GA.METOGENESIS A.NO FK UTILISATION IN NlCMATUS MUESLI. 1 0 1

Gametogenesis and Fertilisation in Nematusribesii.

Li. Doncaster, M.A.,Late Mackinnon Student of the Royal Society; Lecturer in Zoology in

the University of Birmingham.

With Plate 8.

IN a previous paper1 I gave an account of the maturationand behaviour of the polar nuclei in several species ofsawflies which develop parthenogenetically.

In all these species there were two maturation divisions,giving rise to an egg nucleus and three polar nuclei, and insome cases fusion took place between the second polarnucleus and the inner half of the first. The egg nucleussank into the yolk and began to divide to form the embryo,while the polar nuclei in all cases ultimately disintegrated.Since whenever the chromosomes were clearly visible theirnumber appeared to be eight, both in the maturation mitosesand in the later divisions in body-cells, it was concluded thatno reduction in the ordinary sense took place. But iffertilisation ever takes place by conjugation of male andfemale pronuclei, an obvious difficulty arises with regard tothe chromosome number in fertilised eggs, and since theprocess of fertilisation had not been thoroughly examined atthe time when the paper referred to was written, it wasnecessary to leave the question open in the hope of finding asatisfactory answer later. This paper gives an account of

1 'Quart Journ. Micr. Sci.,' vol. 49, 1906, p. 561.

102 L. DONOASTER.

the work done on the fertilised egg in Nematus ribesiiand on the gametogenesis in that and other species.

The methods used were generally the same as before, butit was found that, in searching for male pronuclei in the eggsof impregnated females, thionin or gentian violet were moresatisfactory stains than iron hEematoxylin, since they stainnucleus and cytoplasm but leave the yolk uncoloured. Inthe work on spermatogenesis and the development of theovarian egg, osmic fixatives (e. g. Flemming's fluid) werelargely used in addition to sublimate.

THE FERTILISED EGG IN N. RIBESII.

In some animals, e.g. the bee, the fertilised egg is easilydistinguished from the virgin by the presence of sperm astersin the yolk, but in the sawflies nothing of the kind can befound, and over 200 eggs had to be cut and examined beforeit became certain that conjugation of male and femalepronuclei takes place. In very young eggs I had occasionallyfound minute rod-like bodies in the peripheral protoplasmnear the anterior end, which are probably the heads ofspermatozoa, and in somewhat later eggs bodies whichappeared to be degenerating nuclei sometimes appear in asimilar position. In eggs laid by impregnated females thereare frequently in the yolk in front of the polar region moreor less numerous small radiating patches of protoplasmwhich sometimes appear to contain indistinct nuclei, butprotoplasmic masses not certainly distinguishable from theseare found also in virgin eggs, although with less regularity.In eggs which are probably fertilised there are also fre-quently lines of protoplasm running inward from the edgeof the egg near the point where the spermatozoa had beenfound. But in no case have I been able to recognise withcomplete certainty the male pronucleus before the maturationdivisions of the egg are completed, and after that stagenuclei found in the yolk may always be derived from the

OAMETOGENESJS AND FERTILISATION IN NEMATUS RJBESII. 103

egg-nucleus itself. It is never possible, therefore, to saywith certainty that a given egg is fertilised or not.

But after much time spent in vainly trying to follow theentrance of the spermatozoon aud its conversion into theinale pronucleus, I at last was able to observe the conjugationof the sperm-nucleus with that of the egg, and so to provethat true fertilisation does take place (fig. 1). It occursimmediately after the maturation divisions; the three polarnuclei lie near the edge of the egg (two of them in the samesection as the egg and sperm nuclei), and the fusion of thetwo inner polar nuclei has not yet taken place. The maleand female pronuclei are in contact, the male being distinctlysmaller than the female, but in another egg in which thesame stage is seen the two are of about equal size. Thesubsequent stages of the conjugation and division of thezygote nucleus have not been observed, but the sectionrepresented in fig. 1 leaves no reasonable doubt that normalconjugation takes place. It therefore became necessary toreconsider my previons conclusions with regard to the numberof chromosomes, since never more than eight have beenfound in either fertilised or virgin eggs. I was thus led towork out the spermatogenesis, and to the fresh work on thematuration divisions to be described later.

In my previous paper I mentioned that the behaviour ofthe polar nuclei appeared to be slightly different in fertilisedand in virgin eggs, and subsequent work has confirmed this.

In the virgin egg of N. ribesii the two inner polar nucleifuse and give rise to a group of chromosomes, which isgenerally clearly double, with eight in each half. The twohalves of the group do not lie far apart, and commonlyremain without much change for some time. But in themajority of eggs from impregnated females the chromosomegroups derived from the two inner polar nuclei lie completelyand sometimes widely separated, as if the conjugationbetween the nuclei had been much less complete than invirgin eggs (figs. 2, 3, and 4). Further, in virgin eggs thepolar chromosomes usually do not divide, at least for some

104 L. DONOASTEB.

time, but in fertilised eggs they frequently divide compara-tively early, giving groups containing as many as sixteenchromosomes rather irregularly arranged in the " polarprotoplasm." That this difference in behaviour is reallyconnected with fertilisation is made probable by the factthat it rarely, if ever, occurs in eggs which are certainlyvirgin, but in the eggs laid by impregnated females it isfrequent. Further, in several eggs laid by impregnatedfemales the polar nuclei follow the typical virgin arrange-ment, and in these the little rayed protoplasm masses in theyolk, characteristic of fertilised eggs, are absent; butother eggs laid by the same female have the fertilised typeof polar chromosomes, and in these the rayed protoplasmpatches are also present.

It appears, therefore, that the fertilisation of the egg-nucleus, or the presence of spermatozoa in the egg, in someway influences the behaviour of the polar nuclei.

SPERMATOGENESIS.

When it had been shown that normal fertilization couldtake place in N. ribesii, it became necessary to re-examinethe maturation divisions in order to make certain about thechromosome number, which I asserted in the previous paperto be eight both in the maturation and in the somaticmitoses, and also apparently in fertilized eggs. The matura-tion of the egg begins immediately after it is laid, so that itis very difficult to get good preparations of the early stages,and I therefore decided to examine the matter first iu thedevelopment of the spermatozoa.

In very young male pupae, shortly after the larval skin iscast in the cocoon, the testes consist of compact groups ofcells at the sides of the alimentary canal. These cells(spermatogonia) have relatively large nuclei containing aconspicuous nucleolus (plasmosome) and eight or about eightchromatin masses apparently attached to the nuclear mem-brane (fig. 5). Division figures are scarce, but when found

GAMET0GENES1S AND FERTILISATION IN NEMATUS BIBESH. 105

they show clearly about eight rather large chromosomesin the equatorial plate, which split so that eight traveltowards each centrosome (fig. 6 a, b). At a later stage thetestis becomes larger, and consists of lobes or compartmentsin each of which all the cells are in about the same stage.By the time the colours of the mature fly are beginning toappear the testis contains nothing bnt spermatids and nearlymature spermatozoa, but when the pupa is still white allstages from spermatogonia to spermatids are found in dif-ferent lobes, often in the same section.

In the nucleus before the first maturation divisions thechromatin consists of a number of irregular masses (appa-rently about eight, but they are always rather indistinct).Shortly afterwards it becomes condensed into four moreconcentrated masses, each of which frequently appears doubleor quadruple (fig. 7 a,b, c). A spindle is then formed, andthe four chromatic masses become tightly packed together inthe equatorial plate, which is much smaller than in thesperniatogonial divisions. There are conspicuous centro-somes. The chromosomes in the spindle are so tightly packedtogether that it is difficult to be certain of their number, buta comparison of many mitoses leaves little doubt that thereare four, each of which is bivalent (6g. 8 a, b). The mitosisappears to be of the heterotype form, resembliug the figuresfound by Moore in the cockroach 1 except that the chromo-somes are fewer and very much smaller (fig. 9 ft, b). Theyare, however, appreciably larger than the chromosomes in thematuration mitoses of the egg.

The second maturation division is easily distinguished fromthe first by the fact that the spindle is of about half thediameter; the chromosomes are usually even more tightlypacked, so as frequently to appear as a siDgle body, but inclearer cases there is little doubt that there are four (fig.10 a, b, c). At the telophase a vesicular spermatid nucleusis formed, with the chromatin arranged round the edge

1 ' Quart. Journ. Micr. Sci.,' vol. 48, 1905, pp. 489 and 571,

106 L. DONCASTJill?.

giving it a characteristic appearance (6g. 11). This becomesconverted into the head of the spermatozoon.

It must be concluded therefore that in the male the normalsomatic number of chromosomes is eight; that four " gemini "appear iu the prophase of the first maturation division, andthat finally four chromosomes are distributed by heterotypeand hoinotype divisions to each spermatid nucleus.

There is no trace of the " polar body " formation describedby Meves in the spermatogenesis of the bee.3

OOGENESIS.

Iu the larva before it casts its skin within the cocoon theovaries are much like the testes of the male, but larger, withbigger nuclei. The ovary is enclosed in a cellular sheath,and some ovarian cells are already larger thau others ; thesewill form the eggs, while the more numerous smaller cellsgive rise to the nutritive and probably to the follicle-cells.All the nuclei at this stage contain about eight chromatinmasses and one to three nucleoli (fig. 12). In the youngpupa the egg tubes are already differentiated, and in alongitudinal section of a tube the changes in the nucleus caneasily be followed. At the apex of the tube the nuclei arelike those in the larval ovary; below this zone the chromatinbecomes distributed through the nucleus as fine dots, whichare often aggregated together in one part, as in a sort ofsynapsis (fig. 13). The egg nucleus then enlarges consider-ably, and the chromatin appears as an irregular thread; atthis stage two or three nucleoli are generally conspicuous(fig. 14). After this stage yolk begins to be deposited, andbefore the egg is ripe the nucleus, which has been verylarge, dwindles so that in nearly ripe eggs I have beentotally unable to find it.

In the larval ovary mitoses may be found in the ovariancells and in the sheath; those actually in the ovary appear tohave eight chromosomes (fig. 15 a, b). But in the sheath in

3 ' Anat. Anzeiger,' xxiv, 1903, p. 29.

GAMETOGENESIS AND FERTILISATION IN NEMATDS El BESII. 107

all the mitoses observed the number is more than eight;usually it seems to be sixteen, but in some cases the figuresuggests more than sixteen very small chromosomes (fig.16 a—/). Wilson1 has described spindles with double thesomatic number in the ovary-sheath of Hemiptera, andregards them as abnormal,but the figures seen in N. r i b e s i icertainly suggest that the eight chromosomes in the primi-tive germ-cells are compound, composed of a greater numberof smaller units, possibly more than sixteen.

In the pupal ovary the egg-cells are already definitelyformed, and do not divide further, but merely undei'go theusual growth with, deposition of yolk. The follicle cells arenow quite small, and an occasional mitotic figure is visible;these are rarely clearly defined, but appear to have eightchromosomes. When the egg has reached its full size thefollicle cells become degenerate, with obscure dark-stainingnuclei. Groups of similar degenerating cells are found hereand there in the larval ovary.

The fact that the chromosome number in the ovary, andprobably in the follicle cells, is smaller than that found in thesheath is of considerable interest.

In addition to the case described by Wilson, and referredto above, the same kind of thing has been observed byPetrunkewitsch in the bee,s in which he found the unreducednumber to be sixteen in the egg, but sixty-four in the blasto-derm, and it is still more conspicuous in Ascar i s , which,according to Boveri,has a large number of very small chromo-somes in the somatic cells, but only four in all the cells on the"germ-track," from the fertilised egg up to the maturationdivisions of the germ cells.3 These facts suggest that it mayhappen not infrequently that the chromosomes in cells of thegerm-track may be compound, and consist of a number ofsmaller units which become separated in somatic cells. But

1 " Studies on Chromosomes," iii, Mourn. Exp. Zoo.,' vol. iii, No 1, 1906.3 'Zool. Jahrb.,' vol. xiv, 1901, Anat. und Onto;*., p. 573.' Boveri, 'Ergebnisse fiber die Konstitution der Chromatisclien Substanz

des Zellkerns.' (Fischer, Jena.)

108 J,. DONOASTELJ.

even if this is found to be a phenomenon of general occurrenceit does not necessarily affect the hypothesis of the individualityof the chromosomes in any essential point.

CHROMOSOMES IN THE MATURATION DIVISIONS OF THE EGG.

It has now been shown that in the spermatogonial andoogonial divisions there are eight chromosomes, and that inthe spermatocytes these are reduced to four in the normalheterotype manner. These facts led me to re-investigate thematuration divisions of the egg, since in my previous paper(loc. cit.) I gave evidence that in both first and second polarmitoses the number was eight.

The chromosomes in the maturation of the egg are muchless easy to observe than in the spermatogenesis, for thereare difficulties of technique to be overcome, and the egg hasto be preserved at exactly the right moment. But aftercutting some hundreds of eggs I have been able to convincemyself that while there are two types of maturation. In someeggs no reduction takes place, and eight chromosomes passinto each of the four nuclei produced by the polar mitosis.In other eggs four double chromosomes are found in theequatorial plate of the second maturation division, and theseseparate into their component halves sending four into eachdaughter-nucleus (figs. 17—21).

I have never obtained a sectiou of the first polar mitosis inwhich it is quite certain that there are four " gemini," althoughsome figures strongly suggest this ; but at the close of thefirst division, when the chromosomes are arranging themselvesto form the equatorial plate of the secoud mitosis, four doublechromosomes are sometimes clearly visible (figs. 20, 21). Ihave also several preparations which show only four when thesecond polar mitosis is already begun. A comparison of figs.17 and 19, 18 and 20, respectively, will show the differencebetween the reducing and equational types of maturation.

It must therefore be concluded that in some eggs pairing

GAMETOGENESIS AND FERTILISATION IN NEMATUS E1BESII. 109

(synapsis) of chromosomes takes place before the maturationdivisions; resulting in the separation of complete chromosomesat one of the mitoses, while in other eggs no pairing takesplace, and each chromosome undergoes two equational divi-sions. In connection with this it is noticeable that in theeggs having the equational type the eight chromosomes areabout half the size of the four seen in reduced eggs.

I have found the reduced type in eggs from both virginand impregnated females, so that the view which first sug-gested itself, viz. that reduction only takes place in eggswhich contaiu spermatozoa, is not tenable.

Re-examination of my sections of Pcecilosorna lu teo lumconfirms me in the belief that in that species, which yieldsfemales from virgin eggs, and is normally not fertilised, thereare two equational divisions in all the eggs of which I havesuitable preparations.

In the developing egg the somatic mitoses of fertilisedeggs appear always to have eight chromosomes; a largernumber has never been found. This is what would beexpected if only eggs which undergo reduction are capableof fertilisation.

Tn virgin eggs commonly eight are found, but in some casesthe equatorial plate seems to have four only, showing thatreduced eggs when not fertilised can develop as far as theblastoderm stage (fig. 22). The number of eggs which diebefore hatching varies, in some batches beiug very small, inothers more considerable ; it is possible that the reduced eggsare those which fail to develop to larvse. Since, however, ithas been shown by the mitoses iu tbe ovary sheath that thechromosomes are possibly compound, it may happen thatreduced eggs which are not fertilised restore the normalnumber of chromosomes by division of the compound chro-mosomes, as was asserted by Petrunkewitsch (loc. cit.) withregard to the bee.

The conclusion that the eggs of one species may eitherundergo reduction, or may retain the full number of chromo-

110 I.. DONCASTER-.

somes, although in each case there are two polar mitoses, isof considerable interest. I know of nothing quite parallelwith it hitherto observed in animals, but I think it notunlikely that in the two generations of the Gallflies, one ofwhich is bisexual and the other purely female, a similar stateof things may be found to exist. That there may be twotypes of egg, one of which is reduced and requires fertilisa-tion, and the other not reduced and parthenogenetic, is ofcourse not infrequent, but in such cases the eggs generallyhave obvious external differences, and the unreduced formhas only one polar body. A condition more nearly resem-bling that found in N. r i b e s i i has been observed by Eosen-berg in Hieracium,1 in which the egg-cell in some flowers ona head is reduced and can be fertilised, in others on the samehead not reduced and parthenogenetic. But here again thenumber of maturation divisions is probably not the same inthe two cases. In the bee, according to Petrunkewitsch, allthe eggs ai-e reduced, but if not fertilised, the somaticnumber of chromosomes is restored automatically.

The conclusions here reached may make it necessary toreconsider the provisional hypothesis of sex-segregationsketched in my previous paper, but until further facts areobtained in other species it seems premature to discuss thebearing of my results on the problem of the determinationof sex. I have not found it possible, owing1 to the minutesize of the chromosomes, to determine whether anythingcomparable with Wilson's " heterotropic " chromosome existsin Nematus . In some figures (e. g. the group representedin tig. 18) only seven chromosomes are visible instead ofeight, but when they are so minute it is always possible thattwo are superposed and not distinguishable apart.

Iu conclusion I take this opportunity of expressing mygratitude to Mr. J. E. S. Moore for allowing me to comparesome of my preparations with his, and for valuable help inelucidating my sections.

[NOTE.—In a series of eggs all laid by one insect on one1 Brit. Ass., York, 1906. Discussion ou Fertilisation, Seels. D and K.

GAMETOGENESIS AND FERTILISATION IN NEMATUS RIBES1I. I l l

day the polar mitoses are abnormal. The most extremecase (fig. 23) shows the "polar protoplasm," full of dotsarranged roughly iu lines like iron-filings in a magnetic field.At each pole of the figure is a group of more conspicuousstained bodies which may be chromosomes. Some of theother eggs show a somewhat similar appearance on a smallerscale/and in others nothiug is clearly distinguishable in thepolar protoplasm. In all the eggs the peripheral protoplasmis narrower than usual, and in the most markedly abnormaleggs it is practically absent. I have occasionally foundappearances of the same kind, but much less pronounced,in eggs laid by other insects, but have not sufficient cases tobe able to throw any light on their meaning.]

SUMMARY.

1. True fertilisation (conjugation of male and femalepro-nuclei) may take place in N. ribesii, and the behaviouro£ the polar nuclei is slightly different in fertilised and virgineggs.

2. In the spermatogenesis there are eight chromosomes inspermatogonial divisions; four " gemini" appear at thebeginning of the maiotic phase, and by heterotype and homo-type mitoses distribute four chromosomes to each spermatid.

3. In the oogenesis eight chromosomes appear in oogonialmitoses, but in divisions of nuclei in the ovary sheath morethan eight are found, suggesting that the chromosomes ofthe germ-cells are compound.

4. In the polar mitoses of the egg two types of maturationare found. In some eggs there are successive equationdivisions so that the egg nucleus and each of the three polarnuclei contains eight chromosomes. In other eggs normalreduction takes place, separating entire chromosomes fromone another, and only four are found in each of the daughternuclei.

5. It is probable that only such reduced eggs are capable

112 L. DONCASTBE.

of fertilisation, but when unfertilised they may coutiuue todevelop at least as far as the blastoderm stage.

Birmingham Universi ty;November, 1906.

E X P L A N A T I O N OF P L A T E . 8,

I l lus t ra t ing Mi". L. Doncaster 's paper on " Gametogenesisand Fert i l isat ion "in N e m a t u s r i b e s i i . "

All figures are drawn with an oil-immersion lens, but are not exactly onthe same scale. Those illustrating spermatogenesis are more highly magnifiedthan the remainder. All represent Nematus ribesii except figs. 12, 13,14.

]?IG. 1.—Conjugation of male and female pronuclei. Three polar nucleinear the edge of the egg.

FIGS. 2, 3, 4.—"Polar protoplasm" of fertilised eggs showing chromo-some groups derived from polar nuclei.

FIG. 5.—Nucleus of spermatogonium.FIG. G.—Spermatogouial mitoses, (A) Metaphase, side view; (B) Equ-

torial plate.

FIG. 7, A, J>, C.—Spermatocyte: three propliases of heterotype mitosis.(A) Showing 8 chromosomes; (B and c) Pairing to form & double chromatinmasses.

FIG. 8.—Heterotype mitosis, equatorial plate, (A) Pole view; (u) Sideview.

FIG. 9, A, B.—Heterotype anaphases.

FIG. 10.—Homotype. (A) Pole view of equatorial plate; (B, c) Anaphase,side view.

FIG. 11.—Spermatid.FIG. 12.—Young oogonium, N. lacteus.

FIGS. 13, ]*.—Stages of growth of oogonium, N. lacteus pupa.FIG. 15.—Oogonial mitoses, larval ovary, (A) Pole view; (B) Side view.

FIG. 16, A.—r.—Mitosis in ovary sheath witli more than 8 chromsomes.(A) Equatorial plate, pole view; (B, C) Similar stage seen from side andobliquely; (D, £, r) Anaphases.

GAMET0GENES1SAND FERTILISATION ]N NEMA'l'US RIBESII. 113

FIG. 17.—Second polar mitoses, equational type, with 8 chromosomes.

FIG. 18.—Second polar mitosis, equatorial plate in pole view, with 7 chro-mosomes, some preparing to divide.

FIG. 19.—Second polar mitoses, metaphase; reduced type, with 4 chro-mosomes.

FIG. 20.—Equatorial plate of reduced type, showing 4 double chromo-somes.

FIG. 21.—Stage between first, and second maturation divisions, reducedtype, with 4 double chromosomes each end.

FIG. 22.—Two blastoderm mitoses, each witli 4 chromosomes and con-spicuous centrosomes.

FIG. 23.—Abnormal polar mitosis.

VOI. 5 1 , PART 1. NEW SEKIES.

V #

* « •V \*«

14-.

- , - • « •

16.

a.

"̂ ff

21.! 'urn of !•.

on H a n

GAMETOGENESIS AND FERTILISATION IN NEMATUS RIBESII.