EMBRYONIC DEVELOPMENT OF THE PLATYFISH …gato-docs.its.txstate.edu/jcr:ac37f118-98aa-4b1b-80f1-6828143efb11... · between Platypoecilus development and that of Xiphophorus and the

EMBRYONIC DEVELOPMENT OFTHE PLATYFISH (PLATYPOECILUS),THE SWORDTAIL (XIPHOPHORUS),

AND THEIR HYBRIDS

WILLIAM N. TAVOLGA

BULLETINOF THE

AMERICAN MUSEUM OF NATURAL HISTORY

VOLUME 94: ARTICLE 4 NEW YORK: 1949

EMBRYONIC DEVELOPMENT OF THE PLATYFISH (PLATY-

POECILUS ), THE SWORDTAIL (XIPHOPHORUS ),

AND THEIR HYBRIDS

EMBRYONIC DEVELOPMENT OF THE

PLATYFISH (PLATYPOECILUS ), THE

SWORDTAIL (XIPHOPHORUS ),AND THEIR HYBRIDS

WILLIAM N. TAVOLGA

The American Museum of Natural New York University

History

A DISSERTATION SUBMITTED TO THE FACULTY OF THE GRADUATESCHOOL OF ARTS AND SCIENCE OF NEW YORK UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FORTHE DEGREE OF DOCTOR OF PHILOSOPHY

BULLETINOF THE

AMERICAN MUSEUM OF NATURAL HISTORY

VOLUME 94 : ARTICLE 4 NEW YORK : 1949

BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY

Volume 94, article 4, pages 161-230, textfigures l-48, tables l-7

Issued November 1, 1949

Price: $.85 a copy

CONTENTS

INTRODUCTION - 1 167Acknowledgments

MATERIALS AND METHODS 167

Method of Obtaining Embryos 168

168Species and Strains of Fish Used Maintenance of Fish

168 170

Treatment of Embryos 170EMBRYOLOGY OF Platypoeciluslat 171

Description of Normal Stages in the Embryonic Development of Platypoecilus 171Reproductive Cycle and Embryonic Growth in Platypoecilus 208Discussion of Organogenesis in Platypoecilus 211

Extra-embryonic Membranes 211Cleavage and Gastrulation 212Blastopore Closure 213Notochord Formation : 213Ectodermal Derivatives 213Pigment FormationEndodermal Derivatives

213 214

Mesodermal DerivativesCOMPARISON OF Platypoecilus WITH Xiphophorus HYBRIDS

215 217

Time of Initial Pigment Formation 217Growth RatesGrowth of the Caudal Fin :

218 220

Fecundity of Hybrids 222Discussion of the Effects of Hybridization on Development 223

SUMMARYLITERATURE CITED : :

226227

165

M

INTRODUCTION

IN A PRELIMINARY PAPER,Tavolga and Rugh(1947) described the normal embryonic stagesof Platypoecilus maculatus in terms of grossmorphology. Based upon the stages pre-viously defined, the present work describesthe development of organ systems, pigmentdifferentiation, and embryonic growth.

As described by Gordon (1948 and earlierpapers), the related swordtail (Xiphophorushellerii) can be hybridized readily in thelaboratory with P. maculatus. Peculiar effectsof heterosis and gene reassortment are pro-duced, particularly with regard to the de-velopment of spontaneous melanomas in thehybrids. Consequently, a comparison is madebetween Platypoecilus development and thatof Xiphophorus and the Platypoecilus-Xipho-phorus hybrids, with special emphasis on theeffects of hybridization on embryonic growth.

The previous paper (Tavolga and Rugh,1947) represented the first phase of the prob-lem, i.e., the definition of normal stages.The present work describes the second andthird phases, namely, a detailed account ofdevelopmental morphology and a comparisonof this account with the prenatal develop-ment of hybrids possessing potentialities foratypical pigment cell growths.

ACKNOWLEDGMENTS

The writer is indebted to Dr. Charles M.Breder, Jr., Chairman of the Departmentof Fishes and Aquatic Biology, of the Ameri-can Museum of Natural History, and ofNew York University for his encouragementand sponsorship of this work and for hisreading and criticism of the paper. Thanksare extended to Dr. Roberts Rugh of theRadiological Laboratory of the College ofPhysicians and Surgeons and of New YorkUniversity, and to Dr. Myron Gordon, ofNew York University and the New YorkAquarium, New York Zoological Society, fortheir comments and criticisms.

The laboratory facilities were made avail-able through the courtesy of Dr. Lester R.Aronson of the Department of AnimalBehavior, the American Museum of NaturalHistory. The fishes used here were suppliedby Dr. Myron Gordon from the stocks in theGenetics Laboratory of the New YorkZoological Society. This laboratory is sup-ported in part by grants from the NationalCancer Institute, United States PublicHealth Service.

167

MATERIALS AND METHODS

METHOD OF OBTAINING EMBRYOS

Because of the viviparous nature of repro-duction in the xiphophorin fishes, it wasnecessary to dissect the embryonic broodsfrom the ovaries of gravid females. Themethod of obtaining broods of various ageswas that described by Hopper (1943) andTavolga and Rugh (1947), in which the dateof birth of the previous brood was consideredas the starting point for the development ofthe next complement of eggs. The interval

between broods in all the species used hereaveraged 28 to 30 days, and this period issomewhat longer than the actual gestationperiod, since, as reported by Hopper (1943),fertilization occurs at about seven days afterthe birth of a brood. By dissecting females atvarious times during the brood interval, anestimate of the developmental stage of theembryonic brood at a given time could thusbe determined.

SPECIES AND STRAINS OF FISH USED

The animals used here were viviparousteleost fishes of the tribe Xiphophorini,family Poeciliidae, Order Cyprinodontes(Order Haplomi according to older classifi-cations). These forms, represented only bythe two genera Platypoecilus and Xiphophorus,are commonly kept in home aquaria andpopularly referred to as platyfishes and sword-tails, respectively.

Members of this group have frequentlybeen used in the study of genetics, sex deter-mination, and speciation. In two recentpapers, Gordon (1947b, 1948) reviewed thework on the Xiphophorini with respect topopulation genetics and genetic control oftumor formation.

The original stock fishes were collectedfrom several east Mexican rivers, and thelocalities mentioned below may be found onmaps figured by Gordon (1947b).

Platypoecilus maculatus GUNTHER

The majority of the P. maculatus usedwere obtained from an inbred stock originallycollected from Plaza de Agua at el Tejar,Rio Jamapa, Mexico, by Gordon and Gordonin 1939, and cultured at the Genetics Labora-tory of the New York Zoological Society.The number of wild stock females used was89, which yielded by inbreeding a total of2425 embryos.

Several pattern strains were included.Most of the adults possessed one or two ofthe autosomal allelic series of genes produc-

ing the several tail patterns of micromelano-phores described by Gordon (1947b). Inaddition, some of the strains carried sex-linked genes for macromelanophores (Gordon,1948). The distinction between the two typesof melanophores is important, since it is onlythe macromelanophore type that, by hy-bridization, can be intensified into a melanoticor melanomatous condition (Gordon, 1927,1936).

In addition to the above wild stock ofP. maculatus, 10 females of an inbred domesti-cated strain of golden platyfish were exam-ined. The golden color is produced by a greatreduction in the number of micromelano-phores and an apparent increase in the num-ber of xanthophores. This effect is the resultof an autosomal recessive gene “st,” whichappeared as a mutant during laboratoryculture (Gordon, 1935).

OTHER SPECIES OF Platypoecilus

Embryonic broods were obtained fromfemales of the following species, the strainsall being of the inbred wild type, originallycollected at the localities given:

Platypoecilus variatus Meek (15 females),collected at El Nilo in the Rio Panuco sys-tem, Mexico, in 1940, by the New YorkAquarium Expedition.

Platypoecilus couchianus Girard (five fe-males), collected at Rio Santa Catarina inthe Rio Grande system, Mexico, in 1939, byGordon and Gordon.

168

1949 TAVOLGA: EMBRYONIC DEVELOPMENT IN FISH 169.

Platypoecilus xiphidium Hubbs and Gor-don (five females), collected at Rio Purifica-tion in the Rio Soto la Marina system, Mex-ico, in 1939, by Gordon and Gordon.

The genetic relationships of these speciesto P. maculatus were discussed by Gordon(1946a, 1946b, 1947) and Gordon and Smith(1938b).

Xiphophorus hellerii HAECKELTwenty-three females of a ninth genera-

tion inbred stock of the wild type of X.hellerii were used to obtain embryonic broods.These were descendants of a group originallycollected in 1939 by Gordon and Gordon atArroyo Zacatispan, in the Rio Papaloapansystem. Four of these females were mated tolitter-mate males.

The remaining 19 females were mated tomales of another strain. The latter group ofmales was the result of a cross between afemale of the Zacatispan strain and a male X.hellerrii from a population at Cordova, onthe Rio Blanco, originally collected in 1939by C. L. Turner.

An additional group of gravid femaleswordtails was obtained in preserved statefrom a 1948 collection taken at L’Encero,near Jalapa, in the Rio Jamapa system byM. Gordon, J. W. Atz, and F. G. Wood.Since none of these was available alive, nodeterminations of gestation period or em-bryonic age were possible, but some data onthe embryos of this strain are included in alater section.

Xiphophorus-Platypoeeilus HYBRIDS

The original mating producing all the F1hybrid fishes used in this work was that of aX. hellerii male (Zacatispan strain) and aP. maculatus female (Jamapa strain), bothof wild stock, and the latter possessed a genefor macromelanophores on each of its X-chromosomes. The two genes were Sp (spotson the sides) and Sd (spots on the dorsalfin). Twelve such females were dissected forFl embryonic broods. This cross was firstdescribed by Gordon in 1948, who gave pre-liminary results on the sex ratios of the hy-brids.

TABLE 1

Strain or Species

Pure speciesP. maculatus, wild stock, inbredP. maculatus, golden strainP. variatusP. couchianusP. xiphidiumX. hellerii, laboratory bred, Zacatispan strainX. hellerii, field bred, L’Encero strain

Platypoecilus-Xiphophorus hybridsP. maculatus Q X X. hellerii 3

(female with sex chromosomes XSpXSdFl Sp Q)(RSP d’FlSp QxFlSd 3F1 Sp XP. maculatus C?

F1 Sd XFlSpc?F1 Sd 9 XFdd d’F1 Sd XP. maculatus ~7F1 Sd X X. hellerii 8P. maculatus X F1 Sd 3

Total

Number ofFemales

891015

55

2320

12

1047

12116895

Number ofEmbryos

2425108412132110650370

317

19682

13222597383687

177

5594251

F1 Sp x X. hellerii 8

170 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 94

Mature Fl hybrids possessed one of thetwo pattern genes (a 1:1 ratio of Sp to Sd),as a result of Mendelian segregation. Most ofthese animals were females, and a total of 67of them were mated in various combinations

for the purpose of obtaining embryos.Table 1 summarizes all the strains and

matings used here, the number of females ineach case, and the total number of embryosexamined.

MAINTENANCE OF FISH

The female fish were isolated in two- orfive-gallon aquaria which were maintained ata temperature of 75O to 800 F. in the green-house of the Department of Animal Behaviorat the American Museum of Natural History.Males were introduced for a long enoughperiod to permit insemination a n d develop-ment and birth of an initial brood, then wereremoved and re-introduced into tanks withother females. Removal of males after initial

TREATMENT OF EMBRYOS

In general, the entire ovary was removed,since the embryos develop within the ovarianfollicles. The follicles were dissected and theembryos removed in a .9 per cent saline solu-tion. The embryos were then examined whileliving and allocated to the various develop-mental stages. Most of the embryos werepreserved in 10 per cent formalin for futurestudy. A number of embryos of eachstage, representing each of the several strainsof fishes, were fixed in Bouin’s picro-formoland dehydrated in dioxan. Some of the lattergroup were used for total mounts both in anunstained condition and stained with Harris’haematoxylin. The remainder were sectionedserially at 8 microns and stained with Harris’haematoxylin and eosin or a modification ofthe Masson trichrome stain. In many casesthe embryos were removed from the yolkeither before or after fixation and then sec-tioned after being embedded in paraffin. Inorder to leave the embryo-yolk relationshipundisturbed, it was found necessary to sec-

insemination did not interrupt the productionof future broods, since sperm were stored bythe female in sufficient quantities for thefertilization of several successive comple-ments of ova. Thus a small number of activemales known to be fertile could be utilizedmost efficiently.

The fishes were fed daily on a diet of dried,powdered shrimp mash, alternated withliving Daphnia.

tion some of the embryos with the entireyolk attached. A modification of the Peterfimethod of double-embedding was used here,in which the celloidin was dissolved in di-oxan.

Representative embryos of all stagespossessing an actively beating heart were in-jected with India ink for the study of thecirculatory system. Because of the peculiarposition of the embryonic heart in theseforms, the sinus venosus was exposed andcould be penetrated by a capillary pipet.India ink, diluted with saline solution, wasthen allowed to flow into the vascular system,being pumped through by heart action. Thiswas stopped at the desired point by floodingthe embryo with hot Bouin’s solution. Theinjected specimens were cleared and mountedin Clarite.

The reconstruction drawings were madefrom combined observations on living andpreserved specimens, both total mounts andserial cross sections.

EMBRYOLOGY OF PLATYPOECILUS

DESCRIPTION OF NORMAL STAGES IN THE EMBRYONICDEVELOPMENT OF PLATYPOECILUS

THE FOLLOWING DESCRIPTIONS of embryologi-cal development of Platypoecilus were basedprimarily on the P. maculatus material.However, observations on Platypoecilus vari-atus, P. couchianus, and P. xiphidium areincluded, since no distinctions could be foundin either gross or microscopic morphology inthe development of the four species. This wasdetermined on the basis of the somewhatsmaller amount of material representing thethree latter species (see Materials and Meth-ods section).

The diagnoses of the morphological stagesare essentially the same as those described byTavolga and Rugh (1947) with the additionof observations on internal anatomy and or-ganogenesis. The average dimensions givenhere are to be considered more accuratesince they are based on a much larger sample.Estimated ages given for each stage arebased on the determinations described in alater section.

CONDITION OF OVARY AFTER BIRTH OF A

BROOD AND PRIOR TO FERTILIZATION

The average size of the ovocytes two tothree days after the birth of a brood is 0.7mm. Some yolk is present in the cytoplasm,and the follicular epithelium is cuboidal andvacuolated. The material in these vacuoleshas the same staining properties as the yolk.The vacuoles are frequently seen partiallyextruded into the follicular cavity and incontact with the egg cytoplasm.

The nuclei are central in position in ovo-cytes up to 0.9 mm in diameter. Above thatsize, yolk deposition is completed and thenucleus moves to the periphery, where mat-uration divisions take place. These observa-tions are in agreement with those of Hopper(1943).

STAGE 1

This stage is to be considered as represent-ing 0 time of embryonic development.

The mature ova, just prior to or duringfertilization, average 1.6 mm. in diameter.

They are of a clear yellow color with periph-erally arranged fat globules of various sizes.When the egg is damaged, the globules arefound to remain embedded in a more gelat-inous cortical region. The globules are com-posed of a colorless oil of apparently lowerviscosity.

Unfertilized, degenerating eggs are of fre-quent occurrence, particularly in virgin fe-males, and can be identified easily by theaggregation of oil globules at one pole andthe presence of irregular strands of coagulatedcytoplasm on and below the egg surface,usually near the polar cap of oil globules. Asimilar condition characterizes degenerate,early gastrula stages, but these can be dis-tinguished by the presence of coagulatedcytoplasm on the egg surface only.

STAGE 2

Age, 0.3 day.This stage is defined as representing the

period of cleavage and early blastula forma-tion. It is not readily distinguishable withoutsectioning, since the cleavage cells are broad,flat, and extremely thin and transparent.Hopper (1943) showed figures and photo-graphs of sections of cleavage stages, andhis observations on these were confirmed inthe present work.

STAGE 3Figure lA, B

Age, 0.5 day.The compact blastula is very flat, not

more than 30 to 50 microns thick in thecenter, and averages 0.4 mm. in diameter(fig. 1A). The center is composed of twolayers of irregular polyhedral cells 15 to 20microns in diameter. The cells are smallerand more numerous peripherally, averaging 5to 10 microns, and are arranged in three orfour layers. Around the rim of the blastulathere is a narrow zone of junction. A verythin, flat, segmentation cavity can be dis-cerned between the blastula and the yolksurface (fig. 1 B).

171


zone o f j u n c t i o n

D

C S T A G E 4

FIG. 1. A. Stage 3, top view. B. Stage 3, cross section at level indicated byarrow. C. Stage 4, top view. D. Stage 4, cross section at level indicated. X60.

STAGE 4Figure lC, D

Age, 0.8 day.The diameter of this later blastula or

germ ring is 0.8 mm. (fig. 1C). The thicknessat the center is 30 microns. This region iscomposed of three to four layers of cellsaveraging 10 to 15 microns in size. The pe-ripheral germ ring region is 0.06 mm. thick,0.1 mm. broad, and consists of 10 to 15layers of very small, rapidly dividing cells(fig. 1D). The thinner central region has fewmitotic figures and forms a syncytium. Itmay therefore be termed the periblastic ecto-derm, since there is some doubt as to whetherits origin is the same as that of the periblastdescribed by Wilson (1889). The segmenta-

tion cavity is still very thin, with the undersurface of the blastula often in contact withthe underlying yolk (fig. 1D). A narrow andincomplete zone of junction is present.

STAGE 5Figure 2A, B, C

Age, 0.9 day.At this stage, the periblastic .ectoderm

covers approximately one-quarter of the sur-face of the yolk (fig. 2A). One segment of thegerm ring has expanded and thickened t oform the embryonic shield (or germinalshield) which will form the embryo proper(fig. 2A). The caudal lip of the germinalshield is thickened and may be consideredhomologous to the dorsal lip of the blasto-

TAVOLGA: EMBRYONIC DEVELOPMENT IN FISH 173

FIG. 2. A Stage 5, top view. B. Stage 5, cross section at level indicated byarrow. C. Stage 5, sagittal section at level indicated. D. Stage 6, top view ofembryonic shield. E. Stage 6, cross section at level indicated. F. Stage 6, sagittalsection at level indicated. X6060.

6

pore in other animals (fig. 2C). The lengthof the germinal shield is 0.35 mm. It pos-sesses a thickened central region which re-sults from the proliferation and invaginationof cells to form the neural keel (fig. 2B).

STAGE 6Figure 2D, E, F

Age, 1.7 days.The germ ring has now taken an almost

equatorial position on the surface of the yolk,i.e., about one-half of the yolk surface iscovered by periblastic ectoderm. The embry-

onic shield is 0.6 mm. long and 0.25 mm.broad at the middle (fig. 2D). The neural keelis formed in the anterior two-thirds of theshield (fig. 2E, F).

STAGE 7Figure 3

Age, 1.9 days.The length of the embryonal area is about

0.82 mm. The entire embryo appears to con-sist of the neural keel, which is slightlybroader at the cephalic end (fig. 3E, F).Serial cross sections show the presence of asecond cord of cells extending cephalad


extra - e m b r y o n i cm e m b r a n e

n e u r a l k e e l

d o r s a l l i p

c h o r d a - e n d o d e r m

extra-embryonic m e m b r a n ek e e l

FIG. 3. Stage 7. A-D. Cross sections at levils indicated. E.Dorsal view.F. Lateral view. X90.

from the region of the dorsal lip of the blasto-pore about two-thirds the length of the em-bryo (fig. 3B, C, F). The cord of cells, situ-ated ventral to the neural keel and contigu-ous with it, is provisionally called the chorda-endoderm (see description of stage 8).

A small cavity is formed on the under sur-face of the caudal region of the germinalshield and is identified as Kufler’s vesicle(fig. 3D, F). A fold of periblastic ectodermcovers the anterior tip of the cephalon (fig.3A, E, F). ‘This two-layered fold is the pri-mordium of the extra-embryonic pericardialmembranes. In subsequent stages, the fold isinvaded by mesenchyme, and forms twodouble-layered somatopleural membranes, i.e.

pericardial amnion and pericardial serosa(described by Tavolga and Rugh, 1947).

STAGE 8Figure 4

Age, 2.3 days.

The total length of the embryo at this stageaverages 1.06 mm.

The nerve cord extends for the entirelength except for a short undifferentiatedregion at the caudal end (fig. 4E). The an-terior half of the nerve cord possesses acavity which is widest in the region of theprosencephalon (fig. 4A, E). A pair of lateraloutgrowths, optic buds, are present. Theseare solid at first, but quickly develop smallcavities (fig. 4A, E, F).

The ectoderm of the head is folded ventrad,forming a head fold with an underlying sub-cephalic pocket (fig. 4A, F). The head regionis enclosed in a thin, double-layered sac ofperiblastic ectoderm.

The auditory placode is present at thisstage (fig. 4B, E, F). Mesenchyme is aggre-gated on either side of the neural keel in themidbody region where the first somites willform (fig. 4C, E, F).

The chorda-endoderm appears to split

1949 TAVOLGA: EMBRYONIC DEVELOPMENT IN FISH 175

p o s terior mar ginof extra-embryonic membrane

g e r m ring

nerve cord p e r i b l a s t i c ectoderm

p o c k e t A 8 c vesicleD

FIG. 4. Stage 8. A-D. Cross sections at levels indicated. E. Dorsal view. F. Lateral view. x90.

longitudinally, with the ventral cord of cells closes the head to the level of the auditorybeing the solid primordium of the gut and the placodes (fig. 5H). The blastopore is ellipticaldorsal cord forming the notochord (fig. 4C, F). in shape, about 0.5 mm. in its longest axis.The shallow Kupffer’s vesicle marks the A small tail bud, 0.05 mm. in length, is pres-posterior limit of the chorda-endoderm (fig. ent, overhanging the anterior lip of the4D, F). blastopore (fig. 5G, H).

The germ ring, by this stage, has ad-vanced considerably, leaving about one-fifth of the yolk surface uncovered by peri-blast. This cap of uncovered yolk may beconsidered homologous to the yolk plug ofamphibian eggs.

STAGE 9Figure 5

Age, 3.0 days.The total length of the early tail bud stage

is 1.23 mm. The head has expanded con-siderably, primarily as a result of the growthof the optic vesicles (fig. 5A, G). The widthof the mesencephalon region is 0.08 mm. Theextra-embryonic periblastic membrane en-

The nerve cord is tubular throughout, ex-cept in the relatively undifferentiated tailbud region. The neurocoele is widest in thehead region, where primary divisions intoprosencephalon, mesencephalon, and rhomb-encephalon can be distinguished (fig. 5G).The optic vesicles are large, hollow, and con-nected to the prosencephalon by solid opticstalks (fig. 5A, G, H). In cross section, lensplacodes are shown (fig. 5A). The first cranialplacode of neural crest (anlage of the fifthcranial nerve) is present (fig. 5B, G, H).

The gut endoderm is a cylindrical cord ofcells extending from Kupffer’s vesicle (post-anal gut) anteriorly to the level of the firstcranial placode where the pharyngeal re-


bryonic membrane

A B

m e s o d e r m

FIG. 5. Stage 9. A-F. Cross sections at levels indicated. G. Dorsal view. H. Lateral view. x90.

gion becomes flattened (fig. 5B, C, D, H). Thenotochord can be traced from the blastoporelip region to the level of the auditory placode(fig. 5D, E, F, H).

Five to seven pairs of somites may bepresent in this stage (fig. 5G, H). Occa-sionally somites appear as early as stage 8. Anaggregation of mesenchyme may frequentlybe found ventral to the nerve cord, anteriorto the first cranial placode. This is inter-preted as the anlage of the endocardium.

STAGE 10Figures 6 and 7

Age, 3.4 days.The total length at this stage is 1.45 mm.,

and the width at the mesencephalon is 0.13mm. A bluntly rounded tail bud, 0.20 mm.long, overhangs the still open blastopore(fig. 6A, B). The extra-embryonic membrane

encloses the head as far back as the auditoryvesicle (figs. 6A, B, 7A-F). The subcephalicpocket extends to about the middle of themesencephalon.

The three primary divisions of the brainare easily seen, with the mesencephalicportion possessing the thickest walls (fig. 6A).The telencephalic portion of the prosen-cephalon is small and acuminate (figs. 6A,7A). The diencephalic region is relativelylong and deep. The deepest point representsthe site of the infundibulum (fig. 6B).

The optic vesicles are folded into opticcups with a thickened sensory layer (figs.6A, 7A). The optic stalks are hollow at thisstage. The lens placodes are partially in-vaginated (figs. 6A, 7B). Small olfactoryplacodes are present (figs. 6A, 7A). The au-ditory placode has invaginated as a solidmass and, at this stage, has formed a small

S T A G E 1 0

FIG. 6. Stage 10. A. Dorsal view. B. Lateral view. x90.

Aplacode

d i e n c e p h a l o n

/

1st

a u d i t o r y vesicle

S T A G E 1 0FIG. 7. Stage 10. A-L. Cross sections at levels indicated in figure 6B. x90.

E F G H I J K

S T A G E 1 1FIG. 8. Stage 11. A. Dorsal view. B. Lateral view. x90. I

myelencephalon e x t r a - e m b r y o n i c membrane n e u r a l crest

\ /

e x t r a - e m b r y o n i c m e m b r a n e

FIG. 9. Stage 11. A-L. Cross sections at IeveIs indicated in figure 8B. x90.

m e s e n c e p h a l o n m e t e n c e p h a l o n otic vesicle


FIG. 11. Stage 12. A. Ventral view. B. Lateral view of brain and cranial nerves. X60.

BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 94180

FIG. 12. Stage 12. A. Ventral view of arterial system. B. Lateral view ofarterial system. C. Ventral view of venous system. D. Lateral view ofvenous system. X60.

cavity (figs. 6A, B, 7F). No endolymphaticduct or vestige of it is ever present. Two ag-gregations of neural crest cells are visible inthe cross sections and are shown in figure 6Aand B. The anterior of these is the firstcranial placode (fig. 7E), and the other is thesecond cranial placode (fig. 7F), primordiumof&he seventh and eighth cranial nerves.

Three or four of the first spinal nerve pri-mordia are also present in the form of neuralcrest (fig. 7J).

The cranial portion of the endoderm isflattened and extended dorso-laterad to formthe first evidences of the first three gillpouches, which are thickened grooves at thisstage (figs. 6B, 7D, E, G). Just anterior to


myotome

STAGE 13FIG. 13. Stage 13. A. Dorsal view. B. Lateral view. X60.

the level of the first somite, there is an in-testinal portal opening into a small tubularcavity which extends posteriorly from thatpoint to about the level of the sixth somite(figs. 6B, 71, J). Caudad, the endodermalcord is solid (figs. 6B, 7K, L), and Kupffer’svesicle is greatly reduced.

Twelve or 13 pairs of somites are present(fig. 6). From the second somite to theninth or tenth, a mesomeric region (nephro-geni c cord) can be distinguished (fig. 7J). Theheart anlage is still a diff use mass of mesen-chyme (figs. 6B, 7B, C, D). The dorsal aortamay be found as a thin-walled, flattened tubelying between the notochord and the endo-derm in the region of the first four or fivesomites (fig. 7 J). Blood-cell anlagen arefound inside and around the dorsal aorta.

A pair of thickened grooves of ectodermare present along the lateral limiting sulci,extending from the level of the first cranialplacode up to about the first somite (fig. 7E).These structures are believed to be the pri-

mordia of the rather reduced lateral linesystem typical of the Poeciliidae.


Age, 3.9 days.Total length, 1.67 mm.Width of mesencephalon, 0.25 mm.The tail bud, approximately 0.37 mm. in

length, covers most of the area of the blasto-pore (fig. 8A). The latter is small and ellipti-cal or slit shaped. Occasional embryos havebeen found with much larger blastopores atthis stage. The head lies free within a peri-cephalic (amnionic) cavity only as far backas the beginning of the rhombencephalon(figs. 8B, 9A, B, C). The extra-embryonicmembrane is thicker and possesses somemesenchyme at its margins (figs. 8B, 9A-F).This is the first step in the transformationof the periblastic-ectodermal extra-embryonicmembranes into the definitive inner pericar-dial amnion and outer pericardial serosa,


1st a o r t i c arch(ef ferent s p r i a c u l a r )

0

STAGE 16

p o s t e r i o r v e n t r a l a o r t a

FIG. 14. Left lateral views of aortic arches and ventralaortae of stages 13 to 17.

each a somatopleural membrane.The mesencephalon is widened, and the

rhombencephalon less so. The infundibularregion is deep and distinct (fig. 8B). Therhombencephalon, in the region of the cranialplacodes, possesses a thin roof, i.e., the char-acteristic tela choroidea of the myelencepha-lon (fig. 9D). The lens placode is fully in-vaginated, solid, but not yet separated fromthe superficial ectoderm (figs. 8A, 9B). Theolfactory organs appear as a pair of smallsolid masses of invaginated ectodermal tissue

(figs. 8A, 9B). The lateral-line anlagen ex-tend from the level of the first gill pouch tothe pectoral fin bud (fig. 9D, E, F).

The endoderm shows little change fromthe condition in stage 10. The intestinal por-tion possesses a cavity extending back to thefifteenth somite (figs. 8B, 9H-K). The threegill pouch primordia are more sharply de-fined (figs. 8A, B, 9D, E).

Nineteen to 20 somites are present at thisstage (fig. 8). A pair of pectoral fin budsmake their appearance just anterior and



lateral to the first somite (figs. 8A, B, 9G). Themesomere is distinguishable from the first tothe thirteenth somites. At the level of thesecond to fourth somites, the mesomere isorganized into a nephric duct (fig. 91). Thecirculatory system consists of only a ratherdiffuse heart anlage and a single vessel, thedorsal aorta (figs. 8B, 9A-D, H, I).

STAGE 12Figures 10-12

Age, 4.2 days.Total length, 1.81 mm.Tail length, posterior to anus, 0.48 mm.Width of mesencephalon, 0.38 mm.When embryos at this stage are freed from

their membranes, they exhibit slow twitch-ing movements of the body and tail. Theextra-embryonic membranes are complete,forming a two-layered sac and enclosing thehead of the embryo, including, posteriorly,the auditory vesicles. These membranes,

namely, the pericardial amnion and pericar-dial serosa, are indicated in figure 1OB. Theblastopore is usually closed at this stage,but a faint raphe, marking the closure of thelips, may be distinguished. This raphe islocated postanally and underlies the tail bud.

The mesencephalon is considerably widenedand possesses thickened sides and floor (fig.10A). The metencephalon is small and rela-tively poorly distinguishable from the follow-ing myelencephalon, the latter possessing itscharacteristic tela choroidea and visibleneuromeres (fig. 1OA). The optic stalks arethin (fig. llA, B). The infundibular region ofthe diencephalon is large and deep (fig. llA,B).

Five pairs of cranial nerves can be identi-fied and traced in serial cross sections recon-structed in figure 10B. The trigeminal is thelargest and most anterior, running ventrallyand posteriorly to fuse with the facial (VII).The facial and auditory nerves arise at the


FIG. 16. Stage 14. A. Cross section at level of olfactory bulbs. B. Cross sectionat level of lenses. C. Cross section at level of epiphysis. x90.

same point, just anterior to the auditoryvesicle. The auditory (VIII) branches offand fuses with the otocyst(VII) continues ventrally

tissue. The facialand posteriorly

mordia of the first three pairs of gill pouches.The first and third possess small grooves, and

to fuse with the trigeminal (V). The glosso-pharyngeal (IX) and vagus (X) arise sepa-rately and can be traced for a short distancepostero-ventrad.

The pharyngeal portion of the gut is flat,

the second asecond pouch

small, shallowwill become the

pouch. Thefirst true gill

cleft. Posterior to the third pouch is the in-testinal portal leading into the straight in-testine and ending in a groove-shaped anus.Posterior to the anus is the shallow postanalgut (Kupffer’s vesicle) (figs. lOB, 1lA).

with three pairs of flanges extending Iateradand dorsad (fig. 11A). These are the pri-

Twenty-one to 22 compact somites may bediscerned, the anus being at the level of the

TAVOLGA: EMBRYONIC DEVELOPMENT IN FISH

C S T A G E 1 4

FIG. 17. Stage 14. A. Cross section at level of infundibulum. B. Crosssection at level of hypophysis. C. Cross section at level of otocysts. D.Cross section at posterior pharyngeal level. x90.

eighth (fig. IOA, B). The mesomeric regionis relatively undifferentiated at the level ofthe pectoral fin bud and the first somite.Posteriorly, a pair of nephric ducts extend tothe level of the seventh somite where theytaper off into a solid cord of cells (fig. 1OB).A pair of cords of primordial germ cells areaggregated into gonadial ridges in the regionof the second to fifth pairs of somites. Theposition of these is indicated in figure 11A.

The circulatory system at this stage is acomplete circuit, but little visible blood ispresent. The heart beat is irregular. Theblood circuit can be traced in this stage both

in serial cross sections and in India ink in-jected specimens (fig. 12A-D). The bloodenters the elongate sinu-atria1 region of theheart in front of the tip of the head. Apulsating ventricular region bulges slightlytowards the right side, and empties into ashort conus and ventral aorta (fig. 12A, B).The third aortic arches (first to develop) arewell formed, and a single fourth arch is usu-ally present on the left side (fig. 12A, B).Leading forward from the third arches are apair of internal carotid arteries supplying theextensive optic plexi. Running caudad, thetwo radices aortae fuse into a single dorsal

FIG. 18. Stage 14. A. Cross section at level of esophagus. B. Crosssection at level of gastric region. C. Cross section at level of liver anlage.D. Cross section at midbody level. x90.

l i v e r

FIG. 19. Stage 15. A. Dorsal view. B. Lateral view. X30.

S T A G E 1 5FIG. 20. Stage 15. A. Dorsal view, showing distribution of melanophores.

B. Ventral view, showing viscera and gills. x30.

l i v e r intestine

S T A G E 1 6

FIG. 2.1. Stage 16. A. Dorsal view. B. Lateral view. X30.

I]

y

1,i

ti1,; 188 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 94t!t’1’I::1:I*iI :I’‘8,i/


aorta just posterior to the level of the audi-tory vesicles. At that point, a pair of smallsubclavian arteries are given off. The dorsalaorta extends into the tail region as thecaudal artery, giving off numerous interseg-mentals along the way (fig. 12A, B).

A pair of large anterior cardinal veins leadfrom the optic plexi to the ducts of Cuvier(fig. 12C, D). The caudal vein is drainedby the vitello-caudal vessel emptying into theyolk sac circulation at a level just posteriorto the anus. The caudal vein is also drainedby the postcardinals. The left postcardinal isa vestigial structure, whereas the right is welldeveloped (fig. 12C, D). A circle of smallarterial vessels is present beneath themesencephalon, posterior to the infundibulum(fig. 12C). These blood vessels arise from theinternal carotid plexus in the eye, give off apair of cerebellar arteries, and fuse into asingle median basilar artery. The basilarartery is short and, at this stage, is connectedby a collateral cross connection to the

anterior cardinals (fig. 12C). This circle ofblood vessels, as will be shown later, comes tosurround the infundibulum and consequentlymay be considered comparable to the circleof Willis, present in birds and mammals.

The ducts of Cuvier lead into the extensivevascular yolk sac (fig. 12C, D).

STAGE 13Figure 13

Age, 4.8 days.Total length, 2.02 mm.Tail length, 0.62 mm.Mesencephalon width, 0.42 mm.At this stage the blastopore is completely

closed. Usually a thin streak is visible extend-ing for about 0.4 mm. on the surface of theyolk membrane underlying the tail, markingthe site of closure of the blastopore.

Aside from a general difference in size, themost prominent feature distinguishing thisstage from stage 12 is the presence of the firstindications of visible pigment (fig. 13A). This

c


FIG. 23. Stage 17. A. Dorsal view, showing distribution of melanophores.B. Ventral view, showing viscera and gills. x30.

appears in the outer retinal layer, delineatingthe eyes of living specimens with a thin darkmargin.

In the pharyngeal region, four pairs of gillpouches are present, with the first (spiracularor pseudobranchial) somewhat constricted(fig. 13B). The development of the postanalgut (Kupffer’s vesicle) into the urinarybladder (or vesicle) is indicated by an enlarge-ment and invagination of the former (fig.13B).

Considerable changes have taken place inthe circulatory system. There are usually fourpairs of aortic arches present, i.e., the thirdfourth, fifth, and sixth (fig. 14A). Occasion-ally the fourth or the sixth pairs are in-completely formed. The ventral aorta poste-rior to the fourth arches is double, and it isnoteworthy that the ventricular end of theventral aorta is anterior to the third aorticarch. The blood flow in the ventral aorta,then, is caudad.

The circle of Willis partially surrounds theinfundibulum, and the collateral cross con-nection of the basilar and the anteriorcardinals is greatly reduced or absent. Thebasilar runs caudad and splits into a pair ofvertebral arteries at about the level of thepectoral fin buds.

The pericardial serosa is partially vascular-ized by branches of the common cardinalveins, and is drained by the sinus venosus.


Age, 5.0 days.Total length, 2.31 mm.Tail length, 0.84 mm.Mesencephalon width, 0.56 mm.The pericardial serosa is well vascularized

and is continuous with the extensive yolk sacmembrane (fig. 16C). Pigment is limited tothe retinal layer but is more evident anddarker than in stage 13 (fig. 16B).



The telencephalon is short, extending some-what ventrad between two small olfactorybulbs (fig. 16A). A pair of small, flattenedtelencephalic vesicles are present. In the sametransverse section (fig. 16A), the anteriorregion of the diencephalon is visible, with athin dorsal roof (anterior tela choroidea). Themesencephalic (optic) lobes are large andthick walled, encroaching anteriorly over thediencephalon (figs. 15A, B, 16C, D). Theoptic stalks are vestigial or absent, originat-ing from a broadly flattened optic recess inthe floor of the diencephalon. Extendingdorsad between the optic lobes is a narrowconical projection of diencephalic origin, i.e.,the epiphysis (fig. 16C). The infundibulumprojects posteriorly from a somewhat largerhypothalamic region (figs. 16C, 17A), andjust posterior to the level of the infundibulumis a solid midventral mass of cells growingfrom the superficial ectoderm, i.e., thehypophysis (fig. 17B). The metencephalon isshort and poorly distinguishable from themyelencephalon (fig. 17A, B), the latter being

recognizable by the position of the cranialnerves and the presence of a posterior telachoroidea. The neurocoele is narrow in thespinal cord region and constricted to form adorsal and a ventral canal (fig. 18A).

The retina exhibits three layers (fig. 16B).The outermost is the thin pigment layer. Thesensory layer is divided into an outer nuclearlayer and a paler staining, inner, nerve fiberlayer.

The otocyst is ellipsoidal in shape (fig. 15),with the lateral walls thin and the medianand ventral walls considerably thickened andfused with the fibers of the eighth nerve (fig.17C).

The pharyngeal region possesses four pairsof gill pouches (figs. l5B, 17C). The first isgreatly constricted, and the third and fourthpairs are shallow, dorsolateral evaginations ofthe broad, flat pharynx. The intestinal portalhas moved forward to a position ventral tothe third pair of pouches. Primordia of thefifth pouches are visible in sections as flat,solid outgrowths of the pharynx (fig. 17D).


vascular perc s e r o s a

FIG. 25. Stage 18. A. Cross section at level of telencephalon.B. Cross section at level of lenses. x60.

Posterior to the fifth pouch primordia, theshort esophagus may be identified (fig. 18A).It possesses an irregular lumen lined by asingle layer of cuboidal cells with basophiliccytoplasm. Immediately in front of the liveranlage (figs. 15B, 18C) is a widened region ofthe gut, about 30 micronspossessing a squamous type

in length andof epithelium

arranged in three orThis widened region

four layers (fig. 18B).is presumed to be a

vestigialstages.

stomach, absent in all subsequent

The liver primordium, projecting to theleft side of the intestine with a plexus ofhepatic blood vessels, is a solid mass of un-differentiated cells (figs. 15B, 18C). The swimbladder primordium is not yet visible, but itssite of future development is marked by aplexus of blood vessels on the right side of theintestine at the level of the vestigial stomach(fig. 18B, C).

The intestine is a straight tube, identifiablefrom the level of the liver primordium by itscolumnar, vacuolated epithelium and small


S T A G E 1 8FIG. 26. Stage 18. A. Cross section at level of optic chiasma.

B. Cross section at level of first gill bar. x60.

circular lumen (figs. 15B, 18C, D). The in-testine ends in a slit-shaped anus (fig. 15B).

Twenty-four to 25 myotomes are presentat this stage (fig. 15), and striated musclefibers can easily be identified in the sections.The mandibular arch is large, and the hyoid,third, fourth, and fifth arches are progres-sively smaller. The pectoral fin buds arebroad at the base and have a thin distal

margin (figs. l5A, B, 18A, B, C).The notochord is dorsoventrally flattened,

originating at a level posterior to the otocysts(figs. 17D, 18A-D), extending into the tail,where it becomes more rounded. The poste-rior end is curved dorsad (fig. 15B).

The ventral aorta runs caudad, giving offfive pairs of aortic arches (fig. 14B). Thesecond arch (afferent spiracular artery or

A


S T A G E 18FIG. 27. Stage 18. A. Cross section at level of hypophysis.

B. Cross section at level of third gill bar. x60.

hyoidean artery) is thin and slender. Thethird arch is the largest, and the fourth, fifth,and sixth are progressively smaller, with thesixth arch appearing as a branch of thefourth (fig. 14B). A hepatic plexus is formedaround the liver primordium, drained by ashort hepatic vein, which, in turn, emptiesinto the left common cardinal vein.

The urinary vesicle (formed from the post-anal gut) has invaginated more deeply thanin stage 13 and possesses a slit-shaped open-

ing (fig. 15B). The nephric ducts becomeevident at the level of the esophagus, possessa lumen, and extend caudad to about thesixth or seventh somite, becoming solid atabout the fourth. No tubules are present atthis stage.

STAGE 15Figures 19, 20

Age, 5.5 days.Standard length, 2.52 mm.


FIG. 28. Stage 18. A-B. Cross sectionsat level of mesonephroi. x60.

Length of caudal fin bud, 0.09mm. greatly widened and thickened, forcing theMesencephalon width, 0.68 mm. eyes into an almost frontal position, andThis stage is characterized by the appear- these lobes overlie the metencephalon and

ante of melanophores other than retinal. The diencephalon (fig. 19A, B). The choroidretinal melanophores are obscured by pig- fissure of the retina is distinct and possessesment in the choroid layer, the latter contain- numerous fibers of the optic nerve (fig. 20B).ing numerous iridiodytes in addition to The hypothalamic region of the diencephalonmelanophores (fig. 20A). About 10 to 20 has enlarged so that it partially overlies thesmall, stellate epineural melanophores ap- hypophysis and has pushed the infundibulumpear in the dorsal region above the meten- into a slightly more posterior position. Threecephalon and myelencephalon (fig. 20A). bluntly rounded lobes can be distinguished as

The optic lobes of the mesencephalon are constituting the otocyst (fig. 19A, B): a

>: 1949 TAVULGA: EMBRYONIC DEVELOPMENT IN FISH 195

left duct of C u v i e r

a c c e s s o r y m e s o n e p h r i c duct

/

FIG. 29. Stage 18. A. Cross section at level of ducts of Cuvier.B. Cross section at level of esophagus. C. Cross section at level ofgall bladder. x60.

ventral, saccular portion ; an anterior, medialportion possessing a single large otolith; and aposterior portion with a single small otolith.

Lateral extensions of the mandibular andhyoid arches forming opercula are presenthere. The opercula are short and blunt,concealing only the first of four pairs of gillarches. Twenty-five somites are present, this

number being only one fewer than the fullcomplement (fig. 19). A caudal fin bud ispresent (fig. 19B).

The esophagus is short and distinct. Thegastric region, described for stage 14, is ab-sent. From the anteriormost intestinal re-gion three primordia may be traced (fig.20B). To the right is the short, tubular


FIG. 30. Stage 18. A. Cross section at level of spleen.B-C. Cross sections at midbody level. x60.

primordium of the swim bladder, surrounded of the sixth somite (figs. 19B, 20B).by a plexus of blood vessels. Just posterior to The most important development in thethe above and also extending to the right is a circul;atory. system is in the region of thesolid endodermal outgrowth in which pan-creatic cells and ductules c a n b e dis-

ventral aorta. ‘At this stage, the ventral aortais split into two vessels (fig. 14C). The

tinguished. At the same level, to the left of anterior of these drains a pair of small, effer-the intestine, is the liver (figs. 19B, 20B). The ent, spiracular arteries (first aortic arches),intestine is somewhat S-shaped in ventral and supplies a pair of small efferent spiracu-view and ends in the anus at about the level lars (second aortic arches), and the third and

1949 TAVOLGA: EMBRYONIC DEVELOPMENT IN FISH

cross-wnnection b e t w e e n

caudai ve ins

197

. .FIG. 31. Stage 18. A-C. Cross sections at posterior body level.

D-G. Cross sections at post-anal leveis. x60.

fourth pairs of aortic arches. The posteriorventral aorta is connected to the commonbase of the fifth and sixth aortic arches andalso supplies the fourth arch in part (fig. 14C).

At a level just anterior to that of the pec-toral fin buds, a few mesonephric tubuleshave developed in the mesomere regiondorsal to the mesonephric duct. The meso-nephric ducts possess short extensions 50 to

100 microns anterior to the above region.These extensions may be vestigial homo-logues of the pronephric ducts. Posteriorly,the mesonephric duct can be traced up to thebase of the urinary vesicle. The latter struc-ture has invaginated considerably and lies ina position dorsal to the anal portion of theintestine, its distal (or cephalic) end beingslightly bilobed (figs. 19B, 20B).

198 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY

S T A G E 1 8

FIG. 32. Stage 18. Ventral view, showing viscera and gills. X60.

STAGE 16Figure 21

Age, 6.9 days.Standard length, 2.80 mm.Length of caudal fin, 0.12 mm.Mesencephalon width, 0.79 mm.The small, stellate, epineural melanophores

are more numerous and scattered over thebrain region, extending along the dorsal mid-line slightly beyond the pectoral fin buds.

The opercula are larger, covering the firstthree pairs of gill arches (fig. 21 B).

The full complement of 26 myomeres ispresent (fig. 2lA, B). Beyond the lastmyomere and the upturned tip of the noto-chord is the compressed caudal fin bud withprimordia of six fin rays (fig. 21 B). An analfin primordium is present (fig. 21B).

The intestine is S-shaped in ventral view.At the anterior end of the intestine, nearwhere it gradates from the esophagus, is theliver, connected to the left side of the in-testine by means of a bile duct (fig. 21B). Atthe same level is the pancreas, extending


FIG. 33. Stage 18. A. Ventral view of arterial system. B. Lateral view of arterial system. x60.

ventrally and to the right of the intestine in aseries of nests of cells embedded in the mesen-tery. The anteriormost end of the pancreasconsists of a more compact lobe of relativelyundifferentiated tissue. Connected to theright side of the intestine, at this level, is theductus pneumaticus, characterized by acolumnar epithelium and a fine circularlumen. The swim bladder is small, about thewidth of the intestine, flattened, and in aposition dorsal to the intestine. Its epitheliumis high columnar and highly vacuolated. Theswim bladder extends retroperitoneally andcaudad to about the level of the first myomere(fig. 21B). This structure is described ingreater detail in stage 18.

The gonadial ridges originate near the baseof the swim bladder and are attached to the

dorsal peritoneum. They extend caudad tothe level of the fourth or fifth myomeres. Themesonephric ducts possess a lumen through-out their length and empty into the base ofthe urinary vesicle. Mesonephric tubules arenumerous at the level of the pectoral fin buds,and a few renal capsules and glomeruli arevisible.

The ventral aorta is still double, but theanterior branch is reduced in size. A fine,unpaired external carotid (or lingual) arterybranches off the base of the anterior branchof the ventral aorta (fig. 14D).

STAGE 17Figures 22, 23



--S T A G E 18

:FIG. 34. Stage 18. A. Ventral view of venous system. B. Lateral view of venous system. x60.

Length of caudal fin, 0.23 mm.Mesencephalon width, 0.84 mm.Small, stellate epineural and cutaneous

melanophores are thinly scattered over theentire head and along the middorsal region ofthe body and tail. The melanophores over themesencephalon are somewhat larger. Fine,dot-like melanophores appear in the lateralperitoneal lining (fig. 23A).

Both a dorsal and an anal fin bud arepresent, and the caudal fin possesses at least10 rays (fig. 22B). The operculum, arisingfrom the mandibular and hyoid gill arches,covers the four remaining gill arches (fig.23B). Posterior to the last branchial arch(sixth visceral arch), the pharynx possesses a

pair of shallow lateral evaginations. Thelatter are the vestigial seventh visceralpouches.

At the entrance of the ductus choledochusinto the intestine, a ventral saccular outgrowth representing the gall bladder ispresent (fig. 23B). The diffuse pancreas alsooriginates at this point and extends caudad,scattered in small lobules in the mesenteryand the enteric serosa. The compact cephaliclobe of the pancreas possesses no crypts orducts, and its cells are strongly basophilic(fig. 23B). On the basis of its origin, position,and morphology, it is possible that it repre-sents the homologue of islets of Langerhanstissue of the pancreas. The spleen, a highly



vascular lymphoid structure, is present atthis level to the right of the pancreas (fig.23B).

The swim bladder has enlarged laterally(to the right side) and posteriorly (figs. 22B,23B). This new portion, however, possessesa squamous epithelium, in contrast to thebasal and left regions which possess avacuolated columnar epithelium as describedfor stage 16.

The anterior branch of the ventral aortahas completely degenerated (fig. 14E). Theposterior branch curves dorsally to supplythe sixth, fifth, fourth, and third aorticarches in that order. The hyoidean (second)and efferent spiracular (first) arteries aresmaller. The ventral aorta continues craniadas an unpaired external carotid artery (fig.14E).


Age, 8.4 days.Standard length, 3.57 mm.Length of caudal fin, 0.30 mm.Mesencephalon width, 0.89 mm.Yolk depth, 1.45 mm.The anatomy of this stage is described

more fully than that of other stages for thereason that, in terms of organogenesis, de-velopment here may be considered com-pleted. Descriptions of subsequent stages willbe limited, for the most part, to gross morphology and measurements.

A cephalic flexure is present here, with thetelencephalon underlying the diencephalon inpart (fig. 24A). The head is tightly enclosedwithin the pericardial extra-embryonic mem-branes (figs. 24B, 25A, B). The muscular and


Fig. 36. Stage 20. A. Dorsal view. B. Lateral view. (Labels as in figure 35.) X30.

skeletal systems are well developed, andembryos of this stage, when removed fromthe ovarian follicles, swim about rapidly andwill survive in conditioned aquarium water.

The telencephalon is short, and the lateralvesicles are solid and poorly distinguishable.The olfactory bulbs are solid masses of tissueattached lateroventrally to the telencephalon(fig. 25A). The diencephalon underlies theoptic lobes of the mesencephalon (fig. 25B).At the level of the optic chiasma, a smallmiddorsal extension of the diencephalonprojects through the cleft between the opticlobes, i.e., the epiphysis (fig. 26A). The largehypothalamus extends ventrad and caudadto the level of the medulla (fig. 26B). Themetencephalon is short, overlain in part bythe optic lobes and in part by the myelen-cephalon.

Three layers of the sensory layer of theretina are visible, the innermost fibrous layer,a thick ganglionic layer, and the outermostnuclear layer (fig. 25B). A thin retinal pig-ment layer coats the sensory layer (fig. 25B).The choroid coat of the eye consists only of a

pigment layer one or two cells in thickness(fig. 26B). The optic nerve leaves the retinafrom the posterior ventral quadrantand runsmesiad to form the optic chiasma (fig. 26A).

The otocyst consists of a large anteriorvertical canal primordium, and smallerlateral and posterior vertical primordia. Aprominent ventral saccular region and adorsal medial saccus endolymphaticus arepresent (figs. 24A, 26A, B, 27A, B).

The mandibular arch possesses a set ofsmall cartilages, most prominent of whichare a pair of lateral primordia of the palato-quadrates with ventral extensions, i.e.,Meckel’s cartilages (fig. 25B). The hyoidarch and in part the mandibular form theoperculum enclosing four pairs of branchialarches (i.e., the third, fourth, fifth, and sixthvisceral arches; figs. 24B, 27A, B, 28A, 32).The opercula are separate at the midventralline from the level of the fifth visceral arch,where the ventral aorta enters the base of thegill arches. Each of the branchial arches issupplied by a large, unbranched aortic arch,and blunt, rounded gill filament primordia


FIG. 37. Stage 20. A. Dorsal view, showing distribution of melanophores.B. Ventral view, showing viscera. x30.

are present on the external surfaces of the gillbars (fig. 27A, B). In the posterior region ofthe pharynx, the roof, at the level of thethird gill bar (fig. 27B), and the floor, at thelevel of the fourth gill and back (fig. 28A),possess rows of enamel organs which willform the pharyngeal teeth. Posterior to thefourth gill bar (sixth visceral arch), thepharynx possesses a pair of lateral evagina-tions which are presumed to be vestigialseventh visceral pouches (fig. 28B).

The esophagus is short, lined by cuboidalcells. This lining changes abruptly into avacuolated columnar type at the level of theductus pneumaticus (fig. 29B, C). The latteris a narrow duct with a fine lumen (fig. 30A).The basal portion of the swim bladder, wherethe duct enters, is lined by a high columnarepithelium which is greatly vacuolated andglandular in appearance (fig. 3OA, B). The

remainder of the swim bladder, extending tothe right side and caudad almost as far as theanus, is retroperitoneal and possesses asquamous lining (figs. 3OC, 3lA, 32).

The gall bladder projects ventrad andcraniad in the mesentery mesiad to the liver(figs. 29C, 32). The exocrine portion of thepancreas may be found in irregular massesand lobules throughout the mesentery, scat-tered among the coils of the intestine, andembedded in the surface of the liver (figs.29B, C, 30B, C, 32). The compact lobe of thepancreas is located in the mesentery at thelevel of the spleen, situated between thespleen and the liver (figs. 30A, 32). The cellsof this lobe are tightly packed and possesslarge, darkly staining nuclei, basophilic cyto-plasm, and small, slightly basophilic granules.

The intestine, with a small circular lumenand tall columnar epithelium, possesses a



single loop extending craniad, ventrad, andsomewhat to the right of the midline (figs.29B, C, 30A, C, 31A-C, 32). The characterof the epithelium does not change throughoutits length. The anus, at the level of the sixthmyomere, has a longitudinally oriented, slit-shaped opening (figs. 31 C, 32).

The anterior tip of the notochord issituated at a level just posterior to theauditory vesicle. The notochord is laterallycompressed throughout its length (figs. 28A,30C). Neurocranial cartilages are forming inthe region ventral to the hind brain. Cartilagepartially envelops the otocyst. Sclerotomalcartilages extend laterad from the notochord,forming the initial stages in the developmentof neural arches. The bases of the pectoralfins also contain cartilaginous plates.

The heart extends forward within thepericardial sac located under the head (figs.24B, 25-27). The elongate sinus venosusdrains the blood vessels of the yolk sac andthe pericardial serosa. The atria1 region ispoorly distinguished from the sinus. Theventricle protrudes to the right of the sinu-

atrium. The conus arteriosus region is notdistinguishable, and the long ventral aortacurves dorsally to enter the base of the gillarches at the level of the fourth branchial arch(sixth visceral arch). The ventral aorta turnsanteriorly here and gives off, in the followingorder, the common bases for the fifth andsixth aortic arches, the fourth arch, the thirdarch, and the smaller second and first arches(fig. 33A, B). The median continuation ofthe ventral aorta forms the external carotid.The internal carotids originate from thethird aortie arches. The third and fourthaortic arches enter the dorsal aorta separatelyfrom the fifth and sixth arches (fig. 33A, B).At this point, the mesonephroi surround thedorsal aorta, and numerous renal arteries aregiven off (fig. 28B). Caudad, a pair of largesubclavian arteries are present, and a single,ventral coeliac artery (figs. 28B, 33A, B). Thelatter is relatively short, ramifying out tosupply the viscera. The dorsal aorta con-tinues caudad to become the caudal artery(figs. 31A, D, F, 33A, B).

The caudal veins drain the tail region (figs.


FIG. 39. A. Stage 22, lateral view. B. Stage 23,lateral view. X15.

31D, E, F, G, 34A, B). The dorsal caudalvein runs along the dorsal aorta and incontact with it. The ventral caudal veinpasses through the anal fin primordium (fig.31G). The ventral caudal curves aroundcraniad to the urinary bladder and runsventrad to the right side of the intestine toform the vitello-caudal vein (figs. 3lA-D,34A, B). The long, fine, posterior intestinalvein drains into this vessel (fig. 34B). Thisvessel eventually becomes part of the hepaticportal system, but retains its connection withthe ventral caudal vein. Three vertical crossconnections are present between the dorsaland ventral caudal veins. The anteriormost isat the level of the urinary pore (figs. 31D,34B), the second at the level of the anal fin(fig. 31F), and the third at the base of thecaudal fin. The dorsal caudal vein splitsanteriorly to form two postcardinals (fig.30B), the left one being much the smaller ofthe two (fig. 34A, B). The dorsal caudal andpostcardinal veins actually form the renal

portal system, draining into the numerousrenal vessels. The subclavian veins and theanterior cardinal branches also drain into therenal plexus. Two large ducts of Cuvier col-lect the blood from the renal plexus (fig. 34A,B). A yolk sac portal sinus forms a ringaround the trunk of the embryo, extendingover the head as the posterior margin of thepericardial serosa. The vessel drains thevitellocaudal vein at its anal margin (figs.31A, B, 34B). Laterally it drains the twoducts of Cuvier and a single hepatic veinentering the portal sinus on the left side(figs. 29A, B, 34A, B).

The hepatic portal vein forms a large sinusin the mesentery at the level of the spleen(figs. 30A, 34A, B).

The mesonephroi are limited to regionslateral to the notochord at the level of thebase of the pectoral fin buds (fig. 28A, B).They are highly vascularized, partially en-veloping the fork of the dorsal aorta. Themesonephric ducts become distinguishable


STAGE 26

FIG. 40. A. Stage 24, lateral view. B. Stage 25, lateral view: C. Stage 26,lateral view. X15.

near the caudal margins of the mesonephroi(fig. 29B). Here a small dorsal branch isgiven off by each of the ducts (fig. 29B, C).These accessory Wolffian ducts run caudad,closely appressed to the sides of the noto-chord. They possess a lumen, and an epithe-lium similar to that of the definitive ducts,and extend posteriorly to the level of thebase of the swim bladder, where they endblindly.

The mesonephric (Wolffian) ducts con-tinue posteriorly, curving ventrad in thepostanal region of the coelom, and emptyinto the base of a somewhat bilobed urinarybladder. The elongate urinary pore is situatedposterior to the anus (fig. 31A, D, E).

The gonadial ridges become evident at-tached to the peritoneum undercoating themembranous portion of the swim bladder(fig. 30B, C). They extend caudad as far backas does the swim bladder (fig. 3lA, B). Noevidence of sexual dimorphism in the gonadsis visible at this stage.

The two lobes of the pituitary gland arecompletely separate. The infundibulum hasmoved back to a position posterior to thehypophysis and is poorly distinguishablefrom the hypothalamus. The hypophysis isdetached from the stomodeal ectoderm andexhibits no differentiation into transitional,intermediate,, or anterior lobes (fig. 27A). Aplexus of blood vessels derived from the


“circle of Willis,” described for stage 12,supplies the pituitary.

The epiphysis is small, saccate, and lodgedbetween the optic lobes, and the connectionwith the diencephalon is narrowed downconsiderably (fig. 26A).

The thyroid makes its first appearance inthis stage or in stage 17. It is present as adiffuse group of small follicles around theventral aorta at the level of the third andfourth aortic arches (figs. 26B, 27A). In laterstages it becomes increasingly difficult tolocate, as it migrates caudad in individuallobules or follicles.

STAGE 19Figure 35

Age, 9.6 days.Standard length, 3.68 mm.Length of caudal fin, 0.32 mm.Mesencephalon width, 1.OO mm.Depth of yolk, 1.42 mm.This stage is characterized by the appear-

ance of fin rays in the pectoral fin buds. Thepericardial extra-embryonic membranes be-gin to split here in the region in front of andbetween the eyes. This is the first step in theregression of these membranes.


Age, 10.9 days.Standard length, 3.87 mm.Length of caudal fin, 0.47 mm.Mesencephalon width, 1.03 mm.Depth of yolk, 1.39 mm.The anterior margin of the pericardial

serosa forms a small circle with the mouthand the front of the head protruding (fig.36A, B). The number of melanophores hasincreased considerably, accompanied by thefirst appearance of corolla-type melanophoreson the head (fig. 37A). Since these werepresent in all the strains of platyfish, sword-tails, and hybrids studied, they may beidentified as the micromelanophores.

The intestine possesses a single completecoil. The urinary bladder is large, bilobed,and opens to the exterior by a slit-shapedurinary pore. The mesonephric ducts com-municate with the bladder at its mostposterior margin, near the urinary pore (fig.37B).

STAGE 21Figure 38

Age, 11.5 days.Standard length, 3.95 mm.Length of caudal fin, 0.63 mm.Mesencephalon width, 1.03 mm.Depth of yolk, 1.33 mm.Internally, there is little difference be-

tween this stage and the previous one. Thehead protrudes out of the extra-embryonicmembranes so that the anterior margin of thepericardial serosa encircles the head at thelevel of the lens.

STAGE 22Figure 39A

Age, 13.4 days.Standard length, 4.00 mm.Length of caudal fin, 0.79 mm.Mesencephalon width, 1.05 mm.Depth of yolk, 1.19 mm.The yolk mass is further reduced, and

there is a concomitant regression of the extra-embryonic membranes to form a “neckstrap,” covering the posterior one-third of theeye and extending posteriorly to cover mostof the mesencephalic region. Fin rays in theanal fin make their first appearance here.

STAGE 23Figure 39B

Age, 15.7 days.Standard length, 4.18 mm.Length of caudal fin, 0.92 mm.Mesencephalon width, 1.08 mm.Depth of yolk, 1.14 mm.The “neck strap” and yolk are reduced,

and the former is about one-half the widthof the eye and situated posterior to it. The finrays are more distinct in the anal fin, andrays can be distinguished in the dorsal finbud. Corolla-type melanophores are scat-tered thickly in the epidermis and dermis.Meningeal melanophores are smaller andstellate. Punctate melanophores are mostfrequent in the peritoneum and notochordalsheath, forming almost solid black sheets inthe dorsal coelomic region.

STAGE 24Figure 40A


208 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY

Length of caudal fin, 1.05 mm. STAGE 26Mesencephalon width, 1.11 mm. Figure 40CDepth of yolk, 0.95 mm.The “neck strap” is narrow, less than one-

Age, 21.9 days.

quarter the width of the eye, and it is some-times broken in the middorsal region. Thecephalic flexure is slightly straightened, withthe tip of the lower jaw moving into a moredorsal position.

Standard length, 5.69 mm.Length of caudal fin, 1.76 mm.Mesencephalon width, 1.18 mm.Birth usually takes place at this stage, as

STAGE 25Figure 40B

Age, 19.7 days.Standard length, 5.00 mm.Length of caudal fin, 1.34 mm.Mesencephalon width, 1.15 mm.Depth of yolk, 0.52 mm.Remnants of the “neck strap” are still

visible, the yolk mass is greatly reduced andflattened, and the cephalic flexure is almostcompletely absent.

the embryos rupture the follicle walls andmove into the ovarian cavity and into thegonoduct. The remaining yolk is completelyenclosed within the abdominal cavity, but isstill visible through the thin belly epidermis.The heart has now assumed its definitiveposition, with the sinu-atrium situated poste-rior to the conus and ventral aorta. Growthafter birth is rapid, and the larval fish in-creases to almost 7 mm. within 24 hours afterparturition. Macromelanophores in the wildstock of platyfish do not appear until one ormore days after birth, so that only micro-melanophores are present at stage 26.

VOL. 94

REPRODUCTIVE CYCLE AND EMBRYONIC GROWTH IN PLATYPOECILUS

Normal Platypoecilus females, having at-tained sexual maturity (four to six months ofage), begin producing broods at fairly regularintervals. The intervals between the firstfew broods may be irregular and vary be-tween 26 and 90 days, but once sexual ma-turity is reached, if environmental conditionsremain unchanged, the intervals betweenbroods vary little from the characteristic 28days. In young and very old females, the in-tervals are less regular and the numbers ofyoung per brood are reduced. The averagenumber of young per brood for a female inher prime is 30 to 40 and is often over 60.The reproductive cycle of P. maculatus issimilar to that described for X. hellerii(Bailey, 1933), Lebistes reticulatus (Purser,1938), and Gambusia patruelis (Ryder, 1882).

Insemination of a female is effected bymeans of the modified anal fin of the male.This gonopodium is a sex-hormone influencedstructure, as has been shown by Cohen(1946) and M. C. Tavolga (in press). Fe-males once inseminated will produce as manyas four or more broods at twenty-eight-day

intervals without further contact with males.There is evidence that sperm are stored bythe female in the folds of the oviduct forperiods up to seven months (Gerschler, 1914,Winge, 1922, for Lebistes; van Oordt, 1929,for Xiphophorus), and whenever successivecomplements of ova mature there are spermavailable.

The latter fact gives rise to considerablevariation in the ages of members of the sameembryonic brood. Hopper (1943) describedthe existence of a seven-day interval betweenthe birth of one brood and the fertilization ofthe next complement of eggs. During thisinterval, yolk is deposited and maturationof the egg takes place. Apparently all the ovaof a given complement do not mature at thesame rate, and therefore fertilization and thestart of development are staggered over aperiod of 24 to 48 hours. Consequently, theages of embryos within a single ovarian sacmay vary considerably during early develop-ment. Since development is slower during thelatter portion of the gestation period, theinitial differences become insignificant, and

/

.’


parturition of an entire brood is accomplishedover a period of an hour or two.

Fertilization of the ova takes place througha follicular canal leading from the ovariancavity to each egg. This structure was de-scribed by Bailey (1933) for Xiphophorusand by Purser (1938) for Lebistes. Thefertilized ova are retained within the follicles,and all embryonic development takes placethere. The follicle ruptures just beforeparturition, and the larval fishes find theirway from the ovarian cavity through the

_ oviduct and out through the urinogenitalpore. Towards the end of the gestation period,no young embryos of the next brood arepresent, nor are there any mature ova. Im-mediately after birth of a brood, all the ovaare small and contain little yolk. There is noevidence of superfoetation in this group ofxiphophorin fishes, although superfoetation iscommon in related poeciliids (Scrimshaw,1944a, 1944b; Turner, 1937,1940).

It was recognized that because of theviviparous type of reproduction of the platy-fish, the determination of the intra-ovariangrowth rate would be difficult. An attempt tomake this determination was made byTavolga and Rugh (1947), but the quantity

fof material was insufficient.The following terms were introduced by

the above authors and are retained: theoreticalage, the value, in days, determined from thedate the previous brood was born, less sevendays; morphological age, given in terms ofstage numbers, established for each embryoby comparison with the graded series;chronological age, a calculated value repre-senting the actual developmental time foreach stage. It should be noted that the seven-day interval between birth of a brood andfertilization of the next brood is only anaverage value, since maturation and fertiliza-tion of a complement of ova may be spreadover a period of up to 48 hours (Hopper,1943). Actually the calculated mean value forthe time of fertilization is 6.8 days, as will beshown later.

In a given embryonic brood, the frequencydistribution of morphological ages approachesa normal curve of probability. Consequently,mean values of morphological ages for agiven theoretical age would help to determinethe chronological age. In graphic construc-

tion, plotting the arrays and means of theoret-ical age against morphological age provedto be more effective. Such a graph is shown infigure 41, based upon examination of 2425P. maculatus embryos. In this graph, the ex-tremes for each morphological age group areshown, and it is obvious that there would beconsiderable difficulty in attempting to pre-

5 10 15 20 25

MOROPHOLOGICAL STAGE NUMBERS

FIG. 41. Graph showing correlation betweenmorphological age and theoretical age of Platy-poecilus embryos. Vertical bars represent extremerange with crosses marking the mean age for eachmorphological stage. (See text.)

diet what stage of development would bepresent in the gravid female at any given dayof its cycle. Despite the large variation, thecorrelation is high, with the correlation co-efficient r being equal to .89 and the curvi-linear correlation coefficient (correlation ratio,eta) being equal to .91 (calculations accordingto Simpson and Roe, 1939).

The regression curve resulting from theplotting of data given in table 2 gives areliable estimate of the time lapse betweenany two embryonic stages.

The above values were used in plotting thegrowth curves (fig. 42). The time abscissa,in days, represents the mean values takenfrom the graph in figure 41, after subtracting


StageNos.

123456789

1011121314151617181920212223242526

TABLE 2

Number of Days After Birth ofPrevious Brood I

Means

4.87.17.376.7.78.58.79.19.8

10.210.711.01 l . 611.812.313.714.515.216.417.718.320.222.5241.026.528.7

StandardError

kO.21 0.20 0.20 0.18 0.21 0.21 0.20 0.22f0.24rtrO.23 0.22rto.21 0.25kO.24+0.22f0.23

0.17f0.19kO.17fO.19~0.16kO.1750.14kO.13

0.1250.09

Extremes

4-85 - 95 - 96 - 96-106-116-117-137-157-157-157-147-168-178-179-19

11-181 l-2012-2214-2315-2315-2515-2818-2921-3125-31

Mean TheoreticalAge (in Days)

0 . 00.30 . 50.80 . 91.71.92.33.03.43 . 94.24.85 . 05 . 56 . 97.78.49.6

10.911.513.415.717.219.721.9

the seven-day interval so that fertilizationtime is zero. Actually, the more accurate 6.8-day value was subtracted, since this was thecalculated average time of fertilization afterthe birth of a previous brood.

Three curves are plotted against the sameabscissa in figure 42. The standard length ofthe embryos is shown on the left ordinate,and the width of the mesencephalon anddepth of yolk are plotted against the rightordinate. The data for the graph in figure 42are given in the descriptions of the individualstages.

The standard length measurements weremade with the aid of a camera lucida, and instages where a cephalic flexure was present,the length was taken along the curvature ofthe body axis. Determinations were begunwith stage 5, at which time a measurable

germinal shield first appears. Up to stage 15,standard length is equivalent to total length,but in subsequent stages the caudal fin be-gins to form and is not included in this seriesof measurements.

Head width is given in terms of the maxi-mum width of the mid-brain. These measure-ments were begun at stage 9, at which stage amesencephalon first became distinguishable.This was considered a more accurate deter-mination of growth in width than the distanceacross the eyes, since the latter are pushedlaterad and craniad by the growth of theoptic lobes.

The diameter of the yolk remains sub-stantially constant up to about stage 17,when utilization of yolk material is appar-ently accelerated. The longitudinal axis ofthe yolk, i.e., the one parallel to the long


axis of the embryo, changes little, even bystage 25 when most of the yolk is gone. Thedepth of the yolk, however, is a measurementmore accurately expressing the change inyolk volume. This distance is taken from thepoint of entry of the duct of Cuvier into theportal system directly ventrad, perpendicu-lar to the long axis of the embryo. In stagesbefore the formation of any circulatory sys-tem, the yolk mass is spherical so that all axeswould be equal in length.

The growth curves for both standard lengthand head width exhibit a roughly sigmoidshape, with a considerable reduction in slopeat about the eighth and ninth day, with somerecovery after the fourteenth day (fig. 42).This change in slope corresponds in time toabout stage 18 in terms of morphologicalage. When these growth curves are comparedto a curve showing the change in the size ofthe yolk mass (fig. 42), it is seen that thesharp reduction in yolk volume begins atstage 18 (eighth day of development). Thisstage also marks the most extensive develop-ment of the vascularized extra-embryonicmembranes, and in subsequent stages thesestructures regress rapidly. A temporary

“neck strap” structure, such as was de-scribed by Turner (1940), is formed duringthis regression.

6.0|

EMBRYONIC AGE IN DAYS

FIG. 42. Graph showing growth rates of Platy-Poecilus embryos in terms of standard length(left ordinate) and width of mesencephalon (rightordinate). The regression of the yolk mass is _plotted against the right ordinate. The time scale(abscissa) was determined from data in figure 41.

>

DISCUSSION OF ORGANOGENESIS IN PLATYPOECILUS

EXTRA-EMBRYONIC MEMBRANES

The development of the pericardial amnionand the pericardial serosa, as described hereand by Tavolga and Rugh (1947) for Platy-poecilus, exhibits a remarkable similarity tothe development of the amnion and serosaof avian embryos. Three important differ-ences should be noted. One such diff’erencelies in the initial formation of these mem-branes from a single layered periblastic ecto-

i derm in the platyfish (stage 7), with a laterinvasion by mesenchyme (stage 11) to form atypical double-layered sac, i.e., an inneramnion and an outer serosa, both somato-pleural. Second, there is the presence of ahighly developed vascular plexus in theplatyfish serosa. Third, the membranes en-fold only the head region, and no caudalamniotic fold is present, as in avian embryos.

lematical, but it has been proposed that theypossess an accessory respiratory function byTurner (1940), whose descriptions for Platy-poecilus covered only late “neck strap”stages. Tavolga and Rugh (1947) suspectedthat the pericardial serosa has a nutritive, aswell as a respiratory and excretory function.Support to this view is given here, since it hasbeen shown that the regression of the mem-branes occurs concomitantly with an ac-celeration in the utilization of embryonicyolk. In addition, the work o f Scrimshaw(1944b, 1945) demonstrated that closelyrelated species of Poeciliidae, including -Xiphophorus hellerii, are truly viviparous and -gain nourishment, probably as amino acidsand carbohydrates, from maternal tissue,thus giving support to the contention that thevascular pericardial serosa is a placentalorgan.

The function of these membranes is prob- If the above is assumed to be true, the

-

-- -


simultaneous drop in the slope of the growthcurve leads one to suspect that this is arather critical period of development duringwhich a new set of enzyme systems is broughtinto play and the entire nutritive pattern ofthe embryo is altered. Investigations concern-ing any possible physical and chemicalchanges in the yolk and in the nutritive re-quirements of the embryo have been planned,and results will be reported at a later date.

The nomenclature of the extra-embryonicmembranes in the platyfish is that of Tavolgaand Rugh (1947), rather than that of Turner(1940), for this and related species. Theterm pericardial amnion is used for the inner

. membrane, which is non-vascular and closelyenvelops the head. The outer pericardialserosa is thicker and highly vascularized. Thecavity between the two membranes is con-fluent with the pericardial cavity and istermed the extra-embryonic pericardium.These terms are used in preference toTurner’s “pseudoamnion” and “pseudo-chorion” (1940).

isA similar, but less extensive, pericardiumpresent in the oviparous relatives of the

xiphophorins. In Fundulus (Oppenheimer,1937) and in Oryzias (Kamito, 1928; Rugh,1941) the pericardial sac is expanded ante-riorly, and the head is flexed and partiallyenveloped. Here, too, the pericardial mem-branes are vascularized and may serve as anaccessory respiratory organ.

The origin of the syncytial membrane(periblast) adhering to the yolk was ac-curately described by Wilson (1889), asderived from the peripheral zone of junctionof the early blastula. Both Wilson and Price(1934a) depicted the periblast as a thinmembrane, of little morphogenetic signifi-cance, covering the entire yolk and extendingunder the embryonic region.

In this report the distinction between theperiblast and the closely applied extra-embryonic ectoderm was not possible. In thestage descriptions, therefore, the term peri-blastic ectoderm was used to include theentire thin extra-embryonic tissue, and it isthis tissue, in front of the head, which givesrise to the primordium of the extra-embryonicpericardial membranes. The extent to whichthe syncytial periblast plays a role in the

formation of the definitive extra-embryonicmembranes is not known.

CLEAVAGE AND GASTRULATIONPlatypoecilus possesses a relatively small

blastodisc on the animal pole and cleavagecells that are broad and flat without protrud-ing beyond the circumference of the yolksphere. The more typical situation, withlarge, globular blastomeres, is that describedfor oviparous fishes such as Fundulus (Oppen-heimer, 1937), Curegonus (Price, 1934a),Serranus (Wilson, 1889), Carassius (Battle,1940), and others.

The late blastula (stage 4) begins to spreadover the surface of the yolk as a result of theproliferative activity of the peripheral germring. As one segment of the germ ring differ-entiates to form the germinal shield, the re-mainder of the germ ring migrates over thesurface of the yolk, passing through a “yolkplug” stage, and eventually terminating in asmall, transient blastopore. In Platypoecilus,the closure of the blastopore is accomplishedafter stage 11, with the lateral lips closinginto a temporary lineolate raphe (evidentlyhomologous to the primitive streak of avianembryos).

The posterior margin of the embryonicshield is considered homologous to the dorsallip of the blastopore. It is a region of highlyactive, proliferating, undifferentiated tissue.One of the early descriptions of this structure(Agassiz and Whitman, 1885) called atten-tion to its probable functional significance,but the potencies of this region were notdemonstrated until mapping and graftingtechniques came into use. The work ofPasteels on Salmo (1933, 1934, 1936) and thevital staining and transplanting techniquesof Oppenheimer on Perca (1934a, 1936b),Fundulus (1934b, 1935, 1936a, 1936b, 1936c,1947), and Epiplatys (1938) showed that theposterior margin of the embryonic shield ofteleosts has potencies in common with thoseof the amphibian dorsal lip organizer. Littleis actually known of the mechanics of endo-derm and notochord formation, and whethergastrulation takes place by an involutorymovement as described by Wilson (1889) orsimply by a cell proliferation at the dorsall ip .


BLASTOPORE CLOSURE

The typical situation in teleost fishes issimilar to that of the Amphibia, in that theclosure of the blastopore is accomplished by aconcrescence of the blastoporal lips and aneventual inclusion of the blastopore into theneurocoele during neurulation. Since neurula-tion in the teleosts takes place by an in-vagination of a solid neural keel, a neuren-teric canal, as such, is never present. Thissequence of events, followed by tail budformation, has been described for teleosts ingeneral by Morgan (1893, 1895) and Brachet(1921), and specifically for Coregonus (Price,1934a), Fundulus (Oppenheimer, 1937), Ser-ranus (Wilson, 1889), and numerous otherforms.

The closure of the blastopore in Platy-poecilus occurs considerably later than the de-velopment of a tail bud. At stage 12, theblastopore lips concresce into a lineolateraphe (or primitive streak) somewhat behindand under the well-developed tail. The onlyother report of this phenomenon in a teleostis that of Clapp (1891) on the toad-fish(Batrachus tau), and Balfour (1878) de-scribed the post-caudal blastopore as com-mon in the elasmobranchs.

NOTOCHORD FORMATION

The notochord in the platyfish is describedhere as originating from a longitudinal splitin the chorda-endoderm, a tongue of tissuearising from the dorsal lip and growingcraniad along the ventral midline (stages 7and 8).

The simultaneous origin of notochord andendoderm is apparently typical of theteleost fishes. Wilson (1889) described theformation of two lateral endodermal wingswith a median notochordal primordiumunder which the endoderm later fuses into asingle sheet. Price (1934a) reported es-sentially the same for the whitefish, with theadditional observation that the endoderm isinitially a median cord of cells which laterflattens out into the two lateral sheets andthe median notochord. The derivation of thenotochord in Platypoecilus by delamination issimilar to that described for the trout byHenneguy (1888).

ECTODERMAL DERIVATIVES

The early development of the nerve cordis similar to that described by Wilson (1889)for the sea bass (Sewanus). The formation ofa solid neural keel is characteristic of teleostfishes. The mesencephalon is first to differ-entiate and remains the broadest part of thehead in the form of a pair of optic lobes, asin most other teleosts. The telencephalon issmall and scarcely distinguishable from thediencephalon at all stages. The hypothalamicregion of the diencephalon is the largest,beginning to enlarge as early as stage 12. Inlater stages it projects ventrad and caudad,pushing the infundibulum into a positionposterior to the hypophysis. The epithalamuspossesses a tela choroidea after stage 11, andan anterior choroid plexus after stage 15.Projecting dorsad between the optic lobes isthe epiphysis, originating from the posteriormargin of the diencephalon. This structurebecomes visible at stage 14 as a small conicalprojection. At stage 16 the epiphysial stalkloses its lumen. The metencephalon (cere-bellum) is small and scarcely distinguishable.It becomes hidden by the expanding opticlobes. The medulla possesses a characteristictela choroidea after stage 10 and a posteriorchoroid plexus at stage 14.

The appearance of a neurocoele first takesplace at stage 8, simultaneously with theformation of the optic vesicles. The latterare initially solid buds, as demonstrated forFundulus (Oppenheimer, 1937). By stage 9the opticoele is fully developed and lensplacodes are present. Invagination of thelens is complete at stage 11. The distinctionbetween the thicker, columnar sensory layerand the thinner pigment layer of the retina ispresent at stage 10. Pigment formation doesnot take place until stage 13.

PIGMENT FORMATION

DuShane (1943, 1948) reviewed the ex-perimental work demonstrating the origin ofpigment forming cells from neural crest. Thisderivation has been shown to be common tothe Amphibia, birds, and mammals. The onlyevidence for neural crest origin in the teleostsis an early report by Borcea (1909), in whichthe claim is made that in Uranoscopus,


Fierasfer, and Labrax, pigment develops in of cells that develops a lumen from thethe early neural crest stage, and its subse-quent migrations can be followed. The paperis very brief and gives no description of thetechniques used. The problem of pigmentcell origin in teleosts, then, is badly in needof clarification and confirmation.1

esophageal region caudad, by excavation

pharyngeal pouches begin

rather than by invagination as described byAgassiz and Whitman (1885) and Wilson

to

(1889). This lumen first appears in stage 10,

invaginate at

with an intestinal portal at a level just

stage 11, but the body of

posterior to the pharyngeal region. The

the pharynx doesNeural crest, as such, was observed here

only as cranial and spinal nerve primordia,making its. first appearance as the first cranialplacode in stage 9.

It is noteworthy that, in contrast to thesituation in most oviparous teleosts, theinitial site of pigment formation in the platy-fish is the retina (stage 13), followed closelyby pigment differentiation in the opticchoroid layer and the epineural region of thehind-brain (stage 15). The initial epineuralmelanophores are the dendritic type, and thedefinitive corolla-type micromelanophores donot appear until stage 20. These melano-phores have been interpreted as the micro-melanophore type on the basis of geneticdata both here and previously by Gordon(1931b). The macromelanophores do not ap-pear normally until some time after birth.Punctate peritoneal melanophores first be-come visible in the lateral peritoneum atstage 17.

These melanophores are greatly affectedby the. presence of the “golden” gene (st).st).

Initial pigment development still takes placein the retina at stage 13, but subsequent ap-pearance of melanophores is delayed untilstage 20 or 22. The number and size of thelatter are greatly reduced, but a large num-ber of xanthophores and leucophores arevisible as early as stage 18.

Pigment formation in other teleosts isusually initiated on the yolk and around theheart. In Fundulus this occurs at a stageapproximately matching stages 10 to 13 ofthe platyfish (Oppenheimer, 1937). In Caras-sius eye and head pigment also develop

initially, but at a relatively late stage ofdevelopment (Battle, 1940).

ENDODERMAL DERIVATIVES

In the platyfish the gut is first a solid core1 The writer’s attention was directed recently to a re-

port by G. V. Lopashov (1944, Compt. Rendus (Dok-lady) AcadSci. U.R.S.S., vol. 44, pp. 169472) in whichtransplantation experiments indicated a neural crestorigin of pigment in teleost fishes.

not invaginate until stage 15, when thestomodaeum is well formed. This is in agree-ment with the observations of Wilson (1889).

The first pharyngeal pouch, althoughinitially the largest, becomes greatly reducedby stage 15, i.e., a vestigial spiracle. Fivegill clefts, representing five pharyngealpouches, are formed by stage 15, separatedby four functional branchial arches. Posteriorto the last cleft, a pair of lateral evaginationsfrom the pharynx begin to form at stage 17and reach their maximum at stage 19, afterwhich regression of these structures occurs.It is presumed that the latter represent thevestigial pair of seventh pharyngeal pouches.

Garman (1895) stated that many of thecyprinodonts possess small enlargements ofthe gut just posterior to the origin of theductus pneumaticus and that these swellingsrepresent poorly developed stomachs. Theorder as a whole, then, is characterized by theoccurrence of vestigial stomachs. A similarsituation was reported for a minnow (Campo-stoma) by Rogick (1931), in which the histol-ogy of the gut was-described.

In ontogeny, the platyfish is shown to havea transient gastric region at one stage in itsdevelopment (stage 14). This vestigial seg-ment of the gut loses its identity in subse-quent stages.

The post-anal gut (Kupffer’s vesicle)arises early (stage 7), similar to its develop-ment in the whitefish (Price, 1934a). Itremains shallow and groove shaped untilstage 13 when it begins to invaginate furtherand becomes vesicular. Eventually Kupffer’svesicle contributes to the formation of theurinary bladder in Platypoecilus. ciluscilus. At stage 15,the Wolffian (nephric) ducts empty into thebase of the post-anal (urinary) vesicle. Thevesicle is slightly bilobed and extends craniadand dorsal to the anal portion of the gut.In subsequent stages, up to stage 26, thebilobed nature of the urinary vesicle becomes


more evident, the cranial end of each of thelobes coinciding with the caudal end of thegonadial ridges.

In his original description, Kupffer (as re-ported by Wilson, 1889) compared this earlyendodermal invagination (now called Kupf-fer’s vesicle) to the allantois in the reptilesand birds. Wilson (1889) and Sumner (1904)considered this structure as a post-anal gutwhich represents the site of a “neurentericstreak.” The homology of the latter to theneurenteric canal of the Amphibia has beenmaintained more recently by Rugh (1943).Price (1934a), however, stated that his ob-servations did not confirm the above view.Oppenheimer (1937), in her description ofFundulus development, described the urinaryvesicle as forming early in the embryo andgiving rise to the bilobed urinary bladder, butshe made no mention of the contributionof Kupffer’s vesicle to this structure.

The first indications of thyroid tissue werefound at stage 17, appearing as a loosely ar-ranged group of follicles around the ventralaorta at the level of the third and fourthaortic arches. The follicles possess a cuboidalepithelium whose cells are basophilic (haema-toxylin and eosin stain). Most of the folliclesare solid until stage 18 or 19, when a lumenappears containing a small amount of colloidtypical of thyroid tissue. In subsequentstages the follicles subdivide and may befound scattered in the pharyngeal tissues.The thyroid is not normally encapsulated inteleost fishes, except in the case of the seacatfish (Galeichthys felis) (Fowler, 1942) andthe parrot fishes (Matthews and Smith,1947).

Liver tissue begins to form at stage 14,initially by a series of solid cords of tissueramifying from the left side of the intestineto surround and invade, in part, the left ductof Cuvier. The liver cords soon developlumina and form a consolidated mass oftubules, with a mesodermal reticulum,drained by a single ductus choledochus (stage15). At stage 17 a ventral diverticulum fromthe base of the ductus forms the gall bladder.

Pancreas tissue begins to differentiate fromthe intestine at stage 16, by a large number ofbudding cords of cells. The pancreas lobulespossess lumina by stage 18, and these can betraced back to the intestine via a number of

fine pancreatic ducts, all of which empty intothe intestine around the base of the ductuscholedochus. A single large pancreatic lobe,located anterior to the ductus choledochusand first distinguishable in stage 17, pos-sesses no ductules and is composed of stronglybasophilic cells. This structure, on the basisof its origin and position, may be thehomologue of islets of Langerhans tissue.

The swim bladder formation is pre-indicatedin stage 14 by a vascular plexus on the rightof the intestine at the level of the vestigialstomach. At stage 15 a short duct extendingto the right of the esophagus and intestineinto this plexus represents the swim bladderprimordium. At stage 16 the swim bladderis well formed, extending retroperitoneally,dorsal to the intestine. By stage 18 it reachesthe level of the anus. The initial epitheliumof the duct and that of the bladder at stage 16consist of high columnar cells with basalnuclei and large distal vacuoles. This epithe-lium is retained in subsequent stages only bythe duct and the basal (anterior) portion ofthe bladder. The lateral and caudal expansionof the bladder consists of a one- or two-layered squamous epithelium. The ductcloses and degenerates shortly after birth.

MESODERMAL DERIVATIVES

Non-segmented epimere first appears instage 8, and occasionally as many as threerelatively diffuse pairs of somites may bepresent at this stage. Seven or eight pairs ofcompact somites can be distinguished atstage 9. Thirteen somites is the averagenumber for stage 10, 19 for stage 11, and 20to 21 for stage 12. In the latter, striatedmuscle fibers differentiate, and the embryoexhibits slow, twitching movements of itsbody and tail. Twenty-four to 25 pairs ofsomites are visible at stage 14, these ap-proaching the shape of typical myotomes.The full complement of 26 myomeres ispresent at stage 15, and the caudal finprimordium appears concomitantly with thelast myomere.

The spleen first forms from a vascularplexus originating in part from the plexusaround the ductus pneumaticus and in partfrom the hepatic portal plexus, becoming ahighly vascularized lymphoid organ. Thespleen is distinguishable as a discrete organ at


stage 17, and changes little in character insubsequent stages.

The diffuse heart primordium is evidentfirst in stage 9. Whether or not the heartprimordium in the platyfish is originally apaired structure (as in Gregonus, Price,1934b) is not known, since the endocardialanlagen are too diffuse to be properly in-terpreted. By stage 12 the vascular circuit iscomplete and the heart possesses an elongatesinu-atria1 portion projecting craniad, ashorter bulbous ventricle protruding towardsthe right side, and a short ventral aorta withno distinguishable conus region. At this stagethe heart is beating quite regularly, and verylight pinkish-colored blood is present.

The first blood vessel to develop is thedorsal aorta, which forms in stage 10 alongthe midline, compressed closely between thenotochord and the gut primordium. It is herethat the cellular elements of the blood firstappear, similar to the description given forCoregonus (Price, 1934b).

The nomenclature of the blood vessels asused here was interpreted from the followingsources: E. S. Goodrich (1930), Bridges(1904), Sayles (1938), Hyman (1942), andAllis (1908).

The third aortic arch is first to develop andforms the base for the internal carotid artery(fig. 12). The latter initially supplies only anextensive optic plexus. As shown in figure 14,the fourth, fifth, and sixth aortic arches de-velop next. The last to form are the secondand first arches and the external carotid, inthat order. The second arch (hyoideanartery) apparently functions as an afferentspiracular (or pseudobranchial) artery, andthe first arch as the efferent spiracular (orpseudobranchial), this conclusion being basedon the observed direction of blood flow andflow of injection material (India ink) in livingembryos.

The development of the ventral aorta, asshown in figure 14 and described for stages 13,14, 15, 16, 17, and 18, exhibits a longitudinalsplitting so that in essence the blood flow inthat vessel changes from caudad to craniad.It is noteworthy that the establishment ofthe definitive ventral aorta coincides with astage in development (stage 18) subsequentto which the yolk volume becomes rapidly

reduced. With this reduction in yolk volume,the locus of the sinus venosus describes anarc extending caudad and dorsad until theheart eventually takes up its final positionin the intra-embryonic pericardium with nointervening yolk sac portal circulation be-tween the ducts of Cuvier and the sinusvenosus.

The development of the gonads was tracedup to parturition in this work and was foundto agree in every particular with the de-scription of Wolf (1931). No evidence ofsexual dimorphism was observed up to birth,nor was there any sign of transforma-tion of somatic cells into primordial germcells.

In the development of the kidneys, thenephric duct is the first structure to form inthe mesomere region (stage 11) and may becalled the pronephric duct at this stage.Mesonephric tubules, occupying a limitedregion, appear somewhat later in stage 15, atwhich time the nephric ducts possess alumen throughout and empty into the urinarybladder. At stage 17 or 18 a pair of smallaccessory nephric ducts arise near the caudalmargin of the tubule region. These structuresextend caudad for several segments and endblindly, and they have been found unchangedin form in adult Platypoecilus and Lebistis.The functional significance of the accessorynephric ducts is unknown, but it may bepossible that they represent vestigial homo-logues of true vasa deferentia of othervertebrates, particularly elasmobranchs.

The location and origin of adrenal tissuewere not detected in this work, and thesolution to this question will await futurework involving the utilization of specialtechniques for detection of chromaffin andcortical tissues. However, it was noted thatmost of the visceral blood vessels (stages18 ff.), and particularly the veins, werecoated by three or four layers of lenticularcells. Whether or not this represents adrenaltissue remains to be seen.

The development of the corpuscles ofStannius was found to agree with the descrip-tion of Garrett (1942). Preluminal stageswere found in stage 15 embryos, luminalstages at stage 17 or 18, and post-luminalstages as early as stage 19 or 20.

COMPARISON OF PLATYPOECILUS WITH XIPHOPHORUSAND HYBRIDS

WHEN SIMILARLY COLLECTED EMBRYONICstages of Xiphophorus hellerii, Platypoecilus,and their hybrids were compared, only threepoints of difference were observed, namely,time of initial pigment formation, rate ofgrowth, and rate of caudal fin development.In all other respects organogenesis in the twogenera was found to be the same, the figuresand descriptions for Platypoecilus stagesbeing applicable to Xiphophorus and Xipho-phorus-Platypoecilus hybrids, aside from thedifferences mentioned above.

The laboratory-bred stock of Xiphophorushellerii (described in the Materials andMethods section) was used in all of thiswork, and this was the same stock that wasused in producing the hybrids. Some addi-tional observations are included on a collec-tion of swordtails representing a differentstrain (collected at L’Encero, near Jalapa,

Mexico, in 1948), but these data were keptseparate since no living material of thisstrain was available.

The average interval between broods forXiphophorus hellerii was reported as 30 daysby Bailey (1933) for a domesticated stock ofunknown origin and purity. In the presentreport, the variation in brood intervals washigh (27 to 43 days, average 29.2 days), andtoo few females were available for a reliabledetermination of an average value. However,the variations were well within the limits ofvariability for the platyfish, so that sword-tail embryonic growth stages could be plottedagainst the scale determined for the platyfish.In the case of the hybrids, the brood in-tervals were 25 to 34 days, averaging 28.6.The hybrid growth stages, then, could alsobe plotted against the same time scale.

TIME OF INITIAL PIGMENT FORMATION

As was shown previously for Platypoecilus,pigment is first visible at stage 13 in theouter retinal layer. At stage 15 iridiocytesand additional melanophores are present inthe secondary choroid coat of the eye, andstellate epineural melanophores first appear(fig. 20A).

In the swordtail, however, pigment ap-pears slightly later. Stage 13 exhibits nopigment, whereas stages 14 and 15 possesspigment granules only in the outer retinallayer. Not until stage 16 does any extra-ocular pigment appear. By stages 19 ff. theembryos of the two forms are indistinguish-able with regard to pigmentation.

In all of the hybrids the time of initialpigment appearance and its future differ-entiation follow the time sequence of theplatyfish parent, i.e., first in the retina atstage 13, then epineural melanophores atstage 15. In about 12 per cent of stage 12embryos resulting from a backcross of an Flhybrid to Platypoecilus, some pigment wasevident in the retinal layer. Occasionalembryos of stage 14 were found with choroidand epineural pigment. Corolla-type melano-phores were distinguishable as early as stage

The initial pigmentation is apparently ofthe micromelanophore type, since its amountand distribution are affected only by thepresence of the recessive st gene in a homozy-gous condition, i.e., in the “golden” strainof the platyfish. The macromelanophoreproducing genes are not expressed pre-natally in the platyfish, F1, and F2 hybrids,and the wild stock swordtail does not possessany of these genes. However, the embryosresulting from a backcross to a platyfishexhibit very few prenatal macromelano-phores. The backcross to swordtails has beenshown to have an intensifying effect onmelanosis over that exhibited by F1 and F2hybrids (Gordon, 1937; Gordon and Flath-man, 1943; Gordon and Smith, 1938a). In thehybrid embryos studied here, the intensifica-tion resulted in the appearance of occasionalmacromelanophores on the sides of the bodyin stages 25 and 26, and this occurred onlyif the Fl hybrid parent possessed the S p

217

factor.

19 in these backcross hybrids. The stages atwhich the various types of pigment firstappear in the several strains of fish areshown in table 3.


TABLE 3APPEARANCE OF PIGMENT GIVEN IN TERMS OF MORPHOLOGICAL STAGE NUMBERS

Retinal pigment 13 14 13 13 13 12-13Choroid pigment 15 16 15 15 15 14-15Dendritic epineural melanophores 15 16 15 15 15 14-15Punctate peritoneal melanophores 17 18 17 17 17 17Dendritic dermal melanophores 17 18 17 17 17 17Corolla-type melanophores 20 20 20 20 , 20 19-20

F1 F2Hybrids Hybrids

GROWTH RATES

Figure 43 shows the mean standard lengthsof embryos of Platypoecilus and Xiphophorusplotted against time. After about the ninthday (stages 19 ff.) the growth rate of theswordtail is somewhat faster, and the larvalfish are usually about 1 mm. longer at birththan the platyfish. The change in the slopeof the curve from stage 18 is less well definedthan that of Platypoecilus. The differences

FI xX. F1 xP.hellerii maculatus

Hybrids Hybrids

between the terminal segments of the twocurves are considered significant, since thereis very little or no overlap in the distributionof points about the means.

When the above growth rates were com-pared with those of Platypoecilus-Xipho-phorus hybrids, it was found that, past stage

FIG. 43. Graph showing growth in standardlength of Platypoecilus embryos, as compared tothat of Xiphophorus, and of F1 and F2 hybridembryos.

FIG. 44. Graph showing growth in standardlength of Xiphophorus embryos, as compared tothat of embryos resulting from a backcross of Flhybrids to Xiphophorus.


FIG. 45. Graph showing growth in standardlength of Xiphophorus embryos, as compared tothat of embryos resulting from a backcross of Flhybrids to Platypoecilus.

18, all the hybrids either coincided with orexceeded the growth rate of the swordtail(figs. 43,44,45). The distinction between thehybrids and the platyfish, therefore, is verysharp.

The Fl and F2 embryos followed the growthcurve of the swordtail very closely, althoughthe variability in the hybrids was somewhatgreater. Whether or not the latter differenceis statistically significant is unknown, sinceinsufficient data were available for a com-parison of variabilities. Figure 43 shows asingle curve for combined data on Xipho-phorus Fl, and F2 hybrids.

The backcross hybrids (both types) ex-

StageNos.

5 .35 .35 .35 .34 .35 .346 .64 .62 .64 .64 .65 .657 .82 .83 .80 .80 .81 .828 1.06 1.08 1.06 1.04 1.04 1.069 1.23 1.22 1.22 1.23 1.24 1.25

10 1.45 1.45 1.47 1.43 1.48 1.4611 1.67 1.64 1.64 1.68 1.70 1.6712 1.81 1.78 1.79 1.84 1.81 1.8313 2.02 2.07 2.05 2.01 1.97 2.0514 2.31 2.28 2.28 2.30 2.33 2.3015 2.52 2.58 2.53 2.50 2.49 2.5116 2.80 3.13 2.95 2.89 2.96 2.8817 3.29 3.31 3.24 3.37 3.29 3.3318 3.57 3.55 3.61 3.62 3.54 3.5819 3.68 3.67 3.79 3.81 3.85 3.8720 3.87 3.92 4.13 4.13 4.05 4.1621 3.95 4.11 4.17 4.19 4.14 4.2722 4.00 4.32 4.40 4.39 4.38 4.6023 4.18 4.58 4.56 4.55 4.96 5.0424 4.45 4.96 5.01 5.02 5.40 5.3025 5.00 5.64 5.80 5.83 6.08 5.6926 5.69 6.60 6.47 6.49 6.78 6.54

P. maculatus

Correlationcoefficients .91

219

3.5 *.8 9 IO I I I2 13 I 4 I 5 16 I7 I8 19 2 0 21 22

AGE IN DAYS

TABLE 4

MEAN STANDARD LENGTHS IN MILLIMETERS

X . hellerii

.90

F1 Hybrids

.89

F2 Hybrids

.87

Fl xX.hellerii

Hybrids

.91

F1 xP.maculatusHybrids

.91


ceeded the growth rate of the swordtail inpart (figs. 44,45), but the differences here areof a low order of significance.

The correlation coefficients for the in-dividual curves ranged from .87 to .91, all ofa high order of significance. The mean valuesfor standard lengths of embryonic stages ofthe pure species and the hybrids are given intable 4.

A group of embryos was obtained fromXiphophorus females collected by Gordon,Atz, and Wood at L’Encero, in 1948. Al-though this group of females did not appearto exceed laboratory-raised animals in size,the dimensions of the embryos were con-siderably larger than those of the laboratory

strain. The disparity in size is evident in veryearly stages. There was also a difference inegg size between the two strains. The eggdiameter for the laboratory-bred Zacatispanstrain averaged 1.6 mm. and for the L’Encerostrain 2.3 mm. Table 5 shows a comparisonbetween the L’Encero strain and the labora-tory-bred Xiphophorus hellerii. The com-parison cannot be made graphically sinceno living L’Encero females were available andnothing is known of their developmental rateand the length of the brood intervals. Inaddition, it is not known whether the differ-ences may be ascribed to genetic diversity orto environmental conditions.

TABLE 5MEAN STANDARD LENGTH MEASUREMENTS IN MILLIMETERS OF Xiphophorus hellerii

StageNos.

789

10111213141516171819202122232425

Zacatispan Strain(Laboratory Bred)

.831.081.221.451.641.782.072.282.583.133.313.553.673.924.114.324.584.965.64

L’Encero Strain(Field Bred)

0.791.111.321.581.822.102.452.612.953.874.635.025.275.796.326.617.107.898.68

GROWTH OF THE CAUDAL FIN

As was described in the previous section 21 B). This occurs just after the full comple-on Platypoeclius, the caudal fin begins to ment of myomeres is attained, the normaldifferentiate at stage 15 (eight days; fig. number being 26. (The numbers of somites,19B) and fin rays appear at stage 16 (fig. myotomes, myomeres, and vertebrae are


considered equal.) The normal number ofbody and tail segments in Xiphophorus is 28(Gordon and Benzer, 1943, and the last twosegments are not formed until stage 16, one

FIG. 46. Graph comparing rate of embryoniccaudal fin differentiation in Platypoecilus to thatof Xiphophorus.

StageNos.

P. maculatus X . hellerii F1 Hybrids F2 Hybrids

15 .0916 .1217 .2318 .3019 .3220 .4721 .6322 .7923 .9224 1.0525 1.3426 1.76

Correlationcoefficients .79

stage later than in the platyfish. The initialdifferentiation of the caudal fin in Xipho-phorus is also one stage later, i.e., stage 16

d0.5 -

/

0 l0

X

If� .

☺

.

.

FIG. 47. Graph showing similar rate of em-bryonic caudal fin differentiation in Platypoecilus,F1 and F2 hybrids, and embryos resulting from abackcross of Fl hybrids to Xiphophorus,

TABLE 6

LENGTH OF CAUDAL FIN IN MILLIMETERS

.08

.12

.18

.24

.30

.39

.54

.881.021.331.69

.85

.09 .09

.20 .21

.31 .30

.41 .41

.43 .44

.57 .56

.66 .69

.82 .81

.86 .86

.97 .991.42 1.411.70 1.69

.86 .87

F1 xX.hellerii

Hybrids

.08

.l5

.27

.33

.36

.50

.59

.76

.801.071.381.78

.83

F1 xP.maculatusHybrids

.07

.14

.24

.29

.36

.69

.82

.911 .Ol1.231.611.88

.80


(about nine days), with fin rays appearing atstage 17. This is shown by figure 46, whichalso shows that, although initial differentia-tion of this structure in the swordtail is de-layed, at birth it is not significantly differentin size from that of the platyfish.

In a comparison of the rate of caudal findifferentiation in the pure species with thatof the hybrids, it was seen that initialdifferentiation occurs at stage 15, as in theplatyfish, and this is true even among theembryos resulting from a backcross of an Flhybrid to a swordtail. The latter embryos andthe F1 and F2 embryos follow the platyfishvery closely in terms of rate of caudal fingrowth and its final dimensions (fig. 47). Thegroup resulting from a backcross to a platy-fish exceeds the parental type in caudal fingrowth rate and reaches dimensions above

‘a a a m 1 a a 9 ..I I I., the average for the platyfish and the other5 6 7 8 9 I O 11 1 2 1 3 I4 15 16 I7 18 19 20 21 22

AGE IN DAYS hybrids (fig. 48).FIG. 48. Graph comparing rate of embryonic Table 6 summarizes the data upon which

caudal fin differentiation in Platypoecilus to that the graphs in figures 46, 47, and 48 wereof embryos resulting from a backcross of F1 based.hybrids to Platypoecilus.

FECUNDITY OF HYBRIDS

Among all the Platypoecilus embryonicbroods examined in this work, occasionalabnormalities were found. The most commonof these were the occurrence of over-ripe,unfertilized ova and degenerate gastrulae orblastulae (described under stage 1). Ab-normalities in later stages were exceedinglyrare, represented by a total of nine partiallydisintegrated embryos of stage 13, 16 em-bryos with curved and retarded tails foundat the rate of one or two in a brood, oneacephalic embryo, and one two-headedembryo of stage 16. The average frequencyof abnormalities of all previously mentionedtypes was 4.2 per cent per embryonic brood.

The frequency of occurrence of anomalieswas distinctly higher in the “golden” strainof platyfish, and as many as 23 per cent of theembryos of stages 18 to 26 were found to bedead and degenerating.

Anomalies in embryonic broods of Xipho-phorus were found at a slightly lower fre-quency than in the platyfish, i.e., 3.4 per centper embryonic brood.

The same types of abnormalities were

found in the embryonic broods of the varioustypes of hybrids used in this work, but themost frequent types were delayed or anomal-ous gastrulae and later stages with curled orretarded tails. The abnormal shapes of thetails were the result of an unevenly developednotochord which often assumed a spiralshape. The frequency of occurrence of theanomalies was found to be correlated withthe particular pattern genes involved in thecross (table 7). As described in the Materialsand Methods section, the F1 hybrids pos-sessed either of two genes for macromelano-phores, Sp (spots on the sides) or Sd (spot onthe dorsal fin). The expression of these genesin the hybrids results in a melanotic ormelanomatous condition (as described anddefined by Gordon and Smith, 1938a).Anomalies among Fl embryos were infre-quent. Out of 33 Fl females of the Sp type,only two females proved to be completelysterile, and three possessed aberrant broods,with over 30 per cent anomalous embryos.The total frequency of anomalies in thelatter was 8.7 per cent per embryonic brood.


TABLE 7FERTILITY AMONG FEMALES OF Platypoecilus, Xiphophorus, AND THEIR HYBRIDS

223

Strains or Species

Pure speciesP. maculatus, wild stock, inbredP. maculatus, golden strainX. hellerii, wild stock, inbred

Platypoecilus-Xiphophorus hybridsP. maculatus Q XX. hellerii F1Sp QXFlSp 3F1 Sp QXFdd CTF1 Sp XX. hellerii 3Fl Sp XP. maculatus Cp

891023

12104127

F1Sd XF1Sp 8 11FlSd 9 XFlSd 3 6F1 Sd XX. hellerii 3 9F1 Sd XP. maculatus 8 8

P. maculatus XF1 Sd 3 5

Total No.No. Sterile

200

01001

3345

0

No. withAberrant

EmbryonicBroods

(i.e., withOver 30%Anomalies)

210

01000

3222

0

,

No. Produc-ing Abortive

Broods(i.e., Less

than 6Fish perBrood)

230

0111 0

10011

0

AverageFrequency ofAnomalous

Embryos perEmbryonic

Brood

4.2%23 %3.4%

1.1%

8.7%

31 % 47%

82 %I

0

This held for all types of crosses, backcrossor inbred, using both Sp and Sd hybridmales. An embryonic brood possessing a 30per cent or higher incidence of anomalousstages is considered aberrant.

Out of 34 Sd hybrid females, 15 were com-pletely sterile and 11 possessed aberrantbroods or gave birth to abortive broods ofone to five individuals. The embryonicanomalies occurred with a frequency of 47

cent. In the reciprocal cross, using Sd males,no anomalies were found.

See table 7 for list of crosses used, withtotal number of females, number of sterilefemales, and those females that producedaberrant or abortive broods.

Part of the records on sterility were kindlysupplied by Miss Eugenie Clark of theDepartment of Animal Behavior of theAmerican Museum of Natural History from

per cent per embryonic brood. Sd females unpublished data, and all her records oncrossed to hybrid males or swordtail males sterile female hybrids indicated that motileexhibited an average of 31 per cent anomalies sperm were recovered from the oviducts afterper embryonic brood, and when backcrossed copulation.to platyfish males the frequency was 82 per

DISCUSSION OF THE EFFECTS OF HYBRIDIZATION ON DEVELOPMENT

In 1948 Myron Gordon reviewed a series sary were those producing patterns of macro-of his papers on the genetics of melanomas in melanophores. The melanosis in Fl hybridsfishes. He showed how melanosis may be in- frequently developed into a melanomatousduced in Platypoecilus-Xiphophorus hybrids condition, which could be further intensifiedif certain pattern genes are present in the by backcrossing the hybrid to the swordtail.platyfish parent. In all cases the genes neces- The site of the melanoma could be predicted


on the basis of which of the several genes wasused, i.e., members of the sex-linked, domi-nant, allelic series-Sp (spots on the sides),Sd (spot on the dorsal fin), Sb (spotted belly),N (broad black lateral band), and Sr (stripeson the sides). The genes Sp, Sb, and N pro-duce the most intense melanosis and themost destructive melanomas. The genes Sdand Sr were expressed later in life and pro-duced a more localized melanotic conditionin hybrids (Gordon, 1948; Levine, 1948).

In 1943 Gordon and Flathman showed howthe intensified melanosis produced by a back-cross to Xiphophorus was expressed in lateembryonic stages. The pattern gene utilizedwas N, forming a broad, black, lateral bandof macromelanophores. A similar prenatalmelanosis was described by Gordon andSmith (1938a), in which case the Sp gene wasused. Gordon (1946) showed that the stockfishes used previously were “domesticated”and consequently were not pure Platypoecilusbut possessed swordtail genes ‘acquired byintrogressive hybridization. The N and Spgenes used earlier may have been associatedwith some intensifying modifiers inheritedfrom the swordtail. It is not surprising, there-fore, that no such effects were observed in thepresent paper, since exclusively wild-typemacromelanophore genes were used.

The effect of hybridization on embryonicgrowth rate somewhat paralleled the produc-tion of melanosis. The hybrids all fell in withthe more rapidly developing of the twoparental species, at least with respect togrowth and caudal fin differentiation.

One possible genetic explanation of hybridmelanosis was that offered by Gordon (1948),which involved the assumption that theswordtail possesses dominant modifiers A andB. Either of these genes combined with aplatyfish macromelanophore factor wouldproduce melanosis, and when, by backcross-ing, both A and B are present, the melanosisis further intensified. A somewhat similarexplanation may be applied to results de-scribed here.

In the case of the body growth rate, thegrowth modifiers of the swordtail appear tobe dominant, since both Fl and F2 hybridsfollowed the growth rate of Xiphophorus. Aslight intensification was achieved by back-crossing to Xiphophorus. The fact that the

same result was obtained in P. maculatusbackcross hybrids seems to indicate that notonly were a large number of dominant genesinvolved but that a reassortment with certainplatyfish genes exhibited the effects ofheterosis.

With regard to the growth of the caudalfin, the reverse situation obtained. The F1 andF2 hybrids developed a caudal fin at a ratecoinciding with that of the platyfish parent.Backcrossing to P. maculatus distinctly in-tensified the development of this structure,indicating the accumulation of more platy-fish modifiers.

It was shown here that the swordtailinitially develops pigment at a slightly laterstage than does the platyfish. All the hybridsformed pigment at the same stage the platy-fish did and often earlier. This was partic-ularly true of the P. maculatus backcrosshybrids, indicating again that the oneparental species possessed numerous domi-nant modifiers, in this case the platyfish hav-ing genes for earlier melanin formation. Thishas no apparent direct bearing on hybridmelanosis, since the effect is primarily onthe micromelanophores, and the macro-melanophores do not differentiate until afterbirth. It is quite likely that the site of originof micromelanophores and macromelano-phores is the same, but it is not known justwhen a pigment cell acquires the specificpotency for either one or the other of the twoftypes, or if non-pigmented micromelano-phores are the direct source of macromelano-phores. In 1937 Gordon reported the effect ofthe golden gene (st) on the expression of themacromelanophore genes. The st gene is to befound only in domesticated stocks. It is arecessive mutant, producing a great reductionin the number of micromelanophores. Thismutation was traced back by Gordon (1935)as probably occurring during the early yearsof tropical fish culture and later introducedinto the swordtail by introgressive hybridiza-tion (1946b). Although the st gene affects the micromelanophores primarily, Gordon(1927, 1937) showed that it also produced areduction in the numbers of macromelano-phores. There is, then, some interactionbetween the two kinds of melanophores andthe factors controlling their expression.

In Platypoecilus certain genes have been


found to reduce viability. The behavior oftwo such genes has been described by Gordon(1948). The albino gene, arising as a mutantin domesticated stocks of swordtails, pro-duced aberrant Mendelian ratios as a resultof embryonic lethal effects (Gordon, 1942).The gene for golden, also arising duringdomestication, is described here as beingassociated with poor viability among em-bryos. However, hybridization produced notonly embryonic death but complete sterility,as shown by Gordon (1946a).

Gordon (1946a) reported that F1 hybridfishes exhibited a large percentage of sterilityand highly aberrant sex ratios. The crossoriginally was between a Xiphophorus helleriifemale and a Platypoecilus maculatus male,the latter possessing the Sr gene on the X-chromosome and the Sd gene on the Y-chromosome. As expected, 50 per cent of theoffspring were of the Sr type and were allfertile females. Of the remaining hybrids,possessing the Sd factor, only about 8 percent were fertile males and females, and 92per cent were sterile. Both the sterility andthe aberrant sex ratio were explained on thebasis of a multifactorial sex-determiningmechanism of the swordtail as opposed tothe digametic type in the platyfish, thesterility being associated in the hybrids withthe presence of the platyfish Y-chromosome.

It should be noted that in the present work

the fundamental cross was the reciprocal ofthe above. The female was a platyfish withone X-chromosome marked by the Sd gene,and the other by Sp. No platyfish Y-chromo-some was involved, and yet sterility and poorviability were of frequent occurrence andassociated almost entirely with the Sd gene.The sex ratio was aberrant, with femalespredominating. The same cross was reportedby Gordon in 1948, but the results given atthat time were based on incomplete data. Ad-ditional data have tended to corroborate theobservations made here (Gordon, personalcommunication).

Two possible explanations are presentedhere. The Sd factor itself may possess somelethal effect which is expressed when sword-tail modifiers are present. On the other handthe lethal factor may be linked to the Sd geneby virtue of its being on the same chromo-some, and be expressed when dissociatedfrom its allele as a result of hybridization.The lethal effect is apparently one that isinherent in the female parent rather than inthe embryos, since the use of sd hybridmales in matings with Sp hybrid females orP. maculatus females produced viable off-spring. The final explanation of this linkageof embryonic lethal effects and sterility withthe Sd gene will depend upon additionalcrosses and analyses of sterility data.

SUMMARY

1. EMBRYONIC DEVELOPMENT of the vivip-arous poeciliid fish Platypoecilus maculatusmay be conveniently divided into 26 stagesfor descriptive purposes.

2. The embryonic growth rate for P.maculatus is determined by comparison ofmorphological age (in terms of embryonicstage) with theoretical age (days after thebirth of a previous brood). This comparisonis-based on a series of embryonic broods ob-tained by laparotomy at various times duringthe twenty-eight- to thirty-day reproductivecycle of this species.

3. A study of growth and organogenesis inP. maculatus reveals the following features:

A. The concomitant regression of theextra-embryonic membranes and the yolkmass is accompanied by a drop in the slopeof the growth curve.

B. The initial formation of the extra-embryonic membranes takes place in a man-ner similar to that in the “amniotes.”

C. Cleavage, gastrulation, and neurulationfollow a course basically similar to that ofother teleosts.

D. Blastopore closure occurs after tail budformation.

E. The notochord forms directly fromendoderm by delamination.

F. Pigment formation is evident first inthe retina and subsequently in the choroidand meningeal coats.

G. Kupffer’s vesicle (post-anal gut) con-tributes to the development of the urinarybladder.

H. The formation of the ventral aorta andaortic arches is described with regard to thereversal of blood flow in the ventral aorta andthe change of position of the sinus venosusduring yolk regression.

4. Comparisons with other species ofPlatypoecilus (P. variatus, couchianus, andxiphidium) reveal no significant differencesin embryonic development. 5. Platypoecilus maeulatus development,compared to that of Xiphophorus hellerii andPlatypoecilus-Xiphophorus hybrids, shows thefollowing differences:

A. Initial pigment differentiation in Xipho-phorus occurs one. stage later than in Platy-poecilus. The hybrids form pigment at thesame stage as Platypoecilus or earlier.

B. Over the same period of time, Xipho-phorus embryos attain a larger size thanPlatypoecilus embryos. The growth rates ofthe hybrids all coincide with or exceed those ofXiphophorus.

C. Caudal fin differentiation is one stagelater in Xiphophorus than in Platypoecilus.However, all the hybrids form a caudal fin atthe same rate as P. maculatus, and theF1X Platypoecilus backcross hybrid embryosexceed both parental species in this respect.

6. Sterility and embryonic anomalies arefrequent among hybrid females. This isparticularly true of F1 females possessing theSd (spotted dorsal fin) character, indicating alinkage of lethal factors with this patterngene.

226

LITERATURE CITED

AGASSIZ, L., AND C. 0. WHITMAN1885. On the development of some pelagic

fish eggs. Preliminary notice. Proc.Amer. Acad. Arts and Sci., vol. 20, pp.2 3 - 7 5 .

ALLIS, E. P., JR.1908. The pseudobranchial and carotid ar-

teries in the gnathostome fishes. 2001.Jahrb., vol. 27, pp. 103-134.

BAILEY, R. J.1933. The ovarian cycle in the viviparous

teleost, Xiphophorus hellerii. Biol. Bull.,vol. 64, pp. 206-225.

BALFOUR, F. M.1878. A monograph on the development of

elasmobranch fishes. London, Works ofFrancis Mai tland Balfour, Macmillanand Co., 1885.

BATTLE, H. I.1940. Embryology and larval development of

the goldfish (Carassius auratus L.).Ohio Jour. Sci., vol. 40, pp. 82-93.

BORCEA, M. I.1909. Sur I’origine du coeur, des cellules

vasculaires migratrices et des cellulespigmentaires chez les teleosteens. Comp.Rend. Acad. Sci. Paris, vol. 149, pp.688-689.

BRACHET, A.1921. Traite d’embryologie des vertebretes.

Paris, Masson et Cie.BRIDGE, T. W.

1904. Fishes, exclusive of the systematic ac-count of Teleostei. Cambridge NaturalHistory, vol. 8, pp. 141-537.

CLAPP C. M.1891. Some points in the development of the

toadfish (Batrachus tau). Jour. Morph.,vol. 5, pp. 494-501.

COHEN, H.1946. Effects of sex hormones on the develop-

ment of the platyfish, Platypoecilusmaculatus. Zoologica, vol. 31, pp. 121-128.

DUSHANE, G. P.1943. The embryology of vertebrate pigment

cells. I. Amphibia. Quart. Rev. Biol.,vol. 18, pp. 109-127.

1948. The development of pigment cells invertebrates. In Biology of melanomas.Special Publ. New York Acad. Sci.,vol . 4, pp. l-14.

FOWLER, E.1942. Embryology of the sea catfish, Galeich-

thys felis. (Abstract.) Proc. LouisianaAcad. Sci., vol. 6, p. 81.

GARMAN, S.1895. The Cyprinodontes. Mem. Mus. Comp.

Zool., vol. 19, pp. 1-179.

GARRETT, F. D.1942. The development and phylogeny of the

corpuscle of Stannius in ganoid and tele-ostean fishes. Jour. Morph., vol. 70,pp. 41-67.

GERSCHLER, W. S.1914. Faber alternative Vererbung bei Kreu-

zung von Cyprinodontiden-Gattungen.Zeitschr. indukt. Abstamm. Vererb.,vol. 12, pp. 73-96.

GOODRICH, E. S.1930. Studies on the structure and develop-

ment of vertebrates. London, Mac-millan and Co.

GORDON, MYRON1927. The genetics of a viviparous top-

minnow, Platypoecilus the inheritanceof two kinds of melanophores. Genetics,vol. 12, pp. 253-283.

1931a. Hereditary basis of melanosis in hybridfishes. Amer. Jour. Cancer, vol. 15, pp.1495-1523.

1931b. Morphology of the heritable color pat-terns in the Mexican killifish, Platy-poecilus. Ibid., vol. 15, pp. 732-787.

1935. Discovery of the gold platyfish. Jour.Hered., vol. 26, pp. 96-101.

1937. The production of spontaneousmelanotic neoplasms in fishes by selec-tive matings. II. Neoplasms with macro-melanophores only. III. Neoplasms inday old fishes. Amer. Jour. Cancer, vol.30, pp. 362-375.

1942. Mortality of albino embryos and aber-rant Mendelian ratios. Zoologica, vol.27, pp. 73-74.

1946a. Interchanging genetic mechanisms forsex determination in fishes under do-mestication. Jour. Hered., vol. 37, pp.307-320.

1946b. Introgressive hybridization in domesti-cated fishes. I. The behavior of Comet, aPlatypoecilus maculatus gene, in Xipho-phorus hellerii. Zoologica, vol. 31, pp.7 7 - 8 8 .

1947a. Genetics of Platypoeci lus maculatus IV.The sex determining mechanism in twowild populations of the Mexican platy-fish. Genetics, vol. 32, pp. 8-17.

1947b. Speciation in fishes. Distribution in timeand space of seven dominant multiplealleles in Platypoecilus maculatus. Ad-vances in Genetics, vol. 1, pp. 95-132.

1948. Effects of five primary genes on the siteof melanomas in fishes and the influenceof two color genes on their pigmentation.In Biology of melanomas. Special Publ.New York Acad. Sci., vol. 4, pp. 216-269.

2 2 7


GORDON, M., AND P. BENZER1945. Sexual dimorphism in the skeletal ele-

ments of the gonopodial suspensoria inxiphophorin fishes. Zoologica, vol. 30,pp. 57-72.

GORDON, M., AND F. FLATHMAN1943. The genetics of melanoma in fishes. VI.

Mendelian segregation of melanophorereaction types in embryos of a melanom-atous mother. Zoologica, vol. 28, pp.9 - 1 2 .

GORDON, M., AND G. M. SMITH1938a. Progressive growth stages of a heritable

melanotic neoplastic disease in fishesfrom the day of birth. Amer. Jour.Cancer, vol. 34, pp. 255-272.\

’ 1938b. The production of a melanotic neo-plastic disease by selective matings. IV.Genetics of geographical species hy-brids. Ibid., vol. 34, pp. 543-565.

HENNEGUY, F.1888. Recherches sur la developpement des

poissons osseaux. Embryogenie de latruite. Jour. Anat. Physiol., pp. 413-502.

HOPPER, A. F., JR.1943. The early embryology of Platypoecilus

maculatus. Copeia, pp. 218-224.HYMAN, L. H.

1942. Comparative vertebrate anatomy.Chicago, The University of ChicagoPress.

KAMITO, A.1928. Early development of the Japanese

killifish (Oryzias latipes), with notes onits habits. Jour. Colt. Agric., Imp.Univ. Tokyo, vol. 10, pp. 21-38.

LEVINE, M.1948. The cytology of the typical and the

amelanotic melanomas. In Biology ofmelanomas. Special Publ. New YorkAcad. Sci., vol. 4, pp. 177-215.

MATTHEWS, S. A., AND D. C. SMITH1947. Concentration of I131 by the thyroid

gland of the Bermuda parrot-fish.(Abstract.) Anat. Rec., vol. 99, p. 36,

MORGAN, T. H.1893. Experimental studies on the teleost eggs.

Anat. Anz., vol. 8, p. 803.1895. The formation of the fish embryo. Jour.

Morph., vol. 10, pp. 419472.OPPENHEIMER, J.

1934a. Experimental studies on the developingperch (Perca flavesens Mitchill). Proc.Soc. Exp. Biol. Med., vol. 31, pp. 1123-1124.

1934b. Experiments on early developing stagesof Fundulus. Proc. Natl. Acad. Sci.,

Washington, vol. 20, pp. 536-538.1935. Processes of localization in developing

Fundulus. Ibid., vol. 21, pp. 551-553.1936a. The development of isolated blasto-

derms of Fundulus heteroclitus. Jour.Exp. Zool., vol. 72, pp. 247-269.

1936b. Transplantation experiments on de-veloping teleosts (Fundulus and Perca).Ibid., vol. 72, pp. 409-437.

1936c. Processes of localization in developing

1937.

1938.

Fundulus. Ibid., vol. 73, pp. 405-444. The normal stages of Fundulus hetero-clitus Anat. Rec., vol. 68, pp. l-15.Potencies for differentiation in thetefeostean germ-ring. Jour. Exp. Zool.,vol. 79, pp. 185-212.

1947. Organization of the teleost blastoderm.Quart. Rev. Biol., vol. 22, pp. 105-118.

PASTEELS, J.

1933.

1934.

1936.

La gastrulation et la repartition desterritoires dans la moitie dorsal dublastodisque de trui te (Salmo iridaeus).Comp. Rendus Soc. Biol. Paris., vol.113, pp. 425428.Repartition des territoires et mouve-ments morphogenetiques de la gastrula-tion de l’oeuf de truite (Salmo iridaeus).Comp. Rendus Assoc. Anat. Bruxelles,pp. 45 l-458.Etudes sur la gastrulation des vertebresmeroblastiques. I. Teleosteens. Arch.Biol., vol. 47, pp. 205-308.

PRICE, J. W. 1934a. The embryology of the whitefish, Core-

gonus clupeaformis (Mitchell). Part. I.Ohio Jour. Sci., vol. 34, pp. 287-305.

1934b. The embryology of the whitefish, Core-gonus clupeaformis (Mitchell). Part II.Organogenesis. Ibid., vol. 34, pp. 399-414.

PURSER, G. L.1938. Reproduction in Lebistes reticulatus.

Quart. Jour. Micro. Sci., vol. 81, pp.150-158.

ROGICK, M. D.1931. Studies on the comparative histology of

the digestive tube of certain teleostfishes. II. A minnow (Campostomaanomalum). Jour. Morph. Physiol., vol.52, pp. l-25.

RUGH, ROBERTS1941. Experimental embryology; a manual of

techniques and procedures. New York,New York University.

1943. The neurenteric canal in the frog, Ranapipiens. Proc. Soc. Exp. Biol. Med.,vol. 52, pp. 304-307.

RYDER, J. A.1882. Structure and ovarian incubation of


Gambusia patruelis, a t o p - m i n n o w .Amer. Nat.,vol. 16, pp. 109-118.

SAYLES, L. P.1938. Manual for comparative anatomy. New

York, The Macmillan Co.SCRIMSHAW, N. S.

1944a. Superfetation in poeciliid fishes. Copeia,pp. 180-183.

1944b. Embryonic growth in the viviparouspoeciliid fish, Heterandria formosa. Biol.Bull., vol. 87, pp. 37-51.

1945. Embryonic development in poeciliidfishes. Ibid., vol. 88, pp. 233-246.

SIMPSON, G.G., AND A. ROE1939. Quantitative zoology. New York, Mc-

Graw-Hill Book Co.SUMNER, F. B.

1904. A study of early fish development. Arch.Entwmech. Organ., vol. 17, pp. 92-149.

TAVOLGA, M. C.[In press.] Differential effects of estradiol,

estradiol benzoate, and pregneninoloneon Platypoecilus maculatus. Zoologica.

TAVOLGA, W. N., AND R. RUGH

1947. Development of the platyfish, Platy-poecilus maculatus. Zoologica, vol. 32,pp. l-15.

TURNER, C. L.1937. Reproductive cycle and superfoetation

in poeciliid fishes. Biol. Bull., vol. 72,pp. 145-164.

1940. Pseudoamnion, pseudochorion, follicularpseudoplacenta in poeciliid fishes. Jour.Morph., vol. 67, pp. 59-89.

VAN O O R D T , G . J.1928. The duration of life of the spermatozoa

in the fertilized female of Xiphophorushellerii Regan. Ti jdschr. NederlandscheDierkund. Vereen., ser. 3, vol. 1, pp.77-80.

WILSON, H.V.1889. The embryology of the sea bass (Ser-

ranus atrarius). Bull. U. S. Fish Comm.,vol. 9, pp. 209-277.

WINGE, 0.1922. A peculiar mode of inheritance and its

cytological explanation. Jour. Genetics,vol. 12, pp. 137-144.

WOLF, L. E.1931. The history of germ cells in the vivip-

arous teleost Platypoecilus maculatus.Jour. Morph. Physiol., vol. 52, pp. 115-153.

Documents

EMBRYONIC DEVELOPMENT OF THE PLATYFISH …gato-docs.its.txstate.edu/jcr:ac37f118-98aa-4b1b-80f1-6828143efb11... · between Platypoecilus development and that of Xiphophorus and the