1. A technique is described for the culture of germ-free Drosophila adults on defined diets. The complete larval diet, with the agar base replaced by cotton-wool, was found to be adequate for adults and permitted them to lay more or less normal numbers of eggs during a 16-day test period.

  2. Omission tests showed that casein, the B vitamins (other than B12 and biotin) K and Mg were essential for normal ovary development.

  3. Casein could be replaced with an amino acid mixture. The ten ‘essential’ amino acids were all found necessary for egg production but arginine, histidine and methiohine were apparently synthesized, although at an inadequate rate. The remaining essential amino acids seemed to be the nutrients required in the greatest amounts since egg production stopped soonest when they were omitted. Individual ‘nonessential’ amino acids could be removed from the mixture, but a supply of them was necessary for normal egg production and viability.

  4. Omission of fructose lowers egg production but does not cause its cessation even after 16 days.

  5. The difficulty of determining that essential supplies are not being met by contaminants is illustrated by examination of cholesterol requirements. In this case, and possibly also for choline, the requirement for egg formation is very much less than for larval development, and might be satisfied from contamination of the medium constituents.

  6. Neglecting these trace supplies (which can be measured only in fractions of a microgram per 5 ml. medium) RNA, choline, cholesterol and biotin seem unnecessary for egg formation. The quantitative requirements of the normal adult female must therefore be different from those of the larva.

  7. It was not possible to produce a true P deficiency, or to be certain that traces of Ca and Cl in the medium were not sufficient to permit normal fecundity.

  8. The ovary is shown to be capable of recovery from protein starvation and to respond to omission of single essential amino acids by ceasing to form new chambers. The majority of deficiencies result in inhibition of vitellogenesis, after a period when inviable eggs are laid. Only Mg and pyridoxine omission produced distinguishable pathological changes, which are illustrated.

Although much has been done towards defining the nutritional requirements of immature forms of insects (Dougherty, 1959), very little attention has been paid to the needs of adults. Adults of many species can subsist when fed only simple sugar solutions, and this finding has led to many interesting studies of the suitability of various carbohydrates for the maintenance of adult life (see House, 1961) but also, apparently, to neglect of the examination of the requirements for full adult existence, which in the female involves the production of normal numbers of viable eggs. It cannot be assumed that larval and adult nutritional needs are the same, particularly for a species which inhabits different environments as larva and adult. Further, since ability to produce eggs (and sperm) is of primary significance for survival, it is important that the relationship of nutrition to egg production should be studied more extensively than it has been. The purpose of this paper is to do this for the fruit fly (Drosophila melanogaster) whose larval nutritional requirements have already been defined (Sang, 1956, 1959).

Only one of the many previous studies of adult nutrition was carried out using axenic culture techniques (Singh & Brown, 1957), and this showed that the anautogenous Aedes aegypti required ‘the 10 indispensable amino acids of mammalian nutrition’ in order to lay eggs. The vitamins, salts and other nutrients essential for larval growth and development were not needed for egg formation under the conditions of this test. Unfortunately, egg production was measured only during a relatively short period and was at a low level (nine eggs per female in 6 days) so that it is unlikely that stores carried over from the larval stage had been depleted. That such stores do exist in A. aegypti is evident from the work of Dimond, Lea, Hahnert & de Long (1956) who found that a dietary supply of histidine and methionine was not essential for the production of the first few eggs laid by females which had been reared under non-aseptic conditions but fed a defined diet as adults. When these stores were consumed an essential requirement for the two amino acids could be exposed. It was also found that cystine was necessary for a high level of egg production. It follows that similar reserves of vitamins and salts, etc., may have been sufficient to allow the production of the small number of eggs reported by Singh & Brown (1957), and that their conclusion that ‘vitamins are not required in adult nutrition’ is likely to be true only under the conditions of their tests.

Harlow (1956), who has reviewed much of the literature on adult nutrition, was able to provide evidence for the conversion of the material of the pupal to the adult fat body, and its subsequent transfer to the ovaries at the time of yolk formation in Protophormia terrae-novae. By itself this transfer was not sufficient to permit egg formation when sucrose was supplied for the maintenance of the flies, or if protein was also added to the diet. Accessory factors, thought to be B vitamins and salts, had to be fed along with the protein for eggs to be laid. So there is evidence of a range of adult nutritional types extending from autogenous species (of which D. melanogaster is one) which will lay some eggs without being fed as adults, to anautogenous species which have apparently simple requirements (like A. aegypti) and to anautogenous species which have complex requirements (like P. terraenovae). Egg production is not entirely a matter of nutrition since it is also dependent on hormonal control (Clements, 1956); but in so far as fecundity is regulated by nutrition, the assessment cannot be made dependent on the production of some eggs. Tests of adult nutritional needs must be carried through under conditions which allow more or less normal levels of egg production, and these tests must be extended in time so as to ensure depletion of stores which may have been accumulated during the previous feeding stage. Further, the eggs laid must be viable, since it is possible that eggs may be laid which are none the less deficient in some normal constituent. An attempt has been made to satisfy these requirements in the experiments herein described.

One further difficulty must also be noted. In Harlow’s (1956) work, addition of all the likely B-vitamins and salts to the diet could not substitute for the ‘accessory factors’ essential for egg formation; some limiting substance or substances not being provided, egg production could not take place. It follows that tests of requirements must start from a complete diet which gives normal production of viable eggs, and proceed by omission of individual constituents. Some earlier experiments have been conducted in the reverse fashion, by finding the effects of additions to an incomplete diet, and in such cases negative results may have no significance if some limiting substance was not provided in the basal diet or along with the additive. Again, since the biochemistry of egg formation is clearly a dynamic process and may involve conversion of one nutrient to another, it cannot be assumed that the minimal diet defined by omission tests is in itself adequate for egg production. It is necessary to test this diet and to find if anything must be added to balance the minimal food. In so far as possible, this has also been done. Whenever a dietary omission apparently affected ovarian development, the ovary was examined cytologically.

Some of the technical problems of examining adult nutritional requirements have been indicated above. Axenic larvae have to be grown to provide germ-free adults, and two problems which arise in this connexion relate to the influence of nutrition of these larvae upon the accumulation of larval reserves and to the influence of the genotype of the flies upon egg storage, fecundity and viability. A third problem is that of developing a simple technique for feeding adults so that they produce approximately normal numbers of eggs. The exploratory experiments relevant to these points will not be described in detail, but the reasons for the procedure adopted will be indicated, and the routine finally developed will be outlined.

(a) Larval culture medium

Fourteen nutrients must be supplied to the Drosophila larva if it is to develop into an adult (casein, cholesterol, lecithin (which may be replaced by choline), thiamine, riboflavine, nicotinic acid, pyridoxine, pantothenic acid, folic acid, magnesium, sodium, potassium, phosphorus and chlorine. Addition of biotin, ribose nucleic acid (RNA) and fructose accelerates larval development and reduces mortality; while any trace requirements are apparently met by contaminants from the other constituents. The optimum supplies of most of these constituents have been determined (Sang, 1956), but the usual larval diet (Sang’s medium C, ibid.) contains excess of the vitamins and salts (since this arrangement simplifies preparation of the medium and has no effect on development, whereas excess quantities of the other constituents retard development). As judged by the amount of yolky ooplasm laid down by adults on a protein-free diet, the nutrient reserve of larvae fed on this medium C is about twice that of larvae fed a modified medium in which the vitamins are supplied only in 50 % excess of the determined optimum. For this reason, larvae for experiments were reared on this less efficient medium, which is designated medium D in the Appendix.

(b) Egg storage and genotype

It is well known that copulation stimulates Drosophila to oviposit. Males have therefore to be supplied in the adult cultures, and an attempt was made to keep their numbers equal to those of the females throughout the tests. If males died, and could not be replaced, the culture was discarded.

Two highly inbred wild-type strains (Oregon S and Crianlarich-6) and individuals resulting from the reciprocal crosses between them were tested for the effects of protein deficiency on oogenesis (King & Sang, 1958). At the end of 7 days, females from all strains had synthesized 8–11 eggs. However, both the Crianlarich-6 and the hybrids laid most of their eggs at once, whereas Oregon S stored the majority of the eggs in their ovarioles. Again, a test of these three strains on a Mg-free diet produced abnormalities in the ovaries of the hybrids first, indicating that they depleted their stores of this substance more rapidly than flies from the pure lines. Hybrids have a higher fecundity than the pure lines from which they are derived (Gowen & Johnson, 1946), and tests showed that the eggs laid by hybrids were also more viable. Hybrid adults also survived better under the test conditions. For these reasons, most of the work was done using Oregon S/Crianlarich-6 reciprocal hybrids.

(c) Adult culture medium

The larval food is set in a gel with agar. This method cannot be used for adults since they do not feed efficiently on an agar gel and because the agar is heavily contaminated, particularly with Ca and Mg salts. An attempt was made to replace the agar with 15 % purified starch. Egg production was low in females fed under these conditions, and after 2 weeks their ovaries were abnormal in appearance, showing fusion of adjacent chambers, pycnotic areas in the germaria and abnormal development of the follicle cells surrounding the chambers nearest the germarium. Evidently the females were unable to extract certain nutrients from the starch gel at a rate sufficient for normal oogenesis.

Among the many substrates tested cleaned, non-absorbent cotton-wool proved to be the best vehicle in which to supply the experimental media. Cleaning was performed by soaking the cotton in 1 % EDTA overnight, washing twice with distilled water, and soaking overnight in alcohol and again in ether. It was found that 0-3 g. cotton would then absorb 5 ml. of medium C (in this case prepared without agar and supplemented with Mg) and that females kept on this medium had a reasonably high productivity (Table 1). Further, an increased concentration of certain nutrients in the adult diet caused a lowering of egg production (Table 1) just as the same increase would have depressed the rate of larval development. The complete larval food medium (medium C) in a cotton base was therefore taken as the standard when assessing the effects of a diet. Ashless floc was used as a base in some salt omission experiments, although it was less satisfactory than cotton since its surface tended to dry out quickly.

Table 1.

Egg production and egg viability of females kept for 4 days on medium C

Egg production and egg viability of females kept for 4 days on medium C
Egg production and egg viability of females kept for 4 days on medium C

(d) Technical routine

Eggs were collected from mass matings of Oregon S x Crianlarich-6 flies, sterilized, and handled as described by Sang (1956). The hybrid larvae which hatched 24 hr. later were transferred in lots of 25 to 2·5 ml. of medium D (see Appendix), previously gelled and autoclaved in ordinary test-tubes which were covered with cylindrical metal caps. The adults which emerged exactly 9 days later were shaken into 6 × 1 in. boiling tubes containing the test diet absorbed in cotton-wool plugs and having a folded strip of ashless filter-paper above the plug to prevent the flies from sticking to the wet cotton. Egg production of individual females was usually measured after 4, 8 and 16 days on the test diet. By 8 days larvae were present in large numbers in the control cultures and, as larvae would interfere with adult feeding (Robertson & Sang, 1944), all flies were transferred to fresh food on the 8th day. It should be noted, however, that this simple procedure had one disadvantage : dietary deficiencies which depressed female productivity tended to leave adult feeding conditions less disturbed through larval tunnelling of the medium than were those experienced by control flies which were characterized by high productivity. Consequently, the 8- and 16-day-old flies from different experiments do not always come from strictly comparable environments. On the other hand, the condition of the developing larvae in the feeding tubes provided a check on their nutritional state and hence of the germ-free condition of the diet provided.

Studies of egg production were initiated by sucking the females from the feeding tubes into 3 × 1 in, vials, which were stoppered subsequently with corks each capped with a disk of charcoal-blackened agar (containing 2 % ethanol and 1 % acetic acid as oviposition stimulants). Each fly was then allowed to oviposit in the dark for 24 hr. at 25° C. (the temperature used for adult and larval culture). The females were then removed, the eggs counted, and the developing embryos given at least a further 24 hr. incubation, after which unhatched eggs were scored. No attempt was made to keep conditions aseptic, and therefore no food was provided to the ovipositing females, since it was found for flies fed ad lib. during egg collection that at least 15 % of eggs laid during 24 hr. were formed from food consumed during that same period. Egg counts from typical experiments are given in the tables.

(e) Cytological examination

Experience quickly showed that the histological appearance of ovaries remained normal so long as females were laying eggs. Consequently females were sacrificed only after they had stopped laying. Feulgen-stained whole mounts were then prepared according to the procedure of King, Burnett & Staley (1957), except that the staining time was standardized at 1 hr. The whole mounts were examined at magnifications ranging from × 6 to × 1400. Abnormalities looked for included modification of the stage distribution of the developing oocytes; pycnosis of oogonia, follicle cells, nurse cells or oocytes; fusion of adjacent chambers; inhibition of yolk synthesis as shown by the absence of oocytes in stages 8–11 ; subnormal replication of Feulgen-positive material in nurse cell nuclei; atypical numbers of chambers per ovariole; abnormal differentiation of follicular epithelium; and abnormalities in the shape of maturing oocytes or in the production of the chorion. Where reference is made to stage of development of an oocyte, the classification of King, Rubinson & Smith (1956) is used.

(a) Major food substances

Of the five main nutrients required by D. melanogaster larvae, three are essential for development (casein, cholesterol and lecithin or choline). Surprisingly enough, omission of cholesterol or choline from the adult diet has no significant effect on egg production under the conditions of the test (Table 2). Omission of fructose, which only slows larval development slightly, does reduce egg production, and it should also be noted that by 16 days the great majority of adults died on this otherwise complete diet. Omission of RNA had only a small effect on productivity (see also King & Sang, 1959), whereas it has a more striking effect on larval development than the omission of fructose. That is, adult survival and production of eggs is adversely affected under our conditions by removal of fructose from the diet, whereas cholesterol, choline and RNA may be omitted without notable detriment to the functioning of females, and this would not have been anticipated from their effects on larval development. However, it should be noted here that in these experiments contamination of other nutrients by these substances has not been eliminated (see p. 805), but the amounts present must have been very small since they could not support larval development.

Table 2.

Effects of omission of major nutrients upon fecundity and egg viability

Effects of omission of major nutrients upon fecundity and egg viability
Effects of omission of major nutrients upon fecundity and egg viability

As was anticipated, omission of casein rapidly leads to cessation of egg formation, although 8–12 eggs were laid by females during the first few days on the deficient diet. By the end of a week no oocytes in stages 8–11 are found in the ovary. If the females are then fed live yeast, vitellogenesis restarts immediately and about half the ovarioles contain stage 8–11 oocytes within 24 hr. By 3 days females are laying normally (30 eggs per female day), and they continue to do so for at least 2 weeks. The first dozen or so eggs laid are inviable but subsequent eggs are normal (90% hatch). After 16 days on a protein-deficient diet no oocytes are more advanced than stage 3, many chambers are pycnotic, and recovery is much slower. Of eight females examined after subsequent feeding for 3 days on live yeast, four had no oocytes more advanced than stage 6 and, in all, only three stage-14 oocytes were found. The importance of a protein supply for continued egg production by Drosophila is evident from these observations, as is also the ability of the flies to recover from any temporary deprivation.

(b) Amino acids

Replacement of casein by amino acid mixtures led to many difficulties, for the most part apparently due to contamination of these acids by toxic substances. Eventually, reasonable results were obtained using Sigma brand amino acids, but even with these there is some evidence that removal of particular amino acids (e.g. serine in Table 4) may result in better egg production. There is also a tendency for amino acid to crystallize on the surface of the cotton in a way which would presumably affect adult feeding. However the levels of egg production achieved were sufficient to disclose the effects of omission of single amino acids, all of which were supplied in the natural form.

Table 3 shows in a striking fashion that the ten amino acids essential for the rat are also essential for the adult Drosophila female. Omission of any one leads to a reduction of egg output even in 4 days, and to a lowering of the viability of the few eggs laid. Elimination of methionine, arginine, or histidine, even for 8 days, does not result in a complete cessation of egg production, suggesting that the former can be met in part by synthesis of methionine from cystine, and that arginine and histidine may also be slowly synthesized, as has been found with other organisms (Meister, 1957). After 8 days of feeding on diets lacking single ‘non-essential’ amino acids, no reduction of egg output is evident, nor is infertility notably increased (Table 4). Some significant differences appear by 16 days: notably, the lowered fecundity on cystine-deficient diets and on diets without aspartic or glutamic acids. The first observation might be explained by assuming that the amount of methionine provided in the essential amino acids is insufficient to meet requirements by conversion to cystine (Meister, 1957), and the heightened infertility of the eggs laid would support this assumption. Effects of omission of aspartic or glutamic acid might be a consequence of insufficiency of these non-essential amino acids which are involved in the formation of such key metabolic intermediates as a-ketoglutarate. Direct proof of this possibility would depend on quantitative substitution tests which have not been attempted. It is also interesting to note that elimination of tyrosine affects only hatchability, implying that this substance is synthesized at an inadequate rate from the phenylalanine supplied.

Table 3.

Effects of omission of ‘essential’ amino acids upon fecundity and egg viability

Effects of omission of ‘essential’ amino acids upon fecundity and egg viability
Effects of omission of ‘essential’ amino acids upon fecundity and egg viability
Table 4.

Effects of omission of ‘non-essential’ amino acids upon fecundity and egg viability

Effects of omission of ‘non-essential’ amino acids upon fecundity and egg viability
Effects of omission of ‘non-essential’ amino acids upon fecundity and egg viability

These consequences of the removal of particular non-essential amino acids from the diet imply that essential amino acids alone should not be sufficient to give normal egg production, contrary to Singh & Brown’s (1957) finding with Aedes. That this is so is shown in Table 5. Addition of glycine improves output, demonstrating that a non-essential amino acid is necessary for good egg production. The further addition of glutamic acid results in no eggs being laid, but the ovaries of flies from this third group appear to be normal, and contained stage 14 oocytes. The addition of cystine then releases this inhibition of egg laying, indicating the complexity of the balance of non-essential amino acids required for full egg production. This problem was not explored further since it clearly involved quantitative as well as qualitative aspects of the dietary supply of both essential and non-essential amino acids, and these could not be adequately handled with the techniques described.

Table 5.

Effects of additions to the mixture of essential amino acids upon fecundity and egg viability

Effects of additions to the mixture of essential amino acids upon fecundity and egg viability
Effects of additions to the mixture of essential amino acids upon fecundity and egg viability

(c) Vitamins

All the B vitamins, except possibly B12, are necessary for optimal larval growth (Sang, 1956), and the same qualitative requirement would be expected for egg production. In fact, sufficient larval reserves are carried into adult life to permit the laying of a considerable number of viable eggs, so that the adult requirement is clearly dependent on the status of the larval supply of each vitamin. Under the experimental conditions defined above, the effects of depriving adults of one or other of the vitamins can readily be shown (Table 6), and the results, except for biotin, are as anticipated. Since none of the media contained vitamin B12, it is clear that this vitamin is also not required, except in the trace amounts which may be present as contaminants of the other dietary constituents. Biotin requirements for larval development are conditional on the protein supply, suboptimal protein reducing the requirement (Sang, 1959). So it is possible that the different metabolism of protein involved in egg formation permits normal oogenesis in the absence of this vitamin. It is perhaps worth noting that no larvae grew beyond the second instar from eggs laid in the biotin-deficient medium.

Table 6.

Effects of omission of single vitamins from the adult diet upon fecundity and egg viability

Effects of omission of single vitamins from the adult diet upon fecundity and egg viability
Effects of omission of single vitamins from the adult diet upon fecundity and egg viability

Table 2 shows that omission of RNA has no effect on egg production (see also King & Sang, 1959), whereas removal of the vitamin (folic acid) most involved in the synthesis of RNA lowers both fecundity and egg viability. Diets without both RNA and folic acid permit the same fecundity as those without folic, but the viability of the eggs laid is then lowered (King & Sang, 1959), as might be expected from the known (Travaglini, Levenbook & Schultz, 1958) high nucleic acid content of Drosophila eggs.

These results agree with expectation, but as they relate essentially to one end product (nucleic acids) they do not indicate possible interactions involving more than one reaction system. The effect of omission of pyridoxine concurrently with folic acid, or with RNA, or with both, was examined as an example of the consequences of interfering with more than one system. As before (Table 6) omission of pyridoxine has a more drastic effect than omission of folic acid (Table 7). When both are left out, the consequences are indistinguishable from the effects of not supplying pyridoxine. Omission of pyridoxine and RNA, on the other hand, results in an increase of egg output over that found for omission of pyridoxine alone. Omission of all three substances has the same effect, except that egg viability is lower, as omission of pyridoxine or pyridoxine and folic acid. There is no obvious explanation of the anomalous pyridoxine–RNA result, but the point was not explored further, since this would have involved quantitative as well as qualitative manipulations too laborious to be attempted at this time. The experiments are quoted to show that multiple deficiencies will not necessarily have predictable results and to indicate how natural shortages, which will rarely be limited to a single dietary constituent, may have unpredictable consequences.

Table 7.

Effects of removal of pyridoxine, folic acid and RNA on fecundity and egg viability

Effects of removal of pyridoxine, folic acid and RNA on fecundity and egg viability
Effects of removal of pyridoxine, folic acid and RNA on fecundity and egg viability

(d) Anions and cations

As with the vitamins, different relative amounts of salts seem to be carried over to the adult stage, and depletion times on deficient diets are consequently different. Even after 8 days there is little evidence that either Na or Ca is necessary for the production of normal eggs, whereas the K and Mg deficiencies result in reduction of,egg output before this time, and in mortality of the adults. After 16 days the effect of the deficiencies of Cl and P on viability becomes more obvious, even though P was present in a substantial amount as a contaminant of the casein. In the case of sodium the deficiency eventually causes the death of most adults. It is also interesting to note that the Na and Ca deficiencies do not completely inhibit the development of larvae hatching from eggs laid on the food medium, since in both cases a few pupae were found in the second series (i.e. 8–16 days) of cultures (Table 8).

Table 8.

Effects of omission of some anions and cations upon fecundity and egg viability

Effects of omission of some anions and cations upon fecundity and egg viability
Effects of omission of some anions and cations upon fecundity and egg viability

(e) Cytological changes

The majority of the dietary deficiencies described above produced no degenerative cytological changes in the ovaries, demonstrable with the techniques employed, even when females had ceased to lay. Such ovaries were deficient in oocytes which were in vitellogenic stages, but the structure of those egg chambers present was normal (see Text-fig. 1 B). This result is not surprising in view of the notable homeostatic ability of this organ, as demonstrated by its recovery after prolonged protein deprivation (p. 798) or after aminopterin poisoning (King & Sang, 1959). Put another way, the general response to a particular nutritional deficiency is cessation of oogenesis, not the production of abnormal, and therefore useless, eggs; and this may be an adaptation to the ephemeral character of the natural breeding and feeding sites of the fly. Three exceptions were found to this generalization, and they deserve recording (Sang & King, 1959).

Text-fig. 1.

The changing cytological picture occasioned by omitting any essential amino acid from the diet. A. Two normal ovarioles (drawn from Feulgen-stained whole mounts) showing oocytes in stages of active vitellogenesis (S8, 10 and 11). B. An ovariole at the first phase of deficient development, in which only pre-yolk stages and mature eggs are present. C and 1). A later phase in which only two to three chambers are present, none more advanced than S4. Germaria then contain no germarial cysts, show no mitoses and have little cellular differentiation. Flies fed on sugar water have ovaries as in C and D at or before 18 days. E, squamous epithelium ; F, follicular epithelium ; G, germarium ; K, karyosphere of mature oocyte; k, karyosome of oocyte nucleus; N, nurse cell nucleus; b, border cells; c, chorionic hom; M, micropylar apparatus of mature egg; nt, micropylar cap formed by border cells; S, stage designation. No correction made for shrinkage occasioned by histological manipulation.

Text-fig. 1.

The changing cytological picture occasioned by omitting any essential amino acid from the diet. A. Two normal ovarioles (drawn from Feulgen-stained whole mounts) showing oocytes in stages of active vitellogenesis (S8, 10 and 11). B. An ovariole at the first phase of deficient development, in which only pre-yolk stages and mature eggs are present. C and 1). A later phase in which only two to three chambers are present, none more advanced than S4. Germaria then contain no germarial cysts, show no mitoses and have little cellular differentiation. Flies fed on sugar water have ovaries as in C and D at or before 18 days. E, squamous epithelium ; F, follicular epithelium ; G, germarium ; K, karyosphere of mature oocyte; k, karyosome of oocyte nucleus; N, nurse cell nucleus; b, border cells; c, chorionic hom; M, micropylar apparatus of mature egg; nt, micropylar cap formed by border cells; S, stage designation. No correction made for shrinkage occasioned by histological manipulation.

(1) Deficiencies of essential amino acids

Omission for 8 days of any essential amino acids other than arginine, histidine or methionine results first in the disappearance of yolk-containing stages (Text-fig. 1B, cf. 1 A) and then to the disappearance of all stages later than stage 4 (Text-fig. 1C). As there is no evidence that any chamber can revert to an earlier stage, the assumption must be that stages 5 and higher must develop to stage 14 oocytes and that these are laid as inviable, but superficially normal, eggs. When the ovary has degenerated to this degree and only two to three chambers can be found in each ovariole, it is clear that the structure of the germaria is then also affected (Text-fig. 1D). No germarial cysts are present, few if any mitoses can be identified, and there is little or no cellular differentiation. The ovaries of flies fed only on sugar-water show an identical structure after 18 days; thus protein is required for continued proliferation of chambers from the germarium, as well as for vitellogenesis.

Text-fig. 2.

Camera lucida drawing of a stage 8 chamber from a pyridoxine starved female. The oocyte (O) is centrally located and a nurse cell (N) has an abnormally small nucleus. The follicular epithelium (F) has hypertrophied in a localized area and contains many pycnotic cells (cf. Pl. 1, fig. 1).

Text-fig. 2.

Camera lucida drawing of a stage 8 chamber from a pyridoxine starved female. The oocyte (O) is centrally located and a nurse cell (N) has an abnormally small nucleus. The follicular epithelium (F) has hypertrophied in a localized area and contains many pycnotic cells (cf. Pl. 1, fig. 1).

(2) Pyridoxine deficiency

Pyridoxine is the only vitamin whose deficiency produces abnormal development of cysts, although it is clear from earlier work (King, 1958; King & Sang, 1959) that analogues of other vitamins may be fed at levels which will lead to pathological conditions. As with other vitamin deficiencies, lack of pyridoxine leads first to a disappearance of yolky stages. Later there is an abnormal development of the nurse cell nuclei (Pl. 1, fig. 1) which is characterized by an increase of the size of the nucleus relative to the stage of development, typically one stage larger, and by an apparent decrease of the Feulgen-positive material in the nucleus. Pl. 1, fig. 1 (right) shows that the chromosomal material is more diffuse in pyridoxine-starved nurse cells, and there is also a notable increase in the number of pycnotic nuclei in the follicular epithelium. Occasional cysts also show displacement of the oocyte to a central location and retardation of the development of some nurse-cell nuclei (Text-fig. 2). In the example shown, there is also an hypertrophy of the follicular epithelium into a localized mass containing many pycnotic cells.

(3) Magnesium deficiency

Depriving females of dietary magnesium lowers the frequency of oocytes in active vitellogenesis, but it does so more slowly than removal of an essential amino acid. Replication of Feulgen-positive material is retarded in the nurse cells and no cysts of normal appearance more advanced than stage 6 are found (Pl. 1, fig. 2). However, chambers continue to be produced by thegermarium so that an ovariole may eventually contain 11–13 oocytes, of which the posterior 6–8 will be made up of degenerating stage 6 or 7 oocytes and the anterior 5 will be normal stages 1-5. Presumably the degeneration of the nurse nuclei leads to the failure of vitellogenesis, and the breakdown of the Feulgen-positive material possibly permits some re-cycling of the magnesium to the germarium and to the oocytes in early stages of development.

These results show that the qualitative requirements of the adult female Drosophila are similar to those of the larva, as would be expected since the viable egg must contain everything necessary for the formation of the larva which hatches from it. A difference would be expected only if the adult had synthetic abilities absent in the larva or if the larva had particular requirements for its growth and differentiation. Tests carried out with adults kept for only a few days on restricted diets would often be insufficient to expose these full nutritional needs, for the data presented clearly show that reserves, accumulated during larval feeding, can be drawn from adult tissues and incorporated in the forming eggs. Although the procedures used (p. 796) set a limit to the accumulation of reserves, it cannot be assumed that equivalent amounts of each dietary component would be stored, since the limiting quantity of any one necessary for normal larval development may represent only the maximum required at a particular stage of larval growth. What is carried into the pupa may have no direct relationship to the amounts provided to the larvae, so nothing can be deduced from the data concerning the quantitative nutritional requirements of the adult, except in the broadest terms.

Availability of a protein supply is the requirement most critical for egg production, and ability to recover from protein deprivation is one of the most interesting facts disclosed in these studies. The data (Tables 3 and 4) suggest that this is due to the rapid depletion of reserves of the essential amino acids, other than methionine, arginine and histidine which may be slowly synthesized but at a rate which is insufficient to allow normal egg production. Neither whole protein nor any essential amino acid seems necessary for continued adult existence as such, at least within the limits of the tests described.

The most surprising difference between adult and larval nutritional needs is shown by the apparent ability of females to produce large numbers of viable eggs when cholesterol is omitted from their diet. Cholesterol is essential for larval development (see Clark & Bloch (1959) for comments concerning other species), and no larvae grew beyond the second instar in the adult culture tubes, indicating that the medium was truly deficient in this substance (at least in terms of concentrations necessary for larval development). Chemical tests on large samples of the medium showed that only a small amount of cholesterol or cholesterol-like substances could be present (less than 0·2 μg. per 5 ml. medium). But the amount in eggs laid by yeast-fed females is also low (35 μg. of cholesterol and cholesterol-like substances per g. dry weight) so it seemed possible, if improbable, that contamination might provide sufficient sterol to allow continued egg production. An attempt was made to check this point by careful extraction of medium constituents (Table 9) with warm alcohol and ether prior to compounding them. The medium was unsatisfactory as a result of this treatment, as shown by low fecundity even when cholesterol was added. It was apparently sterol-free, and this resulted in a significant reduction of egg output when compared with the same medium containing added cholesterol.

Table 9.

Effect of addition of 0·03 % cholesterol to a cholesterol-free medium

Effect of addition of 0·03 % cholesterol to a cholesterol-free medium
Effect of addition of 0·03 % cholesterol to a cholesterol-free medium

This example illustrates the difficulty of ensuring that trace amounts of a constituent omitted from the diet are not being provided by contaminants present in the remainder. Consequently, it cannot be concluded that biotin is not required by the adult female (Table 6), but only that if there is a requirement it must be very low. And the same must be said for vitamin B12 (which was not tested, since it was known to be present in trace amounts) and for choline.

For the reasons already noted it is impossible to compare the quantitative requirements of adults and larvae with any accuracy. One calculation is worth making : a larva consumes 3 to 5 times its own weight of yeast during its 5 days of growth (Chiang & Hodson, 1950) and an adult eats about its own weight per day (King & Wilson, 1955), so that in both cases we have about the same average daily intake. Yet the level of cholesterol required by the larva is about 1500 μg./5 ml. of medium; whereas the adult needs not more than 0·2 μg. to produce her daily quota of eggs. Obviously the magnitude of the requirements for larval development and for production of viable eggs are of quite different orders in so far as cholesterol is concerned, and probably the same is true for choline and biotin. This kind of difference might be expected since an egg takes 72 hr. to form and does so in the presence of the enzyme systems of the female; i.e. it is supplied with constituents preformed by the mother (King, 1960). It follows that the quantitative nutritional requirements for egg production should be different from those for larval growth.

This difference is further illustrated by the fact that removal of RNA has no effect on egg production ; whereas it slows larval growth, since the larva’s ability to synthesize adenylic acid is rate-limiting (Sang, 1957). This cannot be the case for egg production, but it apparently can be made so by omission of folic acid. RNA cannot substitute for a folic deficiency as it can in the larva by providing the product in place of the enzyme systems which form it. So either RNA is not used or the rate-limiting process for egg production, involving folic acid, is different from the rate-limiting process for larval growth. Since a great amount of DNA is formed by the nurse cells during the development of an ovum (Jacob & Sirlin, 1959), it seems possible that the rate-limiting process is in the folic-mediated synthesis of DNA. This would imply that DNA is not formed from dietary RNA, which agrees with the fact that it has been impossible to show that dietary DNA spares RNA (Sang, unpublished; cf. House, 1961). The situation is further complicated by the observation that sulphanilamide depresses DNA replication within the nurse cell nuclei (Sang & King, unpublished) implying that the adult can synthesize folic acid to a limited extent.

Failure of nurse cell nuclei to become polyploid seems to result in abnormal vitellogenesis and the formation of abnormal eggs (King & Sang, 1959 ; King & Burnett, 1957; King, 1959). In this situation it is easy to see that a deficiency of Mg, which is necessary for the integrity of the chromosome, would result in the failure of egg development. The abnormalities of chromosome structure induced by pyridoxine deficiency would be expected to have similar results. What is perhaps more surprising is that the ovary can adapt itself to other deficiencies by suspending its activity, but only after a phase during which abnormal (but not cytologically identifiable by the techniques employed) and inviable eggs are laid. That is to say, although a few inviable, and sometimes deformed, eggs are laid when the diet is deficient in some constituent, the ovary preserves its function by a restriction of the formation of pre-yolk stages. This would be the automatic consequence of a system which partitioned nutrients in favour of the yolk-forming stages, and it seems as if the most posterior oocyte in a chamber is given priority in this respect.

Thanks are due to Dr A. W. Greenwood, C.B.E., for generous hospitality to one of us (R. C. K.) and to Dr W. Mclndoe for the sterol analyses quoted above.

APPENDIX

Medium C has the same composition as the standard larval culture medium (Sang, 1956) except that agar is omitted and MgSO4 and NaCl added to allow for the quantities present in agar. When lecithin is replaced by choline it is used at 50 p.p.m. Medium D was set with 3 % Kobe no. 1 agar. Amino acids were neutralized as necessary and kept as individual stock solutions : the concentrations listed are those used in replacement of the casein. For P deficiencies lecithin was replaced by choline, RNA was replaced by adenosine (925 p.p.m.) and cytidine (460 p.p.m.) and an organic buffer replaced the phosphates. For the Ca deficiencies a calcium-free casein and pantothenic acid were used. In the case of sodium and potassium deficiencies organic buffers were used. For Cl deficiencies Cl-free pyridoxin and thiamin were prepared.

Chiang
,
H. C.
&
Hodson
,
A. C.
(
1950
).
An analytical study of population growth in Drosophila melanogaster
.
Ecol. Monogr
.
20
,
173
206
.
Clark
,
A. J.
&
Bloch
,
K.
(
1959
).
The absence of sterol synthesis in insects
.
J. Biol. Chem
.
234
,
2578
82
.
Clements
,
A. N.
(
1956
).
Hormonal control of ovary development in mosquitoes
.
J. Exp. Biol
.
33
,
211
23
.
Dimond
,
J. B.
,
Lea
,
A. O.
,
Hahnert
,
W. F.
&
De Long
,
D. M.
(
1956
).
The amino acids required for egg production in Aedes aegypti
.
Canad. Ent
.
88
,
57
62
.
Dougherty
,
E. C.
(
1959
).
Axenic culture of invertebrate metazoa: a goal
.
Ann. N.Y. Acad. Sci
.
77
,
25
54
.
Gowen
,
J. W.
&
Johnson
,
L. E.
(
1946
).
On the mechanism of heterosis. I. Metabolic capacity of different races of Drosophila melanogaster for egg production
.
Amer. Nat
.
80
,
149
79
.
Harlow
,
P. M.
(
1956
).
A study of ovarial development in its relation to adult nutrition in the blowfly, Protophormia terraenovae
.
J. Exp. Biol
.
33
,
777
97
.
Housb
,
H. L.
(
1961
).
Insect nutrition
.
Ann. Bev. Entomol
.
6
,
13
26
.
Jacob
,
J.
&
Sirlin
,
J. L.
(
1959
).
Cell function in the ovary of Drosophila melanogaster. I. DNA classes in the nurse cell as determined by autoradiography
.
Chromosoma
,
10
,
210
28
.
King
,
R. C.
(
1958
).
Further studies of oogenesis in D. melanogaster
.
Drosophila Inform. Serv
.
32
,
131
King
,
R. C.
(
1959
).
Oogenesis in mr2
.
Drosophila Inform. Serv
.
33
,
143
.
King
,
R. C.
(
1960
).
Oogenesis in adult Drosophila melanogaster. IX. Studies on the cytochemistry and ultrastructure of developing oocytes
.
Growth
,
34
,
265
323
.
King
,
R. C.
&
Burnett
,
R. G.
(
1957
).
Oogenesis in adult Drosophila melanogaster. V. Mutations which affect nurse cell nuclei
.
Growth
,
21
,
263
80
.
King
,
R. C.
,
Burnett
,
R. G.
&
Staley
,
N. A.
(
1957
).
Oogenesis in adult Drosophila melanogaster. IV. Hereditary ovarian tumours
.
Growth
,
21
,
239
61
.
King
,
R. C.
,
Rubinson
,
A. C.
&
Smith
,
R. F.
(
1956
).
Oogenesis in adult Drosophila melanogaster
.
Growth
,
20
,
121
57
.
King
,
R. C.
&
Sang
,
J. H.
(
1958
).
Modification of ovogenesis in D. melanogaster
.
Drosophila Inform. Serv
.
32
,
131
2
.
King
,
R. C.
&
Sang
,
J. H.
(
1959
).
Oogenesis in adult Drosophila melanogaster. VIII. The role of folic acid in oogenesis
.
Growth
,
33
,
37
53
.
King
,
R. C.
&
Wilson
,
L. P.
(
1955
).
Studies with radiophosphorus in Drosophila. V. The phosphorus balance of adult females
.
J. Exp. Zool
.
130
,
71
82
.
Meister
,
A.
(
1957
).
Biochemistry of the Amino Acids
,
485
pp.
New York
:
Academic Press Inc
.
Robertson
,
F. W.
&
Sang
,
J. H.
(
1944
).
The ecological determinants of population growth in a Drosophila culture. I. Fecundity of adult flies
.
Proc. Roy. Soc. B
,
132
,
258
77
.
Sang
,
J. H.
(
1956
).
The quantitative nutritional requirements of Drosophila melanogaster
.
J. Exp. Biol
.
33
.
45
72
.
Sang
,
J. H.
(
1957
).
Utilization of dietary purines and pyrimidines by Drosophila melanogaster
.
Proc. Roy. Soc. Edinb. B
,
66
,
339
59
.
Sang
,
J. H.
(
1959
).
Circumstances affecting the nutritional requirements of Drosophila melanogaster
.
Ann. N.Y. Acad. Sci
.
77
,
352
65
.
Sang
,
J. H.
&
King
,
R. C.
(
1959
).
Nutritional requirements for normal oogenesis in D. melanogaster
.
Drosophila Inform. Serv
.
33
,
156
8
.
Singh
,
K. R. P
&
Brown
,
A. W. A.
(
1957
).
Nutritional requirements of Aedes aegypti L
.
J. Insect Physiol
,
1
,
199
220
.
Travaglini
,
E. C.
,
Levenbook
,
L.
&
Schultz
,
J.
(
1958
).
Nucleic adds and their components as affected by the Y chromosome of Drotophila melanogatter. II. Nucleotides and related compounds in the add soluble fraction of the unfertilized egg
.
Exp. Cell Res
.
15
,
62
79
.

Fig. 1. Photomicrographs of cysts from a normal (left) and pyridoxine-deficient ovaries (right). The abnormal chambers are equivalent in volume to stage 8, but the diameters of the nurse cell nuclei are characteristic of the next stage. The nuclei contain far fewer Feulgen-positive strands than normal, and they are less compactly organized. Note also that the epithelium around the follicle is characterized by pycnotic areas never found in the controls.

Fig. 2. Preparation of an ovary from a female deprived of Mg for 16 days. The chambers have continued to proliferate and as many as a dozen may be found in an ovariole (cf. Text-fig. 1). None develop beyond stage 6 (pre-vitellogenesis), and the nurse cell nuclei are deficient in Feulgen-positive material. The posterior half of an ovariole is composed of degenerating cysts.

Fig. 1. Photomicrographs of cysts from a normal (left) and pyridoxine-deficient ovaries (right). The abnormal chambers are equivalent in volume to stage 8, but the diameters of the nurse cell nuclei are characteristic of the next stage. The nuclei contain far fewer Feulgen-positive strands than normal, and they are less compactly organized. Note also that the epithelium around the follicle is characterized by pycnotic areas never found in the controls.

Fig. 2. Preparation of an ovary from a female deprived of Mg for 16 days. The chambers have continued to proliferate and as many as a dozen may be found in an ovariole (cf. Text-fig. 1). None develop beyond stage 6 (pre-vitellogenesis), and the nurse cell nuclei are deficient in Feulgen-positive material. The posterior half of an ovariole is composed of degenerating cysts.