Cleavage in Cecidomyidae (Diptera) is characterized by the elimination of chromosomes from presumptive somatic nuclei. The full chromosome complement is kept by the germ-line nuclei.

The course of cleavage in Mayetiola destructor (Say) is described. After the fourth division two nuclei lie in the posterior polar-plasm and become associated with polar granules, and fourteen nuclei lie in the rest of the cytoplasm. All the nuclei possess about forty chromosomes. During the fifth division the posterior nuclei do not divide and the polar-plasm becomes constricted to form primordial germ cells (pole cells). The remaining fourteen nuclei divide and lose about thirtγ-two chromosomes so that twenty-eight nuclei are formed containing only eight chromosomes. These are the presumptive somatic nuclei. During subsequent divisions the pole cell nuclei retain the full chromosome number; these divisions occur less frequently than those of the somatic nuclei.

Experiments were performed on early embryonic stages to elucidate the properties of the posterior end during the time that chromosome elimination was taking place from the presumptive somatic nuclei. Ultraviolet irradiation, constriction, and centrifugation techniques were used.

The polar granules are concerned with the non-division of the germ-cell nuclei during the fifth division, since if the granules are dispersed by centrifugation, or if nuclei are prevented by constriction from coming into contact with them before the fifth division, all the nuclei divide with chromosome elimination at this division. With each technique it is possible to obtain embryos possessing germ cells with only eight chromosomes in their nuclei.

Individuals possessing germ-line nuclei with only eight chromosomes were allowed to develop to maturity. Abnormalities were confined to the germ cells only and were the same regardless of which technique had been used to produce the deficient germ line. An ovary containing germ-cell nuclei with only eight chromosomes is unable to form both oocytes and nurse cells. A testis containing germ-cell nuclei with only eight chromosomes is unable to form spermatocytes but cells which come to resemble gametes are formed. Experimental males and females are both sterile.

The results are discussed in relation to other experimental work on Cecidomyidae and the following main conclusions are reached: (a) the polar granules are responsible for preventing an irreversible loss of chromosomes from the germ-cell nuclei by preventing the mitosis of these nuclei during the fifth division; (b) the chromosomes normally retained in the germ line are required for gametogenesis, particularly for oogenesis.

The significance of chromosome elimination is discussed.

In some animals the germ cells originate during embryonic cleavage and it is well established that experimental procedures which interfere with the normal association between the germ-line nuclei and the cytoplasm these occupy can result in sterility (e.g. Geigy, 1931).

In the Cecidomyidae (Diptera) germ cells form as a result of one or more of the posterior cleavage nuclei migrating into the cytoplasm at the posterior end of the embryo (polar-plasm) and the polar-plasm becoming constricted off to form germ cells (pole cells). There is no evidence that in this family of insects the pole cells give rise to any tissue other than the germ cells.

Synchronous with, or immediately after, the formation of pole cells in Cecidomyidae are atypical mitoses of all of the presumptive somatic nuclei. These enter mitosis and a large but specific number of the chromosomes remains at the equator at anaphase and fails to enter the daughter nuclei. This is known as chromosome elimination and was first observed in Miastor by Kahle (1908). The chromosomes which are eliminated are designated E chromosomes, and the remainder S chromosomes. Somatic nuclei have only the S number and the primordial germ cells have the full E+S number (White, 1950).

In many male cecidomyid embryos there is a further elimination of chromosomes from the presumptive somatic nuclei. White (1950) considers that this difference between the elimination pattern of the two sexes is important in sex determination in Cecidomyidae.

The formation of pole cells in Cecidomyidae is marked not only by the elimination of chromosomes from the presumptive somatic nuclei but also by the onset of a lower frequency of mitosis of the germ-line nuclei. Chromosome elimination is a particularly clear example of nuclear differentiation at the genetic level and has been subjected to experimental analysis by Geyer-Duszynska (1959, 1961, 1966), by Nicklas (1959) and by the present author (1961). The aim of the present work is an attempt to throw some light on the causes of the retention of the E chromosomes by the germ-line cells, on the factors causing the germ cells to divide less frequently than the somatic nuclei, and on the functions of the E chromosomes. The Hessian Fly Mayetiola destructor (Say) was chosen since it was found to be possible to breed this species continuously through the year.

The eggs of Mayetiola are laid in longitudinal rows on the leaves of wheat plants in the 2-leaf stage. The eggs are deposited always with the posterior end of each nearest the axil of the leaf. Generally the eggs from any one female will develop into adults of only one sex. The eggs hatch in 3–5 days and the larvae migrate to a position under the ligule. This position is maintained throughout larval and pupal life and the last larval and pupal instars are spent in a puparium. The life-cycle is completed in 47–50 days during the summer months in England. Under natural conditions there are three generations a year, the last generation overwintering in the puparia. Emergence takes place early in the morning; egg-laying starts within an hour or two of mating and may continue for as long as 48 h.

Breeding techniques

Development can be temporarily arrested by storing puparia at 5 °C in Petri dishes containing moist filter paper, but it was not possible to prevent emergence once pupation had occurred. When adults were required puparia were transferred to glass cylinders 10·2 cm in diameter closed at the top end by muslin-covered metal rings, the cylinders resting on moistened peat fibre and kept at 20 °C. Adults emerged in 3–14 days. Those adults required for maintaining the stocks were transferred to potted Peko wheat plants in the 2-leaf stage. The plants were enclosed by glass cylinders closed with muslin at the top. There is a mortality rate of approximately 40% in young larvae before they become established under the ligule; they are particularly susceptible to desiccation. Humidity was maintained in the cylinders by covering the tops with wet cloths when oviposition was completed. The cylinders were removed and the plants covered with muslin bags when all the eggs had disappeared from the leaves.

When eggs were required for experimental purposes short pieces of leaf were placed in glass tubes containing mated females. Embryos and young larvae were handled with silver-plated entomological pins set in matchsticks. Post-experimental embryos required for analysis in later stages in the life-cycle were transferred to a drop of water in the axil of a leaf, from which the larvae were able to take up their normal position. All experimental embryos were kept at 20 ± 1 °C.

Cytological techniques

Embryos were examined chiefly by means of sections. Embryos were fixed in a mixture of formalin, glacial acetic acid, absolute alcohol, and distilled water (6:1:16:30) for 10–24 h. The chorion, though permeable to the fixative, is impermeable to cedar-wood oil and paraffin wax, and was pricked with a tungsten needle set in a micromanipulator while the embryos were still in the fixative. The embryos were taken through 50 % cellosolve to a saturated solution of eosin in cellosolve, where they were left for 24 h. This stained the embryos bright red so that they could be orientated more easily during embedding. The embryos were cleared in cedar-wood oil and embedded in a mixture of 56 °C and 58 °C paraffin wax (1:1) containing 3 % beeswax. Sections were cut at 10 μ. In order to facilitate finding the cut sections on the slide, the wax ribbons were outlined with an indian-ink mixture developed by Pusey (1939) before dewaxing. Sections were stained with Heidenhain’s iron haematoxylin, leucobasic fuchsin counterstained with eosin being substituted in certain cases.

For the histochemical studies the same procedure was used except that Helly’s or Zenker’s fixatives were used and that in some cases sections were cut at 5 μ. The chromosomes of the embryos were examined by means of smears; each embryo was pricked on a slide so that the cytoplasm ran out in a fine film. The smears were fixed in Zenker’s fixative and stained with iron haematoxylin. The chromosomes of the later stages were examined by means of acetoorcein squashes, and sometimes by means of sections of material fixed in San Felice’s fluid and stained with crystal violet.

The chromosomes of Mayetiola

The chromosomes and early cleavage divisions of Mayetiola destructor were investigated by Metcalfe (1935). Although this account is inaccurate in several respects it correctly establishes that chromosome elimination occurs at the fifth division in this species and also that there is no subsequent elimination of chromosomes from the presumptive somatic nuclei in male embryos.

The pre-elimination chromosome number

Over 120 embryos were examined by means of smears but it has not been possible to count the full chromosome number accurately. This is due to the fact that in division the chromosomes are long and thin, and though the chromosomes of different nuclei never become mixed, the chromosomes of any one nucleus frequently overlap. The chromosome number was always in the range 35–45, and in one particular nucleus between 39 and 42 chromosomes could be counted.

The primordial germ-cell chromosome number

Divisions of the pole cells were examined in post-elimination embryos squashed on slides. The chromosome number was in the same range as in preelimination nuclei: 35–45. The larval gonads were also examined but the germ cells were too small to be squashed at all satisfactorily. Metcalfe (1935) estimated the germ-cell chromosome number entirely by means of 5μ sections and gives a number of 16 for both sexes. The present work covers a wider range of stages and indicates that the full chromosome number is a far larger one than Metcalfe’s estimation. Many sections of gonads appear to contain germ cells with a small number of chromosomes, but a careful study of the slides shows that this is due to the nuclei being cut through and therefore incomplete.

The somatic chromosome number

The female somatic chromosome number is eight; this is characteristic of many cecidomyids. A large part of the larval ovary consists of cells of somatic origin, and when these were observed in division eight chromosomes could be counted. There is strong evidence that the male somatic chromosome number is also eight, unlike the majority of male cecidomyids, which have six chromosomes in the somatic nuclei. When stages in divisions in embryos between the fifth division and 4 h of development were examined, eight chromosomes could frequently be counted. Since Mayetiola frequently but not always produces unisexual batches of eggs, over 200 embryonic smears were examined from over thirty different batches of eggs. At least half of each batch was allowed to develop to maturity so that some indication of the sex of the embryos could be given. Even when batches of embryos, some of which had been used for chromosome study, yielded only male midges, eight chromosomes could be counted in the somatic prophases.

Cleavage and pole cell formation in Mayetiola

The eggs are cylindrical, approximately 400μ long and 80 μ in diameter, and are bounded by a transparent chorion. The pattern of cleavage divisions can be followed clearly in living embryos since the nuclei are surrounded by a dense red material (Fig. 1A–D). This material is at first distributed throughout the cytoplasm but gradually accumulates round the synkarion. During the early anaphases each red mass can be seen to divide, and between the divisions the centre of each mass appears to be hollow; this is due to the presence of a large interphase nucleus. That the characteristic behaviour of the red masses corresponded to nuclear divisions was shown by sectioning embryos at definitive stages.

FIGURE 1.

A–D. Four living unstained embryos. The cytoplasm is pale orange in colour and a red material surrounds the nuclei.

(A) After the first division. The two nuclei appear as two reddish bodies about a third of the distance from the anterior end (Ant.). At the posterior end the chorion extends slightly beyond the cytoplasm.

(B) After the third division. There are eight nuclei in a single row down the embryo.

(C) After the fourth division. There are sixteen nuclei. Two of these are at the posterior end in the polar-plasm.

(D) After the fifth division. One of the two pole cells at the posterior end appears as a single vesicle. In the main body of the embryo there are twenty-eight nuclei, only twenty-four of which are in focus. These nuclei have undergone chromosome elimination during the fifth division, but the eliminated chromosomes are not visible in living embryos.

E–l. Longitudinal sections through the posterior end of embryos showing the polar granules and the formation of the primordial germ cell. The sections are stained with iron haematoxylin.

(E) An uncleaved fertilized egg. The polar granules (p.#.) can be seen at the extreme posterior end.

(F) Before the fourth division. The most posterior nucleus has entered the polar-plasm and the polar granules have begun to form a crescent at the posterior end. A circular constriction has begun to appear round the polar-plasm.

(G) The fourth division, showing prophase in the most posterior nucleus.

(H) After the fourth division. The circular constriction has been carried further so that the polar-plasm, the polar granules, and the two posterior nuclei, will be separated from the rest of the embryo.

(I) During the fifth (eliminating) division. A binucleate primordial germ cell is almost completely formed, and the polar granules have spread out over the surface of the two nuclei which are not dividing. Part of an elimination metaphase can be seen lying in the somatic part of the embryo.

FIGURE 1.

A–D. Four living unstained embryos. The cytoplasm is pale orange in colour and a red material surrounds the nuclei.

(A) After the first division. The two nuclei appear as two reddish bodies about a third of the distance from the anterior end (Ant.). At the posterior end the chorion extends slightly beyond the cytoplasm.

(B) After the third division. There are eight nuclei in a single row down the embryo.

(C) After the fourth division. There are sixteen nuclei. Two of these are at the posterior end in the polar-plasm.

(D) After the fifth division. One of the two pole cells at the posterior end appears as a single vesicle. In the main body of the embryo there are twenty-eight nuclei, only twenty-four of which are in focus. These nuclei have undergone chromosome elimination during the fifth division, but the eliminated chromosomes are not visible in living embryos.

E–l. Longitudinal sections through the posterior end of embryos showing the polar granules and the formation of the primordial germ cell. The sections are stained with iron haematoxylin.

(E) An uncleaved fertilized egg. The polar granules (p.#.) can be seen at the extreme posterior end.

(F) Before the fourth division. The most posterior nucleus has entered the polar-plasm and the polar granules have begun to form a crescent at the posterior end. A circular constriction has begun to appear round the polar-plasm.

(G) The fourth division, showing prophase in the most posterior nucleus.

(H) After the fourth division. The circular constriction has been carried further so that the polar-plasm, the polar granules, and the two posterior nuclei, will be separated from the rest of the embryo.

(I) During the fifth (eliminating) division. A binucleate primordial germ cell is almost completely formed, and the polar granules have spread out over the surface of the two nuclei which are not dividing. Part of an elimination metaphase can be seen lying in the somatic part of the embryo.

The egg nucleus first appears as a red body lying a third of the distance from the anterior end. The first division occurs about 2 h after oviposition and the subsequent divisions occur at intervals of about 20 min between each anaphase. The first two divisions take place synchronously in all the nuclei but the subsequent divisions are asynchronous and take place in a gradient of division, so that the anterior nuclei start dividing slightly before those situated in more posterior regions. The primordial germ cells are formed during the fifth division, and in living embryos they can be seen at the posterior end and are characterized by a reduction in the amount of red material found round the nuclei (Fig. 1D).

The cytoplasm is at first irregularly vacuolated, and is slightly more dense round the egg nucleus. This concentration of the cytoplasm round the nuclei increases during the early divisions and ultimately results in the complete separation of the cytoplasm from the yolk when the somatic nuclei move to the periphery to form the blastoderm at 4 h of development. The polar-plasm is indistinguishable from and continuous with the rest of the cytoplasm and is characterized by the possession of polar granules (Fig. 1E). These granules stain very heavily with iron haematoxylin and later become spatially associated with the most posterior cleavage nucleus, which enters the polar-plasm. It was found that these granules stained red with the Baker modification of the Altmann-Metzner method for mitochondria. In so far as this procedure is a test for mitochondria the result indicates that at least some or part of the polar granules are mitochondrial in nature, or closely associated with mitochondria. The granules also stained red with the pyronin and methyl green method for RNA. Bradbury’s saliva technique (Carleton & Drury, 1957) for the preparation of RN-ase was used as a control, and it was found that the polar granules then failed to stain. It may thus be concluded that the polar granules contain RNA and probably also mitochondria. Mahowald (1962) has shown that the polar granules in Drosophila contain RNA, and it is worth pointing out that the germinal plasm in frogs and toads appears to contain mitochondria and RNA (Blackler, 1958; Czolowska, 1969).

The equatorial plates of the first three divisions are orientated perpendicular to the long axis of the embryo, so that by the end of the third division eight ovoid nuclei lie equally spaced in a straight row along the embryo. The appearance of a gradient during the third or fourth division is evident; the more anterior nuclei enter prophase slightly before those near the posterior end of the embryo. Immediately before the fourth division the polar granules lying in the polar-plasm assume the form of a crescent at the posterior end, with the most posterior nucleus lying in the polar-plasm near the granules (Fig. 1F). At this time a slight constriction appears round the embryo at the posterior end, so that a polar bud is formed containing the polar-plasm, the polar granules, and a nucleus. During the fourth division all eight nuclei divide with the equatorial plates parallel to the long axis. Immediately after this division the polar granules spread out and become closely associated with the nuclear membranes of the two nuclei lying in the polar-plasm. During the fifth division these two nuclei do not divide; the fourteen nuclei lying in the somatic part of the embryo divide with chromosome elimination. During this division the posterior transverse constriction is carried farther, so that a binucleate primordial germ cell is formed by the time that chromosome elimination is completed in the rest of the embryo. A cell membrane appears between the two presumptive germ-line nuclei and thus two pole cells are formed. The main stages in the formation of the germ line can be seen in Fig. 1F–-I. From the time of the fifth division the pole cells divide less frequently than the somatic nuclei and stop dividing altogether by of development.

The gradient which appears at the third or fourth division persists during the fifth division so that chromosome elimination takes place as a wave starting at the anterior end. At the fifth metaphase all the chromosomes lie closely packed on the equatorial plates, which are not noticeably different from those of earlier divisions. Owing to the extreme rapidity of the process and to the great number and small size of the chromosomes during anaphase, it is difficult to ascertain the exact sequence of events during elimination. A small space appears across the equator, suggesting that all the chromatids enter anaphase. However, only those chromosomes (S) destined to occupy the presumptive somatic nuclei reach the poles; most of the chromosomes (E) return to the equator of each spindle and coalesce to form a lump of strongly Feulgen-positive material.

The subsequent divisions of the small presumptive somatic nuclei are normal ; no elimination of chromosomes has been observed from any nucleus after the fifth division. The ultimate fate of the eliminated chromosomes is uncertain. During embryonic life they break up into smaller pieces, but there is no evidence either that this takes place in any regular way or that it is accompanied by any DNA synthesis. The lumps remain strongly Feulgen-positive and finally disappear with the yolk when the young larva starts to feed.

Experiments on chromosome elimination

It is not known whether the nucleus which enters the polar-plasm before the fourth division is predetermined to become the germ-cell nucleus. If, however, it is assumed that mitosis results in initially identical nuclei it follows that it is extremely likely that all of the sixteen nuclei in the Mayetiola embryo before the fifth division are identical. If this is the case the sudden dichotomy in nuclear behaviour between those nuclei lying in the somatic part of the embryo and those lying in the polar-plasm must result from some influence outside the nuclei themselves. The temporary cessation of mitosis in the two nuclei lying in the polar plasm during the eliminating division in the rest of the embryo, and the characteristic behaviour of the polar granules at this time, suggest that the latter may be concerned with this inhibition of mitosis. The following two types of experiment were performed to analyse the function of the polar granules and their possible relationship to the non-division of and the non-elimination of chromosomes from the germ-cell nuclei: (i) the first was an attempt to affect the normal potentiality of the posterior end by ultraviolet irradiation before any nucleus had migrated into it; (ii) the second was an attempt to alter the normal spatial relationships of the nuclei, the polar granules, and the polar-plasm, so that a consistently abnormal situation prevailed at the time of the fifth division. Constriction and centrifugation procedures were used for this.

Experiments with ultraviolet irradiation

Pieces of embedding wax were used as a base on which to support the embryos during the experiment. Parallel grooves, 2 mm apart, were made on one surface and the embryos transferred to these grooves and arranged with the same orientation in as straight a row as possible, one in each groove. The embryos were then covered with a glass coverslip wrapped in silver paper; the position of the coverslip was adjusted so that only the posterior end of each embryo projected from under the silver paper. The whole piece of wax was then placed in an ultraviolet beam from a Hanovia 100 W medium-pressure mercury arc lamp. The lamp was used without a filter, but with a quartz condenser, and arranged so as to give a downwardly directed beam. [The bulk of the output of the lamp was at 2537 Å, and it is likely that it was this wavelength that was operative in altering the potentialities of the posterior end (J. B. Gurdon, personal communication).] The embryos were placed 12 mm from the source, and were all in the 2- or 4-nuclei stage. Any embryo which had reached the 8-nuclei stage by the end of the irradiation was discarded as it was likely that the most posterior nucleus in these cases had received at least some irradiation. Three experiments were carried out; the first two were controls.

  • Seventeen embryos were transferred to wax blocks and completely covered with silver-paper-wrapped coverslips. The blocks were irradiated for 12 min. Subsequent development of all the embryos was completely normal.

  • Twenty-three embryos were irradiated with no protection whatever for doses of between 3 and 12 min. In nearly all cases the treatment proved lethal before the fourth division.

  • Three hundred and twenty embryos were irradiated in batches of about six at a time for 7 min doses, with only the posterior end exposed to the ultraviolet beam. The following list gives the stages at which these embryos were analysed :

The first five categories are dealt with here.

  1. Twelve embryos died shortly after irradiation. This was probably due to damage received during transference to and from the wax blocks. It was noticed that the chorion is particularly fragile in very young embryos (2- or 4-nuclei stages).

  2. Eleven embryos were fixed within 5 min of the completion of irradiation. The polar granules in these embryos had lost their clearly granular structure; they appeared to be undergoing disintegration. This was accompanied by a decrease in their staining properties. The polar-plasm itself appeared to be unaffected.

  3. In the sixteen embryos fixed during the fifth division the polar granules were either barely visible or had disappeared completely. In no case could there be observed any association of the granules with the nuclei in the polar-plasm. The irradiation appeared to slow down the formation of primordial germ cells, since although one or two nuclei were always present in the polar-plasm by the time of the fifth division, there was generally little sign of cytoplasmic constriction. Since the fifth division takes place as a wave of mitosis starting at the anterior end, it has been difficult to determine exactly what happens in the polar-plasm when the wave of mitosis reaches the posterior end. There is no sign that those nuclei lying in the polar-plasm of an irradiated embryo divide at the fifth division. Some of the sixteen embryos had been fixed just after this division; a variable number of small vesicles containing diffusely scattered chromatin can be seen in the polar-plasm of these embryos. It is probable that these vesicles are derived non-mitotically from the nucleus which entered the polar-plasm after the third division.

  4. In the fifteen embryos fixed 4 h after irradiation there was no sign of large pole cells (Fig. 2A). The posterior end was occupied by small cells indistinguishable from somatic cells in the rest of the blastoderm (Fig. 2B). It is not possible to say whether the nuclei of these cells are derived from the original primordial germ cell nucleus or from nuclei of somatic origin which have migrated into the polar-plasm after the fifth division.

  5. Five irradiated embryos fixed at 17 h of development showed gonad primordia consisting entirely of small cells with the reduced number of chromosomes in their nuclei.

FIGURE 2.

Longitudinal sections through normal and experimental embryos. The sections are stained with iron haematoxylin.

(A) A normal 6 h embryo. Small somatic nuclei are becoming surrounded by ceil membranes so that the blastoderm will be formed. At the posterior end are large pole cells containing the polar granules.

(B) A 6 h embryo in which the polar-plasm was irradiated with ultraviolet light at the 2-nuclei stage. The polar granules have disappeared and the polar-plasm is occupied by small nuclei.

(C) A 3 h embryo constricted during the third, fourth and fifth divisions so that no nucleus could enter the polar-plasm. All the nuclei divided with chromosome elimination at the fifth division. Small nuclei then entered the polar-plasm and became associated with the polar granules. These nuclei then divided at the slow rate of mitosis characteristic of normal pole cells. The small nuclei of three of these pole cells can be seen as pale vesicles completely surrounded by the polar granules. The arrow indicates the position of the original constriction.

(D) A 312 h embryo centrifuged at the 4-nuclei stage with the polar-plasm placed centripetally. All the nuclei divided with chromosome elimination at the fifth division. During the course of their redistribution after centrifugation small nuclei enter the polar-plasm and small pole cells (r.p.c.) are formed. These do not contain the polar granules. The chromatin of the small pole cells is diffusely scattered within the nuclear membrane; this is different from the somatic nuclei where the chromatin is concentrated in the centre of the nucleus.

(E) A 4 h embryo centrifuged at the 4-nuclei stage with the polar-plasm placed centripetally. All the nuclei divided with chromosome elimination at the fifth division, and small pole cells (r.p.c.) are forming in a lateral position.

FIGURE 2.

Longitudinal sections through normal and experimental embryos. The sections are stained with iron haematoxylin.

(A) A normal 6 h embryo. Small somatic nuclei are becoming surrounded by ceil membranes so that the blastoderm will be formed. At the posterior end are large pole cells containing the polar granules.

(B) A 6 h embryo in which the polar-plasm was irradiated with ultraviolet light at the 2-nuclei stage. The polar granules have disappeared and the polar-plasm is occupied by small nuclei.

(C) A 3 h embryo constricted during the third, fourth and fifth divisions so that no nucleus could enter the polar-plasm. All the nuclei divided with chromosome elimination at the fifth division. Small nuclei then entered the polar-plasm and became associated with the polar granules. These nuclei then divided at the slow rate of mitosis characteristic of normal pole cells. The small nuclei of three of these pole cells can be seen as pale vesicles completely surrounded by the polar granules. The arrow indicates the position of the original constriction.

(D) A 312 h embryo centrifuged at the 4-nuclei stage with the polar-plasm placed centripetally. All the nuclei divided with chromosome elimination at the fifth division. During the course of their redistribution after centrifugation small nuclei enter the polar-plasm and small pole cells (r.p.c.) are formed. These do not contain the polar granules. The chromatin of the small pole cells is diffusely scattered within the nuclear membrane; this is different from the somatic nuclei where the chromatin is concentrated in the centre of the nucleus.

(E) A 4 h embryo centrifuged at the 4-nuclei stage with the polar-plasm placed centripetally. All the nuclei divided with chromosome elimination at the fifth division, and small pole cells (r.p.c.) are forming in a lateral position.

Ultraviolet irradiation of the posterior end during the 2- or 4-nuclei stage appears to have the following effects on embryonic development :

  • The polar granules show a progressive disintegration and reduction in staining power. They do not associate themselves with primordial germ-cell nuclei.

  • Constriction of the polar-plasm to form cells is delayed so that it takes place after the fifth division.

  • The two primordial germ-cell nuclei appear to break up into smaller vesicles during the fifth division.

  • Somatic nuclei appear to replace the abnormal germ-cell nuclei so that a germ line is formed which consists entirely of cells with the reduced chromosome number in their nuclei.

Experiments involving constriction of embryos

These experiments were designed to prevent any nuclei from entering the polar-plasm before the fifth division, so that all the nuclei were lying in the somatic part of the embryo at the time of the fifth division. Embryos with two or four nuclei were used, and each embryo was treated separately. Fine human hair was used for the constriction; a hair was weighted at one end with a small piece of plasticine and the other end was attached by adhesive tape half-way along the long edge of a glass slide. The slide was arranged under a binocular microscope so that the hair lay across the slide. One embryo was placed on the slide a few millimetres from the edge and by means of a pin the hair was placed across the embryo in the required place. The weight of the plasticine tightened the hair and caused a constriction to appear across the embryo. After the fifth division the hair was removed and the embryo returned to its original shape. One hundred and sixty embryos were constricted; 109 of these survived and continued to develop after the hair had been removed. In 43 of the 109 embryos which survived the constriction the most posterior nucleus passed under the constriction into the polar-plasm before the fifth division so that these embryos had the normal distribution of nuclei at the time of the fifth division. It was later ascertained by sections that these embryos possessed pole-cell nuclei with the full chromosome number. In sixtγ-six embryos the constriction was tight enough to prevent the migration of the posterior nucleus into the polar-plasm at the normal time; in these embryos all sixteen nuclei lay in the somatic part of the embryo before the fifth division and divided at that division. This was followed by the migration of two posterior nuclei under the constriction into the polar-plasm. These sixtγ-six embryos were analysed at the following developmental stages :

The first two categories are dealt with here.

  1. In the twelve embryos fixed at the blastoderm stage a germ line was present containing nuclei with the reduced number of chromosomes. Owing to the reduced size of the nuclei occupying the polar-plasm, the polar granules tend to surround them completely instead of forming an irregular crescent (Fig. 2C). The low number of these small pole cells indicates that they were dividing at the rate of normal pole cells; that is, much less frequently than the nuclei in the somatic part of the embryo.

  2. These eight embryos showed the same features as embryos in which the posterior end had been irradiated; gonad primordia were present containing cells with small nuclei.

By preventing the posterior nucleus from entering the polar-plasm at the normal time all the nuclei undergo elimination at the fifth division. Nuclei with only the somatic chromosome number then enter the polar-plasm and become associated with the polar granules. These nuclei subsequently divide at the same frequency as normal pole-cell nuclei, and germ cells are established which are normal in every way except that their nuclei contain the somatic chromosome number only.

Experiments with centrifugation

Since Mayetiola lays its eggs all with the same orientation in longitudinal rows on wheat leaves, and since large numbers of eggs are laid at any one time, it was found that it was possible to centrifuge large numbers of embryos of the same age without removing them from the leaves. Each leaf was attached with adhesive tape to a strip of perspex which just fitted into a small centrifuge tube. Embryos were centrifuged at 2-, 4-, 8- or 16-nuclei stages, and were centrifuged in batches of between eight and twenty-five at a time. Eight series of experiments were carried out, summarized in Fig. 3, involving a total of 470 embryos. It was found that the most consistent results were obtained at relatively faster centrifuge speeds maintained for short periods of time (3–5 min), rather than for longer periods of centrifugation at slower speeds.

FIGURE 3.

Series 1–IV are embryos with 2–16 nuclei centrifuged with the posterior end orientated centripetally. Series V–V1II are embryos with 2–16 nuclei centrifuged with the posterior end centrifugal. Centrifugation produced dispersal of the contents of the embryo in the form of a density gradient. In all the embryos the cytoplasm became stratified into three zones. In series I, II, V and VI, the polar granules were dispersed and could not be located in sections after centrifugation. With series HI, IV, VII and VIII (in which the polar granules were associated with a nucleus or nuclei at the moment of centrifugation) it was not possible to separate the posterior nucleus or nuclei from the polar-plasm and polar granules after the association between these had been established (after the third division). The subsequent development of embryos in series III, IV, VII and VIII was perfectly normal. A germ line was formed with the full chromosome number in the nuclei. The subsequent development of embryos in series I, II, V and VI was abnormal. All the nuclei divided at the fifth division and underwent chromosome elimination. Small pole cells were formed when the somatic nuclei redistributed themselves within the embryo; some of them entered the polar-plasm, which may have been displaced from the posterior end. All these embryos died before gastrulation.

FIGURE 3.

Series 1–IV are embryos with 2–16 nuclei centrifuged with the posterior end orientated centripetally. Series V–V1II are embryos with 2–16 nuclei centrifuged with the posterior end centrifugal. Centrifugation produced dispersal of the contents of the embryo in the form of a density gradient. In all the embryos the cytoplasm became stratified into three zones. In series I, II, V and VI, the polar granules were dispersed and could not be located in sections after centrifugation. With series HI, IV, VII and VIII (in which the polar granules were associated with a nucleus or nuclei at the moment of centrifugation) it was not possible to separate the posterior nucleus or nuclei from the polar-plasm and polar granules after the association between these had been established (after the third division). The subsequent development of embryos in series III, IV, VII and VIII was perfectly normal. A germ line was formed with the full chromosome number in the nuclei. The subsequent development of embryos in series I, II, V and VI was abnormal. All the nuclei divided at the fifth division and underwent chromosome elimination. Small pole cells were formed when the somatic nuclei redistributed themselves within the embryo; some of them entered the polar-plasm, which may have been displaced from the posterior end. All these embryos died before gastrulation.

At the faster centrifuge speeds the embryonic cytoplasm became stratified into three zones. This took place irrespective of both the age of the embryo and the direction of the centrifugal force in relation to the longitudinal axis. At the centrifugal end of the embryo an orange zone appeared, occupying one-third of the embryo; sections showed that this zone consisted of a number of large globules, probably of yolk. The middle zone was a clear yellow colour consisting of what appeared to be granular cytoplasm and the centripetal zone was opaque and grey in colour and consisted of a mass of uniform cytoplasm. In all embryos with up to four nuclei the polar granules were dispersed so that they did not stain with iron haematoxylin. This dispersal of the polar granules took place at short periods of centrifugation and also at slower speeds; the granules appeared to be among the first things affected by centrifugation. The polar-plasm is distinguishable from the rest of the cytoplasm only by the possession of polar granules, and since these granules are dispersed at the onset of centrifugation it is not possible to determine the degree of displacement of the polar-plasm in embryos before the 8-nuclei stage. In embryos with eight or more nuclei the polar granules were not removed from the polar-plasm, nor were they separated from the nucleus or nuclei lying in the polar-plasm; in these embryos the primordial germ cell was moved as a single unit.

The degree of displacement of the nuclei was found to be dependent on the direction and duration of centrifugation, and on the age of the embryo. Up to the 4-nuclei stage all the nuclei were displaced into the centrifugal orange zone; this was irrespective of the direction of centrifugation. In older embryos the displacement was similar except that the primordial germ cells (or cell) were moved only into the central yellow zone if they were initially centripetal, but they remained at the posterior end if they were initially centrifugal.

Centrifugation did not alter the course of the nuclear divisions; the nuclei sometimes divided during centrifugation. Immediately after centrifugation the nuclei began to redistribute themselves and the cytoplasmic stratifications gradually disappeared. It was noticed that in all cases the gradient of division was maintained or reconstituted after centrifugation; the nuclei always divided in a wave which started at the anterior end.

SeriesI and II. (Fiftγ-seven embryos with two or four nuclei, centrifuged with the polar-plasm centripetal.)

In these embryos the nuclei did not reach the posterior end until between 3 and 4 h after centrifugation ; the most posterior nucleus lay about a quarter of the way from the posterior end at the time of the fifth division. This division took place as in normal embryos, as a wave which started at the anterior end. All the nuclei divided at the fifth division, and sections showed that chromosome elimination had occurred throughout. Sections of later stages showed that cells which resembled small pole cells were formed when the nuclei, towards the end of their posterior migration, reached the posterior end. These can be seen in Fig. 2D. In some embryos of these series what appeared to be small pole cells were formed in a lateral position at a short distance from the posterior end (Fig. 2E). In these embryos it seems likely that the polar-plasm itself has been moved during the centrifugation and that when small nuclei enter it, the polar-plasm becomes constricted to form pole cells before a posterior position has been reached.

All the embryos in Series I and II died before gastrulation ; the nuclei stopped dividing at about of development. It was thus not possible to follow the fate of these small pole cells beyond the point of their formation.

Series III and IV. (188 embryos with eight or sixteen nuclei, centrifuged with the polar-plasm centripetal.)

In these embryos the position of the primordial germ cell or cells could be located immediately after centrifugation, since once a nucleus had come into contact with the polar granules and the polar-plasm it was not possible to destroy the association by centrifugation. Only the presumptive somatic nuclei divided at the fifth division; the germ-line nuclei retained the E chromosomes and divided less frequently than the somatic nuclei. They reached the posterior end in about an hour after centrifugation. In spite of the fact that the embryos in these series were centrifuged for the same time and speed as those in series I and II where the effect proved lethal, all these survived and were capable of giving rise to normal adults. In other words, centrifugation of embryos in series III and IV produced no irreversible effect on development.

Series Vand VI. (Eightγ-one embryos with two or four nuclei, centrifuged with the polar-plasm centrifugal.)

The subsequent development of these embryos was similar to that of series I and II. Although the polar-plasm was centrifugal the polar granules became dispersed and could not be seen in any of the sections. All the nuclei divided at the fifth division and chromosome elimination occurred throughout. Small pole cells were formed when these nuclei reached the posterior end. As with series I and II, all these embryos died before gastrulation.

Series VII and VIII. (144 embryos with eight or sixteen nuclei centrifuged with the polar-plasm centrifugal.)

It was not possible to force any more nuclei into the polar-plasm by centrifugation. The primordial germ cell was relatively undisturbed by the centrifugation; it was sometimes shifted to a partly lateral position. The subsequent development of these embryos was normal, as with series III and IV.

The following conclusions can be reached concerning centrifugation of Mayetiola embryos:

  1. Centrifugation causes stratification of the cytoplasm into three layers. This takes place independently of the age of the embryo and of the direction of the centrifugal force.

  2. In embryos centrifuged before a nucleus enters the polar-plasm, the polar granules disappear, probably by becoming dispersed into the somatic part of the embryo.

  3. The gradient of division of the nuclei is not changed by centrifugation.

  4. If no nucleus is in contact with the polar-plasm and polar granules at the time of the fifth division, all the nuclei divide with chromosome elimination at this division. In such embryos small pole cells are formed when nuclei with the somatic chromosome number enter the polar-plasm. All these embryos die before gastrulation.

  5. Centrifugation of embryos in which the primordial germ cell is already formed does not have any permanent effect on development.

Analysis of post-embryonic stages possessing germ-line nuclei with the reduced number of chromosomes

Three hundred and seven experimental embryos were allowed to develop to maturity. Fortγ-six of these were derived from constriction experiments and 261 from ultraviolet irradiation experiments. As has already been mentioned, there is an approximately 40% larval mortality at the immediately post-embryonic stage; of the fortγ-six embryos subjected to constriction procedures, only twenty-six survived to maturity (56% survival), and of the 261 embryos subjected to irradiation only 150 survived to maturity (57 % survival). A control with exactly 300 normal embryos gave 193 adults (64% survival), suggesting that the high mortality rate of experimental embryos is due to factors other than those to which they have been subjected experimentally.

Both irradiation and the constriction experiments thus produced some embryos which became adults. When these embryos developed to maturity it was found that they showed exactly the same degree of abnormality, but it is important to appreciate the difference between the two types of germ line. The germ line of a constricted embryo differs from the normal germ line only in the absence of the E chromosomes. The germ line is otherwise unchanged; it has both polar granules and an undamaged polar-plasm. On the other hand, in an irradiated embryo the polar granules are destroyed and the polar-plasm itself is sufficiently damaged to cause the disintegration of the nuclei which occupy it during the fifth division. The polar-plasm is later occupied by somatic nuclei, and the polar-plasm is subsequently able to provide an adequate environment for the normal division of the nuclei lying in it. The irradiation does not affect the potentiality of the polar-plasm for forming a germ line, and neither does the destruction of the polar granules affect this. The abnormalities found in older stages derived from irradiated embryos are the same as those in individuals developed from constricted embryos and evidently result from the absence of the E chromosomes from the germ line.

For the greater part of larval life the gonads are identical in male and female, and consist of small groups of cells with nuclei with the reduced chromosome number. The gonads lie in the ninth trunk segment and are attached to the Malpighian tubules by fine tracheae.

1. Females

Of the 176 experimental embryos which survived to post-embryonic stages, 159 developed into females (presumably by chance). Fig. 4E shows the appearance of the ovary in a larva just before pupation ; the larva is derived from an embryo subjected to irradiation. The ovary consists entirely of cells with the somatic number of chromosomes; there are none of the oogonia which are found in the normal ovary (Fig. 4A). The greater part of the normal ovary consists of follicle cells of somatic origin. These later form follicles which surround groups of oogonia derived from the primordial germ cells, and also form the inner end of the oviduct (Fig. 4B). All but one of the oogonia in each follicle form large nurse cells with deeply staining cytoplasm (Fig. 4C). One oogonial cell in each follicle becomes an oocyte; each of these subsequently enlarges so that the whole ovary fills the abdomen.

FIGURE 4.

Transverse sections through the ovaries of normal and experimental individuals. All the sections are stained with iron haematoxylin except A, which is stained with crystal violet. A–D, Normal individuals. E–H, Experimental individuals, with ovaries lacking the E chromosomes.

(A) The ovary in the last larval instar. A large mass of follicle cells (f.c.) can be seen on the left. The larger primordial germ cells, or oogonia (oog.), form a smaller group to the right.

(B) Section through the ovaries in the middle of the pupal instar. The follicles contain groups of oogonia and are attached to the inner end of the oviduct (ov.).

(C) Transverse section through the abdomen of a late pupa, showing the two ovaries much enlarged due to the formation of ooplasm (oop.) in the follicles (f). Each follicle contains a group of nurse cells (n.c.) with deeply staining cytoplasm.

(D) The same individual as C, showing the follicles composed of follicle cells (f.c.) containing the nurse cells (n.c.) and ooplasm. The ooplasm of the oocyte in each follicle is lightly stained and the nurse cells have large spherical nuclei with deeply staining irregular chromosomes.

(E) The ovary in a late larva derived from an embryo in which the polar-plasm was ultraviolet irradiated at the 2-nuclei stage. The ovary consists entirely of small cells (r.). The oviduct (ov.) can be seen.

(F) Transverse section through the abdomen in the middle of the pupal instar. This individual was derived from an embryo in which the polar-plasm had been ultraviolet-irradiated at the 2-nuclei stage. The follicles (f.) are attached to the oviduct (ov.) and have remained small at the stage when they would normally have enlarged and filled the abdomen. The follicles contain groups of cells with only the somatic number of chromosomes. The large abdominal space was originally filled with red pigment; in normal individuals this pigment is deposited in the ooplasm of the developing oocytes.

(G) Transverse section through part of the ovary towards the end of the pupal instar. This individual was derived from an embryo in which the posterior end had been constricted at the 2-nuclei stage so that no nuclei could enter the polar-plasm. The follicles (f.c.) are enlarged and contain groups of closely packed cells (r.n.c.) with only the somatic chromosome number. There are no oocytes and hence no ooplasm. ov. = oviduct.

(E) Transverse section of part of the ovary of an adult derived from an embryo which was constricted at the 2-nuclei stage. The groups of small cells (r.n.c.) in the follicles have broken down and most of these cells have been passed into the oviduct (ov.), and now lie in the vagina.

FIGURE 4.

Transverse sections through the ovaries of normal and experimental individuals. All the sections are stained with iron haematoxylin except A, which is stained with crystal violet. A–D, Normal individuals. E–H, Experimental individuals, with ovaries lacking the E chromosomes.

(A) The ovary in the last larval instar. A large mass of follicle cells (f.c.) can be seen on the left. The larger primordial germ cells, or oogonia (oog.), form a smaller group to the right.

(B) Section through the ovaries in the middle of the pupal instar. The follicles contain groups of oogonia and are attached to the inner end of the oviduct (ov.).

(C) Transverse section through the abdomen of a late pupa, showing the two ovaries much enlarged due to the formation of ooplasm (oop.) in the follicles (f). Each follicle contains a group of nurse cells (n.c.) with deeply staining cytoplasm.

(D) The same individual as C, showing the follicles composed of follicle cells (f.c.) containing the nurse cells (n.c.) and ooplasm. The ooplasm of the oocyte in each follicle is lightly stained and the nurse cells have large spherical nuclei with deeply staining irregular chromosomes.

(E) The ovary in a late larva derived from an embryo in which the polar-plasm was ultraviolet irradiated at the 2-nuclei stage. The ovary consists entirely of small cells (r.). The oviduct (ov.) can be seen.

(F) Transverse section through the abdomen in the middle of the pupal instar. This individual was derived from an embryo in which the polar-plasm had been ultraviolet-irradiated at the 2-nuclei stage. The follicles (f.) are attached to the oviduct (ov.) and have remained small at the stage when they would normally have enlarged and filled the abdomen. The follicles contain groups of cells with only the somatic number of chromosomes. The large abdominal space was originally filled with red pigment; in normal individuals this pigment is deposited in the ooplasm of the developing oocytes.

(G) Transverse section through part of the ovary towards the end of the pupal instar. This individual was derived from an embryo in which the posterior end had been constricted at the 2-nuclei stage so that no nuclei could enter the polar-plasm. The follicles (f.c.) are enlarged and contain groups of closely packed cells (r.n.c.) with only the somatic chromosome number. There are no oocytes and hence no ooplasm. ov. = oviduct.

(E) Transverse section of part of the ovary of an adult derived from an embryo which was constricted at the 2-nuclei stage. The groups of small cells (r.n.c.) in the follicles have broken down and most of these cells have been passed into the oviduct (ov.), and now lie in the vagina.

In the experimental females follicles are formed but these contain only groups of cells with nuclei with the somatic chromosome number. These cells are presumably derived from the small germ cells formed in the embryo. It is during the pupal instar that the effects of the absence of the E chromosomes become apparent. The cells in the follicles show no differentiation into nurse cells and oocytes. Even though there are no oocytes in the follicles, these enlarge so that their walls become folded and spaces appear between the follicle walls and the groups of cells in each follicle (Fig. 4G). Oocyte enlargement in normal pupae is accompanied by the appearance of a secretion which stains strongly with iron haematoxylin and which appears to pass from the nurse cells into the ooplasm in each follicle. The whole of the ovary becomes red in colour due to the appearance of red pigment which is deposited in the ooplasm and which later persists round the nuclei of the developing embryo. The deeply staining secretion is absent from the ovaries of experimental embryos. The red pigment, however, does appear, so that the pupal abdomen becomes swollen and red in colour. It was found that the pigment, instead of being deposited in the follicles, was secreted in the large abdominal space (Fig. 4F) which in the normal pupa would have been occupied by the enlarged ovary.

Towards the end of pupal life the remainder of the reproductive system develops. This consists of the egg-laying apparatus, accessory glands and spermatheca, and all were normal in the experimental individuals. It was found that in late experimental pupae the small cells occupying the follicles began to separate (Fig. 4H) and were eventually discharged irregularly into the oviduct after emergence of the imago. Later they were passed as far back as the vagina, by which time they had broken up and their nuclei had assumed the form of deeply Feulgen-positive globules. It was noticed that the cells began to be discharged into the oviducts at the time that the oocytes are released into the reproductive tract of a normal female individual, and that this was continued during the time that normal egg-laying would have taken place.

Fiftγ-four female midges were obtained from the experimental embryos which survived. These adults appeared to be perfectly normal except in reproductive capacity. They were capable of mating with normal and experimental males and after mating attempted to lay eggs by walking up wheat leaves and extending their ovipositors. Occasionally small droplets of colourless fluid were deposited.

The absence of the E chromosomes from the germ line of the female is thus associated with the following developmental abnormalities:

  • Only the germ cells are affected. The rest of the body, including the somatically derived part of the ovary and the rest of the reproductive system, is normal.

  • Ovarian follicles are formed but these contain only small cells which show no differentiation into oocytes and nurse cells.

  • The oocyte growth stage is deficient and the midges are sterile; sterility is manifest from the moment that gametes would have begun to develop had the individuals been normal.

  • In the adult female midge the cells occupying the follicles become free and are discharged into the oviducts. They accumulate in the ovipositor, by which time they have broken down and disintegrated into Feulgen-positive globules.

  • The behaviour of the female midges is unaffected. They mate and attempt to lay eggs.

2. Males

Of the 154 experimental embryos which survived, only seventeen developed into males and so information concerning the development of abnormal testes is based on small numbers. Up to the middle of the last larval instar the testes in the normal male consist of a spherical mass of primordial germ cells, surrounded by a thin wall of somatic cells. Early in the pupal instar some of the primordial germ cells become large ‘nutritive’ cells; the remainder are spermatogonia (Fig. 5A). The spermatogonia develop by two divisions into spermatozoa. The first division involves unequal cytokinesis and results in the formation of small inter-kinetic cells with four chromosomes and an equal number of residual cells containing the remainder of the chromosomes possessed by the original spermatogonia. The second division concerns only the inter-kinetic cells and results in the formation of spermatids. These develop tails before the end of pupal life and the resultant spermatozoa are passed by vasa deferentia to a seminal vesicle lying near the posterior end of the abdomen. The nucleus of the spermatozoon is divisible into two parts (Fig. 5B, C).

FIGURE 5.

Sections through the testes in normal and experimental individuals. All the sections are stained with iron haematoxylin. A–C, Normal individuals. D–F, Experimental individuals, with testes lacking the E chromosomes.

(A) The testis early in the pupal instar. Four large ‘nutritive’ cells (nat.c.) can be seen, with the primordial germ cells, or spermatogonia (5.), arranged loosely round them.

(B) The testis (t) in a pupa just before emergence. Most of the spermatozoa (sp.) have already passed to the seminal vesicles, but some can be seen entering the vas deferens (vas.). The residual cells (res.) and the ‘nutritive’ cells remain in the testis.

(C) Part of a seminal vesicle in a pupa just before emergence. Spermatozoa (sp.) can be seen entering the seminal vesicle (s.v.) at the anterior end. Spermatozoa can also be seen in the vas deferens (vas.). The nucleus of each spermatozoon is divided into two parts.

(D) The testes in a mid-pupa derived from an embryo which was constricted at the 2-nuclei stage. Each testis consists of a mass of small cells (r.).

(E) A testis (t.) towards the end of the pupal instar of an individual derived from an embryo in which the polar-plasm was ultraviolet-irradiated at the 2-nuclei stage. The testis is sac-like in form and leads into the vas deferens (vas.). It contains cells the nuclei of some of which have assumed a resemblance to spermatid nuclei (‘sp.).

(F) Part of the seminal vesicle in an adult derived from an embryo in which the polar-plasm was ultraviolet-irradiated at the 2-nuclei stage. The anterior end contains Feulgen-positive particles (‘sp.) derived from the cells originally occupying the testis.

FIGURE 5.

Sections through the testes in normal and experimental individuals. All the sections are stained with iron haematoxylin. A–C, Normal individuals. D–F, Experimental individuals, with testes lacking the E chromosomes.

(A) The testis early in the pupal instar. Four large ‘nutritive’ cells (nat.c.) can be seen, with the primordial germ cells, or spermatogonia (5.), arranged loosely round them.

(B) The testis (t) in a pupa just before emergence. Most of the spermatozoa (sp.) have already passed to the seminal vesicles, but some can be seen entering the vas deferens (vas.). The residual cells (res.) and the ‘nutritive’ cells remain in the testis.

(C) Part of a seminal vesicle in a pupa just before emergence. Spermatozoa (sp.) can be seen entering the seminal vesicle (s.v.) at the anterior end. Spermatozoa can also be seen in the vas deferens (vas.). The nucleus of each spermatozoon is divided into two parts.

(D) The testes in a mid-pupa derived from an embryo which was constricted at the 2-nuclei stage. Each testis consists of a mass of small cells (r.).

(E) A testis (t.) towards the end of the pupal instar of an individual derived from an embryo in which the polar-plasm was ultraviolet-irradiated at the 2-nuclei stage. The testis is sac-like in form and leads into the vas deferens (vas.). It contains cells the nuclei of some of which have assumed a resemblance to spermatid nuclei (‘sp.).

(F) Part of the seminal vesicle in an adult derived from an embryo in which the polar-plasm was ultraviolet-irradiated at the 2-nuclei stage. The anterior end contains Feulgen-positive particles (‘sp.) derived from the cells originally occupying the testis.

During the larval instars and the first part of the pupal instar a testis lacking the E chromosomes consists of a mass of small cells surrounded by a wall of somatic cells (Fig. 5D). The central cells are presumably derived from the small germ cells formed in the embryo, and they fail to show any sign of differentiation into ‘nutritive’ cells and spermatogonia. During the pupal instar the wall of each testis becomes more distinct and the central cells become separated from each other. The nuclei of these cells do not show any single consistent appearance; sometimes they are divided into two parts, as in a normal spermatid nucleus, but more usually they are irregular in shape and size (Fig. 5E). They are approximately twice the size of normal spermatid nuclei and the cytoplasm of these cells is sometimes elongated into tails, but this is not a consistent feature. It appears then that cells with the somatic chromosome number in the male gonad attempt to form gametes in the sense that some of them come to resemble spermatids. This appears to take place without division.

The rest of the reproductive system develops towards the end of the pupal instar and is completely normal in experimental males. Just before emergence of the male midge the testis tends to become emptied of the cells lying in it; these are passed through the vas deferens and sometimes a few of them reach the anterior end of the seminal vesicles. By this time the cells have disintegrated and their nuclei have assumed the form of Feulgen-positive globules (Fig. 5F).

Experimental males were capable of mating with both normal and experimental females; the tendency to mate was not diminished in either sex. In matings between normal individuals, spermatozoa are passed to the spermatheca of the female. When a sterile male mated there was no transmission of gametes or of cells of any kind; the spermatheca was quite empty after such a mating.

The absence of E chromosomes from the male germ line is associated with the following abnormalities:

  • Only the gonads are affected. The rest of the body, including the rest of the reproductive system, is normal.

  • The testes fail to enlarge at the usual time during the pupal instar, and large ‘nutritive’ cells are not formed.

  • Some of the small cells occupying the testes assume a slight resemblance to spermatids. This takes place apparently without division. The nuclei become compact and occasionally divided into two parts. They are larger than normal spermatid nuclei. The cytoplasm of some of these cells tends to become elongated into a tail-like structure.

  • Functional gametes are not formed, and the midges are sterile. The cells in the testes are passed into the vasa deferentia, where they begin to disintegrate ; a few deeply staining dead nuclei are passed as far back as the seminal vesicles.

  • The mating behaviour of the male midges is normal.

The observations on the development of normal and experimental individuals of Mayetiola establish that:

  1. A cleavage gradient exists during the first 4 h of development; the nuclei towards the anterior end of the embryo enter prophase slightly before those nearer the posterior end of the embryo.

  2. Chromosome elimination occurs at the fifth division when fourteen nuclei divide; each nucleus loses between 31 and 34 chromosomes. The eliminated chromosomes are called the E chromosomes.

  3. The somatic nuclei contain eight chromosomes, the S chromosomes.

  4. The moment of the elimination of chromosomes from the presumptive somatic nuclei cannot be experimentally altered; it always occurs at the fifth division.

  5. The germ line develops from pole cells which are formed at the posterior end of the embryo. The entry of nuclei into the polar-plasm at the posterior end is marked by a reduction in the frequency of mitosis of these nuclei which retain the E chromosomes during their divisions.

  6. The two germ-line nuclei do not divide during the fifth division of the presumptive somatic nuclei.

  7. The posterior end of the embryo contains polar granules which become associated with the membranes of the germ-line nuclei during the fifth division.

  8. The granules cannot be regenerated if they are destroyed during cleavage.

  9. The polar granules consist at least in part of RNA and probably mitochondria.

  10. The polar granules are not necessary for pole cell formation.

  11. The E chromosomes are not necessary for the formation of pole cells.

  12. The E chromosomes are necessary for the formation of gametes.

  13. All the nuclei are capable of undergoing chromosome elimination at the fifth division. This occurs if nuclei are prevented from entering the polar-plasm before the division.

The various aspects of the work will be discussed in turn.

The cause of chromosome elimination

The present work does not elucidate the causes of chromosome elimination, but it is useful to consider the past approaches to this phenomenon, and also to the comparable case of Ascaris (Nematoda), where parts of chromosomes are lost from the somatic nuclei during cleavage (chromosome diminution). The first experimental work on Ascaris was performed by Boveri (1910). On the basis of abnormal cleavages after centrifugation of eggs and embryos of Ascaris, Boveri came to the conclusion that diminution was caused by cytoplasmic zones situated perpendicular to the animal/vegetal axis, and that these zones caused diminution in the somatic part of the embryo and prevented it from occurring in the germ-line nucleus. With respect to chromosome elimination in Miastor, Kraczkiewicz (1936) suggested that there were cytoplasmic zones which were responsible for changing the state of the spindle and which could cause elimination. He considered that the existence of a gradient of elimination was evidence of the presence of cytoplasmic zones. Du Bois (1933), working with Sciara, suggested that there were concentrically arranged zones in the embryo, and that elimination took place when the nuclei lay in the appropriate zones. The present work on Mayetiola shows that chromosome elimination can take place anywhere in the somatic part of the embryo and does not appear to be linked to the existence of cytoplasmic zones or centres in the embryo. Geyer-Duszynska (1959) showed that in Wachtliella chromosome elimination can take place anywhere in the cytoplasm, and even in the polar-plasm provided that the polar granules have been removed from this area. The constriction experiments show that in Mayetiola chromosome elimination takes place at the fifth division irrespective of the number of nuclei lying in the somatic part of the embryo. It is thus likely that the entire embryonic cytoplasm allows elimination to take place in it.

The immediate causes of elimination are unknown. All other authors (White, 1950; Geyer-Duszynska, 1959; Nicklas, 1959, 1960) have been of the opinion that the E chromatids are attached normally to the spindle fibres during anaphase. In Monarthropalpus the E chromatids do not separate and White (1950) suggests that the E chromatids are held together by some matrix and that this was the immediate cause of elimination. In both Mayetiola and Wachtliella the E chromatids separate completely but fail to continue anaphase, and Geyer-Duszynska (1959) suggests that functional defects in the centromeres of the E chromatids were responsible for causing elimination. By means of the ultraviolet irradiation of small areas of cytoplasm round the nuclei before elimination, Geyer-Duszynska showed that in the absence of a spindle the E chromatids were capable of completing anaphase during the division in which they would never normally have reached the poles. Since a normal spindle is thus necessary for chromosome elimination it is likely that elimination is due to some active factor which is centred in or around the spindle. At present the origin and nature of this factor is unknown.

The causes of the retention of the E chromosomes by the germ-line nuclei

The formation of the binucleate primordial germ cell (pole cell) is marked by two events. The first is an increase in the time between each mitosis of the germline nuclei and the second is that the germ-line nuclei retain the E chromosomes during their divisions. A similar situation occurs in Monarthropalpus, Wachtliella and Miastor.

If it is accepted that elimination is caused by the action of some factor round the cleavage nuclei it follows that this factor must either be absent from the polar-plasm, or that if it is present, its action must be inhibited in some way if the germ-line nuclei are to retain the E chromosomes. Since in Mayetiola the polar granules become associated with the nuclear membranes from the time of the fifth division, when the germ-line nuclei do not divide, up to 14 h of development, during which time the germ-line cells are dividing only infrequently, it is likely that the granules are in some way concerned with both the retention of the E chromosomes when the germ-line nuclei divide and with the low frequency of these divisions. The constriction experiments on Mayetiola indicate that the retention of the E chromosomes is caused by a posterior factor. When a cleavage nucleus is prevented from entering the polar-plasm before the fourth division it divides with loss of chromosomes at the time when it would normally have retained them and not divided. Nuclei with the reduced chromosome number then enter the polar-plasm, become associated with the polar granules, and subsequently divide less frequently than somatic nuclei. Geyer-Duszynska (1959) found that in Wachtliella division with elimination took place in the polar-plasm provided that the polar granules had been removed from the area. She also found that in some of the embryos the polar granules were moved by centrifugation into the somatic part of the embryo as a single mass. If this mass came into contact with a nucleus with the full chromosome number the nucleus divided less frequently and retained the E chromosomes. This suggests that the polar granules are directly concerned both with the non-division of the germ-line nuclei during the time that chromosome elimination occurs in the rest of the embryo, and also with the onset of the lower frequency of mitosis of the germ-line nuclei. In Mayetiola the polar granules become displaced and are dispersed very easily by centrifugation. This is unusual since it has been found that in other Diptera the polar granules are not moved very easily (Howland, 1941 ; Nicklas, 1959). Because the polar granules are dispersed in Mayetiola any effect they may have had is lost and all the nuclei divide with chromosome elimination at the fifth division in spite of the presence of dispersed polar granules in the somatic part of the embryo. Nuclei with the reduced chromosome number sometimes enter the polar-plasm after centrifugation, but since these embryos unfortunately die it is not possible to ascertain whether the small pole cells that are formed, lacking both polar granules and E chromosomes, divide at a different rate from the somatic nuclei.

It is interesting to speculate as to how the polar granules carry out their apparent function of causing the retention of the E chromosomes. The way they do this must depend on the original source of the factor causing elimination. If this factor is already present in the cytoplasm the granules could prevent it from coming into contact with the chromosomes by preventing it from entering the germ-line nuclei. If on the other hand the elimination inducer is produced by the cleavage nuclei themselves then the polar granules could function by preventing such a factor from leaving the germ-cell nuclei. In either case it is probable that the granules function by blocking the activity and penetration or release of an elimination inducer. Or in other words the granules interfere with some nucleo-cytoplasmic interaction which would otherwise result in the appearance of irreversible changes to the nuclei in the form of loss of chromosomes.

Cleavage gradients and chromosome elimination

Dipteran development is characterized by the presence of a cleavage gradient (Krause & Sander, 1962); generally this is centred at the point of formation of the first nucleus and causes the cleavage nuclei nearest this point to enter prophase slightly before those farther away. A cleavage gradient is present in Cecidomyidae and it follows that elimination takes place in a gradient. Kracz-kiewicz (1936) considered that the presence of a gradient in cecidomyid development was evidence of the presence of cytoplasmic zones which were responsible for causing chromosome elimination. However, since a cleavage gradient exists in insects which do not undergo elimination, and since the gradient operates at divisions other than those at which elimination occurs, it is likely that the two phenomena are unrelated. It is unlikely that the polar granules (which are associated with a lower frequency of mitosis of the germline nuclei) are responsible for the appearance of a cleavage gradient since the gradient exists after centrifugation even when the polar granules are dispersed or moved in a mass into the somatic part of the embryo.

The function of the E chromosomes

After their elimination from the presumptive somatic nuclei the E chromosomes form strongly Feulgen-positive lumps lying in the somatic part of the embryo. These lumps break up but do not undergo mitosis; they eventually disappear when the yolk is digested completely. Nicklas (1959) has shown by Feulgen cytophotometry that DNA is not synthesized in the E chromosomes after their elimination in Miastor and it is likely that this is true for all Cecidomyidae. The E chromosomes are thus unlikely to contribute to the development of the soma after their elimination. On the other hand, the present work on Mayetiola shows that the E chromosomes are concerned exclusively with gamete formation and confirms the work of Geyer-Duszynska (1966) on Wacht-liella persicariae. Mayetiola individuals lacking the E chromosomes from their germ-line nuclei are normal in every respect other than in gamete formation. In females the effects of the absence of the E chromosomes become manifest when oocytes should be formed; there is a total failure to form both ooplasm and some of the accompanying secretions. Painter (1966) suggests that the E chromosomes serve to increase the polysome-forming capacity of the nurse cells; in Drosophila the nurse cells show a series of endomitotic divisions which are associated with an increased yield of nucleolar material. It is not known whether the E chromosomes of the nurse cells in Cecidomyidae are concerned with polysome formation, but the fact that the effects of the absence of the E chromosomes become manifest at the time that ooplasm would be expected to develop suggests that this may be so. In males the abnormalities associated with the absence of E chromosomes are less drastic since some cells are formed which resemble gametes. These cells are unable to fertilize normal eggs.

The significance of chromosome elimination

An understanding of the significance of chromosome elimination must be related to that of the evolution of the cytogenetic system found in the Cecidomyidae. As White (1950) points out, on morphological evidence the Cecidomyidae belong to the suborder Nematocera of the Diptera, and most probably form a natural group together with the Mycetophilidae and the Sciaridae. The Mycetophilidae are considered to be the most primitive and in this family the germ line and somatic chromosome number are the same. In those Sciaridae which have been analysed cytologically the elimination of two or three autosomal chromosomes takes place from the somatic nuclei during cleavage. In the Cecidomyidae the full chromosome number is very much larger than it is in the Sciaridae and in some but not in all is a multiple of the somatic number. Nicklas (1959) has shown by measurements of the centromere positions and the relative lengths of the chromosomes that half the germ-line chromosome number in Miastor cannot be homologous with the somatic number. On the other hand the possibility of tetraploidy of the haploid somatic number is not excluded. Since the E chromosomes do not form bivalents during meiosis it is likely that they form a genetically distinct chromosomal group, and since in the more primitive Mycetophilidae the somatic and germ-line nuclei have the same chromosome number it is likely that the large chromosome number in Cecidomyidae is of a secondary nature and has arisen by an increase of an originally smaller number. In other words, a system seems to have arisen whereby the genes necessary for gametogenesis have accumulated in particular chromosomes which are for some reason eliminated from the somatic nuclei. There is no explanation as to why this separation should have occurred between somatic and germ-line genes, and there is no explanation as to the origin of the extra chromosomes. It is also not obvious why the necessary rearrangement within the genotype which would give the extra chromosomes their particular functions in the germ line should have been selected for. Nicklas (1960) considers that chromosome elimination concerns the preservation of a certain nucleo-cytoplasmic ratio based on the volume of the nuclei relative to the volume of cytoplasm surrounding them. This interpretation could apply to the Cecidomyidae and the Orthocladiinae, but hardly to the Sciaridae, where the eliminated chromosomes are so few in number.

In conclusion it can be said that chromosome elimination represents an extreme example of the loss of totipotency of somatic nuclei during development and must be regarded as an extreme specialization not having any obvious relation to the development of animals which do not show similar changes.

Expériences sur des éliminations chromosomiques chez Cécidomyde Mayetiola destructor

La segmentation chez Cecidomyidae (Diptera) se caractérise par des éliminations de chromosomes dans les noyaux somatiques présomptifs. La formule chromosomiale complète est conservée dans les noyaux de la lignée germinale.

Le déroulement de la segmentation est décrit chez Mayetiola destructor (Say). Après la 4ème division, deux noyaux se trouvent dans le plasme polaire postérieur et s’associent aux granules polaires, tandis que 14 noyaux se situent dans le reste du cytoplasme. Tous les noyeaux possèdent à peu près 40 chromosomes. Pendant la 5ème division, les noyaux postérieurs ne se divisent pas et le plasme polaire se pince pour former les cellules germinales primordiales (cellules polaires). Les 14 autres noyaux se divisent et perdent à peu près 32 chromosomes de sorte que les 28 noyaux ainsi formés ne possèdent que huit chromosomes. Ce sont là les noyaux somatiques. Au cours des divisions subséquentes, les noyaux des cellules polaires gardent la formule chromosomiale complète; ces divisions sont moins fréquentes que celles des cellules somatiques.

Il a été fait des expériences sur les stades embryonnaires précoces afin d’élucider les propriétés de la partie postérieure de l’œuf au cours de la période d’élimination chromosomique à partir des noyaux somatiques. On a utilisé l’irradiation aux rayons ultra-violets, la constriction et la centrifugation.

Les granules polaires jouent un rôle dans l’absence de division des cellules germinales au cours du 5ème cycle mitotique car, lorsque ces granules sont dispersés par la centrifugation ou lorsque les noyaux sont empêchés de venir à leur contact, avant la 5ème division, à la suite d’une constriction, tous les noyaux se divisent avec élimination chromosomique au cours de cette 5ème division.

Des individus ayant une lignée germinale avec seulement huit chromosomes ont été menés à maturité. Les anomalies n’ont été observées que dans la lignée germinale et ont été identiques quelle qu’ait été la technique utilisée pour produire cette déficience de la lignée germinale. Un ovaire contenant des cellules germinales à huit chromosomes seulement est incapable de constituer des oocytes ni des cellules nourricières. Un testicule contenant des cellules germinales à huit chromosomes est incapable de former des spermatocytes mais bien des cellules qui ont quelque ressemblance avec les gamètes. Ces mâles et femelles expérimentaux étaient également stériles.

Les résultats sont discutés dans le cadre d’autres travaux expérimentaux sur les Céci-domydes et on a pu arriver aux conclusions suivantes: (a) les granules polaires sont responsables d’empêcher la perte irréversible des noyaux de cellules germinales et les empêchent de se diviser au cours du 5ème cycle; (6) les chromosomes retenus normalement dans la lignée germinale sont requis pour la gamétogénèse, en particulier l’oogénèse.

La signification de l’élimination chromosomique est discutée.

This work was carried out in the Department of Zoology of Oxford University during the tenure of a Christopher Welch Scholarship. I am grateful to many members of the Department for the help I was given, and in particular to Professor M. Fischberg and to Dr J. B. Gurdon for their advice, supervision, and many helpful suggestions. I would like to thank Mr R, Haveley and Miss L. Hartwell of the Department of Biology and Geology of the Northern Polytechnic for their assistance with the Plates.

Bantock
,
C.
(
1961
).
Chromosome elimination in Cecidomyidae
.
Nature, Land.
190
,
466
467
.
Blackler
,
A. W.
(
1958
).
Contribution to the study of germ cells in the Anura
.
J. Embryol. exp. Morph.
6
,
491
503
.
Boveri
,
T.
(
1910
).
Über der Teilung Centrifugierter Eier von Ascaris megalocephala.
Arch. EntwMech. Org.
30
,
101
145
.
Carleton
,
H. M.
&
Drury
,
R. A. B.
(
1957
).
Histochemical Technique.
Oxford University Press
.
Czolowska
,
R.
(
1969
).
Observations on the origin of the ‘germinal cytoplasm’ in Xenopus laevis.
J. Embryol. exp. Morph.
22
,
229
251
.
Du Bois
,
A. M.
(
1933
).
Chromosome behaviour during cleavage in the eggs of Sciara coprophila (Diptera) in relation to the problem of sex determination
.
Z. Zellforsch. mikrosk. Anat.
19
,
595
614
.
Geigy
,
R.
(
1931
).
Action de l’ultra-violet sur le pôle germinal dans l’œuf de Drosophila melano-gaster
.
Revue suisse Zool.
38
,
187
288
.
Geyer-Duszynska
,
I.
(
1959
).
Experimental research on chromosome elimination in Cecidomyidae (Diptera)
.
J. exp. Zool.
141
,
391
448
.
Geyer-Duszynska
,
I.
(
1961
).
Spindle and chromosome behaviour after partial embryo irradiation
.
Chromosoma
12
,
233
247
.
Geyer-Duszynska
,
I.
(
1966
).
Genetic factors in oogenesis and spermatogenesis in Cecidomyidae
. In
Chromosomes Today,
vol.
1
(ed.
C. D.
Darlington
&
K. R.
Lewis
), pp.
174
178
.
Edinburgh
:
Oliver and Boyd
.
Howland
,
R. B.
(
1941
).
Structure and development of centrifuged eggs and early embryos of Drosophila melanogaster.
Proc. Am. phil. Soc.
84
,
605
616
.
Kahle
,
W.
(
1908
).
Die Paedogenesis der Cecidomyiden
.
Zoológica, Stuttg.
21
,
1
80
.
Kraczkiewicz
,
Z.
(
1936
).
Etudes cytologiques sur l’oogénèse et la diminution de la chromatine dans les larves paedogénétiques de Miastor metraloas (Meinert) (Diptera)
.
Folia morph.
6
,
1
37
.
Krause
,
G.
&
Sander
,
K.
(
1962
).
Ooplasmic reaction systems in insect embryogenesis
.
Adv. Morphog.
2
,
259
303
.
Mahowald
,
A. P.
(
1962
).
Fine structure of pole cells and polar granules in Drosophila melanogaster.
J. exp. Zool.
151
,
201
207
.
Metcalfe
,
M.
(
1935
).
The germ cell cycle in Phytophaga destructor, Say
.
Q. J! Microsc. Sci.
77
,
585
603
.
Nicklas
,
R. B.
(
1959
).
An experimental and descriptive study of chromosome elimination in Miastor spec
.
Chromosoma
10
,
301
336
.
Nicklas
,
R. B.
(
1960
).
The chromosome cycle of a primitive cecidomyid—Mycophila speyeri. Chromosoma
11
,
403
419
.
Painter
,
T. S.
(
1966
).
The role of the E-chromosomes in Cecidomyidae
.
Proc. natn. Acad. Sci., U.S.A.
56
,
853
855
.
Pusey
,
H. K.
(
1939
).
Methods of reconstruction from microscopic sections
.
JI R. microsc. Soc.
59
,
232
244
.
White
,
M. J. D.
(
1950
).
Cytological studies on Gall Midges (Cecidomyidae)
.
Univ. Texas Publ. No.
5007
.