The function of the Corpus Allatum in the Growth and Reproduction of Rhodnius prolixus (Hemiptera)

It was further suggested that the corpus allatum is the source of both these factors. The main object of the present paper is to bring forward more evidence in support of this suggestion, and to describe the relation of the corpus allatum to the activity of the reproductive organs in the adult insect (Wigglesworth, 1935). But, first, there are a number of additions to be made to the subject matter of the earlier paper.

It has already been shown that in Rhodnius a certain degree of distension of the abdomen is necessary to bring about the secretion of the moulting hormone. The question arises whether the hormone can induce moulting in an insect which has not been distended in this way.

This has been tested by decapitating a number of unfed 4th-stage nymphs within a day after moulting, and joining them to 5th-stage nymphs decapitated 12 days after feeding. The 4th-stage nymphs were caused to moult, and, as was to be expected from the previous observations, they developed adult characters.

This experiment proves that distension in Rhodnius is a necessary prelude to moulting only because it provides the stimulus for the secretion of the moulting hormone. If the hormone is supplied artificially the insect can use the undigested blood which it carries forward from the previous instar for the development of a new cuticle.

It follows from this that the moulting hormone must normally be removed from the body fluid as soon as moulting has taken place; otherwise the insect would moult again. Perhaps that is why, when two insects are joined and one moults first, moulting in the other is arrested (Wigglesworth, 1934).

In all the experiments previously described, moulting has been induced in decapitated insects by joining them directly to insects which had passed the ‘critical period ‘. The possibility was not excluded that moulting was caused by contact with the growing epidermis of the second insect.

This has been tested by decapitating a number of 4th-stage nymphs at 7 days after feeding, and joining them by way of capillary tubes about 4 mm. long to 4th-stage nymphs decapitated 24 hours after feeding (fig. 1, Pl. 4). The second insects were duly caused to moult—proving that the moulting hormone circulates freely in the blood.

In the earlier paper it was noted that when nymphs of Rhodnius are joined together they nearly always moult simultaneously. From this it was inferred that the whole process of growth was co-ordinated by hormones circulating in the blood. But when longitudinal sections of insects which have moulted simultaneously are cut, it is found that the new cuticle and epidermis continue without a break from one insect to the other (Text-fig. 3 B). While in insects which have failed to moult simultaneously this epithelial continuity is not completely established. And in the experiments described above, where the insects were connected by a capillary tube, they always moulted independently.

It appears, therefore, that the continuity of the epidermis between the two insects is the essential condition for simultaneous moulting; though by what mechanism co-ordination is effected remains obscure.

The same phenomenon was observed by Bodenstein (1933) in the moulting of transplanted appendages in caterpillars at the same time as their new host. And Crampton (1899) noted the simultaneous development of Saturniid pupae which had been joined together, and demonstrated the union of the integument in the resulting moths.

Nymphs of Triatoma rubrofasciata and Triatoma infestans (blood-sucking Reduviids related to Rhodnius) decapitated 24 hours after feeding do not moult. But if they are joined to Rhodnius nymphs which have passed the critical period they are caused to moult simultaneously with the Rhodnius; and the two insects grow together like two Rhodnius nymphs, so that in sections it is not possible to tell where the epidermis of one ends and the other begins.

Similar experiments have been made with the bed-bug Cimex lectularius (Hemiptera: Cimicidae). In this insect it is convenient to cut through the body behind the prothorax. Text-fig. 1 shows that if this is done 24 hours after feeding (in insects kept at 24° C.) none of them moult: the ‘critical period ‘, after which the head is no longer necessary for moulting to occur, is from 2 to 3 days after feeding. Fifth-stage nymphs of Cimex transected behind the prothorax 24 hours after feeding were joined to 3rd-stage nymphs of Rhodnius decapitated 6 days after feeding. The Cimex nymphs were caused to moult at the same time as the Rhodnius, and in sections it was found that the epidermis and cuticle of the two insects (although belonging to different families) had become continuous (Text-fig. 2).

The moulting factor is therefore non-specific (cf. Bodenstein, 1933); later (p. 98) it will be shown that the same is true of the inhibitory factor.

In the earlier paper (Wigglesworth, 1934) it was concluded that the corpus allatum is the probable source of the moulting hormone because the cells of this gland become swollen, the cytoplasm increasing greatly, during the ‘critical period ‘when the moulting hormone is being secreted. Conclusive experimental evidence was not given.

Rather indirect evidence for this view has since been obtained as follows: 4th-stage nymphs of Rhodnius in two groups (A and A’) were decapitated 7 days after feeding and joined to groups of 4th-stage nymphs (B and B’) 24 hours after these had fed. Of these, B were transected through the prothorax, the corpus allatum being removed, while B’ were transected through the posterior part of the head behind the brain but in front of the corpus allatum.

Groups B and B’ both showed that moulting was beginning¹ at the end of 5 days. The pairs were then separated and B and B’ were joined to further groups of 4th-stage nymphs (C and C’) transected through the prothorax 24 hours after feeding. At the end of 8 days the group C’, i.e. insects whose partners retained the corpus allatum, were obviously beginning to moult; the group C was probably just starting to moult. The pairs were then separated once more and joined to groups D and D’cut through the prothorax 24 hours after feeding. At the end of 7 days D’ were clearly beginning to moult; but D died several weeks later without showing any sign of moulting beginning.

Thus in the series A, B, C, D in which no corpus allatum was present, the moulting hormone appeared to diminish in concentration until it ceased to induce moulting at all. Whereas in the series A’, B’, C’, D’, in which B’ retained the corpus allatum (and was presumably able, therefore, to supply an adequate amount of the hormone to O’), there was no sign of diminished activity in the moulting hormone—the group D’ being caused to show signs of moulting as soon after joining to C’ as B’ did after joining to A’.

In the earlier paper it was shown that if a 4th-stage nymph of Rhodnius 6 days after feeding, that is, after the ‘critical period’, had only the tip of the head removed, and was then joined to a 5th-stage nymph decapitated 24 hours after feeding, the latter moulted to an extra nymphal stage instead of becoming an adult. The head is therefore necessary for the production of the ‘inhibitory hormone’.

If this experiment is repeated, but the brain and all the other structures lying in front of the corpus allatum are cut away in the 4th-stage nymph (Text-fig. 3 A), the same results are obtained; and in sections the corpus allatum is seen to be hypertrophied (Text-fig. 3 B).

Further, if a 5th-stage nymph is allowed to pass the ‘critical period’ and then (9 days after feeding) is decapitated behind the corpus allatum and joined to a 4th-stage nymph without the brain but with the corpus allatum intact, the 4th-stage nymph is caused to moult (by the hormone from the 5th-stage nymph) and metamorphosis of the 5th-stage nymph is inhibited (presumably by the secretion from the corpus allatum of the 4th-stage nymph) and it develops once more into a nymph.

This same result is obtained if a 3rd-stage nymph of Triatoma is used in place of the 4th-stage nymph of Rhodnius (fig. 3, PL 4). That is, the corpus allatum of the earlier nymphal stages of Triatoma will prevent metamorphosis in the 5th- stage nymph of Rhodnius. And in the experiment described above (p. 94) in which 5th-stage nymphs of Cimex were transected behind the prothorax and caused to moult by joining them to 3rd-stage nymphs of Rhodnius which had passed the ‘critical period’, it was noted that metamorphosis of the Cimex was partially inhibited: the genitalia were imperfectly developed (Text-fig. 4 c). If the 3rd-stage nymphs of Rhodnius were allowed to retain the corpus allatum the characters of the Cimex were still more nymph-like (Text-fig. 4 D).¹

Further proof that the corpus allatum secretes the inhibitory factor was obtained by experiments in which the corpus allatum from 3rd- or 4th-stage nymphs was implanted in the abdomen of 5th-stage nymphs.

The technique adopted was as follows: a small circle about 5 min, in diameter is traced with melted paraffin wax on one side of the dorsum of the abdomen of the 5th-stage nymph 24 hours after feeding. The object of this is to prevent haemolymph which escapes from the wound from flowing over the body surface. A small incision is made inside this circle. The posterior extremity of the head of the 4th-stage nymph (removed a day or two after feeding) which contains the corpus allatum but not the brain, is then cut off and inserted through the incision, and the wound sealed with melted paraffin. There is almost no mortality among insects operated on in this way, and although moulting often requires a week or two longer than usual it always takes place in those which survive.

Out of fifty-nine 5th-stage nymphs so treated, twenty-three developed into normal adults, thirty-six developed nymphal characters to a variable extent. In some this amounted only to a little patch of nymphal cuticle (see Wigglesworth, 1933) immediately around the site of the transplant, gradually merging into the adult cuticle beyond (Text-fig. 5 E, F, and fig. 16, Pl. 5). In others, although the wings and genitalia were those of an adult, the cuticle over the entire abdomen was intermediate in character. Others again had reduced wings (Text. fig. 5 D) and imperfect genitalia but nymphal cuticle all over the abdomen (figs. 13, 14, and 15, Pl. 5).

In the most successful experiments the insects retained all the characters of nymphs (fig. 12, Pl. 5). The leathery wing lobes, though larger than those of the 5th-stage nymph (Text-fig. 5 A), were of the same type (Text-fig. 5 B). The external genitalia in the female were only slightly more differentiated towards the adult form (Text-fig. 6 c). In the male there were rudimentary claspers and a pale, soft, membranous structure, presumably a rudimentary aedeagus, which was kept permanently extruded. The dimensions of the wings, limbs, &c., increased to just about the same extent as in the change from 5th-stage nymph to adult. There was no apparent difference in the characters of these ‘6th-stage’ nymphs whether they were produced from the 5th-stage nymph by implanting the corpus allatum from the 3rd or the 4th stage. The implantation of corpora allata from 5th-stage nymphs, even when six were implanted in a single insect, had no effect upon metamorphosis.

By the technique described above, a part of the nerve-cord and suboesophageal ganglion and sometimes a small part of the brain are implanted along with the corpus allatum. In order to prove that the ‘inhibitory effect’ is due to the corpus allatum alone the following experiments were performed.

(i) The corpus allatum and the attached sympathetic ganglion were dissected away from the nervous system of 4th-stage nymphs, in Ringer’s solution, and implanted in the abdomen of 5th-stage nymphs. Out of five such insects two retained nymphal characters on moulting. If better methods of handling the material were devised a higher proportion of successes could doubtless be obtained.

(ii) Many 5th-stage nymphs received implantations of the brain and other structures in the head, without the corpus allatum, but in no case were nymphal characters developed.

(iii) Sections of the implants were cut and the state of the transplanted tissues compared in the successful and unsuccessful experiments. In both groups the tissues were alive and growing, though for the most part quite disorganized. The nerve-cells were often conspicuous. But the only constant difference between the two groups was that where ‘inhibition’ had failed and adult characters developed the corpus allatum was shrunken and atrophic (Text-fig. 7 A) ; while in the ‘6th-stage’ nymphs it retained its form as a compact gland made up of healthy swollen cells (Text-fig. 7 B).

The only other tissue which appears to remain organized after transplantation is the epidermis. This always grows in such a way as to establish continuity with itself (Text-fig. 7 c). The epithelium from one end of the cylindrical implant grows outwards and backwards to join with the epithelium from the other end. When the host moults the old cuticle of the implant is shed and is encysted within the reflected slieves of cuticle laid down by the continuous epidermis. Sometimes the tracheae show this same curious form of growth, but in them the reflection takes place inwards down the lumen of the tube (Text-fig. 7 D).

The ‘6th-stage’ nymphs are capable of undergoing another moult after feeding. Unfortunately none of them has succeeded in freeing itself completely from the old cuticle, and therefore all have died during the process. The characters developed are variable, and the experiments have been too few to provide a satisfactory explanation. Three ‘6th-stage’ nymphs which had been produced by the implantation of the corpus allatum of a 4th-stage gave rise to adults when they moulted again; whereas three produced by implanting 3rd-stage corpora allata (which had now presumably reached the 4th stage) developed nymphal or intermediate characters. Text-fig. 6 D shows the external genitalia of a female ‘7th-stage’ nymph which had characters intermediate between nymph and adult. These results were to be expected if the number of instars in Rhodnius is determined by the corpus allatum. But the insect shown in fig. 14, Pl. 5, intermediate between nymph and adult, which was produced by implanting a 4th-stage corpus allatum into a 5th- stage nymph, developed nymphal characters when it moulted again. This phenomenon clearly needs further investigation.

The secretion of the corpus allatum is thus responsible for preventing the development of adult characters until the moulting of the 5th-stage nymph. The next question is whether this secretion plays a part in determining the characters of each nymphal stage. Although the change from 5th-stage nymph to adult is far more striking than the change at any previous moult, there are slight morphological changes at the moulting of each instar—notably in the external genitalia (Gillett, 1985) and in the size of the wing lobes. In other hemimetabolic insects there are still more definite differences between the instars, and in Locusta Key (1936) has shown that when, as sometimes happens, an extra instar occurs in a given individual, it may be, morphologically, either a 3rd instar or a 4th.

The experiments made to test this question consisted in joining together 3rd- and 4th-stage nymphs in different relations. The chief experiments are set out diagrammatically in Text-fig. 8, i–iv. In each case the lower, shaded, insect provides the moulting factor, i.e. is decapitated at 6 days after feeding; the upper, unshaded, insect is decapitated at 24 hours after feeding. The black spot indicates that the insect in question, though deprived of its brain, retains the corpus allatum. Females were used throughout.¹

In experiment (i), the 4th-stage nymph develops into a normal 5th stage (Text-fig. 8 B) ; the 3rd-stage nymph develops into a 4th (Text-fig. 8 A). From this it is evident that the corpus allatum of the 4th stage does not specifically determine the development of 5th-instar characters—for if that were so the 3rd-stage nymph would develop into a 5th.

In experiment (ii), where the 4th-stage nymph is deprived of its corpus allatum but the 3rd-stage nymph retains this gland, the 3rd-stage nymph develops into a 4th-stage nymph, and the 4th-stage nymph develops characters intermediate between those of a 4th instar and those of a 5th (Text-fig. 8 c).

In experiment (iii), where the 3rd-stage nymph provides both the moulting factor and the corpus allatum, both insects develop 4th-instar characters (Text-fig. 8 D). Here the 4th-stage nymph has been prevented by the corpus allatum of the 3rd- stage nymph from developing 5th-instar characters.

In experiment (iv), where neither insect retains the corpus allatum, both nymphs suffer a premature metamorphosis and develop adult characters (Text-fig. 8 E).

From these experiments it appears that the characters of the various instars are controlled by the corpus allatum in the same way as metamorphosis is controlled: the secretion does not ensure that the characters of a given instar shall be developed, but seems to restrain development in such a way as to limit the degree of differentiation towards the adult form. This restraint appears to be lessened slightly in each instar, so that at each moult the body form advances a little toward the adult state. This process continues until the moulting of the 5th- stage nymph; then all restraint ceases and complete metamorphosis occurs.

What the nature of this inhibition of differentiation may be will be considered later, in the discussion.

In the earlier paper (Wigglesworth, 1934) it was shown that during the moulting of the 5th-stage nymph the corpus allatum shows signs of activity, that is, the cells become swollen and the cytoplasm uniformly dense, during the ‘critical period’ when the moulting hormone is being secreted. Thereafter the cells become shrunken, with vacuolated spaces between. This state persists until the insect moults (Text-fig. 9 A) ; but soon after it has become adult the cells acquire once more the same increase of dense cytoplasm (Text-fig. 9 B). NOW at the time of moulting the ovaries of the female contain no ripe eggs; but, provided sufficient undigested blood remains in the stomach, the eggs begin to develop soon after moulting, and within a week or 10 days, at 24° C, a few ripe eggs are usually formed.

Similarly, after prolonged fasting, the cells of the corpus allatum revert to the shrunken condition (Text-fig. 9 c). A few days after feeding they again become swollen (Text-fig. 9 D) and the reproductive organs become active.

These observations suggest that the secretion of the corpus allatum in the adult may be responsible for the activity of the reproductive organs. This would agree with the observation of Ito (1918) that in Lepidoptera, which mature their eggs largely during the pupal stage, the corpora allata show signs of maximum activity in the pupa. And Holmgren (1909) noted that in queen Termites the corpora allata attain a colossal size, regarded by him as a hypertrophic degeneration.

The question has been tested (i) by decapitating insects within a day or two of becoming adult; (ii) by selecting females in which there are no ripe eggs in the ovaries,¹ feeding them, and decapitating them 24 hours later. Of thirty adult females in which the entire head containing the brain and the corpus allatum was removed, none developed eggs (fig. 6, Pl. 4). Of twenty-four adult females in which the brain was removed, but the posterior extremity of the head which contains the corpus allatum was left intact, twenty-two developed batches of eggs (fig. 8, Pl. 4).

That this effect of the corpus allatum is due to an active substance circulating in the blood, was proved by joining together (by the same technique as that employed in the nymphal stages) adult females with a corpus allatum and adult females without (fig. 4, Pl. 4). Of twenty-five such pairs of females, eggs were developed in both individuals in twenty; eggs were developed in neither individual in five; in no case did eggs develop in one individual and not in the other.

Like the hormones of the nymphal stages, the active substance in the adult female is non-specific: adult females of Triatoma infestans, with a corpus allatum, joined to adult females of Rhodnius without, caused the latter to develop eggs (fig. 5, Pl. 4).

Each ovary in Rhodnius contains seven ovarioles of the acrotrophic type. That is, the nurse cells are confined to the apical region, where they surround a nutritive cavity connected by fibrillar nutritive cords to the developing oocytes below. The oocytes are nourished in this way until they attain a length of about 0·46 mm. and a diameter of about 0 30 mm. (the ripe egg is about 1 5 mm. by 0·65 mm.) then the connexion with the nurse cells is broken, and the oocyte is completely surrounded by follicular cells which complete the formation of the yolk. When the chorion is laid down and the egg is fully formed, the follicular cells at the lower pole separate and the egg is passed on to the oviduct. The empty follicle then collapses to form the so-called ‘corpus luteum ‘and the cells undergo necrosis and absorption. By the time the next oocyte is ripe the preceding follicle has practically disappeared (Text-fig. 10 B).

In the female Rhodnius deprived of the corpus allatum this process continues normally until the stage at which the oocyte loses its connexion with the nurse cells. At that point its development is arrested; the oocyte dies, and the follicular cells, instead of producing yolk, seem to proliferate amitotically and absorb the dead oocyte. Ultimately, they themselves break down and become absorbed, like the ‘corpus luteum’ of the normal ovariole. Meanwhile new oocytes are undergoing partial development followed by degeneration; so that in a single ovariole there may be several follicles and oocytes in different stages of necrosis (Text-fig. 10 c).

We have seen that in the starved adult the corpus allatum reverts to what is regarded as the inactive condition with the cells shrunken. It is not surprising, therefore, that in the starved female the ovary shows the same changes as in the fed female deprived of the corpus allatum: fresh oocytes continue to be produced, only to die later instead of ripening.

If the 5th-stage nymph, 10 days after feeding, has the brain removed but the corpus allatum left intact, and is then joined to an adult female deprived of the corpus allatum before egg development has begun, the 5th-stage nymph completes the moulting process but the adult female does not develop eggs. The ‘moulting hormone’ of the 5th-stage nymph is, therefore, distinct from the hormone which brings about the development of eggs in the adult female.

Conversely, if female adults are fed and, with the corpus allatum intact, are joined to 4th-stage or 5th-stage nymphs decapitated 24 hours after feeding (fig. 2, Pl. 4), the adults duly develop eggs but the nymphs are not caused to moult.

If adult females with the corpus allatum intact are joined to female 5th-stage nymphs decapitated after the critical period— say, at 15 days after feeding—so that the nymphs contain in their blood both the ‘moulting hormone’ of the nymphal stage and ‘egg forming hormone ‘of the adult female, then at the time of moulting (about 28 days after feeding) they are found on dissection to have their ovaries at the same stage of development as in normal newly moulted adults. Thus, the ovaries of the developing 5th-stage nymphs do not produce ripe eggs even in the presence of the secretion of the corpus allatum of the adult female.

In the earlier paper (Wigglesworth, 1934) an account was given of the external structure of ‘precocious adults ‘produced by joining young nymphs of Rhodnius decapitated soon after feeding to 5th-stage nymphs which had passed the critical period. The internal structure has now been examined in sections of female ‘adults’ produced by joining 3rd-stage to 5th- stage nymphs.

Text-fig. 11 shows for comparison the ovaries of a normal 4th-stage nymph soon after moulting, and of one of these precocious adults. In the latter the oviducts are greatly developed and often distended with a neutrophil fluid. The ovarioles are larger than those in the normal nymph and the follicular cells, nurse cells, and oocytes are becoming differentiated.

Text-fig. 12 A shows the internal reproductive organs and accessory glands of the male Rhodnius dissected out.¹ In the newly moulted male (Text-fig. 12 B) the testes are fully developed but there are few spermatozoa in the vesiculae seminales, and the tubular accessory glands are small and thin. During the first 2 or 3 weeks after moulting, as the blood in the stomach is digested, the vesiculae become distended with an opaque white mass of spermatozoa, the accessory glands with a clear fluid (Text-fig. 12 c).

In adult males decapitated 24 hours after moulting the vesiculae become distended with spermatozoa in the usual way, but the secretion in the accessory glands is wanting and they retain their original attenuated form (Text-fig. 12 D). On the other hand, if the insects are transected in front of the corpus allatum, the accessory glands develop normally; and if insects which retain the corpus allatum are joined to insects in which it has been removed, normal development occurs in both.

The secretion which is necessary for the normal activity of the follicular cells in the female ovary and of the accessory glands in the male seems to be identical in the two sexes. For if females are decapitated and joined to males with the corpus allatum intact they develop ripe eggs (fig. 7, Pl. 4), and if males without the corpus allatum are joined to females which retain the gland they develop the normal distension of the accessory glands.

When, under the stimulus of the ‘moulting hormone’, the epidermal cells of the 4th-stage Rhodnius nymph detach themselves from the cuticle and begin to divide, they are capable of producing either another nymphal instar or an adult. The type of development depends upon the hormones present. In the absence of the corpus allatum (as in nymphs decapitated at the ‘critical period ‘) the epidermis gives rise to a cuticle of the adult type: the nymph undergoes metamorphosis. If the corpus allatum remains intact a cuticle is laid down which differs very little from that which preceded it: metamorphosis is ‘inhibited’.

Now the change of form which metamorphosis involves requires a greater amount of cellular proliferation than does the formation of another nymphal cuticle. This is particularly obvious in the development of the wings, the external genitalia, and the lateral pleats in the abdomen characteristic of the adult. The moulting glands of the abdomen (Wigglesworth, 1933), the climbing organ on the first two pairs of legs (Gillett and Wigglesworth, 1932), the elaborate structure of the prothorax, and the ocelli which now appear for the first time, all exhibit the extensive remodelling which metamorphosis entails.

This differentiated growth requires a longer time for its completion. The normal interval, at 24° C., between feeding and moulting in the 1st- to the 4th-nymphal stages is 12-16 days. The moulting of the 5th-stage nymph into an adult requires 28-30 days; whereas a 5th-stage nymph caused to moult into a ‘6th-stage ‘nymph by joining it to 4th-stage nymph with the corpus allatum intact (Wigglesworth, 1934) required only 17 days. Conversely, when 1st- to 4th-stage nymphs are decapitated around the critical period, some turn into nymphs, some into adults (Wigglesworth, 1934); and although there are occasional exceptions, those which develop nymphal characters are generally ripe for moulting sooner than those which are suffering metamorphosis. Again, in the moulting of the ‘6th- stage’ nymphs (p. 103), one insect which gave rise to a ‘7th- stage ‘nymph moulted in 21 days, while those which gave rise to adults moulted in 28–34 days.

It is possible that this time factor may be important in determining the structural characters. For in the course of moulting there are two successive processes: (i) the epidermal cells detach themselves from the old cuticle, multiply, and group themselves so as to produce the form of the next instar; (ii) they then lay down the new cuticle in the manner already described (Wigglesworth, 1933). Now if the second process supervenes soon after the first has begun, the form of the cuticle will differ very little from that of the preceding instar; in other words, metamorphosis will be inhibited. Whereas, if the laying down of the cuticle is delayed, sufficient time will have been given for the formation of the adult organs, and hence metamorphosis will take place. What we have called the ‘inhibitory factor’ may be pictured as arresting the first process (differentiation) by initiating the second (cuticle deposition).

This conception may be represented graphically upon lines similar to those adopted by Goldschmidt. In Text-fig. 13, which represents the moulting of a 4th-stage nymph, the oblique line A represents the velocity of the first process—differentiated growth towards the adult form; the vertical lines mark the times at which the second process—the deposition of cuticle— supervenes. If this occurs at B, a 5th-stage nymph will be developed; if at B’, an adult; between B and B’ the characters will be intermediate between adult and nymph ;¹ before B they will be intermediate between a 4th-stage nymph and a 5th. In each nymphal instar a little more differentiation is permitted; but this amount is so slight compared with what takes place at the final moult that it is justifiable to describe the final moult as metamorphosis.

It has been shown in the present paper that the hormone responsible for the early occurrence of process B is secreted by the corpus allatum. Whether this hormone is chemically distinct from the moulting hormone which initiates process A, or whether the early termination of process A is merely the result of a higher concentration of moulting hormone, cannot be decided from the experiments described: these could be equally well explained upon either hypothesis. That the moulting hormone is present at a higher concentration in the earlier nymphal stages is, however, very probable. For decapitated 4th-stage nymphs are caused to start moulting in about five days if they are joined to 4th-stage nymphs that have passed the critical period, but not until seven or eight days if they are joined to 5th-stage nymphs at a corresponding time.

The foregoing is not, of course, the only explanation of the inhibition of metamorphosis: it is merely a convenient hypothesis which brings the phenomenon into line with the theory of intersexes, and of wing patterns in butterflies, elaborated by Goldschmidt (1932).

Nor is it a complete explanation; for there is another side to the matter which has not yet been considered in this discussion. There seems to be a profound physiological difference between the epidermal cells according to the type of cuticle they have laid down. If they have laid down nymphal cuticle, even though the body-size has reached its natural limit (as in the ‘6th-stage’ nymphs), the cells are capable of responding again to the moulting hormone, multiplying, and producing either a nymphal or an adult cuticle. Whereas, if they have laid down cuticle of the adult type, the moulting hormone has no effect upon them :¹ after metamorphosis the insect is incapable of further growth. The ‘inhibitory factor’ seems, indeed, to keep the epidermal cells to some extent in an ‘embryonic state’. So long as they are restrained in this way there seems to be no limit to the capacity for growth: the normal body-size appears to be determined by the alteration of activity inherent in the corpus allatum.

From the work of Seidel (1934) and Schnetter (1934) it appears that in the embryonic development of all insects there exists in the first thoracic segment or in one of the posterior segments of the head a centre from which every kind of subsequent differentiation spreads like a wave throughout the developing egg. It is interesting to note the identical position of this ‘differentiation centre’ and the corpus allatum—the centre which, in Rhodnius, controls differentiation in post- embryonic life.

In Rhodnius the oocytes remain at an early stage of development until after the final moult. It has therefore been possible in this insect to differentiate clearly between the hormonal control of moulting and of egg development. But in many Lepidoptera, for example in the Sphingidae, the eggs become fully formed during the pupal stage. Bytinski-Salz (1933) showed that female pupae resulting from the cross Celerio ga 11 ii d × Celerio euphorbiae?fail to develop into imagines. But if the ovaries of these hybrid females are transplanted into the hybrid males (which develop normally) full-grown eggs are matured within them. Bytinski-Salz suggests that this effect is due to a hormonal influence by the new host; but in his material it was not possible to draw a distinction between the factors influencing general and ovarian growth.

Recently Iwanoff and Mestscherskaja (1935) studied the permeability of the oocytes in the immature and mature ovaries of insects and concluded that the essential change at maturity is an increased permeability. They believe that this change is normally brought about by a hormone from the fat-body, and that the reverse change, a temporary arrest of egg growth due to diminished permeability, is brought about by a secretion from the ‘corpus luteum’. The methods employed by these authors are so different from those described in the present paper that no useful comparison can yet be drawn between the results.

1. The stretching of the cuticle which initiates moulting is necessary only as a stimulus for the secretion of the moulting hormone. The hormone so produced will induce moulting in unfed nymphs.

2. The simultaneous moulting of joined insects is dependent on the continuity of the epidermis established between them. But moulting can be induced by the medium of the blood λvith- out contact between the tissues.

3. The moulting hormone from Rhodnius will induce moulting in bugs of the allied genus Triatoma and in the bed-bug, Cimex. The epidermis of these joined insects grows together during the process.

4. Some rather indirect experimental evidence that the corpus allatum secretes the moulting hormone is given.

5. The corpus allatum secretes the ‘inhibitory hormone’ which prevents metamorphosis in the earlier nymphal stages. Fifth-stage nymphs with the corpus allatum of 3rd- or 4th-stage nymphs implanted in them give rise to ‘6th-stage’ nymphs. These ‘6th-stage’ nymphs may give rise to ‘7th-stage’ nymphs when they moult again.

In some transplantation experiments the characters all over the body may be intermediate between nymph and adult; or nymphal characters may only be developed near the implant.

6. The inhibitory effect is non-specific as between Rhod- nius, Triatoma, and Cimex.

7. The corpus allatum also determines the characters of each nymphal instar by limiting the degree of differentiation towards the adult form which occurs during the moult.

8. In the phenomenon of ‘inhibition of metamorphosis’ there seem to be two elements: (i) deposition of the new cuticle follows rapidly upon the initiation of growth and hence differentiation of the adult characters is arrested; (ii) so long as differentiation is arrested in this way the cells are capable of renewed growth and will respond again to the ‘moulting hormone ‘.

9. Whether the ‘moulting hormone’ is chemically distinct from the ‘inhibitory hormone ‘is not proved.

1. In the adult female the corpus allatum is necessary for the production of ripe eggs. The secretion from the corpus allatum of Triatoma females will cause egg development in Rhodnius.

2. In the absence of this secretion the oocytes continue to grow so long as they are connected to the nurse cells. They die and are absorbed when their nutrition is taken over by the follicular cells.

3. The moulting hormone of the nymphal stages will not cause egg development in the adult female; nor will the egg-forming hormone induce moulting.

4. The corpus allatum in the adult male is necessary for the normal activity of the accessory glands. Its secretion will induce egg development in the adult female; and the secretion from the female will activate the accessory glands of the male.

PLATE 4.

Fig. 1.—Two decapitated 4th-stage nymphs connected by a capillary tube.

Fig. 2.—Adult female with corpus allatum joined to decapitated 4th- stage nymph.

Fig. 3.—Fifth-stage nymph of Rhodnius (below) joined to 3rd-stage nymph of Triatoma infestans.

Fig. 4.—Female Rhodnius with corpus allatum joined to female without.

Fig. 5.—Female Triatoma infestans with corpus allatum joined to female Rhodnius without.

Fig. 6.—Female Rhodnius 1 month after decapitation: no development of eggs in ovaries.

Fig. 7.—Female Rhodnius showing egg development in ovaries induced by joining to male Rhodnius with brain removed but corpus allatum intact.

Fig. 8.—Female Rhodnius 1 month after removal of the brain; the corpus allatum left intact.

PLATE 5.

Fig. 9.—Normal 4th-stage Rhodnius unfed.

Fig. 10.—Normal 5th-stage Rhodnius.

Fig. 11.—Adult Rhodnius.

Fig. 12.—’Sixth-stage’ nymph of Rhodnius.

Figs. 13-15.—’Sixth-stage’ Rhodnius with nymphal cuticle all over the abdomen; wings, thorax, &c., intermediate between nymph and adult.

Fig. 16.—Adult Rhodnius showing an area of nymphal cuticle around the site of implantation of the corpus allatum from a 4th-stage nymph.