ABSTRACT
The investigation of the regeneration field of the larval legs of cockroaches (Leucophaea maderae), which commenced in a previous paper with extirpation experiments, has been continued by transplantation experiments. The extirpation experiments showed that there are two regions near the leg which are indispensable for leg regeneration: the basal sclerites, and the membranous region extending behind the leg up to the spiracle of the next segment, called Teg-inducing membrane’ (LIM). The LIM is followed by ‘sclerite-inducing membrane’ (SIM) which, upon contact with sclerites, only allows formation of sclerite structures.
The results of the extirpation experiments have been confirmed by transplantation experiments. When the whole leg including the basal sclerites is removed, no leg regeneration occurs. The regenerative ability can be restored by implantation of part of the basal sclerites (for instance, the trochantin), but leg regeneration takes place only when the implantation area is covered by LIM. When the sclerites are transplanted to a region which is covered by SIM only additional basal sclerites are formed.
Whole sets of basal sclerites have been implanted at different distances behind the uninjured hindleg. Additional legs are regenerated only in the anterior half of the membranous field extending between the hindleg and the first abdominal segment. Thus, there is a distribution of LIM and SIM in the region of the hindleg, similar to that near the midleg.
Whole sets of basal sclerites have been implanted at various sites on the dorsal or ventral surfaces of the abdomen. Legs are formed on both surfaces, but only when the transplanted sclerites contact the intersegmental membranes. This means that the intersegmental membranes of the abdomen also have leg-inducing capacities.
The implantation of a trochantin into a field of LIM is followed by the development of two regenerates -a normal one at the posterior border of the field, and one with reverse anterior-posterior polarity at the anterior border. When the trochantin is transplanted together with the praecoxa in a similar way, only one normal regenerate is formed at the posterior margin of the trochantin. The praecoxa prevents contact of the anterior margin of the trochantin with LIM, and contact of the anterior margin of the praecoxa with LIM does not promote leg regeneration.
INTRODUCTION
In a previous paper on the regeneration field of the legs of cockroaches (Bohn, 1974a) it was shown that leg regeneration is possible only when at least two parts of the leg surroundings are allowed to interact -namely, part of the sclerites of the leg base and part of the membranous area extending behind the coxa of the leg, which was called ‘leg-inducing membrane’ (LIM). A third region, making up the most anterior part of the ventral surface of a thoracic segment, when combined with sclerites does not initiate leg formation but only allows regeneration of sclerites. This membranous region therefore was called ‘sclerite-inducing membrane’ (SIM).
The previous results were obtained by excising various parts of the tissues surrounding the leg; those results are confirmed by transplantation experiments presented in the present paper. Either whole sets of basal sclerites or parts thereof (the trochantin) were transplanted to the desired region. The first series of experiments (series D) attempted to learn whether regenerative ability, which was lost after the extirpation of the midleg and its basal sclerites, could be restored by the implantation of a sclerite. A second series of experiments (E) was undertaken to analyse the regeneration field of the hindleg. Finally, an attempt was made (series F) to find tissues with leg-inducing capacities in the segments of the abdomen, which normally does not bear legs.
MATERIALS AND METHODS
Most experiments have been done with larvae of Leucophaea maderae. Operations were performed on freshly moulted third instar larvae, which were observed after operation for at least three further moults. In experimental series D and F the operations were made in two steps; first, the left midleg, including its basal sclerites, was removed from third instar larvae. After two moults, when it was clear that no legs had been regenerated, sclerites of different type or orientation were transplanted to the site. Thus, the final operation was performed on fifth instar larvae. The age of the animals from which transplants were taken differed according to the desirable size of the transplanted tissues; in most cases third instar larvae were used.
Since experiments of the previous paper will often be referred to in this and the following paper, it seemed desirable to arrange the experimental series of all three papers in a continuous order to avoid confusion. Thus the three experimental series of paper I (A-C) (Bohn 1974a) are continued by the three series (D-F) in this paper and by three more (G-J) in the one which follows (Bohn. 1974a).
EXPERIMENTS AND RESULTS
Series D (Fig. 1, Table 1)
When the left midleg including all basal sclerites and part of the anterior membranous area was removed, a leg was seldom regenerated (Expt. A6, Bohn, 1974a). The absence of regenerative capabilities could be due either to the severe disturbance of the tissues by the large wound area, or to the complete absence of sclerites. To answer this question, part of the basal sclerites, namely the trochantin of the left midleg of another animal, was transplanted into the membranous, experimentally leg-free area between fore- and hindleg. Two positions were selected for the transplants (Fig. 1). In Expt. D1 the trochantin was implanted at the level of the spiracle of the mesothoracic segment -a region which is considered to belong to the SIM (Bohn, 1974a). In Expt. D2, the trochantin was implanted in the middle of the field between the spiracles of the meso- and metathoracic segments, which is expected to be covered by the LIM.
Design of experimental series D. The experiment was performed in two steps. First (1) the left midleg was removed completely. The figure at the left shows a situation where only the distal parts of the leg were removed (cross-hatched area: cut surface). In addition to this, the basal sclerites and part of the anterior membranous area were also cut out (heavily dotted area). After two moults (II), during which time no leg regeneration occurred, the trochantin (heavily dotted) of a left midleg of another animal was implanted in one of two different positions (D1 2, indicated by the heavily dotted outlines) into the membranous area covering the former leg area. The results are seen at the right. When the trochantin is implanted far anteriorly, i.e. near to the spiracle (D1), only basal sclerites are regenerated; a double set of basal sclerites is formed. When the trochantin is implanted more posteriorly, one or two legs are formed. The results confirm that leg formation depends on the presence of part of the basal sclerites and of a suitable membranous region, the ‘leg-inducing membrane’, which normally extends between the posterior border of the coxa and the spiracle of the next segment. It is reached by the transplant only in Expt. D2 (for further explanation see text and Fig. 11). Lightly dotted area, membranous parts; clear area, sclerotized elements, consisting of sternites (in the middle), the legs and their basal parts; longitudinally hatched area, wing anlagen.
Design of experimental series D. The experiment was performed in two steps. First (1) the left midleg was removed completely. The figure at the left shows a situation where only the distal parts of the leg were removed (cross-hatched area: cut surface). In addition to this, the basal sclerites and part of the anterior membranous area were also cut out (heavily dotted area). After two moults (II), during which time no leg regeneration occurred, the trochantin (heavily dotted) of a left midleg of another animal was implanted in one of two different positions (D1 2, indicated by the heavily dotted outlines) into the membranous area covering the former leg area. The results are seen at the right. When the trochantin is implanted far anteriorly, i.e. near to the spiracle (D1), only basal sclerites are regenerated; a double set of basal sclerites is formed. When the trochantin is implanted more posteriorly, one or two legs are formed. The results confirm that leg formation depends on the presence of part of the basal sclerites and of a suitable membranous region, the ‘leg-inducing membrane’, which normally extends between the posterior border of the coxa and the spiracle of the next segment. It is reached by the transplant only in Expt. D2 (for further explanation see text and Fig. 11). Lightly dotted area, membranous parts; clear area, sclerotized elements, consisting of sternites (in the middle), the legs and their basal parts; longitudinally hatched area, wing anlagen.
No legs were regenerated in Expt. D1, but there was regeneration of basal sclerites ; anteriorly, a praecoxa was added, posteriorly the whole set of basal sclerites was reduplicated symmetrically (Fig. 2). Expt. D2 showed three types of results. About half of the animals had structures similar to those resulting from Expt. D1 and no leg regenerates; in the other animals legs had been regenerated. In the latter case, either a complete leg was formed, the basal sclerites also being completed by the formation of a praecoxa and epimeron (Fig. 3), or the completion of the basal sclerites did not take place; instead, a second, symmetrical, trochantin and leg were formed anteriorly (Fig. 4).
Figs. 2-4.Examples of the main results of experimental series D. Only a double set of basal sclerites (p, praecoxa; t, trochantin; broken line, axis of symmetry) is formed after transplantation of the trochantin close to the spiracle (sp) (Expt. D1. When the trochantin is transplanted to a more posterior region (Expt. D2) a complete normal leg (Fig. 3) or two symmetrical legs (Fig. 4) are formed, c, Coxa of the unaffected right midleg.
Figs. 2-4.Examples of the main results of experimental series D. Only a double set of basal sclerites (p, praecoxa; t, trochantin; broken line, axis of symmetry) is formed after transplantation of the trochantin close to the spiracle (sp) (Expt. D1. When the trochantin is transplanted to a more posterior region (Expt. D2) a complete normal leg (Fig. 3) or two symmetrical legs (Fig. 4) are formed, c, Coxa of the unaffected right midleg.
Series E (Fig. 8, Table 1)
The distribution of the two different kinds of membranous tissues (SIM and L1M) covering the ventral surface of the mesothoracic segment, as determined by extirpation experiments (series A-C) was shown in Bohn (1974a, Fig. 1). This has been further confirmed by the results of series D. Series E was designed to study the possibility of a similar distribution behind the hindleg. Complete sets of basal sclerites of midlegs were implanted at varying distances from the left hindleg, which was left intact (Fig. 8). Since regenerates only occurred in transplants of this type at the posterior border of the trochantin (see also Expt. G1 in Bohn, 1974b) the sclerites had to be implanted with reversed anterior-posterior polarity in some of the experiments (Expt. E1,2 and E5) to allow investigation of the regenerative capacities of the membranous field in the immediate proximity of the hindleg. In these cases the basal sclerites of the right legs were used as grafts, and by rotation, the medio-lateral polarity of the transplant came into accordance with that of the host tissues.
When the sclerites were transplanted to the membranous area near the hindleg an additional leg was often regenerated (Expt. E1,2; Fig. 6). But the frequency of leg regeneration rapidly decreased with increasing distance from the hindleg (Expt. E3−6). The legs regenerated in E1,2 had reversed anterior-posterior polarity, as was to be expected from the polarity of the sclerites. As usual, when no legs were regenerated, the sclerites had been reduplicated symmetrically (Fig. 5). Therefore the posterior part of the regeneration field of the hindleg seems to be similar to that of the other legs. An area with LIM is followed posteriorly by a membranous field which only allows regeneration of sclerites (SIM). In Expt. E5 there was one exception -an animal which had regenerated two legs (Fig. 7). Obviously there had been an enlargement of the wound surface. The sternite of the first abdominal segment had been split into two parts. Because of this gap the transplant came into contact with the intersegmental membrane of the next segment, which seems to have leg-inducing capacities (as will be seen in series F).
Figs. 5-7. Examples of results for experimental series E. Reduplicated basal sclerites (p, praecoxa; e, epimeron; t, trochantin) after transplantation of sclerites near the first abdominal sternite (a). The parts below the broken line (axis of symmetry) represent the transplant, those above the line the regenerate.
Figs. 5-7. Examples of results for experimental series E. Reduplicated basal sclerites (p, praecoxa; e, epimeron; t, trochantin) after transplantation of sclerites near the first abdominal sternite (a). The parts below the broken line (axis of symmetry) represent the transplant, those above the line the regenerate.
A leg (r) is regenerated when the basal sclerites are implanted near the coxa of the hindleg. The coxa of the regenerate is mostly hidden beneath the proximal parts of the hindleg (hl) whose distal parts were removed for photography.
Exceptional case with leg regenerates (c, coxae of the two regenerated legs) after implantation of the basal sclerites near the first abdominal sternite. By obliteration of part of the first abdominal sternite (a) the transplant had come into contact with the intersegmental membrane between the first and second abdominal sternite (arrow).
Exceptional case with leg regenerates (c, coxae of the two regenerated legs) after implantation of the basal sclerites near the first abdominal sternite. By obliteration of part of the first abdominal sternite (a) the transplant had come into contact with the intersegmental membrane between the first and second abdominal sternite (arrow).
Design of experimental series E. The basal sclerites of the left or right midleg (lower figure, heavily dotted area) were implanted at different distances (heavily dotted outlines in upper figures) behind the left hindleg, in normal or anterior-posterior reversed position. For drawing, the left hindleg was turned anteriorly to allow full view of the implantation area. The left midleg was left intact but is drawn only up to the base of the coxa (broken line). Leg regeneration only occurs when the cut surface of the transplanted trochantin is lying in the anterior half of the membranous region between the hindleg and the first abdominal sternite; in the other cases (below the horizontal line) only a symmetrical reduplication of the basal sclerites occurs. Therefore, similar to the situation at the base of the midleg, the membranous field behind the hindleg consists of two areas with different morphogenetic capacities: anteriorly, a ‘leg-inducing membrane’ (LIM), posteriorly a ‘sclerite-inducing membrane’ (SIM). Lightly dotted area, membranous parts; clear area, sclerotized parts; longitudinally hatched area, wing anlagen.
Design of experimental series E. The basal sclerites of the left or right midleg (lower figure, heavily dotted area) were implanted at different distances (heavily dotted outlines in upper figures) behind the left hindleg, in normal or anterior-posterior reversed position. For drawing, the left hindleg was turned anteriorly to allow full view of the implantation area. The left midleg was left intact but is drawn only up to the base of the coxa (broken line). Leg regeneration only occurs when the cut surface of the transplanted trochantin is lying in the anterior half of the membranous region between the hindleg and the first abdominal sternite; in the other cases (below the horizontal line) only a symmetrical reduplication of the basal sclerites occurs. Therefore, similar to the situation at the base of the midleg, the membranous field behind the hindleg consists of two areas with different morphogenetic capacities: anteriorly, a ‘leg-inducing membrane’ (LIM), posteriorly a ‘sclerite-inducing membrane’ (SIM). Lightly dotted area, membranous parts; clear area, sclerotized parts; longitudinally hatched area, wing anlagen.
Two legs were regenerated in this case because of the obliteration of the medial part of the praecoxa (see series G in part III).
Series F (Table 2)
This series of experiments attempted to discover whether leg-inducing capacities were restricted to the corresponding membranes in the thoracic segments, or whether there were other tissues -for instance, on the abdomen -which in combination with basal sclerites would allow regeneration of legs. These experiments were suggested by an exceptional result of experiment E5, in which legs had been regenerated by a trochantin which had come into contact with the intersegmental membrane between the first and second abdominal segments.
Compilation of the results of experimental series F. Transplantation of basal sclerites to abdominal segments

Complete sets of hindleg basal sclerites were transplanted in normal orientation to the ventral (Expt. F1;2) or dorsal (Expt. F3,4) surface of the abdomen. The posterior margin of the trochantin either bordered the anterior margin of the intersegmental membrane (Expt. F1,3) or a sternite or tergite (Expt. F2,4).
When no leg was regenerated, there was a reduplication of the basal sclerites (Fig. 9). In 10–14% of cases where the posterior margin of the trochantin bordered the anterior margin of the intersegmental membrane, legs were formed -irrespective of whether the sclerites had been transplanted to the dorsal or the ventral surface (Fig. 10).
Examples of results for experimental series F. Legs may be regenerated after transplantation of basal sclerites to abdominal segments (r, Fig. 10) provided that the sclerites are combined with the intersegmental membrane. In most cases, however, only symmetrical basal sclerites are formed (Fig. 9). Broken line, axis of symmetry; above the line, transplant; below the line, regenerate.
Examples of results for experimental series F. Legs may be regenerated after transplantation of basal sclerites to abdominal segments (r, Fig. 10) provided that the sclerites are combined with the intersegmental membrane. In most cases, however, only symmetrical basal sclerites are formed (Fig. 9). Broken line, axis of symmetry; above the line, transplant; below the line, regenerate.
Examples of results for experimental series F. Legs may be regenerated after transplantation of basal sclerites to abdominal segments (r, Fig. 10) provided that the sclerites are combined with the intersegmental membrane. In most cases, however, only symmetrical basal sclerites are formed (Fig. 9). Broken line, axis of symmetry; above the line, transplant; below the line, regenerate.
Examples of results for experimental series F. Legs may be regenerated after transplantation of basal sclerites to abdominal segments (r, Fig. 10) provided that the sclerites are combined with the intersegmental membrane. In most cases, however, only symmetrical basal sclerites are formed (Fig. 9). Broken line, axis of symmetry; above the line, transplant; below the line, regenerate.
DISCUSSION
The conclusions regarding the regeneration field of cockroach legs reached in a previous paper (Bohn, 1974a) have now been confirmed by transplantation experiments. After the loss of regenerative ability following an extirpation of leg plus sclerites, the regenerative capacities could be restored by implantation of sclerites -for instance, the trochantin (series D). But the success of the implantation depended on the site of implantation. When the trochantin had been implanted far anteriorly (Expt. D1) only sclerites were regenerated. At a more posterior position (Expt. D2), either sclerites, or one or two legs were formed (Fig. 1). These results are interpreted as follows (see Fig. 11) : the sclerite parts of the legs are bordered by membranous areas, anteriorly by the sclerite-inducing membrane (SIM), and posteriorly by the leg-inducing membrane (LTM). After extirpation of the leg including all basal sclerites, the former leg area is covered by a homogeneous membrane which should partly consist of LIM and partly of SIM. The results of the implantation experiments suggest an approximate boundary between these two kinds of tissue. In Expt. D1 the trochantin obviously had been implanted within a region covered by SIM, therefore sclerites were regenerated at its anterior and posterior margins (position 1 in Fig. 11). The site of implantation in Expt. D2 seemingly was in the region of the boundary between SIM and LIM. When the trochantin happened to lie within the SIM (position 2, Fig. 11) only sclerites were regenerated, as in Expt. D1. The cases where one complete leg was regenerated, can easily be explained by an intermediate position (position 3 in Fig. 11) of the trochantin; anteriorly it bordered the SIM, therefore the basal sclerites were completed; posteriorly it had contact with the LIM, so a leg was regenerated. The third possibility was that the trochantin was completely surrounded by LIM (position 4, Fig. 11); in this case two legs were regenerated, a normal one at the posterior margin of the trochantin and another at the anterior margin, with reverse anterior-posterior polarity.
Interpretation of the results of experiments D1 and D2, (compare with Fig. 1). The type of regenerate depends on the site of implantation of the trochantin (triangular structure). In Expt. D4 all transplants (1) were lying within the region of the SIM, therefore only sclerites were regenerated. The site of implantation in Expt. D2 was in the region of the boundary between SIM and LIM, so the trochantin happened to lie either completely within the SIM (2), or within the LIM (4), or in between (3). Therefore in case (2) only basal sclerites were formed, as in Expt. D2 (1); in case (3) sclerites were regenerated at the anterior (here upper) margin of the trochantin; but a leg was regenerated (indicated by the arrow) at its posterior margin which had contact with the LIM. Because of the contact of the trochantin with the LIM at both margins, two legs developed in case (4), one at the anterior and one at the posterior margin.
Interpretation of the results of experiments D1 and D2, (compare with Fig. 1). The type of regenerate depends on the site of implantation of the trochantin (triangular structure). In Expt. D4 all transplants (1) were lying within the region of the SIM, therefore only sclerites were regenerated. The site of implantation in Expt. D2 was in the region of the boundary between SIM and LIM, so the trochantin happened to lie either completely within the SIM (2), or within the LIM (4), or in between (3). Therefore in case (2) only basal sclerites were formed, as in Expt. D2 (1); in case (3) sclerites were regenerated at the anterior (here upper) margin of the trochantin; but a leg was regenerated (indicated by the arrow) at its posterior margin which had contact with the LIM. Because of the contact of the trochantin with the LIM at both margins, two legs developed in case (4), one at the anterior and one at the posterior margin.
Most of the previous experiments for the study of regeneration fields in the legs of Leucophaea (Bohn, 1974a) had involved the midlegs. It was not considered likely that there were great differences in the regenerative fields of the different legs; nevertheless, it seemed desirable to make at least one trial experiment with the field of another leg-for instance, of the hindleg. Instead of extirpating tissues, basal sclerites were implanted at different distances from the hindleg (series E). Regeneration only occurred in the anterior half of the membranous field which extends between the hindleg and the first abdominal segment. Therefore it seems that -as in the case of the midleg -the coxa of the hindleg is followed posteriorly by a strip of LIM, which then is followed by an area consisting of SIM.
In series E there was one exceptional case (Expt. E5, Fig. 7) with a leg regenerate arising from a very posterior transplant, in a region which was thought to be covered by SIM. However, in this animal the implantation of the transplant obviously was followed by extended necrosis of the surrounding host tissues. The enlarged wound surface caused a splitting of the sternite of the first abdominal segment into two parts. By means of this gap the transplant was able to come into contact with the intersegmental membrane between the first and second abdominal segments. The simplest explanation for the exceptional leg regenerate is that there is a leg-inducing capacity in the intersegmental membranes of the abdomen. This assumption was tested in the last experimental series (F). Basal sclerites were transplanted either to a membrane region or to a sclerite region of both ventral and dorsal surfaces of the abdomen. Legs were regenerated on both surfaces, but only when the sclerites had been combined with the tissues of the intersegmental membranes. Therefore, leg-inducing capacities are not restricted to the surroundings of a leg, but are also present in the intersegmental membranes of the abdominal segments, which normally have no legs. The rather low incidence of regenerates (10-14%) could indicate that, as in the case of the thoracic segments, only part of the membranous area has leg-inducing capacities.
When the three experimental series of this paper are compared with respect to the number of regenerated legs per transplant, a striking difference appears between those experiments in which only the trochantin had been transplanted (series D) and the experiments where a complete set of basal sclerites (i.e. trochantin plus praecoxa; series E and F) was used as the transplant. In series D, 44% of the animals with leg regenerates had double legs, while in the two series E and F only 5 % (two cases) had double legs. The development of double legs seems to be caused by contact of the trochantin with LIM both at its posterior and at its anterior margin; one leg is formed at each margin. In series E and F, only the posterior margin of the trochantin has a free border which can be brought into contact with LIM; the anterior margin is covered by the other basal sclerites which, despite their contact with LIM, do not form leg regenerates. A similar situation with similar results was described in Bohn (1974a) (series C), where the basal sclerites of an intact leg had been brought into contact with the LIM of the previous segment. Leg regeneration only occurred after removal of the praecoxa, when the trochantin had contact with the LIM. The simplest conclusion which could be drawn from these results -that the praecoxa has no capacity for leg regeneration -is not valid, since regeneration occurred from a posterior cut surface of the praecoxa (Expt. A2−4) with high frequency. The difference in reaction seems to be due to the different orientation of the cut surface of the praecoxa. In Expt. A2−4 it was a posterior cut surface; in series C, D and F it was an anterior cut surface. The reason for the different capacities of differently orientated cut surfaces is not clear. However, the two cases with double legs in series E and F need a satisfactory explanation. The second leg had formed although the greater part of the praecoxa was present, and prevented the trochantin from extensive contact with the LIM. But it seemed that the medial-most wedge of the praecoxa was missing. Possibly the contact of the trochantin with the LIM over such a restricted length was still sufficient to elicit leg regeneration. Therefore, the question arises: to what extent and at which point must the trochantin be exposed to the LIM to allow regeneration at the anterior margin ? Experiments of this kind have been performed and are presented in Bohn (1974b).