area exists at the first dorsal fin, and two behind, as in the common form (short- spined Cottas). The breast-fins are large, reaching behind the middle of the body. The ventral fins are still comparatively short. Four prominent spines exist on the front gill-cover (pre-operculum). The four tubercles on the top of the head are well marked, besides a supra-orbital and a turbinal spine. On each side of the dorsal fins anteriorly is the series of elevated scales which are continued backward to the termination of the second dorsal. A somewhat smaller example (10 mm.) is care- fully described by Mr Holt1. THE RED GURNARD. (Trig la cuculus, L.) This species, which is not common on the North-East coast, though frequent on the southern shores, both east and west, spawns about the same period as the grey gurnard. Its eggs have been described both by Mr Cunningham2 and Mr Holt3, and have probably been occasionally procured in the tow-nets off the Forth and the neighbourhood. The eggs (Plate I, fig. 10) measure from T47 mm. to T61 mm. — a considerable variation, while the oil-globule ranges from "30 to %33 mm., and is copper-coloured. In the earlier stages, several oil-globules are present, but by-and-by they coalesce, so that, on the escape of the larva, only a single globule remains, and its colour gradually fades. Cunningham was successful in hatching the eggs, and found the larva to be 37 mm. long. It bears a general resemblance to that of the grey gurnard, though the oil-globule is not thrust so close to the posterior border of the yolk as in the latter species. The pigment (black and orange or yellowish orange) appears to correspond very closely in the two species. 1 Tram. E. D. S. v. 2, p. 119. - Jour. Mar. liiol. Assoc. No. 1, p. 12, 1889. 3 Sc. Trans. Roy. Dub. Soc. v. 2, p. 31, 1893. THE SEA-SCORPION FAMILY. 135 THE SAPPHIRINE GURNARD. (Trigla lucerna, L.) Though by no means a rare fish, the eggs have hitherto received little attention. Day is of opinion that the breeding- season is during the first six months of the year. Couch found mature ova in December and February. Risso gives the spawning-period as spring. In July (18th) the ovaries were much enlarged, the eggs having an average diameter of •031 inch. Even so late as the 14th September some ripe eggs have been found in the ovaries at St Andrews. THE GREY GURNARD. (Trigla giirnardus, L.) Amongst the pelagic eggs of numerous species found on the East coast, that of the grey gurnard is perhaps the most conspicuous, not only because of its occurrence in great numbers but also because it has a wide distribution in point of time. Though not one of the earliest to put in an appearance, it is found fairly frequently in April and in some parts even earlier, and from this month onwards through the summer to the end of August it increases in numbers, so that in June it is by far the commonest egg, though that of the sprat is found with it in great quantities. From this it will be seen that the spawning- season of the gurnard extends over a considerable period, a fact which is further borne out by an examination of the structure of the female reproductive organs. In this fish, as in the cod and some others, the ovaries ripen by degrees, so that very rapid extrusion of all the eggs is not likely, but they are dis- charged in sections as they become ripe. Hence if a female gurnard be examined in May, the ovary will show a number of ova in different stages of development, and, as one observer states : — " In the gurnard the gradual process of ripening and the co-existence of perfectly mature and microscopic ova is even more marked than in the whiting, and shows that the spawning- process is a prolonged one," though it is possible that the microscopic forms are not all utilised the same season. A single ovary contains roughly about 120,000 eggs, so that 136 THE SEA-SCORPION FAMILY. the gurnard is not one of the most prolific of the ' pelagic ' fishes. The male organs in such as the sapphirine gurnard, men- tioned by Mr Holt, are proportionally large, so that some resemble females by the distension of the abdomen. As regards the spawning-habitat the gurnard often seems to frequent the inshore waters. Mature gurnards are most abun- dant just outside the three-mile limit, though they are also found in the shallower parts of St Andrews Bay and high up the Frith of Forth. The distribution of the pelagic eggs them- selves, as found by surface-netting, agrees with these facts, the eggs of the gurnard and the sprat being found further up the Frith than any other pelagic eggs. In the autumn the ' spent ' fishes appear again to migrate seawards, so that the migratory habit of the gurnard is to some extent the reverse of that of the cod and the haddock. One observer states that the gurnard spawns twice every year, once in the winter and once in the summer, but although there is an isolated case of the egg of the gurnard occurring in Moray Frith in January, yet an examination of the females, as above alluded to, gives no indication of spawning taking place twice every twelvemonths. The females of this species are considerably larger than the males, a feature in which they agree with the cod and haddock. In the salmon the same occurs, the difference in this case being due to an earlier maturation of the male : whether the difference in comparative size of the two sexes of gurnards is due to the same cause cannot be stated with certainty. The egg of this fish (Plate I, fig. 9), which closely resembles that of the red gurnard1, is large, being T52 mm. in diameter, with a large ('28 mm.) and conspicuous oil-globule of a smoky or sometimes salmon-hue. It is less delicate in appearance than those of the cod, rockling or flounder. The perivitelline space is often fairly large before the egg is fertilised, but after this process the germ usually swells up and comes in close contiguity with the egg-capsule. The period of incubation varies from eight to fourteen days 1 J. T. Cunningham, Join: Mar. Rial. Assoc. I. iv. s. p. 11. THE SEA-SCORPION FAMILY. 137 in May to six days in June or July, the greater mean tem- perature of the latter having the usual effect of hastening the development. The yellowish pigment is developed both on the body and the yolk-sac in the egg before hatching, but it appears somewhat later in the embryo in May than in the gadoids and flounders. It consists of pale yellow spots which have a delicate sea-green tinge in certain lights. They are sparsely scattered over the trunk proper, but form a rude line along the back, and an undulating line along the sides, around the eyes and over the yolk-sac. The pigment in the latter is preceded by pale protoplasmic stellate corpuscles. The young larva when just hatched (Plate V, fig. 12) is a glassy transparent form with a large yolk-sac, at the posterior border of which is seen the tinted oil-globule, surrounded by a thin layer of protoplasm. It measures about 3'S mm. The larval fish is conspicuously marked with pigment of two kinds. The dull yellow, sometimes greenish, spots already referred to in the embryo, are scattered over the head, back and abdomen, but are absent from the tip of the tail. The front (dorsal) part of the embryonic fin is picked out, at its edge, by yellow spots, with which are mingled a few black ones, forming a row which invades the fin further behind and meets the pigmentation of the body just in front of the tail. There is a similar but less distinct line on the ventral part of the marginal fin. The pectoral fin has a peculiar arc, of black and yellow stellate spots, which is very striking. Other larval fish described later on have the same banded arrangement upon the pectoral fins. The whole of the black pigment is later in appearance than the yellow, and then to some extent follows the lines of the latter colour on back and fins and is diffusely scattered over the body and yolk-sac in minute black dots. Lastly, the pigment, besides being distributed over the surface of the larva, as detailed above, spreads out over the surface of the yolk, which presents scattered yellow and black chromatophores. In this feature the gurnard agrees with the whiting, sole and ling, whilst, on the other hand, it disagrees with the cod, haddock and rockling. The pigment-corpuscles 138 THE SEA-SCORPION FAMILY. appear to be really in the protoplasmic layer enveloping the yolk, and differ from those found on the body of the larva by the fact that they are much branched, and often anastomose. A good deal of individual variation in the amount of the pigment exists, probably because some are hatched at a slightly earlier stage than others. The larva, as in other species, hangs in the still surface-water of vessels with the yolk-sac uppermost, often with the head downwards. During the next three or four days the pigmentation increases, the individual chromatophores become branched, and the yolk-sac rapidly diminishes. One remarkable feature in the early stages of the grey gurnard is the great development of the pectoral fins, as, indeed, Cunningham also found in the red gurnard. This is evident upon the second day, and it is accompanied by a marked decrease in the region between them and the embryonic ears. The snout also becomes prominent, and the mouth is open from the first, although the gullet does not acquire a lumen till later. At 5 days the little fishes have become active and are about '165 inch long. The eyes have black pigment, with a greenish lustre. The yellow colour is conspicuous about the head, the yolk-sac, the pectoral fins, the anterior dorsal region and the angle of the lower jaw. The black colour is found in abundance at the base of the abdomen, and a few spots occur on the snout and ventral border of the muscle-plates. The pectoral fins have now developed into a pair of large fan-like paddles projecting at right angles to the body and are concave backwards, forming efficient organs of locomotion. During the following days there is an increase of the black pigment and a still further hypertrophy of the pectoral fins. The mouth still gapes widely, the gurnard being a fish which essentially develops with its mouth open, a habit which gives the larva a diagnostic appearance, and no doubt subserves respiration to a great extent when combined with a rapid forward movement of the body in the somewhat quiescent water of the vessels in the laboratory. The last stage reached in confinement is about 3 weeks old ; in this we note the huge pectoral fins, the yellow pigment on THE SEA-SCORPION FAMILY. 139 the latter, the head, and the yolk-sac. The marginal fin is still continuous, and a few pigment-spots are dotted over it and upon the tail. The next stage, about -235 inch long, is represented by specimens caught in the sea on August 31st, which clearly belong to the same series. The head, with eyes and brain, is large, and the scoop-shaped indented snout is prominent. The yolk-sac has nearly disappeared. The pectoral fins have still further increased and the tail shows signs of becoming heterocercal by the greater development of the lower rays. By this time the little fishes have, even though so small, left the surface-water and taken up the habitat of their older brethren, namely the still water of the open sea at a depth of about 25 fathoms. With a further growth in length, the head and snout become proportionately larger and longer, and the little gurnard is protected by the development of a number of conspicuous spinous processes which make it a prickly morsel for predaceous enemies. Two spines on the back of the head first appear, and later, others are seen upon the operculum, the angle of the lower jaw and the facial region. In the large pectorals there are thirteen fin-rays all joined by membranes, but as the fish progresses in size the three most ventral rays separate from the others and eventually become free, both from the pectoral fin and from each other. They form the three pairs of moveable ' feelers ' which are a well- known feature of the adult gurnard, and which have been carefully investigated at St Andrews by Mr H. C. Williamson1. This is a good instance of the repetition, in the development of the individual, of a modification in structure which the species has adopted in the course of time. Thus it is believed that the gurnard is descended from ancestors which had all the fin-rays united together as in most fishes, but that the three ventral rays, being gradually specialised as feeling organs, became separated from the rest until at last the free condition of these rays, in the gurnards of the present day, was reached. Hence we find that the young gurnards pass through the 1 Vide 12th Keport Fishery Board for Scotland, p. 322, Plates 13—15. 140 THE SEA-SCORPION FAMILY. same stages on their way to the adult condition. Instances of this law, even more striking, will be noticed in the life-history of the flat fishes. It is still an undecided point as to how far the gurnard uses its three pairs of feelers only as tactile organs, or combines with this function that of supporting the weight of the body when at rest, and of assisting locomotion. Returning to the young forms of gurnards, fig. 13, p. 57 shows a post-larval stage of about f inch in length. The tail is markedly heterocercal and the enormous pectoral, and small ventral, fins are noticeable. The marginal fin has a dorsal indentation, which indicates the future splitting up into first and second dorsal fins, though only embryonic rays are present at this stage. Black pigment has increased over the pectoral fins, head, cheeks and abdomen. At this and later stages there is often to be found attached to the head a young form of the parasitic crustacean, Caligus. Further development proceeds in the direction of the break- ing up of the marginal fin to form the caudal, the dorsals and the anals, and the appearance of adult fin-rays ; the ventral fins also elongate till they reach beyond the vent. A post-larval gurnard at this stage (Plate V, fig. 13) is one of the most grotesque little animals which one meets with, its long angular snout, large greenish eyes, huge pectoral fins and numerous little spines all adding to its unique ap- pearance. The huge pectoral fins form a drapery for the entire body when folded back, only the tip of the tail extending be- yond them. They are indeed proportionally as large as in the southern flying gurnards, but in these the fins reach full de- velopment only in adult life, while in the young stages they are comparatively small, exactly the reverse happening in the grey gurnard of our seas. At f inch, the black and brownish pigment-spots, dusted over the pectoral fins, may be noticed to have a tendency to group themselves into three transverse bands which later be- come very marked. At a little over | of an inch, the little gurnard assumes THE SEA-SCORI'ION FAMILY. 141 the leading adult characters (Plate V, figs. 14 and 15). The three pectoral rays are free; the general pigmentation has increased, probably correlated with its change of habitat to the shallower inshore waters ; while the pigment upon the pectoral fins is conspicuously disposed in three bands, a basal and two distal, and that on the hinder part of the body forms a pair of prominent V-shaped vertical bands. The head arid dorsal fins are also boldly marked. Later, the spines on each side of the dorsal fin and on the lateral line are distinct, and the membrane connecting the base of the three free pectoral rays gradually disappears. During September and October little gurnards from \ to £in. in length are found, and again at the end of October a specimen about 29 mm. or 1£ inch occurs. In this specimen the pigment shows the same general arrangement as in the last (Plate V, figs. 14 and 15) but is more abundant. The tail has brownish pigment-spots dusted over its distal half and a thick mass of brown pigment at its base, and these two areas are divided by a broad transverse band without pigment, so that the general effect is that of two dark bands and a light one crossing the rays of the tail. The back and sides have blotches of brownish pigment and that on the dorsal fins is black. A dense black blotch is found on the first dorsal, and two faint black longitudinal lines run on the second dorsal fin. The barred condition of the pectorals is no longer evident, for the deep black pigment covers the whole fin. The pigment has increased under the eyes. Four longitudinal rows of spines running down the trunk are now conspicuous, two on the back, on either side of the dorsal fins, and two along the lateral line. The pectoral fins now bear a smaller proportion than here- tofore to the size of the fish and approach those of the adult. A young gurnard of 2 inches caught at the end of November shows the black pigment confined to the distal half of the pectoral fins, and the first dorsal fin has the same change, the black pigment being now concentrated in one large spot. The pigment of the body is also more diffuse and closely resembles that of the adult. A specimen almost identical in size and appearance occurs 142 THE SEA-SCORPION FAMILY. in June, together with two others, 3 in. in length. In these the first three rays of the first dorsal protrude beyond the fin-membrane as in the adult ; they are rigid and sharply pointed. The depth of the body is greatly increased and the head is smaller in proportion. The next stages occur in July, still on the bottom, and are about 3 to 3^ in. long. Their appearance and structure are practically those of the adult. As regards the rate of growth the little gurnards appear to grow fairly rapidly, and by June of the following year two series are noticed, some ranging from 2| to 3 inches and others from 4f to 6^ inches. Another series is represented by little fishes ranging from 4£ to 61 inches in length and oc- curring a month earlier than the above (i.e. in May). At present we have no definite means of confirming the supposition that these are all fishes which were hatched during the previous season ; but this is probably the case, in spite of the great disparity of size. The latter feature may be largely due to the lengthened period during which the gurnard spawns. Thus, larvae hatched at the end of April might easily be three months older than those hatched at the end of the spawning-period, which, placing them under more advantageous circumstances for their earlier development, might, apart from the mere difference in age, cause great disparity of size in the following summer. More data are however required before one can pronounce with certainty upon this point. Thus, at the end of August, certain young gurnards range from ^ to f inch and of course differ in point of develop- ment to the same degree. They are all of the same season's growth, but whereas the former are a few weeks old, the latter must have been hatched in the earlier part of the spawning-period. Again at mid-September specimens of 1^ inch are found, which must also be regarded as of the same season's brood. We are thus presented with a series of gurnards at the end of summer ranging from J inch to 1^ inch or more, which vary from an age of a few weeks to one of nearly six months (April to September). Taking this series as a basis, those in May referred to THE ARMED BULL-HEAD FAMILY. 143 above (4£ to 6i inch) would be from 10 months to 13 months old, whilst the series of 2f to 3 inches occurring in June may be from 9 to 10 months and the second series in the same month (4| to 6^ in.) would probably contain individuals from two broods, the smaller being about 14 months and the larger 26 months. There is still uncertainty, however, on these points. The curve of growth from specimens measured by Mr Williamson, of this Laboratory, gives an average length of 4 inches at the age of one year and of 6| inches at two years, which is increased to 9 inches or more at the end of the third year. It is probable that the gurnard, like most food-fishes, comes to maturity during its third year or a little later. The number of females appears to be in excess of the number of males in the proportion of 4 to 1 and the male is, on an average, slightly smaller than the female. The Armed Bull-Head Family. Cataphracti. THE ARMED BULL-HEAD. (Agonus cataphractus, L.) No form is more familiar in sandy bays in the naturalist's, or small-meshed, trawl than this species, yet its spawning was more or less unknown until recently. Couch gives nothing more than that the spawning-period is stated to be spring ; while Day describes a female, 5 inches long, from Southend, in February, as having its comparatively large eggs nearly ripe. In the Researches1 it was mentioned that the females caught in the sprat-nets in the Tay showed nearly ripe ovarian eggs on the 16th December, and that they had a dull golden or dull yellow colour, their diameter being '07 in. and that of the oil-globule 0216 in. The capsule is minutely dotted with punctures arranged in linear series, and the surface has well- marked reticulations (Plate I, fig. 8). It is added that the species seems to spawn from January (or perhaps December) to April, females with nearly ripe ova occurring in March. 1 Mclntosh and Prince, loc. cit. p. 674. 144 THE ARMED BULL-HEAD FAMILY. The eggs are often deposited on the rocks at low water-mark in straw-coloured masses. In the recently published Scandinavian Fishes1 the authors observe that " the males are so rare that neither Kroyer nor Ekstrom has met with a full-grown example ; nor, during late years, has the Royal Museum received a single specimen. It is probable that it is only during the spawning-season that they live in so shallow water as to be in any danger from the nets used in shore-fishing. The spawning- season is in spring ; in March and April, or the beginning of May. We have no information as to the way in which the roe is deposited, or the development of the fry." It is also mentioned that Kroyer estimates the number of eggs at 3,000. So far as our observa- tions go the males are by no means rare, and they accompany the females into the shallow water, and pass into the estuaries of the Eden and the Tay. The remarks published on the subject have escaped the notice of the authors. The occurrence of small masses of ova of this species, at the beginning of October, considerably extends the spawning- period. They are procured at low water near the pier-rocks at St Andrews, attached to the roots of tangles and passing into the interstices. The eggs measure from 1*7526 mm. ('0705 inch) to T9050 mm. ('075 inch), showing that the ovarian examples, mentioned in the Researches, us indeed had often been observed, were nearly ripe. Each egg at this date (to take an instance on 1st October) presented a large oil-globule and an advanced embryo. The circulation was in active operation on the 27th November, the exterior of the egg being, further, coated with many parasites. In January and February the water containing the eggs was frozen, and remained so nearly a month, yet, on the 2nd March, some of the embryos were alive, and, though the water could not be changed, as the pumping apparatus was under repair, were safely hatched. The re- markable hardihood of such eggs is in contrast with pelagic ova, which, as a rule, would have been killed. The egg-capsule (Plate I, fig. 7) is very tough, and rebounds from the needle under pressure like a ball of india-rubber. 1 op. cit. 210, quoted Rept, S. F. B. THE ARMED BULL-HEAD FAMILY. 145 It is thus difficult to extrude a perfect embryo. Under a low power the torn edge shows layer upon layer of the secretion. The larva immediately after hatching closely corresponds with that captured in the bay, and figured by Prof. Prince (Researches, Plate XVIII, fig. 11). The body is about 7 mm. long, the tail, which has embryonic fin-rays, being somewhat longer than in the sketch. The head has large silvery eyes, with greenish-yellow pigment behind them, and around and below the widely-open mouth. The snout is blunt, and the ear-capsules large. The trunk, generally, is dotted over with the same greenish-yellow pigment (ochreous by transmitted light), which is also present in streaks on the basal region of the breast-fins, on the yolk-sac, and on the prominent cone at the vent. The three dorsal patches in the marginal fin are distinctly separated, the first being small, and nearly in a line running upward from the posterior border of the vent; the second about the widest part of the fin, and the third about the posterior region of the same dilatation. In numerous examples obtained in the tow-net in March the three yellow dorsal patches had opaque- white tips, thus giving them a striking- aspect. Opposite the two latter is a corresponding patch ventrally. A similar touch of yellow occurs at the commence- ment of the tail. Numerous stellate black pigment-spots are present on the yolk-sac, a few on the breast-fin, and a series along the dorsal and ventral margins of the body as far as the third patch of yellow on the marginal fin. One or two occur on the border of the marginal fin, and a considerable number in the ventral patch of yellow (second) in the latter, while only a very few are found in the dorsal patch opposite it. Two exist in the ventral touch behind the former and only traces in the dorsal patch opposite. One or two chromatophores also occur ventrally in the patch at the root of the tail. The black pigment would seem to be better developed ventrally than dorsal ly. The larval examples appear in the various nets in the beginning of April, and measure 7 to 8 mm. in length. They soon become abundant, especially in the bottom-nets, their M. P. 10 146 THE ARMED BULL-HEAD FAMILY. condition being represented in Plate VI, fig. 1. As they get a little older their length seems to diminish rather than increase, and the black chromatophores occur on the cheeks. The bases of the breast-fins show rows of them springing from the distal edge of the basal process. The fin-rays everywhere are still embryonic. The sides of the body from the tips of the breast- fins backward almost to the tail show a series of isolated black chromatophores arranged in a double row, viz. towards the dorsal and the ventral edges. A group of similar pigment- specks characterises the median or wider part of the marginal fin dorsally and ventrally (region of the dorsal and the anal fin). On the sides and ventral surface of the abdomen the same black corpuscles are scattered, and they run along the ventral surface to the base of the tail. A slight narrowing — dorsally and ventrally — is evident in front of the tail, which likewise has only embryonic rays. The vent is very prominent, and in front of it another projection occurs — a feature characteristic of this form. The notochord (rudimentary back-bone) is quite straight. When G mm. in length (in spirit), and thus shorter than the larval form, as in the case of the paradoxical frog of North America, a great change occurs in regard to the surface, e.g. on the 19th April. The fish is now considerably broader and thicker, and more densely pigmented on the ventral than the dorsal surface, as in the dragonet and other forms, and some specks also are evident behind the breast-fins. Rows of sharp and slightly curved spines pass along the dorsum, on the lateral crests of the head and on the gill-covers. True rays appear on the lower part of the tail, besides a basal thickening (hypurals); and permanent rays are also developing in the breast-fins. When about 8 mm. long, as in the end of April and beginning of May, the head has now acquired a predominance in bulk — from its elevated dorsal (occipital) and ocular ridges, its spines on the gill-cover, and a tendency to the occurrence of tubercles generally. The spines on the trunk and abdomen are prominent. All these are much more developed in propor- tion than in the adult. Slight papillae on the chin mark the future 'cilia.' The head is boldly differentiated from the THE ARMED BULL-HEAD FAMILY. 147 slender body by the huge breast-fins which project outward like a pair of fans. The black pigment on the under surface is less marked, except at the vent. A V-shaped series of these specks occurs at the breast-fins. The larval tail with the up-turned notochord is still present above the permanent caudal. The dorsal and anal fins, in which true rays are developing, are connected by a broad marginal fin with the tail, and are not so distinctly differentiated as in Prof. Prince's figure (Plate VI, fig. 2). No trace of ventral fins yet exists. In colour this stage approaches the figure just mentioned. When 12 mm. long (in spirit) the breast-fins are larger, so that they form great fans which fold backward and envelop nearly a third of the body, like those of the young gurnard. Moreover, while they are boldly speckled with black, the sides of the body which they cover when adpressed, are pale. The ventrals (pelvic fins) are now present on the flattened abdominal surface. The head reminds us of that of the Moloch lizard of Australia from the great development of spines and tubercles besides the ocular and occipital crests. The double row of dorsal tubercles in the adult is here represented by prominent rows of soft conical spines and so with the two lateral and abdominal rows, the latter making the sides coarsely dentate. Only papillae are still visible on the chin and neighbouring region (branchiostegal). The permanent rays of the caudal are complete and the larval tail at the upper border has disappeared. This stage is also somewhat less than that figured by Prof. Prince but it resembles it in coloration. The little fish is a truly pelagic one and finds safety and food amidst the swarms of minute forms carried about by every current. Those from 14 to 18 mm. (May to July), which are also truly pelagic, fall under the stage represented in Prof. Prince's figure (Plate VI, fig. 2). The spines retain the character of the previous (12 mm.) stage, though they are now considerably harder. Thus from the occiput to the tail the body is armed with four lateral rows of tooth-like processes which are directed backward. In those which have reached 18 mm. the breast- fins seem to be shorter, from the elongation of the body. The papillae on the chin are longer, but do not project much. The 10—2 148 THE ARMED BULL-HEAD FAMILY. chief coloration is fine chrome yellow with black, especially on the breast-fins, spines and dorsal-fins. In the breast-fins the yellow pigment follows the line of the rays, with black points here and there, the intermediate regions being yellow and black. Some a little older reach 19 mm. towards the end of May, and are stouter and hardier fishes with a greater development of dark pigment on the dorsnm, while three deep brown bands, separated by pale bars, occur on this region of the slender body. The cephalic ridges are prominent and the anterior dorsal spines are large and powerful. The papilla?, under the chin show little change. In specimens of this length (19 mm.), with the huge fan-shaped breast-fins, obtained from the tow- nets in the open water offshore, in the midst of a vast assemblage of the pelagic stages of other fishes, the spines are remarkably acute all over. On the other hand, examples of the same length, captured on the bottom within the estuary of the Forth, already have blunt spines throughout, the cephalic ridges flattened and the general aspect more closely like that of the adult, especially in regard to the dense greyish pigment on the upper surface. The breast-fins, however, are much larger, but the ' cilia ' beneath the chin are short. The black bar on the second dorsal, and on the anal fin, is very distinct, while in spirit the first dorsal is pale. The young armed bull-heads reach 25 to 26 mm. at the end of June and beginning of July, and the pigment-bars of the body are distinct. The ventral fins are more elongated. The dorsals are variegated with black pigment (in spirit) and the bar still occurs in the anal. The tail-fin is likewise variegated. The authors of the Scandinavian Fishes observe that Kroyer thought the young reached 90 to 100 mm. in September, but the captures hitherto made in this country do not seem to corroborate this opinion. It is more likely that the specimen of 32 mm. in the Royal Museum at Stock- holm was a large example of that season, just as that of 50 mm. in December would be an older form of the same series. Those procured by Ekstrbm in spring on the coast of Bohuslan, and measuring 80 to 100 mm., were probably a year and some months old. THE ANGLER FAMILY. 149 The next stage, which is separated by a considerable gap from the foregoing, is that procured in May of the following year, and which is considered by Mr J. R. Tosh in his paper1 as one year and three months old. It measures from 50 to 67 mm. (in July) svith marked adult characters, such as the flattening of the ridges on the head and of the spines on the body, the prominent condition of the nasal and opercular spines, the barred appearance of the breast-fins and their proportionally smaller size, and the bleaching of the anal fin. Other features now present are the variegated condition of both dorsals, and the pale border to the darkly-tinted caudal. The filaments on the chin and adjoining region (branchiostegal) are now con- siderably longer. In July, those of the same series reach from 67 to 76 mm., and in December 67 to 87 mm., and it may be that those of 110 mm. pertain to the same group, making the age, according to Mr Tosh, of 1 year and 10^ months in the latter case (110 mm.). Well-grown female adults of 138 to 144 mm., with ripe eggs in December, are probably, as the author just quoted says, about 2 years and 8 or 10 months old. The life-history of this species seems to be as follows : — The adults haunt the sandy flats of the bays and estuaries, and at the breeding-season leave these to deposit their eggs amongst the roots of tangles. From the eggs issue larvae at an advanced stage of development, which soon become brightly coloured and assume a pelagic existence for a considerable time ; then they settle on the bottom like their parents. Meanwhile, how- ever, they have been carried to new districts — it may be at great distances from their birthplace. The Angler Family. Pediculati. THE ANGLER-FISH. (Lophius piscatorius, L.) This species, also called the Frog-fish, is somewhat isolated by its peculiar structure. Its huge gaping jaws, large flattened head crowned by three free fin-rays of a peculiar shape, the 1 12th Ann. Rcpt. S. F. B. Part in, p. 384. 150 THE ANGLER FAMILY. single gill-opening on each side and the short tapering tail are all features which give the angler-fish a unique appearance. The stories concerning the habits of this fish are full of interest but are scarcely in place here. It comes under the category of a food-fish and is treated in a similar way to the wolf-fish. The head is cut off and only the trunk is sent into the market. This is sold under the name of ' croan ' and sometimes of ' John Dory/ a name that of right belongs to a very different fish. Donovan states that its flesh is ' delicate, like that of a frog ' ! whilst Couch pro- nounces it to be delicious. Parnell states that ' the flesh is considered good, especially near the tail.' The angler-fish belongs to that group of fishes which have floating (pelagic) eggs and agrees with most of this group in that the number of eggs carried by one female is great (Thompson gives 1,420,000 and Fulton 1,345,848) and in that the females are larger than the males. The eggs, however, do not float separately in the open water but are surrounded by, and packed closely together in, a mass of mucus, and carried about by the currents, as in the case of Fierasfer. Baird states that the floating sheet of egg- laden mucus may be from 60 to 100 feet square; Agassiz describes the same as a gigantic band 12 or 13 yards long, of a violet grey colour and somewhat blackish, from the pigment of the contained embryos ; Prince describes a specimen from the Frith of Forth 12 yards in length, from 6 to 10 inches broad and -j-L inch in depth ; each egg is about T^ inch in diameter, so that the band appears to be, so to speak, only one egg thick. It is a peculiar accidental coincidence that the frog and its allies have eggs enclosed in masses or bands of mucus, and Prince also compares the young frog-fish larva in external appearance to the Iarva3 of Amphibians. The later larva and even the adult have a certain amount of superficial resemblance in shape to an Amphibian tadpole, which latter fact probably gave rise to the name before the features of eggs or larvae were known. The large, floating masses of eggs are somewhat rarely met with in comparison with the abundance of the adult, and Day THE ANGLER FAMILY. 151 and others have noticed that " considering the wonderful number of eggs they deposit, but the comparatively few young that are seen, they must be subject to some wholesale de- struction as fry, or else when in the ova state." The masses are sometimes caught by the ropes and nets of the salmon- stake nets off the East Rocks at St Andrews. The ova sent to the St Andrews Laboratory were ex- amined and described by Prof. Prince, and the larval forms up to fifteen days after hatching were followed. The very early stages of development have yet to be described, the eggs noted above containing embryos at an advanced stage (Plate I, figs. 11 and 12). The young embryos differ very much from the type of pelagic embryos found among the Gadoids and Pleuronectids. They are short and stunted with a stout obtuse head, and a ' dagger-shaped tail only partially encircling the yolk.' The blunt outline of the head is caused by the enor- mous optic lobes of the brain. The whole character of the larva recalls the ' robust embryos of Teleosteans with demersal eggs and a rich perivitelline circulation.' There is a great aggregation of black pigment, mainly covering the brain and spinal cord and the intestinal region of the trunk. Fig. 3, Plate VI, represents the appearance of the newly- hatched larva. In it we can note the characters above indi- cated and also the peculiar rounded yolk-sac with a conspicuous oil-globule situated at the posterior end. This globule is covered with a mass of protoplasm in which are a number of large black pigment-spots. Another feature of the larval frog- fish not so well seen except in sections, is a very abundant development of serous spaces just under the skin. By the fifth day after hatching (Plate VI, fig. 4) the yolk- sac is considerably reduced, and the head stands out more clearly from it : the body is more elongated and the first dorsal spine commences to appear. The ventral fins are very con- spicuous as rod-like bodies of immense length, the pectorals being also present but small. The flattening of the larva dorso-vcutrally is not yet very evident. During the next few days the first two free dorsal rays become very prominent as elongated filaments above the head, the third being smaller. 152 THE ANGLER FAMILY. The serous space above the head becomes distended. The ventral fins are also long and filamentous with a second short ray appearing at the base of the first. Plate VI, fig. 5, shows these features in the post-larval form 15 days after hatching, and in views from above it is apparent that the flattening of the head and trunk, which is so characteristic of the adult, has already taken place to some extent. By this flattening the eyes come to lie completely upon the dorsal surface, and look upwards. At this point one reaches the stage where the post- larval phase of development succeeds the larval, and the young fish no longer has yolk from which to obtain its supplies of nutriment. It is difficult in tanks to provide the suitable food and surroundings, and hence our knowledge of the further stages has to be gleaned piecemeal by the tow-net. Here we encounter a peculiar developmental feature of the angler-fish, for during all the continued trawlings of the East coast waters which have been conducted for the last few years, — one and only one post-larval stage of the angler has been forthcoming. This curious little fish has been described in detail elsewhere1 and is figured in Plate VI, fig. 6. As will be seen it differs in several important respects from the 15-days-old larva. It was caught in the mid- water net, 15 miles off the Isle of May, at a depth of 25 fathoms. It is somewhat over | inch (7 mm.). The large flattened head and tapering little tail give it a tadpole-like aspect. The gape is large and the snout very blunt — with a median notch. The eyes are large and round and look directly upwards. Part of the continuous embryonic median fin is still present, and the paired pectoral fins are enormously developed, stretching backwards and outwards as two fan-like processes ; the ventrals on the other hand are small, and although broken, appear to have been poorly developed, a feature in marked contrast to the condition of the pectoral and ventral fins of the 15-day larva, where their relative sizes and importance are exactly reversed. The gill-slits are still large and long and differ greatly from the little con- tracted pore-like apertures of the adult. From these observations it will be seen that there are 1 Trans. R. S. E. vol. 35, pt. in. p. 869. THE ANGLER FAMILY. 153 important gaps in the stages of development both between this post-larval form and the earlier stage, and between the same and the adult. The well-known ichthyologist, Day, states that ' the young of these fishes when alarmed seek for shelter in the branchial pouches in the axilla of the pectoral fin.' Such a habit must apply to a later stage than fig. 6, when the pelagic habit has been given up. Another observer has as- serted that the food of the frog-fish is stowed away temporarily within the gill-pouches : the question naturally suggests itself — can the presence of young anglers in the parental gill-pouches be explained in this way ? All accounts of the voracity of this fish would lead us to suppose that it would not refrain from feeding upon its own young. On the other hand we do not know upon what observations the deceased author based his statements : the above expla- nation would only apply if his remarks rested simply upon the presence of young anglers in the gill -pouches of the captured adults and not upon direct observation of the habits of parents and young. Fishes with pelagic eggs are generally characterised by an entire absence of all parental instincts as displayed in the care or protection of the young, and the latter have to make their way in the world with nothing for which to thank their parents but their mere existence, so that from the fact that the ova of the frog-fish arc pelagic we would expect that the young would grow up quite independently of their parents. The eggs, however, although pelagic have one feature characteristic of demersal eggs, namely, the mucus enveloping them in a common mass. The larvoe also have certain features in common with demersal larvae, as stated above, so that to some extent the eggs of the angler form a type transitional between the pelagic and the demersal. It is possible that they may have been demersal in the earlier history of the species, but have secondarily acquired a pelagic habitat. Valenciennes describes the young fish, 2 inches in length, as having long pectoral and ventral fins, with some of the rays of the former extending beyond the fin-membrane. The dorsal spine also had more numerous and longer tentacles than in the adult. 154 THE ANGLER FAMILY. There seems to be a great deal of uncertainty concerning the stages between the post-larval fish already described and figured (fig. 6) and the stage at which the adult characters are assumed. Both Day and Giinther in their treatises give figures of a young angler-fish which agree the one with the other. The former author remarks, ' in the young the head is broader than long, but less depressed than in the adult. The de- pression increases with age while the spines of the first dorsal fin have lateral soft-branched enlargements along their anterior and posterior edges.' Giinther also observes that the ' first dorsal spine bears a ciliated bifid appendage and the next four rays are of a racemose feathered type. There are huge ventral s with four free rays, extending to behind the tail.' This description to some extent agrees with that given by Diiben and Koren of a young angler from Norway which had the first dorsal spine terminating in a transverse cylindrical knob, provided with cilia, and being scarcely above half the length of the second spine. This young form so differed from the adult angler that the above-mentioned observers supposed it to be a separate but allied species and even gave it the name of Lophius eurypterus. The gradual broadening and flattening of the head spoken of by Day is a process we should expect to occur at quite a late stage of development as being a feature in which the angler differs from the general Teleostean type, and which has been acquired comparatively recently in correlation to its environment ; the peculiar feathered tentacles and the enormous ventral fins, on the other hand, must be post-larval characters of a transitory nature as they are not found in fig. 6 and are also absent in the adult. The long ventrals are present at the age of 15 days. One cannot, however, pronounce a definite opinion upon this point until more certain knowledge is obtained. The rarity of the post-larval and young forms when compared with the common occurrence of the adults only tends to the supposition that there must be some element in the environment or habit of these stages of which we are at present in ignorance. THE ANGLER FAMILY. 155 The next stage met with during many years' observations measures 5^ inches, and was tossed on shore after a storm in March. Unfortunately it had been considerably injured by gulls, and also partially dried. The head at this stage appeared to be more elongated than in the adult, and the spinous processes were more conspicuous. The three long anterior spines of the first dorsal were well developed. The teeth were long and sharp as in the adult and similarly arranged, while the mucous membrane of the tongue and that within the lower jaw was tinted brownish. The tip of the lower jaw seemed to be somewhat more acute than in the adult. Both pectorals and ventrals had at the free edge a series of long white elastic processes (the tips of the rays), which were all the more conspicuous in the pectorals from the deep brown colour of the adjoining region dorsally and the blackish belt ventrally ; whereas in the ventral fins it was the outer aspect which was blackish, the inner being whitish. The long fin-rays of the younger stages had thus been considerably modified to suit life on the bottom, the elastic nature of the tips of the rays and the tough integument intervening being adapted in a most efficient manner for locomotion on soft sand. From the colour of shreds of skin which here and there remained the frog-fish of this size is a brilliantly tinted form. Such an example is probably about 9 or 10 months old. A young specimen 7£ inches long was trawled on the 17th February, and another 6^ inches in length occurred in May ; both they and the preceding probably belong to the previous year's brood, but this is conjectural. The adult is found as long as 4 or 5 feet, but data with regard to its rate of growth are not to hand. This much with regard to the habitat we can say with certainty, namely, that the eggs are pelagic and that the early larval forms are also pelagic and descend to the bottom-water during the post-larval period. The adults are, as a whole, both shallow-water shore dwellers, and inhabitants of the deeper areas beyond — lurking in the sand or mud where there are abundance of large and small fishes to be lured into the gaping jaws. 156 THE WEEVER FAMILY. The Weaver Family. Trachinidae. THE LESSER WEEVER. (Trachinus vipera, Cuv. and Val.) For the first account of the pelagic eggs of this species we are indebted to the late Mr George Brook, who, with rare enthusiasm, fitted up tanks of sea- water near his home at Huddersfield, and studied the life-history of this form1 and that of the rockling. He found that his captive weevers in June deposited their eggs during the night, just as was subsequently found in the spawning-pond at Dunbar in the case of the plaice. They are buoyant eggs having a diameter of 1'25 to 1'37 mm., and containing from 11 to 30 small oil- globules (Raffaele says 4 — 10). They were subsequently got in considerable abundance by Mr Holt, in Blacksod Bay and other parts off Ireland, in June and July, and are common in St Andrews Bay, at the surface, from May till July. The yolk is clear and transparent and the oil-globules have a greenish- yellow hue (Plate I, fig. 14). The egg-capsule was thought by Mr Brook to consist of two layers, and in section Mr Holt ob- served that the outer region only took the yellow stain ; while the inner was faintly affected by carmine. Practically, however, the egg-capsule agrees with that of other pelagic eggs in structure, as also does the micropylar opening. Mr Brook gives a minute account of the early stages of development of the egg as they were seen in April, and the late stages as they occurred in July. They hatched in from 9 to 11 days. Dr Raffaele's examples hatched in 8 days in spring, the eggs lying on the bottom during the last three or four days. The spawning-period would seem to range from April to July. Mr Holt found pale yellow pigment, in minute round specks, in the developing embryo, before the appearance of the free tail-region. As development proceeds, the yellow chromatophores become stellate and have a brilliant orange hue. They extend along the upper and lower margins of the tail, and along the gut on either side. Small black chromato- phores have also appeared : others by and by occur over the 1 Journ. Linn. Soc. vol. 18, p. 274. Pis. 3—5, 1884. THE WEEVER FAMILY. 157 brain, about the eye and ear-capsules, and on the fins, as well as over the yolk and other parts of the body. Thus, as Mr Holt observes, at a comparatively early stage of development in the egg the yellow pigment is in excess of the black, whereas subsequently it diminishes. In the recently-hatched larva of Mr Holt, measuring 3'27 mm. (Plate VI, fig. 7) the black pigment of the trunk forms a con- spicuous line from the pelvic fins to the posterior fourth of the tail, while, in those hatched by Mr Brook in his aquarium, the pigment was much less developed. Then the black pigment becomes " concentrated into two bars, one above the vent and the other midway between this and the tail, and they remain till the post-larval stage — when all yellow pigment, except three touches at the margin of the dorsal fin, disappears." Before hatching, the pelvic fins are distinct, and they are soon rotated and carried downward and forward to a ventral position, a little behind the breast-fins. This condition, viz. the early appearance of these fins (pelvic) is remarkable amongst bony fishes. As a rule they do not appear till much later. In three days Mr Holt found that the fishes reached the post- larval stage, by which time the yolk had been absorbed and the mouth and the vent were open. A fine example (Plate VI, fig. 8) about 15 mm. long (in spirit) was procured by the ' Garland ' probably at the end of summer or beginning of autumn — to judge from the condition of a young rockling accompanying it 1. Many of the adult characters are now present, though the body is much flattened. Black pigment occurs in specks on the snout, a black spot is placed beneath each eye, and each opercular region has the same coloration, which is connected superiorly with that tinting the brain and the snout. Patches of similar pigment are present in front of the breast- fins, and a broad black band extends from the opercular spines rather beyond the middle of the body, and from the dorsum as far down as the lateral line. A few isolated specks exist on the side between the latter and the ventral margin. The lower jaw is directed obliquely upward and 1 W. C. M. 9th Ann. Kept. S. F. B. p. 324, 1891. 158 THE WEEVER FAMILY. forward, and the upper jaw has a well-marked notch in the centre. Five long and strong spines guard the front margin of the gill-cover (pre-operculum) as in Coitus, the second from above being the strongest. They are sheathed in skin almost to the tip, and thus from their conical shape are very con- spicuous. Two spines, as in Gottus, occur 011 the gill-cover proper (operculurn), the upper, long and powerful, directed backward and very slightly downward, and a shorter inferior spine slanting downward and very slightly backward, and therefore at a considerable angle to the former. Two spines project just above the upper angle of the gill-slit. None occurs over the eyes. The breast- fins are elongated, and seem to have undergone a rotation downward, so that the angle formed by the rays with the lateral line, is small. So far as can be observed, the lower agree in structure with the upper rays. The large fan-shaped pelvic (ventral) fins stand boldly out from the prominent median keel, and their bases are distinctly in front of the breast-fins ; while they are deeply tinted with black pigment both dorsally and ventrally. The spine is well- developed, but is sheathed in skin to the point. The contour of the belly, as seen from below, differs much from that of the young Cottus, since it forms only a small swelling behind the large pelvic fins ; whereas in the latter the large tumid abdomen is conspicuous — the pelvic fins forming minute pro- cesses at the anterior border. Moreover a somewhat V-shaped mass of black pigment, with the angle directed forward, occurs in the middle line immediately behind the pelvic fins. The first dorsal fin is black, with the first three spines well-developed, but sheathed in skin almost to the tip. It is joined by mem- brane to the second dorsal, the short spines in this part being still covered. The second dorsal and anal resemble those in the adult. A young Crustacean parasite adheres to the anterior surface of the right breast-fin. Precisely similar forms about 14 mm. in length were obtained by Mr H. C. Williamson at Naples in May. In a specimen (in spirit) measuring about 26 mm. (1 inch) the chief changes are in the pigment and the proportionally larger size of the body. The black pigment now has a tendency THE WEEVER FAMILY. 159 to form a series of spots along the sides. These form three longitudinal rows in the upper half, viz., a row along or just above the lateral line, another along the base of the dorsal fin and an intermediate more or less double row. Ven- trally a row occurs along the base of the fin and scattered specks above it. The two opercular spines are still sharp and prominent, and five long pre-opercular spines are conspicuous, the median being slightly curved as well as longest, and the most anterior being directed forward. At least three spines occur above the opercular slit. The papillae along the upper edge of the lower lip are distinct. When 34 mm. (in spirit) the rows of black specks along the sides are distinct, the median being least marked. Only a single opercular (gill-cover) spine, the upper and larger, is present, and the pre-opercular spines are shorter, indeed, on one side only four can be made out. One is directed forward. The breast- and pelvic fins are proportionally larger. The body has increased considerably in depth. At 36 mm. the three rows of pigment-specks are very distinct, and the scattered ventral series is also present. At 45 mm. (If inch) in length (July) the aspect from the dorsum is mottled, while laterally the most distinct row of black pigment-spots is above the lateral line. The eyes are more prominent. Only a single spine below the pre-operculum is directed forward ; and two point backward. The opercular spine is larger and longer. The other characters are appar- ently those of the adult. This specimen was procured by digging in the sand while searching for sand-eels. Both species when in the sand escape capture by trawl or ground-net, but both are dislodged by storms, and tossed on the beach. The eyes are proportionally nearer the tip of the snout and look more upwards than in the earlier stages. This approximation of the eyes to the tip of the snout and the protrusion upwards of the whole anterior part of the head are still more marked features in an example 70 mm. (2f inches) long. The bands of pigment in this stage approximate more closely to the adult condition. It is interesting that in the earlier stages Gottus and 100 THE MACKEREL FAMILY. Trachinus should have so many features in common in regard to the armature of the head and the general shape. Couch1 describes a young example of what he considers to be the greater weever, measuring only f of an inch in length and captured with a drift-net in August. The opercular spine was not much developed, and the same may be said of those at the anterior superior angles of the orbit. Pigment occurred on the head and anterior portions of the body, diminishing towards the posterior region — which was translucent. " The lower part of the pectoral fins have colour, but the upper portion is devoid of it. Two or three rays on the inner portion of the ventral fins are black." The Mackerel Family. Scombridae. THE MACKEREL. (Scomber scombrus, L.) The mackerel forms an instance of a fish, whose pelagic habits are not, as in the case of the haddock or the cod, confined to the embryonic and larval stages, but the surface-water in the open seas is the haunt of this species throughout life, with perhaps the exception of its very young stages. This exception is fraught with interest to the investigator, for it might with some reason be conjectured that a fish living in mid:ocean and having a floating egg would thereby be enabled to spend the whole of its life Avithout approaching the proximity of land, in which case it would not be to any great extent available as food for man. Fortunately this is not the case, for the mackerel congregates in vast shoals, and these leaving the open sea migrate towards the land and discharge their eggs in the inshore waters. The periodic migrations of myriads of fishes, as occurs in the case of the mackerel and the herring, are not by any means well understood, but the tendency amongst naturalists is to regard them as migrations connected with the discharge of the sexual function : thus some may be inclined to hold that the mackerel has become secondarily adapted to a pelagic life i Vol. II. p. 46. THE MACKEREL FAMILY. 161 and is descended from shore-living forms in the past, and although the adult mackerel is perfectly suited to a pelagic existence, yet the eggs and young stages for their successful development still require the proximity to land, hence the periodic migration of the adults landwards and that of the adolescent forms seawards. The life history of the frog with its familiar aquatic tadpole-stages has usually a similar interpreta- tion put upon it. Granted, however, that the primary cause of the migration of these fishes is due to demands made by the needs of the young there is no doubt that perhaps in a lesser degree, the direction of migration is determined by the sources of food ; thus shoals of mackerel are frequently found in keen pursuit of millions of young clupeoids and ' mackerel-midges ' (see Rockling). This pursuit of food has been assumed by some to be the primary cause of the migration. We may approach the question from another point of view. Along with, and contemporaneous with the migration is the crowding together, in a close mass, of myriads of individuals : this massing is directly favourable to the fertilisation of the eggs, it being obvious that the greater the number of fishes that shed their sexual elements in a given area, the greater the probability of fertilisation ; on the other hand it is directly unfavourable to the nutrition of each fish, for the supply of food in a given area is limited. Hence the fulfilment of the repro- ductive function furnishes a factor tending to shoaling whilst that of the nutritive function supplies, on the other hand, a factor tending to keep the individuals apart. If it were not that the latter function is largely suspended in most fishes at the reproductive period shoaling would be almost impossible to the vast extent that occurs in many clupeoids. The periodic appearance and disappearance of the mackerel, as in the case of the herring, has formed the basis for a good deal of speculation and for the initiation of many curious legends and superstitions. The two most generally prevalent theories in the past may here be mentioned. It was supposed by many that the Arctic regions were the home of the mackerel (cf. Herring) and that from the Polar seas emanated the shoals to our shores. Lacepede and others relate M. F. 11 162 THE MACKEREL FAMILY. on the authority of a French admiral, who apparently derived his information from his seamen, that mackerel could be seen through the clear water on the shores of Greenland with their heads buried in the mud, like the proverbial ostrich : they even went so far as to relate how the fishes, on emergence from their temporary concealment, were partially blind and could be easily caught ! Needless to say this story has not been corroborated, and Fabricius and others have been unable to find any mackerel in either Greenland or Iceland. The other suggestion was that the mackerel retired to deeper water in the winter and passed the colder months in a state of lethargy at the bottom of the sea, a theory peculiarly suitable to the case by its very difficulty of confutation or corroboration. Curiously enough, this point was brought up some years ago in connection with the claim of United States boats to fish in Canadian waters, the right to follow and capture shoals migrating northwards being under dispute. It is now pretty well ascertained that the middle and eastern parts of the Atlantic form the home of the mackerel which migrate thence into British waters. Towards the end of winter they commence their movements to the coast, and by May and June large shoals of spawning fishes are to be found close inland. Couch states that in these migrations the males precede the females. Prof. Sars found that the process of spawning is effected quite near the surface of the sea, and he first described the egg as slightly larger than that of the cod, but with a conspicuous oil-globule. It was also described and figured by A. W. Malm of Gb'teborg. The spawning-season in Norway is in June and July. A mature female may contain about half a million eggs, not an extremely prolific proportion. The egg of the mackerel has not been definitely described as occurring off the Scotch coast, but Mr Cunningham has followed the early stages at Plymouth, and Mr Holt in specimens from the W. coast of Ireland. The latter states that the egg (Plate I, fig. 16) is 1'22 mm. in diameter with a colourless oil-globule of '32 to '33 mm. diameter. Cunningham describes the oil-globule as occupying a position ventral to the posterior of the embryo, THE MACKEREL FAMILY. 163 but Holt shows that it is really situated at the vegetative pole exactly opposite the middle of the embryo and that it gradually migrates posteriorly as development proceeds. He also remarks upon the 'cloudy' appearance of the globule. The period of hatching extended over six days, with a mean temperature of 14'5°C. (Cunningham). The yolk is large and elongated and the head of the embryo is closely pressed upon it ; there is no trace of a mouth. The pectoral fins may be noticed as small semicircular folds far behind the head. The vent opens a little way behind the yolk-sac. The notochord is niulticolurnnar. The marginal fins are broad and devoid of all pigment. Yellow pigment is found upon the oil-globule and behind the eyes, black pigment also occurs upon the oil-globule. Stellate black chromatophores cover the dorsal surface of the head and trunk, with a tendency to the formation of a median dorsal row. A few black dots are also seen ventrally. Holt notes that the chromatophores are ' blue-black' and not ' dead-black' as in the rocklings. The oil-globule, although at first loose in the yolk, early becomes fixed in position by the growth round it of a layer of protoplasm. It may now be seen anterior to the hind end of the embryo. At this stage (5th day) bright green pigment in patches occurs, first behind the eyes and on the tail, near the oil-globule. This pigment and the absence of all yellow colour would at once distinguish the species from any of the gadoids or pleuronectids. Holt described the pigment as of a yellow colour by reflected light, and in a later paper Cunningham explained the 'green' appearance as being due to the admixture of black and yellow chromatophores. This is not a surprising- result when we recollect that Holt has noted that the black pigment in this species is of a 'blue-black' tint. The green appearance does not lose diagnostic value on account of its being only an effect and not due to the presence of actual green pigment. By transmitted light Holt found that the yellow pig- ment appeared to be reddish-brown. In Plate VII, fig. 1 is shown a lateral view of the newly hatched larva; 3'5 — 3'9 mm. (Holt), 4'23 mm. (Cunningham). In four to six days the larva has grown to 4'16mm. H 11—2 164 THE MACKEREL FAMILY. (Holt). The mouth has appeared and is open. The eyes are intensely black. There is a tendency, in other parts, to lose the black pigment. The oil-globule has now only yellow pigment, and the black pigment on head and trunk is sparse and confined to the median dorsal and ventral regions. A black mass, however, appears on the dorsal wall of the abdominal cavity. The yolk is rapidly disappearing. At the ninth day the larva (Plate VII, fig. 2) has increased to 4'88mm. in length. The yolk has disappeared and the oil-globule is reduced to very small dimensions. The lower jaw projects and the eyes are, by reflected light, a brilliant blue. The black pigment has increased in the dorsal region of the abdomen but decreased elsewhere. At the 10th or llth day the post-larval condition is reached, and of the stages beyond this we know nothing. Holt points out the strong resemblances of the larval mackerel to the larval Coitus. He emphasises in his comparison the following characters in common : — the absence of pigment on the marginal fins, yolk-sac, and general surface of the yolk ; large stellate black spots on the head and anterior region generally ; a mass of black pigment on the roof of the abdominal cavity, and the tendency of the black pigment on the trunk to be arranged longitudinally (not in transverse bars, as in the gadoids generally). If this could be construed into any racial connection between the mackerel and the cottoids, then the latter might be regarded as relations of the former who have not become adapted to a pelagic life bat have remained littoral in habit, but this is mere conjecture. There seems good reason for holding that the embryos in the egg and the larvre drift landwards in the way described for the cod and haddock and that the post-larval (Plate VII, fig. 2) and young stages are passed in the inshore waters. Dunn states that at Mevagissey the young fishes, about 3 inches long, occur in the bays during August and September, leaving for the deep sea in November when they are from 6 to 7 inches long, and by the time of their reappearance the following June they reach a length of 8 or 9 inches. Prof. Sars believed that at the end of the first year the THE SCAD FAMILY. 165 young are as long as one's finger, and that in two years they reach the size of a herring, and are ' fully grown' in three years. He states that for the h'rst two years they remain in the open water near shore. The facts given by these observers can only be reconciled by assuming that the ' young' of about 3 inches described by Dunn must be about one year old, but a difference in length of from 3 inches in September to 6 or 7 in November makes the continuity of the series rather improbable. Perhaps the series 4 inches — 8 inches — 11 inches may be taken to express the average rate of growth of the mackerel in the first three years of its life. Sexual maturity appears to be attained at a length of 11 — 12 inches but individuals have been known to exceed 18 inches in length. The Scad Family. Carangidae. THE HORSE-MACKEREL. (Caranx trachurus, Lacep.) Couch found the scad with ripe roe in the middle of July, Day giving this month as well as June and August as the spawning-season. In the Mediterranean, according to Risso, it is earlier, viz. in spring. Holt found ripe males and almost ripe females on the 16th June at Penzance. The eggs measured •84 mm. in diameter, with several oil-globules — which, by and by, fused into one. He imagined that they were pelagic, and similar to those of the mackerel except in size. The oil-globule is proportionally large. In September the bays in Cornwall swarm with young, an inch in length, and in October about 2 inches long. Couch also observes that he has found them, not exceeding an inch in length, in the stomach of various fishes, from the end of December to the middle of February ; while he has again seen them 3 or 4 inches in length in August and September. Cunningham took similar specimens during the same months. The spawning-period, therefore, as usual, would seem to extend over a considerable time, but chiefly in June and July. The fish is not uncommon, singly or in small 166 THE SCAD FAMILY. numbers in Scotland, but opportunities for the thorough inves- tigation of its eggs and young have hitherto been denied. Young scads were found by A. W. Malm many years ago in company with the jelly-fishes, Ct/anea and Rhizostoma, feeding on their eggs, and ranging, from the 22nd July to the 18th August, from 12 mm. to 36 mm. They appear to leave the jelly-fishes when 70mm. in length. They are caught in seines from the latter length up to 100 mm. This habit of frequenting jelly-fishes for shelter and food is likewise found in the cod- tribe, as well as in other forms. The only instance, in which a young fish resembling this form and about 9 mm. in length occurred, was in the Forth on the 25 th of October. A specimen 34 mm. long captured by Mr H. C. Williamson at Naples, on the 8th August, is a shorter and a deeper fish than the adult, with a larger head and eye. The pigment of the body is much less developed than that of the adult, and the two dorsal and the caudal fins are minutely dotted with black. The second dorsal and the anal almost touch the base of the tail. The base of the ventral is behind that of the pectoral of the same side, whereas in the adult they are nearly in the same perpen- dicular, and the abdomen has considerably lengthened. The lateral line shows indications of the hard plates. This specimen would seem to have been spawned in spring, probably in April or May, and therefore the season is earlier than that mentioned by Day, viz. July. BOAR-FISH. (Capros aper, L.) Risso long ago stated that this species, which is common on the southern coast of Britain, shed its eggs in April, and Mr Dunn found it spawning on the 20th July, 1880. The eggs appear to be pelagic. Mr Cunningham1 considered that they approached the shore in July and August for the purpose of spawning, and he describes the egg (Plate I, fig. 13) as hyaline, and having a diameter of '97 to '98 mm., with an oil- globule '19 mm. He did not succeed in hatching them, but 1 Jour. Mar. Biol. Assoc. 1889—90, p. 10. THE DORY AND THE GOBY FAMILIKS. 167 observed that the embryo showed black chromatophores near the dorsal median line. The Dory Family. Cyttidae. THE DORY. (Zeus faber, L.) This is a rare fish in northern waters, an occasional specimen only being captured. Couch considered the dory a fertile fish, and "young ones of small size are often met with, but they soon become scattered." Day records that Mr Dunn (Mevagissey) thinks it spawns in winter: and that the young are 2 or 3 inches in length in August. Mr Cunningham l, on the other hand, is of opinion that spawning takes place about June and July, since in spring the roe was ripening, while in autumn it was " shotten." He gives a table showing that on Oct. 3 the young are 1'7 inches, and apparently, he thinks, about 3 mouths old. Next June (June 3), they are 4-9 to o'o inches and 11 months old. On August 20, 5'3 to 61 inches and a year and 1 month old. The following April they reach 9'6 inches with an estimated age of 1 year and 10 months ; while in September they are 11 1 inches and 2 years and 2 months. The average size, according to this author, is from 15 to 18 inches, which he says is probably not reached in less than 3 years. The Goby Family. Gobiidae. THE FRECKLED GOBY. (Gobius minutus, Pall.) Though of little use to man as food, yet this and other members of the genus are readily preyed on by the cod and other fishes, and from their wide distribution and great numbers they are of greater importance than at first sight appears. The remarkable features of their eggs also claim some at- tention. 1 Jour. Mar. Biol. Assoc. 1891—92, p. 111. 168 THE GOBY FAMILY. The eggs of the freckled goby (Plate II, figs. 6, 7 and 8) were figured by Hoffman1, and indeed had probably long been familiar to marine zoologists. Parnell and Day give the month of June as the spawning-period, but little definite was known of the eggs and larv* until lately. Those of this species not infrequently occur at St Andrews in May, attached to the inner surface of the valves of Lutraria and Solen, and they have been carefully figured and described by Mr E. W. L. Holt2. Dr Petersen has recently given excellent figures and a descrip- tion of the eggs of this species, which occurred on the interior of the shells of My a. arenaria (Plate II, fig. 8), fastened by the same network as already described. The eggs (Plate II, fig. 7) are somewhat pyriform, though as in the black goby, the outline alters with the developing embryo, and they measure -9 to 1 mm. in their long diameter. Mr Holt observes that the egg is elongated — varying from 1'4 mm. to 1*2 mm. The contour is somewhat pyriform, and the narrow end is blunt, almost truncated. The larger end, on which the egg rests, tapers rapidly to a small facet or pedicle of attachment. It will thus be seen that the British examples are larger than the Danish, and differ slightly in outline, being less tapered distally, but a considerable amount of variation probably occurs. The space round the yolk (perivitelline space) is large and is principally in the lower region of the egg. The yolk is bean-shaped, with numerous oil-globules of various sizes. The embryo lies in the long axis of the egg. The capsule is very thin, but has the usual structure. From the facet or pedicle of attachment, Mr Holt goes on to say, springs a hyaline structure which spreads outward in the form of an umbrella. It is pierced by alternate con- centric rows of diamond-shaped or ovoid apertures, which increase in size the further they lie from the pedicle, whilst on the contrary the hyaline intermediate substance is more massive than in the distal rows. Finally the structure is continued in the form of a fringe of long tapering threads, which adhere to the shell and to the threads of the adjacent 1 Verhandel. d. Komnk. Akad. 1883. - Ann. Nat. Hist. July, 1890, p. :U. THE GOBY FAMILY. lb'9 eggs. The latter with this exception are completely isolated. Mr Holt likewise traced the development of this curious process for attachment in the roe of the parent fish, so that there can he no doubt about its identity. The larval gobies 3 to 3'5 mm. (Plate VII, fig. 3) appear in the bottom-nets towards the end of April and in May, and they are readily distinguished by the air-bladder, and by their color- ation. Of a living one of the latter size from St Andrews, Mr Holt1 observes " the vent is slightly anterior to median, the pectorals are large and fan-shaped, and the embryonic dorsal commences opposite the pectoral girdle. Larval fin-rays occur in the tail. There is a considerable pre-anal fin.... The eye is greenish yellow-." Black stellate chromatophores, with yellow amongst them, occur below the anterior end of the notochord, extending as far back as the air-bladder. The latter is greenish, with black dendritic pigment. " Above the vent and between that and the air-bladder are two large masses of gamboge- yellow pigment (reddish-brown by transmitted light), over each of which is a large black stellate chromatophore. Small black chromatophores extend along the ventral edge of the anterior two-thirds of the abdomen and along the ventral edge of the post-anal region to a point a little short of the tail. Midway between the vent and the tail is another yellow patch, overlaid by stellate black pigment, and a similar but smaller dorsal patch lies just above it. The diminished yolk seems to con- sist almost entirely of small oil-globules." In those of 5 mm., in the bottom-nets, the yellowish coloration is lost in spirit, but a group of black chromatophores occurs on the occiput and immediately behind, while the post-anal region has four symmetrical black touches — two on the dorsal edge and two at the ventral. A single row of black specks occurs along the middle line of the abdomen. The upward slope of the mandible is marked. In June they range from 7 mm. to 15 mm., but numerous examples are only from 3 to 3'5 mm. not only in inshore waters, but in the deeper areas, such as the Moray Frith, and generally 1 02). cit. p. 37. - From the dorsum, silvery when viewed laterally. 170 THE GOBY FAMILY. in the bottom-nets. This variation is in all probability due to the prolonged period during which spawning of the various individuals takes place, and also to the differences in species. Gobies become common in the nets in July and they vary from 3'5 to 17 mm., and the same may be said of August, when many reach 18 mm. In September they are likewise abundant in the bottom- nets, and range from 7 to 30 mm., but they do not uniformly increase in size as the month advances, since many towards the end of it are only from 8 to 10 mm. Groups of larger forms from 17 to 25 mm. (1 inch), however, are more common. In October, gobies of 4'5 to 14 mm. still occur in con- siderable numbers, both in-shore and off-shore, the average size of the groups being larger than in the earlier months. Towards the end of the month some reach 38 mm. (1^ in.), and such would appear to be the young of the season. Again, those of 33 mm. in February, of 33 to 47 mm. in March and of 42 to 50 mm. in April would appear to be the young approaching the end of their first year. In June one specimen may be noticed as reaching 2f inches. In July the latter reach from 42 (If in.) to 62 mm. (2^ in.) ; in September from 62 to 75 mm. (3 in.) or even to 80 mm. Some of the second year however only reach from 55 to 60 mm. in October, so that considerable variety exists. In the following January they attain from 76 to 80 mm. (3 to of in. ). THE BLACK GOBY. (Gobins niger, L.) So little notice was taken of the eggs of this form in Britain, that Day in 1884 does not refer to them, while Parnell observes that they spawn in June. Couch says that he found enlarged roe (ovaries) in February and young ones in autumn. They were incidentally referred to by Mr Saville Kent in the Hand- book to the Fisheries Exhibition of 1883, and he mentioned that over these eggs the male mounts guard. They were also found by Prof. Prince in the pools near the Beginnish Islands on the West coast of Ireland in 1890. Long before, however, A. W. Malm had referred to and figured the peculiar eggs of THE GOBY FAMILY. 171 these fishes, and so had Hoffman. The southern forms would appear to spawn considerably earlier in the year than those from the north, to judge from the size of the eggs in a female obtained in Guernsey in August. When describing the eggs of the freckled goby in July 1890, Mr Holt referred to the spindle-shaped egg-capsules of the black goby and gave a figure of a preserved specimen. They are not uncommon in the Channel Islands, where the species is frequently met with between tide-marks, and the eggs are fixed in rows to various bodies by the peculiar network so characteristic of the eggs of gobies (Plate II, figs. 11 and 12). An excellent account has been given of the eggs of this and other species of goby by Dr C. G. Petersen1, the able Director of the Biological Station of the Danish Government. He also refers to the spindle-shaped out- line of these eggs, and the narrowing towards the surface of attachment, viz. mussels, stones, ascidians, wood, and sea-weeds. They measure 1*5 mm. long, and have the same loose network of attachment as in the other species ; the edge of the network ends in long filaments as formerly described. As soon as these eggs are laid in salt water, the membrane in which they are enveloped bursts, remaining at the lower pole as the network just mentioned. The free end of the egg is able to swing when the water is in motion. The yolk is opaque, lemon-coloured, and has many oil-globules. The Iarva3 when hatched are 3 mm. long (Plate VII, fig. 4), and show a series of black chromato- phores along the central edge of the muscle-plates from the vent to the tail. Only a single spot occurs about the posterior fourth on the dorsal edge. Dr Petersen found that this species was two years in reaching full size, whereas he thought the other species attained their full growth in one year, He is also of opinion that this species is less partial to pure, clear and cool water than the doubly-spotted and freckled gobies. Recently Mr Holt figured a post-larval form of 11 mm. (Plate VII, fig. 5) procured on the West coast of Ireland, on the llth June, which would show that the spawning- period is similar to that of the other forms. In this the air- bladder is large ; the eyes are black with a dark greenish sheen 1 Fiskeri-BcrctniiKj, 1891—92. Kjbeukavn. 1 72 THE GOBY FAMILY. superiorly. " The surface of the head and body is covered with a dull olive-green pigmentation, which is only absent from the pectoral and marginal fins, opercular flap and tips of the jaws. The green colour is somewhat darker on the top of the head and abdomen than elsewhere, and dotted with small black chromatophores. Four bands of reddish-brown cross the body at various points, viz. behind the air-bladder, at the vent, at the base of the tail, and between the latter and that at the vent. The body is thus more or less opaque1." THE DOUBLY-SPOTTED GOBY. (Gobius flavescens, Fabr.) The eggs (Plate II, fig. 13) are somewhat similar to those of the freckled goby, and are attached to dead shells of mollusks and cirripedes (sea-acorns), the male, as in the other gobies, keeping guard over them. In the Scandinavian Fishes the spawning-season is stated to extend from May to August, and Mr Roberts, who kept them in the Scarborough Museum, found that the eggs hatched on the 16th day. Dr Petersen observed this species spawning in June in the clear and pure water of the enclosed Fjords of Denmark in the grass- wrack region, the eggs being attached to smooth spruce-stakes, to loose grass-wrack, tangles (Plate II, fig. 14) and other structures. Their long diameter is '8 mm. and where broadest '(J mm., the British examples being larger. They are also fastened by the characteristic network of the other species. The yolk is opaque, gritty, and slightly brownish, with many reddish-brown oil-globules. He thinks that it deposits its eggs in deeper water than the other two species mentioned, an opinion in agreement with experience in this country. The early stages of the young of this species have not yet been described with sufficient care to warrant separate con- sideration at present. Suffice it to say that they occur in swarms in the rock-pools all round our eastern shores in August and September, reaching in the latter month from 20 to 30 mm., while in the following July they attain the length of 47 mm., 1 E. W. Holt, op. cit. p. 441, PI. 47, f. 12. THE GOBY FAMILY. 173 the adult length being, according to Day, about 2£ inches, while the authors of the Scandinavian Fishes state that it is from 60 to 65 mm. ; they also give a very complete account of the coloration of this species. NILSSON'S GOBY. (Crystallogobivs Nilssonii, Dlib. and Kor.) The free eggs of this species have not yet been described, though they are visible enough in the body of the ripe female, in spirit. The post-larval stages are met with in July. Thus on the 20th July a specimen 7 mm. in length1 presented a head with a large pair of eyes and a prominent mandible below the open mouth. From the somewhat truncated condition of the premaxillary region, however, the mandible, when closed, is probably bent obliquely upward, and it is also truncated, the margin indeed in each being nearly straight, with a prominence at the outer end, from which one or two sharp teeth project beyond the mucous membrane. The truncated margin is shorter in the premaxillary region, and after an interval there is another prominence with a smaller tooth. From what appeal's to correspond to the lower margin of the preopercular region, two prominent spines project outward and backward. The branchia? have simple papillae. The body tapers gently from the head to the tail, the notochord passing nearly straight out in the latter, though a considerable thickening (hypural) exists beneath. The dorsal marginal fin begins behind the head and joins the caudal. Ventrally it extends from the vent to the caudal. The latter shows only embryonic rays and these are less distinctly marked throughout the marginal fin— dorsally and ventrally. The pectorals are rather large and long, with true rays, the upper and lower being short, the median much longer. They are deeply tinted with black pigment on the inner surface, or what probably is more or less ventral when in motion. It is chiefly developed in the inter- radial integument in the form of isolated squares or patches. 1 W. C. M. Ninth Rept. S. F. B. 1891, p. 322. 174 THE GOP.Y FAMILY. Two pigment-specks occur near the middle line in front of the eyes, four symmetrically placed over the mid-brain, with several chromatophores in the middle line behind, the only others being a few at the base of the pectorals ventral ly, and along the ventral margin in the posterior third of the body. Another example, only 4'5 mm. long, in spirit, was got on the 16th of the same month, but it differed very little in structure from the foregoing, for the pectorals had true rays and deep black pigment, the caudal had embryonic rays, and sharp teeth were present. In all probability these are the post-larval stages of males of Crystalloyobiiis nilssonii, which, at that date, had only once been found in Scotland, viz. by Mr Edward, in a rock- pool at Banff. The very early stage at which the special sexual characters occur in this species is interesting, yet Dr Day considered that in the young males the head is more pointed, indeed, almost as in the females — in which the jaws are short, straight, and toothless. It is clear that such a statement requires amendment, especially the supposition that the teeth (' canines ') are only developed as maturity is reached. It is possible, however, that the young of both sexes may have male secondary characters. In the small example from St Andrews it was the presence of these characteristic teeth and the shape of the mandible that- attracted attention, and yet the larval marginal fin was still present. The description just given may be supplemented by the remark that in the adult male the pectoral fins are remark- able for their " broad (multiradiate) and somewhat lobate form, with their semicircular muscular root, and elongated, roundish shape. When expanded the pectoral fins are as deep as the body." The remarkable coloration of these fins in the post- larval stage is an addition to the series in developing Teleo- steans. Since the first specimen was captured Mr Cunningham has procured many adults of both sexes near the Eddystone Lighthouse, while Mr Holt has been equally successful off the West coast of Ireland, and on the East coast of England. It appears to be generally distributed round British shores, though the adults have not yet been secured here. It is sometimes THE DRAGONET FAMILY. 175 found in as great abundance in Norway as in the South of England. The authors (Fries, Ekstrom and Simdevall) of the recently published work on Scandinavian Fishes, observe that June is approximately the spawning-season of this species, and the development of the post-larval examples would appear to point to May or early June as the probable period on our shores. In spirit-preparations about 32 mm. long procured in August, and forwarded by the Royal Dublin Society, the black pigment is arranged as follows : — In both sexes, a continuous median line passed from the hyoidean region to the rnid- abdomen, while from the vent backward an interrupted series of black chromatophores proceeds to the tail. From the pectorals backward to the vent a series of similar pigment-specks is dotted on the sides of the abdomen. The female in this case was nearly ripe, the ova being visible. The Dragonet Family. Callionymidae. THE DRAGONET OR SKULPIN. (Callionymus lyra, L.) This fish stands in a somewhat isolated position in its relationship to the other food-fishes, and though it is not directly used for food by man, it is utilised indirectly in that it forms an item in the diet of the cod or other food-fishes. Although its development does not closely resemble that of the gobies, there are no good reasons for altering its present alliance with them. Its scientific name is given to it on account of the brilliant colouring of the males. The females besides being smaller are of a sombre brown colour, which is doubtless a protective factor as the fishes haunt the bottom where stones and sand abound. The males are said usually to frequent deeper water though we cannot corroborate this opinion. The eggs are pelagic and fertilisation is effected in the water. So different do the sexes appear that it is not very long since they were described in works on Fishes as separate species. The male dragonet differs from the female, 176 THE DEAGONET FAMILY. not only in its gorgeous tints but also in other features such as longer dorsal fin-rays, and a larger head. These differences between the sexes, apart from those of the sexual-organs, are known as secondary sexual characters, and such characters as a rule are not found (except disparity of size) in the fishes which lay pelagic eggs. This is probably connected with the fact that the fishes do not pair but merely shed their milt or ova indis- criminately in the water. The dragonet forms an exception in that the egg of this species is pelagic and yet the secondary sexual characters are very marked. In accordance with this, we find that the dragonets instead of congregating in shoals at the breeding season, pair off and move about in couples whilst discharge of the sexual products is effected. The males are said to be rarer than the females, in the proportion of about 1 to 3. Two theories have been suggested with regard to the origin of sexual coloration. The bright colour of the males is explained as the result of continued selection by the females of the most brilliantly accoutred males, and thus by the law of heredity the brightest males only reproduce their species, and on the other hand, a protective and therefore sombre garb is more necessary to the females than the males in order that the next generation may be preserved. Young males are occasionally found present- ing the brown coloration usually characteristic of the females only. The dragonet appears to spawn from May to August. Eggs are found in most abundance in June, July and August, but some occur in early May (Smith Bank), even at an advanced stage of development1. This being the case, the dragonet seems to spawn throughout the summer months, and its breed- ing-season is long. It is one of the species which occasionally discharges eggs in proximity to bhe shore, i.e., within the territorial waters. The egg (Plate I, fig. 15) is comparatively small (average diameter '025 — '03 inch) and perfectly translucent, but is at once distinguished by the fact that outside the egg-membrane are ' a series, for the most part, of hexagonal reticulations like 1 A. T. M., llth Ann. Rep. S. F. B. THE DRAGONET FAMILY. 177 those of a honeycomb1. The septa bounding the reticulations stand out very distinctly and their edges show minute striae.' The cause and raison d'etre of these peculiar ridges covering the whole of the exterior of the egg are unknown. They are doubtless part of the egg-membrane thus modified whilst still in the early soft condition. It is curious that in an allied species of dragonet (C. festivus) the egg-capsule is smooth, as specimens ('55 to '57 mm. in diameter) forwarded by Mr H. C. Williamson prove ; and this led the able Italian observer Raffaele to conclude that the original description of one of us was erroneous. Cunningham2, however, showed that such was not the case. This peculiar reticulate capsule is said to occur in Uranoscopus and Sanrus lacerta. The appearance of the capsule of the large egg of the former3 (" star-gazer") is almost identical with that of the skulpin. A coarsely reticulate appearance is also present in the large thick-shelled egg of Macrurus sent from Naples by Mr Williamson. There is no oil-globule. The development of the embryonic dragonet, within the egg, has been observed by Prof. Prince4 and Mr Holt5 in the same year, the former in February, the latter in the autumn. Prof. Prince gives the earlier stages, from the second or third day onward, while Mi- Holt commences with the larva. The former found that the cortex of the yolk was vesicular or reticulated externally, so that it appeared to be irregularly segmented. Pigment of a rich yellow colour, deeper than canary yellow, and approaching orange, early appears in the embryo. " It occurred as small grains upon the head, before and behind the eyes, and dispersed in this anterior region beyond the embryonic trunk proper over the yolk, but to a very limited extent. A girdle of amorphous yellow spots occurred just in front of Kupffer's vesicle, and others were grouped at the caudal end of the fish. In a more advanced form black pigment-spots had made their appearance." Mr Holt states that the marginal fins are of moderate size, and that the ventral sends forward a narrow 1 W. C. M. Ann. Nat. Hist. 1885, p. 480. 2 Jour. Mar. Blol. Assoc. I (N. S.) pp. 21 and 37. s Kindly sent by Mr H. C. Williamson from Naples. 4 9th Ann. Pep. S. F. B. p. 34*J. 5 op. cit. p. 442. M. F. 12 178 THE DRAGONET FAMILY. strip along the posterior border of the yolk-sac. The notochord is uni-columnar, as in the gobies. The newly-hatched larva (Plate VII, fig. 6) is 2'OSmm. long. The pigment is bright orange, and occurs on the snout ; it forms a bar in the middle of the free caudal region, various touches on the marginal fin and scattered points on the yolk-sac. When twelve hours old the yolk has slightly diminished, while the marginal fin has undergone considerable expansion. The body-pigment is now more definite, a dorsal and a ventral touch in the respective marginal fins being opposite the conspicuous bar in the mid-caudal region. Another dorsal touch of yellow occurs in front, and various specks on the head and yolk- sac. When forty-eight hours old1, the snout is pointed and unusually prominent. The breast-fins form pad-like areas midway between the eyes and the vent, a somewhat interesting position, since in many species they appear much nearer the head. The yellow pigment is still largely confined to the upper half of the trunk, and the sparse black specks are strictly so limited. The dorsal fin is still deeper than in the previous stage. The surface of the larva presented a peculiarly reticu- lated aspect, which Holt considers to be pathological. Only rarely is a larval dragonet secured in the nets in May. In June post-larval forms 3'4 to 4 mm. long have been obtained in St Andrews Bay and the Forth, the tail forming a long slender process — feathered with the embryonic rays, and, in the preparations, bent upward. With such a tail, shortening of the body can readily take place, since the thickening at the base (hypural) leads to the absorption of the long larval tail above it. In July, a length of 9 mm. is reached, though some are only 3'5 mm. The mid- water net used in deep water, such as S.E. of the Isle of May, captures just above the bottom many young skulpins from 5 to 12 mm., and they also are found more sparingly in St Andrews Bay. Similar forms occur in the nets in September and October, and since their length has not much increased it is probable that the larger examples have taken to the ground. A series of post-larval forms of this species, varying from 1 E. E. Prince, op. cit. p. 350, THE DHAGONKT FAMILY. 179 T\j inch to 4- inch in length have been obtained by the mid- water net in St Andrews Bay during August. Thus they still retain their pelagic habit, but a few are found near the rocks, shewing a tendency to assume the adult habitat. The smallest (^ inch) are mainly conspicuous for a large head and trunk and a long whip-like tail. The lower jaw projects at an angle (Plate VII, fig. 7). ' Head and body are speckled with brownish-black pigment-corpuscles, which attain their greatest development on the ventral surface of the abdomen1/ The predominance of a dark coloration on the under surface is rarely found amongst fishes, this part being usually white or pale, as it is also in the adult dragonet. When the little larva has nearly doubled its length (± inch) the head has become flatter, and by a growth outwards of the cheeks, the eyes are carried towards the top of the head, so that they look upwards more than laterally. The upper jaw has grown forwards and the arch of the mandible has increased. The dark pigment on the under side is more prominent than before. At a little over ^ inch in length, the same coloration obtains. The head and trunk are still broad when seen from above. A side view shews the fish to have a fusiform outline. The embryonic continuous median fin has broken up into two dorsals, a tail fin and an anal. The pectoral and ventral fins approach those of the adult in shape, no marked hypertrophy taking place in either pair. There is a spine upon each gill- cover which, at the stage of about J inch in length, becomes a formidable armature. By this time also the ventral fins have gained greatly upon the pectorals, being markedly longer. Shortly afterwards (Plate VII, fig. 8) the adult characters have been mostly assumed. The peculiar protrusible mouth is conspicuous. The paired fins are pale, except for a mass of dark pigment at the fleshy base of the pectorals. The lower surface ' is still coloured with black pigment/ and dark bands extend over the greenish back, one behind the head, and three across the trunk. ' The figure here referred to is from a sketch made some days after confinement in a glass tank in the laboratory, hence the coloration is modified. The 1 Researcites, p. 86-4. 12—2 180 THE DRAGON ET FAMILY. abdominal region is pinkish, from the contained food. The eyes are lustrous and greenish. In the laboratory the young fishes lay at the bottom, keeping the pectoral fins in active motion, while the ventrals were spread out like a pair of wings.' One specimen about | in. long, taken from the stomach of a cod, shews the spines on the gill-covers very well developed, and has a close similarity in general features to the adult. The cod is a ground-feeder and this specimen had no doubt taken to the habitat frequented by the adult, namely the sea-bottom in- shore. This example was obtained at the end of August and it was therefore not more than from 2 to 3 months old, though sup- posing that it was spawned in early May, there is a margin of nearly four months. The other specimens described, running down in a series to ^ inch in length are all evidently from the same season's eggs, and their diversity in size (though all occur in August) is to be accounted for, to a large extent at any rate, by the fact that the spawning of the dragonet extends over a considerable period. An example, mentioned in the Trawling Report, £in. long, occurred in the stomach of a haddock in October ; while others, from 50 to 78 mm., were found in the stomach of the cod, no date unfortunately being mentioned, though in all probability late in autumn. A young dragonet 2J inches in length, also without date of capture given, shews most of the adult characters. The gill-chambers open by the small opercular apertures situated dorsally, and the opercular spines con- spicuously project outwards and backwards on either side of the head. The three spines on each side are mounted on a bony pedicle which protrudes well away from the surface of the gill-cover, the whole forming a powerful weapon of defence. With growth in size the spines become less con- spicuous and more closely apposed to the surface of the gill-cover. A specimen of about 78mm. (3 inches) occurs from the West coast of Ireland in the month of July ; this is probably rather less than a year old, belonging to the last of the preceding season's brood. At this stage the trans- verse bands of dark pigment are still to be made out, especially THE SUCKER FAMILY. 181 in the dorsal region, but there is a tendency towards a breaking up of the bands into a number of isolated spots, so that the next stage has the uniform tint of the adult, with dark spots over the body and fins. Another of 5J inches from St Andrews Bay in November also belongs to the brood of the preceding season but will have entered its second year, as the breeding-season does not appear to continue beyond August. Its age might be from 15 to 20 months, and is probably nearer the latter estimate. Sexual maturity is reached in some females at a length of 6 inches. The adult seldom exceeds one foot in length, and is usually rather less. The Sucker Family. Discoboli. THE LUMP-SUCKER, (Cyclopterus lumpus, L.) The eggs of this species (Plate II, fig. 1) are very abundant along the rocky margins from February to the end of May1 or even somewhat later, one specimen having been procured on the 12th July, 1888. Their colour varies from reddish to straw-tint and even greenish, while in a ripe ovary in the body of the fish it ranges from a beautiful amethystine lustre through the various shades of dull yellow to greenish. They appear to vary in diameter from '09 — '1 inch, with numerous oil-globules, the largest of which are "041 inch in diameter. When the eggs are partly immersed in a crevice of the rocks and partly exposed, those on the latter surface are straw tinted, while those in the pocket of the rock are of a faint lilac hue. The capsule is dense and minutely punctured, but presents no special thickenings or superficial wrinkles, except where the facets of attachment are situated (Plate II, fig. 2). They are fixed together in considerable masses on the rocks, the water percolating readily through the interstices ; but the hatching in confined tanks is somewhat uncertain since decaying debris or the death of a few leads to putrefaction of the whole. 1 W. C. M. 3rd S. F. B. Report, p. GO. 182 THE SUCKER FAMILY. In the majority of instances the exposed surface of the masses of eggs presents peculiar and smoothly rounded depres- sions as if some of the eggs had been scooped out by a predatory fish or mollusk. Such depressions are bowl-shaped, that is, wider at the top or surface, and gently narrowing to the bottom. An attentive examination, however, shows no mark of injury, such as would be caused by the teeth of a fish or by the rasping ' tongue ' of a shell-fish. The eggs lie evenly together and their capsules are uninjured. In all probability, therefore, the depressions are due to pressure applied by a blunt surface immediately after deposition ; when the mass is soft, it may be by the snout or other part of the male as he fertilises them. The eggs are chiefly to be found from low-water mark to half- tide mark, often deposited in corners or in holes in rocks. They generally are exposed to the wash of the sea, at St Andrews for instance, chiefly facing the east. In this position they are eagerly eaten by rooks, starlings, and rats. The food-fishes and others are also extremely partial to them1. Thus, at the end of April, it occasionally happens that codling caught off the rocks have their stomachs distended with the eggs of the lump- sucker. Even such small fishes as Yarrell's blennies take the same food. The care which certain male bony fishes take of the eggs is well known, while Dr Giinther mentions only two cases in which females do so. In. this country the males of the river bull-head, the lump-sucker, and the marine and fresh-water sticklebacks are familiar instances. Most authors who have treated of Cyclopterus have observed this feature in the male; indeed, it is sufficient, under ordinary circumstances, to try to push him off guard with a stick to bring out the feature clearly. Various interpretations, however, have been placed on the habit, some supposing that the mere fact of the male being in the neighbourhood at deposition sufficed to account for its subsequent appearance near the eggs; while others, after Fabricius, bestowed considerable attention on the description of the instinct. In regard to the remarks 1 W.