Adequate measures for conservation of our fishery resources and the production of the maximum quantity of food with the minimum of expenditure through proper propa-gation methods require as their basis a reasonably complete knowledge of the life his-tories and habits of the fishes. The first step in this direction is the determination of the character of the eggs and young, so that they may be recognized at any stage of develop-ment. Many of our important marine food fishes--e. g., cod, haddock, hake, mackerel, and halibut--have floating eggs, which may be collected with an ordinary tow net with much less effort than is required to locate and capture the spawning fish. Thus this knowl-edge may serve to locate the spawning grounds and also the schools of spawning fish. The immediate value of knowledge of this character has been well illustrated by Dr. Hjort, of Norway. Knowing the character of cod eggs, he applied this method to the coast banks off northern Norway and thereby "succeeded in finding enormous shoals of cod on certain banks where no fishing was carried on, and where, as a consequence of our discovery, millions of cod were afterwards taken."

The present paper embodies the results of a study of the embryology and larval development of teleostean fishes taken in the region of Woods Hole, Mass., during July and August, 1915.

The majority of the species, especially the more important ones, common to this region spawn earlier in the season. Little is known of the breeding habits of the bonito (Sarda sarda), the menhaden (Brevoortia tyrannus), the butterfish (Poronotus triacanthus), and the hake (Urophycis chuss). Females of these species in which the eggs were nearly or quite mature and males from which the milt flowed freely were taken. However, all attempts at artificial fertilization failed. The eggs of the bonito and the hake were never taken in the plankton. With the exception of the menhaden, butterfish, and whiting, the eggs of all the species described herein were artificially fertilized and hatched in the laboratory. The eggs of the whiting were artificially fertilized, but all died during early cleavage. Both eggs and young of all the species with pelagic eggs were taken in the plankton.

The embryology and early larval life of several species included in this paper were early studied by Agassiz  (Agassiz and Whitman)^a. The embryology of one species, viz, Tautogolabrus adspersus, is described and illustrated in great detail by these authors in their fundamental work on the development of osseous fishes.b Their observations on other species included in this paper are more or less fragmentary. The pelagic eggs and larvae identified by them as Cottus groenlandicus are doubtless Prionotus carolinus, as a comparison of their figures with the figures here presented of the eggs and larvae of the latter species will show. The eggs and larvae described by them as species allied to Motella are probably eggs and larvae of the butterfish. Eggs apparently identical with the eggs described by them as those of the Sienna flounder were taken throughout July and August. They were taken in greatest abundance off Gay Head on August 24. These eggs were not identified by the present writers. Observations on the early development of one species (Apeltes quadracus) included in this paper are recorded by Ryder.^b

The observations recorded herein were made almost exclusively on living material. It is not the purpose of this paper to discuss in detail the embryological development of each species studied, but rather by means of illustrations and descriptions to afford a ready means of identifying eggs or larval fishes at any time during embryonic or larval life. With this purpose in view, the inclusion of several species upon which more or less complete observations have been previously recorded seems justifiable.
 

Spawning.--The principal spawning month for the tautog is June. Although the majority of the fish taken after July 1 were spent, eggs were abundant in the plankton as late as July 15. During the latter half of July they became gradually less abundant, but were taken in small numbers as late as August 20.

The tautog is prolific, but difficulty is experienced in obtaining eggs from captured fish in quantities sufficient for successful fish-cultural operations. Little difficulty, however, was experienced in obtaining and artificially fertilizing the eggs required for embryological study.

Eggs.--The eggs are highly transparent, spherical in form, and 0.9 to 1 mm. in diameter. The egg membrane is thin and horny. The yolk sphere contains no oil globule. The protoplasm which invests the yolk sphere in a very thin layer is finely granular and hardly perceptible until fertilization has taken place and the process of concentration that results in the formation of the blastodisc is initiated. As soon as fertilization has taken place a relatively small space, the perivitelline space, becomes apparent between the egg membrane and the delicate vitelline membrane which incloses the yolk sphere.

Blastodisc.--As soon as fertilization has taken place, the protoplasm becomes concentrated at one pole of the yolk sphere into a lenticular mass, the blastodisc. During this process the protoplasm slowly flows toward the pole of concentration. The "streaming" movements early described by Ryder a that occur in the protoplasmic layer during the process of concentration are less apparent in the eggs of this species than in the eggs of many other species of teleosts by reason of the extremely minute size of the protoplasmic granules. The process of concentration occupies less than one-half hour. The fully differentiated blastodisc (fig. I, BD.) comprises nearly all the protoplasm contained in the egg. It is circular in outline and of nearly uniform thickness throughout the central area, thinning out abruptly near the periphery. At the periphery it thins out gradually into a very thin layer of protoplasm, which continues to invest the yolk sphere.
 

Segmentation.--The first act of cleavage occurs less than one hour after fertilization. Later acts of cleavage follow each other in rapid succession. Blastoderms in advanced stages of cleavage may be observed within four hours after fertilization.

As the moment of cleavage approaches, one axis of the blastodisc becomes somewhat longer than the other. The first plane of cleavage cuts the blastoderm at right angles to the longer axis (fig. 2). The second plane of cleavage cuts the first at right angles.

During the four-cell stage (fig. 3) the two axes of the blastoderm are approximately equal. The third planes of cleavage cut the blastoderm approximately parallel with the first (fig. 4). As the third act of cleavage occurs one axis of the blastoderm again becomes distinctly longer than the other (fig. 4). Typically, the eight blastomeres formed by the third act of cleavage lie in two symmetrical series of four cells each. As the fourth act of cleavage occurs, the two axes of the blastoderm again become approximately equal. The blastoderm now becomes more or less circular in outline and approaches true radial symmetry more and more closely as cleavage advances.

The first two or four blastomeres are usually approximately equal in size and quite symmetrical. As the third act of cleavage occurs, symmetry is usually disturbed. Early blastoderms of more than four cells show a marked lack of symmetry and frequently some disparity in the size of the constituent cells. However, blastoderms of 8, 16, and 32 cells are found occasionally which remain almost ideally symmetrical. Beyond the 64-cell stage symmetry or lack of symmetry in the arrangement of the cells is not easily observed. Blastoderms in advanced stages of cleavage usually appear radially symmetrical.

Formation o/ the periblast. The cells at the margin of the blastoderm are not sharply limited peripherally, but remain continuous with the thin layer of protoplasm at the surface of the yolk. As segmentation advances this layer of protoplasm becomes concentrated at the periphery of the blastoderm into a somewhat flattened protoplasmic ridge that gives rise to the periblast (fig. 5, PB). Before this ridge of protoplasm has become fully differentiated, nuclei become apparent near the margin of the blastoderm and gradually become distributed throughout the entire protoplasmic ridge. The periblast nuclei as observed by Agassiz and Whitman,a doubtless are derived from the peripheral cells of the blastoderm. When fully differentiated the periblast consists of a flattened syncytial ridge of protoplasm with nuclei apparently like those of the cells in the blastoderm distributed throughout its entire extent.
 

Until nuclei are present throughout the peripheral area of the periblast it remains continuous with the peripheral cells of the blastoderm. As segmentation advances further the peripheral cells of the blastoderm become completely cut off from the periblast. A thin sheet of protoplasm, the central periblast, which is also invaded by nuclei now advances centripetally from the periblast beneath the blastoderm.

During early cleavage the blastoderm is essentially a lenticular mass of cells. As segmentation advances, it becomes distinctly dome-shaped, leaving a cavity beneath its central area. This cavity, which is the cleavage cavity, now lies between the blastoderm and the central periblast.

Formation of the germ ring.--The germ ring, when fully differentiated, appears as a thickened peripheral zone of the blastoderm (fig. 6, GR). This zone becomes roughly outlined before the marginal cells of the blastoderm are completely cut off from the periblast. The thickening is at first more apparent than real, being due primarily to the thinning of the central area of the blastoderm, by reason of which its under surface becomes concave. After the blastoderm is completely cut off from the periblast, cells at the periphery grow inward (invaginate), thus adding somewhat to the thickness of the germ ring. Before invagination begins the cells forming the surface layer of the blastoderm become distinctly flattened. This layer plays no part in invagination. The cells which grow inward from the periphery are derived from the deeper layers. The full extent of the ingrowth of cells from the periphery of the blastoderm can not be determined in living material. For a detailed discussion of the role of invagination in the formation of the germ ring and the embryonic shield based on a careful study of histological sections the reader is referred to Wilson's paper^c on the embryology of the sea bass? As the blastoderm gradually grows larger the germ ring, which in its earlier stages involves but a narrow zone, increases somewhat in width by the centrifugal growth of the blastoderm as well as by the invagination of the marginal cells.

Formation o[ the embryonic shield and differentiation o[ the embryo. Before the germ ring is fully differentiated it becomes apparent that invagination advances more rapidly at one pole than rottad the rest of the periphery of the blastoderm. This is the posterior or embryonic pole (fig. 6, PP). At this pole a broad tongue of cells is pushed forward into the cleavage cavity. Viewing the blastoderm from above, there soon appears at the posterior pole a roughly triangular area which is obviously thicker than the adjacent areas. This triangular area marks an early stage in the differentiation of the embryonic shield (fig. 6, es).

The blastoderm now increases in size more rapidly than in the earlier stages, and the germ ring gradually advances arotmd the yolk sphere. As the blastoderm spreads over an increasingly greater area of the surface of the yolk, the embryonic shield grows larger and becomes more definitely outlined. Soon there occurs a linear thickening along its anteroposterior axis that marks the axis of the future embryo (fig. 7, EA). The embryonic shield is now differentiated into an embryonic and an extra-embryonic area. The further differentiation of the embryo begins in the anterior or head region and gradually advances posteriorly. Before the embryonic axis is well differentiated, the blastoderm covers more than half the surface of the yolk sphere, and the circumference of the germ ring is actually decreasing. As development advances much of' the material contained in the germ ring becomes incorporated in the embryo. The part played in this process by concrescence in the sense of His ^d and confluence in the sense of Sumner ^e can not be discussed in this paper. This entire process is doubtless but a part of a larger process by which much of the material contained in the embryonic shield becomes incorporated in the body of the embryo.

By the time the embryo is well formed the blastoderm covers approximately three-fourths of the surface of the yolk sphere. As development advances the blastoderm soon covers the entire yolk sphere and the blastopore is closed.

The closure of the blastopore occurs within 18 hours after fertilization. At this time the embryo extends approximately halfway round the circumference of the yolk sphere and segmentation of the body has already begun. Figure 9 illustrates an egg shortly after the blastopore is dosed. The embryo remains highly transparent and shows no evidence of pigmentation. The beginning of pigmentation is observed in embryos with 15 to 20 somites. The chromatophores first appear as minute rounded black dots scattered over the dorsal aspect of the embryo. As the time of hatching approaches, the chromatophores become somewhat larger and show irregular pigmented processes. However, the embryo remains highly transparent (fig. 10). The extra-embryonic blastoderm remains free from pigment.

Larval development.--Incubation at laboratory temperature i.e., in water at approximately 22° C.--occupied 43 to 45 hours. In the tidal hatching boxes at the same time incubation occupied approximately 48 hours.

The newly hatched larvae (fig. 11) are approximately 2.2 mm. in length. The head is slightly deflected. The yolk sac remains relatively large. It is ovate-elliptical in form and free from pigment. The vent is located at some distance from the posterior margin of the yolk sac and a little more than half the length of the body from the anterior end. The depth of either dorsal or ventral fin fold is less than the depth of the body just posterior to the vent. The chromatophores have grown somewhat larger, but have

not increased materially in numbers. They remain confined more or less closely to the dorsal and dorsolateral aspects of the body. The fin folds and the posterior caudal region of the body remain entirely free from pigment.

One day after hatching (fig. 12) the larvae have grown to a length of 2.8 to 3 mm. The yolk sac is greatly reduced and the head is no longer deflected. The chromatophores have increased materially in size and show well-developed pigmented processes, but are apparently fewer in number than in the newly hatched larvae. Individual pigment cells, doubtless, have become intimately associated with each other to form larger chromato- phores. The larvae now have a distinctly blackish color. Four days after hatching (fig. 13) the larvae have grown to a length of 3.2 to 3.5 mm. The yolk is completely absorbed. Larvae kept in dishes of sea water as well as those hatched in the tidal hatching boxes now begin to die rapidly. The critical period for this species, therefore, comes about the fourth day after hatching. At this stage black chromatophores are more or less uniformly distributed over the dorsal and lateral aspects of the body. However, the posterior caudal region remains free from pigment.

Figure 14 illustrates a young fish 5 mm. in length taken in the plankton. In young fishes at this stage growth is indicated more especially by the increase in the depth and thickness of the body than by the increase in length. The distribution of pigment remains essentially the same as in larvae four days after hatching. However, the chromatophores are larger and have increased materially in numbers.

In young fish 10 mm. in length (fig. 15) the dorsal, anal, and caudal fins are becoming well differentiated. The distribution of pigment remains essentially as in the early stages. However, the number of chromatophores, as well as the quantity of pigment, has materially increased.

As development advances the young fish gradually assume adult characters. Young fish 30 ram. in length (fig. 16) exhibit nearly all the diagnostic characters of the species. The depth of the body in proportion to its length is rapidly increasing and the back is becoming strongly arched. The ground color of the body at this stage is greenish. The black chromatophores have become aggregated to form heavily pigmented areas, which are roughly arranged in transverse bands and give the body the transversely banded appearance characteristic of the adult.

^b Ryder. J. A.: A contribution to the embryography of osseous fishes * * * . Report United States Fish Commission 1882, p. 455-605.
 

^c Wilson, H, V.: The embryology of the sea bass (Serranus atrarius). Bulletin United States Fish Commission, vol. IX, x889, p. :2~77.

^d His, W.: Zur Frage der Langsverwachsung von Wirbelthierembryonen. Verh. d. anat. Ges., I89r, p. 70-83.

^e Sumner, F. B.: Kupffer's vesicle and its relation to gastrulation and concrescence. New York Academy of Sciences. Memoirs. vol. 1I, pt. n, z9oo, p. 47-8a.