Some Developmental Effects of X-Rays in Drosophila

Several different genotypes were used: wild type, straw-3, dachs, split singed-3 and morula blistered-2/Curly. Most of the experiments were made with the first three types, but no differences in behaviour were found with any of the genotypes, and the results of all the experiments can therefore be grouped together for description.

Larvae were collected by allowing a number of flies to lay eggs on normal food in 3×1 in. vials for a given period, the cultures then being kept at 25°C. until irradiation, after which they were returned to an incubator at the same temperature. Pupae were collected as ‘white prepupae’ from the sides of culture bottles; as the puparium darkens within 3−4 hr. of its formation, their age was known within those limits.

Irradiation was in all cases with a dose of 7000 r. units. The X-rays were generated at 150 kV. constant potential, 7 mA., and filtered by 0·7 mm. copper, 1·2 mm. aluminium.

The half-value layer was 1·0 mm. copper, corresponding to a mean wave-length 0·15 A. The intensity was 79 r./min., and the duration of the exposure 89 min.

The results can be described in groups, according to the age of the larvae or pupae at irradiation.

A considerable number of wild-type pupae of this age were irradiated. Mortality was comparatively low, and some of the flies emerged unaided from the pupa cases. No important morphological abnormalities were found, except for the rare absence of a microchaete.

Mortality was much higher; none of the flies emerged unaided, and many died before attaining the emergence stage. Those which survived were dissected from the pupa cases. The most marked effect was on the bristles, many of the macro- and microchaetae being absent. In most flies the proportion of microchaetae affected was somewhat larger than that of macrochaetae, and in a few the microchaetae were almost entirely removed while the macrochaetae were present in the full normal number. In nearly all affected hairs, both the hair and its socket were absent, but cases were found both of hairs present with no socket, and sockets with no hair.

A few of the eyes appeared to be slightly rough, but the structural basis of this was not investigated in detail.

The largest series of these had been timed from laying, and were aged 105–113 hr. at 25°C.

The bristles were very extensively affected, but the microchaetae were again the most severely damaged (Fig. 1). Only one case was found in which one of the two components was affected alone; in that there was a socket with no hair. Certain abnormalities in the hairs were also found: a few macrochaetae were doubled (e.g. post-verticals in Fig. 1), there was an occasional forked or singed type, and rather more frequently the T-shaped shaven type (Lees & Waddington, 1942).

The anterior-dorsal parts of the eyes were sometimes rough, but this effect was much stronger in some of the younger groups. The wings were highly abnormal. Their shape and size was that characteristic of a 48 hr. pupa (stage P 2d), and they had not increased in size to any great extent after the formation of the pupal sheath, since they lay more or less flat within it, without showing the normal folding. Their smallness was clearly due to the death of a considerable number of cells. This could be deduced from two facts: many of the characteristic marginal bristles were missing, so that there were many gaps in the normal row; and, in spite of the small size of the wings, the size of the cells, as judged by the distance between the cell hairs, was not proportionately reduced. The cell hairs were, however, much more variable in size than usual, suggesting that many of the surviving cells were somewhat damaged by the treatment.

Although the wings were small, the pupal wing sheaths were of normal size, and, as Mas been said, the wings lay more freely than usual within them. These sheaths are, of course, secreted by the wing itself at the time of its greatest inflation, just before true pupation (about 15. hr. after puparium formation). This suggests that the cells do not die until after the inflation period; and this is confirmed by direct examination. In late prepupal wings, no sign of degenerating cells can be found. In wings of the contraction phase of the pupa (stage P2d), on the other hand, the whole surface is covered with a layer of feebly staining jelly-like debris, many necrotic and pycnotic nuclei can be seen in the epithelium, while the vein lacunae, which are much wider than usual, are nearly filled with granular lumps of dead material. The death of the cells is therefore postponed for some hours after the irradiation.

After the application of somewhat smaller doses of irradiation in other organisms, it has commonly been found that death of the cells takes place at the next division following the exposure to the X-rays. In Drosophila the cells of the wing are very unfavourable histological material, and the occurrence of a division phase at the time of contraction has not been directly demonstrated. Such a period of division is, however, known to occur in other insects at a comparable stage, and it is most probable (cf. Waddington, 1940 a) that Drosophila behaves in the same way. If that is the case, the postponement of necrosis from the time of irradiation till the contraction phase would fall into-line with the phenomena described in other material.

Mortality was considerably higher than in the pupal classes. No flies emerged, but enough survived to the emergence stage for their development to be followed.

The bristle effects were still very strong. In most flies both macro- and microchaetae were affected, but in this group the former were more sensitive, and the latter less so than in the older groups (Fig. 2). Again, occasional bristles showed only one of the components affected, and shaven and forked-like types occurred. In one fly extra dorso-central bristles were present, but the significance of this is doubtful, since the number of these bristles is somewhat variable in normal wild-type animals.

The eyes were nearly always rough, and sometimes were extremely reduced in size. Their anatomical structure was very varied. In some cases the pigment cells were almost lacking, in others the rhabdomeres were absent, while in others again, patches of retinal tissue appeared to be transformed into post-retinal material. The histology of these different abnormalities has not yet been studied in detail, but it is probable that different components of the eye have slightly different periods of maximum sensitivity, which could only be disentangled with accurate timing: further investigation of this problem is in progress.

A few of the wings were affected as in the last group, but many of them appeared to be fairly normal; owing to the difficulty of expanding them, after dissecting the flies out of the pupa cases, it was impossible to examine them in detail, but there is no doubt that they were of at least approximately normal size. The legs, on the other hand, were very badly crippled. They were bent, frequently very short, and often bloated, tit segmentation of the tarsus being sometimes almost completely obliterated. The abnormal ties strikingly resembled those found in the mutant ‘comb-gap’. The most important event leading to their formation is a disturbance of the normal process of eversion of the leg rudiment from its peripodial sac in the early prepupa. Any malformations produced by difficulties of eversion tend to become even more exaggerated as pupal development proceeds. The legs undergo the usual process of inflation at the end of the prepupal period, but frequently fail to carry out the subsequent contraction, since their irregular shape, sharp bends and mutual interference, obstructs the flow of body fluid out of them into the main body of the thorax.

Several different age groups falling within this period were separately irradiated, but as they gave essentially the same results they may be treated together. Mortality was always very high, and this is probably the period of maximum sensitivity during larval life. The first notable point is that bristle effects are nearly completely absent in this group; both macro- and microchaetae are formed in full numbers. On the other hand, a new type of effect is found which did not occur in the older animals. This is a tendency for the overgrowth of various regions of the body, which sometimes results in reduplication of the part.

This effect has only rarely been found in the antennae, but there is one case (Fig. 3) in which the basal part of the right arista is thickened and transformed into-a somewhat leg-like tissue; the arista resembles those produced by the weakest expression of the gene aristopedia (e.g. typically by the allele ss^s-B). This certainly involves some overgrowth, but of course no reduplication.

The eyes are very constantly affected. In one or two individuals they are very much reduced (on both sides), only a few malformed ommatidia being formed. The histology of these has not been studied. A very much commoner effect, found to some extent in nearly all the remaining individuals, is the formation of a palp protruding from part of the eye surface. In the commonest and ‘weakest’ form, this is merely a flattish plate of chitin, bearing a few hairs, covering the anterior-dorsal region of the eye (Fig. 4). In other cases the plates are large and protruding, bearing a number of macrochaetae as well as some microchaetae (Fig. 6). In such flies, the size of the eye is not reduced by an area equal to that of the palp, although a certain degree of reduction does sometimes occur. Usually, however, it appears that the eye is of normal size but is displaced downwards, towards the ventral surface of the head. The palp must, at least in part, represent additional material. This is particularly clear in a few cases in which the eye is exceptionally large, with a band of palp-like material running across it from front to rear (Fig. 7). were the lower part of the eye is of normal full size, while the part above the palp is additional, and represents a reduplication.

The palp usually shows no very definite resemblance to any particular part of the body; and this is normally the case also for the eye palps sometimes found in Lobe flies and regularly produced by the gene ophthalmopedia. The suggestion, implied by the latter name, that they represent a malformed leg, must not be taken too seriously. In a few of the X-rayed cases, definite organs can be detected in the palps. In one case (Fig. 7), a flat expanse of palp on the dorsal surface of the head bears an extra ocellus. In another, a large protruding palp on the left eye contains a Johnson’s organ, which identifies it as an antenna, while a further palp-like expansion on the same side of the head contains a second similar organ, in this case accompanied by a small plumose spike resembling an arista (Fig. 3).

The formation of palp-like outgrowths is sufficiently common for one to be able to identify it in sections of earlier stages. Such outgrowths can be traced back to the early prepupa. In sections through the anterior end of the animal at that time, the eye-forming region can be identified by the regular grouping of its cells (see Steinberg, 1941 ; Pilkington, 1942) at the posterior end of the cephalic imaginal buds ; it normally forms a smoothly curved surface which fits snugly against the brain. In the X-rayed animals, however, the central region of this smooth sunace is sometimes deeply folded (Fig. 8). In the tissue lining the fold, no trace of the usual pre-ommatidial arrangement of cells can be detected. It appears certain that these folds represent material which would later have developed into a palp. An exactly similar appearance has been found in ophthalmopedia. The lateral palps, such as that seen in the left eye of Fig. 4, would presumably be represented by abnormal folding of the edge of the eye forming region in the prepupal disk, but these have not been detected in the sections.

Passing to more posterior regions of the animal, overgrowths of various parts of the thorax can sometimes be found. They normally produce only swelling, and the appearance of extra bristles. In a few cases, real reduplications may occur, as of the left half of the scutellum in Fig. 4. Fig. 5 shows one example of the interesting phenomenon of the reduplication of the wing, together with a region of thorax around its articulation. Since the fly was, like all those described in this group, dissected out of its pupa case after death, the extra wing cannot be flattened, and its shape and venation are therefore unknown. Its histology is perfectly typical; it is covered with cell hairs, and also bears the characteristic marginal bristles, both the short thick ones of the anterior margin and the thin longer ones of the posterior margin. It is therefore definitely a wing, and not a mere persisting larval spiracle, such as those described by Enzmann & Haskins (19386), although such persisting spiracles have also been seen.

Friesen (1936) found, particularly after irradiations at this stage, a considerable number of cases in which marginal parts of the wing were missing, as in the well-known scalloped mutants. Owing to the fact that the wings were not expanded in the flies described here, such modifications could not be detected in them.

Overgrowth also takes place within the wings themselves. Owing to the difficulty of flattening the wings, it is impossible to give accurate pictures of their shape, but even an inspection of the pupal sheaths will show that comparatively large extra lobes may be formed; these may in fact be so large that they suggest that partial reduplication has occurred, rather like that found in blot (Waddington, 1939). A somewhat more peculiar result has been seen both in whole mounts and in sections. That is, the formation by the wing membrane of a thick chitinous layer, resembling the normal body surface, and, like it, bearing fairly large bristles, complete with sockets (Fig. 11). No developmental stages of this are available. The process has not been described in any mutant stocks. It might be supposed to be due to the formation of isolated spheres of wing epithelium, but in fact such structures, at least when they occur in bloated (Waddington, 1940 a), always develop the normal wing histology.

A phenomenon which may perhaps be related to the last is the conversion of the entire wing into a lump with a normal body surface bearing chaetae. This has been seen in several adult flies. It occurs occasionally in certain stocks, for instance, one of expanded which was kept at Pasadena, but whether the effect is due to a separate gene, or what its mechanism may be, remains undetermined.

The legs are often extremely irregularly shaped, with the usual comb-gap-like. features described for the older groups. In most cases it is difficult to make out whether they are increased in size or not, but in one or two cases, clear duplications have been found (Fig. 9). Incomplete fusion of the halves of the thorax is also fairly common in these flies, and is probably due to some irregularity of eversion similar to that involved in the leg malformations. Total lack of eversion of legs or thorax parts may occur.

In larvae X-rayed at ages less than 60 hr., the survivors show progressively less profound modifications of development the younger they were at irradiation, and the flies developed from larvae aged less than 48 hr. at raying were more or less normal.

It will be convenient to begin by discussing separately the mechanisms of the effects of irradiation on the various organs.

Only two cells are concerned in the development of each bristle and socket, and the developmental system is therefore very simple. It is clear from the 27 hr. pupae that this system is not very sensitive to irradiation at that time, although we know (Lees & Waddington, 1942) that the secretion of the hair and socket are by no means complete at that stage. The 19 hr. pupae, which are the next group, exhibit considerable sensitivity. At that time, the hair cells have already divided to form a group of two, but these are still quite small. The decrease in sensitivity between 19 and 27 hr. would be expected (cf. Stadler, 1932) as a consequence of the increasing polyploidy of the cells, which clearly accompanies the growth in size of the nuclei (Lees & Waddington, 1942): it might also be connected with the onset of secretion, but there is no evidence of this.

The sensitivity of the hair cells extends from the 19 hr. pupae back till about 12 hr. before puparium formation (90 hr. larvae). This agrees fairly well with the data of Friesen (1936). Using somewhat lower doses of irradiation, he found the highest percentage of ‘Borstenreduktion’ in the groups within 24 hr. before or after puparium formation. It is impossible to make out exactly what was involved in all these cases of bristle reduction, since the one protocol which he provides shows that some of the flies exhibited a lack of many bristles, while in others only a single one of the rather variable head bristles was missing. If the present material were described in the same way, the sensitive period of the bristles would certainly be extended farther backwards towards the earlier parts of the larval life. But the most striking feature, which is rather concealed in Friesen’s account, is the very high sensitivity of the period just before and just after puparium formation.

It is clear that within this general period, the microchaetae and macrochaetae have different phases of maximum sensitivity, the macrochaetae being on the whole more severely affected in the earlier part, before puparium formation, and the macrochaetae in the later part. This is, of course, exactly what would be expected if the decrease in sensitivity towards the end of the period is due to increasing polyploidy, since there is little doubt that the rapidly growing macrochaetal cells become much more polyploid than the comparatively small microchaetal ones.

The great majority of the effects are complete absences of the chaetal elements, and are presumably caused by the killing of the cells. The occurrence of a period before which the cells are insensitive might indicate one of two things. It might be, first, that definitely determined bristle cells were present throughout larval life, but only became sensitive towards the end of it; or, secondly, it might be that the cells.only become determined as bristle-forming cells at the time when they first exhibit sensitivity, cells knocked out at an earlier stage being replaceable by others. If the former alternative were true we should expect to find that the maximum sensitivity corresponded to the period just before the division into the two components, trichogen and tormogen, from which the bristle and socket arise. This, however, cannot be the case. The sensitivity is still marked at 19 hr. after puparium formation, when the division is known to have taken place; moreover, the division phase can hardly be supposed to last so long as the phase of sensitivity. It seems probable that a pre-division sensitivity, if it occurs in these cells, would only cause a minor variation in the general reaction with which we are dealing. The second alternative given above is therefore to be preferred; and we may fake the evidence to indicate that during the early part of larval life the bristle-forming cells are not irrevocably determined, so that if some of the presumptive bristle cells are destroyed, other cells may take their place.

Contributory evidence that the bristles become determined at about 80–90 hr. after laying may be found in the experiments of Plunkett (1926), Goldschmidt (1935), Child (1935), Ives (1935, ‘1939), and Child, Blanc & Plough (1940) on the temperature effective periods of various bristle-removing genes. In all cases they found that the removal of bristles is not produced by heat treatments administered earlier than about $312$ days after laying. It is, however, not at all clear just what is going on in larvae subjected to heat treatment, whereas one can be fairly certain that at least the primary effect of the irradiation is, to kill a certain number of cells. But, whatever may be the modus operandi of heat treatment, its lack of effect in early larval life can be most plausibly explained by the supposition that the bristle-determining process has not occurred at that stage.

It is noteworthy that the X-ray effects very, commonly lead to the disappearance of both the bristle and its accompanying socket. This would be only to be expected if the Irradiation led to the death of the bristle cell before it had divided into trichogen and tormogen. But in the 19 hr. pupa this division has certainly already occurred. If the effect of the radiation was confined to the killing of individual cells, it is easy to calculate the proportion of cases in which only one, or both, the trichogen and tormogen should be affected. If the proportion of hairs absent was x, and that of sockets absent was y, it should only be in the proportion xy of cases that both a hair and its own socket were missing. But actually both components are usually affected. The proportion of isolated hairs or sockets is highest in the 19 hr. pupae, but even there it is less than that of cases in which both structures are missing. For instance, in one 19 hr. pupa there were fifty-seven microchaetae with both elements, seven isolated sockets and three isolated hairs on the dorsal surface of the thorax in a region in which the total normal number of microchaetae is about 200; in another fly the figures were fifty-two complete microchaetae, sixteen isolated sockets and seven isolated hairs. Thus the failure of both elements had occurred much more frequently than that of only one of them. This must indicate that the killing of one component inhibits the development of the other, even if the latter remains potentially viable. It is probably significant that isolated sockets are commoner than isolated hairs, since the development of the socket, which lies on the surface, is, one would imagine, less inhibited by the death of the underlying trichogen than the latter would be by the death of the tormogen.

The exact time of division into trichogen and tormogen has not been determined histologically, owing to the difficulty of identifying the hair cells before their growth phase. The meagre evidence that was available suggested to Lees & Waddington (1942) that it occurred about 15 hr. after puparium formation. It might be possible to date the division more accurately by the occurrence of isolated hairs or sockets in these irradiated flies, since, if the above explanation is all that is involved, these should not occur in flies irradiated before the division. The isolated components are, however, found, though very rarely, even in the youngest groups which gave any bristle effects at all. The same fact can be unearthed from Friesen’s account, since his ‘Reduktion von o-Typus’ can certainly be interpreted as the formation of a socket without a hair. This might be taken to indicate that the division occurs almost immediately after the cell is determined as a bristle cell; the comparative rarity of the phenomenon in the young groups would then be attributed to the greater time available for a dead cell to inhibit the development of its surviving partner. But an alternative explanation is possible. It is known that irradiation may produce chromosome rearrangements of such a kind that a subsequent division leads to the formation of two daughter cells of differing genetic constitution, one of which may be viable while the other is not; and the rare early examples of isolated components may be due to a process of this kind. The attempt to date the time of division does not, therefore, lead to a conclusive result.

The most common effect on the legs is to cause a crumpling and bending which can, as was stated above, be traced back to irregularities in the eversion of the leg rudiment from its peripodial sac. Similar malformations are well known in mutant races. They are probably exemplified in crippled, and I have investigated their development in comb-gap. In the latter, the crippling is accompanied by an irregularly increased growth. This is probably also true of the X-ray malformations induced in larvae aged 90 hr. or less, although there the legs never attain the enormous size sometimes found in comb-gap. It is more doubtful, however, whether overgrowth can have taken place in the short interval between irradiation and eversion in the flies issuing from treated larvae aged more than 90 hr.; and in these the legs, although showing the same kind of malformations, did not seem to be any larger than normal. The disturbance of eversion was in these cases probably brought about not by increase in size but rather by decrease, consequent on the killing of some of the cells. It is clear that such disturbances can be fairly easily caused by a number of agencies, since even normal untreated animals which develop in unhealthy conditions may have one or more malformed legs of the typical comb-gap type.

In a few cases among the early irradiated animals, overgrowth has been followed by duplication; and there are some cases of atypical histogenesis, e.g. the tarsal claw in a non-terminal position in Fig. 10. These phenomena are discussed in more detail later.

The later irradiations produced a roughening of the eye surface. The histology of these eyes has not yet been investigated in detail, but the roughness is presumably caused by the death of some cells, which would lead to the development of irregular ommatidia. It should be noted that in some cases only part of the eye is affected, and when this is so, it is the anterior-dorsal border which seems most sensitive; this is the same region as that favoured for palp formation in the earlier irradiated groups.

In these earlier groups, the formation of palps is extremely frequent. The evidence summarized above (p. 106) shows that the palps are sometimes derived from regions of the eye-rudiment which have been abnormally folded; and that they are often accompanied by increases in size. It is simplest to assume that the formation of a palp is in all cases preceded, and presumably caused, by an atypical folding. Overgrowth would then be one condition which would naturally lead to some abnormality in folding and would thus tend to produce palps.

The formation of reduplicated eyes, such as that in Fig. 7, is not difficult to understand in the light of the above discussion. Overgrowth of the eye region has certainly taken place, and part of it, in a strip across the whole width, was presumably folded inwards so that it developed as body surface, thus separating the lower, more or less full-sized, ommatidial region from the upper smaller one.

The dorsal palps, such as those of Fig. 4, may perhaps be attributed entirely to extra material formed by overgrowth, but in many cases, such as that of Figs. 3 and 6, there has certainly been a conversion of presumptive eye material into palp. The majority of the palps develop as rather unspecific ‘body surface’ bearing some chaetae. The minimum hypothesis is to suppose that, once ommatidial development has been inhibited, the development of body surface proceeds without needing any specific stimulus; and this is supported by the fact that similar tissue can be formed on the wing (p. 106). But in a few cases more definite organs have been formed, ocelli or antennae of various degrees of completeness. Some definite stimulus, or at least some definite level on a gradient, must be necessary to produce these specific structures. The exact nature of these stimuli cannot yet be determined, but the general phenomenon of atypical histogenesis is discussed again later.

The main effect noticeable on the thorax, apart from disturbances of eversion, is an increase in the size of some parts. This may produce a swelling, such as that of the right shoulder of Fig. 4. Less often, it causes a reduplication, such as that of the left half of the scutellum in the same figure or of the wing-bearing region in Fig. 5. The conditions under which overgrowth of imaginal tissue causes reduplication in Drosophila are not at all clear. Enormous overgrowth of some of the imaginal buds is found in some combinations of mutants, for instance in dachsous-dachs or dachsous-fourjointed (Waddington, 1942). In both these, we tend to find reduplicated antennae, often combined with the absence of the corresponding eyes, while the thorax is produced into very large forwardly-reaching shoulders, like the right of Fig. 4 but much larger. The antennal effect is due to an overgrowth in the antennal region of the cephalic disk, and regularly leads to duplication; the thorax effect, which is caused by an overgrowth of the dorsal mesothoracic bud, does not lead to duplication, and no cases of extra wings have been observed. This appears to indicate that the antenna is a more closely organized morphogenetic unit than the thorax, and cannot easily be increased in size relative to the rest of the organism; if an unduly large region is determined as antenna, it can only develop into two antenna. The thorax, on the other hand, appears to be more capable of modification to incorporate extra material. Similar differences in the ‘extensibility’ of individuation fields are well known in amphibian development. Thus, for instance, the main dorsal axis can easily be caused to incorporate additional tissue, whereas the limbs show a much stronger tendency to become reduplicated.

This modification of the thorax does not occur so regularly in the irradiated flies as in the mutant types such as combgap-fourjointed; in the former we find duplications of the wing and thorax regions which do not occur in the latter. The explanation of this difference is obscure: it may be connected with the general unhealthy condition of the tissue caused by the irradiation.

The gross anatomy of the wings cannot be made outin detail in these flies dissected out of the pupa cases. Two phenomena, however, require notice. One is the formation, by wing epithelium, of typical body surface, as in Fig. 11. This has never been described in any mutant. It is an example of the general phenomenon of atypical histogenesis, which is discussed later. The conversion of the entire wing into a lump of body surface occurs sporadically in some mutant stocks, but nothing is known of its mechanism ; one might guess that it is connected with a failure of the folding of the wing into the cavity of the bud.

The second phenomenon is the formation of extra outgrowths of the wing. This is, of course, another example of overgrowth, which is also discussed in general below. The outgrowths seem, as far as one can judge them in their folded state, to be duplications of parts of the wing. Duplications, springing from the proximal end of the posterior margin of the wing, are characteristic of the mutant blot, The condition m the imaginal buds of this mutant have not been investigated, but one can hardly doubt that the duplication supervenes on an overgrowth. Morgan, Schultz & Curry (1940) have pointed out that the mutant blot is located very near the breakage points of the chromosome rearrangement Xasta, which produces the phenotypic effect of an apical incision into the wing. This apical incision is due to an abnormal position of the wing fold in the imaginal bud (Waddington, 1939, 1940 a), and the above authors remarked that the very different developmental processes leading to the blot and Xasta effects made it difficult to consider the latter as a position-effect modification of the former. This difficulty may, however, be mitigated if one considers two facts; we have evidence in the data presented here that overgrowth may lead to reduplication, and we know, from mutants such as vestigial, that undergrowth may cause shifts in the position of the wing fold. It is only necessary to reverse the latter datum, and assume that overgrowth may also sometimes lead to shifts of the wing fold, to reduce both Xasta and blot to similar primary effects, namely, increases in growth. The question of whether, in a given case, an overgrowth leads to a reduplication of part of the wing, or to a shift in the wing fold, is presumably dependent on the exact quantitative relations.

The effects produced by the irradiations in the later stages are of a kind which might have been expected. They seem to be easily explicable in terms of the killing of individual cells, modified to a rather slight extent by secondary effects due to the interaction of cells in development, as exemplified in the unduly high proportion in which both components of the hairs are removed. The effects of the earlier irradiations are much more peculiar and provide a considerable amount of quite new information about the developmental mechanisms of Drosophila.

The first point which must be mentioned is that the irradiations of earlier stages produce morphological changes which affect quite large groups of cells, comprising considerable areas of tissue. These areas are much too large to be derived, in the time available, from successive divisions of a single original cell. Patches of altered tissue formed in the latter way are well known in experiments in which somatic mutations have been produced by X-rays. With doses of a few hundred to one or two thousand r. units, patches derived from a single somatically mutated cell never include more than a few score adult cells even when the irradiation takes place shortly before the hatching of the larva from the egg. With the much later irradiations described here, the patches should be still smaller. Since there is definite evidence of increased growth in the surviving cells in these experiments it is possible to ask whether this increase might have been sufficient to increase the size of the altered patch up to that found; but the answer must almost certainly be in the negative. Moreover, the alterations produced (duplications of organs, conversion of one tissue into another) are not of a kind which can easily be attributed to mutations of individual genes; they are much more akin to the normal developmental changes which are carried out by the genotype as a whole. The explanation of these ‘mass effects’ must be sought along other lines than the multiplication of a single original affected cell.

On the other hand, there is no doubt that the immediate effects of X-rays in living tissue are to cause ionizations, each of which occupies a very small space and affects only a single cell. The production of the same developmental change in all members of a large group of cells can only be attributed to mutual interactions of the cells subsequent to the original X-ray effects. The injury caused by the irradiation must, in fact, produce changes in the physical conditions within the mass of cells, or the release of substances capable of. diffusing throughout the mass, and it is to such pervasive secondary effects that we must look for the immediate causes of the alterations in development. Since we know, both from many other investigations and also from the effects of later irradiations on the bristles, that large doses of X-rays cause the death of a considerable number of cells, it is simplest to imagine that it is the dead or necrotic cells which release the diffusing morphogenetic substances or cause the general alteration in physical conditions; but it is of course possible that one of the immediate effects of the radiation is to alter the metabolism of the surviving cells in such a way as to bring about the secretion of an abnormal morphogenetic substance.

The detailed mechanism by which the immediate effect on individual cells becomes transformed into the final alteration of the development of a large mass of tissue must remain for further investigation. But the mere fact that such alterations can be produced at these stages is of considerable importance for our understanding of the mechanics of development in Drosophila. The changes are of three main kinds : alterations in histogenesis, suck as the production of eye palps; increases in growth; and the production of duplications. As was made clear in the discussion of the alterations in the eye and on the thorax (pp. 110, 111), the duplications seem to be one of the possible results of increased growth in the imaginal buds, and the two last of the three categories above are therefore intimately related. Again, the factors which determine whether increased growth will lead merely to the development of an over-sized organ, or whether duplication will occur, remain for more thorough investigation in the future. The fact, however, that the increase in growth can not only compensate for the destroyed cells, but can result in a considerable excess of material, seems to indicate a greater degree of indeterminacy than might have been expected. The only previous experimental evidence of a capacity for overgrowth in the imaginal buds showed that this may occur in the eye buds of Bar flies (see Bodenstein (1939), Steinberg (1941) for recent discussions). That gene, however, causes the development of abnormally small eyes, and an ability of the reduced eyes to grow towards the normal size does not at all imply the ability, which is shown by the present material, for wild-type imaginal buds to grow larger than normal.

Much the most important phenomenon, however, is the first of those mentioned above, the alteration in histogenesis. It has been for long believed that the development of Diptera is highly ‘mosaic’, that is to say, determined at a very early age. It is clear that from an epigenetic point of view, the embryonic development of the larva and the pupal development of the adult are two different systems; and the greater part of the evidence for the mosaic character of dipteran development admittedly relates to the former, and can only be extrapolated to the latter with some hesitation. But the slight evidence available as to pupal development also seemed to indicate that it was rather rigidly determined, in considerable detail, at an early stage. Thus Geigy (1931) observed the formation of defects in imagos developed from eggs which had been injured by ultra-violet puncture a few hours after fertilization. Howland & Child (1935) obtained similar results with mechanical injuries. Wolsky & Huxley (1936), on the other hand, suggested that the determination.of ommatidia in the eye only took place during pupation, but Chevais (1937) pointed out that good development of ommatidia occurs in the eye disks transplanted from late larvae into abnormal situations in the body. More recently, Bodenstein (1939) pushed back the earliest time at which ommatidial development could be obtained to the 50 hr. (after hatching) stage; and Steinberg (1941) has since obtained good development of eye disks aged only 24 hr. after hatching.

Although, on the basis of the experiments just mentioned, Steinberg is impressed with the early determination of the imaginal buds in Drosophila, he brings forward evidence that this determination is not in all cases absolute. In Bar eyes the number of ommatidia is much reduced, some of the normally ommatidial material being converted in body surface. Steinberg shows that the area which is affected by this conversion can be changed by procedures acting even in the later half of larval life. He therefore supposes that the imaginal bud is determined as a whole, to develop into an eye with the associated part of the head, but that the exact fate of the individual cells is still susceptible of modification. Exactly parallel cases are, of course, well known in amphibian development, for instance, in the eye cups discussed by Spemann( 193 8, Chap. III). It is clear, however, that this degree of lability does not go nearly far enough to explain the present results. Steinberg contemplated only the possibility of shifting slightly the boundary between tissue which develops into ommatidia and that which becomes body surface. The formation of eye palps would already be a more exaggerated case of this than any he was concerned with. A still more radical reorganization would be necessary for the production of duplications, such as those of the wings. Even they could be regarded as redistributions of potentialities already present within the bud. The greatest degree of modification obtained in these experiments went even further; in the formation of antenna-like palps on the eye, a change of developmental fate outside the range of the eye region has occurred, and the same is probably true of the formation of body surface on the wings.

It appears then to be incontrovertible that very considerable alterations in development are still possible as late as the $312$ day larva. Moreover, as was seen above, the immediate causes of such alterations must be sought in abnormal pervasive conditions (chemical or physical) which can arise within the tissue. That is to say, the cells of the imaginal tissue can themselves fairly easily produce substances or conditions which can spread into the surrounding tissue and alter its developmental fate. Such a phenomenon is obviously similar, in general terms, to the well-known organizer process which plays such a prominent role in the development of vertebrates.

There is no necessary contradiction between this conclusion and the earlier experimental results. Experience with the processes of vertebrate development has abundantly shown that the development of a mass of tissue may remain unaltered under one treatment (whence the material appears ‘determined’), while some other treatment may change it (whence the material appears still labile). Determination, in fact, is experimentally always a relative and not an absolute term. We need not be at all surprised to find that, even though an imaginal bud is still labile, its fate is not altered by mere isolation into an abnormal situation.

A few words are necessary about the increased growth which has been described in these irradiated animals. No measurements of growth rate have been made, nor has there been any exact measurement of the sizes of the various parts of different flies. Any attempt to give a quantitative estimate of the increase in growth would meet with very great difficulties, both because of the difficulty in isolating the derivatives of the various imaginal buds, and because of the well-known dependence of the absolute size of the animals on the nutritive conditions. Mere inspection, however, reveals that, in a given individual, the derivatives of a certain imaginal bud may be over-sized in comparison with the rest of the body. It is this relative hypertrophy of particular regions that has been spoken of as an overgrowth. It must certainly be taken to indicate that the over-sized imaginal bud has formed a greater mass of tissue than it would have done if untreated. The alternative explanation would be that the irradiation had tended to cause a general diminution in size, but that certain imaginal buds were less strongly reduced than the majority. But this can be excluded, since the reduction in size of the whole animal is not sufficient to account for the relative hypertrophy of the overgrown parts.

An increase in growth is not a common effect of irradiation by X-rays. In fact, it seems doubtful if any other satisfactory case can be found in the literature. The phenomenon is, of course, not to be confused with the transitory increase in growth rate, which may follow and compensate for the decline in mitotic frequency caused by small doses of radiation (Spear, 1931). In the earlier literature, a few cases of accelerated growth following irradiation have been claimed. Thus Gilman & Baetjer (1904) claim that irradiated Amblystoma eggs became larger than the controls during the first few days after radiation; but no data as to the dose employed are given, and since the embryos (presumably irradiated immediately after fertilization?) would be living on stored yolk during this period, the significance of the observation is obscure. Bardeen (1911) has described monsters in radiated amphibian embryos in which there were some outgrowths of the epidermis; but it appears from the descriptions that the animals were always extensively deformed, and the outgrowths were probably due rather to the failure of the normal elongation of the dorsal axis than to any excessive formation of epidermal tissues.

Thus no case at all closely analogous to the present one seems to have been described. Presumably the excessive growth in irradiated Drosophila must be attributed to a stimulative effect of some of the substances released by the cells killed by the X-rays. It is perhaps worth pointing out that the greatest growth of the imaginal tissues normally takes place in an environment containing a large amount of necrotic tissue, that is, inside the puparium where active histolysis of larval tissues is proceeding. It may well be that the growth of dipteran imaginal tissues is favoured by such an environment.

No increases in total growth occurred in the animals irradiated at very early stages. This indicates the activity, in those stages, of a. growth-regulating system capable oil suppressing the overgrowth found in the animals irradiated at an older stage. The occurrence of such a regulation in growth during the early period is perhaps connected with the regulation of bristle-forming cells which also occurs during the same phase of development.

The absence of abnormalities of histogenesis in the early irradiated animals which show complete growth regulation, strongly suggests that the effects on histogenesis are causally dependent on the growth effects. In the particular case of the conversion of eye tissue to palp tissue, it has already been argued that the histojogical change is a consequence of the abnormal folding which is a sequel to overgrowth. The general parallelism in the occurrence of abnormal histogenesis and abnormal growth makes it probable that mechanisms of a similar kind may be concerned in the development of other organs, but definite anatomical evidence of this is not yet available.

1. Drosophila larvae of various ages and several genotypes were subjected to heavy doses of X-rays (7000.r. units).

2. (a) In pupae aged 27 hr. after puparium formation, no morphological abnormalities were produced.

   (b) In late larvae (aged more than 90 hr. after laying) and young prepupae, the main effects were: roughening of the eyes, destruction of many macro- and microchaetae and of cells of the wing, arrest of morphogenesis of the wing, and faulty eversion of the legs leading to crippling.

   (c) In mid-larvae (aged 60–90 hr. after laying) the bristles were usually normal. The main effects were: overgrowth of parts, leading to the formation of over-sized organs and/or excrescences, e.g. on the wings; reduplication of organs, e.g. of wings, eyes, parts of the thorax, legs, etc.; changes in histogenesis, e.g. the formation of palps from presumptive eye material, of leg-like thickenings of the arista, of body-surface material from wing epithelia, etc. The eye palps may have the character of antennae, or extra ocelli may be formed.

   (d) In younger larvae the effects were less marked, and the youngest groups irradiated gave quite normal adults.

3. The effects on late larvae (groub b) can be simply explained as the results of the killing of individual cells, although the unduly frequent destruction of both a bristle and its accompanying socket indicate some interaction of the trichogen and tormogen cells during development.

4. The effects on mid-larvae demonstrate that the determination of the imaginal buds is by no means final at this stage, and a considerably greater degree of plasticity, both in growth and histogenesis, must be assumed than has previously been suggested. Moreover, the fact that the abnormalities affect large masses of tissue indicates that they are caused by agents (chemical substances or physical conditions) capable of spreading throughout the affected region. These agents must be producible by the affected cells themselves. This suggests that determination in normal development occurs fairly late, and through the agency of tissue interactions.

5. Attention is drawn to the occurrence of an increased quantity of growth following irradiation, but no explanation is offered for this phenomenon.