We have isolated a laboratory strain of Chironomus samoensis in which determination of the anteroposterior egg polarity is disturbed. Most conspicuous is the spontaneous formation of ‘double abdomen’ embryos where head and thorax are replaced by a mirror image of the abdomen. Such double abdomens are found in about half of the egg clusters in this strain, which we call the spontaneous double abdomen (sda) strain as opposed to the normal (N) strain. Also observed in the sda strain, although less frequently, are ‘double cephalon’ embryos showing a mirrorimage duplication of cephalic segments in the absence of thorax and abdomen. Moreover, embryos from the sda strain tend to form cells at the anterior pole resembling the pole cells at the posterior pole. Reciprocal crossings between the sda and the N strain indicate that the sda trait is inherited maternally. Spontaneous double abdomen formation is correlated with signs of disturbed egg architecture, including extruded yolk and detached cells. Double cephalons can also be generated by centrifuging embryos from the N strain, whereas centrifugation of sda embryos produces mostly double abdomens. Double abdomen formation can be induced experimentally by anterior u.v. irradiation of embryos from either strain. The sda trait and u.v. irradiation act in a synergistic fashion. The data suggest that the sda trait may be caused by one or more genomic mutations interfering indirectly with the activity of anterior determinants, i.e. cytoplasmic RNP particles necessary for the development of anterior segments. The sda defects may be ascribed to alterations in cytoskeletal components involved in anchoring anterior determinants and segregating them into anterior blastoderm cells.

The determination of early embryonic cells frequently involves localized cytoplasmic determinants (Freeman, 1979; Davidson, 1986). The mechanisms by which localized determinants control embryonic gene expression are still obscure. Also unresolved is the question of how localized determinants become anchored in their positions. Studies on ascidian eggs have shown that a cortical lamina of F-actin is involved in the segregation of pigment granules into a yellow crescent, which becomes incorporated into the lineage of muscle and mesenchyme cells (Jeffery & Meier, 1983, 1984). While these studies do not clarify the nature of muscle and mesenchyme determinants, they demonstrate a possible role for cytoskeletal components in localization processes.

In dipteran eggs, there seem to be three sets of localized maternal gene products which are involved in the determination of primordial germ cells (pole cells), the dorsoventral axis and the anteroposterior axis (see Kalthoff, 1983). The anteroposterior polarity appears to be controlled by mutually repressive anterior and posterior determinants. Anterior determinants which are necessary for the development of cephalic segments have been investigated extensively in Drosophila and in chironomid midges. The localization of anterior determinants has been demonstrated by combined centrifugation and u.v. irradiation experiments (Kalthoff, Rau & Edmond, 1982) and by ectopic transplantation as well as rescue experiments (Frohnhöfer & Nüsslein-Volhard, 1986; Kalthoff & Elbetieha, 1986). In the chironomid midge, Smittia sp., anterior determinants appear to be localized in the form of a shallow gradient with a maximum near the anterior pole (Kalthoff et al. 1982). In the same species, anterior determinants may become translocated during nuclear migration from a central to a more spread-out distribution beneath the surface of the embryo (Ripley & Kalthoff, 1983). A similar translocation has recently been visualized by in situ hybridization of a localized RNA sequence in Drosophila embryos with a cloned DNA probe (Frigerio et al. 1986). Here, we describe a genetic variant of Chironomus samoensis that appears to be deficient in the anchorage and translocation of anterior determinants.

Animals

The Chironomus strains used in this study were raised from egg clusters kindly provided to us by Dr Hideo Yajima, Ibaraki University, Mito, Japan. On the basis of larval and adult morphology, Yajima (personal communication) identified the material as Chironomus samoensis Edwards 1926. His identification was confirmed by Dr Wolfgang Wúlker, University of Freiburg, FRG, who also analysed the karyotype of the material (personal communication).

Laboratory culture

Larvae were kept in Petri dishes and plastic dish pans with tap water and some cellulose. They were fed a mixture of 80% nettle powder, 10% baker’s yeast, and 10% (Sergeant’s) goldfish food flakes. The larvae used both cellulose and food particles to build tubes around themselves on the bottoms of the dishes. When larvae left their tubes and formed pupae, the pans were placed in large (2′ ×2′×5′) cages with strips of filter paper soaked in 2% sucrose suspended as a food source. The air in the cages was humidified to about 90% and the water containing the larvae was aerated. The whole culture was kept at room temperature and a 16 h L:8h D light cycle. Females deposited egg clusters on solid support at the waterline, such as the walls of the dish pans or strips of filter paper hanging into the water.

Centrifugation

Embryos were centrifuged within 40 min after the two-polecell stage (P2) using a force of 6000g for 5 min at 20°C. To keep the embryos properly oriented, they were lined up in a snugly fitting groove formed by two coverslips mounted in a cuvette which was floated on a saturated sucrose solution (see fig. 1C in Kalthoff, Hanel & Zissler, 1977). The assembly was spun in a swinging bucket rotor (HB-4) in a refrigerated centrifuge (Sorvall RC5B). The embryos were oriented with either the anterior or the posterior pole facing in the direction of the centrifugal force. This treatment is referred to as anterior or posterior centrifugation, respectively. Other embryos were centrifuged laterally, i.e. with the long axis perpendicular to the centrifugal force or in oblique orientations. Embryos were analysed approximately 30 h after centrifugation and again one day later. They could be classified rather easily as normal embryos (Fig. 1A), double abdomens (Fig. 1B-G), or double cephalons (Fig. 2).

Fig. 1.

Normal and double abdomen embryos of Chironomus samoensis at germ band stages; anterior pole to the left, dorsal side up; eo, extraovate; Ic, loose cells; s, serosa. (A) Normal embryo from N strain; (B) double abdomen obtained after u.v. irradiation of early embryo from N strain; (B) double abdomen obtained after u.v. irradiation of early embryo from N strain; (C–H) spontaneous double abdomen embryos from sda strain; (C) most common symmetrical type; (D) with extraovate anteriorly; (E) with loose cells posteriorly, embryo rotated in egg shell; (F) with anterior extraovate and loose posterior cells, germ band occupying asymmetrical position with both abdominal ends shifted anteriorly; (G) hairpin-shaped germ band, both abdominal ends pointing anteriorly. Embryos A and G have developed a serosa, while the others have not. Note similarity between embryos shown under B and C. Egg length: 0·3 mm. Live photographs, Nomarski contrast.

Fig. 1.

Normal and double abdomen embryos of Chironomus samoensis at germ band stages; anterior pole to the left, dorsal side up; eo, extraovate; Ic, loose cells; s, serosa. (A) Normal embryo from N strain; (B) double abdomen obtained after u.v. irradiation of early embryo from N strain; (B) double abdomen obtained after u.v. irradiation of early embryo from N strain; (C–H) spontaneous double abdomen embryos from sda strain; (C) most common symmetrical type; (D) with extraovate anteriorly; (E) with loose cells posteriorly, embryo rotated in egg shell; (F) with anterior extraovate and loose posterior cells, germ band occupying asymmetrical position with both abdominal ends shifted anteriorly; (G) hairpin-shaped germ band, both abdominal ends pointing anteriorly. Embryos A and G have developed a serosa, while the others have not. Note similarity between embryos shown under B and C. Egg length: 0·3 mm. Live photographs, Nomarski contrast.

Fig. 2.

Spontaneous double cephalon embryos from the sda strain of Ch. samonsis; eye, eye pigment; la, labrum; st, stomadeum. The focus is on an approximately median plane in specimens A and B, and on a paramedian (lateral) plane in specimen C. Live photographs, Nomarski contrast.

Fig. 2.

Spontaneous double cephalon embryos from the sda strain of Ch. samonsis; eye, eye pigment; la, labrum; st, stomadeum. The focus is on an approximately median plane in specimens A and B, and on a paramedian (lateral) plane in specimen C. Live photographs, Nomarski contrast.

U.v. irradiation

Embryos were irradiated with ultraviolet light (u.v.) obtained from a monochromator as described previously (Ripley & Kalthoff, 1981). Embryos were oriented with their long axes parallel to, and with their anterior poles facing, the u.v. beam. Irradiation was at 285 nm wavelength and a fluence rate of 5 W m−2. The embryos were irradiated within 60 min after the two-pole-cell stage at two fluence levels (300 J m−2 and 600 J m−2). From each egg cluster, several hundred embryos were kept as an unirradiated control group in which the percentage of surviving embryos and spontaneous double abdomen frequency could be determined quite accurately. After u.v. irradiation, embryos were handled in filtered light (Schott OG 515) to prevent uncontrolled photoreactivation and stored in lightproofboxes at room temperature. Irradiated embryos were examined approximately 30 h after irradiation when segmented germ bands had formed and most embryos could already be scored as normal, double abdomen or undifferentiated (i.e. no germ band formed). Embryos of questionable nature were allowed to hatch or were released from the egg shell after 3 days with tungsten needles. Most of these embryos either had anterior defects or were asymmetrical double abdomens (Percy, Kuhn & Kalthoff, 1986). Embryos with head structures at one end and the typical set of terminal abdominal structures at the other end were scored as normal. Embryos with terminal abdominal structures at both ends were scored as double abdomens.

Spontaneous formation of double abdomen and double cephalon embryos

Females of Ch. samoensis, as well as other chironomids, lay clusters containing several hundred eggs. They are deposited within a few minutes and are embedded a gelatinous matrix. Thus, all embryos within a cluster are certainly derived from one female and, possibly, from one pair of parents. A female produces only one cluster. A mating cage in our culture typically contains a hundred adults which, at any given time, are derived from about a dozen egg clusters. Therefore, a small fraction of the matings probably occurs between sibling pairs.

The spontaneous development of double abdomen embryos was detected (by K.L.K.) during a routine selection of clusters to be used for breeding. In these embryos (Fig. 1), head, thorax and anterior abdominal segments were replaced by a mirror-image duplication of middle and posterior abdominal segments. Before this discovery, our Ch. samoensis culture had gone through a period of high mortality with many apparently unfertilized clusters. The first five clusters with double abdomens appeared over a period of 8 days. Together, they contained approximately 1900 double abdomens, 86 undifferentiated eggs and 1000 embryos of normal appearance (normal siblings). The individuals from these five clusters constituted the first generation of our spontaneous double abdomen (sda) strain (Fl in Table 1). The trait disappeared in the following generation but reappeared thereafter. From the fifth generation on, we maintained the sda strain by inbreeding only normal siblings from clusters containing double abdomen embryos. The original strain, from which the sda strain has been isolated, is referred to as the normal (N) strain. Both the N strain and the sda strain were always kept next to each other under the same culture conditions.

Table 1.

Developing clusters and double abdomen penetrance

Developing clusters and double abdomen penetrance
Developing clusters and double abdomen penetrance

We also observed the spontaneous formation of double cephalon embryos which showed a mirrorimage duplication of head structures in the absence of thorax and abdomen (Fig. 2). The first cluster containing such embryos was found in the sda strain. It contained 42 embryos showing a clear double cephalon morphology, including symmetrical duplications of eyes, labrum and stomodaeum. In addition, the same cluster included 45 embryos with pigmented eyes anteriorly and posteriorly but lacking other double cephalon characteristics, 258 undifferentiated eggs and 327 normal embryos. The latter were raised together with normal larvae from other sda clusters. This first generation produced 36 clusters with developing embryos, of which 8 contained only normal embryos, whereas 27 contained normal embryos and double abdomens. One cluster contained four double cephalon embryos, two double abdomen embryos, and eight embryos described as ‘dwarfs’ (Percy el al. 1986), as well as normal embryos and undifferentiated eggs. Attempts to breed a strain that would regularly produce clusters with double cephalons were not successful. However, clusters with double cephalon embryos have since appeared sporadically in the sda strain, but rarely in the N strain.

In addition, we found embryos in the sda strain that showed distinctive pole cells at both the anterior and the posterior egg pole. Pole cells are normally formed only at the posterior egg pole and these are primordial germ cells. The timing of early mitoses in Ch. samoensis suggests that the first pole cell encloses one out of four cleavage nuclei which moves precociously to the posterior egg pole (Kalthoff, unpublished data). This pole cell divides twice, giving rise to four distinctive pole cells. When the somatic nuclei reach the periplasm, which seems to occur during the ninth nuclear cycle, the pole cells become less conspicuous. However, they can still be recognized during the following preblastoderm stages by their complete separation from the egg cytoplasm, their round shape, large size, large nuclei and yolk inclusions (Fig. 3B). The ectopic pole cells found at the anterior pole of Ch. samoensis embryos from the sda strain perfectly mimicked their posterior counterparts at the four-pole-cell stage (Fig. 3A). After nuclear migration (Fig. 3C) and blastoderm formation, their appearance became irregular and less distinctive. Such embryos developed either normal germ bands or double abdomens. The latter were more frequent in clusters with a high overall frequency of double abdomens among the developing embryos. Thus, there was no apparent correlation between the formation of ectopic pole cells and the double abdomen pattern.

Fig. 3.

Pole cell formation at both the anterior and the posterior pole of embryos from the sda strain of Ch. samoensis. ape, anterior pole cells; m, micropyle, n, nucleus, ppc, posterior pole cells, v, yolk. (A) Embryo shortly after the formation of four pole cells at either pole; the micropyle marks the anterior pole; (B) posterior pole region of the same embryo after nuclear migration but before complete cellularization of the blastoderm. Note round shape, large size, large nuclei and yolk inclusions of pole cells as compared to blastoderm. (C) Anterior pole region of the same embryo; note smaller size, irregular shape and missing yolk inclusions of ectopic pole cells as compared to normal pole cells shown in B. Live photographs, A with crossed polarizing filters to visualize the micropyle, B and C with Nomarski contrast.

Fig. 3.

Pole cell formation at both the anterior and the posterior pole of embryos from the sda strain of Ch. samoensis. ape, anterior pole cells; m, micropyle, n, nucleus, ppc, posterior pole cells, v, yolk. (A) Embryo shortly after the formation of four pole cells at either pole; the micropyle marks the anterior pole; (B) posterior pole region of the same embryo after nuclear migration but before complete cellularization of the blastoderm. Note round shape, large size, large nuclei and yolk inclusions of pole cells as compared to blastoderm. (C) Anterior pole region of the same embryo; note smaller size, irregular shape and missing yolk inclusions of ectopic pole cells as compared to normal pole cells shown in B. Live photographs, A with crossed polarizing filters to visualize the micropyle, B and C with Nomarski contrast.

Fig. 4.

Spontaneous double abdomen frequencies in egg clusters from the sda strain of Ch. samoensis. Two samples of 61 clusters and 54 clusters were collected in 1982/3 (immediately after the isolation of the sda strain) and 1984, respectively. These samples included only clusters with double abdomens among the developing embryos. The frquency of double abdomens was classified as 0 < fsp ⩽ 10 %, 10 % < fsp < 20 %, etc. More than half of all clusters had double abdomen frequencies of either up to 10 % or more than 90 %. Note a shift towards lower frequencies between the first and second sample; the shift probably reflected the use of clusters with low double abdomen frequencies for breeding.

Fig. 4.

Spontaneous double abdomen frequencies in egg clusters from the sda strain of Ch. samoensis. Two samples of 61 clusters and 54 clusters were collected in 1982/3 (immediately after the isolation of the sda strain) and 1984, respectively. These samples included only clusters with double abdomens among the developing embryos. The frquency of double abdomens was classified as 0 < fsp ⩽ 10 %, 10 % < fsp < 20 %, etc. More than half of all clusters had double abdomen frequencies of either up to 10 % or more than 90 %. Note a shift towards lower frequencies between the first and second sample; the shift probably reflected the use of clusters with low double abdomen frequencies for breeding.

Fig. 5.

Reciprocal crossing experiment showing the maternal inheritance of the sda (spontaneous double abdomen) trait in Chironomus samoensis. Males and females from our mutant (sda) and wild-type (N) strain were put together in mating cages. Egg clusters deposited in these cages were analysed and classified according to the percentage of developing embryos (i.e. embryos reaching the germ anlagen stage) and according to the double abdomen frequency (i.e. percentage of double abdomens among the developing embryos). The data in the upper panel indicate very clearly that double abdomen frequency is controlled by the female parent. The data in the lower panel indicate that matings with sda males produced fewer clusters containing only nondeveloping, i.e. presumably unfertilized, eggs. Matings with N females, on the other hand, produced more clusters in which 90–100 % of all embryos developed.

Fig. 5.

Reciprocal crossing experiment showing the maternal inheritance of the sda (spontaneous double abdomen) trait in Chironomus samoensis. Males and females from our mutant (sda) and wild-type (N) strain were put together in mating cages. Egg clusters deposited in these cages were analysed and classified according to the percentage of developing embryos (i.e. embryos reaching the germ anlagen stage) and according to the double abdomen frequency (i.e. percentage of double abdomens among the developing embryos). The data in the upper panel indicate very clearly that double abdomen frequency is controlled by the female parent. The data in the lower panel indicate that matings with sda males produced fewer clusters containing only nondeveloping, i.e. presumably unfertilized, eggs. Matings with N females, on the other hand, produced more clusters in which 90–100 % of all embryos developed.

Fig. 6.

Correlation of anterior extraovates (filled circles, solid line) and detached posterior cells (open triangles, hatched line) with double abdomen frequency. Frequencies of extraovates and detached posterior cells are expressed as the percentages of double abdomen embryos showing these features among the total double abdomens in a cluster. Both characteristics showed a positive correction (r= +0·3) with the double abdomen frequency, i.e. the percentage of double abdomens among the developing embryos, in the same cluster.

Fig. 6.

Correlation of anterior extraovates (filled circles, solid line) and detached posterior cells (open triangles, hatched line) with double abdomen frequency. Frequencies of extraovates and detached posterior cells are expressed as the percentages of double abdomen embryos showing these features among the total double abdomens in a cluster. Both characteristics showed a positive correction (r= +0·3) with the double abdomen frequency, i.e. the percentage of double abdomens among the developing embryos, in the same cluster.

Morphological similarity between spontaneous and experimentally induced pattern abberations

The development of u.v.-induced double abdomen embryos in Smittia sp. and Ch. samoensis was observed directly and by time-lapse cinematography (Kalthoff & Sander, 1968; Kalthoff, 1975 and unpublished observations). A typical embryo obtained after anterior u.v. irradiation of an egg from the N strain of Ch. samoensis is shown in Fig. 1B. The specimen comprised two sets of six abdominal segments joined in mirror-image symmetry to a midpiece, which was longer than a normal segment. Each terminal segment showed the morphological characteristics of the 10th abdominal segment, including the rudiments of posterior prolegs, a proctodaeum and four anal papillae. The subterminal segments were formed earlier by fusion of the 8th and 9th abdominal segments (Kalthoff & Sander, 1968; Percy et al. 1986). The proximal segments could not be distinguished individually but most probably represent the 4th through 7th abdominal segment on either side. The plane of symmetry and polarity reversal would then be located in the 3rd abdominal segment. Most of this segment, together with its mirror image, appeared to form the midpiece. The same abnormal segment pattern was observed in the specimen shown in Fig. 1C, which represents a typical spontaneous double abdomen from the sda strain.

In most double abdomen embryos, both spontaneous and u.v.-induced, the germ band was symmetrical about the egg equator, with the germ band occupying the ventral (more convex) side of the egg. However, some of the spontaneous double abdomens in Ch. samoensis formed with the germ band on the dorsal (less convex) side of the egg (Fig. IE). Others developed in a position symmetrical to a meridional plane (Fig. 1G) or in a position between equatorial and meridional symmetry (Fig. 1D,F). Development in these orientations relative to the egg shell was also observed in u.v.-irradiated eggs from the N strain of Ch. samoensis (Kalthoff, unpublished data) and from Smittia (Kalthoff & Sander, 1968). Independently of the orientation of the germ anlage relative to the egg shell, the segment pattern of most of the first instar double abdomen larvae was symmetrical, i.e. showed equivalent sets of abdominal segments on both sides of a midpiece. In Ch. samoensis, we also observed asymmetrical double abdomens showing unequal numbers of segments on both sides of the plane of polarity reversal. The asymmetries, ranging from moderate to extreme, were observed in both spontaneous and u.v.-induced double abdomens (Percy et al. 1986). Similarly, the spontaneous double cephalons observed in the sda strain of Ch. samoensis (Fig. 2) were indistinguishable from double cephalon embryos obtained after centrifuging eggs from the N strain (Fig. 7), except that the stratification of yolk materials persisted in the centrifuged specimens. The double cephalon embryos of Ch. samoensis resembled most closely the double cephalons generated by centrifugation of eggs from Smittia sp. (Kalthoff et al. 1982).

Fig. 7.

Effect of spontaneous double abdomen frequency on the frequency of centrifugation-induced abnormal body patterns. The spontaneous double abdomen frequency is the percentage of spontaneous double abdomens among the developing embryos in a cluster. Results obtained using clusters with frequencies of 0, and average frequencies of 3 % and 20 %, were pooled (abscissa). The results are also grouped according to the orientation of the embryos in the centrifuge. For example, ‘Anterior’ indicates that the anterior pole was pointing in the centrifugal direction. Centrifugation was for 5 min at 6000g and 20°C. The surviving embryos were classified as normal, double cephalon or double abdomen. Note that clusters from the N strain (double abdomen frequency = 0) tended to produce double cephalons, whereas clusters with spontaneous double abdomen frequencies of 3 % or 20 % (sda strain) produced very high double abdomen yields after centrifugation. The percentage of developing embryos was 94, 84, 70 and 72 % in uncentrifuged controls and after lateral, anterior and posterior centrifugation, respectively.

Fig. 7.

Effect of spontaneous double abdomen frequency on the frequency of centrifugation-induced abnormal body patterns. The spontaneous double abdomen frequency is the percentage of spontaneous double abdomens among the developing embryos in a cluster. Results obtained using clusters with frequencies of 0, and average frequencies of 3 % and 20 %, were pooled (abscissa). The results are also grouped according to the orientation of the embryos in the centrifuge. For example, ‘Anterior’ indicates that the anterior pole was pointing in the centrifugal direction. Centrifugation was for 5 min at 6000g and 20°C. The surviving embryos were classified as normal, double cephalon or double abdomen. Note that clusters from the N strain (double abdomen frequency = 0) tended to produce double cephalons, whereas clusters with spontaneous double abdomen frequencies of 3 % or 20 % (sda strain) produced very high double abdomen yields after centrifugation. The percentage of developing embryos was 94, 84, 70 and 72 % in uncentrifuged controls and after lateral, anterior and posterior centrifugation, respectively.

Index and frequency of double abdomen formation

In the following, we use the term ‘double abdomen index’ for the percentage of clusters containing one or more double abdomen embryos among the clusters containing any developing embryos. The term ‘double abdomen frequency’ is used for the percentage of double abdomen embryos among the developing embryos of a given cluster. These terms simply describe time-dependent properties of our strains, rather than genetic alleles, since we have no independent measure of the genetic purity of our strains.

The double abdomen index changed over time in our two strains of Ch. samoensis (Table 1). In the sda strain, the index increased slowly within a year to about 14 %. Thereafter it rose quickly to above 50 % where it seems to be holding steady now. The levelling off may be, at least in part, a result of our breeding technique (see below). In the N strain, the double abdomen index increased to more than 10 % during the year following the isolation of the sda strain. Since then, the index declined and seems to hold steady now around 4% in the N strain. The continued occurrence of clusters with double abdomen embryos in the N strain is somewhat surprising because this strain is being propagated only with clusters that do not contain any double abdomens. We do not think that stray animals alone can account for this observation. We presume that the ability to produce double abdomen offspring is coupled with other traits which lead to a favourable selection for such individuals under our culture conditions. For instance, sda males appeared to be more successful at fertilization in a reciprocal crossing experiment (Fig. 5).

The frequency of double abdomen embryos was analysed in a sample of 61 clusters taken during the first months after isolation of the sda strain, and in another 54 clusters sampled more than a year later. These samples included only clusters containing double abdomen embryos. In more than half of the clusters, the double abdomen frequencies ranged either below 10 % or above 90 %, while intermediate frequencies were spread about evenly (Fig. 4). There was also a shift from a preponderance of high double abdomen frequencies in the first sample to a preponderance of low double abdomen frequencies in the later sample. This shift may have resulted from our breeding practice. To propagate the sda strain, we do not normally use clusters with high double abdomen frequencies, because there would not be enough normal siblings to sustain sizeable mating swarms. Since clusters with intermediate double abdomen frequencies are rare, clusters with low frequencies are used most of the time. The same breeding practice may have caused the stabilization of the double abdomen index around 50 % in the sda strain (Table 1).

Maternal inheritance of spontaneous double abdomen formation

In order to determine whether the spontaneous double abdomen trait was maternally inherited, reciprocal crosses were performed between the N strain and the sda strain. Pupae were collected as they swam to the surface of the water and allowed to eclose individually in capped test tubes. The sex of each adult was determined on the basis of antennal morphology. Over a period of 2 weeks (in Aug./Sept., 1984), 251 sda females were released together with 438 N males into a mating cage and 257 N females were released with 366 sda males into another cage.

At the same time, the sda and N strains were maintained in the same room as usual. The clusters obtained from the reciprocal crossings were collected and analysed, as were clusters from the N and sda strain which were deposited during the same period. For each cluster, the percentage of developing em: bryos and the double abdomen frequency were estimated.

The double abdomen index depended very clearly on the origin of the females from the sda strain (Fig. 5, upper panel). The index was as high as 87 % in the sda females × N male cross as compared to only 4 % in the N female × sda male cross. In the sda and the N strain, the index levels during this experiment were 64 % and 8 %, respectively, which were close to the levels listed in Table 1 for the corresponding 6 month period (July–Dec., 1984). It came as a surprise that the index was significantly higher (87 % versus 64 %, P< 0·005) in the sda females × N male cross as compared to the sda strain. The double abdomen frequency was also higher in clusters from the sda females x N male cross as compared to the sda strain. Moreover, as diagrammed in the lower panel of Fig. 5, the percentage of clusters with developing embryos was significantly (𝒳2 test, P< 0·001) higher in the crosses involving sda males (86 % and 81 %) than in the crosses with N males (63% and 61%), suggesting that sda males were more successful at fertilization. The percentage of clusters with more than 90% developing embryos was significantly (P < 0·005) higher in the crosses involving N females (35 % and 23 %) as compared to sda females (17 % and 13%), suggesting that N females were more likely to produce clusters of uniformly high egg quality.

Correlation of spontaneous double abdomen formation with indications of abnormal egg architecture

In a sample of 36 sda clusters, we found no correlation (r = +0·02) between the percentage of undifferentiated eggs and the double abdomen frequency. However, weak correlations were observed between the double abdomen frequency and the frequency of some structural defects among the double abdomen embryos from the same cluster. An average of 6 % of all double abdomen embryos showed an extraovate of yolk-rich material at the anterior pole (Fig. 1D,F). The frequency of anterior extraovates among the double abdomen embryos of a given cluster was positively correlated (r = +0 3) with the double abdomen frequency in the same cluster (Fig. 6, solid circles). The extraovates appeared to form after blastoderm formation since they were not observed at earlier stages.

Some spontaneous double abdomen embryos in Ch. samoensis showed small lumps of loose posterior cells which might be derived from pole cells. The loose cells were also limited to a small fraction of double abdomen embryos and their frequency showed a slightly positive correlation (r = +0·3) to double abdomen frequency (Fig. 6, open triangles). The loose cells occurred in conjunction with (Fig. IF), or independently of (Fig. ID), anterior extraovates. Neither anterior extraovates nor loose posterior cells were observed in normal siblings of double abdomens or in embryos from the N strain. Some of the spontaneous and the u.v.-induced double abdomens developed an amnioserosa as the normal embryos do (Fig. 1A,G), while most double abdomens developed without embryonic membranes.

Centrifugation and u.v. irradiation of embryos from the sda versus N strain

The double abdomen and double cephalon patterns (Figs 1, 2) which develop spontaneously in the sda strain can also be induced by centrifugation or u.v. irradiation. We tested whether embryos from the N versus sda strain reacted differently to these experimental treatments. Samples of 100 to 200 eggs from individual egg clusters were used for centrifugation while the remainder served as a control to determine survival rate and spontaneous double abdomen frequency. In none of the clusters used for centrifugation did we observe spontaneous double cephalon formation.

The results of the centrifugation experiments depended strongly on the spontaneous double abdomen frequency (fsp) in the cluster from which the experimental embryos were taken (Fig. 7). Most of the double cephalons were formed in eggs from the N strain (fsp = 0) after anterior centrifugation. Otherwise, centrifugation mainly seemed to amplify the spontaneous double abdomen frequency to higher levels, with lateral centrifugation being least effective and posterior centrifugation being most effective. Thus, posterior centrifugation boosted the double abdomen rates from spontaneous levels of 3 % and 20% to 93% and 100%, respectively.

There was considerable variation in the effects of anterior centrifugation among the clusters without spontaneous double abdomens (fsp = 0). Five out of nine clusters in this category showed a regular pattern of about 20% double cephalons, 80% normal embryos and no double abdomens. One cluster showed a record 86 % (32 out of 37 survivors) double cephalons while another cluster yielded 97 % (39 out of 40 survivors) double abdomens. Yet another cluster showed 14% double abdomens. Only one cluster showed both double cephalons (23 %) and double abdomens (6 %).

The results of anterior u.v. irradiation also depended strongly on the spontaneous double abdomen frequency in the egg cluster used. The double abdomen frequency among the irradiated embryos was corrected by subtracting the spontaneous double abdomen frequency (fsp) from the total double abdomen frequency as indicated in Fig. 8; the corrected value is called the u.v.-induced double abdomen frequency (fuv). In embryos from clusters with fsp = 0, fuv was very low for both u.v. fluence levels used (Fig. 8). This result was observed with clusters from both the N strain and the sda strain. By contrast, almost all irradiated embryos developed into double abdomens if taken from clusters with fsp = 2 % or more. An intermediate reaction was shown by embryos from clusters with fsp values between 0 % and 2%.

Fig. 8.

Effect of spontaneous double abdomen frequency (fsp) on the frequency of u.v.-induced double abdomens (fuv) in Chironomus samoensis. fsp is the frequency of spontaneous double abdomens among the developing embryos of an egg cluster, fuv is the frequency of u.v.-induced double abdomens surviving after u.v. irradiation. This value is corrected for the contribution of fsp as indicated. If clusters from the sda strain with fsp values between 2 and 55 % were used, u.v. irradiation with doses of 300 or.600 J m-2 was sufficient to cause double abdomen formation in virtually all embryos. The same treatment caused virtually no double abdomen formation when applied to eggs from clusters from the N strain lacking spontaneous double abdomens (fsp = 0). U.v. irradiation of clusters with very low spontaneous double abdomen frequencies (0 < fsp < 2 %) still boosted the fuv values to 45 % (300 J m − 2) and 90 % (600 J m − 2). The percentage of developing embryos was 93, 90 and 86%,among the unirradiated controls, the embryos irradiated with 300 J m − 2, and the embryos irradiated with 600 J m − 2, respectively.

Fig. 8.

Effect of spontaneous double abdomen frequency (fsp) on the frequency of u.v.-induced double abdomens (fuv) in Chironomus samoensis. fsp is the frequency of spontaneous double abdomens among the developing embryos of an egg cluster, fuv is the frequency of u.v.-induced double abdomens surviving after u.v. irradiation. This value is corrected for the contribution of fsp as indicated. If clusters from the sda strain with fsp values between 2 and 55 % were used, u.v. irradiation with doses of 300 or.600 J m-2 was sufficient to cause double abdomen formation in virtually all embryos. The same treatment caused virtually no double abdomen formation when applied to eggs from clusters from the N strain lacking spontaneous double abdomens (fsp = 0). U.v. irradiation of clusters with very low spontaneous double abdomen frequencies (0 < fsp < 2 %) still boosted the fuv values to 45 % (300 J m − 2) and 90 % (600 J m − 2). The percentage of developing embryos was 93, 90 and 86%,among the unirradiated controls, the embryos irradiated with 300 J m − 2, and the embryos irradiated with 600 J m − 2, respectively.

Abbreviated translocation of anterior determinants in sda embryos

The results described in the previous section are compatible with the hypothesis that eggs from the sda strain are deficient in the quantity or activity of anterior determinants. Redistribution of already diminished determinants by centrifugation would be expected to cause, in most cases, insufficient activity in both anterior and posterior egg halves and, therefore, double abdomen formation. The hypothesis would also explain the synergistic effect between u.v. light and the sda trait. Alternatively, sda embryos might be abnormal in any cofactor(s) required for anterior determinant activity or localization. For instance, sda embryos might be deficient in some cytoskeletal components that keep anterior determinants localized and help to segregate them into anterior blastoderm cells. This interpretation would be more in line with other signs of disturbed egg architecture, i.e. the extraovates and unattached posterior cells described above.

A critical test to distinguish between the two interpretations was modelled after experiments carried out earlier with embryos of another chironomid, Smittia sp. (Ripley & Kalthoff, 1983). The response of Smittia embryos to anterior u.v. irradiation (as diagrammed in Fig. 8) increased dramatically during nuclear migration. Between 5 and 7h after egg deposition, the u.v. fluence required to induce 50% double abdomens decreased from 295 to 72 J m − 2, translating into a fluence reduction factor of 0·24. These data, together with results from microbeam irradiations, suggested that anterior determinants shifted, during the intervening time, from a central position to a more spread-out distribution beneath the surface of the embryo. The following experiments with Ch. samoensis embryos were carried out to determine whether there was a similar stage-dependent change in the response to anterior u.v. irradiation and, if so, whether this change was diminished in the sda as compared to the N strain.

Ch. samoensis embryos were u.v. irradiated as described above. From each egg cluster, four samples of about 50 embryos each were used for irradiation at the two-pole-cell stage (P2), and at hourly intervals thereafter. The remainder of the cluster was set aside to determine the spontaneous frequency of double abdomen formation (fsp), which was used to determine the u.v.-induced double abdomen frequency (fuv) as indicated in Fig. 8. The u.v. fluence levels were gauged so that the fluence required to reach an fuv value of 50% could be determined by linear interpolation (Fig. 9).

Fig. 9.

(A) U.v. induction of double abdomens by anterior u.v. irradiation in Chironomus samoensis embryos. Eggs were from N strain clusters that did not form any spontaneous double abdomens. The u.v. wavelength was 285 nm, the fluence between 75 and 1200 J m −2 as indicated. Note the sharp increase of the double abdomen yield between the two-pole-cell stage (P2) and an early preblastoderm stage 3 h later. The increase coincides approximately with the nuclear migration phase, when nuclei and surrounding cytoplasm move toward the egg surface. (B) Experiment similar to the one shown in A, except that the eggs were from clusters showing spontaneous double abdomen frequencies of more than zero but less than 50 %. The frequency of u.v.-induced double abdomens (ordinate) was corrected for the spontaneous duoble abdomen frequency, as indicated in Fig. 8. In comparison with A, note that, for a given u.v. dose, the double abdomen yields were generally higher but increased less steeply with time.

Fig. 9.

(A) U.v. induction of double abdomens by anterior u.v. irradiation in Chironomus samoensis embryos. Eggs were from N strain clusters that did not form any spontaneous double abdomens. The u.v. wavelength was 285 nm, the fluence between 75 and 1200 J m −2 as indicated. Note the sharp increase of the double abdomen yield between the two-pole-cell stage (P2) and an early preblastoderm stage 3 h later. The increase coincides approximately with the nuclear migration phase, when nuclei and surrounding cytoplasm move toward the egg surface. (B) Experiment similar to the one shown in A, except that the eggs were from clusters showing spontaneous double abdomen frequencies of more than zero but less than 50 %. The frequency of u.v.-induced double abdomens (ordinate) was corrected for the spontaneous duoble abdomen frequency, as indicated in Fig. 8. In comparison with A, note that, for a given u.v. dose, the double abdomen yields were generally higher but increased less steeply with time.

With embryos from the N strain (fsp = 0, Fig. 9A), the fluence (F50) required for fuv = 50 % decreased from 1190J m-2 at stage P2 to 220J m-2 3h later, which translates into a fluence reduction factor of 0·18. Using embryos from clusters with spontaneous double abdomen formation (0 < fsp < 50 %, Fig. 9B), the F50 values decreased from 288 J m −2 to 150 J m −2, translating into a fluence reduction factor of 0·52. Thus the fluence reduction, which is presumably caused by a deshielding of anterior determinants, was 2·9 times stronger in the N strain as compared to the sda strain. In particular, the decrease in the F50 value stopped around 2 h after the P2 stage in sda embryos, whereas in N embryos there was a continued decrease between 2 and 3 h after P2. This observation suggests that the presumed translocation of anterior determinants is abbreviated in sda embryos.

Double abdomen and double cephalon embryos in other dipterans

Double abdomen and double cephalon patterns were first described by Yajima (1960, 1964), who studied the effects of centrifugation and u.v. irradiation on embryos of Chironomus dorsalis. In a later study, Yajima (1983) used the Ch. samoensis material, which he also supplied to us. Although his centrifugation procedure was somewhat different, he still obtained results that were quite similar to ours. Both double abdomens and double cephalons were obtained after either anterior or posterior centrifugation. Anterior centrifugation was generally more effective in producing the abnormal patterns as was the case in our experiments (Fig. 7). Yajima’s results also depended considerably on the source of egg material. Eggs collected in the field tended to produce more double abdomens and fewer double cephalons than eggs from his laboratory culture. Yajima obtained double abdomen embryos after anterior u.v. irradiation and double cephalons after posterior u.v. irradiation. Despite considerable effort, we have been unable to reproduce the latter result.

Our results with Ch. samoensis embryos parallel earlier data obtained with embryos of Smittia sp. (Kalthoff et al. 1982). Centrifugation of Smittia embryos also resulted in the formation of both double cephalons and double abdomens, independently of the orientation of the embryos in the centrifuge. However, posterior centrifugation was generally more effective than anterior centrifugation in this species. Also, centrifuged Smittia embryos showed a stronger propensity to form double cephalons than double abdomens. A strong cluster dependence in the results of centrifugation experiments was observed with Smittia, as well as with Ch. samoensis as described here. The formation of extraovates in spontaneous double abdomen embryos (Fig. 1D,F) was also reminiscent of an earlier observation in Smittia embryos, which were punctured at the anterior pole while they were submerged in tap water. If this operation was done before nuclear migration (i.e. before the P2+lh stage according to Fig. 9), the embryos tended to produce extraovates later during blastoderm formation. About 40% of the surviving embryos developed into double abdomen embryos (Schmidt, Zissler, Sander & Kalthoff, 1975).

Evidence of a genetic basis for the sda trait in Ch. samoensis

We suggest that our sda strain of Chironomus samoensis is a genetic variant, i.e. that it differs from the N strain in one or more genomic mutations which is (are) expressed maternally. We rule out environmental causes because these could not explain the results of the reciprocal crossing experiment (Fig. 5) and because the two strains have been kept next to each other in the same room and under the same culture conditions. A propagation of the sda trait by cytoplasmic components, such as mitochondria or microorganisms, cannot be ruled out.

Our view of the Chironomus sda strain as a maternal-effect mutant is further supported by the parallelism with the bicaudal mutant in Drosophila (Nüsslein-Volhard, 1977), which is also expressed maternally and is characterized by a very similar range of phenotypes as those of Chironomus (Percy et al. 1986). In addition, squash preparations of salivary glands showed a considerable polymorphism in the fourth chromosome of Ch. samoensis (data not shown). This indicates a considerable degree of structural heterozygosity and, therefore, a potential for segregating homozygous mutants under conditions of inbreeding. Moreover, Southern blots of genomic DNA, probed with cloned Drosophila caudal cDNA (kindly provided by W. Gehring), show a 10 kb EcoRI fragment that is present in the Chrionomus sda strain but absent from the N strain (Kalthoff, unpublished data). Finally, the discovery of the sda trait was preceded by a period of high mortality during which there may have been inbreeding combined with a strong selection for male mating behaviour or other traits possibly associated with the sda trait.

Evidence for a defective localization of anterior determinants in sda embryos

For the purpose of the following discussion, we define anterior determinants as localized cytoplasmic components which are necessary for the development of anterior segments. In Smittia eggs, anterior determinants were characterized as RNP particles on the basis of u.v. action spectra, sensitivity to enzymes, and data from combined centrifugation and u.v. irradiation experiments (see Kalthoff, 1983). More recently, we have established a bioassay for anterior determinant activity in Ch. samoensis embryos. These embryos are programmed for double abdomen development by anterior u.v. irradiation, as diagrammed in Figs 8, 9. Such embryos can be ‘rescued,’ i.e. normal development can be restored, by microinjecting cytoplasmic or RNA fractions near the anterior pole. The extent of the rescue, i.e. the frequency of normal embryos among the injected survivors as compared to a buffer-injected control sample, is a measure of anterior determinant activity. This activity was found in anterior but not in posterior cytoplasm, and in poly(A)+ RNA (Kalthoff & Elbetieha, 1986).

Most of the data presented here could be explained by the assumption that embryos from the sda strain have a reduced amount of, or molecular alteration in, anterior determinants which would render them less active. However, this assumption would not account for the abbreviated translocation of anterior determinants, as indicated by the data shown in Fig. 9. This observation would be better explained by assuming that sda embryos are defective in some cytoskeletal components involved in the apparent translocation of anterior determinants, which appears to be necessary for segregating them into anterior blastoderm cells. The latter hypothesis would still account for the general synergism between anterior u.v. irradiation and the sda trait. The decreased frequency of double abdomens in centrifuged sda embryos would also be explained because centrifugation is likely to exacerbate the consequences of anterior determinant anchorage to defective cytoskeletal components. Moreover, defective cytoskeletal components might be related to the extraovates and/or loose posterior cells observed in spontaneous double abdomen embryos (Figs 1, 6). Finally, a defective anchorage of anterior, and perhaps other, localized determinants would be well suited to account for the occurrence of double cephalons, and ectopic pole cell duplicates, in the same clusters with double abdomen embryos. It might also have been the reason why our attempts to breed a spontaneous double cephalon strain failed.

The preceding discussion, as well as the design of the experiment shown in Fig. 9, have been based on the assumption that anterior determinants in chironomid embryos do in fact undergo a translocation from central to peripheral sites in the embryo. Until recently, this assumption had only been inferred from changes in double abdomen yields obtained after u.v. irradiation of different target areas at different stages of development (Ripley & Kalthoff, 1983). It was therefore reassuring to observe this type of translocation directly in Drosophila eggs and embryos for an RNA sequence hybridized in situ by a cloned cDNA probe (Frigerio et al. 1986). This sequence, designated as prd 4, probably represents the product of the bicoid gene which is necessary for head development (Frohnhófer & Nüsslein-Volhard, 1986). While the homology between the bicoid gene product and anterior determinants in chironomids is an open question, the in situ hybridization data of Frigerio et al. directly show that translocations of localized RNA sequences, as inferred from our u.v. irradiation data (Ripley & Kalthoff, 1983), do occur.

The presumed involvement of cytoskeletal components in the anchorage and translocation of anterior determinants also helps to explain some data obtained in previous experiments. Puncturing Smittia embryos at the anterior pole while they were submerged in water caused double abdomen development and extraovate formation as mentioned above (Schmidt et al. 1975). The osmotic disturbance caused by the influx of water may disrupt cytoskeletal components, which in turn may cause both extraovate formation and a disruption in the normal translocation of anterior determinants. In accordance with this interpretation, extraovate and double abdomen formation were observed after puncturing before nuclear migration but not later when the translocation of anterior determinants is under way or completed (Schmidt, 1974).

The frequency of double abdomen embryos obtained after anterior u.v. irradiation of Smittia was found to be strongly dependent on the temperature during subsequent incubation. When samples of embryos were irradiated and then split and raised at 8°C versus 20°C, the resulting double abdomen frequencies differed as much as 90% and 10%, respectively (Kalthoff, 1971). Temperature-shift experiments showed that the incubation temperature was most critical between 5·5 and 8h, the time during which the translocation of anterior determinants occurs. The frequency of double abdomen embryos obtained after u.v. irradiation of Smittia eggs was also significantly increased by incubation in NaCN (Kalthoff, 1971). Taken together, the data suggest that the u.v. damage to anterior determinants might have been exacerbated by an inhibitory effect of low temperature or NaCN on the translocation mechanism.

We thank Dr H. Yajima for providing the initial egg clusters of Chironomus samoensis, and Dr W. Wülker for the karyotype analysis. This work was supported by a research grant (No. 1-1003) of the March of Dimes Birth Defect Foundation.

Davidson
,
E. H.
(
1986
).
Gene Activity in Early Development (3rd ed
.).
New York
:
Academic Press
.
Freeman
,
G.
(
1979
).
The multiple roles which cell division can play in the localization of morphogenetic potential
. In
Determinants of Spatial Organization
(ed.
S.
Subtelny
), pp.
53
76
.
New York
:
Academic Press
.
Frigerio
,
G.
,
Burri
,
M.
,
Bopp
,
D.
,
Baumgartner
,
S.
&
Noll
,
M.
(
1986
).
Structure of the segmentation gene paired and the Drosophila PRD gene set as part of a gene network
.
Cell
47
,
735
746
.
Frohnhôfer
,
H. G.
&
Nüsslein-Volhard
,
C.
(
1986
).
Organization of anterior pattern in the Drosophila embryo by the maternal gene bicoid
.
Nature, Lond.
324
,
120
125
.
Jeffery
,
W. R.
&
Meier
,
S.
(
1983
).
A yellow crescent cytoskeletal domain in ascidian eggs and its role in early development
.
Devi Biol.
96
,
125
143
.
Jeffery
,
W. R.
&
Meier
,
S.
(
1984
).
Ooplasmic segregation of the myoplasmic actin network in stratified ascidian eggs
.
Wilhelm Roux’ Arch, devl Biol.
193
,
257
262
.
Kalthoff
,
K.
(
1971
).
Temperature effects on embryogenesis in Smittia spec. (Diptera, Chironomidae): Q10-values of normal development and frequency of “double abdomens” after UV irradiation
.
Wilhelm Roux’ Arch, devl Biol.
168
,
85
96
.
Kalthoff
,
K.
(
1975
).
Smittia spec. (Diptera) Normale Embryonalentwicklung. Aberration des Segmentmusters nach UV-Bestrahlung. Gottingen: Encyclop Cinematograph Film E 2158/1974
.
Kalthoff
,
K.
(
1983
).
Cytoplasmic determinants in Dipteran eggs
. In
Time, Space, and Pattern in Embryonic Development
(ed.
W. R.
Jeffery
&
R. A.
Raff
), pp.
313
348
.
New York
:
A. R. Liss
.
Kalthoff
,
K.
&
Elbetieha
,
A.
(
1986
).
Transplantation of localized anterior determinants in Chironomus eggs by microinjection
.
J. Embryol. exp. Morph.
97
Suppl
.,
181
196
.
Kalthoff
,
K.
,
Hanel
,
P.
&
Zissler
,
D.
(
1977
).
A morphogenetic determinant in the anterior pole of an insect egg (Smittia spec., Chironomidae, Diptera). Localization by combined centrifugation and UV irradiation
.
Devl Biol.
55
,
285
305
.
Kalthoff
,
K.
,
Rau
,
K.-G.
&
Edmond
,
J. C.
(
1982
).
Modifying effects of ultraviolet irradiation on the development of abnormal body patterns in centrifuged insect embryos (Smittia sp., Chironomidae, Diptera)
.
Devl Biol.
91
,
413
422
.
Kalthoff
,
K.
&
Sander
,
K.
(
1968
).
Der Entwicklungsgang der Missbildung “Doppelabdomen” im partiel! UV-bestrahlten Ei von Smittia parthenogenetic (Diptera, Chironomidae)
.
Wilhelm Roux’ Arch. Entwmech. Org.
161
,
129
146
.
Nüsslein-Volhard
,
C.
(
1977
).
Genetic analysis of pattern formation in the Drosophila melanogaster embryo. Characterization of the maternal effect mutant bicaudal
.
Wilhelm Roux’ Arch, devl Biol.
183
,
249
268
.
Percy
,
J.
,
Kuhn
,
K. L.
&
Kalthoff
,
K.
(
1986
).
Scanning electron microscopic analysis of spontaneous and UV-induced abnormal segment patterns in Chironomus samoensis (Diptera, Chironomidae)
.
Wilhelm Roux’ Arch, devl Biol.
195
,
92
102
.
Ripley
,
S.
&
Kalthoff
,
K.
(
1981
).
Double abdomen induction with low UV-doses in Smittia sp. (Chironomidae, Diptera): Sensitive period and complete photoreversibility
.
Wilhelm Roux’ Arch, devl Biol.
190
,
49
54
.
Ripley
,
S.
&
Kalthoff
,
K.
(
1983
).
Changes in the apparent localization of an anterior determinant during early embryogenesis (Smittia sp., Chironomidae, Diptera)
.
Wdhelm Roux’ Arch, devl Biol.
192
,
353
361
.
Schmidt
,
O.
(
1974
).
Der Einfluss mitochondrialer Staffwechselinhibitoren auf frühembryonale Entwicklungsprozesse im Ei von Smittia spec
.
Staatsexamensarbeit Biol. Inst. I der Univ. Freiburg
.
Schmidt
,
O.
,
Zissler
,
D.
,
Sander
,
K.
&
Kalthoff
,
K.
(
1975
).
Switch in pattern formation after puncturing the anterior pole of Smittia eggs (Chironomidae, Diptera)
.
Devl Biol.
46
,
216
221
.
Yajima
,
H.
(
1960
).
Studies on embryonic determination of the harlequin-fly, Chironomus dorsalis. I. Effects of centrifugation and of the combination with constriction and puncturing
.
J. Embryol. exp. Morph.
8
,
198
215
.
Yajima
,
H.
(
1964
).
Studies on embryonic determination of the harlequin-fly, Chironomus dorsalis. IL Effects of partial irradiation of the egg by UV light
.
J. Embryol. exp. Morph.
12
,
89
100
.
Yajima
,
H.
(
1983
).
Production of longitudinal double formation by centrifugation or by UV-partial irradiation in the Chironomus samoensis egg
.
Entomol. Gener.
8
,
171
191
.