ABSTRACT
A grasshopper hatches from its egg, which is laid in soil, as a vermiform larva. This larva continues the stereotyped hatching behaviour as it digs through the egg pod, which provides a passageway to the soil surface. Once at the surface, shedding, or ecdysis, of the vermiform cuticle is initiated. When this process is complete, the first-instar cuticle is expanded to assume the form of the first-instar hopper. We have demonstrated, using immunocytochemical techniques, that these behaviour patterns are associated with dramatic increases in intracellular levels of cyclic GMP in sets of identified neurones in the ventral central nervous system.
The most prominent cyclic-GMP-expressing cells are 34 neurones that appear to contain crustacean cardioactive peptide (CCAP). These CCAP cells show no detectable cyclic GMP at hatching or while the vermiform larva digs through the soil. Upon reaching the surface and freeing itself, the larva initiates ecdysis and associated air-swallowing and tracheal filling within about 1 min. These changes are immediately preceded by the appearance of cyclic GMP in the CCAP cells. Cyclic GMP levels in these neurones peak by 5 min and then decline back to basal levels by 20–30 min. Conditions that cause ecdysing animals to resume digging prolong the elevation of cyclic GMP levels. Once animals have assumed their ‘hopper’ form, however, external stimuli can no longer affect the time course of the cyclic GMP response. The neurones containing elevated cyclic GMP levels probably influence the air-swallowing, tracheal filling and circulatory changes that are associated with ecdysis behaviour.
Pairs of descending midline neurones in abdominal segments 2–4 also become cyclic-GMP-immunoreactive, but they show peak expression after cyclic GMP levels in the CCAP cells have declined. Also, neurones in the caudolateral region of the abdominal ganglia often become cyclic-GMP-immunoreactive when ecdysing animals are forced to resume digging for an extended period.
Introduction
Ecdysis in insects involves a stereotyped sequence of behaviours that is initiated with the release of the neuropeptide eclosion hormone (EH; Truman, 1992). Studies on various Lepidoptera have shown that this peptide acts directly on the central nervous system (CNS) to evoke the behavioural cascade that brings about the shedding of the old cuticle (Truman, 1978; Hewes and Truman, 1991). Besides its behavioural effects, EH also causes the release of other neuropeptides that orchestrate the physiological changes that accompany ecdysis. Among these effectors are the cardioacceleratory peptides (CAPs) that increase heart rate at the time of ecdysis (Tublitz and Truman, 1985; Tublitz and Evans, 1986). One of the CAPs (CAP2a) in Manduca sexta was subsequently sequenced and found to be identical to crustacean cardioactive peptide (CCAP; Cheung et al. 1992), a peptide first discovered in crustaceans (Dircksen and Keller, 1988) but later found to be distributed widely in insects (Dircksen, 1994). Neurones that contain CCAP in M. sexta show a striking increase in intracellular 3’5’-cyclic guanosine monophosphate (cyclic GMP) levels shortly before the onset of ecdysis (Ewer et al. 1994). This rise in intracellular cyclic GMP levels causes an increase in excitability of these neurones (Gammie et al. 1994; S. C. Gammie and J. W. Truman, unpublished observations), presumably facilitating the release of CCAP. Diverse insect species, with the exception of members of the higher Diptera, show a similar increase in intracellular cyclic GMP levels in CCAP-containing neurones at ecdysis (Ewer and Truman, 1996).
One of the best timed and characterized ecdyses in insects is that exhibited by locusts shortly after hatching. The hatching locust is covered by the second embryonic cuticle and is termed the vermiform larva (Bernays, 1971). The vermiform larva utilizes a specialized behaviour to escape from the chorion and dig upwards through the egg pod to the surface of the soil (Bernays, 1971, 1972a). As the anterior end of the larva comes free, the animal ceases hatching/digging movements and rapidly shifts to ecdysis behaviour. The second embryonic (vermiform) cuticle is shed within minutes (Fig. 1), and the larva assumes the shape and proportions of the first-stage hopper. This dependence of ecdysis on the freeing of the larva makes it possible to control precisely the time of onset of ecdysis behaviour and makes hatching locusts excellent animals in which to explore further the relationship between ecdysis and the changes in cyclic GMP levels that occur in the CCAP neurones.
Materials and methods
Experimental animals
A laboratory colony of Locusta migratoria was maintained on a diet of wheat germ and sprouted wheat seedlings. Breeding females were provided access to cups of moist sand into which they deposited their egg pods. The oviposition cups were removed weekly and maintained at 30 °C. Intact individual egg pods were removed from the sand and maintained between layers of moist cheesecloth in Petri dishes. Individual eggs were dissected at intervals to determine the developmental stage of each clutch. Staging of embryos was based on the system of Bentley et al. (1979), with additional criteria added for late embryos (E. E. Ball and J. W. Truman, in preparation).
Fixation and immunocytochemistry
In order to obtain an accurate picture of the cyclic GMP response at specific times, we dissected animals as rapidly as possible in 4 % formaldehyde (in phosphate-buffered saline, PBS). Embryos or hatching larvae were stripped out of the chorion, immersed in fixative at room temperature and rapidly opened mid-dorsally. Following removal of the gut, the body wall was pinned out flat. Fixation was generally for about 2 h at room temperature or overnight at 4 °C. To improve reagent access after fixation, the CNS was dissected free from most of the body wall and processed as an isolated nerve cord. Adhering muscle tissue in the meso-(T2) and metathoracic (T3) regions of some nervous systems reduced the quality of immunostaining in these areas. Consequently, our analyses focus primarily on the prothoracic (T1) ganglion and the unfused abdominal ganglia (A4–A8).
Cyclic GMP levels were determined in situ using the immunocytochemical methods developed by De Vente et al. (1987). After repeated washes, fixed tissues were pre-blocked in 5 % normal goat serum (NGS) in PBS with 0.3 % Triton X100 (PBS-TX), and then incubated with a 1:4000 dilution of a rabbit anti-cyclic-GMP antiserum (generously provided by Dr Jan De Vente) in PBS-TX with 1 % NGS. After 12–36 h at 4 °C, the tissues were repeatedly washed and then incubated overnight with a peroxidase-conjugated goat anti-rabbit IgG (Kierkegaard and Perry Labs) in PBS-TX with 1 % NGS. After additional rinses, the location of the antibody complexes was revealed by reaction with diaminobenzadine (DAB, Sigma) and glucose oxidase (Watson and Burrows, 1981) or H2O2. The developing solution contained 0.03 % nickel chloride to yield a black reaction product. Tissues were then dehydrated, cleared in methyl salicylate, and mounted in Permount (Fisher Labs). We scored the intensity of cyclic GMP staining in the cell body of some of the neurones as follows: 0, no staining; 1, cell body can be distinguished as pale grey; 2, cell body is a moderate grey; 3, cell body is dark grey to black.
Behavioural manipulations
Individual egg pods were partially opened and monitored to assess the progress of development. When hatching began, larvae were timed from the beginning of emergence from the chorion or from the time that they initiated ecdysis movements. Ecdysing larvae were placed in individual containers until the time of their dissection.
Hatching animals showed sustained ‘digging’ behaviour when placed under two layers of moist cheesecloth in a plastic Petri dish. The metathoracic legs of digging vermiform larvae retained the position that they had occupied while in the egg, with the femur extended posteriorly but the tibia pointed anteriorly. We obtained digging ecdysing larvae by placing animals that had just initiated ecdysis under cheesecloth. Their posture differed from that of digging vermiform larvae in that the metathoracic tibiae projected posteriorly and trailed the digging animal with the exuvium collected around the distal end of the legs.
Image processing
Video images of the histological preparations were captured using a Sony DXC-960MD video camera mounted on a Nikon Optiphot microscope. Two-dimensional projections of neurones within a ganglion were constructed from stacks of optical sections that were imported into Adobe Photoshop 3.0. The stack was assembled using the layering palettes with out-of-focus elements being deleted. The flattened images were then printed on a Tektronix dye sublimation printer.
Results
Neurones that show cyclic GMP accumulation around the time of ecdysis
CCAP cells
Thirty-four cells distributed from the suboesophageal ganglion (SOG) to A7 (Ewer and Truman, 1996) show elevated levels of intracellular cyclic GMP in ecdysing locusts (Fig. 2). Four pairs of neurones are present in the SOG, two pairs each in ganglia T1–T3 and one pair in abdominal ganglia A1–A7.
The cell body positions and axon trajectories of all 34 cells identify them as members of the group of cells that contain CCAP (Dircksen et al. 1991; Dircksen and Homberg, 1995). The posterior three pairs of SOG neurones and the segmental neurones from T1 to A7 are type 1 and type 2 CCAP neurones. In the maxillary SOG neuromere (S2), one pair of neurones sends axons posteriorly through contralateral descending lateral tracts (Fig. 2A) and, thus, these are type 2 neurones. The S2 and S3 type 1 cells appear to be CN2 and CN3, respectively (Dircksen and Homberg, 1995); these neurones project via the brain to pharyngeal dilator muscles (Braünig, 1990). In each thoracic ganglion, one of the two pairs of cyclic-GMP-immunoreactive neurones is also of type 2 since these cells have the characteristic contralateral axonal bifurcation, with one axon exiting to the periphery via nerve 5 and the other axon joining the SOG axon in the descending lateral tract (Fig. 2B). The axons of these thoracic type 2 cells project the length of the CNS and arborize in the terminal ganglion. The abdominal ganglia also have type 1 and type 2 CCAP cells (Dircksen et al. 1991). In contrast to their thoracic homologues, the abdominal type 2 cells failed to show detectable levels of cyclic GMP at ecdysis, whereas the abdominal type 1 cells showed a robust increase (Fig. 2D). The type 1 cells in both the thoracic and abdominal segments are characterized by an initial process that extends ventrally to the midline and then projects dorsally and forms a contralateral axon. In most instances, the axon then splits, with one branch projecting posteriorly to the next ganglion where it exits to the periphery via the contralateral intersegmental nerve. The other axon projects anteriorly to the intersegmental nerve in the ganglion of origin (Fig. 3A). The posterior axon projects to the alary muscles of the heart (Dircksen et al. 1991).
This splitting of the axon of the type 1 cell has not been observed except in the S3 neuromere (Dircksen et al. 1991; Dircksen and Homberg, 1995). Indeed, in a few of our animals, the anterior or posterior axon in a particular segment was missing. We do not know whether this is a stage-dependent (Dircksen et al. 1991 used adults) or a strain-dependent difference.
The descending axons of the thoracic type 2 neurones appear to contact the initial processes of the type 1 neurones. A similar feature is evident in M. sexta and a number of other insects (Ewer et al. 1994; Ewer and Truman, 1996). This anatomy suggests that the cells may interact directly with one another.
Midline cells
Paired cells near the ventral midline in ganglia A2–A4 also show increases in intracellular cyclic GMP levels (Fig. 3B). These cells do not contain CCAP. Their axons descend the ipsilateral connectives to unknown termination sites.
Caudolateral cells
The last set of cells that shows cyclic GMP immunoreactivity consists of a group of two or three paired cells situated in the caudolateral regions of ganglia A1–A7. They have ipsilateral axons that loop around the lateral tract and leave the ganglion via the ipsilateral segmental nerve (Fig. 3C). These cells are in the position of the lateral neurosecretory cells that project to the perivisceral neurohaemal organs (Taghert and Truman, 1982; Dircksen et al. 1991).
Time course of cyclic GMP accumulation at the time of ecdysis
Normally, a hatching locust digs through the soil as a vermiform larva and initiates ecdysis only after it reaches the surface. If the anterior end of the egg is exposed, however, the hatching larva can dispense with digging altogether. While still partially in the chorion, it switches to ecdysis behaviour, which then frees it from both the chorion and the old cuticle (Bernays, 1971, 1972a,b).
We collected larvae that were in various stages of hatching from partially dissected egg clutches and examined their cyclic GMP responses. The progression of cyclic GMP staining in the CCAP cells as a function of time after the start of ecdysis is shown in Table 1. We focused on the type 1 and 2 cells in T1, the type 2 cell axons (from the SOG and T1–T3) in the abdominal lateral tract, and the type 1 cells in A4–A7. The neurones of the SOG are omitted from quantitative consideration because this ganglion was often damaged during dissection. We did not quantify the intensity of cyclic GMP expression for individual cells in this group.
Of 20 animals examined within the first minute after starting to emerge from the chorion, seven had begun ecdysis and were dissected immediately following their first ecdysis movement. All had air in their gut or tracheal system and all showed cyclic GMP staining in their CCAP neurones. The remaining 13 hatching larvae were fixed prior to the onset of ecdysis movements. There was no air in either the gut or tracheae of these animals. Eight of the 13 were devoid of any trace of cyclic GMP immunoreactivity (cGMP-IR) in the CCAP network (e.g. Figs 4, 5; −1 min). The remaining five animals all showed distinct cyclic GMP staining in the CCAP neurones. In the thoracic cells, cGMP-IR was evident only in the medial neurites, with little or no staining in the cell bodies. The type 2 cell axons in the abdominal portion of the lateral tract were stained along their entire length. The type 1 cells in A4–A7 showed a weak response in their cell bodies and processes without any evidence of an anterior-to-posterior gradient. In four of five animals, cGMP-IR was also evident in some of the SOG cells, including axons that projected to the brain.
Animals at the start of ecdysis, fixed only seconds after the preceding set of animals, showed the same pattern of staining (Table 1; Figs 4, 5, 0 min). Over the next 2–5 min, the cyclic GMP levels continued to rise within the network (Figs 4, 5; 2–5 min). In the thoracic cells, cGMP-IR invaded the cell bodies by 2 min and by 5 min the cyclic GMP levels appeared maximal. The abdominal cells also reached their highest levels of intracellular cyclic GMP by about 5 min after the start of ecdysis. By 10 min, the overall intensity of the cyclic GMP response had waned. In the thorax, the cyclic GMP staining was faint in both the cell bodies and the medial neurites (Fig. 4; 10 min). The axons of the type 2 cells that projected through the abdominal ganglia still showed moderate cyclic GMP staining, but their terminations in the terminal ganglion were no longer stained. The abdominal type 1 cells showed moderate cyclic GMP staining (Fig. 5, 10 min). By 20 min, cyclic GMP levels in the thoracic ganglia were reduced to traces of staining in the medial neurites (Fig. 4, 20 min). The descending axons of these cells showed a gradient of weak staining, with cGMP-IR falling to undetectable levels before reaching A7. The abdominal type 1 neurones stained weakly (Fig. 5, 20 min), with those in the posterior ganglia being the most weakly stained. By 30 min, the cyclic GMP levels in the descending axons had fallen below the level of detectability. A few abdominal neurones still showed very low levels of cyclic GMP staining confined to the cell body.
Behavioural modification of the cyclic GMP response in the CCAP neurones
When eggs were placed under moist cheesecloth, the newly hatched vermiform larvae did not ecdyse but rather showed the abdominal movements and cyclical extension of their cervical ampullae characteristic of digging behaviour (Bernays, 1971). We examined five vermiform larvae after 10–60 min of continuous digging. None had air in either their tracheae or gut at the time of dissection. Four were completely lacking cGMP-IR in their CCAP cells (Fig. 6A; Table 2) and one animal at 60 min after hatching showed a trace of immunostaining in the type 1 cells in A4 and A5. Thus, stimuli that delay ecdysis and cause the larva to remain as a vermiform larva also delay the induction of cyclic GMP.
When a digging vermiform larva encountered an open space, it switched from digging to ecdysis behaviour. If it was then covered again, it resumed digging, but this time as a digging ecdysing larva. As described in Materials and methods, these larvae differed from vermiform larvae in their leg position during digging. Also, when dissected, they always had air in both their gut and their tracheal systems. Such partially ecdysed diggers maintained detectable cyclic GMP levels in their CCAP network for at least twice as long as animals freed immediately after ecdysis (Fig. 6; Table 2). In spite of this prolongation of high levels of cyclic GMP expression, the pattern of disappearance of cyclic GMP was similar to that seen in normal animals; namely, a fading of cGMP-IR from the thoracic cell bodies and from the descending axons, starting from their posterior ends. By 120 min, cyclic GMP levels in these digging animals had dropped below our level of detection.
Digging animals have a difficult time walking or even standing when they first reach the surface of the soil. Within a minute or two of being freed, however, these larvae expand and straighten their new cuticle to its proper hopper proportions. These ‘hoppers’ can now walk and hop, but they can no longer dig. When placed under cheesecloth, they show kicking and struggling movements but no coordinated digging behaviour. As seen in Table 2, the types of confinement that prolonged cyclic GMP expression in the partially ecdysed digging larvae did not do so in these larvae that had become hoppers. Thus, once ecdysed animals make the transition to being hoppers, cyclic GMP levels in the CCAP cells are no longer affected by external stimuli.
Changes in cyclic GMP levels in other neurones during ecdysis
The two other types of neurones that showed cyclic GMP expression around the time of hatching were the midline cells and the caudolateral cells. Cyclic GMP staining in the midline cells was observed only rarely prior to hatching. In embryos taken from egg pods at 3 h (N=8), 11 h (N=9) or 28 h (N=10) prior to hatching, we observed weak cyclic GMP staining in midline cells in only four individuals (all from the -28 h collection). At hatching, over 50 % of the animals showed some level of cyclic GMP in the A4 midline cells. The cyclic GMP response in these cells, however, did not depend on whether the CCAP cells had started to respond. By 2 min after hatching, midline cells were stained in essentially all the animals. The levels of cyclic GMP in these cells continued to increase while those in the CCAP network rose and then fell. Cyclic GMP levels began to wane in the midline cells by about 45 min post-hatching.
Expression of cyclic GMP in the caudolateral cells was seen only in partially ecdysed larvae that were showing digging behaviour. Immunoreactive cells were observed in one of five animals that had been digging for 20–30 min, two of five that had dug for 40–50 min, and three of five that had done so for 110 min.
Changes in cyclic GMP levels during premature ecdysis behaviour
In Lepidoptera, the mechanical stimulation of peeling the pupal cuticle from developing adult moths can trigger the precocious onset of ecdysis behaviour (Blest, 1960; Truman, 1976; Kammer and Kinnamon, 1974). Likewise, late locust embryos that were removed from their egg shells and placed in PBS often showed weak ecdysis movements after a delay of a few minutes. These ecdysis movements were also associated with the appearance of cyclic GMP in cells of the CCAP network. As seen in Table 3, the appearance of cyclic GMP in these cells could be induced only during approximately the last 25 % of embryogenesis. Initially, the response was typically confined to the abdominal cells. The number of these cells that responded ranged from only one or two in the younger embryos (Fig. 7A) to the entire abdominal set. Within the last 4–6 h before the hatching of the clutch, the embryos became very sensitive to disturbance. Removing them from the chorion resulted in a very rapid onset of ecdysis movements and activation of the entire network from the SOG to the seventh abdominal ganglion (Fig. 7B). We did not examine how the strength of the precocious ecdysis behaviour was correlated with the extent of the cyclic GMP response in the CCAP cells.
Discussion
Relationship of changes in cyclic GMP levels to the events of hatching
A hatchling locust passes through three behavioural phases during the transition from a hatching vermiform larva to a normally proportioned first-instar hopper. First, the vermiform larva performs a stereotyped hatching behaviour and continues with this behaviour as it digs up through the confining space of the egg pod (Bernays, 1971). Second, when the anterior end of the animal becomes free, the vermiform larva initiates the ecdysis of the embryonic cuticle (Bernays, 1972a). However, given the appropriate conditions, this ecdysing animal can then revert back to digging, albeit with a different leg posture. Third, upon escaping confinement, the animal expands the first-instar cuticle and assumes the proper form of the hopper. Walking and hopping are now possible but digging behaviour can no longer be elicited. Each of these three phases is variable in length and the transition from one to the next is triggered by particular stimuli (or lack thereof). Once made, a transition appears to be irreversible.
The dramatic increase in intracellular cyclic GMP levels seen in the CCAP cells was invariably linked to entry into the ecdysis phase. If the anterior end of the hatching vermiform larva did not contact an obstruction, the insect switched to starting ecdysis within about a minute. Cyclic GMP began to appear in the CCAP cells, and these were invariably cyclic-GMP-immunoreactive by the time of ecdysis (Table 1). This increase in cyclic GMP levels could be delayed for at least an hour by forcing the vermiform larva to dig (Table 2).
The appearance of intracellular cyclic GMP in the CCAP network prior to the onset of ecdysis behaviour has also been observed for larval ecdysis in M. sexta (Ewer et al. 1994). This relationship in two different species raises the intriguing possibility that the activation of the CCAP network might be the immediate trigger for the start of ecdysis behaviour.
Once initiated, the duration of the cyclic GMP response was subject to modulation by external factors. In animals that were not confined, ecdysis was completed rapidly, and within 2–3 min the hopper had expanded its cuticle to the proper proportions and had started pigmentation. Under these conditions, the cyclic GMP response in the CCAP cells waned within 20–30 min. By contrast, when ecdysis and digging were prolonged, the CCAP cells maintained elevated intracellular cyclic GMP levels for at least twice the normal duration.
In M. sexta, the anatomy and peptide content of the cells that show the cyclic GMP increase suggest that these cells mediate physiological changes that occur at ecdysis, such as air-swallowing and increases in heart rate (Ewer et al. 1994). In Locusta migratoria, both physiological (Braünig, 1990) and anatomical (Dircksen and Homberg, 1995) studies show that the SOG type 1 cells project to dilator muscles of the pharynx. Importantly, cyclic GMP begins to appear in these cells immediately before the onset of air-swallowing, supporting the involvement of these cells in this behaviour. With respect to circulatory changes, the locust abdominal type 1 CCAP cells project to the heart alary muscles and CCAP has stimulatory effects on the heart (Dircksen et al. 1991). Moreover, ecdysing locusts show a dramatic increase in blood pressure, which then gradually falls over the next 30 min (Bernays, 1972a). This change in blood pressure mirrors the changes in cyclic GMP levels in the abdominal CCAP cells. Besides air-swallowing and circulatory changes, the CCAP cells may also play a role in the filling of the tracheal system. Resorption of fluid and the entry of air into the new tracheae is an important aspect of every ecdysis (Reynolds, 1980). It is especially dramatic, however, at the ecdysis of the embryonic cuticle since the tracheal system of the vermiform larva is filled with fluid, which is rapidly replaced by air when ecdysis begins. The timing of the filling of the tracheal system and the fact that CCAP neurones also project to the spiracles (Dircksen et al. 1991) suggest a possible role for CCAP in these tracheal events.
The functions of the medial cells and the caudolateral cells are not known. The time course of cyclic GMP accumulation in the medial cells is intriguing because it starts at about the time of hatching but does not peak until about 30–40 min later, after the CCAP cells have finished their response. These cells are probably involved in some events that occur during the early part of the hopper phase. The caudolateral cells are also interesting because we saw them responding in only one context – as partially ecdysed larvae were digging through the layers of moist cheesecloth. In L. migratoria, one pair of these cells also contains CCAP (Dircksen et al. 1991). In M. sexta, the corresponding cells contain CAPs (Tublitz and Sylwester, 1990) and diuretic hormone (Chen et al. 1994) in the larvae and bursicon in the adult (Taghert and Truman, 1982). These cells also occasionally show cGMP-IR during ecdysis in M. sexta (J. Ewer and J. W. Truman, unpublished observations). The possibility of affecting diuretic hormone neurones is especially intriguing in the context of an animal digging through a moist environment while showing vigorous swallowing behaviour.
Spread of cyclic GMP through the CCAP network
The ability to induce ecdysis within a minute on command made it possible to examine the time course of changes in cyclic GMP levels within individual cells. The suboesphageal, thoracic and abdominal cells turn on as a unit. Our incomplete staining of the neurones in the SOG precludes a detailed description of the appearance of cyclic GMP in these cells. For the thoracic cells, the first appearance of cyclic GMP was in their medial neurites and their descending axons. Only later did cyclic GMP appear in the cell body. This pattern suggests that the increase in cyclic GMP levels may be initiated by inputs onto the medial neurites of these cells. In M. sexta, transection experiments show that the activation of the cyclic GMP response in the abdominal CCAP cells depends on a descending input from the anterior end of the larva (Ewer et al. 1994). The axons of the type 2 neurones are perfectly situated to supply such a signal, as they project in the lateral tract through the entire length of the CNS and appear to contact the type 1 cells in each abdominal ganglion. In L. migratoria, the appearance of cyclic GMP in these type 2 axons was always accompanied by the induction of cyclic GMP in the type 1 CCAP cells in A4–A7. In this respect, the embryos that showed precocious ecdysis behaviour were interesting because we saw two types of responses in this network. In some cases, we saw a pattern similar to normal ecdysis, with the axons in the dorsal lateral tract and the abdominal type 1 cells showing the appearance of cGMP-IR. In other cases, we saw only scattered type 1 cells expressing cyclic GMP, but this was in the absence of a response in the descending axons. We think it likely that the rapid, coordinated activation of the abdominal cells is through input from the thoracic type 2 cells via their axons in the lateral tract. Consistent with this hypothesis, we find that cyclic GMP in the abdominal cells appears simultaneously in the cell bodies and the medial neurites. These two regions are similar distances away from the proposed contact point with the type 2 axons in the lateral tract.
Relationship of the CCAP network to eclosion hormone
Direct demonstration of the role of EH in triggering the behaviour patterns associated with ecdysis has been made only for Lepidoptera (Truman, 1992). In M. sexta, for example, release of EH is followed by the performance of the pre-ecdysis behaviour, with the subsequent activation of the CCAP network and ecdysis occurring 30–40 min later. EH bioactivity has been found in grasshoppers (Truman et al. 1981), and recent work using the polymerase chain reaction has provided a partial nucleotide sequence for the EH gene of L. migratoria (F. M. Horodyski, personal communication). Nevertheless, we do not know the timing of EH release in L. migratoria relative to the events of hatching and ecdysis. EH is probably released prior to hatching. According to this scenario, the hatching and digging behaviour of the vermiform larva would then be comparable to the pre-ecdysis behaviour of M. sexta. Although the pre-ecdysis behaviour of larval M. sexta is generally considered to be relatively fixed in duration, ecdysis is delayed when pre-ecdysing larvae are placed on smooth substrata that they cannot grip with their prolegs (J. W. Truman, unpublished observations). Thus, even in M. sexta, the duration of this initial phase is somewhat flexible and sensitive to external conditions. In both M. sexta and L. migratoria the onset of ecdysis behaviour is tightly linked to the activation of the cells in the CCAP network. Whether this relationship results from both systems responding to a common input or whether CCAP release is actually the proximal trigger for ecdysis remains to be established.
ACKNOWLEDGEMENTS
We are grateful to Dr J. DeVente for the anti-cyclic-GMP antiserum. These studies were partially funded by a grant from the National Science Foundation (J.W.T.).