The first larval instar of Nemeritis lasts longer in young caterpillars of Ephestia than it does in older caterpillars. First-instar Nemeritis larvae in young hosts feed and grow very slowly, but they remain capable of fast growth if transferred to older host caterpillars.
Measurements of the protein concentration, the amino acid concentration and the freezing-point depression show rapid changes in the composition of the haemolymph of Ephestia caterpillars at that stage of larval development at which they first become capable of supporting the fast development of first-instar larvae of Nemeritis.
It is suggested that the rate of development of Nemeritis larvae depends on their rate of feeding, and that their rate of feeding is determined by a behavioural response to the composition of their food, the host’s haemolymph; they feed slowly in young host caterpillars because the concentration of solutes (notably amino acids) in their food is so high.
A similar situation may hold when the development of insect parasites is delayed in diapausing hosts.
The larvae of the ichneumonid Nemeritis canescens, endoparasitic in caterpillars of Ephestia kuehniella, remain in their first instar until their host attains its last larval instar. This is an example of a widespread phenomenon; in many other endoparasitic species of Hymenoptera and Diptera growth is arrested in the first instar, as it is in Nemeritis (see, for example, Coppel & Maw (1954), Dowden (1934), Smith & Langston (1953), Thompson (1934); and Salt (1963, p. 667) discusses the persistence of Nemeritis larvae in their first instar in two species of host insects). In many other species, as in Nemeritis, further development of the parasite depends on the host’s reaching a particular stage of growth. For instance, the development of the parasite may begin at the last larval moult of the host, at pupation, or at metamorphosis (examples are listed in Schoonhoven, 1962).
In this paper the hypothesis is put forward that in young caterpillars of Ephestia the first-instar larvae of Nemeritis grow slowly because they feed slowly; that their feeding rate is determined by their response to the composition of their food (the host’s blood), a response possibly mediated by chemoreceptors on the head; and that the composition of the host’s blood at a given time depends on the relative rates of a number of interrelated metabolic processes in the tissues of the host, some or all of which are, directly or indirectly, under hormonal control.
MATERIALS AND METHODS
The experimental animals used in this study are referred to here as Ephestia kuehniella and Nemeritis canescens because these names are well known to experimental zoologists; but the names of both species have been changed recently. Ephestia kuehniella is now Anagasta kuehniella (Zeller) (see Heinrich, 1956), and Nemeritis canescens has been changed to Devorgilla canescens (Gravenhorst) (see Townes, Townes & Gupta, 1961).
The animals used in this study originated from stock cultures maintained in Cambridge by Dr G. Salt. The history of these cultures is described in Salt (1965) (for Ephestia) and in Salt (1964) (for Nemeritis). The culture of Ephestia was kept on Allinson’s wholewheat flour, and the culture of Nemeritis was maintained solely on Ephestia caterpillars. The life-history of Ephestia is described in Corbet (1966), and that of Nemeritis in Corbet & Rotheram (1965).
In this work it has been necessary to define a number of morphologically recognizable stages in the development of caterpillars of Ephestia (Corbet, 1966). The Ephestia caterpillars in the stock cultures normally pass through four larval instars which can be distinguished by the width of the head capsule. Some individuals go through a fifth instar; in head capsule width, the supernumary instar is intermediate between the third instar and the last. Caterpillars in this supernumerary instar were not used in experiments.
Caterpillars weighing less than 5 mg. are in their first, second or early third instar. The third and fourth (final) instars were further subdivided. In mid third-instar caterpillars the head capsule has darkened after the last moult and the neck membrane is still folded under the head capsule. In late third-instar caterpillars the neck membrane is stretched taut and is clearly visible as a pale area between the head capsule and the prothoracic tergite. New fourth-instar (new final-instar) caterpillars can be recognized because the head capsule is still pale after the recent moult. Mid fourthinstar (mid final-instar) caterpillars, in which the head capsule has darkened, still weigh less than 16 mg. Late fourth-instar (late final-instar) caterpillars, which weigh more than 16 mg., have been classified into five stages depending on the degree of retraction of the pigment from the ocelli (Kühn & Piepho, 1936). The first of these stages lasts considerably longer than any of the others. Certain late final-instar caterpillars, which leave the food and crawl up the sides of their jar, are described as wandering caterpillars.
During the experiments caterpillars of Ephestia were kept singly, with rolled oats as food, in 3 in. × 1 in. glass vials or in in. . plastic boxes in the dark in an incubator at 25 +°C. The ages of parasites are given as the number of days since parasitization. For instance, a 4-day-old Nemeritis larva is a larva that has hatched from an egg that was laid 4 days earlier. To avoid the complications that might result from superparasitism, all the hosts used in these experiments were individually parasitized. An Ephestia caterpillar was exposed to several Nemeritis adults, and as soon as the caterpillar was seen to be parasitized it was removed. This method was successful in only about 80 % of cases, and it was necessary to check by dissection that each host contained only one parasite. Results from unparasitized or superparasitized hosts were not used.
To trace the duration of the stages of Nemeritis in hosts of different stages, five larvae of each host stage were dissected on each successive day after parasitization. The times from parasitization are correct ± 4 hr.
When eggs and larvae of Nemeritis were to be transferred from one host to another, they were dissected out from their first host under saline (composition: NaCl, 7·5 g.; KCl, 0·1 g.; CaCl, 0·2 g.; NaHCO3, 0·2 g. per litre). They were taken up in a compound micropipette (Salt, 1955) and injected through a cut in the body-wall of a fresh host caterpillar. When larvae were transferred, a close-fitting glass rod was inserted in the capillary tube of the micropipette; this helped to push the larva out of the capillary (Schneider, 1950).
The freezing-point depression of the haemolymph of Ephestia was determined by the method of Ramsay & Brown (1955) on haemolymph that had been withdrawn under liquid paraffin from a punctured thoracic leg (larvae) or wing pad (pupae). The concentration of proteins in the haemolymph of Ephestia was estimated colori-metrically using the method described by Gornall, Bardawill & David (1949). The solutions were centrifuged to remove haemocytes, and the optical density was measured at 540 mμ on a Unicam SP500 spectrophotometer. The results are expressed in arbitrary units of optical density per microlitre of haemolymph. The ninhydrin-positive solutes were estimated by the method of Rosen (1957), after the proteins had been precipitated with 80% ethanol. The optical density of the solutions was measured on a Unicam SP 500 spectrophotometer at 570 mμ, and the results are expressed in arbitrary units of optical density per microlitre of haemolymph.
Caterpillars whose brains were to be removed were held under saline and the neck membrane was stretched out and pierced so that when the caterpillar was released and gently manipulated, the brain popped out through the hole and could be removed entire. After the operation the caterpillars were blotted dry.
Two techniques were used to study the feeding of parasite larvae in various solutions in vitro. In the earlier experiments (type I) (Fig. 5) a 4-day-old larva of Nemeritis was dissected out from its host under distilled water (in the sorbitol experiment) or under saline (in the experiments on glycerol and sodium chloride). The larva was at once transferred on the tip of a needle to a drop of the experimental solution under liquid paraffin, where its feeding movements were observed for 9 min. In the dextrose experiment the larva was dissected out under the experimental solution in which its feeding was to be observed. In later experiments (type II) (Fig. 6), 4-day-old larvae of Nemeritis were dissected out from their hosts under distilled water. They were at once picked up in a pipette and blown out onto filter-paper, and then transferred, on the tip of a needle, to a dish of experimental solution. The anterior end of the larva was taken into a compound micropipette with an internal diameter just greater than the diameter of the parasite’s thorax. The pipette was arranged in a drop so that the parasite was wholly immersed in the experimental solution, and the parasite’s swallowing movements were observed for 2 min. In all the feeding experiments parasites that made no swallowing movements during the observation period were assumed to be injured and data derived from them were discarded.
THE INFLUENCE OF EPHESTIA ON THE GROWTH OF NEMERITIS
When eggs of Nemeritis are laid in caterpillars that are in the wandering stage of their final instar, the larvae that hatch from them spend only a short time in their first instar. The time-course of development of larvae in these hosts is shown in Fig. 1 (a). When eggs of Nemeritis are laid in caterpillars that are at an earlier stage of their final instar, the first larval instar of the parasites lasts longer, but development from the second instar onwards is as fast as it is in the parasites of late final-instar caterpillars (Fig. 1 (b)). When eggs of Nemeritis are laid in caterpillars that are in one of their first two larval instars or in the first part of their third instar, the first larval instar of the parasites always lasts very much longer than it does in older hosts (Fig. 1 (c)).
It appears, then, that the duration of the first instar of Nemeritis depends on the stage of development of its host. Similar relationships between insect parasites and their hosts have been reported in several other cases (Johansson, 1951; Mellini, 1962; Takahashi, 1957; Taylor, 1937, pp. 197-8). The first instar of Nemeritis is affected much more severely by its host than are its later instars.
Evidence from a transplanting experiment indicates that this influence of the host on the duration of the parasite’s first instar is a reversible one, acting mainly on the first-instar parasite itself, not on the egg. The experiment is summarized in Fig. 2 and Table 1. Adults of Nemeritis were allowed to lay one egg in each of nineteen small caterpillars of Ephestia (weighing less than 5 mg.). After 3 days in these hosts the eggs were dissected out from ten of the hosts and each was injected into a late final-instar caterpillar (group C). The other eggs, group B, were allowed to continue development undisturbed in their small hosts. On day 8 the parasites were retrieved from the surviving hosts. Those that had been transplanted to late final-instar hosts (group C) had developed further (Table 1): those that were retrieved from small hosts (group B) were still at an early stage of their first instar. These nine first-instar parasites of group B were then injected into late final-instar hosts. Five days later, on day 13, they were retrieved from their hosts. Five hosts had died, and in all of the four survivors the parasite had developed beyond the first instar. In other experiments small Ephestia caterpillars (each weighing less than 5 mg.) were individually parasitized and the parasites were retrieved on day 13. All thirteen of them were still in their first instar (group A). The parasites in group E were left undisturbed for 8 days in the late final-instar hosts in which the eggs were laid. Each of the parasites in group D was transferred as a mature egg from one late final-instar host to another; when they were retrieved by dissection on day 8, they had not developed as far as the parasites in group E.
Since the parasites in group C had developed further by day 8 than had the parasites in group B, it is clear that the small hosts affected their parasites in such a way as to prolong their first instar. (Evidently the operation itself did not cause the transferred larvae to grow more quickly, since the transferred larvae in group D were less advanced on day 8 than were the undisturbed larvae of group E.) Further, although this influence of young hosts on the development of Nemeritis increased the duration of the egg stage a little, its main effect was on the duration of the parasite’s first instar (cf. Figs, 1 a and 1 c). The influence of the young hosts on the parasites was short-lived; it did not prevent the subsequent rapid development of parasite larvae which had been transferred to suitable hosts (cf. groups B and A).
Late final-instar caterpillars of Ephestia support the rapid development of their parasites, whereas young caterpillars cause delay in their parasites’ development (Fig. 1). An experiment was performed to discover at which stage of the hosts’ development this change took place. Caterpillars of Ephestia were parasitized individually, and on the eighth day after parasitization the stage of development of both host and parasite was recorded (Fig. 3). Most of the parasites were still in their first instar on day 8 if the hosts were in their mid final instar or at an earlier stage; but if the hosts were at a later stage, the parasites had developed beyond their first instar. The results of this experiment indicate that the critical change in the properties of the host takes place towards the end of the mid final instar or early in stage one.
THE FEEDING BEHAVIOUR OF LARVAE OF NEMERITIS
First-instar larvae of Nemeritis, whose development is delayed in young hosts, feed more slowly than do parasites in late final-instar hosts. Delayed development of parasite larvae seems always to be associated with slow feeding. The evidence is this:
The midgut of larvae whose development has been delayed in young hosts looks nearly empty, whereas that of developing larvae can be seen to be full of opaque material. In the early larval instars of Nemeritis an occlusion separates the lumen of the midgut from that of the hindgut (Rietra, 1932), and ingested material that cannot penetrate the gut wall must accumulate in the midgut. The nearly empty gut of larvae whose development is delayed therefore indicates that they have eaten little, whereas the distended appearance of the midgut of developing larvae indicates that these have fed much more.
Developing larvae of Nemeritis feed faster in the blood of late final-instar hosts than do delayed larvae in the blood of young hosts. Four-day-old (first-instar) larvae of Nemeritis were dissected out and each was at once placed in a clear drop of its host’s blood, under liquid paraffin. They were allowed to feed, and their swallowing movements were observed and counted. These swallowing movements consist of backwards-moving peristaltic waves in the posterior region of the foregut, clearly visible through the transparent tissues of the parasite. Apparently material can pass along this region of the gut only when it is distended in a peristaltic wave; and it is clear that material does enter the midgut with each wave, because when a first-instar larva is submerged in liquid paraffin, a highly refractive droplet of paraffin can be seen to pass along the fore-gut with each swallowing movement. The peristaltic waves in the foregut always appear to be about the same size; and it therefore seems reasonable to take the frequency of these swallowing movements as an indication of the rate of feeding. The mean number of swallowing movements made in 9 min. by ten parasites from wandering final-instar hosts was 24 ±6·4; for ten parasites from mid third-instar hosts the mean number of swallowing movements was 12 ±2·5 (P = < 0·001).
An analysis of the causes of the delayed development of the parasites in young hosts therefore requires an explanation of the slow feeding of these parasites.
Changes in the feeding behaviour of the parasites are likely to be traceable to changes in the quality of their immediate environment or of their food. Since first-instar larvae of Nemeritis live free in the haemocoele of their hosts, surrounded by, and feeding on (Rietra, 1932), their hosts’ haemolymph, their feeding behaviour may be expected to be influenced by changes in the composition of their hosts’ haemolymph.
Three properties of the haemolymph of Ephestia caterpillars have been examined, and all were found to exhibit a sudden change at about the time of the critical change in the host’s influence on the development of its parasites, in or just before stage one of the final instar. The properties were these: the concentration of proteins; the concentration of amino acids and other ninhydrin-positive substances; and the concentration of all solute particles, measured in terms of the depression of the freezing point (Fig. 4).
The concentration of proteins in the haemolymph of Ephestia caterpillars was low in the third instar and in the early part of the final instar. Later in the final instar, at the time when parasite larvae are able to develop, the protein concentration reached a high value, and it fell again before pupation. The concentration of free amino acids and other ninhydrin-positive solutes was found to be inversely related to the protein concentration in final-instar caterpillars. It was lowest in stage two of the final instar, when the protein concentration was at its highest. The freezing-point depression of the blood was high in the middle of the instar, when the caterpillars were feeding, and fell steeply late in the final instar, when feeding had stopped.
From Fig. 4 it can be seen that at about stage one of the final instar important changes occur in the concentration of constituents of the haemolymph in Ephestia. Although the lowered concentration of ninhydrin-positive solutes probably contributes largely to the observed fall in the freezing-point depression (see Sutcliffe, 1963), much of the variation in the freezing-point depression probably reflects changes in the concentration of other solutes. Thus the parasites are subjected to a very complex environment in which the concentration of numerous components changes at about the same time.
When the parasites increase their feeding rate in stage-one hosts, are they responding to changes in the concentration of one or a few specific ingredients of their hosts’ haemolymph, or are they responding to some more general property of their environment? The parasites’ feeding cannot be dependent on the presence of a specific stimulant, because parasites feed quickly in artificial solutions of single solutes (Figs. 5 and 6) and in liquid paraffin. The slow feeding of parasites in young hosts might be due to inhibition of their feeding by specific substances in the hosts’ blood. This possibility has not been eliminated, but since general concentration effects have been shown to influence the parasites’ rate of feeding in vitro (Figs. 5 and 6) it seems unnecessary to postulate a specific inhibitor to explain the slow feeding of parasites in young hosts.
The coincidence in time between the fall in the freezing-point depression of the hosts’ blood and the resumption of growth of their parasites points to the possibility that the parasites’ feeding rate is governed by a general response to the concentration of solute particles in their food. This idea received further support from experiments, both in vivo and in vitro, in which alteration of the total concentration of solute particles in the parasites’ food was followed by changes in the rate of feeding and growth of the parasites.
(i) In vivo experiments
In the experiments conducted in vivo the hosts were subjected to three types of treatment that were found to raise or lower the freezing-point depression of their blood, and the effect on their parasites was examined. The treatments were: decerebration, ligature and crowding (see Table 2).
When the brain was removed from late final-instar caterpillars of Ephestia, the freezing-point depression of their blood 3 days after the operation was much higher than that in unoperated controls or in sham-operated controls. The stage of development reached by their parasites on day 8 of the experiment was correspondingly less advanced. (The presence of the 3-day-old parasite made little difference to the freezing-point depression of the host’s blood. But on day 8, whereas the freezingpoint depression of the blood of unparasitized decerebrate larvae was still high, the blood of parasitized larvae was of nearly the same freezing-point depression as that of the unoperated parasitized controls. Apparently the parasite had caused the freezingpoint depression of the host’s blood to fall, in some way that is not understood.)
When wandering larvae of Ephestia were ligatured on the first or second abdominal segment, the anterior part usually died within a few days. In the posterior part, which remained alive for much longer, the freezing-point depression of the blood on day 8 of the experiment was higher than that of the controls. The parasites grew a little more slowly in ligatured hosts than they did in unligatured hosts. (If parasites were present in the ligatured abdomen, the freezing-point depression of the blood on day 8 differed little from that of unoperated controls. Here again, the parasites appear to have lowered the freezing-point depression of the blood of unoperated hosts.)
When fifteen late final-instar caterpillars of Ephestia were crowded together in a plastic box (), in the presence of food, the freezing-point depression of their blood was a little higher on day eight than it was in isolated controls. Parasites in crowded hosts grew rather more slowly than parasites in control hosts. When caterpillars were similarly crowded in the absence of food, the freezingpoint depression of their blood did not rise; it may even have fallen. The parasites did not grow more slowly than those of control hosts; they may even have grown faster.
In all these experiments treatments that raised the freezing-point depression of the blood of host larvae caused delay in the development of their parasites. The extent of the delay seems to be related to the severity of the effect on the freezing-point depression of the host’s blood.
(ii) In vitro experiments
First-instar larvae of Nemeritis were exposed in vitro to solutions of various solutes, to see whether their feeding rate was modified by changes in the concentration of the solution on which they were feeding. In experiments of type I the parasite larvae were lying free in a drop of the solution; in experiments of type II, however, the anterior end of each larva was confined in a glass capillary tube. The results of these feeding experiments are shown in Figs. 5 and 6. The pattern of change of the feeding rate with concentration varies according to the nature of the solute and the experimental method; but in most of the solutions, over the concentration range which corresponds, in terms of thermodynamic activity, with the concentration of the blood of an Ephestia caterpillar, the parasites’ feeding rate falls as the concentration increases.
These feeding experiments are complicated by a response to tactile stimulation of the ventral surface of the head. It was noticed that parasites fed most when the ventral surface of their head capsule was pressed against a glass surface. One of the differences between experiments of types I and II is that in type I the parasites only rarely come to rest with the ventral surface of the head capsule in contact with glass, whereas in experiments of type II the larvae have this region touching the inner surface of the glass tube for most of the time. A tactile response could account for the different feeding rates of larvae, particularly noticeable in dilute solutions, in experiments of type I and type II. To test this idea, a larva was immersed in distilled water and placed alternately free in the drop, and in a tube; its swallowing movements were recorded. Table 3 shows the results of an experiment of this kind. The presence of a glass tube can make a difference to the feeding rate of a parasite in vitro. But it seems unlikely that the tactile stimulus provided by late final-instar caterpillars would be sufficiently different from that provided by young caterpillars for this response to play an important part in the control of feeding in parasites when they are in their hosts.
The initiation of rapid feeding by larvae of Nemeritis when their hosts reach stage one of the final instar could be an integrated response (involving several chemosensory endings) to the over-all fall in the concentration of ions and molecules in solution in the host’s haemolymph at that time. Even if the parasite’s feeding rate is governed in the host by a response to the change in concentration of a single solute, this is not a specific effect in that it can be engendered in vitro by similar changes in the concentration of other solutes.
It has frequently been recorded in the literature concerning delayed development in parasitic insects that the parasite’s resumption of growth coincides with the termination of diapause in the host (see Salt, 1941; Schoonhoven, 1962). This has been noted often enough to indicate that the ability to delay the growth of parasites may be a general feature of insect diapause. If this is so, an analysis of the means whereby a diapausing insect delays the growth of its parasites might throw light on the physiological basis of diapause.
Schoonhoven (1962), who studied the relationship between a tachinid larva, Eucarcelia rutilla, and its host, the geometrid Bupalus piniarius, suggested that when the host’s diapause ended the parasite larva was activated by one of the host’s hormones, possibly the prothoracic gland hormone. Similarly, Schneider (1950, 1951) proposed that the resumption of growth of the endoparasitic ichneumonid Diplazonfissorius at the end of the diapause of its host, the syrphid Epistrophe bifasciata, might be due to the parasite’s response to hormones in the host’s haemolymph. It appears from my work that the feeding behaviour of Nemeritis in Ephestia is influenced by the concentration of non-hormone constituents of the blood in such a way that the observed changes in these constituents could account for the observed changes in the rate of feeding; and the observed changes in the rate of feeding may be great enough to account for the observed changes in the parasites’ rate of growth. Therefore, in this case at least, it seems unnecessary to postulate a direct effect of the host’s hormones on the parasite. However, the growth of Nemeritis is, presumably, ultimately governed by the hormones of the host, since the host’s hormones play a part in the control of its own metabolic activities and these activities modify the composition of the haemolymph, the parasites’ food.
One of the primary aims of this work has been to find a clue that might help towards an understanding of the cause of the delay of endoparasitic insects in diapausing hosts. The delay of Nemeritis is related to the high concentration of solutes in the blood in the feeding stages of caterpillars of Ephestia; it may be that in diapausing hosts, too, parasites fail to feed because the concentration of solutes in the host’s blood is too high. Remarkably high concentrations of small-molecule solutes such as glycerol, sorbitol or trehalose have been reported in the blood of diapausing or aestivating insects (Salt, R. W., 1961; Somme, 1965; Asahina & Tanno 1964).
An incidental finding that may be of some importance is that the composition of the haemolymph of Ephestia larvae shows a consistent pattern of change related to the moulting cycle. This is only to be expected in view of the important changes that must occur in the metabolic activities of the tissues in contact with the haemolymph. Yet, despite the finding by Jeuniaux, Duchâteau-Bosson & Florkin (1961) of a somewhat similar pattern of change in the freezing-point depression of the blood of final-instar caterpillars of Bombyx, the phenomenon is not widely recognized. The mean freezing-point depression of the blood in Ephestia can vary from 0·48°C. to nearly twice that value within the final instar. If comparable changes occur in other insects, it would appear that a statement of the freezing-point depression of the blood of a species is of much greater use if the instar and the stage in the moulting cycle are specified.
I am indebted to Dr George Salt for his careful supervision of this work and for reading and commenting on the manuscript. I am grateful to the Nature Conservancy for financial support.