A heat shock (of 15 min at 48°C) given to early embryos of the locust, Schistocerca gregaria, results in localized abnormalities in the segment pattern subsequently formed. Most defects involve two consecutive segments of the thorax or abdomen, and these are analysed in detail. The abdominal defects fall into three main classes each of which involves the absence of a particular region of the segment pair and, in one class, duplication of the region which remains. The thoracic defects similarly involve absence of parts of the segments and the formation of a single limb base from which one, two, or three limbs develop.

Heat shock may result in the absence of parts of segments in two distinct ways. It may interfere with the process of segmentation or it may delete parts of already formed segment primordia. These possibilities are discussed although, at present, neither can be excluded.

The duplication observed in some abdominal disruptions and the formation of triple limbs indicates that the absence of parts of embryonic segments is followed by pattern regulation similar to that occurring in regeneration studies on larval segments and appendages of other insects. Two out of the three classes of abnormality can be explained in terms of intercalary regeneration restoring pattern continuity, but it is possible that discontinuities persist in the remaining class.

Early in insect development the ventral part of the blastoderm cell layer, the ‘germ anlage’, becomes divided into the head, thoracic and abdominal segments of the embryo. This process of segmentation has been most thoroughly studied in Drosophila, a ‘long germ’ insect. The results of ligation (e.g. Schubiger & Wood, 1977), cell transplantation (e.g. Simcox & Sang, 1983) and clonal analysis (e.g. Wieschaus & Gehring, 1976) show that the segment primordia become determined around cellular blastoderm stage (at approximately 3|h). This process may be related to the further restriction of cells to anterior or posterior compartments which occurs at about the same time. Segment primordia seem to be evenly spaced on the blastoderm, each about three to four cells in anterior-posterior length (Lohs-Schardin, Cremer & Nusslein-Volhard, 1979; Technau & Campos-Ortega, 1985). The segmental organization of the embryo becomes visible later, at approximately 5 h (see Martinez-Arias & Lawrence, 1985).

In ‘short germ’ insects such as the locust, the embryo forms from a small region of the blastoderm, and the segments become visible sequentially during the elongation of the posterior part of the germ anlage. At present there are no estimates of the size of the segment primordia and it is not clear when they become determined. The results of ligation studies indicate that segmentation may occur in an anterior-to-posterior sequence just in advance of the visible appearance of segments, although it is also possible that segmentation could occur much earlier (Mee, 1986). It seems likely that the mechanism of segmentation in short germ insects depends on growth of the germ anlage, and therefore differs fundamentally from that in long germ insects, but there is no unequivocal evidence that this is the case (see Sander, 1976).

Although segments are determined early in embryogenesis, the pattern of cuticular structures subsequently formed depends upon cellular interaction within the segment. These interactions have been studied by damage experiments on embryos and, particularly, by excision and grafting experiments on the abdominal segments of larval insects (reviewed Lawrence, 1981). Many grafting experiments show that the epidermal cells have stable ‘positional values’ (Wolpert, 1971) related to their position on the anterior-posterior axis of the segment. Positional values form a regular sequence (or a gradient) which is repeated in each segment and if cells with different values are confronted they will interact to restore pattern continuity.

Following the excision of a strip of epidermis from a segment, Wright & Lawrence (1981) showed that the pattern regenerated is precisely related to the position of the cells confronted by the operation. The original pattern was reformed following the removal of a half-segment length or less, but replaced by other structures in reversed orientation when more than half a segment length was excised. When a piece equivalent to a segment length was excised across a border, so confronting cells of equivalent position from the two segments, the remaining regions healed forming a chimaeric segment. These patterns of regeneration were explained by intercalary regeneration via the ‘shortest route’ (see French, Bryant & Bryant, 1976) whereby discontinuities are removed by inserting cells with the shortest possible sequence of intervening positional values.

The segment border was regenerated when excised as part of a narrow strip, and an extra segment border was formed ectopically when a broad strip was removed from the middle of a segment. Hence, the segment border can be regarded as part of the repeating pattern, rather than a special boundary responsible for maintaining the gradient within the larval segment.

As a means of studying segmentation in short germ insects we have heat-shocked locust embryos at different stages. In the preceding paper (Mee & French, 1986) we show that segmental abnormalities are induced in a predictable location, two or three segments posterior to the segments visible at the time of heat shock. Here the structure of the abnormal segments is analysed and discussed in relation to the mechanism of segmentation and to pattern regulation within embryonic segments.

Disruptions of the segment pattern were generated following a 15 min heat shock of 48 °C delivered at stages prior to and during visible segmentation of the embryo of the locust, Schistocerca gregaria (see Mee & French, 1986). Usually the segmental abnormalities were examined in pigmented first instar hoppers. However, some animals were fixed as hatchlings (prior to shedding the embryonic cuticle) or as late embryos which were dissected out of the egg after fixation. Abnormal segments were examined in the animal or as whole-mount preparations of cuticle cut from fixed animals digested with 10 % KOH (at 60 °C for 2 to 3 h), dehydrated and mounted in Euparol.

(A) Disruptions of the abdomen

(1) The normal abdomen

The cuticular pattern characteristic of most abdominal segments of the first instar hopper is shown in Fig. 1. Some segments differ slightly in structure; the first abdominal segment (Al) is not distinct from the thorax ventrally and bears a pair of tympanal organs and a pair of pleuropodia; A8 (in females only) and A9 bear rudimentary genitalia and A10 and All are fused and bear the cerci. The features of the segment pattern examined in heat-shocked animals were the extent of the pigmented regions of the tergite and sternite, the extent of the ‘intersegmental membrane’ (ism), and the polarity and location of bristles. The pleura was not examined for abnormalities as it is relatively featureless and highly folded.

Fig. 1.

The cuticular pattern of abdominal segments. Camera lucida drawing of the circumference of two segments, cut along the pleura on the left side, opened out and mounted flat. The dorsal sclerotized tergite is mostly darkly pigmented (stippling) with a narrow unpigmented stripe at the dorsal midline (D) and also two irregular unpigmented bands (bl and bit) laterally. Bristles are absent from the anterior margin of the tergite (ter) and elsewhere they are sparsely distributed and directed posteriorly. The ventral sclerotized sternite (ster) consists of two lightly pigmented patches separated by a wide upigmented band at the ventral midline (V). Bristles occur in pigmented areas (except at the anterior margin) and point posteriorly. The lateral unsclerotized pleura and the ‘intersegmental membrane’ (ism) are unpigmented and the pleura contains the spiracles (sp). The border between ism and sternite is indistinct. In this preparation the anterior-posterior extent of the ism is reduced slightly by folds and creases. In the intact animal the posterior region of each segment overlaps the anterior region of the next (posterior) segment and the ism is folded away.

Fig. 1.

The cuticular pattern of abdominal segments. Camera lucida drawing of the circumference of two segments, cut along the pleura on the left side, opened out and mounted flat. The dorsal sclerotized tergite is mostly darkly pigmented (stippling) with a narrow unpigmented stripe at the dorsal midline (D) and also two irregular unpigmented bands (bl and bit) laterally. Bristles are absent from the anterior margin of the tergite (ter) and elsewhere they are sparsely distributed and directed posteriorly. The ventral sclerotized sternite (ster) consists of two lightly pigmented patches separated by a wide upigmented band at the ventral midline (V). Bristles occur in pigmented areas (except at the anterior margin) and point posteriorly. The lateral unsclerotized pleura and the ‘intersegmental membrane’ (ism) are unpigmented and the pleura contains the spiracles (sp). The border between ism and sternite is indistinct. In this preparation the anterior-posterior extent of the ism is reduced slightly by folds and creases. In the intact animal the posterior region of each segment overlaps the anterior region of the next (posterior) segment and the ism is folded away.

(2) The location of defects

The abdominal disruptions resulting from heat shock most frequently involved a pair of consecutive segments (Mee & French, 1986) and affected all or only part of their circumference. For convenience, the segment circumference was divided into quadrants, consisting of the two hemisternites and two hemitergites, and two-segment defects are defined as those in which one or more quadrants are affected in both segments. Many (31 %) of the disruptions involved all four quadrants of both segments. Slightly less common were defects affecting one lateral half of the animal (23%) or the whole ventral surface (23%). Defects affecting the whole dorsal surface, one hemitergite, one hemisternite, a hemitergite and the contralateral hemisternite, or three quadrants of the circumference were rare (5 % or less).

(3) The structure of two-segment defects

The abdominal defects analysed in detail resulted from heat shock between formation of the germ anlage and the appearance of the 7th abdominal segment (sA7 stage) and all involve two consecutive segments. Defects involving segments A10 and All were excluded as these segments are fused and difficult to score accurately. The defects are described with reference to a pair of consecutive hemitergites or hemisternites (and the intervening ism) and can be grouped into four classes.

Class 1 (loss of a segment)

This abnormality appeared to involve the absence of one of the hemitergites or hemisternites of the pair; the remaining structures were normal (Fig. 2). However, it was not clear whether (i) one of the hemitergites or hemisternites and the ism were absent or (ii) the posterior of one hemitergite or hemisternite, the ism, and the anterior of the next hemisegment were missing, with the remaining parts fused to form a single, chimaeric, segment (see thoracic defects below).

Fig. 2.

Abdominal segment disruption -class 1. Camera lucida drawings of dorsal (i) and ventral (ii) surfaces of an animal with a segment missing on the left side. Dorsally, the single segment gradually increases in anterior-posterior length towards the midline.

Fig. 2.

Abdominal segment disruption -class 1. Camera lucida drawings of dorsal (i) and ventral (ii) surfaces of an animal with a segment missing on the left side. Dorsally, the single segment gradually increases in anterior-posterior length towards the midline.

Class 2 (loss of the ism)

A dorsal class 2 defect was characterized by the absence of ism for some distance along the transverse axis and the fusion of consecutive hemitergites. The total anterior-posterior length of the hemitergites, at the site of fusion, was usually less than twice the normal tergite length, and all bristles appeared to be normally orientated (Fig. 3). The defect was usually restricted to the midline region, but occasionally the ism was absent in lateral regions or across the whole hemisegment. In a ventral class 2 defect there was a disturbance in the normal overlap between consecutive hemisternites in the vicinity of the midline, suggesting loss of ism in this region (Fig. 3). It was usually not clear whether the ism was completely absent because the border between ism and sternite is not distinct and consecutive pigmented regions were not fused. In a few cases however, the ism was clearly missing between the hemisternites, and there was a single enlarged pigmented region with bristles in normal orientation (Fig. 3B).

Fig. 3.

Abdominal segment disruptions -class 2.

(A) Dorsally (i), the ism is absent and the hemitergites are fused in the vicinity of the midline (see arrows). Bristle polarity is normal and the total length of the fused tergites is similar to that of two tergites. Ventrally (ii), the sternites do not overlap normally at the midline and appear to be fused (arrow), although the pigment patches are separate.

(B) An animal with class 2 defects on the right. Dorsally (i), the ism is absent and the hemitergites fused from bll to the midline (arrows). The total length of the hemitergites is reduced and bristle orientation is normal. Ventrally (ii), there is a single enlarged patch of pigmentation bearing bristles with normal orientation (arrow).

Fig. 3.

Abdominal segment disruptions -class 2.

(A) Dorsally (i), the ism is absent and the hemitergites are fused in the vicinity of the midline (see arrows). Bristle polarity is normal and the total length of the fused tergites is similar to that of two tergites. Ventrally (ii), the sternites do not overlap normally at the midline and appear to be fused (arrow), although the pigment patches are separate.

(B) An animal with class 2 defects on the right. Dorsally (i), the ism is absent and the hemitergites fused from bll to the midline (arrows). The total length of the hemitergites is reduced and bristle orientation is normal. Ventrally (ii), there is a single enlarged patch of pigmentation bearing bristles with normal orientation (arrow).

Class 3 (loss of part of the ism, polarity reversal)

In class 3 defects the anterior-posterior extent of the ism and the two hemitergites or hemisternites (particularly that posterior to the ism) was reduced. In the posterior segment, bristles of reversed or abnormal orientation were located near the anterior margin, in a region normally devoid of bristles (Fig. 4). Dorsal class 3 disruptions usually had all of these characteristics, while ventral class 3 defects often had only a reduction in the size of the pigmented region of the hemisternites. As in class 2 defects, it was difficult to assess the extent to which the ism was reduced in length ventrally but occasionally it was completely absent and there was a single large patch of pigment, with a marked transverse ridge bearing bristles in reversed orientation (Fig. 4B).

Fig. 4.

Abdominal segment disruptions -class 3.

(A) Dorsally (i), the ism is reduced in length in some places and totally missing elsewhere. The hemitergites of the posterior segment are reduced in length and bear a reversed bristle (arrow). Ventrally (ii), there is no overlap between sternites and the posterior hemisternites are small and bear a bristle with abnormal orientation (arrow).

(B) An animal with a class 3 defect on the left side. The dorsal pattern (i) is normal at the midline but laterally there is no overlap between the segments, and bristles with reversed or abnormal polarity (arrows) occur along the anterior margin of the posterior hemitergite. On the ventral surface (ii) the ism is absent (except laterally) and the two hemisternites are fused to give a single large patch of pigmentation with a marked ridge bearing bristles with reversed polarity (arrows).

Fig. 4.

Abdominal segment disruptions -class 3.

(A) Dorsally (i), the ism is reduced in length in some places and totally missing elsewhere. The hemitergites of the posterior segment are reduced in length and bear a reversed bristle (arrow). Ventrally (ii), there is no overlap between sternites and the posterior hemisternites are small and bear a bristle with abnormal orientation (arrow).

(B) An animal with a class 3 defect on the left side. The dorsal pattern (i) is normal at the midline but laterally there is no overlap between the segments, and bristles with reversed or abnormal polarity (arrows) occur along the anterior margin of the posterior hemitergite. On the ventral surface (ii) the ism is absent (except laterally) and the two hemisternites are fused to give a single large patch of pigmentation with a marked ridge bearing bristles with reversed polarity (arrows).

Class 4 (defects at the dorsal midline)

In most class 4 disruptions right and left hemitergites were incorrectly aligned so that noncorresponding segments were fused at the dorsal midline (Fig. 5). In a few cases the hemitergites of a single segment did not extend to the midline to form a normal tergite.

Fig. 5.

Abdominal segment disruption -class 4. In this example, showing dorsal mismatch, both hemitergites on the left are fused with a single hemitergite on the right (arrowed). The posterior right hemitergite is isolated.

Fig. 5.

Abdominal segment disruption -class 4. In this example, showing dorsal mismatch, both hemitergites on the left are fused with a single hemitergite on the right (arrowed). The posterior right hemitergite is isolated.

Although most abdominal segment abnormalities can be classified in this way, the classes may not be distinct. For example, in class 1 defects the single ‘abnormal’ sclerite may be of greater than normal length at the midline where it is fused with the two normal sclerites from the opposite side of the animal (Fig. 2), while class 2 defects may have a fused sclerite of less than twice the normal length (Fig. 3B). Also, some class 3 defects resemble class 1 or 2 in that the ism remaining occurs intermittently and is absent in some regions (Fig. 4B).

Table 1 shows the frequency of each of these classes of disruption dorsally and ventrally. Frequently hemisegments have defects ventrally and dorsally, although an abnormal ventral pattern is often associated with a normal dorsal pattern. In general, class 2 defects are associated with normal patterns, and class 1 and 3 defects with a similar defect.

Table 1.

The occurrence of abdominal segment defects (classes 1–3) on each hemisegment of 660 out of 678 experimental animals with two-segment defects

The occurrence of abdominal segment defects (classes 1–3) on each hemisegment of 660 out of 678 experimental animals with two-segment defects
The occurrence of abdominal segment defects (classes 1–3) on each hemisegment of 660 out of 678 experimental animals with two-segment defects

(4) Disruptions involving more than two segments

Extensive defects usually involved a number of consecutive segments (see Mee & French, 1986) but sometimes there were two regions of disruption separated by a normal segment border (Fig. 6) or one or more normal segments. Within a disrupted area, defects between adjacent segments could be classified in the same way as the two segment defects (Fig. 6).

Fig. 6.

Disruptions affecting more than two segments. Dorsal (i) and ventral (ii) surfaces of an animal with a disruption affecting five segments (7 to V). There are two regions of disruption, of two and three segments, separated by a normal segment border. Defects affecting pairs of consecutive segments are labelled class 1 to class 3 (1-3) and reversed bristles are indicated by arrows.

Fig. 6.

Disruptions affecting more than two segments. Dorsal (i) and ventral (ii) surfaces of an animal with a disruption affecting five segments (7 to V). There are two regions of disruption, of two and three segments, separated by a normal segment border. Defects affecting pairs of consecutive segments are labelled class 1 to class 3 (1-3) and reversed bristles are indicated by arrows.

(B) Disruptions of the thorax

(1) The normal thorax

The structure of the thoracic segments and their appendages is shown in Fig. 7. Dorsally, the cuticular pattern is similar to that of the abdomen but ventrally the segments are not clearly delineated. The metathoracic leg has a double row of spines on the lateral face of the tibia but, on the pro- and mesothoracic legs, the spines are smaller and located on the medial face of the tibia. The four spurs present on the distal tibia were also larger on the metathoracic leg.

Fig. 7.

The normal thorax of the first instar hopper. (A) Lateral view of the thorax. The prothorax is covered by the saddle-like pronotum (T1) which extends laterally and partly covers the mesonotum (T2). The lower posterior angles of meso- and metano turn (T3) give rise to the wing buds. The pleura of each segment bears an anterior and a posterior sclerotized plate (1 and 2). T3s and Als are the spiracles of the metathoracic and first abdominal segments. (B) Anterior views of (i) prothoracic and (ii) metathoracic left legs, each with an enlarged medial view of the tarsus, illustrating the segments of the limb -coxa (co), trochanter (fr), femur (fe), tibia (ft) and tarsus (ta); and the circumferential markers -the two rows of tibial spines (s), four tibial spurs (sp), pairs of tarsal pads (tap) and the two claws (cl). The mesothoracic limb is similar to the prothoracic. The faces of the limb are labelled medial (M), anterior (A), lateral (L) and posterior (P).

Fig. 7.

The normal thorax of the first instar hopper. (A) Lateral view of the thorax. The prothorax is covered by the saddle-like pronotum (T1) which extends laterally and partly covers the mesonotum (T2). The lower posterior angles of meso- and metano turn (T3) give rise to the wing buds. The pleura of each segment bears an anterior and a posterior sclerotized plate (1 and 2). T3s and Als are the spiracles of the metathoracic and first abdominal segments. (B) Anterior views of (i) prothoracic and (ii) metathoracic left legs, each with an enlarged medial view of the tarsus, illustrating the segments of the limb -coxa (co), trochanter (fr), femur (fe), tibia (ft) and tarsus (ta); and the circumferential markers -the two rows of tibial spines (s), four tibial spurs (sp), pairs of tarsal pads (tap) and the two claws (cl). The mesothoracic limb is similar to the prothoracic. The faces of the limb are labelled medial (M), anterior (A), lateral (L) and posterior (P).

In heat-shocked animals the dorsal cuticular pattern was examined and scored in the same way as that of the abdomen. Limb structures were scored for the presence of the leg segments and the cuticular markers illustrated in Fig. 7B. The segmental composition of limbs associated with meso- and metathoracic segment defects could often be determined from the location of the tibial spines.

(2) Thoracic defects

The thoracic disruptions analysed in this section resulted from heat shock prior to the sA7 stage and involve pairs of segments and their appendages. There were 151 scoreable disruptions (right and left sides scored separately); 4 involving the labium (G3) and T1; 61, segments T1 and T2; 44, segments T2 and T3 and 42, segments T3 and Al.

In 34 other animals thoracic disruptions involved the dorsal or ventral surface but not the limbs. Disruptions involving the limb of a single segment (these included minor pattern defects and also occasional cases of partial duplication or truncation of the limb) and disruptions involving three or more segments are not described here.

(3) Abnormal limb patterns

In disruptions affecting a pair of thoracic segments there was a single leg base, usually associated with two pleural plates (but occasionally three or four) and loss of the metathoracic spiracle in defects involving segments T2 and T3. These defects therefore involved loss of at least the regions of the thorax normally separating successive leg bases. The limbs had between one and three complete sets of limb structures and usually separated into a corresponding number of branches.

The number of sets of pattern elements could only be scored in the tibia (two rows of spines, four spurs per set), the tarsus (two rows of tarsal pads per set) and at the distal tip (two claws). The total number of sets present on all branches often remained constant from the proximal to the distal end of the limb (Figs 8, 9), but sometimes the number increased (the pattern diverged -see Fig. 10) and occasionally it decreased (the pattern converged).

Fig. 8.

A class LI limb resulting from a disruption of segments T1 and T2. The leg has a slightly enlarged coxa (co) but is otherwise normal. Dorsally, there is a single tergite (T2/T1) similar to the pronotum.

Fig. 8.

A class LI limb resulting from a disruption of segments T1 and T2. The leg has a slightly enlarged coxa (co) but is otherwise normal. Dorsally, there is a single tergite (T2/T1) similar to the pronotum.

Fig. 9.

Class L2 and class L3 limbs, both involving segments T1 and T2. (A) shows a limb which branches at the proximal femur (fe), to give two normal distally complete branches (class L2). The common limb base, coxa (co) and trochanter, are enlarged. (B) shows a class L3 limb with three distally complete branches each with a set of two claws (cl). The limb branches at the femur/tibia joint and the anterior branch divides again at the proximal tibia to give a normal anterior limb and a middle branch which is distally complete but lacks the tibia. The common leg base, coxa, trochanter and femur are considerably enlarged.

Fig. 9.

Class L2 and class L3 limbs, both involving segments T1 and T2. (A) shows a limb which branches at the proximal femur (fe), to give two normal distally complete branches (class L2). The common limb base, coxa (co) and trochanter, are enlarged. (B) shows a class L3 limb with three distally complete branches each with a set of two claws (cl). The limb branches at the femur/tibia joint and the anterior branch divides again at the proximal tibia to give a normal anterior limb and a middle branch which is distally complete but lacks the tibia. The common leg base, coxa, trochanter and femur are considerably enlarged.

Fig. 10.

A class L3 limb, involving segments T1 and T2, which was fused for most of its proximal-distal length. The coxa, femur and tibia are enlarged and the limb branches in the tarsus (ta) to give two complete branches and a branch which is probably damaged. The pattern diverges; the tibia has two sets of pattern elements -four rows of medial spines (s) and eight spurs (sp -seven shown); proximally, the tarsus has two and a half sets -five rows of tarsal pads (tap), and one branch with three rows of pads terminates with two complete sets of claws (cl).

Fig. 10.

A class L3 limb, involving segments T1 and T2, which was fused for most of its proximal-distal length. The coxa, femur and tibia are enlarged and the limb branches in the tarsus (ta) to give two complete branches and a branch which is probably damaged. The pattern diverges; the tibia has two sets of pattern elements -four rows of medial spines (s) and eight spurs (sp -seven shown); proximally, the tarsus has two and a half sets -five rows of tarsal pads (tap), and one branch with three rows of pads terminates with two complete sets of claws (cl).

Limbs were classified according to the number of sets present at the distal tip of the limb (all branches). If two claws were present the set was scored as ‘distally complete’, while sets with a single claw, no claws, or missing distal segments were scored as ‘incomplete’. The segmentation of distally complete limbs was usually normal, but the tibia was sometimes absent or much reduced in length.

The limbs fell into three classes (Table 2).

Table 2.

Abnormal limb patterns following fusion of segments T1 with T2, and T2 with T3

Abnormal limb patterns following fusion of segments T1 with T2, and T2 with T3
Abnormal limb patterns following fusion of segments T1 with T2, and T2 with T3

Class LI

Class LI limbs often appeared to be completely normal (Fig. 8), with one complete set of pattern elements, but in other cases had an enlarged circumference or bore a spike on the appendage or pleural plates. The limb was orientated normally with respect to the body.

Class L2 and L3

As shown in Table 2, these limbs had, at most, two (class L2) or three (class L3) distally complete sets of pattern elements. Of these 73 limbs, 47 branched proximally, at a level between the leg base and the distal end of the femur (Fig. 9) and the majority of the remaining limbs branched within the tarsus (Fig. 10). The class L3 limbs usually branched proximally into one normal limb and a second branch, which often branched again at a more distal level. A few class L2 and L3 limbs did not branch (Fig. 11 A) and some class L3 limbs only branched once (Fig. 10) so that the tip of a limb could bear up to six claws. Many of the limb branches were fused or twisted so that it was difficult to determine their orientation, but in scoreable cases the medial-lateral axes of all limb branches were approximately parallel and orientated normally with respect to the body.

Fig. 11.

Segmental origin of limbs in T2/T3 defects. (A) A class L3 limb with a single row of tibial spines (5) and four spurs (sp) typical of T3. The limb terminates with three pairs of claws (cl). (B) A class L3 limb which branches proximally to give a normal, distally complete, T2 branch and posteriorly, a branch with an enlarged tibia (ti) bearing three distinct rows of T3 spines and eight T3 spurs (two views of the tibia and tarsi are shown). A further short row of smaller spines is present (arrow) and these may be T2 structures. (C) A class L2 limb with a composite tibia (ti) bearing two rows of spines (s) and three spurs (sp) typical of T3, and two rows of spines (5) and three spurs (sp) typical of T2. (D) A class L2 limb branching in the proximal tibia to give one complete branch, with spines and spurs characteristic of T3, and another complete, but composite, branch with a lateral row of T3 spines and a medial row of T2 spines.

Fig. 11.

Segmental origin of limbs in T2/T3 defects. (A) A class L3 limb with a single row of tibial spines (5) and four spurs (sp) typical of T3. The limb terminates with three pairs of claws (cl). (B) A class L3 limb which branches proximally to give a normal, distally complete, T2 branch and posteriorly, a branch with an enlarged tibia (ti) bearing three distinct rows of T3 spines and eight T3 spurs (two views of the tibia and tarsi are shown). A further short row of smaller spines is present (arrow) and these may be T2 structures. (C) A class L2 limb with a composite tibia (ti) bearing two rows of spines (s) and three spurs (sp) typical of T3, and two rows of spines (5) and three spurs (sp) typical of T2. (D) A class L2 limb branching in the proximal tibia to give one complete branch, with spines and spurs characteristic of T3, and another complete, but composite, branch with a lateral row of T3 spines and a medial row of T2 spines.

(4) The segmental origin of limbs

The criteria for distinguishing between limbs are confined to the tibia. In pro- and mesothoracic limbs the double row of tibial spines occurs on the medial face; in the metathoracic limb on the lateral face. Metathoracic tibial spines and spurs are also larger than those of the pro- and mesothorax (Fig. 7). Thus, the segmental origin of abnormal limbs could be investigated only where the disruption involved segments T2 and T3 (see Table 2). The following patterns were found.

  • (i) A class LI limb with characteristics of either segment T2 or T3 (5 cases).

  • (ii) Class L2 or L3 limbs with characteristics of either segment T2 or T3 (4 cases). These limbs branched in the tarsus and had only one or two rows of tibial spines and four or five spurs (Fig. 11 A).

  • (iii) Class L2 limbs separating proximally into one T2 and one T3 branch (10 cases).

  • (iv) Class L3 limbs separating proximally into a normal T2 or T3 branch and a second branch bearing more than one set of pattern elements but markers of only T3 or T2 respectively (Fig. 11B; 5 cases).

  • (v) Class LI limb with one row of T2 tibial spines and one row of T3 tibial spines (1 case).

  • (vi) Class L2 or L3 limbs with composite tibiae bearing rows of spines and spurs characteristic in location and size, of both segments T2 and T3 (7 cases; Fig. 11C,D).

In a further 12 class L2 or L3 limbs only one set of cuticular structures could be identified and these were usually T2 in origin.

Thus most (23/32) of the scoreable T2/T3 limbs had structures characteristic of both the T2 and the T3 limbs, although they were present on the same branch of the limb in only 8 cases.

(5) Disruptions of the dorsal thorax

In order to relate the abnormal limbs to the disrupted patterns of segments observed in the abdomen, the accompanying dorsal thorax patterns were examined. Of the 105 abnormal limbs, however, only 26 occurred in pigmented animals in which the dorsal pattern could be scored in detail. The dorsal disruptions were similar to dorsal abdominal defects of class 1 (16 cases) and class 3 (10 cases). Both classes of dorsal defect were found with all types of abnormal limbs, from class LI limbs to class L3 limbs with three distally complete branches.

(6) Disruptions involving segments T3 and Al

Disruptions involving segments T3 and Al were also often accompanied by abnormal appendages (Fig. 12A). However, the pleuropodia of Al lack specific markers and are normally lost on hatching, so the composition of abnormal appendages was difficult to assess. They were usually deformed and swollen proximally but often had two or four tibial spurs and a tarsus ending with one or two claws.

Fig. 12.

(A) Fusion of segments T3 and Al. The example shown was dissected out of the egg as a late embryo and is only partially pigmented. The femur (fe) and tibia (ti) of the abnormal appendage are grossly enlarged and swollen. Distally the limb is normal in size and bears four spurs (sp -two visible here) and a single claw (cl). The T2 leg has been removed. (B) Fusion of segments G3 and Tl. The segmental appendages are fused proximally. The palp (pa) and paraglossa (par) of G3 and the Tl leg are present. Between them there is a supernumerary appendage, limb-like in shape and size, with limb segments corresponding to the trochanter, femur and tibia. The tibia bears three spurs (not visible here) and the limb terminates with a single claw.

Fig. 12.

(A) Fusion of segments T3 and Al. The example shown was dissected out of the egg as a late embryo and is only partially pigmented. The femur (fe) and tibia (ti) of the abnormal appendage are grossly enlarged and swollen. Distally the limb is normal in size and bears four spurs (sp -two visible here) and a single claw (cl). The T2 leg has been removed. (B) Fusion of segments G3 and Tl. The segmental appendages are fused proximally. The palp (pa) and paraglossa (par) of G3 and the Tl leg are present. Between them there is a supernumerary appendage, limb-like in shape and size, with limb segments corresponding to the trochanter, femur and tibia. The tibia bears three spurs (not visible here) and the limb terminates with a single claw.

(7) Disruptions involving segments G3 and T1

In four animals the posterior head segment (labium) and the prothorax were disrupted and formed two or three appendages fused at the base. The three appendages (2/4 cases) were an anterior labium, a posterior prothoracic leg and, between them, a short swollen appendage with two or three tibial spurs and one tarsal claw (Fig. 12B). When only two appendages were formed they were either the swollen appendage (with one claw) and the leg, or the labium plus a malformed leg (with no spurs and a single claw).

(C) Disruptions of the head

Individual head segments are not clearly defined (Fig. 13A) and in heat-shocked animals only the labrum, eyes and appendages could be scored. Disruptions of the eyes, antennae and labrum often involved either duplication or fusion and loss of structures (Fig. 13C,D). In the gnathos, disruptions included absence of appendages and, or, fusion of the appendages of adjacent segments (Fig. 13B) and sometimes duplication of appendages was observed.

Fig. 13.

Disruptions of the head. Heads of late embryos are shown in anterior view. Only the eyes and the teeth of the mandibles, and laciniae were pigmented (indicated by stippling). (A) Normal head showing the eyes, labrum (lab), antennae (ant) and mouthparts: mandibles (G1), the lacinia (la), galea (ga) and palp (pa) of the maxillary segment (G2) and the paraglossa (par) and palp (pa) of the labial segment (G3). (B) Fusion of appendages. On the left of the animal the appendages of segments G1 and G2 were fused proximally (bracket) and the antenna was absent. (C) Duplication of pattern elements. This animal bore one extra eye dorsally (not visible) and one extra antenna. (D) Fusion and loss of pattern elements. The eyes and antennae were fused and the labrum absent.

Fig. 13.

Disruptions of the head. Heads of late embryos are shown in anterior view. Only the eyes and the teeth of the mandibles, and laciniae were pigmented (indicated by stippling). (A) Normal head showing the eyes, labrum (lab), antennae (ant) and mouthparts: mandibles (G1), the lacinia (la), galea (ga) and palp (pa) of the maxillary segment (G2) and the paraglossa (par) and palp (pa) of the labial segment (G3). (B) Fusion of appendages. On the left of the animal the appendages of segments G1 and G2 were fused proximally (bracket) and the antenna was absent. (C) Duplication of pattern elements. This animal bore one extra eye dorsally (not visible) and one extra antenna. (D) Fusion and loss of pattern elements. The eyes and antennae were fused and the labrum absent.

Dorsally, the head and thorax (and sometimes anterior abdomen) were occasionally fused in the vicinity of the midline, but it was not possible to determine which head segments were involved.

A brief heat shock to the locust embryo at stages just before and during the anterior-to-posterior progress of visible segmentation frequently produces a discrete disruption in the segment pattern of the hatching larva. The defect usually affects two consecutive segments and is located two or three segments posterior to the last segment visible on the embryo at the time of heat shock (Mee & French, 1986). The abnormal patterns are most readily analysed in the abdomen where the segments are fairly simple in structure and clearly delineated dorsally and ventrally. In the thorax, the presence of more numerous cuticular markers and the morphological differences between T2 and T3 legs allows the segmental composition of the disrupted area to be assessed. Head defects were infrequent (Mee & French, 1986) and only the appendages, labrum and eyes could be readily scored.

Segmental abnormalities

Heat shock results in three main classes of pattern abnormality in the abdomen, all of which appear to involve the absence of regions of the segment;

Class 1 -loss of a segment (Fig. 2). This may correspond to either absence of a single segment (including the ism) or absence of the posterior of one segment, the ism and the anterior of the following segment (see below).

Class 2 -absence of the ism (Fig. 3).

Class 3 -absence of most of the ism plus the anterior part of the following segment (Fig. 4). In class 3 patterns the presence of bristles with reversed orientation suggests that deletion may be followed by pattern regulation (see below).

The remaining defects (class 4) appear to involve abnormalities of the late process of dorsal closure, during which the two edges of the germ band extend around the remains of the yolk and fuse at the dorsal midline.

The dorsal surface of the thoracic segments is very similar to that of the abdomen, and the defects found in the heat-shocked animals are also similar, and involve loss of a segment (class 1) or loss of part of a segment with disturbed polarity (class 3). Both classes of dorsal defect were found with all classes of abnormal limbs.

Differences in leg structure enable the segmental composition of limbs formed in ventral meso-/metathoracic segmental disruptions to be determined. Assuming that thoracic and abdominal segments respond similarly to heat shock, as suggested by the similarities in dorsal defects, the leg structures also indicate the probable composition of the abdominal defects. Conclusions about composition must be tentative, however, since markers are restricted to the tibia and consist only of the size of structures and their lateral or medial position. The majority (23/32, 72 %) of the scoreable appendages formed in meso-/metathoracic defects bore structures characteristic of both segments, suggesting that segmental disruptions usually result from loss of parts of both segments and fusion of the remnants. Of the six scoreable single (class LI) limbs however, only one had structures from both segments, implying that these (and perhaps abdominal class 1) defects often result from the deletion of an entire segment.

The classes of abnormalities induced in the short germ locust embryo following heat shock appear to be similar to the segmental defects induced in long germ dipteran larvae by X-irradiation (Pearson, 1974) or localized microcautery (Bownes, 1976) at approximately blastoderm stage, and by severe heat shock at blastoderm and later stages (Maas, 1949). Polarity cannot readily be assessed in these larvae but Bownes (1976) examined the resulting adults and found deletions associated with mirror-image duplications of the posterior or, occasionally, the anterior parts of the segment. She also found duplicated leg structures (Bownes, 1975). Similar abdominal and limb defects were also found after X-irradiation of intermediate germ cricket embryos at stages before and during visible segmentation (Heinig, 1967). Higher X-ray doses resulted in some animals lacking all or most of the abdomen. Locust embryos completely lacking the anterior or posterior part of the segment pattern were occasionally found after heat shock (Mee, 1984).

Pattern regulation

The abnormal cuticular patterns observed in the locust abdomen following heat shock to the early embryo can be explained in the same way as the results of excision experiments performed on larval segments of other insects (e.g. see Wright & Lawrence, 1981). However, it is not necessarily the case that heat shock alters a pattern which is already laid down (e.g. removing tissue by causing cell death). It is possible that the process of segmentation may be affected directly (e.g. by inappropriate specification of some positional values), resulting in the formation of segments which lack some positional values and consequently undergo intercalary regeneration to form the observed patterns.

Class 1 abdominal defects could result from the absence of one segment length of tissue, either corresponding exactly to one morphological segment or lying across the segment border. Cells of equivalent position would be confronted and the pattern would be stable. Absence of one to one and a half segments would confront cells with slightly different positional values and the resulting intercalary regeneration would give rise to a class 1 defect (Fig. 14B). The pattern would be restored if less than half a segment were absent.

Fig. 14.

Interpretation of the formation of cuticular pattern in two abdominal segments in normal (A) and heat shocked (B-E) animals. (A) The positional values (pv) of each segment are represented as a graded series, JO to 0, with the segment border at 0/10. The dorsal cuticular pattern corresponding to the positional values is given below and bristle orientation shown by arrows. (B) In the absence of slightly more than one segment length of tissue, cell with values 5 and 3 are confronted and intercalary regeneration inserts value 4, restoring continuity and producing a normal-sized composite segment, as in a class 1 defect. (C-E) If rather more than half a segment is missing, intercalary regeneration inserts the shorter set of intervening values in reversed order. The precise pattern formed depends on the location of the defect. Confrontation of cells with values 2 and 6 (C) leads to a mirror-image duplication of posterior tergite (5, 4) and ism (3), as in a class 3 defect. Confrontation of positions 3 and 7 leads to a similar duplication, but complete absence of ism, and confrontation of 6 and 10 leads to a duplication of the anterior tergite (7, 8, 9) which bears few bristles and might be scored as class 2 (absence of ism, no polarity reversal).

Fig. 14.

Interpretation of the formation of cuticular pattern in two abdominal segments in normal (A) and heat shocked (B-E) animals. (A) The positional values (pv) of each segment are represented as a graded series, JO to 0, with the segment border at 0/10. The dorsal cuticular pattern corresponding to the positional values is given below and bristle orientation shown by arrows. (B) In the absence of slightly more than one segment length of tissue, cell with values 5 and 3 are confronted and intercalary regeneration inserts value 4, restoring continuity and producing a normal-sized composite segment, as in a class 1 defect. (C-E) If rather more than half a segment is missing, intercalary regeneration inserts the shorter set of intervening values in reversed order. The precise pattern formed depends on the location of the defect. Confrontation of cells with values 2 and 6 (C) leads to a mirror-image duplication of posterior tergite (5, 4) and ism (3), as in a class 3 defect. Confrontation of positions 3 and 7 leads to a similar duplication, but complete absence of ism, and confrontation of 6 and 10 leads to a duplication of the anterior tergite (7, 8, 9) which bears few bristles and might be scored as class 2 (absence of ism, no polarity reversal).

Class 3 defects may result if half to one segment length of tissue is missing, confronting cells from very different positions. For example, in the absence of most of the ism and anterior margin of the next segment, intercalary regeneration would lead to the duplication of the posterior part of the segment (Fig. 14C). This corresponds to the most frequent form of class 3 defect (Fig. 4A). Absence of this length of tissue from other positions could result in the complete loss of the ism and an enlarged sclerite with a midregion of bristle reversal (Fig. 14D), or in the absence of most or all of the ism and a mirror-image duplication of the anterior region of the sclerite (Fig. 14E). The first of these patterns, but not the second, was found among the heat-shocked animals (Fig. 4B). However, since bristles are found mainly on the posterior part of the sclerite, the pattern shown in Fig. 14E may have been scored as class 2 (with no obvious polarity reversal). Class 2 defects could also result if the loss of the ism is not followed by intercalary regeneration to restore pattern continuity.

The abnormal limbs induced by heat shock can be interpreted in a similar way. In some insects, extirpation of most of the ventral thorax between successive larval legs leads to the formation of an extra leg in reversed anterior-posterior orientation (Bohn, 1974; French & Rowlands, in preparation). These results suggest that the larval leg and the surrounding ventral thorax bear a polar coordinate map of positional values (see French et al. 1976; Bryant, French & Bryant, 1981). If the embryonic limb primordium contains some of these positional values, the absence of thoracic tissue may stimulate intercalary regeneration, resulting in the formation of an abnormal limb base and multiple limbs (Fig. 15).

Fig. 15.

Interpretation of the formation of abnormal limbs after heat shock. (A) The anterior-posterior values (10 to 0, as in Fig. 14) are shown on two thoracic segments (T1 and 72). Each segment has a leg primordium (Ip) with proximal circumferential positional values (1′ to 72′), as in the polar coordinate model (French et al. 1976). During development distalization will complete the primordium, forming proximal-to-distal levels A to E. (B) If more than one segment is absent, intercalary regeneration will form value 4, restoring continuity and completing one leg base which will grow out to form a single class LI leg. (C,D) In the absence of half to one segment, intercalary regeneration will insert the alternative shorter set of values (as in Fig. 14C-E). If the deletion extends into two limb bases (C), values 4, 5,6 are formed, leading to an enlarged leg base (with three copies of values T2′ and 6′) within which intercalary regeneration will lead to divergence into three distal tips (class L3). A similar absence involving part of only one of the primordia (D) will lead to regeneration of values 2,3,4 and the formation of a mirror-image double-half limb base. In the course of distal outgrowth this may lose or gain values at the plane of symmetry, forming either (a) a duplicated limb or (b) an incomplete truncated limb (see Bryant et al. 1981).

Fig. 15.

Interpretation of the formation of abnormal limbs after heat shock. (A) The anterior-posterior values (10 to 0, as in Fig. 14) are shown on two thoracic segments (T1 and 72). Each segment has a leg primordium (Ip) with proximal circumferential positional values (1′ to 72′), as in the polar coordinate model (French et al. 1976). During development distalization will complete the primordium, forming proximal-to-distal levels A to E. (B) If more than one segment is absent, intercalary regeneration will form value 4, restoring continuity and completing one leg base which will grow out to form a single class LI leg. (C,D) In the absence of half to one segment, intercalary regeneration will insert the alternative shorter set of values (as in Fig. 14C-E). If the deletion extends into two limb bases (C), values 4, 5,6 are formed, leading to an enlarged leg base (with three copies of values T2′ and 6′) within which intercalary regeneration will lead to divergence into three distal tips (class L3). A similar absence involving part of only one of the primordia (D) will lead to regeneration of values 2,3,4 and the formation of a mirror-image double-half limb base. In the course of distal outgrowth this may lose or gain values at the plane of symmetry, forming either (a) a duplicated limb or (b) an incomplete truncated limb (see Bryant et al. 1981).

Intercalary regeneration provoked by the absence of one to one and a half segments would lead to the formation of a normal limb base and a single normal limb (Class LI), as shown in Fig. 15B. The absence of half to one segment would result in duplications of remaining structures, including parts of the limb bases. Such a loss extending across the segment border into both limb primordia would lead to formation of a large fused limb base within which interactions could complete three sets of circumferential positional values forming a triple branched leg, as in class L3 (Fig. 15C). The class L2 limbs are less readily explained. Absence of half to one segment across the segment border, but affecting only one limb primordium, could lead to the formation of a symmetrical double-half base. This could converge to form a truncated limb or diverge to form two distal parts (Fig. 15D). A few limbs like these were observed but were not scored as ‘two-segment’ defects as they only affected one limb base. The absence of one and a half to two segments, however, could result in the complete absence of one limb base plus a diverging symmetrical limb, which would be scored as class L2. This requires the absence of a large and precisely positioned region and therefore class L2 limbs would be expected to occur infrequently. However, class L2 limbs were as frequent as the other classes. An alternative possibility is that they were formed by the fusion of two consecutive limb bases, without intercalary regeneration.

This interpretation of the abnormal limbs induced by heat shock is consistent with the relative positions and mediolateral orientation of the limbs and limb branches. It predicts that the middle branch of a class L3 leg (and perhaps one branch of a class L2 leg) is reversed in the anterior-posterior axis, but this cannot be resolved due to the absence of suitable markers. No strict predictions are made about the composition of the limbs. If intercalary regeneration occurs equally from both sides of the deletion, class LI limbs and the middle branch of class L3 limbs would be expected to be of dual origin. The composition of class L2 legs would depend on their mode of origin. The limbs in T2/T3 defects show, however, that (i) most LI and occasional L2 and L3 limbs have markers of only one segment and (ii) some L3 legs have middle branches with markers of only one segment type. However, as explained above, the available markers can only give an approximate indication of limb origin.

The analysis suggests that pattern regulation is similar in classes 1 and LI and classes 3 and L3. In class 2 and L2 defects it seems that intercalary regeneration may not have occurred and pattern discontinuities persist. It is perhaps relevant that limb buds do not regenerate when amputated at slightly later stages (Whitington & Seifert, 1982) and intercalary regeneration does not occur in limbs or abdominal segments of locust larvae (unpublished results).

Segmentation and heat shock abnormalities

Heat shock given at early embryonic stages results in an absence of regions of the cuticular pattern formed in the late embryo and scored in the first instar hopper. It is not clear by what mechanism heat shock produces the segmental disruptions or whether it acts upon the process of segmentation or by altering an existing normal segment pattern.

In the short germ locust embryo segmentation may occur sequentially, just posterior to the visible segments (see Mee, 1986; Mee & French, 1986) and may be directly affected by the heat shock. In Drosophila, and many other organisms, heat shock evokes a characteristic temporary cellular response, consisting of the transcription and translation of a small number of ‘heat-shock’ genes and the repression of the normal pattern of RNA and protein synthesis (see Mitchell & Lipps, 1978; Ashburner & Bonner, 1979). This response develops in Drosophila only after the blastoderm stage and is correlated with ability to survive the heat shock (Bergh & Arking, 1984). Characteristic bristle abnormalities are induced by heat shocks given at particular times during pupation and these may result from the temporary failure of protein synthesis occurring at critical times (Mitchell & Lipps, 1978). In the locust embryo, heat shock temporarily stops cell division (M. Bate, personal communication) and, if segmentation depends on cell division (as in the Progress Zone Model -see Mee & French, 1986), then heat shock could directly interrupt the formation of segments. If this is the case, the formation of partial segments (as in class 2 and 3 defects) implies that segmentation involves more than the specification of a series of segment borders, between which segmental positional values are subsequently formed. Similarly, because variable amounts of the segments appear to be formed, segmentation cannot just consist of establishing an alternating pattern of anterior and posterior compartments (Kornberg, 1981).

It is possible, however, that segmentation occurs much earlier in the locust embryo (see Mee, 1986) and that heat shock produces defects by removing parts of already formed segment primordia. Segmental abnormalities similar to the locust heat-shock defects follow the localized cell death caused by microcautery of Drosophila embryos at around the time of segmentation (Bownes, 1975, 1976), and defects following severe X-irradiation at later stages have been interpreted in the same way (Postlethwaite & Schneiderman, 1973). The heat-shock-induced disruptions are also similar to the phenotypes of many of the Drosophila ‘segmentation mutants’ (Nusslein-Volhard & Wieschaus, 1980). In the ‘pair-rule’ mutants a segment length is missing from within each pair of segments and the remaining tissue is fused to form a single composite segment, as in locust class 1 and LI defects. In the ‘segment polarity’ mutants each segment has part of the pattern missing and the remainder duplicated, as in locust class 3 and L3 defects. In several of these mutants pattern deletions are associated with cell death occurring in approximately the corresponding regions, well after segmentation (Martinez-Arias, 1985; Ingham, Howard & Ish-Horowicz, 1985). It seems that duplications in the segment polarity mutants may arise through pattern regulation after cell death (Martinez-Arias, 1985).

If the locust heat shock defects are due to cell death within segment primordia, the localization of defects implies that only segments at a particular developmental stage are vulnerable. Class 2 and L2 defects may then result from cell death delayed until a stage when pattern regulation is no longer possible, so discontinuities persist. There is, however, no indication that heat shock does cause cell death in locust embryos (M. Bate, personal communication).

To establish whether heat shock affects the process of segmentation or the already formed segment primordia, more information is required about the time of segmentation in short germ insects and about the effects of heat shock on cell proliferation and cell death.

This work was supported by an SERC studentship to J.E.M. We thank David Wright, Mike Bate and Neil Toussaint for useful comments over the years, and especially we thank Klaus Sander for advice, encouragement, and inspiration.

Ashburner
,
M.
&
Bonner
,
J. J.
(
1979
).
The induction of gene activity in Drosophila by heat shock
.
Cell
17
,
241
254
.
Bergh
,
S.
&
Arking
,
R.
(
1984
).
Developmental profile of the heat shock response in early embryos of Drosophila
.
J. exp. Zool
.
231
,
379
391
.
Bohn
,
H.
(
1974
).
Extent and properties of the regeneration field in the larval legs of cockroaches (Leucophaea maderae). I. Extirpation experiments
.
J. Embryol. exp. Morph
.
31
,
557
572
.
Bownes
,
M.
(
1975
).
Adult deficiencies and duplications of head and thoracic structures resulting from microcautery of blastoderm stage Drosophila embryos
.
J. Embryol.-exp. Morph
.
34
,
33
54
.
Bownes
,
M.
(
1976
).
Larval and adult abdominal defects resulting from microcautery of blastoderm staged Drosophila embryos
.
J. exp. Zool
.
195
,
369
392
.
Bryant
,
S. V.
,
French
,
V.
&
Bryant
,
P. J.
(
1981
).
Distal regeneration and symmetry
.
Science
212
,
993
1002
.
French
,
V.
,
Bryant
,
P. J.
&
Bryant
,
S. V.
(
1976
).
Pattern regulation in epimorphic fields
.
Science
193
,
969
981
.
Heinig
,
S.
(
1967
).
Die Abànderung embryqnaler Differenzierungsprozesse durch totale Rôntgenbestrahlung im Ei von Gryllus domesticas
.
Zool. Jb. (Anat.)
84
,
425
492
.
Kornberg
,
T.
(
1981
).
Compartments in the abdomen of Drosophila and the role of the engrailed locus
.
Devi Biol
.
86
,
363
372
.
Lawrence
,
P. A.
(
1981
).
The cellular basis of segmentation in insects
.
Cell
26
,
3
10
.
Ingham
,
P. W.
,
Howard
,
K. R.
&
Ish-Horowicz
,
D.
(
1985
).
Transcription pattern of the Drosophila segmentation gene hairy
.
Nature, Lond
.
318
,
439
445
.
Lohs-Schardin
,
M.
,
Cremer
,
C.
&
Nusslein-Volhard
,
C.
(
1979
).
A fate map for the larval epidermis of Drosophila melanogaster: Localized cuticle defects following irradiation of the blastoderm with an ultraviolet laser microbeam
.
Devi Biol
.
73
,
239
255
.
Maas
,
H.
(
1949
).
Über die Auslosbarkeit von Temperaturmodifikationen wàhrend der Embryonalentwicklung von Drosophila melanogaster
.
Wilhelm Roux Arch. EntwMech. Org
.
143
,
515
572
.
Martinez-Arias
,
A.
(
1985
).
The development of fused embryos of Drosophila melanogaster
.
J. Embryol. exp. Morph
.
87
,
99
114
.
Martinez-Arias
,
A.
&
Lawrence
,
P.
(
1985
).
Parasegments and compartments in the Drosophila embryo
.
Nature, Lond
.
313
,
639
642
.
Mee
,
J. E.
(
1984
).
The development of segments in locust embryos
.
Ph.D. thesis, University of Edinburgh
.
Mee
,
J. E.
(
1986
).
Pattern formation in fragmented eggs of the short germ insect, Schistocerca gregaria
.
Wilhelm Roux Arch. devlBiol. (submitted)
.
Mee
,
J. E.
&
French
,
V.
(
1986
).
Disruption of segmentation in a short germ insect embryo. I. The location of abnormalities induced by heat shock
.
J. Embryol. exp. Morph
.
96
,
000
000
.
Mitchell
,
H. K.
&
Lipps
,
L. S.
(
1978
).
Heat shock and phenocopy induction in Drosophila
.
Cell
15
,
907
918
.
Nusslein-Volhard
,
C.
&
Wieschaus
,
E.
(
1980
).
Segmentation in Drosophila-. Mutations affecting segment number and polarity
.
Nature, Lond
.
287
,
795
801
.
Pearson
,
M. J.
(
1974
).
The relation between larval and adult abnormalities in the abdominal segmentation of Calliphora erythrocephala (Diptera)
.
J. Embryol. exp. Morph
.
32
,
533
555
.
Postleth Waite
,
J.
&
Schneiderman
,
H.
(
1973
).
Pattern formation in imaginai discs of Drosophila melanogaster after irradiation of embryos and young larvae
.
Devi Biol
.
32
,
345
360
.
Sander
,
K.
(
1976
).
Specification of the basic body pattern in insect embryogenesis
.
Adv. Insect Physiol
.
12
,
125
238
.
Schubiger
,
G.
&
Wood
,
W. J.
(
1977
).
Determination during early embryogenesis in Drosophila melanogaster
.
Am. Zool
.
17
,
565
576
.
Simcox
,
A.
&
Sang
,
J.
(
1983
).
When does determination occur in Drosophila embryos?
Devi Biol
.
97
,
212
221
.
Technau
,
G. M.
&
Campos-Ortega
,
J. A.
(
1985
).
Fate-mapping in wild-type Drosophila melanogaster. II. Injections of horse-radish peroxidase in cells of the early gastrula stage
.
Wilhelm Roux Arch, devl Biol
.
194
,
196
212
.
Whitington
,
P. M.
&
Seifert
,
E.
(
1982
).
Axon growth from limb motor neurons in the locust embryos: the effect of target limb removal on the path taken out of the central nervous system
.
Devi Biol
.
93
,
206
215
.
Wieschaus
,
E.
&
Gehring
,
W.
(
1976
).
Clonal analysis of primordial disc cells in the early embryo of Drosophila melanogaster
.
Devi Biol
.
50
,
249
263
.
Wolpert
,
L.
(
1971
).
Positional information and pattern formation
.
Curr. Top. devl Biol
.
6
,
183
224
.
Wright
,
D. A.
&
Lawrence
,
P. A.
(
1981
).
Regeneration of the segment boundary in Oncopeltus
.
DevlBiol
.
85
,
317
327
.