ON THE 
DEVELOPMENT 
OF 
VIVIPAROUS OSSEOUS FISHES 
AND OF 
THE ATLANTIC SALMON. 
BY 
JOHN A. RYr>ER. 
[EXTRACTED FROM PROCEEDINGS UNITED STATES NATIONAL MUSEUM, 1885, 
PAGES 128-162, PLATES VI-XIL] 
WASHINGTON: 
GOVERNMENT PRINTING OFFICE. 
18 8 5. ON THE 
DEVELOPMENT 
OF 
VIVIPAROUS OSSEOUS FISHES 
AND OF 
THE ATLANTIC SALMON. 
BY 
JOHN A. RYDER. 
[EXTRACTED FROM PROCEEDINGS UNITED STATES NATIONAL MUSEUM, 1885, 
PAGES 128-162, PLATES VI-XII.] 
WASHINGTON. 
GOVERNMENT PRINTING OFFICE. 
1885. 128 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
ON THE DEVELOPMENT OF VIVIPAROUS OSSEOUS FISHES. 
By JOHN A. RYDER. 
INTRODUCTORY. 
The following paper is intended to give a summary of our knowledge 
respecting the best known of the truly viviparous osseous fishes char- 
acterized by an intrafollicular or intraovarian development. The only 
memoirs of importance which have not been consulted in the prepara- 
tion of this series of notes is that by Rathke,* on Zoarces viviparus, and 
that by Duvernoy on Pcecilia, surinamensis, which is cited at another 
place. 
The preliminary notice published by me on the development of Gam- 
•busia appeared nearly four years ago, and left the subject in a very 
incomplete state. What is here added on the anatomy of the embryos 
;and the follicles may, I trust, be of interest to morphologists. 
The new observations here recorded on the changes undergone by 
the embryos of the Ebiotocoids during gestation, relate to the devel- 
opment of the intestine and the vascular supply of the median fins, 
both of these organs also undergoing other changes which have appeared 
to the writer to have considerable significance. 
Some apology may be necessary for the long extracts which I have 
incorporated from the paper by Professor Wyman and Dr. Girard's re- 
port, but I think the student who wishes to be spared the vexatious 
task of finding and consulting authorities will be rather gratified than 
■otherwise to find the American contributions to the subject thus brought 
together. 
I.-The development of Anableps. 
Wyman's memoir f on Anableps contains such valuable observations 
;and reflections on the viviparity of fishes that I will here reproduce the 
most important parts of his paper entire, as follows: 
Extended observations have proved that a large number of species 
•of fishes, belonging to many genera, are truly viviparous, the foetus 
passing through a real gestation by the parent before its development 
is complete. These viviparous fishes may be divided into two groups, 
according to the position occupied by the embryo during the period of 
its growth. 
"I. In the first group may be arranged those fishes, in which the 
* Bildungs- und Entwickelungsgeschichte des Blennius viviparus oder des Schleim- 
stisches. Abhandl. zur Entwickelungsgeschichte. Zweiter Theil, Erste Abh.,pp. 1-68 
pls. 5, 4to, Leipsic, 1833. 
t Observations on the development of Anableps Gronovii (Cuv. & Vai.). By Jeffries 
Wyman, M. D., Bost. Journ. Nat. Hist, vi, 1850-'57, pp. 432-443, pl. xvii. 1885.] 
PROCEEDINGS OE UNITED STATES NATIONAL MUSEUM. 
129 
Vol. VIII, Io. 9. Washington, 1>. €. Hay 25, 1885. 
ovum leaves the ovary in an undeveloped state, and in which the process of 
evolution is not commenced until it reaches the lower portion oftheovidtict. 
The species which this group comprises are nearly all, if not all, Pla- 
giostomes. The best known are Spinax, Carcharias, Mustelus, Galeus, 
and 'torpedo. Although they are usually classified among the lowest 
of fishes, it is in some of them that the process of reproduction be- 
comes most nearly analogous to that of the highest Vertebrates. Not 
only does the yelk reach proportions like the yelk of birds, but the 
yelk-sac itself plays the part of an allantois and forms an organ analo- 
gous to a placenta. In Spinax the vessels on the surface of the vitel- 
line sac are brought into close contact with the highly-vascular folds 
which line the oviducts. But in Carcharias, as Muller has demonstrated 
in his memoir on the subject, not only is there an approximation of the 
foetal and maternal vessels, but the surfaces of the yelk-sac and of the 
oviduct are both deeply convoluted, and the projections of the one are 
admitted into and embraced by the concavities of the other, and the 
opposing surfaces become adherent even. In both Spinax and Car- 
charias the necessary conditions exist for the reaction of maternal 
and foetal blood upon each other, as in the case in the mammalia, but 
to a much more limited extent.* 
'{H. In the second group those fishes are comprised in which the ges- 
tation is either wholly or in part ovarian, the last stages only of the pro- 
cess usually occurring in the oviduct. Among the genera included in 
this division are Silurus,] Blennius,$Anableps,§ Pcecilia,\\and Embiotocaf\ 
In all of these genera impregnation takes place in the ovary, and, as 
seems probable, while the ovum is still invested with its original envel- 
opes. In Blenny, Rathke has shown, the ovarian gestation having con- 
tinued about three weeks, that about the end of September the sac rup- 
tures, and that the embryo is discharged into the central cavity of the 
ovary, which is in fact the oviduct; here the foetus remains till the be- 
ginning of January, when it is born. In Poecilia the foetus is liberated 
and escapes into the oviduct towards the end of gestation. Valenciennes 
has given several details in relation to the development of Anableps 
Gronovii, made for the most part upon specimens in an advanced stage 
of foetation, the smallest embryo being more than an inch long. He 
* Dr. John Davy has shown that in Torpedo the embryo is nourished at the expense 
of materials furnished by the parent, since the mature fcetus weighs more than twice 
as much as the egg at the time development commenced. Philos. Trans., 1834. On 
the Development of the Torpedo. 
t Cuvier et Valenciennes, Hist. Nat. des Poissons, t. i., p. 540, 1828. 
+ Rathk6, Mem. sur la Develop, de 1'Homme et des Animaux, 2me partie, Leipsic, 1833. 
§Cuv. and Valenciennes, Hist. Nat. des Poissons, t. xviii, p. 245, Paris, 1846. 
HDuvemoy, Ann. des Sc. Nat., t. i, 3e, set., p. 313, 1844. 
T Agassiz, Am. Journ. of Science, xvi, 2d set., Nov., 1853. 
Proc. Nat. Mus. 85 9 130 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM 
[1885. 
found only seven or eight foetuses in the so called uterus of each female, 
and each of the young was surrounded by a distinct sac, which he re- 
gards as simply an enlargement of the original envelope of the ovum- 
The mature foetus he found to be more than one-fourth as long as the 
parent, and except for the non-development of the ovary was constituted 
in every respect like the adult, as regards both its internal and external 
structure. 
" For the specimens of Anableps Gronovii upon which the following ob- 
servations were made, I am indebted to the liberality of Dr. Francis W. 
Cragin, United States consul at Paramaribo, in Surinam. Among them 
were three males and five females, four of which last were in different 
stages of gestation. The different individuals varied from 3| to 
inches in length, the females being much longer than the males. 
"I. The smallest female measured 3| inches in length, but on careful 
examination no traces of an ovary were discovered; its development did 
not appear to have commenced as yet. 
"II. The next specimen measured 7 inches in length and the ovary 
was in a state of gestation; the foetuses, four or five in number, meas- 
ured but five-eighths of an inch. The ovary appeared single externally ; 
was invested with peritoneum, which was supported by a more firm but 
thin membrane of condensed areolar tissue; on cutting through this the 
interior was found filled with sacs corresponding in number to the foe- 
tuses and united to each other and the ovarian walls by a very loose 
areolar tissue. They had no communication of any kind with each 
other. With the aid of the point of a needle the sacs were easily de- 
tached and removed entire with the inclosed foetus; the envelope was 
much larger than was necessary to hold the embryo, and the space be- 
tween the two was filled with a fluid, a portion of which (albumen ?) had 
been coagulated by the action of the alcohol. In each instance it was 
ascertained that the young had no connection whatever, vascular or 
otherwise, with the walls of the sac which inclosed it. 
" The external characters of the embryo even at this early stage, as 
regards its general form and the fins, resemble those of the adult; but 
no longitudinal black bands were yet visible on the sides; the eye had 
not acquired the prominence of that of the adult, the cornea was not 
divided by a transverse band, and the pupil existed in the form of an 
oval, with its long diameter in a vertical direction, but the sides of the 
iris had just commenced extending towards the center in order to form 
the two laminae, which in the adult give the pupil its singular shape. 
The umbilical sac forms a spheroidal mass about one-fourth of an inch 
in diameter, and is sufficiently transparent to allow the folds of the in- 
testine which fill it to be visible. Externally the sac is covered with 
what appear to be parallel projecting lines, extending from the sides of 
the abdomen to its most prominent part. These Valenciennes describes 
as 'vascular strue' (stries vascutiformes).* Such was not the nature of 
* Op. cit., ex fol. de planche, 539. 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
131 
these markings in the specimens which I examined; but, when placed 
under the microscope, were found to form a peculiar structure, which 
possibly may have some connection with the process of nutrition in their 
peculiar mode of gestation. 
"The sac itself seemed nearly homogeneous in structure, but the striae 
are made up of spherical, or in some cases pyriform or cylindrical, 
papillae or villi, projecting from the surface and arranged so nearly 
together in a linear series as to give the appearance of a continuous 
band. In regard to the minute structure of these papillae, as far as it 
could be determined from an alcoholic specimen, they consist externally 
of an exceedingly thin membrane, inclosing a vast number of minute 
granules; no vessels were seen in connection with them nor in the mem- 
brane to which they were attached. There was no appearance of any 
communication between the cavity of the papilla and that of the mem- 
brane to which it was attached. Within the latter, but more nearly in 
contact with the intestines, there was a second more delicate membrane, 
which seemed to be a continuation of the parietal peritoneum. No 
traces of the yelk were found in connection with the intestines. 
" III. The third specimen is much longer and measures 10 inches in 
length ; the ovary had been ruptured, so that some of the foetuses had 
escaped into the cavity of the abdomen, but the whole number of young 
taken from the parent was much greater than in the preceding case, 
namely, eighteen ; one of these was projecting from the genital opening. 
They generally resembled the preceding except in size, though the eyes 
had become more prominent, and the iris now exhibited its lateral pro- 
jections sufficiently developed to give the pupil the shape of a dumb- 
bell. The umbilical sac has become much larger than in the embryos 
first described, and measures three-eighths of an inch in diameter. The 
papillae of the yelk-sac are much more distinct, and contain colored 
granules. The yelk-sac communicates with the cavity of the abdomen 
by a long fissure extending from a point just behind the union of the 
opercula nearly as far as the anal opening, consequently beyond the ven- 
tral fins. It is from the circumstance just mentioned, doubtless, that 
an explanation is to be found of the non-union of the ventrals in the 
adult. The scales terminate abruptly at the edge of the fissure. The 
intestine, as in the first described embryos, were invested by the internal 
sac, which was regarded as parietal peritoneum. No bands were visible 
on the flanks of the body, nor were the anal fins yet modified to mark 
the sexes. A rudimentary liver is visible in these specimens, extending 
backwards on the left side of the intestinal mass. The intestinal canal 
is of almost uniform size throughout, there being no distinction between 
intestine and stomach. 
" IV. This specimen measured a little less than 10 inches in length, but 
the embryos were of much larger size, having a length of inches; 
the umbilical sac had disappeared, but the fissure on the under side of 
the abdomen still remained, and, what seems quite remarkable, had 
grown in dimensions just in proportion to the entire fcetus, so that in 132 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
these specimens it measured 1 inch in length, and was consequently 
longer than the whole embryo of the first specimen noticed above. The 
edges of the fissure were united by the intermedium of a thin membrane, 
without scales, on which no papillae were noticed, and was sufficiently 
lax to allow the edges of the fissure to separate from each other to a 
slight extent. The transverse band upon the cornea was now distinct, 
though it had not yet become as opaque as in the adult. 
"Seven foetuses were found in the ovary; on the sides of them one or 
two dark longitudinal 1 nes were now visible; the general form of the 
body had assumed more precisely that of the adult, and, as noticed by 
Valenciennes, the intestines had obtained their permanent form. The 
external sexual characters were not visible in any of the specimens ex- 
amined, though they were seen and figured in specimens of about the 
same size by Valenciennes. 
"All of the foetuses of this female had escaped from their original 
sacs (no traces of which were now visible) and were all contained in one 
large cavity formed by the dilated ovary which now had become analo 
gous to an uterus and extended from the genital opening as far forwards 
as the bases of the pectoral fins. The walls of this ovarian sac were suf- 
ficiently thin to allow the foetuses to be seen through them. On its in- 
ner surface, as well as on that of some of the other specimens, were to 
be seen numerous immature ova, some of them microscopic and others 
as large as the sixteenth of an inch in diameter. The co-existence of 
immature ova on the walls of the ovarian cavity with foetuses in it cor- 
responds with what was noticed by Duvernoy in his investigations of 
the embryology of Pcecilia.* The more minute ovarian eggs, though 
for a long time macerated in alcohol, yet preserved their microscopic 
characters to a remarkable degree. The smallest consisted of a cell, in 
the center of which a nucleus was visible, and around this last were a 
few granules. In the larger ova the granules have become more and 
more abundant, and in some instances obscure the nucleus or germ-cell. 
After the egg has increased to a certain size, a clear space appears ex- 
terior to the vitelline membrane, which gradually increases to nearly 
twice the diameter of the egg itself. This clear space is limited by the 
substance of the stroma, which becomes condensed around it and thus 
forms a distinct sac. If the ovum be compared to that of a mammal, 
then the sac just described may be said to be analogous to a Graafian 
vesicle; that is, the egg of the fish floats free in a sac much larger than 
itself, just as the mammiferous egg does in the vesicle of De Graaf. 
There were no intermediate conditions between this and the impreg- 
nated condition to enable me to determine whether or not it is this sac 
which foimed the external covering of the foetus. Valenciennes seems 
to adopt the idea that it does, and compares it to a chorion.t If this view 
* Sur la developpement de la Pcecilia surinamensis, Ann. des Sci. Naturelies, 3e ser., 
i, p. 313, 1814. 
tLa cellule qui contient un oeuf fdcondd s'aggrandit et finit par former une sorte 
de chorion. Op. cit., t. xviii, p. 261. 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
133 
of its nature be true, then there seems no alternative, since develop- 
ment advances so far before the sac ruptures, but to suppose that im- 
pregnation must take place through its parietes and that the sperma- 
tozoon cannot enter bodily into the substance or even come in direct 
contact with the vitelline membrane of the egg except through the walls 
of its outer covering, which is not probable. It would seem that it must 
act simply by its presence on the surface of the egg-sac or by an endos- 
mosis of its fluid contents through the membranes by which the ovum 
is invested.* 
"A microscopic examination of the egg-sacs in the advanced foetuses 
proves conclusively that they do not consist of loose areolar tissue only, 
as stated by Valenciennes,! but that while the tissue in question forms 
the basis of them, they are in reality highly vascular, large trunks and 
minute ramifications of vessels being easily traced by the aid of the 
coagulated blood which they contain. 
" In comparing foetuses of different stages of development together, 
a very interesting question is presented to us in connection with their 
growth. In the smallest specimen examined, the yelk was no longer 
visible; it had been wholly consumed in supplying materials for the for- 
mation of the embryo; and yet subsequent to this disappearance of the 
yelk, the embryo, while still in its ovarian sac and cut off from all ex- 
ternal communication, continues to increase in size, and grows until it 
acquires the length of an inch and a quarter, which gives the size of 
the longest foetus which our specimens furnished. Even the umbilical 
sac and the fissure which succeeds it continue to grow after the yelk 
has disappeared. As a geueral rule among oviparous fishes, the yelk 
supplies all the material required for the growth of the foetus; and the 
same holds good with regard to nearly all Batrachians,t to scaly rep- 
tiles and birds. So general has this rule been believed to be, that none 
but mammals have been supposed to contribute anything beyond the 
materials of the egg to the support of the young. But recent observa- 
tions go to prove that some fishes, such as the Torpedo among the 
Plagiostomes, the Embiotoca among osseous fishes, are to be placed in 
the same category as mammals, in relation to the fact of being nourished 
by the parent during gestation, although neither a placenta is formed nor 
does any direct vascular communication whatever exist between the 
foetus and the maternal circulation. We cannot explain the growth of 
the foetal Anableps by any other hypothesis than that it is nourished by 
a fluid secreted by the walls of the sac in which it is lodged in the 
earlier stages, or by the parietes of the general ovarian cavity in which 
the foetuses are received towards the end of gestation. The high de- 
gree of vascularity of the egg-sac is favorable to this supposition. As 
[*See description below of the follicular pore in Gambusia.] 
t Op. tit., p. 261. 
t The only exception among Batrachians, as yet noticed, is found in the Pipse of 
South America. See Observations on Pipa Americana, by Jeffries Wyman, M. D., in 
American Journal of Science, 2d series, vol. xvii, p. 369. 134 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
the body of the foetus at a very early period becomes covered with scales, 
absorption could ouly take place through the intestinal canal or by the 
surface of the yelk-sac, which invests the viscera and increases in size 
for a long period after the yelk itself has wholly disappeared. In the 
later stages of gestation even the yelk-sac is out of the question, since 
it in turn wholly disappears, while the foetus occupies the general cav- 
ity of the ovary." 
Great interest attaches to the peculiar development of the rows of 
papillae on the empty yelk-bag of Anableps as described above by Wy- 
man. Its continued growth with the growth of the foetus up to a cer- 
tain stage of advancement is also remarkable; in fact, so far as yet 
known, it is unique amongst osseous fishes. The peculiar character of 
the villi on the yelk-bag remind one somewhat of what has been ob- 
served by Osborn in the structure of yelk sac of the uterine foetus of 
Didelphys, and by Owen in Macropus, in which the foetal membranes 
fitted into uterine furrows, but were not adherent to the uterus and 
without villi. This structure in Anableps also reminds one of the hol- 
low villi developed in certain mammalian blastodermic vesicles de- 
scribed by Bischoff. It is therefore unfortunate that Professor Wyman 
was not in a position to describe the minute histological structure of 
this abdominal sac in the foetus of Anableps more fully, so that a more 
exact comparison could have been instituted between it and the yelk- 
bag or inferior pole of the blastodermic vesicle of the Marsupialia. A 
knowledge of the embryonic layers which enter into the formation of 
the yelk-bag of Anableps would also be a desideratum. 
The extension downwards of the intestine of the embryo of Anableps 
into the empty yelk-bag is also interesting and reminds one somewhat 
of the peculiar protrusion of the hind-gut of Embiotocoid embryos into 
an inferior saccular diverticulum of the back part of the abdomen, 
which is obviously uot homologous, however, with the globular, bag- 
like structure seen in the former. 
Wyman does not state whether he examined the surfaces of the ova- 
rian membranes in the earliest stages of development of the papillifer- 
ous sac to see if they did not present pits or crypts into which the 
papillae may have fitted. These papillae are obviously in some way con- 
nected either with the respiration of the foetal Anableps or with its nu- 
trition in the same way as are the marginal lobes of the fins of the foetal 
Embiotocoids. * 
II.-The development of the Viviparous Surf-perches or Em- 
BIOTOCID35 OF THE PACIFIC COAST. 
Dr. Girard* states that the discovery that these fishes were vivipa- 
rous was made in May, 1852, in San Diego Bay, California, by Dr. 
Thomas IT. Webb, whose manuscript journal is quoted by Girard. 
*Explorations and Surveys for a Railroad from the Mississippi River to the Pacific 
Ocean. IV. Fishes- 4to. Washington, 1858, 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
135 
The following observations of his own have been put upon record in 
the latter's report, cited below; but it is evident that he overlooked 
the peculiar vascular lobes appended to the vertical fins of the embryos 
noticed by Blake, and also drew some erroneous inferences as to the 
identity of the yelk-sac, which will be noticed later. From pages 165 
to 166, inclusive, of Girard's report the following paragraphs are 
quoted: 
"But the most remarkable trait in the organization of the fishes of 
this family consist in the mode of reproduction. The eggs, instead of 
being laid, as is the case in most fishes, are retained within the body of 
the female, where they undergo not only their embryonic growth, but 
likewise a growth which might be termed larval, it being subsequent 
to their escaping from the egg envelope, until they have attained a 
size sometimes of several inches in total length. Nevertheless this 
peculiarity of habit is not altogether without analogy in the class of 
fishes. Moreover it bears no resemblance whatever to the marsupia- 
lian gestation; in the first place the eggs develop in the ovary, not in 
a uterus, and there is no placental connection of any sort; secondly, 
the young are not at liberty to quit and enter the ovarian cavities alter- 
nately and at pleasure. 
" The observations which we have made upon the genital apparatus 
of the female have satisfied us that there exists an ovarian sheath or 
sac, which during the early period of pregnancy is an elongated and 
subcylindrical tube, containing the ovaries proper, two in number, 
each of which consist of two, three, or more vascular membranes, at- 
tached by their upper edges to the upper floor or roof of the sheath, 
forming either one or two pouches (according to the number of these 
membranes) of the same length as the sheath itself, widely open be- 
neath, though not in direct communication with one another, since the 
membranes hang loosely down, reaching the lower floor of the sheath. 
" The eggs are formed within the texture of the ovarian membranes 
themselves. We have examined the ovaries of Ennichthys heermanni 
and Embiotoca argyrosoma when the sheath within which they were 
contained was not larger than an ordinary quill. Numerous eggs could 
be observed in a very immature state, appearing to the unarmed eye 
like minute dots. Under the microscope they exhibited evident traces 
of the germinal vesicle, surrounded as yet with a very scanty supply of 
vitelline substance. 
" The sheath and the ovaries are gradually increasing in bulk, as 
the eggs themselves first increase in size and the embryos afterwards. 
The sheath is chiefly a muscular membrane, while the ovaries, we have 
stated, are altogether vascular. 
" When mature the eggs either fall into the space between the mem- 
branes or ovarian pouches, or else remain attached to the ovaries until 
the embryos issue out of them. We are inclined to think that they 
drop into the pouches as eggs. At any rate we found very young enp 136 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
bryos loosely contained in the ovarian pouches when no trace of the egg- 
membrane could be seen within the tissues of the ovaries in the shape 
of a corpus luteum or a Graafian vesicle. Whatever be the case, numerous 
eggs or embryos may be observed within one pouch. The young thus 
remain together until grown to a considerable size, when, filling up the 
space in a more compact manner, the ovarian membranes, in their 
nature very expansive, will extend a fold between each embryo. In 
this manner every individual young, when removed sideways from the 
ovary, appears to the operator as though inclosed in a separate cavity, 
pouch, or follicle of the ovary, whilst in reality the membranes may be 
stretched out or extended, and the entire progeny loosened from all 
adherence or connection with them. 
"The male organs of generation consist of two spermaries, a right 
and a left, perfectly independent from one another, having each its sep- 
arate duct, discharging their contents into an elongated cloaca, into 
'which the bladder likewise empties its contents. This cloaca commu- 
nicates with the exterior by a subcircular opening, the edge of which 
is rather protruding. Such is that apparatus, the same in its general 
structure as in the other osseous fishes. There is no sheath inclosing 
the two spermaries, and this fact throws considerable light upon the 
morphology of the ovaries; the latter being, in fact, two in number, but 
so closely connected together as to simulate a single organ. Thus the 
general disposition, not the plan of structure, of these organs is adapted 
to the mode of reproduction; a single sheath being a more simple 
adaptation than two, one for each ovary. 
" How the mechanical act of fecundation takes place we are not pre- 
pared to say from direct observations; the eggs themselves must be 
fecundated within the ovarian sheath; a copulation of some sort is, 
therefore, required, and it is not improbable that at this period the 
eggs have dropped from the ovarian membranes into the pouches or 
spaces between these membranes in which they are freely floating." 
Then follow some observations on the development of the embryos 
which are to a great extent erroneous, the " abdominal bag," alluded 
to as the yelk-sac, being probably the same as what I have determined 
to be the greatly hypertrophied hind-gut which protrudes from the ven- 
tral aspect of the body of the embryos. The cleft of the mouth, it is also 
stated, is not apparent at the time the fins began to grow out. This is 
also to be seriously doubted, and is a statement which probably rests 
upon defective observation, as I find the vent to be perforate at the time 
the embryos are an inch in length. 
Farther on Dr. Girard, speaking of the embryos of Ditrema jacksoni, 
six to seven tenths of an inch long, " The yelk-bag was still to be ob- 
served in the shape of a hernia under the abdomen." 
In Embiotoca perspicabilis, Girard states, op. tit., page 179, that " a 
female on being opened was found to contain about eighty young of 
an average size of half an inch, the whole embryo consisting of cells 1885 ] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
137 
with no signs of the mouth. A layer of black pigmentum constituted 
the eye. Membranous ridges above and below showed the first steps 
in the formation of fins. The caudal itself was a mere membranous ex- 
pansion of the cellular substance of the body. As such, the embryos 
had made their escape from the eggs." 
Under Phanerodon page 185, Girard again mistakes the pro- 
truding hind part of the intestine for a yelk-bag. On page 195, in em- 
bryos of Holconotus from three-quarters to an inch in length, 
he speaks of this part of the intestine as " a vitelline abdominal sac." 
Au examination of the figures of the immature ovarian ova of Embio 
tocoids has served to convince me that those figured by Dr. Girard were 
very far from mature, and can give us but a slight conception of what 
they are like when ripe, because he represents the germinative vesicle 
as still central. .The general rule that the nucleus breaks up and leads 
to the formation of a new nuclear center in an eccentric position in the 
egg will probably be found to hold in the Embiotocoidsas in other fishes.* 
As Girard states, I find them embedded in the substance of the thick 
membranes which depend from the roof of the ovarian cavity. 
The youngest foetus figured by Girard is also pretty well advanced, 
all of the vertical fins being already defined. The comparison also of 
these youngest foetuses hitherto figured with some somewhat older 
which I have had the opportunity to investigate, leads me to believe 
that the former were much farther advanced than was supposed by that 
author. His figures of the foetuses are poor, and give a very imperfect 
idea of what it must actually have been possible to see. His statement 
that the mouth was still closed I am also inclined to doubt, while his 
observations on the development of the eye are simply calculated to 
force a smile. He has clearly mistaken the protruding and hypertro- 
phied hind-gut in all of his figures for a yelk-bag, the latter having prob- 
ably vanished long before, and at a point somewhat in advance of what 
he regards as the yelk. 
III.-The hypertrophied hind-gut of Embiotocoid embryos 
AND ITS SUBSEQUENT DIMINUTION IN RELATIVE SIZE. 
The hypertrophied hind-gut of Embiotocoid embryos which protudes 
into a sac-like downward projection of the abdominal profile is the 
most important of the embryonic visceral organs. Upon opening the 
abdominal cavity it is found that it fills up fully two-thirds of the lat- 
ter, and that its dimensions, especially the transverse diameter of the 
gut, is far greater than the portion of the canal which subsequently be- 
comes the stomach. Upon cutting this swollen hind-gut open its walls 
are found to be thickly clothed with crowded villi of the most extraor- 
dinary length, many being entangled together at their tips, and much 
* See the law of nuclear displacement and its significance in embryology, Science, 
i, 1883, pp. 277, 278. [1885. 
138 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
bent. If extended some of these villi would more than reach across the 
lumen of the intestine. This is their condition in the foetuses of several 
species of Embiotocoids examined by me ranging from 1 to inches 
in total length. 
As development proceeds the gut begins to assume its normal dimen- 
sions posteriorly. It becomes narrower, and the sigmoid turn or flexures 
of the intestine more pronounced. The villi also diminish in length, 
and become far less numerous in foetuses inches long. At this stage, 
also, the projection of the rectum below the abdominal profile becomes 
less marked, and in fact is scarcely apparent. 
In the adults the posterior part of the intestinal canal exhibits no 
such dilatation and dense internal garniture of villi as is seen in the 
foetus; in fact, the hind-gut of the fully-grown fishes presents nothing 
unusual when compared with that of the usual type. 
It is therefore obvious that this hypertrophy of the hind-gut and re- 
markable development of elongated villi in the embryos of the surf- 
perches has some important function to subserve during foetal life. As 
suggested at another place, that function seems to be digestive; and 
as it is not at all improbable that the fluids secreted by the walls of the 
ovarian sac are nutritive; it seems likely that the foetal fishes swallow 
such nutriment while still in the ovary. If Dr. Blake is correct in stat- 
ing that the ovarian duct is occluded by a mucous or membranous 
plug during gestation, this interpretation of the function of the gut of 
the embryo would seem all the more likely to be the correct one. 
IV.-The intra-ovarian respiratory function of the verti- 
cal fins of Embiotocoid fish embryos. 
Two papers on the intra-ovarian gestation of the embryos of the Em- 
biotocidce, Holconoti, or surf-perches of the Pacific coast have been pub- 
lished by Dr. James Blake,* in the first of which the singular vascular 
lobes extending beyond the rays of the vertical fins, as productions of 
the inter-radial membranes, are for the first time described and figured. 
I reproduce the following from Dr. Blake's second paper, as it will 
serve to indicate the extent of bis contributions to the subject under 
consideration: 
" The manner in which the young of the Embiotocoid fishes is nour- 
ished until it escapes from the ovary has not, that I am aware, been 
satisfactorily explained. In this class of fishes the young remain in the 
ovary until they are apparently as perfect as the adult fish. As during 
the process of gestation the ovary is cut off from all communication 
with the water, the external orifice being sealed up by a dense layer of 
epidermis or inspissated mucus, it is evident that the foetal fish must 
* 1. On the nourishment of the fcetus in the Embiotocoid fishes. Proc. Calif. Acad, 
of Nat. Sci., iii, 1867, pp. 314-317, 3 figs, in text. 
2. On the nourishment of the foetus in Embiotocoid fishes. Jourp. Anat. and Phys- 
iol. ii, 1868, pp. 280-282, 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
139 
receive the elements of its growth from the interior of the ovary, and 
that it must be furnished with organs of absorption and respiration 
suited to the medium in which it is developed. A large supply of nour- 
ishment must evidently be required, as the ovary contains frequently 
forty or fifty young fish, which, when fully developed, measure more 
than 2 inches in length, and wrhich together will weigh from one-twelfth 
to one-sixteenth as much as the parent fish. The ovary consists of a 
membranous bag partially divided at the upper (anterior) part, and 
terminating below in a narrow canal, which, in some species, opens ex- 
ternally in a cloaca common to it, and the urethra in others by a sepa- 
rate opening situated between the anus and urethra. The sac has 
three coats-a peritoneal, a muscular, and a mucous. The mucous 
layer is thrown into a number of longitudinal folds, and on the inner 
surface of the ovary the ova are found; and as the foetal fish grows 
these folds extend so that each foetus is in contact by each of two sur- 
faces with the mucous membrane. The organ is well supplied with 
blood-vessels, an artery entering at each horn of the ovary. They arise 
from the aorta immediately after the union of the branchial arteries, and 
at the latter end of gestation each is larger than the descending aorta. 
The posterior part of the uterus (ovary) also receives vessels from the 
abdominal aorta. In the early stages of development the foetus is 
embedded in a very tenacious secretion, which resembles para-albu- 
men in that it can be drawn out into threads of a foot or 18 inches 
long. As gestation advances the secretion of the ovary becomes more 
fluid, and its quantity is increased so that as much as three drams 
can be obtained from one ovary towards the end of gestation. It then 
forms an opalescent fluid, specific gravity, 1.025; does not coagulate on 
heating; acetic acid throws down a white precipitate, mixed with equal 
parts of distilled water, and filtered, it has no action on polarized 
light; heated in an open vessel, a pellicle forms on the surface. It 
probably contains some albumen-compound, fat, salt, phosphates, and 
iron ; but it has not been submitted to a careful analysis.* The foetus 
during the earlier stages of its development goes through changes ap- 
parently analogous to those which take place in the ova of oviparous 
fishes, and it is not until the fins are formed that any departure from 
the ordinary plan shows itself. As soon as the fins were well formed 
the dorsal, caudal, and ventral (anal) fins became edged with a delicate 
membrane formed apparently entirely of capillary blood-vessels. As 
the foetus grows this membrane is split up into processes or digitations, 
which extend a considerable distance beyond the margin of the fin, 
sometimes as much as a quarter of an inch. They retain the same 
structure during the whole period of development, and are so extremely 
* This fluid is usually expelled from the ovary by the struggles of the fish when 
takeu from the water, so that it is extremely rare to find one in our market in which 
the membrane closing the orifice of the ovary has not been ruptured. This is proba- 
bly the reason that it has not been already noticed, 140 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
delicate that I have been unable to detect them in any specimens pre- 
served in alcohol. As soon as the young fish is expelled from the ovary 
these processes rapidly diminish in size, and I have no doubt they 
entirely disappear after a few hours. Their use is evidently to absorb 
nourishment from the fluid in the ovary; they also serve to aerate the 
blood, for I have observed that on heating the fluid from the ovary, or 
on mixing it with ether, there is an abundant escape of gas. I think 
these facts serve to explain the manner in which the foetus of the Em- 
biotocoid fishes is nourished during its intra-ovarian development. 
They have been frequently observed in the three varieties of this sort 
of fish which are found in our waters." 
Within a very recent period, through the kind offices of Miss Rosa 
Smith, of San Diego, Cal., I have been enabled to obtain some well- 
preserved material for purposes of study, which, while it enables me to 
add considerably to our knowledge of the development of these singu- 
lar fishes, also impels me to differ to some extent from the conclusions 
of Dr. Blake. Miss Smith, at no little trouble to herself, obtained for 
me gravid females of Ampliistichus arenatus, Ditrema jacksoni, and Mi- 
crometres aggregates. Unfortunately this material, as well as some in 
addition which I have obtained through the kindness of Dr. Bean from 
the collections of the National Museum, represents only the more ad- 
vanced stages of foetal development, so that I am unable to add any- 
thing to the very little which is known of the earliest stages of devel- 
opment. Judging from the dates upon which the specimens collected 
by Miss Smith were taken, it would seem probable that during the 
months of October and November one would probably find the earlier 
stages which are so desirable in order to clear up what must evidently 
be a most interesting chapter in vertebrate embryology. 
The most striking characteristic of the foetal surf-perches, as they lie 
in the ovarian sac of the parent fish, is the exaggerated development 
of the vertical or median fins, all of which combined present a lateral 
area almost or quite as great as the united areas of the side of the 
head, body, and tail. These fins present an exaggerated length and 
height nowhere approached by the adults, and not exceeded in respect 
to the width and relative length of the caudal fin by any fishes except 
perhaps some of the domesticated Japanese breeds of Carassius. such 
as the Kin-yi-ko. This extension of the area of the vertical fins is still 
further increased by the production of the interradial membrane into 
thin, highly vascular processes or marginal digitations. These Dr. 
Blake has already described, but he has not mentioned the very extra- 
ordinary way in which these structures acquire their blood-supply, 
which is, so far as I am aware, unique amongst young fishes, though 
faintly approximated by the arrangement of the blood supply seen in 
the enormously developed translucent caudal of the Kin-yi-ko, as 1 have 
had the opportunity to observe in the living fish. 
Description of the vascular supply of the vertical fins of Embiotocoid 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
141 
embryos will be necessary before it is possible to comprehend the mean- 
ing and importance of the vascular digitations at the edges of those fins. 
All of them receive their blood-supply from the median aortic trunk. 
The trunks given off to the dorsal fin traverse the median plane of 
the body and tail, and pass up in a nearly straight direction to the base 
of the dorsal, at irregular intervals of one, two, or even six muscular 
segments. Upon reaching the base of the fins they subdivide into from 
two to six branches, which pass up a little to one edge of the interradial 
space, but give off smaller trunks along their entire course through the 
interradial membrane, which is highly vascular. Continuing they pro- 
ceed to the margin of the fin. They end in a flat sieve like capillary 
mesh, which fills up almost the whole of the marginal lobes formed by 
the prolonged interradial membranes. The tissue of the walls of the 
vessels within these terminal interradial lobes seems to be mesoblastic, 
but their external walls next the exterior seem to be mostly epidermic 
and exceedingly thin, probably not more than three or four cells thick. 
The capillary mesh is most complex and finest in the terminal lobes, 
less so in the interradial spaces. 
The digitations on the spinous portion of the dorsal are very feebly 
developed, in fact scarcely apparent between the tips of the nine spines 
of the spinous part of the dorsal of a fcetus of Micrometrus aggregatus, a 
species in which these spines develop, by the way, in a continuous fold, 
and not in separate pockets or diverticula as in Lophius and Gasteros- 
teus. * Four principal vessels supply the interradial and terminal capil- 
lary meshes of the soft dorsal with blood. These four vessels pass in 
an almost vertical direction from the aorta to the base of the fin, and 
do not follow the curvatures of the intermuscular septa. No vessels 
of unusual caliber pass to the spinous part of the dorsal. 
The terminal or marginal interradial lobes of the anal of the foetus of 
Micrometrus are twenty-six in number, or three in excess of the dorsal, 
while the two spaces between the anterior spines of the anal are without 
produced vascular lobes. Five principal vascular trunks pass verti- 
cally downward to the base of the anal, where they subdivide and sup- 
ply each of the interradial spaces. The marginal lobes of the anal and 
dorsal are about alike in length and all have a marginal vessel circum- 
scribing the capillary mesh at their edges. In this last respect they 
are like the terminal lobes of the caudal, in which they are, however, 
somewhat larger and longer. 
From what has been said above it will be noticed by the reader that 
the number of vascular trunks passing outward to the bases of the 
vertical fins does not correspond to the number of rays which they con- 
tain. This fact indicates that this singular vascular supply of the ver- 
tical fins of Embiotocoid embryos has attained great specialization and 
must be of very great physiological importance. 
This fact is rendered still more striking when we come to consider 
the vascular supply of the caudal fin, which is quite unique so far as I 
* See Am. Naturalist, 1885, p. 415. 142 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1886. 
am aware amongst existing fish embryos. Tn the caudal, the posterior- 
extremity of the aorta is prolonged backward, slightly bent upward, to 
beyond the anterior ends of the rays where it divides into two branches, 
a ventral and a dorsal one, each running nearly vertical to the axis of the 
body along the base of the caudal. From their posterior sides these two 
branches each give off about eleven or twelve secondary twigs which 
run out along and between the rays through the interradial spaces to 
supply the capillary mesh-work of the latter and the terminal lobes. 
The membranous fold which is continued forward dorsally and ven- 
trally from the base of the caudal, and which is a remnant of the con- 
tinuous median fin-fold of an earlier stage, is also very vascular, but 
gets its blood supply from vessels arising from the aorta in front of the 
caudal. The foregoing is the arrangement of the vascular system of 
the fins in foetuses about 1 inch in length. On comparing this stage of 
Micrometrus with Ampliistichus 1| inches long, it was observed that the 
vascular apparatus above described had already begun to diminish in 
importance, and in the adults nothing could be found which exactly 
represented it. The conclusion, therefore, is that like the vessels which 
render the skin of the foetal Embiotocoid so highly vascular, this system 
atrophies for the most part after having temporarily subserved some 
important function. 
The skin of the foetal Micrometrus was highly vascular, and it was 
observed that special trunks passed outwards to the skin along the 
sides, these vessels arising for the most part above the aorta from the 
sides of the vessels which passed to the dorsal and the muscles of the 
tail. The vascular network beneath the skin is remarkably wrell de- 
fined and could obviously not be for the purpose of supplying the skin 
alone with blood. 
The complex and unique vascular apparatus of the vertical fins of 
Embiotocoid fish embryos it seems to me to be mainly for the purpose 
of effecting respiration through the skin, and not for the purpose of 
absorbing nutriment from the ovarian space as Dr. Blake, has supposed. 
My reason for entertaining such an opinion is the fact that the mouth 
of the foetus is now open and that the throat is perforate, and that the 
back part of the intestinal canal is widened and clothed internally with 
extraordinarily long and densely crowded villi. In fact this part of the 
intestine has been hypertrophied in the foetus, and while there was no 
food found by me in the alimentary canal, I cannot but believe that 
this part of the intestine is already functional and subserves all of the 
purposes of the alimentary apparatus of the adult, and that the al- 
buminous substances secreted in the ovarian cavity are swallowed by 
the foetus, and finally digested by the peculiarly organized hind gut 
which 1 have described. There is, therefore, it seems to me, nothing left 
to do for the produced vascular membranes of the fins and the skin 
except to aid the gills in respiration, the interchange of gases between 
the parent and the foetus taking place through the fluids contained in the 
ovarian cavity, and in much the same manner as in Gambusia. The 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
143 
structure of the median fins of these embryos is such that, while their 
marginal lobes somewhat resemble the foetal villi of the mammalian 
placenta, they differ from the latter in not being received into maternal 
crypts, which I diligently but vainly sought for on the surface of the 
folds which depend from the dorsal wall of the ovary. 
The vascular net work of the ovary by which the maternal bipod is 
brought into indirect contact with the marginal lobes of the vertical 
fins of the foetuses is derived from a strongly-developed pair of vessels 
which enter the two anterior coruna of the ovarian sac. These give off 
branches to the dorsal, thickened wall of the sac and others to the 
pendent membranes which extend downwards from its roof, between 
which the advanced foetuses are packed in a somewhat irregular 
manner. 
V.-The development and intrafollicular gestation of Gam- 
BUSIA PATRUELIS. 
During a temporary residence at Cherrystone, Va., in August, 1881, 
my attention was directed by Col. M. McDonald to the existence of 
a small Cyprinodont in the fresh-water streams of the vicinity. The 
females were gravid, the ovaries of most consisting of twenty to twenty- 
five yellowish eggs about one-twelfth of an inch in diameter. This 
species, which proved to be Gambusia patruelis of Baird and Girard, 
does not, as do most other fishes, commit its ova to the care of the ele- 
ment in which it lives, but carries them about in the ovary, as do most 
of the members of this family, where they are impregnated and where 
they develop in a very remarkable manner. 
Of the manner of impregnation of the female we know very little ex- 
cept from the observations of Mr. Duly, who has related to mewbat he 
has noticed in the actions of the adults kept in aquaria in the National 
Museum. I have appended Mr. Duly's observations at the end of this 
paper. 
Tn the adult male, which measures of an inch in length, the anal 
fin is strangely modified into an intromittent organ for the conveyance 
of the milt into the ovary of the female; a tubular organ appears to be 
formed by three of the foremost anal rays, which are greatly prolonged 
and united by a membrane. At the apex these rays are somewhat 
curved toward each other, and thus form a blunt point, but the fore- 
most one of the three rays is armed for its whole length with trans- 
verse ridges, and with sharp recurved hooks at its tip, the other two at 
their tips similarly with hooks, and between their tips are two small 
fenestra or openings which possibly communicate directly with the 
sperm duct from the testes. The basal interspinous bones of the anal 
fin are bound together by fibrous tissue into a cylindrical columnar 
bony mass, truncated above, and is prolonged upwards from the in- 
ferior abdominal parietes behind the anus into the cavity of the air- 
bladder for the distance of nearly the eighth of an inch; from the upper 
truncated end and the posterior side of this column a series of fibrous 144 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
bands pass to the dorsal and posterior wall of the air-bladder, Io be in- 
serted in the median line. Whether this bony column serves to steady 
the fin in the act of copulation, or whether it serves to give passage to 
the sperm duct are unsettled questions with the writer. The modified, 
almost styliform, anal fin of the male measures a third of an inch in 
length, or nearly one-third of the total length of the fish.* Other pecu- 
liarities of the male are noticeable, for instance, the more abbreviated 
air-bladder, which also occupies a more oblique position than in the 
female. The most remarkable difference presented by the male as com- 
pared with the female, however, is his inconsiderable weight, which is 
only IGO milligrams, while that of the gravid female is 1,030 milli- 
grams, or nearly six and one-half times the weight of the male, an 
unusual difference in the relative sizes of the two sexes amongst fishes. 
The adult female, as already stated, is heavier than the male, and 
measures If inches in length. The liver lies for the most part on the 
left side. The intestine makes one complete coil or sigmoid flexure 
upon itself in the anterior half of the body-cavity ; its posterior third 
passes back along the floor of the abdominal cavity, and at the poste- 
rior end of the latter the Wolffian ducts traverse it vertically, to be en- 
larged near their outlets into a fusiform urinary bladder. 
The ovary is a simple, unpaired organ, the greater part of which lies 
on the right side of the body-cavity below the air-bladder, and serves 
to fill up the greater part of the inferior moiety of the former when de- 
veloped to maturity with its follicles gravid with embryos. The ova, 
when full grown, are each enveloped in a sack or follicle supplied with 
blood from a median vascular trunk, which divides and subdivides as it 
traverses the ovary lengthwise in a manner similar to that of the stem 
to which grapes in the bunch are attached. The arrangement of the 
ova and their attachment to the median vascular rachis is well shown 
in Fig. 11, Plate VIII, in which the immature ovarian eggs are also rep- 
resented as small whitish bodies attached to the sides of the vessels. 
In Fig. 10, Plate VIII, two of these immature ovarian ova are also shown 
close to the follicular pore of a follicle containing an advanced foetus. 
Every fully-grown ovum, by means of the preceding arrangement, has 
its own independent supply of blood from the arterial system of the 
mother, the ovarian arterial trunk being a branch of the median dorsal 
aorta. Each egg and egg-sac is thus supplied with materials for its 
growth and maturation, the latter eventually becoming specialized into 
a contrivance by which the lives of the developing embryos are main- 
tained while undergoing development in their respective follicles. The 
young or unripe eggs which are found together in the same ovary with 
the developing foetuses are, as stated above, enveloped in a cellular and 
fibrous stroma, which serves not only to strengthen the vessels, but 
*The anal fins of the males of closely allied species of Gambusia and of Limia, from 
Cuba, are similarly modified, according to Prof. Felipe Poey, who has figured those 
observed by him iu his Memorias sobra la Hist. Nat. de la Isla de Cuba, i, pp. 
382-387, pls. 31 and 32. Habana, 1851. 145 
1885.] 
FEoCEEDINGtS OF UNITED STATES NATIONAL MUSEUM. 
Vol. VIII, No. 10. Washington, D. C. May 25, 1885. 
also afterwards enters into the structure of the walls of the ovarian 
sacs or follicles externally, as these grow in size. 
The immature ova measure from a hundredth to a fiftieth of an inch, 
and enlarge to something more than a line in diameter, when they may 
be said to be mature. The median vessel of the ovary extends back- 
ward and slightly downward, and arises very far forward in the roof of 
the body-cavity from the dorsal aorta. 
The ova after developing a little way are each inclosed in a follicle 
or ovisac, membrana granulosa of Von Baer, or membrana cellulosa of 
Coste. As the egg develops there seems to be a space gradually formed 
about it in the same way as described by Wyman in Anableps. This 
space is filled with fluid, and in this liquid, which increases in quantity 
somewhat as development proceeds, the embryo Cyprinodont is con- 
stantly bathed. 
There is no trace tchatever in the egg follicles of Gambusia of an inde- 
pendent egg membrane, such as is present in the ova of all known forms 
of osseous fishes which spawn directly into the water, the egg mem- 
brane or zona radiata of the latter, in all cases being perforated by a 
minute pore or micropyle for the passage of the spermatozoa. The 
condition in Gambusia raises the question whether the zona is not more 
or less completely aborted, sooner or later, in all viviparous fishes. 
Wyman, however, points out that a membrane exists, covering the 
immature ovarian eggs of Anableps, and I find that the mature eggs of 
Fundulus are provided with a strong zona radiata during the early 
stages of their intra-ovarian development; how long it persists, how- 
ever, has not been determined, as it is possible that it is disintegrated 
and absorbed during the later stages of gestation. Rathke's description 
of the ovarian follicles of Zoarces agree in some respects pretty closely 
with my account of the egg-follicles of Gambusia given farther on. 
The narrow elongate stigma, devoid of vessels, on the follicle, spoken 
of on page 4 of his memoir on Zoarces, probably corresponds to what 
I have called the follicular pore. lie has described a vascular net-work 
in the follicle, a stalk joining it to the vascular rachis, and a space 
around the yelk much as in Gambusia. 
The zona radiata, or covering of the eggs of osseous fishes, is said to 
be secreted from the cells lining the follicles, and is composed of a strong, 
somewhat gelatinoid substance, and is usually, though not always, per- 
forated all over by a vast number of extremely fine tubules, called pore 
canals by tbeir discoverer, Johannes Mueller. In the yellow perch 
(Perea flavescens), the zona radiata is well marked, but external to it 
there is a highly elastic layer many times thicker, which is traversed 
by what seem to be tubules, which are, however, much less numerous 
than those of the true zona. In the ova of closely allied forms, such as 
the white perch and r ck-fish or striped bass, this extra external envel- 
Proc. Nat. Mus. 85 10 146 
PROCEEDINGS OR UNITED STATES NATIONAL MUSEUM. 
[1885. 
ope is wanting. In the ova of the Ganoids, Amia and Lepidosteus, the 
zona is composed, in the first instance, at least, of short, parallel, elastic 
fibers disposed in a plane vertical to that of the membrane, these fibers 
being fused at their ends or just below the inner and outer surfaces of 
the membrane. Sections through the egg membrane of Lepidosteus seem 
to indicate the same condition of things as in Amia, in fact, Dr. E. L. 
Mark, of Cambridge, Mass., has kindly shown me drawings which show 
the fibers of the zona of the former isolated in the same condition as I 
have been able to separate those forming the egg-membrane of the 
latter. 
The egg-cases of the oviparous Plagiostomes have a different struct- 
ure from the zona investing the ova of osseous fishes in that they 
are fibrous in structure and of a horny consistency, the fibers running 
in a direction parallel with the surfaces of the case. The ovoviviparous 
forms, judging from the character of the egg-cases of Squalus, have the 
walls of the case extremely thin, apparently to facilitate the interchange 
of the gases needful in foetal respiration during development in the 
oviducts. 
The skate's eggs have two hollow processes at each end of the pecu- 
liarly formed case, which are perforated at the sides and which afford 
passage for the water in and out of the case needful for the respiration 
of the developing embryo. The egg-case of Ueterodontus or Gyropleu- 
rodus, a specimen of which has been sent me by Miss Rosa Smith, may 
be compared to a skate's egg, which is narrower at one end than at the 
other, but which has been molded into a spiral form during its forma- 
tion in the oviduct, and which has traces at either end of openings 
which are apparently homologous with those found in the latter. 
The preceding data render it evident that there is a wide range of 
variation in the character of the investments which cover the ova of 
different species of fishes, and it is obvious that, while all of the various 
forms are for the purpose of giving protection to the ovum, some of them 
are modified to minister to the respiration of the embryo. In other cases 
a modification of the membrane such as exists in the adherent ova of 
Amiurus, may have still another function to subserve, viz, that of en- 
abling the male parent to aerate the eggs without injury during his ap- 
parently violent movements over them with his anal and ventral fins. 
In Amiurus a double membrane invests the egg; an outer highly 
elastic one separated from the inner or true zona by a shallow space 
filled with water, but attached to the latter by irregularly disposed 
elastic pillars of t lie same material as the outer envelope. The outer 
envelope is also very adhesive, so that the masses of ova in their elastic 
investments may be repeatedly moved about or shaken up by the old 
fish without injury or risk of detachment. 
The variability of the egg-membranes of the class Pisces, as restricted 
by Jordan, is rendered still more striking when the types with filamen- 
tous appendages, such as Menidia, Belone, Exocoetus, and Gobius, are 
considered. No trace of such a structure existing as a covering for the 1885. J 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
147 
egg of Gambusia, we are in a position to ask the question why such an 
unique condition of affairs should exist in this case. The answer, it 
would appear, we need not go far to seek. In the case of eggs which 
ordinarily hatch in the water it is necessary that they should be sup- 
plied with a covering more or less firm and capable of protecting the 
contained embryo, which, in the case of Gambusia, is not needed, be- 
cause the embryo is developed so as to be quite competent to take care 
of itself as a very well organized little fish when it leaves the body of 
its parent. Nature will not waste her powers in an effort to make use- 
less clothes for such of her children as do not need them ; on the con- 
trary, she is constantly utilizing structures economically, and often so 
as to serve nfore than one purpose at one and the same time. 
The ovarian follicles of Gambusia containing mature ova or foetuses 
are built up internally of flat or squamous polygonal cells of pave- 
ment epithelium, and externally of a net-work of multipolar, fibrous 
connective tissue cells and minute capillary blood-vessels with cellular 
walls, which radiate in all directions over the follicle. From the point 
where the main arterial vessel enters it, this vessel, together with its 
accompanying vein and investment of fibrous tissue, constitutes the 
stalk by which the follicle and its contained naked ovum is suspended 
to the main arterial trunk and vein. The minute structure of the fol- 
licular membrane is shown in Fig. 14, Plate IX; the capillary system 
of the follicles converges by way of the veinules which join the large 
median ovarian vein which follows alongside the course of the ovarian 
artery back to the heart. The very intricate mesh-work of fine vessels 
which covers the follicle supplies the developing foetus with fresh oxy- 
gen, and also serves to carry off the carbonic dioxide in much the same 
way as the placenta or after-birth performs a similar duty for the young 
mammal developing in the uterus of its parent. There is this differ- 
ence, however, between the foetal fish and the foetal mammal: In the 
former there is no uterus; the development takes place in the follicle 
in which the eggs have grown and matured; there is no placenta, but 
respiration is effected by a follicular mesh-work of blood-vessels, and 
the interchange of oxygen anol carbonic dioxide takes place through 
the intermediation at first of the fluid by which the embryo is sur- 
rounded in its follicle, anol later, when blood-vessels and gills have de- 
veloped in the embryo (see Fig. 17, Plate IX), they, too, become acces- 
sories to aid in the oxygenation of the blood of the foetus. 
In the mammal, on the other hand, there is a uterus; the egg must leave 
its ovarian follicle ; be conveyed to the uterine cavity before a perfectly 
normal development can begin; there is a fully developed, richly vascular 
placenta joined to the foetus, the villi or vascular loops of which are in- 
sinuated between those developed on the maternal surface of the uterine 
cavity. In both fish and mammal, however, the following general like- 
ness remains, viz, that there is no immediate vascular connection be- 
tween mother and embryo. In both the respiration of the embryo is 
effected by the transpiration of gases through the intermediation of 148 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
membranes and fluids, oxygen being constantly supplied and carbonic 
dioxide carried off by means of a specialized portion of the blood-sys- 
tem of the maternal organism. 
There is still another character which distinguishes the development 
of Gambusia from that of the mammal. The body of the former is built 
up by a gradual transformation or conversion of the substance of the 
yelk into the various structures which make up its organization. In 
other words, the young Gambusia obtains no nutrition from its parent; 
there is merely an incorporation byr the embryo of the stored proto- 
plasm of the yelk-sac. In the Eutherian mammal, on the other hand, 
hand, the embryo receives nourishment through the placental apparatus; 
by far the greatest proportion of the embryo being built up from the 
protoplasm supplied and conveyed to it from the blood-system of the 
parent. Judging from the large size of the young of some viviparous 
fishes, such as those of Anableps and the Embiotocidoe, it is obvious that 
there are exceptions to the method of development characteristic of 
Gambusia. 
Besides the very intricate net-work of capillary vessels which covers the 
follicles of the ovary of Gambusia, a large opening, Figs. 9, 10, 11, and 12 
mp, Plate VIII, and Fig. 13, Plate IX, of a circular, though usually 
more or less oval, form makes its appearance in the wall of each one at 
or near the point of attachment of the vascular stalk by which they are 
supported. This opening appears to increase in size as the young fish 
develops; whether it is present during the earliest stages of the intra- 
follicular development of the embryo I do not know, as I did not have 
an opportunity to see those phases. A branch from the main nutritive 
vessel frequently lies near the margin of the opening, curving around 
it. Whether this opening serves the same purpose as the micropyle of 
ova provided with a membrane would appear very probable, as it is 
difficult to see in what other manner the milt, which is probably intro- 
duced into the abdominal cavity by the male, could reach the ovum 
through the wall of its follicle. The ovary itself seems to have no ex- 
terior investment, so that the follicles lie directly within the abdominal 
cavity, the young fishes upon the completion of their development rup- 
ture them and escape into the latter, and from thence through an 
abdominal pore into the outer world. The opening into the follicle 
licle may be named the follicular pore; Through it the cavity in which 
the embryo lies is brought in direct communication with the cavity of 
the abdomen, because it is extremely doubtful if the ovary as a whole 
is invested by a distinct membrane. If it is, its tenuity is extreme, 
since the follicles when gravid are readily separable as if entirely un- 
covered, or as if there were no external peritoneal investment reflected 
over them. 
The nearly globular vitellus of the mature egg measures about a line 
in diameter. The germinal protoplasm probably occupies a peripheral 
position covering the nutritive or vitelline portion of the egg as a con- 
tinuous envelope with strands of germinal matter running from it down 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
149 
through the yelk substance. At the time of fertilization it is probable 
that the germinal matter is aggregated at one pole of the yelk. The 
latter in Gambusia is orange colored, and embedded between it and in 
the periblast superficially are a great number of refringent oil globules 
of small size. 
The body of the young fish lies in a groove or furrow on the surface 
of the yelk at about the time the tail is beginning to bud out, as shown 
in Fig. 13, Plate IX, which represents the youngest stage of Gambusia 
which I have been able to observe. The somites or segments of muscle 
plates had been developed for some time, and the heart, brain, intes- 
tine, and organs of sense were well defined. 
The next stage observed was that represented in Fig. 1, Plate VI, figured 
from the fresh embryo removed from its follicle. 
The mouth is not yet well open and the pericardiac cavity seems to 
have its anterior or inferior wall supported by the end of the snout. 
The heart lies on the floor of this cavity on the yelk, with its venous end 
directed downward and forward. The vitelline net-work of vessels arise 
behind from the caudal vein, which bends down when it reaches the 
yelk, when it divides and sends a trunk forward along either side of 
the latter to join the rudimentary Cuvierian duct which runs down upon 
the yelk in the vicinity of the liver where the portal vein anastomoses 
with the latter. 
The air bladder ab, Fig. 1, does not yet fill up the peritoneal cavity 
jpp, and the liver L is quite large and lies on the left side. The intestine 
is not yet much bent or coiled upon itself. 
The auditory vesicle and otoliths are well defined, and the eye and top 
of the head are pretty well pigmented. 
The lateral line organs, or the neuromasts of R. Ramsay Wright, are 
present at this stage over each muscular segment, as shown by the small 
circles nh of Fig. 1. But at a later stage similar organs are found at 
the anterior border of the opercles, lodged in depressions of the skin as 
shown in section much enlarged in Fig. 20. These last-mentioned ter- 
minal sensory disks or neuromasts in the skin over the cheeks appear 
to derive their nerve supply from the great fifth pair. 
The fins at this stage consist of the pectoral, dorsal, anal, and cau- 
dal; the ventral is not yet developed. Rays have commenced to form 
in the caudal; the others do not yet present any well-marked indica- 
tions of rays, which only become well defined during the later stages 
shown in Figs. 2 and 3, when the ventral first appears as a papilliform 
rudiment on either side a little in front of the vent. No sexual differ- 
entiation of the anal fin was observed in any of the embryos taken 
from the follicles, and it is therefore inferred that such differentiation 
occurs during post-fee tai life. 
The next stage observed is that represented in Fig. 2, which is drawn 
from a specimen hardened in chromic acid, and is consequently a sur- 
face view. The striking peculiarity which this specimen renders ap- 
parent is the prolongation upwards and forwards of a process of 150 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
the yelk-bag over the opercles at py. In other hardened specimens 
this pouch was observed to be prolonged upwards so as actually to 
meet its fellow of the opposite on the top of the head. In this way it 
happens that a sort of girdle is formed by the hollow cornua arising 
from the sides of the yelk-sac, which thus surround the head. 
The next stage. Fig. 3, is from a living foetus liberated from its fol- 
licle. In this specimen the scales were already developed, though the 
yelk was not yet absorbed, but was considerably diminished in amount. 
The heart still had its venous end directed downwards, gathering the 
blood from the net-work of superficial vitelline capillaries through a 
great median ventral vein, the same as in Fig. 1. In a cross-section, 
Fig. 15, from a specimen a little more advanced than that represented 
in Fig. 3, the dermal pouches in which the scales are developed are 
shown at sp • the section has also cut through the more advanced ven- 
tral at vt and the anal at a. 
- A remarkable characteristic of the embryos of Gambusia is to be noted 
in the fact that they do not develop continuous fin-folds, but, as shown in 
Figs. 1, 2, and 3, the vertical fins at once appear as distinct folds. This 
type is therefore not perfectly lophocercal or provided with a continuous 
eradiate dorso-ventral fin-fold such as is found in the embryos of Gadus 
of the same relative stage of development. 
Chondrocranium.-The cartilaginous skull of Gambusia at about the 
stage represented in Fig. 3 is shown in detail in Fig. 21, Plate X. The 
brain-box has an incomplete roof; there is a cartilaginous bridge over 
the pineal region T cr, and one over the medulla oblangata oc. On its floor 
there is wide pituitary interval, as seen in Figs. 21 and 28, into which 
the hypophysis Uy, the infundibulum In, and optic nerve II, and crus 
ch project downward more or less decidedly. The chorda projects a con- 
siderable distance under the hind-brain, as may be inferred from Fig. 
16, representing a longitudinal section of its anterior end, the vertebrae 
nr, vr being already differentiated with their intervertebral ligaments 
Z, I, defined. 
The auditory vesicle Aw is covered externally by cartilage, and the 
orbit is limited above by an imperfect supraorbital curved cartilagi- 
nous bar, below by the pterygopalatine and hyomandibular bars. The 
nasal fossae are limited behind by the cartilaginous orbito nasal septum, 
below by the rostrum. The hyomandibular Hm is remarkably long 
and slender when embraced together with the symplectic Sy. Meckel's 
cartilage, Mk, is short; the glosso-hyal G. Hy is long; the hypo-hyal, 
Hy, nearly globular; the cerato-hyal, C. hy, wide and somewhat pro- 
longed and joined to the hyomandibular by a short interhyal, 1H. The 
branchiosteges Brs originate in perichondria! membrane, and have no 
cartilaginous rudiments. There are five branchial arches, I, II, III, IV, 
V, and detached suprabranchial nodules, & br. 
Behind the branchial arches, but more superficially, lies the cartilag- 
inous coraco-scapular rudiment Cor, Sc, with the cartilaginous plate 
from which the actiuosts are formed, abutting against it posteriorly. 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
151 
The posterior branchial arches support clusters of teeth, which, in 
the sections studied by the writer, presented the forms represented in 
Figs. 26 and 27. A saccular organ, df, was developed from the deeper 
layer of the epidermis, in which the conical crown of the tooth, d, is 
first formed, as shown in Fig. 26, but with the progress of development 
the perichondrium pch, Fig. 27, which invests the cartilaginous bran- 
chial bar J5r, forms a basal cushion, s, s, s, for each tooth, the conical 
crowns d, d, d being evidently articulated to the summits of these 
cushions, which give rise to the so-called cementum plates. It would 
thus appear that the crowns of the pharyngeal teeth, like the dermal 
denticles of certain forms, arose separately from the basal parts, which it 
is very clear develop later than the crowns in this instance, in intimate 
relation to the perichondrium investing the branchial bar, and there- 
fore have a deeper origin than the crowns. The saccular organ, df, is 
developed from a more or less nearly solid diverticulum of the deep 
layer of the epidermis, which, in some cases, finally loses its connection 
with the latter, and becomes hollowed out inferiorly into a cap-like organ 
in which the germ of the crown appears. 
The branchial filaments in the advanced embryo shown in Fig. 3 
already have cartilaginous axes, c, as shown in Fig. 17 ; the base of the 
axis of the filament being furcate, and resting astride of the branchial 
vein vb, and not quite in contact with the branchial arch bh itself. 
These filaments are already pinnate and vascular. 
The vertebral column.-The anterior part of this organ has already 
been described. The vertebrae are best differentiated anteriorly; in 
the tail the membrane in which the ossification of the centra occurs is 
apparent but not so strongly marked as anteriorly. The chorda bends 
upward but slightly at its hinder end, so that the urostyle is finally very 
short, as shown in Fig. 18. 
The ribs are developed in cartilage as a single line of longitudinally 
compressed cells, as shown in Figs. 19 and 23 at rb. In the anterior 
thoracic region the heads of the ribs are elevated above the level of the 
chorda and are embedded in the fibro cartilaginous rudiments sic of the 
neural arches, as shown in Fig. 23; posteriorly the ribs have their 
heads lowered more and more and assume a more ventral and intimate 
relation to the skeletogenous sheath of the chorda. 
The fin-rays of the caudal fin in the most advanced foetuses are al- 
ready pretty well developed in membrane, beneath the epidermis; their 
relations to the mesoblast me internally and externally to the malpighian 
layer si and epilthelium ep being shown in the cross-section repre- 
sented in part in Fig. 24, and in Fig. 9 at t. 
Visceral anatomy.-The way in which the foetus is coiled up in the 
follicle is shown in Fig. 4, in which the process of the yelk-bag py, which 
extends upward over the opercles and to the top of the head, is also 
shown. In Fig. 9 a section through a follicle with the embryo in place 
and ent through at three different points is shown. This section also 
shows the open lolljcular pore mp with a thickened margin and the fol- 152 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
licular artery fv at the outer border of the thickened edge of the mem- 
brane. 
The yelk y, liver Z, intestinei,i, i, the air-bladder ab, the Wolffian ducts 
Wd, Wd, the aorta ao, chorda ch, muscles of the pectoral pf, &c., are also 
shown in Fig. 9. But in order to comprehend the visceral anatomy of 
the fcetal stages of Gambusia, Figs. 5, 6, 7, 23, and 25 are more sat- 
isfactory. 
These sections will render it obvious once for all that in this form at 
least there is no connection whatever of the yelk with the intestine. 
That there is a wide space, sc, between the intestine posteriorly and the 
vitellus y is plainly shown in Figs. 5, 6, and 7. 
The yelk also presents a well-marked periblastic layer, yh, in which 
the oil globules are embedded superficially. This periblastic stratum 
is thickest on the outer surface of the yelk and in the vicinity of ves- 
sels, but quite thin where it forms the floor of the abdominal cavity. 
In the cross-section, Fig. 5, the periblast sends up a median septum 
which divides the cruder yelk or deutoplasm into two moieties. Super- 
ficially the periblast has the vitelline vessels impressed into its surface, 
and in consequence of the fact that I have been unable to find a distinct 
cellular wall on their inner or periblastic sides I infer that the yelk is 
taken up from the periblast directly by the vessels. The cleavage 
cavity is persistent as in other Teleosts, and is apparent on either side 
of the vitellus in Fig. 9. 
The heart, the details of which are shown in transverse section in 
Fig. 8, extends down into the anterior end of the yelk-sac, as shown in 
Figs. 6 and 7, having been first extended into the cleavage cavity, as 
indicated in Fig. 13. The most inferior chambers of the heart receive 
the venous blood, and all are thin walled except the ventricle ve and 
the bulbus aortae ba, Fig. 6. The heart, as well as the liver and intes- 
tine, is more or less impressed into the surface of the yelk. 
The oesophagus is somewhat bent upward posteriorly, from which ex- 
tremity the air-bladder arises somewhat to one side. The pneumatic 
duct is open in the foetus and is short, leading abruptly into the air- 
bladder ab, which is a depressed sac, the floor of which is very much 
thicker than the roof. The liver is a very massive organ relatively, and 
opens into the alimentary canal in the vicinity of its fii>t turn. A 
large portal vessel passes from the liver outwards to the left, as shown 
in Fig. 5, and communicates with the vitelline system of vessels. What 
seems to represent a pancreas is seen in some sections adherent to the 
intestine, as shown in Fig. 6. In the posterior and upper part of the ab- 
dominal cavity, behind the intestinal coil, are placed the reproductive 
organs o, as indicated in Figs. 6 and 7; with greater magnification of 
that part of a cross-section the reproductive organ or genital folds have 
the appearance indicated at o in Fig. 25, which shows the organ in its 
indifferent stage, when it is quite impossible to tell whether the large 
nucleated cells embedded in it are going to give rise to ova or to sperma- 
tozoa. They lie on either side of the mesenteric suspensor of the intes- 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
153 
tine; are, in fact, thickened ridge-like swellings of the sides of the mes- 
entery. 
The renal system lies almost wholly exterior to the abdominal peri- 
toneum of the embryo, and is the only organ, with the partial exception 
of the air-bladder, which does so. The renal or Wolffian organ of Gam- 
busia consists of a remarkably well-developed anterior mesonephric por- 
tion pn, Fig. 7, which is crowded up against the auditory vesicle poste- 
riorly and richly supplied with blood-vessels. The pronephros was not 
observed, as the stages studied by me were already too far advanced, 
but the posterior part of the segmental duct was still essentially pro- 
nephric in character, as no segmental tubules opened into it. At the 
hinder aspect of the urinary vesicle al the segmental duct of the one 
side joined its fellow of the opposite side, and shortly after opens into 
the bladder, which has a distinct outlet, which lies in a line continuous 
with that of the segmental tubes when the latter are viewed from the 
side. In Fig. 23, Pn, a cross-section of the anterior mesonephric organ 
of Gambusia is shown, and in Fig. 25 the segmental or Wolffian ducts 
Wd. are shown in cross section through the region of the posterior part 
of the abdomen. 
The brain.-A vertical longitudinal section of the brain of a fetus of 
Gambusia, which has attained about the development of the one shown 
in Fig. 3, is shown in Fig. 28. The cerebellum cer is unusually wide 
antero-posteriorly, resembling in this respect the cerebellum of Amiurus 
of about the same age. The oral epithelium is shown in place at oep, and 
at Hi there is an involution which seems to represent the last vestige of 
the hypophysial pouch from which the hypophysis Hy has been de- 
veloped, and from which that structure has become disconnected. 
The muscular system presents nothing essentially different from that 
1o be met with in other fish embryos. The so-called lateral dermal mus- 
cle Dm, Fig. 23, is distinguishable as a distinct layer of fibers external 
to the myotomes m, m, m, but it is remarkable that this muscle still 
shows the segmented condition seen in the latter, a feature which it does 
not share in common with the dermal muscles of higher animals. It is 
therefore seriously to be doubted if it has any very strong claim to be 
called a dermal muscle. 
The courses of the fibers of the muscles of the body are shown in 
Fig. 7, which represents a longitudinal section considerably off of the 
median line. 
The muscles of the pectoral develop as usual on either side of the 
coraco-scapular plate cs, Fig. 7, and are probably derived as in the 
Elasmobranchs from buds given off' by the muscular segments above 
the rudiment of the girdle. It will be noticed, however, that there are 
no muscles developed around or below the yelk-sac, so that there are 
no recti-abdominales muscles yet formed, the only investment of the 
yelk being the outer epiblastic covering and the deeper mesoblastic one 
developed in connection with the vascular net-work which ultimately 
unites with the heart anteriorly. 154 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
Conclusion.-The foregoing account greatly expands and finally com- 
pletes the history of the development, as far as my materials permit, of 
this remarkable form, preliminary notices of which I published in For- 
est and Stream in August, 1881, and in the American Naturalist* for 
February, 1882. While this history is still far from complete, especially 
as respects the very early stages, we have at least cleared the ground 
and prepared the way for further investigations. A number of points 
which had been left undecided have been determined. These are the 
following: 
1. That fertilization of the egg of Gambusia occurs within the ovarian 
follicle, the spermatic fluid being apparently introduced into the ovary 
or abdominal cavity by the male, which is provided with an intromittent 
organ consisting of the anal fin much modified, and that the spermatozoa 
find access to the egg through a wide opening in the follicle which 
answers to a micropyle, but which may be called the follicular pore. 
2. That there is no evidence, as in the case of Anableps and the 
that the ovarian follicles are ruptured until the develop- 
ment of the young embryos is approximately completed, since the most 
advanced foetuses of Gambusia studied by me have the scales, fins, fin- 
rays, and cranium remarkably well developed, even before the yelk is 
all absorbed, as it is known that the young are born not a great while 
after the stage of development represented in Fig. 3 has been reached. 
3. It is also known that little or in nutriment is derived from the 
parent, as in Anableps and the Embiotocidw ; or, in other words, that the 
embryo of Gambusia grows entirely at the expense of the material con- 
tained in the yelk-sac, and does not form villi upon the latter nor en- 
large after the yelk has been absorbed as in Anableps; neither does the 
rectum or hind gut hypertrophy, nor do the fins expand and develop 
prolongations of the interradial membranes as in the Embiotocidce. 
4. As is the case with all viviparous forms, the number of embryos 
produced seems to be diminished in correlation with the protection 
which the young receive in consequence of their peculiarly complete 
development within the body of the parent. The embryos leave the 
parent as active little fishes about half an inch in length; in fact this 
fish begins an independent career as far developed as when the shad, 
cod, Spanish mackerel, catfish, and many other fishes are from three to 
six weeks old. By so much it has the advantage over those species in 
the struggle for existence in that it is ready to feed, to pursue its prey7 
discriminately, as soon as it is born, while the other forms alluded to 
are comparatively helpless until some time after they are born; hence 
nature makes up in the fertility of such species for the advantages 
enjoyed by the viviparous ones, so that finally the chances of the 
survival of the two types, which differ so widely in their breeding 
habits, are about equal. 
* Structure and Ovarian Incubation of Gambusia patruelis, a top-minnow. Am, 
Nat., 1882, pp. 109-118, 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
155 
VI.-Habits of Gambusia during the breeding season. 
Mr. A. A. Duly has informed me that he has witnessed the act of copu- 
lation and the birth of the young of Gambusia. In coitus the male's head 
is turned in the direction of the tail of the female, the prolonged anal fin 
seeming to be thrust into the external opening of the ovarian duct or 
genital pore of the female, which lies just in advance of the anal tin. 
The young when born are stated by Mr. Duly to be about three- 
eighths of an inch in length, and to be expelled in a single mass, con- 
sisting of eight to eleven young fishes at a single effort. This mass as 
soon as it escapes is seen to be composed of the infant Gambusias, which 
at once separate and swim away. No membranes seemed to be expelled 
together with the mass of young, so that it is probable that in this species 
as in Anableps and the Embiotocidce the foetuses rupture the follicles in 
which they were developed a short time before birth. I say a short time 
before birth, because our observations indicate that, unlike Anableps and 
Micrometrus, the development of Gambusia is essentially completed 
within the follicles, and no yelk-sac remains outwardly visible when 
the young are set free. 
My informant also tells me that the parent fishes devoured their 
young as soon as they were born if they were not separated, by trans- 
ferring one or the other at once to another aquarium. Fright seemed 
to hasten or precipitate the parturition, which Mr. Duly tells me actually 
took place under such circumstances. He also noticed that more than 
one brood seemed to be produced by the same parent consecutively 
and during the same season, and he has reason to think that more may 
have been produced, as his observations only extended over the latter 
part of summer with adults brought from Cherrystone, in August and 
September, which he kept in aquaria in the National Museum. 
VII.-The viviparity of Fundulus. 
On Plate XI, Figs. 29 and 30, I figure two views of another type of 
Cyprinodont embryos, viz, Fundulus majalis, also obtained at Cherry- 
stone, Va., July 18, 1881. In this case a well-marked zona wras de- 
veloped investing the developing embryo. The oil drops were super- 
ficially embedded in the yelk the same as in Gambusia. A great net- 
work of vessels surrounded the yelk, as may be seen in both figures; 
these spread over the yelk from the two Cuvierian veins which pass 
out from the side of the body of the embryo just in front of the pectoral 
fin folds, and are gathered inferiorly, as represented in Fig. 30, into a 
great median vein which is extended from the venous end of the heart. 
A hood is probably formed at the sides of the head of this species the 
same as in as indicated by the vascular membrane at the 
sides of the head raised from contact with the yelk, as shown in Fig. 30. 
The body cavity under the axis of the embryo is a deep oval sinus 
extending down a little distance into the substance of the yelk, as shown 
in Fig 29, behind the rudiments of the pectoral fins. 156 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
ON CERTAIN FEATURES OF THE DEVELOPMENT OF THE SALMON. 
By .10115 A. KYDEK. 
So much has been written upon the anatomy and development of this 
fish by' eminent authorities that I approach the subject with a cer- 
tain hesitancy. The development of the skull has been elaborately 
worked out by W. Kitchen Parker. The skeleton of the adult has been 
figured in great detail by Bruch in a magnificent monograph, while 
the general development has been repeatedly discussed by investi- 
gators during the last century with more or less thoroughness. Not- 
withstanding this, it may be truly said that our knowledge of the exact 
details of some features of its development is still imperfect, even 
though such able embryologists as (Ellacher, Balfour, His, Hoffmann, 
and Ziegler have devoted considerable attention to it and its allies 
within a period extending over scarcely more than the past decade. 
The early stages of development have been investigated by (Ellacher, 
His, and Ziegler, with such opportunities that can only be enjoyed by 
one who is near a locality where the spawning or oviposition of the 
adults is in progress. I can therefore add nothing to the information 
given us by those writers, but all that will concern us at present is the 
arrangement of the blood-vascular system at the time of hatching, 
some of the impairments which this system suffers when the young 
fishes are under the care of the fish-culturist, and the development of 
the fins. 
The material used in this investigation consisted of recently-hatched 
embryos of the land locked salmon, Salmo salar, var. sebago. I have 
carefully drawn a live specimen several times enlarged by the help of 
the camera lucida, as represented in Fig. 1, in order especially to show 
the arrangement of the vessels on the vitellus, the distribution of the 
rose-colored oil drops in the latter, and the vessels and venous sinus 
in the tail. 
The mode of development and outgrowth of the fins is especially in- 
teresting, the more so since Professor Cope has recently reached the 
conclusion that many of the so-called "Ganoids" of the Palaeozoic 
rocks seem really to be affiliated to a great and important order of ex- 
isting fishes embraced by that author under the term Isospondyli, which 
includes the existing Salmonoids, Clupeoids, Hyodonts, Albulids, &c. 
As it therefore seems that the salmon belongs to a very ancient series 
of forms dating back phyletically to the Devonian, it may be well for 
us to examine into the development of the fins to see if that process 
would really give countenance to Professor Cope's views. 
I. The vertical jins.-This set of fins is developed from a median fold 
extending from a vertical slightly behind the pectorals back over the 
end of the tail, thence forward on the ventral side to the posterior side 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
157 
of the yelk-sac, the ventral part of the fold being interrupted only by 
the posterior part of the gut at v. Fig. 1. This fold at the time of 
hatching contains simple embryonic rays throughout its entire extent, 
these being indicated in Fig. 1 by the fine linear striation apparent for 
the whole length of the fold in the engraving. These embryonic rays, 
as they have been called by A. Agassiz, I will call actinotrichia, from 
their slender, unsegmented, hair-like appearance; they are developed 
from special cells of the mesoblast, which I have named pterygoblasts 
elsewhere. These actinotrichia exceed in number the permanent rays 
of the adult at least ten-fold. The latter are in fact in great part 
formed by the fusion of a number of these actinotrichia lying side by 
side. 
The extent to which actinotrichia are developed in the median fin-fold 
at the time of hatching, however, varies very greatly in different genera. 
In the recently hatched salmon, which passes through a prolonged pe- 
riod of incubation, these rudiments of the future permanent rays are 
very numerous, far more so than is the case with most other forms at 
the time the embryo escapes from the egg. Tn the Spanish mackerel 
and in Gadus the actinotrichia are not well defined until sometime after 
hatching, and then only in the posterior end of the median tail-fold and in 
the pectoral fin-fold. As a rule, actinotricbia, (horny-fibers of Balfour), 
appear first in the anterior paired fins ; this is the case in the salmon, 
in which these primitive rays are well marked in the pectoral about the 
time the ventral fin-folds vt become well defined. 
As urged in my paper " On the origin of heterocercy," the presence 
of actinotricbia throughout the whole extent of the fin-folds of the em- 
bryos of the salmon is an illustration of the Haeckelian principle, viz, 
that the ontogeny of a form is usually an epitome of the phylogeny of 
the same. The persistence of the pneumatic duct and the presence of 
adipose fins are also to be considered in this connection, since both are 
archaic characters, the first especially. 
The permanent dorsal rays of the salmon are formed from the actino- 
trichia developed in the anterior part of the median fold d,fig. 1, to- 
wards which radial muscles are shoved out at an early stage, as shown 
by the evenly stippled intervals at the base of this part of the fold. 
The fold in the interval between the dorsal and soft dorsal atrophies 
with the further growth of the young fish, the actinotrichia of this 
interval also disappearing with the fold. 
The next portion sd of the median vertical fold which is perceptibly 
widened, gives rise to the soft dorsal or " adipose fin." The aotino- 
trichia of this fin never pass beyond their embryonic condition, so that 
it is said by the comparative anatomists to contain horny fibers. The 
whole of the Plectospondyli except the Characinidce'have lost their adi- 
pose fins, and thus have but one dorsal remaining. The herrings and 
salmonoids or Isospondyli would therefore seem to stand in an ancestral 
relation to the carps, suckers, and minnows, or Plectospondyli. It is at 158 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
any rate obvious that the protopterygian condition of the adipose fin is 
a less specialized and more archaic one than that of its complete atrophy. 
The interval between the soft dorsal and the caudal is in part atro- 
phied. The posterior part of this interval, however, gives rise to the 
short accessory rays of the dorsal edge of the .caudal. 
The truly dorsal part of the median fin-fold of the embryo salmon 
ends at the notch n, towards which the urochord, later the urostyle, is 
directed. At this point the proximal ends of the dorsal and ventral 
actinotrichia embedded in the fold converge in a penniform manner. 
The actinotrichia below the notch n and extending out into the caudal 
lobe cd give rise to the prolonged caudal rays, while those in the ventral 
part of the fold just in front of the caudal lobe give rise to the inferior 
accessory caudal rays. 
The fold in the interval between the caudal and anal a atrophies, to- 
gether with its contained actinotrichia, while those in the widened anal 
part of the fold a as far as the vent v give rise to the permanent rays 
of the anal, into which the muscles of the fin grow at an early stage or 
at about the same time that those of the dorsal are developed. 
The preanal part of the foldjpa in front of the vent, together with its 
contained actinotrichia, atrophies entirely during further growth and 
development. 
It is thus shown by the development of the salmon that the most 
primitive type of distribution of the vertical fins of osseous fishes was a 
continuous one, because of the development of a continuous series of 
actinotrichia or primitive rays, and that the forms which now exist and 
most nearly realize this distribution of the rays of the vertical fins are 
the Dipnoans, in which the rays are also scarcely more than well-devel- 
oped actinotrichia, several of which, taken in succession, are homolo- 
gous with a single ray of an adult Teleost. 
The Dipnoans present many other embryonic characters which seem 
to be partially paralleled by what is transitory in the embryo salmon. 
When we shall know more of the embryology of Ceratodus through the 
efforts of Mr. Caldwell, who has succeeded in obtaining its ova, further 
comparisons may be instituted between the ordinary Teleostean embryo 
and that of the singularly specialized mud fishes.* 
Some of the oldest "Ganoids" had the vertebral axis persistently 
chordal, unmodified anteriorly, and with an extended series of fins, in 
these respects paralleling somewhat the condition which is transient in 
the embryo salmon. These seem to me to be good reasons for accept- 
ing Professor Cope's views as to the affiliation of certain of the Palaeo- 
zoic fishes with the existing Isospondyli, in the absence, as the latter 
* It is important to examine the fin-folds of the tails of the larvae of such Amphibians 
as Pseudis and Dactylethra to determine whether or not actinotrichia are present. 
Should these earliest representatives of fin-rays be found in any of the Amphibia the 
abrupt hiatus now existing between the Dipnoans and the latter would, in a great 
measure, be bridged. 1885.] 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
159 
authority states, of all other characters, except the scales, adequate to 
separate the two. 
II. Paired jins.-The paired fins of the salmon develop at wide inter- 
vals from each other, that is, the ventral pair is separated from the pec- 
toral by an interval of not less than sixteen muscular segments. In 
Lophius the number of muscular segments opposite the interval be- 
tween the pectorals and ventrals are reduced to four, according to the 
figures given by A. Agassiz. In other types this interval between the 
rudiments of the paired fins of the embryo is still further reduced, 
as for example in Trachinus vipera* and in Motella mustelal in which, 
according to Brook, but two muscular segments intervene between 
the earliest rudiments of the pectoral and pelvic fins, so that in 
these types it may be said that the paired fins develope from almost 
continuous rudiments. These and Lophius seem to develop the rudi- 
ments of the paired fins almost synchronously, which is far from being 
the case with the larvae of the Physostomous orders, Ginglymodi, Glan- 
iostomi, Nematognathi, Plectospondyli, Isospondyli, Haplomi, and Enchel- 
ycephali,]. in which the ventral pair of fins appears late, often after the 
pectoral is well developed and in active functional use. In most, if not 
in all, the larvae of Physostomous species of Actinopteri, as limited by 
Cope, the ventral pair of fins is later in appearing than the pectoral; 
on the other hand in the Percomorph and Pediculate divisions of the 
Physoclistous Actinopteri it seems that in many species the pectoral 
and ventral limbs appear almost synchronously, the pectoral usually a 
little more developed than the ventral pair, and separated serially by 
only two to four myotomes, while this interval in the Physostomous 
forms may embrace over sixteen muscular segments. 
Finally it may be said that the ventral pair of limbs is almost always 
undergo shifting or translocation forwards in the Physoclistous groups 
mentioned. The researches of Mr. Brook, cited above, afford additional 
evidence of the truth of the principle laid down by me in a prelimi- 
nary notice § recently published. 
The Physoclistous genera, Gadus, Cybium (Scomberomorus') and Pare- 
phippus seem to be exceptions to the rule spoken of above, as holding 
in the development of some forms. How far the rule held in the devel- 
opment of the Palaeozoic fishes with ganoid scales and ventral or ab- 
dominal pelvic fins, we, of course, now have no means of knowing. 
These, with the exception of Dorypterus and Blochius, seem to have had 
* Preliminary account of the development of the Lesser Weever-fish Trachinus 
vipera, by George Brook, F. L. S., Journ. Linn. Soc., xviii, pp. '274-291, pls. 4, 1884. 
t On some points in the development of Motella mustela, Linn., by George Brook, 
F. L. S., Journ. Linn. Soc., xviii, pp. 299-307, pls. 3, 1885. 
tThe eels are of course permanently without ventrals; the youngest obtainable 
larvae of Anguillidae and the Leptocephalid stages of marine eels show no traces of 
ventral limb-folds; these forms are in fact permanently apodal. 
§ On the Translocation forwards of the Rudiments of the Pelvic Fins in the Em- 
bryos of Physoclist Fishes, Am. Naturalist, 1885, pp. 315-317. 160 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
the pelvic fin not displaced forward, but normal in position, as in the 
majority of existing Physostomous Actinopteri. 
The belated development of the pelvic fins of some of the Isospondyli 
and Ilaplomi is quite remarkable; in fact, the pectoral in these forms 
may have the permanent rays pretty well developed before the ventral 
fin-fold has done much more than begun to develop. The ventral fin- 
fold, however, in those forms in which it appears synchronously with the 
pectoral is at first always the smallest, thus showing the effects of an 
inherited tendency to retard the development of the pelvic fins. 
In the salmon the pectoral is functionally developed with well de- 
fined actinotrichia lying beneath the investing epidermis, as shown in 
Fig. 1, by the time the ventral is appearing as a flat, immobile lobe 
without developed actinotrichia, behind and above the yelk-bag. 
III. The embryonic blood vascular system.-In the embryo salmon this 
attains great importance apparently in consequence of the presence of 
a voluminous yelk which is absorbed by a system of vitelline vessels. A 
large median aortic trunk ao, Fig. 3, is developed under the notochord 
c/z, Fig. 4. It is formed from two convergent suprabranchial arteries an- 
teriorly, which receive their blood from the branchial trunks coming 
from the gills. The aorta extends backward, giving off "intercostal* " 
and intersegmental branches s along either side and terminates under 
and at the end of the urochord. A recurrent vessel then bends down 
from it and divides into several loops which converge in the caudal 
venous sinus sc, Fig. 3, incorrectly referred to as a " caudal heart" by 
some writers, as it exhibits no independent pulsations of its own. 
From the caudal sinus the caudal vein or cava arises and passes for- 
ward towards the body-cavity, where it divides anteriorly into the 
paired cardinal veins cv, Fig. 3. These pass forward to the venous sinus 
of the heart formed by the Cuvierian ducts into which they empty their 
contents. The heart lies behind and below the branchial frame-work, 
and forces only venous blood through the gills. Its anterior chamber or 
bulbus aorhe is prolonged forward into a truncusarteriosus, which gives 
off a pair of vessels to each pair of gills. These vessels, after break- 
ing up into a plexus of capillaries, send the arterialized blood through 
another set of branchial vessels which join a pair of longitudinal trunks, 
of which the radices aortae are prolongations posteriorly and the caro- 
tids anteriorly. The venous blood from the head is carried back to the 
heart by way of a pair of jugular veins or anterior cardinals to the 
venous sinus. The foregoing describes in outline the systemic circu- 
lation proper, exclusive of the great portal system, to which the net- 
work covering the yelk-bag also properly belongs. 
The portal system of the young salmon with the yelk-bag still at- 
tached may be said to comprise no less than three successive sets of 
capillaries. The first of these arise from the caudal and cardinal veins 
er, and pass from above downward on either side of the intestine to 
join and fill with blood a large azygous or median subintestinal vein, 1885.] 
proceedings of United states National museum. 
161 
Vol. VIII, Vo. 11. Washington, D. C. June 1, 1885. 
si, Fig. 3. This vessel passes forward under and into the liver L, where 
it again breaks up into a plexus of smaller vessels; these then again 
blend into larger trunks, as shown at pv, in Fig. 4, which emerge from 
the liver to again break up into the capillary net-work cc, Fig. 3, which 
is still better shown in Fig. 1, on the surface of the yelk-bag. This 
vitelline net-work then joins the vitelline veins vv and vv', which blend 
into a common trunk before joining the venous end of the heart. It 
will thus be seen that we have no less than five capillary systems in 
the young salmon, as shown in Fig. 1, if we reckon those of the portal 
system together with those belonging to the systemic system of vessels. 
These in their order are: (1) The branchial, (2) the systemic, (3) the in- 
testinal, (4) the hepatic, and (5) the vitelline capillaries. 
What the subsequent history of the third and fourth sets may be I 
have not made out, but the fifth or vitelline set has only a temporary 
existence, remaining only as long as there is yelk in the yelk-bag. As 
the yelk is absorbed this system disappears, when the vitelline veins vv 
and vv' become portal veins; that is, they carry all of the blood which 
passes through the viscera back to the heart. 
A study of sections of the yelk:sac of the salmon leads to the follow- 
ing conclusions: A well-marked periblastic stratum of plasma, p, Fig. 
4, invests the yelk. Beneath the periblast lie the oil-drops o, o, which 
are largest at the upper part of the yelk, the greater buoyancy of these 
larger, superior oil globules tends to keep the young fish buoyed up, 
and functions much in the same way that an air bladder would, a struct- 
ure which, by the way, is not yet functional in the young salmon at 
this stage of growth. In the earlier stages these larger oil-drops, which 
lie just under the blastodisk or germinal mass, by their buoyancy con- 
stantly keep the germ rotated or turned toward the top of the egg. 
External to the periblastic layer of the yelk comes the vascular net- 
work of capillaries, the walls of which are formed, apparently, by a 
thin sheet of splanchnic mesoblast vni, which invests the yelk but 
which has grown down over the latter at a later period, possibly, than 
the thin epidermic or epiblastic investment ep, Fig. 4. This vascular 
net-work is obviously the apparatus by means of which the yelk is ab- 
sorbed superficially from the external plasmodium or periblast of the 
yelk, a stratum, which, as is well known, contains scattered free nuclei. 
The epidermis or epiblast of the young salmon is remarkable amongst 
fish embryos for the peculiar goblet-shaped cells which are found dis- 
tributed over almost the entire surface of the embryo. These are shown 
in a section of the epidermis in Fig. 5, much enlarged. Their function 
is apparently to secrete a mucilaginous substance for the purpose of 
protecting the skin of the embryo. 
IV. Diseases or abnormalities which involve the yelk-sac and vitelline 
Proc. Nat. Mus. 85 11 162 
PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
circulation of salmon embryos.-A discussion of some of the pathological 
phenomena observed while salmon embryos are under the care of the 
fish-culturist seem to me to be in place here. 
Shortly after hatching many of the young fish in the hatching troughs 
sometimes show whitish spots on the yelk-sac. If examined under 
the microscope it is plain that these spots consist of coagulated or dead 
yelk material. Very often the capillaries in the vicinity of these spots 
are found to be occluded and filled with clots of blood, as if the vessels 
had been bruised. In other cases it is found that the capillaries of the 
liver are occluded so that that organ, which is visible on the left sile of 
the embryo through the integument, assumes a whitish, abnormal color. 
Closer examination reveals the fact that the blood no longer circulates 
through the liver, and that the tissues of the organ are practically dead, 
as indicated by the white color which they assume. These conditions lead 
to the death of the affected embryos in great numbers. The causes which 
seem productive of such abnormalities have not been determined with 
certainty, but it would seem probable that blows or knocks received by 
the sac from careless handling or the violent and too rapid flow of 
water over the young fish, so as to carry them violently against fixed 
objects in the trough, are probably very hurtful and productive of the 
changes noted. 
Still other abnormal changes in the yelk-sac may be noticed here. 
The most serious is that characterized by the distension of the epiblastic 
covering of the sac with fluid so that it is lifted up from contact with 
the yelk more or less extensively. Usually, this distension only affects 
the posterior extremity of the sac, but occasionally specimens are ob- 
served in which there is a space all round the yelk between the latter 
and the epiblastic sac. Sometimes free-blood corpuscles which have 
escaped from ruptured vessels are found floating about in the fluid con- 
tained in the cavity described. At other times, an extension of the back 
part of the yelk proper may be prolonged backward into the outer sac, 
which may become constricted so as to embrace part of the yelk. As 
the anterior part of the yelk is then absorbed, the posterior constricted 
part is finally left hanging to the abdomen by a sort of pedicle formed 
by the outer sack. This finally drops off and the young fish survives. 
The spot where the stalk breaks off on the under side of the embryo 
heals up and the young fish seems none the worse for having lost part 
of its yelk, except that it has probably not grown quite so large or so 
rapidly as its more fortunate fellows. 
The plate illustrating the foregoing article is number XII of the pres- 
ent volume. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
EXPLANATION OF PLATE Vi. 
Fig. 1.-Side view of foetal Gambusia recently removed from the ovarian follicle, x 22. 
Fig. 2.-Side view of a somewhat older embryo taken from the follicle and hardened in chromic acid, 
x 11. 
Fig. 3.-Side view of a still older embryo, with the yelk-sac still apparent, taken from the follicle and 
figured while alive, x 22. Proceedings Nat. Mus., Vol. VIII, 1885. Ryder on Viviparous Fishes. 
PLATE VI. 
DEVELOPMENT OF GAMBUSIA PATRUELIS. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
EXPLANATION OF PLATE VH. 
References a, anal fin; ab, air-bladder; al, urinary vesicle; aw, auditory vesicle; ba, bulbus aorta; 
c, pericardiac cavity; cs, coraco-scapular cartilage,-/, pecten; i, intestine; L,liver: m, 
muscular somites; mb, mid-brain; Mk, Meckel's cartilage; n, nasal pit; o, generative 
gland; pn, mesonephros: pq, palato-quadrate cartilage; pp, peritoneal cavity; r, muscles 
of the eye: sc, cavity between the viscera and yelk; sv, sinus venosus; v, vent or anus; 
vv, vitelline vessels; w, ventricle: Wd, Wolffian or segmental duct; y, yelk; yh, yelk 
hypoblast or periblast. 
Fig. 4.-Foetus hardened, in its follicle to show the way in which it is coiled up. x 15. 
Fig. 5.-Transverse section of an advanced embryo through the yelk-sac and body in the vicinity of 
the pectoral fins, x 29. 
Fig. 6.-Longitudinal vertical section of an advanced embryo somewhat off of the median line, x 29. 
Fig. 7.-Longitudinal vertical section of a similar embryo nearer the median line than the preceding, 
x 29. Proceedings Nat. Hus., Vol. VIII, 1885. Ryder on Viviparous Fishes. 
PLATE VII. 
DEVELOPMENT OF GAMBUSIA PATRUELIS. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
EXPLANATION OF PLATE VIIL 
Refebences: ab, air-bladder; ao, aorta; au, auricle; ba, bulbus aortae; ch, chorda dorsalis; ep, 
epiblast of yelk-sac; fm, follicular membrane; fv, follicular vessel; i, intestine; L, 
liver; mp, follicular pore; ms, medulla spinalis; pf, muscular mass at base of pectoral ; 
sv, sinus venous; t, pectoral in section; vc, venae caudalis and cardinales; ve, ventricle; Wd 
Wolffian duct; y, yelk; yhy, yelk hypoblast or periblast. 
Fig. 8.-Transverse section through the heart, showing its chambers, x 35. 
Fig 9.-Transverse section through a follicle containing an embryo, cutting through the plane of the 
follicular pore, x 32. 
Fig. 10.-A follicle containing an embryo with two immature eggs attached near the follicular-pore. 
xl5. 
Fig. 11.-Ovary of Gambusia, showing the way in which the ripe ova are attached to the median ves- 
sel. x 3. 
Fig. 12.-Follicular pore and adjacent parts of membrane of a follicle, showing the course of the ves- 
sels, from a transparent preparation in balsam, x 65. Proceedings Nat. Mus., Vol. VIII, 1885. Hyder on Viviparous Fishes. 
PLATE VIII. 
DEVELOPMENT OF GAMBUSIA PATRUELIS. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
References : a, anal fin; ab', branchial artery; &. basicranial end of chorda; bh, branchial arch; c, 
cartilaginous axis of branchial filament; ch, chorda dorsalis ; e, end organs or neuromasts 
on the cheeks; hs, haemal cartilages; I, intervertebral ligaments; ms, medulla spinalis; 
n, nerves to neuromasts; ns, neural spines; r, rays; rb, rib; sh, membranous basis of 
vertebral centrum; sp, scale pouches; w, haemal cartilage into which the urochord pro- 
jects ; vr, vertebral centra; vt, ventral fin. 
Fig. 13.-Earliest embryo of Gambusia as it lies in the follicle, x 22. 
Fig. 14.-Part of follicular membrane very much magnified, x 365. 
Fig. 15.-Cross-section through the tail of an advanced embryo, x 35. 
Fig. 16.-Longitudinal section through the anterior end of the incipient vertebral column of an ad- 
vanced embryo, x 96. 
Fig. 17.-Longitudinal section through a branchial filament, x 183. 
Fig. 18.-The tail of an advanced embryo, drawn from sections, x 35. 
Fig. 19.-A vertebral segment and rib in place, from an advanced embryo, x 365. 
Fig. 20.-Neuromasts with their nerves from a cross-section through the cheek of an advanced embryo, 
x 200. 
EXPLANATION OF PLATE IX. Proceedings Nat. Mus., Vol. VIII, 1885. Ryder on Viviparous Fishes. 
PLATE IX. 
DEVELOPMENT OP GAMBUSIA PATRUELIS. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
EXPLANATION OF PLATE X. 
References: ab, air-bladder; ao, aorta; Au, auditory vesicle; B.B, basibranchial cartilages; Br, 
branchial bar; Brs, branchiosteges; Ob, cerebrum; Cer, cerebellum; ch, chorda; Oh, optic 
crus; Ohs, chorda sheath; Cor, coracoid end of scapular arch; d, denticles; df, dental fol- 
licle; Dm, dermal muscle; ep, epithelium; F, cranial fontanelle; Ghy, glosso-hyal; Hi, 
hypophysial involution ; Hm, Hyomandibular; Hy, hypophysis; H. Hy, Hypo-hyal; i, 
intestine; I. H, interhyal; In, infundibulum; it, iter a tertio ; m, myotomic segments; mb, 
mid-brain; me, mesoblast; mes, mesentery; Mk, Meckel's cartilage; ms, medulla spinalis; 
nc, canalis centralis; ns, cartilaginous or fibro-cartilaginous neural arch; o, genital folds; 
oep, oral epithelium; Oc, occipital cartilage ; Or, orbit; Om, ocular muscles; pp, peritoneal 
cavity ; pp', peritoneal or lymph cavity on either side of the air-bladder; pi, pigment layer; 
Pch, parachordal element; pch, perichondrium; Pv, Pons varolii; Pn, mesonephros ; pn, 
pineal gland ; P. Pt, palato-pterygoid; Qu, quadrate; r', r', rays; rb, cartilaginous axis of 
ribs; rb', fibro-cartilaginous investment of ribs ; rao, radix aorta; s, perichondrial tooth- 
sockets (cementum plates of pharynx); Sc, scapular part of shoulder girdle; & br, supra- 
branchial elements ; Sk, skeletogenous tract from which the vertebra and arches develop ; 
ssv, supra-spinal vessel; si, deep layer of epidermis; Sy, symplectic ; Tr, trabecula (an- 
terior part); T. er, tegmen cranii ; vc, vena cardinales ; TFd. Wolffian ducts. 
Fig. 21.-Chondrocranium of advanced embryo of Oambusia. x 35. 
Fig. 22.-Cross-section through chorda, medulla and incipient vertebrae of the tail, x 65. 
Fig. 23.-Cross-section through an embryo of Oambusia just behind the occiput, x 65. 
Fig. 24.-Cross-section through the tail of a similar embi'yo. x 365. 
Fig. 25.-Cross-section through the posterior part of the air-bladder and adjacent organs of an embryo 
of Oambusia. x 96. 
Fig. 26.-Section through a dental follicle and tooth from the pharyngeal region, x 365. 
Fig. 27.-Section through a group of pharyngeal teeth and their supporting plates and branchial bar. 
x 365. 
Fig. 28.-Median section through the brain of an advanced embryo of Oambusia. x 35. Proceedings Nat. Mus., Vol. VIII, 1885. Ryder on Viviparous Fishes. 
PLATE X. 
DEVELOPMENT OF GAMBUSIA PATRUELIS. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 
[1885. 
EXPLANATION OF PLATE XL 
Fig. 29.-Transparent view of a developing ovum of Fundulus forced from the ovary by pressing the 
abdomen of the living fish, x 32. 
Fig. 30.-A similar view of another ovum of the same species; the embryo viewed from below and 
anteriorly, instead of from above, as was the preceding, x 32. Proceedings Nat. Mus., Vol. VIII, 1885. Ryder on Viviparous Fishes. 
PLATE XI. 
DEVELOPMENT OF FUNDULUS. PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM 
[1885. 
EXPLANATION OF PLATE XE 
References : a, anal lobe; ao, aorta; c, vitelline capillaries; ca, carotid; cd, caudal lobe; cv, cardinal 
or caudal vein; ch, chorda dorsalis; d, dorsal lobe; ep, epiblast of yelk-sac; g, goblet 
cells of epiblast; h, heart; i, intestine; j, jugular; L, liver; ll, lateral line ; m, cut ends of 
dorsal halves of the muscular segments; ms, medulla spinalis; n, caudal notch; o, round 
spaces in sections of vitellus in which oil has been contained ; p, periblast; pe, pectoral; 
pv, portal vessel, preanal lobe; r, ribs, cut through obliquely in the section; s, segmental 
intermuscular vessels; sc, caudal sinus; sol, subclavian artery; sd, soft dorsal lobe; si, 
subintestinal vessel; v, vent; vv and vv', vitelline veins; vc, venae cardinales; vm, meso- 
blastic investment of vitellus; vt, ventral fln; TFd, Wolffian duct. 
Fig. 1.-Recently hatched embryo of the Scboodic or landlocked salmon, viewed from the left side and 
figured from the living specimen, x 7 +• 
Fig. 2.-Same viewed from the right side, x 7 +• 
Fig. 3.-Diagram of the circulatory system of the young salmon. 
Fig. 4.-Cross-section through the body and yelk of a young salmon, through the region of the liver, 
x 16. . 
Fig. 5.-Section through the epidermis of a salmon embryo to show the goblet cells g. x 200. Proceedings Nat. Mus., Vol. VIII, 1885. Ryder on. Salmo salar. 
PLATE XII. 
CERTAIN FEATURES OF THE DEVELOPMENT OF THE SALMON.