Interactions between Ovary and Lateral Oviduct in Drosophila melanogaster

In Drosophila the genital ducts arise from the so-called genital imaginai disk which in the larva lies at some distance from the gonads. In the male, interactions between these two components of the reproductive system are well known. The study of gynandromorphs of Drosophila simulans led Dobzhansky (1931) to suggest that the spiral form which the testes normally assume in the adult of these animals is only produced when the gonad makes contact with the genital duct. Stern (1941 a and b) investigated the matter by transplanting the testes and amply confirmed Dobzhansky’s suggestion, finding that the genital duct induces the testes to grow into a form which, surprisingly enough, is characteristic of the inducer and not of the reacting material. Further, Stern & Hadorn (1939) have revealed a reciprocal reaction by which the testis affects a duct to which it is attached in so far as it supplies to the latter pigmented cells which clothe it in a coloured epithelium. Hadorn & Bertani (1948) have shown that similar cells can pass on to the duct from an attached ovary and then form typical male pigment. We have, however, rather little information about interactions between the gonads and ducts in the female, although Hadorn & Gloor (1946) have shortly recorded the results of an investigation into the properties of the various parts of the female genital disk.

The aim of the present investigation was to detect any interactions there may be between the gonads and the genital ducts in the female, using the methods of extirpation of the ovaries or transplantation of ovaries or disks. All the experiments were carried out on late third instar larvae of various laboratory stocks of Drosophila melanogaster. The transplantations were made by the usual method of Ephrussi & Beadle (1935). Ovaries were extirpated by making a small incision at the sixth abdominal segment with a fine pair of forceps, pulling out part of the fat body containing the ovary and cutting it off. It is difficult but possible to remove the ovaries on both sides, but the percentage of survival following such a drastic operation is very small. The same technique has been applied for the removal of testes.

Eleven out of forty-four operated larvae reached the adult stage and were dissected about 2 days after hatching. The diagrams of Text-fig. 1 show the arrangement of ducts and ovary found. It will be noted that no lateral oviduct develops on the side on which there is no ovary left, although a rudiment of the lateral oviduct is clearly present. The uterus, oviduct, and the single lateral oviduct form a more or less straight tube. In two exceptional cases, in which both lateral oviducts developed, they were both joined to the single remaining ovary.

The conclusion was tentatively drawn that the presence of the ovary is a necessary factor for the growth of the lateral oviduct. The evidence was further strengthened by the following experiment.

Only two out of a small series of larvae, both ovaries of which were removed, hatched as imagos. Both lateral oviducts failed to develop. The rest of the genital duct system was normal.

Large numbers of larvae were thus operated and several stocks were used. The results showed that an additional ovary cannot cause the growth of an additional lateral oviduct, and either remains unattached or shares a lateral oviduct with one of the host’s own ovaries. In some cases the implant forms a compound ovary with one of the host’s gonads. In others it remains distinct and connects to the lateral oviduct directly. In two cases one lateral oviduct was found to be considerably longer than the other and joined to a compound of two ovaries. Presumably, as a result of the operation, the two ovaries on the operated side were displaced and the lateral oviduct grew out to them. This is a further confirmation of the reality of the influence of the ovary on the development of the lateral oviduct. The arrangements obtained are shown diagrammatically in Text-fig. 2 and their incidence in a typical series were as follows:

In the above list only the cases where the implanted ovary achieved attachment are included. In about one-fourth of the hatching imagos, however, the implanted ovary was either not found or was unattached and smaller than the attached ones. Also, where all three ovaries were attached, not all were of the same size, usually the two on the same side remaining smaller than the other and also unequal among themselves. It seemed doubtful, however, whether attachment to the oviduct is necessary to the ovary’s growth, and the small size of the implant in many cases could be attributed either to damage during the operation or to competition. That the presence of the lateral oviduct is not necessary for ovary growth was confirmed by the results obtained by the following transplantation experiment.

In most out of twelve operated males which hatched, the ovary was found in the abdomen, with differentiated egg strings but of small size. In one case, however, the ovary had become attached to the male duct, excluding the testes from it, and had grown to full size as shown in Text-fig. 3. The testes remained rudimentary and uncoiled, but the accessory glands were normal. In section sperm bands were seen in the testes, and the sheath of the accessory glands or ‘para-gonia’ showed the typical binucleate cells. In another specimen too the implanted ovary had grown to normal size, but was not connected with the ducts. The testes were normal in size, but their coiling was atypical.

It appeared possible that, although as Stern showed the vas deferens regulates testis growth after it establishes connexion with it, the testis might, as the ovary proved to be doing in the female, exert an effect on the growth of the vas deferens so that it reaches the testis. By removing one of the testes it was found that the vasa deferentia had grown normally on both sides, and actually that the one on the operated side ended in a thin filamentous structure. In several cases the remaining testis showed atypical coiling or reduced growth. In one case the accessory gland was also missing on the operated side.

It appeared then that the testis exerts no influence on the growth of the vas deferens. It could still be the case, however, that such an influence is necessary, and that in this experiment it was exerted on both vasa by the remaining testis. The removal of both testes was therefore attempted.

It proved difficult to achieve a reasonable survival rate in this experiment and only one adult hatched. Both vasa deferentia were normal, although there was no testis present (Text-fig. 4). The filamentous structures at the end of both vasa were large, and in fact appeared to be tubular.

Since only one specimen was obtained in the previous experiment, it was attempted to obtain more genital ducts developing without attached testes by transplanting male genital disks into male host larvae. A small series of such operations was performed and four hatching imagos were obtained. In three of them the implant had given rise to a complete system of genital ducts and (inverted) external genital plates. Again the vasa deferentia1 were present and ended in a filament. In the fourth specimen the segment carrying the host’s external genitalia was abnormally elongated. On dissection, one of the host’s accessory glands was found to be reduced in size and the testis on that same side was rudimentary. The implant was also poorly developed; the genital duct was not dilated at the distal end and there were no vasa deferentia or accessory glands.

This experiment confirmed the conclusion that the vasa deferentia develop independently of the testes.

The evidence shows that the ovary exerts an influence governing the growth of the lateral oviduct on the corresponding side. If a similar influence on the other side is missing, the remaining ovary may affect both lateral oviducts. It does not appear that the lateral oviduct exerts an influence on the growth of the ovaries. The size of any supernumerary ovary, experimentally introduced, is decided by competition. It seems likely that if the implant comes from a different stock, genetic factors might favour the ovary or ovaries of one genotype compared with the other.

These phenomena contrast with those in the case of the male, where (a) the testis does not influence the growth of the vasa deferentia, and (b) once connexion between these structures is established, the vas governs the growth of the testis. Stern attributed this effect to a substance produced by the vas and diffusing to the attached testis. In this connexion too there is a difference in the case of the female. If one assumes that the influence of the ovary is exerted via a diffusible substance, this must be reaching the duct without a direct attachment between the two structures.

The author is indebted to Professor C. H. Waddington for his interest and suggestions during the course of this work, and to the Melville Trust for a Research Fellowship.