A method is described for using a reflecting microscope to irradiate small precisely located areas of insect embryos with lethal micro-beams of ultraviolet light, so that cell lineages may be traced from histological studies of subsequent development.
The gamete precursors in Drosophila melanogaster were shown to be those pole cells which are carried into the posterior mid-gut rudiment during gastrulation. The interblastodermal-migrating pole cells are not incorporated into the embryonic gonad but probably play a role in the formation of mid-gut epithelium. The mesodermal rudiment of the gonad is derived from the posterior half of the mid-ventral mesodermal invagination, and is formed independently of precursor gamete cells. Bilaterally and unilaterally castrated flies resulted after pole-cell irradiation.
Support for certain parts of Poulson’s anlagen plan of the Drosophila egg was obtained from studies of development after irradiation of certain areas of blastema and blastoderm. The non-cellular blastema stage was more easily damaged by ultra-violet radiation than the blastoderm stage. Gastrulation was a critical stage, but the study supported Sonnenblick’s ideas concerning the morphogenic factors governing endodermal invagination. No departure from mosaicism was noticed, except that the results indicated some slight regulative capacity in the presumptive pole-cell area at the 1-hour stage.
Recent developments in the study of the biosynthetic processes Of oogenesis in Drosophila melanogaster (Jacob & Sirlin, 1959; King & Sang, 1959; Zalokar, 1960; Sirlin & Jacob, 1960) and a more extensive knowledge of the submicro-scopic structure of the developing egg (King & Devine, 1958; Waddington & Okada, 1960) have renewed interest in the ‘mosaic’ character of the freshly deposited egg. This has long been a subject of investigation for insect embryologists, but the available experimental techniques have not provided the necessary refinement for detailed studies.
Previously, the most successful method of experimental investigation was that of partial egg irradiation with ultra-violet light (Geigy, 193la, b). The ultraviolet was particularly suitable for this type of work because of its comparatively low penetration into protoplasm: roughly 85 per cent, absorption occurred in 50 μ, at a wavelength of 300 mμ (from data of Bachem & Read, 1931, for the stratum malpighii), and stronger absorption was obtained at shorter wavelengths. Geigy achieved localized irradiation by a method of shielding, and his results included evidence of the developmental fate of at least some of the pole cells. The work also left no doubt about the practicability of using ultra-violet light as a means of ‘marking’ cells so that they could be traced histologically through later development. Unfortunately, the small size of the Drosophila egg makes it almost impossible to perform precisely these micro-irradiations with only shielding devices.
Recent development of a micro-beam apparatus has altered this picture in that precision focus and easy administration afford the experimenter a much better tool. This type of apparatus was first perfected by Zirkle & Bloom (1953), and these authors were able to work with such refinement that it was possible to treat only parts of cells. Needless to say, such a technique would certainly be adequate for localized irradiation of parts of an egg. The present work, involving microspot ultra-violet irradiation by means of a reflecting microscope, was a preliminary investigation of certain aspects of morphogenesis and cell-lineage patterns in early embryonic development.
APPARATUS AND TECHNIQUE
The D. melanogaster used throughout the experiments were of a highly inbred line of Oregon K maintained at this Institute. They were reared at a constant temperature of 25° C., and all developmental stages denoted in this paper are in accordance with the findings of Poulson (1950) and Ede & Counce (1956) for development at that temperature. By dechorionation in 3–4 per cent, sodium hypochlorite it was possible to render the eggs transparent to visual observation and thus to enable accurate timing of developmental stages to within 15–20 minutes.
The micro-beam apparatus uses a reflecting microscope constructed by R. & J. Beck Ltd. of London from specifications developed and described by Norris, Seeds, & Wilkins (1951). The microscope is a completely reflecting optical system; thus the user avoids the focusing problems associated with refracting systems in which the refraction, properties and hence the focal lengths are wavelength-dependent. In the optical system, which is shown diagrammatically in Text-fig. 1, the mirrors all have aluminized surfaces. The spherical mirror pairs have adjustments for centring and mutual separation. The aperture consisted either of an iris diaphragm or, alternatively, of an interchangeable series of apertures each of fixed size. A quartz lens was used to collect light from the ultra-violet source and to illuminate uniformly the aperture with it. The reflecting objective produces a reduced image of this aperture in the plane of the biological specimen on the stage. The size of the ultra-violet irradiated spot on the specimen depended upon the size of the aperture and the magnification of the objective. Either of two objectives were used, of magnifications × 54 and × 74 respectively. This apparatus was used to produce ‘micro-spots’ of ultraviolet light, down to a minimum diameter of 5 μ. Smaller spots could have been produced, but for this work they were not needed. The period of irradiation was controlled by means of a variable-speed photographic shutter, situated alongside the illumination aperture, and operated by cable-release. Before any experiment was made, the ultra-violet spot was first centred with the cross-hairs of the ocular, using either a mirror or a fluorescent screen in the object plane.
The part of the specimen which it was desired to irradiate with ultra-violet light was first centred on the microscope stage with the cross-hairs of the ocular. This was done with transmitted visible light, using as a condenser the reflecting objective not in use above the specimen (see Text-fig. 1), or alternatively, a standard refracting condenser.
The ultra-violet source used in the present work was a 250-watt Osram ME/D mercury vapour lamp in a quartz envelope. Its principal ultra-violet emission was between wavelengths of 290 mμ and 370 mμ, but a considerable quantity of visible light was also emitted. In the absence of a monochromator, a wide band of wavelengths was incident on the irradiated area, but such an arrangement was, nevertheless, completely satisfactory for its present use as a precisely located lethal spot. The spot had a high intensity and only brief exposures were required to produce, on the irradiated area, a definite observable effect. The water-filter was used to absorb infra-red radiation and thus to prevent heat damage from being a cause of any developmentally observed alteration. All optical parts were mounted on an optical bench.
In all cases the eggs were obtained from rapidly laying females. The embryos were then allowed to develop for a specific period of time, after which they were dechorionated and their developmental stage checked by visual observation. For the pole-cell irradiations the eggs were usually placed on an improvised cavity slide (with a depression about the thickness of a Drosophila egg) which was submerged in distilled water. While still submerged, a quartz coverslip was carefully placed over the eggs. The treatment slide with eggs and coverslip was then removed from the medium and placed on the microscope stage. The eggs were individually aligned using transmitted visible light for observation and then treated with the ultra-violet light from above (Plate 1, fig. 1). The irradiations of selected areas of blastema and blastoderm were performed with the embryo in air, the positioning having been done while the embryo rested on an agar block. After the ultra-violet irradiation, the eggs were removed from the treatment slide or agar and placed in distilled water in solid watch-glasses. After the specified period of development at 25° C. they were fixed and prepared for histological examination. All eggs were fixed in formol acetic alcohol (6 : 1 : 15), sectioned at 5 μ, and stained with iron haematoxylin.
General criteria for recognizing damage caused by ultra-violet light
Preliminary experiments were made, mostly by irradiation of pole cells, using micro-beams of different diameters between 6 μ and 60 μ and employing different doses of ultra-violet light. Some experience was necessary before the various cytological changes which cells undergo after treatment with doses of ultra-violet light could be recognized with certainty. The most easily recognized damage was observed 1 hour after treatment of the pole cells which are normally segregated from other tissue after 2 hours’ development. The injured cells stain considerably darker, lack distinct nuclear structure, and begin to lose their customary cellular outline (Plate 1, fig. 2). Damage observed several hours after treatment took a variety of forms and was considerably more difficult to interpret. In pole-cell-treated embryos the pole cells are normally segregated in the posterior mid-gut after 5 to 7 hours of development and pycnotic cells can be seen there (Plate 1, fig. 5). There may also be vacuolated nuclei containing beaded chromatin adhering to the nuclear membrane in less severely damaged cells (Plate 1, fig. 4). A third type of damage observed was the presence of general cell debris (Plate 1, fig. 3). It was often difficult to distinguish between such debris and normal cytoplasm in sections where the nuclei were not included in the plane of the section, and study of adjacent sections was then necessary.
The study of definitive embryonic gonads was much less ambiguous because the pole cells, if present, were quite conspicuous when compared with a gonad lacking them (compare figs. 6 and 7 of Plate 2). Determination of the number of nuclei present required careful study under high magnification of successive sections. Embryos with more severe defects were also obtained, although usually such material gave little information of biological value beyond an approximate estimate of the age at which development had ceased.
This study was undertaken to determine the migratory paths of the pole cells. The role assumed by these cells in later development is dependent upon which of two migratory paths is taken; one will result in gonadal inclusion and the other perhaps in mid-gut formation (Poulson, 1950). The experiment was planned to distinguish between the various hypotheses which have been made concerning pole-cell fate. Batches of embryos were treated at each of three stages of development. In the first batch the posterior poles of the embryos were treated prior to pole-cell delimitation and before the migration of the cleavage nuclei to the pole-cell area of the periplasm, i.e. at approximately 1 hour. In the second batch the posterior poles were treated after pole-cell formation was complete but before the interblastodermal migration of some of them into the yolk (and perhaps eventually into the gonad), i.e. at approximately 2 hours. In a third batch the posterior poles of the embryos were treated after the interblastodermal migration had taken place and before the gastrulation movement of the posterior mid-gut (PMG) and the overlying pole cells, i.e. at approximately 3 hours. In embryos of this third batch pole cells which have migrated interblastodermally will therefore escape treatment with ultra-violet light. The micro-beam in all cases was 60 μ in diameter (Plate 1, fig. 1) and was administered for 30 seconds. All these eggs were fixed at 15 hours, sectioned, and examined for gonad and mid-gut structure.
By comparison with the controls (as shown in Text-fig. 2) it is obvious that irradiation at all three developmental stages has markedly reduced the number of pole cells which are eventually incorporated into the embryonic gonad. The results show quite definitely that it is those pole cells which are carried at gastrulation into the posterior mid-gut rudiment, rather than the pole cells concerned in the earlier interblastodermal migration, which play a major role in normal gonad formation, for if the pole cells of the interblastodermal migration had been destined for gonadal inclusion then we should have found more pole cells subsequently in the gonad after treatment at 3 hours than after treatment at 2 hours; however, this was not the case. The embryonic mesodermal gonadal rudiment, on the other hand, was invariably present (Plate 2, fig. 7), and this corroborates the findings of Geigy (1931a, b) and Aboim (1945). It can be shown, from the data illustrated in Text-fig. 2, that the mean number of pole cells per embryo incorporated into the gonads after treatment at 1 hour, i.e. 7· 83±2· 25, is greater than the mean number of pole cells per embryo similarly incorporated after the same treatment at 2- and 3-hour stages, i.e. 2· 15±0· 58, the difference between the means being significant at the 1 per cent, level. This suggests that at the 1-hour stage, but not later, there may be some capacity for recovery or regulative power in the presumptive pole-cell region.
The adult survivors of treated embryos also supplied interesting results. There was definitely a larger proportion of castrated, or practically castrated, flies in the experimental specimens than in the controls. The condition of these flies almost exactly resembled those pole-cell irradiated embryos depicted by Geigy. This author concluded, after careful histological study, that the degenerate ovary stubs represented the only adult tissue originally derived from the mesodermal element of the embryonic gonad. Gametes were the only cell type missing from his sectioned material. Although serial sections of these defective ovaries were not made in the present study, their external characteristics could not be distinguished from those shown in Geigy’s drawings.
After irradiation of the pole cells, castration is sometimes unilateral and sometimes bilateral. In several of the bilaterally castrated animals (as shown by subsequent dissection after death) sterility was displayed in simple mating experiments. Also, it appeared that the healthy ovary of the unilaterally castrated adults had an unusually high number of ovarioles, for in the five cases reported here they numbered 10, 14, 14, 16, and 16, respectively. The range in normal flies included such high counts, but not consistently. This might suggest a hormonal control mechanism leading to compensatory function of the remaining ovary.
The lethal effect of pole-cell irradiation at the three stages was clearly demonstrated. In each group at least 45 eggs were treated, and the percentage lethalities after 15 hours’ embryonic development were as follows : 33 per cent, (irradiation at 1 hour); 57 per cent. (2 hours); 0 per cent. (3 hours); and 9 per cent, (controls). Those treated embryos which did survive until 15 hours’ embryonic development were often retarded. This retardation was most evident in the embryos treated at 2 hours, so there seems little doubt that this particular stage of development is one of epigenetic crisis for the embryo. However, all three treated groups, including the 3-hour group, showed approximately 60 per cent, mortality when counts were made of emergence of flies from treated embryos. The corresponding controls exhibited a 39 per cent, death-rate.
In the hope of finding evidence for the participation of the interblastodermal pole cells in mid-gut formation, as Poulson (1947) at one time suggested, all material was systematically analysed along these lines. Unfortunately, the stain most suitable for gonadal studies (iron haematoxylin) was not too satisfactory for gut studies, but some interesting conditions were observed. Tn embryos which had their posterior poles treated before the interblastodermal migration an occasional blatant anterior mid-gut malformation was observed (Plate 2, figs. 9, 10, 11). In embryos which had been similarly treated after interblastodermal migration, however, no such abnormality was found. This evidence strongly suggests that the interblastodermal pole cells play a role in mid-gut development, so that it seems possible to account for the two differently migrating pole-cell groups in terms of their roles in later embryonic development.
Blastema and blastoderm surface irradiation
In order to trace the developmental fate of various regions of the blastema and blastoderm, six different areas of the egg were irradiated at these two stages of development, that is, at 1 and hours respectively. The treated areas, on both the ventral and dorsal surfaces, were as shown diagrammatically in Text-fig. 3. As in the pole-cell irradiations, the micro-beam was approximately 60 μ in diameter as it was projected on to the surface of the egg. The limits of the radiation damage could have been slightly greater than is indicated by a circular area of diameter 60 μ, due to scattering of the ultra-violet beam, some penetration of the beam at an oblique angle into the egg interior, and possible diffusion of toxic substances produced as a result of the radiation. Eggs were treated in air resting on an agar block, and the same dose was given in each case.
In a survival test approximately thirty eggs were treated at each of the six regions and then transferred to culture medium. Only after irradiation of the ventral 3 region of the blastoderm (for position see Text-fig. 3) was there more than one survivor to the adult stage. In this case, 7 of 31 matured, so that the ventral 3 region of the blastoderm may be relatively insensitive to irradiation. These seven flies exhibited no external morphological irregularities, but were lost before a dissection could be made. The other groups suffered mainly embryonic or larval deaths, and in the group irradiated in the ventral 2 region a very small number of unhatched pupae were observed.
Embryonic lethality data were based on a study of 108 serially sectioned embryos fixed at 6, 10, or 15 hours after the same dose of ultra-violet light as was used in the adult survival test. The most obvious result was the extreme sensitivity of the blastema to the ultra-violet light. The dorsal 1 and 3 regions were the least affected, but in all but one case the ventral 2 and dorsal 2 regions were severely damaged and death occurred before 6 hours of embryonic development. The embryos irradiated in the ventral 3 region displayed an intermediate sensitivity.
Irradiations at the blastoderm stage generally caused less damage to embryonic development, at least up to 15 hours of the normal 22– 24-hour embryonic period, but again the dorsal 2 area was the most sensitive. This finding recalls the somewhat vague ‘differentiation centre’ postulated by previous workers (e.g. Seidel, 1935). The ventral irradiations of the blastoderm caused the least lethality up to the 15-hour stage. Undoubtedly here the developmental impairment was manifested at a later stage, as the survival studies demonstrated.
Because of this lack of lethality by 15 hours, a considerable amount of information was gained concerning cell lineage of the ventral surface of the blastoderm. The main results are illustrated in Table 1. The mosaic nature of the egg was especially well demonstrated by head-region irradiations. The phenomenon of involution of the head was in all cases prevented or severely retarded. Involution normally occurs at 11– 15 hours, but in treated specimens it was often only partially completed, totally absent, or extremely disorganized at 15 hours. The remainder of the embryo, however, usually exhibited no anomalies; thus, head development seemed quite isolated from total embryonic development in these early stages.
Blastoderm irradiation of the ventral 2 area resulted in rather widespread abnormalities. As shown in Table 1 and Text-fig. 3, they could be classified according to the development of the primary germ layers. In no case was there evidence for the existence of any regulating mechanism such as might allow damaged tissue to be replaced by a proliferation of similar neighbouring tissue of equipotentiality. Most of the defects of this group were easy to understand from the morphogenetic point of view, but the occurrence of posterior and middle mid-gut maldevelopment was unexpected and has no immediately evident interpretation. It seems likely that some yolk cells and yolk were damaged by penetrating ultra-violet light. If it is assumed that these yolk cells were destined to become part of the yolk-encircling ‘primitive gut’ (Poulson, 1950) and would eventually participate in the gut formation of that particular area, there might then be a localized deficiency of mid-gut epithelial precursors. A failure of gut musculature development (normally derived from the adjacent splanchnic mesoderm on the ventral surface), due also to radiation damage, would only magnify the extent of the gut malformation. Unless one assumes a direct participation of mesodermal cells in the epithelium itself this seems to be the only plausible interpretation. Briefly, the presumptive potential of the ventral 2 region of the blastoderm seems to be mesoderm (and later fat-bodies and gut and body musculature) from the median cells; ectoderm (and later hypoderm, tracheae, and nervous tissue) from the more lateral cells along the mid-ventral line; and perhaps, indirectly, endodermal tissue, as related to ‘primitive-gut’ development (Table 1 and Text-fig. 3).
Irradiations of the dorsal and ventral 1 and 2 regions of the blastoderm did not seem to alter the morphogenesis of the posterior endodermal invagination. Moreover, the blastodermal irradiation of the embryos in the ventral 3 region affected the endodermal invagination in only one case, the damage being restricted to mesodermal tissue in the rest. Since the micro-spot was administered before mesoderm formation (via the ventral furrow invagination), those cells lying along the mid-ventral line would be presumptive mesodermal tissue. The slightly lateral and unirradiated cells of this region, on the other hand, would be presumptive ectodermal cells and would migrate to the midline during mesodermal invagination. The ectodermal cells of the subsequent midline would therefore be undamaged and capable of rapid mitotic activity just before antero-dorsal migration of the PMG rudiment. This also accounts for the lack of defect observed in the hind-gut and anus, which apparently develop from this region (Poulson, 1950).
Without doubt, the most interesting mesodermal derivative defect was the absence of the mesodermal gonad rudiment in embryos irradiated in the ventral 3 region, as shown by careful study of the serially sectioned embryos. This occurred several times unilaterally, and once bilaterally. In view of the fact that this rudiment was known to form independently of the presence of pole cells in the pole-cell irradiation experiments reported above, it was particularly significant that this mesodermal embryonic gonadal structure could also be prevented from developing. In fact, almost the exact converse of the pole-cell treated embryos was found in one case, for pole cells were seen free in the bodycavity near the site normally occupied by the mesodermal gonad rudiment.
DISCUSSION AND CONCLUSIONS
The carefully timed irradiation of pole cells at various stages and the subsequent study of the definitive embryonic gonad proved beyond doubt that in normal development the gamete precursors are derived from those pole cells which are carried into the posterior mid-gut (PMG) invagination during gastrulation. The fact that treatments after completion of the interblastodermal migration had the most pronounced effect not only pinpoints the PMG pole cells as being those entering into the gonad, but it also discounts the possibility of a general equipotentiality of both groups of cells, for if this equipotentiality existed, a deficiency of PMG pole cells would probably be compensated for by the earlier migrating pole cells being incorporated into the gonad. Such a lack of equipotentiality on the part of both groups of pole cells does conflict, however, with the explanation which has been put forward to account for the curious gonad development in the two mutants, Lff 11 (Ede, 1956) and nasrat A (Counce & Ede, 1957), in which normal gonads formed in the absence of a normal PMG invagination. Again, Poulson & Waterhouse (1959) concluded that the earlier migrating cells, which pass between the blastoderm cells to the interior of the egg, give rise to germ cells in both Drosophila and Lucilia. It would, however, appear that in the case of Drosophila the conclusions were based on work with embryos treated at stages no later than hours (Poulson, personal communication). But the interblastodermal migration takes place between 2 and hours and occasionally at a slightly later stage while pregastrular cell movements are taking place. Since the only way to distinguish between the two separately migrating groups of cells is to treat them after the first migration is complete, it seems likely that Poulson & Waterhouse treated the embryos at too early a stage to fully justify their conclusion. The conclusion of Geigy & Aboim (1944) and Aboim (1945) was that pole cells carried into the posterior mid-gut later entered the gonad.
The repeated occurrence of mesodermally derived gonads without pole cells is very strong evidence of the independent formation of the mesodermal rudiment. No inductive influence by the pole cells was needed, contrary to Counce & Selman’s (1955) interpretation of the results of ultrasonic treatment. In a small number of cases Counce & Selman found that a disorganization of the egg’s constituents resulted in a displaced but complete gonad. From this they inferred that the pole cells, having been displaced into regions of mesoderm with which they are not normally associated, had nevertheless been able to induce them to form gonad sheath and interstitial elements. It could, however, also be argued that the pole cells had migrated or had by chance become situated next to the gonadally determined but displaced mesoderm after the movements caused by ultrasonic treatment. The independent mesodermal derivation of the gonadal rudiment was again demonstrated by its absence after irradiation of the posterior third of the presumptive mesoderm cells of the blastoderm. In fact, free pole cells but no mesodermal rudiment were seen in the body-cavity near the normal gonad region. These results are similar to those that Haget (1955) obtained with the Coleopteran Leptinotarsa by microcautery and micro-vivisection. Haget showed firstly the mutual independence of the pole cells and mesodermal rudiments of the gonad, secondly the regional origin of the mesodermal rudiment, and thirdly the lack of endodermal influence on the competence of that mesodermal rudiment.
The puzzling fate of the interblastodermal migrating pole cells remains an open question although a little evidence has been presented above to support the earlier suggestion of Poulson (1947) that they play a key role in mid-gut epithelium, formation where the anterior and middle mid-gut rudiments fuse. To account for the rather infrequent occurrence in sectioned material of embryos with defective mid-gut after pole-cell treatment at early stages, two possible explanations exist. Either there is an abundance of interblastodermal pole cells which are capable of participating in mid-gut formation, so that it is only in rare cases that a sufficient number are damaged to such an extent that there are none available for this function; or, alternatively, it is possible that slightly damaged pole cells may undergo the normal interblastodermal migration and actually assume their mid-gut role before degenerating completely. In either case, unequivocal histological evidence would seldom be obtained because the development of such a defect could occur over a rather extensive period of time and its morphological manifestation would be transitory. A more complete histological study of carefully stained embryos after ultra-violet treatment might provide a satisfactory answer to this question.
The presumptive areas of the blastoderm, as deduced by tracing histologically the ultra-violet ‘marked’ cells, or their absence, substantiated in general the anlagen plan put forward by Poulson (1950). Furthermore, there was no evidence of changes occurring in the location of these presumptive areas between the blastema and blastoderm stages. This, of course, begs the important question as to whether the mosaicism is continuous and unchanged from deposition onward. Further studies are needed, firstly on the effects of surface irradiations on freshly deposited eggs, and secondly on the detailed cell-lineage patterns in overall embryonic development. In connexion with the latter it would be wise to use a smaller micro-beam so as to mark smaller regions of the embryo.
The present study supports Sonnenblick’s (1950) suggestion that the force giving rise to the posterior mid-gut migration during gastrulation lies in the rapidly proliferating ectodermal cells which are then in the postero-ventral region. These cells originally would have lain laterally to the ventral midline, and at the time of the mesoderm formation via the mid-ventral longitudinal furrow invagination at hours, these laterally situated cells would migrate medially to the midline. Bearing in mind this sequence of events, it follows that the ventral 3 midline irradiation at hours should primarily affect presumptive mesodermal tissue, and the ectodermal cells destined for the rapid proliferation would not be damaged; thus posterior mid-gut migration would be unimpaired. This was the case, and embryos irradiated in the ventral 3 region exhibited mainly mesodermal defects and a normal posterior mid-gut invagination. Sonnenblick also considered that the temporary dorsal foldings between the dorsal lip of the posterior mid-gut invagination and the anterior oblique cleft were also caused by this same expansion of ectoderm in the postero-ventral region, occurring as the posterior mid-gut rudiment passes from the posterior tip of the egg anteriorly along the dorsal surface. However, ectoderm in this region of temporary folding was damaged by the dorsal 2 and 3 irradiations, often leading to herniation of the yolk in that area.
Evidence is rapidly accumulating in the literature dealing with oogenesis in Drosophila which emphasizes the developmental continuity from the earliest stages of the egg’s biosynthesis to its early stages as an independent organism after fertilization. The present results demonstrate that the mosaicism is nearly complete by 1 hour after deposition. It is possible that slight regulative powers exist during this first hour but not to any considerable degree. However, after treatment of posterior pole areas at 1 hour, embryos generally had a higher number of pole cells incorporated into the definitive gonad by 15 hours than did embryos similarly treated at 2 and 3 hours. It can be suggested, therefore, that there was some recovery or regulative power in the presumptive pole-cell area at approximately 1 hour of development. Poulson (personal communication) feels that his results suggest the same thing. Further experimental evidence is needed to substantiate this hypothesis. The present study leaves untouched the allimportant first hour of development. Study of this period will provide important information regarding the relationship between oogenesis and the mosaic character of the egg. The adequacy of the reflecting microscope and microbeam ultra-violet irradiation for this type of investigation has been clearly demonstrated.
In terms of the present investigation, it is helpful to distinguish between two general types of embryonic lethality which resulted from experimental interference in normal development. The first type of embryonic death appeared to be a direct result of the experimental treatment. This was illustrated by the results of the blastema irradiations when development stopped immediately. The exact nature of the cause of death could not be determined, but it is not unreasonable to assume that a toxic substance was produced by the ultra-violet radiation and was able to diffuse freely through the egg because of the absence of cellular membranes to act as protective barriers. The second or indirect kind of embryonic death was similar to that often encountered in studies with genetic lethals and results from the maldevelopment of a particular area or structure with subsequent impairment of function or structure of other related parts of the organism. Damage inflicted on certain precursor cells by the ultra-violet light later proved lethal when the derivatives of these precursors failed to assume their normal role in development. During certain epigenetic crises the failure of these defective structures was sufficiently serious to cause death. Other workers have compiled extensive data on these critical periods (reviewed by Imaizumi, 1958), which are blastoderm formation , mesodermal and endodermal tissue differentiation, or gastrulation , shortening of the germ-band (9– 11 hours), and eruption of gas into the tracheae (20 hours). In the present study, dorsal 2 and 3 and ventral 2 irradiations usually exhibited gastrulation defects when the embryo died. These abnormalities seemed to arise from the malfunction of a small region of irradiated cells which upset the biophysical factors involved in normal gastrulation. Again, all those irradiated groups which produced no adults and which were alive at 15 hours of embryonic age must have suffered an indirect lethal effect at a later stage. The survival results showed that after irradiation of almost all the treated areas the embryos exhibited this late lethality to some degree. If a similar experiment were conducted with a lesser dose, information on imaginal disk development might be gained by examination of pupal and adult survivors.
Quelques résultats obtenus par l’emploi du dard ultra-violet chez des embryons de Drosophila melanogaster au point de vue des filiations cellulaires et de la morphogenèse
On décrit une méthode utilisant un microscope à réflexion pour irradier chez un embryon d’insecte, à l’aide d’un micro-faisceau létal de lumière ultraviolette, de petites régions précisément localisées, de telle sorte qu’on puisse déceler histologiquement, par l’étude des stades ultérieurs du développement, la filiation de certaines cellules.
Il a pu être ainsi démontré que chez D. melanogaster les précurseurs des gamètes sont des cellules polaires qui pendant la gastrulation sont reportées dans la partie postérieure de l’intestin moyen. Les cellules polaires migrant à travers le blastoderme ne sont pas incorporées dans la gonade mais jouent probablement un rôle dans la formation de l’épithélium revêtant l’intestin moyen. La partie mésodermique de l’ébauche gonadique provient de la moitié postérieure de l’invagination mésodermique médio-ventrale et se forme indépendamment des cellules précurseurs qui sont les gamètes.
Certaines parties du plan d’ébauches de l’œuf de Drosophila, tel que l’a établi Poulson, ont pu être confirmées par l’étude, après irradiation, de la destinée de diverses régions du blastème et du blastoderme. Le stade du blastème non cellulaire a été plus aisément lésé par les radiations ultra-violettes que le stade du blastoderme. La gastrulation s’est révélée un stade critique, mais les données recueillies ont été en faveur des idées de Sonnenblick concernant les facteurs morphogènes qui régissent l’invagination de l’endoderme. Rien n’a été observé qui s’écarte du type mosaïque de développement, sauf que les résultats indiquent une légère capacité de régulation au stade d’une heure pour la région présomptive des cellules polaires.
The authors are grateful to Professor C. H. Waddington, F.R.S., for suggesting and providing facilities for the present investigation and for discussion. One of us (D. S. H.) is indebted to Dr. William L. Gaines and the Fulbright Commission in the United Kingdom for the grant which made possible this period of study.
EXPLANATION OF PLATES
Magnification: In figs. 2– 11 inclusive, the scale represents 20 μ