go to previous section of this publication

Waterston (in Uvarov et al., 1951) and Dempster (1957) have listed a number of egg-predators of Dociostaurus in Cyprus and commented on their relative importance and the effect of their predation on locust numbers. The two principal insect predators are a Clerid beetle, Trichodes laminatus Chevr. var. cyprius Reitt., and a Bombyliid fly, Systoechus autumnalis Pall. A third predator, the Tenebrionid Apontanodes globosus Reich., was discovered in 1955, but the amount of damage it causes is uncertain. In the following sections, the data on predation obtained from the analyses of egg-pod samples are discussed and a description is given of what is known of the life-history and ecology of Trichodes and Systoechus. No work was done on the bionomics of Apontanodes.

No detailed descriptions of the morphology of the predators are given in this paper, which is principally concerned with their ecology. However, drawings have been made of those stages of the life-history which have not been hitherto described and very brief descriptions of them are given. The most important of these are the early larval stages of Systoechus and Trichodes. Preserved material of these has been lodged with the Commonwealth Institute of Entomology.

The data obtained from the analyses of egg-pod samples are presented in such a way as to show the development of our ideas on predation rather than our later conclusions only. This is because the early samples were not classified as fully as the later ones and because there is still doubt about the importance of Apontanodes.

In the early stages of the investigation, it was not known what species of Bombyliid were involved (see p. 100), and it was not until 1954 that S. autumnalis was recognised to be the most important and probably the only Bombyliid predator of Dociostaurus in Cyprus. Nevertheless, for the sake of clarity, Bombyliid predation is always referred to below as predation by Systoechus.

Each year from 1951 to 1955, samples of egg-pods were examined from a number of sites, to determine the percentage destroyed by internal predators (i.e. those predators which entered the pods to consume the eggs) and in later years to ascertain how many of the unconsumed eggs remained viable until hatching. Latterly, sampling was carried out at Mia Milea in such a way as to provide an estimate of the total number of pods on the site (Dempster, 1957). The method of collection and the size of the sample varied, but in general the samples were small in the earlier years and were taken at frequent intervals, while in the later years they were large and taken less frequently. However, in 1955-1956 large samples were taken at fortnightly intervals. Most of the samples were obtained by a deliberate search for pods within the chosen area. The locust populations concerned were of low density, and pods were consequently scarce and often very difficult to find. Even with experience it was difficult to obtain 100 pods in a day, and half that number was more usual. In sampling, bare ground where pods were likely to occur was dug over and searched. In 1952 and subsequently, the loose earth was then passed through a coarse sieve. The only truly random sampling, apart from some early trials by Curry, was that carried out by Dempster at Mia Milea (Dempster, 1957).

In samples collected in July and August pods attacked by Trichodes or Systoechus could easily be distinguished. The larvae were large and readily identifiable; the adults had died out earlier and the maximum predation rate was apparently attained. There were, in addition, a few pods attacked by unknown predators which no longer remained within them. The characteristic appearance of pods attacked by Trichodes and Systoechus is shown in Fig. 19. Those attacked by Trichodes are characterised by a small entry hole in the lid, or, more usually, immediately below it, the complete destruction of all eggs and eggshells, the formation of a cell, in which the larva remains, in the basal half of the pod, and a certain amount of debris in the upper half consisting of broken fragments of the froth of the pod and pink larval frass.

Fig. 19.-The appearance of locust egg-pods (all, except C, in section).
A. A normal, healthy egg-pod, with eggs below, froth above.
B. Egg-pod attacked by Trichodes. A diaphragm shuts off the cell containing the larva. The frass is above it.
C. & D. Pods attacked and vacated, by predator originally thought to be Systoechus. C shows the external appearance of a pod with one hole, D a section of a pod with three holes.
E. Pods containing one or more Systoechus larvae (two in the pod figured). The empty egg-shells are packed around them.
F. & G. Pods attacked and vacated by Systoechus. There is a large exit-hole and an earth fill. In F the froth is apparently undamaged: in G the froth is largely destroyed.
H. Trichodes-attacked pod suffering superpredation. Similar to B above, but with no larva and an exit hole. A pod with two diaphragms and an earth fill is figured. A single diaphragm and no earth fill is more common.
J. Pods attacked and vacated by unknown predators. This is one of the commonest types in this category; it results from attack by an unknown predator following Systoechus attack. A few egg-shells remain at the base.

In pods attacked by Systoechus, there is no visible entry hole. The larvae, of which up to three have been found in a single pod, lie towards the base. All the eggs are eaten (except, occasionally, a few at the upper end, immediately below the froth), but the contents are sucked out of the egg-shells and the empty shells are packed together into a wad at the base of the pod, against which the larvae lie. The froth sometimes has the appearance of being untouched, but it usually appears somewhat damaged.

The percentage of pods attacked by these two predators varied from site to site and year to year. Figures from the analysis of samples collected from the three main sites in the latter half of the summer are given in Table 22. The total number of pods examined was low in the 1951 samples.

Table 22 Percentage infestation by Trichodes (T%) and Systoechus (B%) larvae in egg-pod samples collected from the three main sites during late summer.

Trichodes-infested pods appear soon after the locusts oviposit. The percentage predation increases until the beginning of August (see Tables 23 and 24). Systoechus-infested pods may appear in June or not until September. The percentage attacked by Trichodes can be high; in a sample from Soros, a locality at Phrenaros, collected on 18.viii.54, 62 out of 83 egg-pods (74.7%) were attacked. This was the highest figure recorded, but between 40% and 50% was not uncommon. All large samples collected after June showed at least some infestation.

Table 23 Analysis of egg-pod samples collected at Mia Milea in 1954-1955, showing the increase in predation. (In the 13-14.ix samples, the Systoechus pods were all BB; in the figure for Systoechus predation in the last sample, BB, BR and BS pods are included, but not BX. For explanation of symbols, see Table 25)

Table 24 Percentage of Trichodes-infested egg-pods in samples collected at Sotira in 1952

Towards the end of August or the beginning of September, egg-pods very different in appearance were found in the samples. There were no larvae in them and no eggs or eggshells, and the froth was badly damaged. The lid was intact. In the wall of the basal half of the pod, there were one to three holes. These varied in size inversely with their number; there was usually one large hole or three small ones. At the same time, free-living Bombyliid larvae began to be found in the adjacent soil. These pods were at first thought to be those abandoned by Bombyliid larvae and were classified as such. Their appearance is shown in Fig. 19, C and D.

Few samples were collected during the winter of 1952. In those collected during the winter of 1953, several pods were found which had no larvae inside, but which had a single rather large exit hole in the basal wall of the pod and an earth plug which filled the whole of the base. The earth plug had obviously been intruded through the hole in the wall, not washed into the pod. Above and below this plug, or sometimes only above it or below it, were the shells of the locust eggs, empty and flattened, and compacted into a wad in a similar fashion to the egg-shells in a pod containing a living Bombyliid larva. In some of these pods, the froth remained apparently intact, in others it was slightly damaged. They were classified as BS and BR respectively (see Table 25) and are illustrated in Fig. 19, F and G. It is probable that such pods had been discarded in 1951 as pre-1951 pods, because of their earth filling.

Table 25 Analysis of egg-pod samples collected at Akhna and Mia Milea in 1953-1954. (The actual numbers of pods examined are given above, the percentages below. The figure for P%. which is the percentage of pods attacked, is less than the total of the remaining columns because pods attacked by both predators are included in both T% and B%.)
Ts = Pods attacked by Trichodes alone
Td = Trichodes-attacked pods suffering superpredation
T = Total number of pods attacked by Trichodes
BB = Pods containing one or more Systoechus larvae
BS = Systoechus-attacked pods which have been vacated; froth undamaged
BR = Systoechus-attacked pods which have been vacated; froth damaged
BX = Pods attacked and vacated by predator, originally thought to be Systoechus
B = BS+BR+BX
U = Pods attacked and vacated by unknown predators.

Samples were collected on the upper site at Mia Milea in May, June, July and December, 1953, and at the end of February, 1954, before locust hatching began. An analysis of the samples is given in Table 25. It appeared probable that BX, BR and BS were all Systoechus-attacked pods and that the presence of an earth plug in BR and BS could be explained by assuming the larvae had left the pod after the soil had been moistened by the winter rains. The importance of the absence of the compacted egg-shells in BX pods was not fully realised at this time.

However, if this were a true analysis, the greatly increased Systoechus predation rate in the winter samples, in comparison with the summer samples, was difficult to explain. An alternative possibility is that BR and BS were pods which had been attacked by Systoechus larvae that had left their original pods, but if this were so, and if BX pods represented those primarily attacked by Systoechus, then even the original attack was far higher than that in samples taken from the same population earlier in the year. That they were consistent samples was suggested by the observed constancy of the Trichodes predation rate.

Further sampling in February 1954 (see Table 25) showed that the predation rate continued to increase, although no Systoechus larvae were found in egg-pods. Confirmation that there was continuing attack was given by the Trichodes-infested pods, some of which had suffered superpredation. These pods had a diaphragm, above which was the typical pink frass of the Trichodes larva, but there was a large hole in the wall of the cell (rarely more than one) and the larva had gone. Sometimes the pod contained earth, more usually not. Occasionally fragments of the skin of the Trichodes larva remained. Such pods were classified as Td (Fig. 19, H).

There was, therefore, undoubted superpredation, leaving damage of a type similar to BR and BS, which were assumed to be, in part at least, due to a second Systoechus attack. If, however, the figure for primary predation from the summer samples, when the Systoechus larvae were still in the pod, was compared with the totals for BR, BS and BX in the winter samples, on the assumption that all were due to Systoechus, the conclusion was reached that each larva must consume something of the order of five to ten pods, which, since all the eggs were eaten, seemed very improbable

The analysis of the winter samples collected at Akhna in 1953-54 gave a similar picture to that from Mia Milea (Table 25). The percentage of BR, BX and BS in the winter samples was much higher than that of Systoechus-infested pods in the summer. In the February sample, however, the percentage of Trichodes-infested pods dropped from 35.0% to 25.5%.

By the spring of 1954, it had been established that the principal Bombyliid predator (and probably the only one) was Systoechus autumnalis. The entomological collection in the Department of Agriculture at Nicosia was found to contain a number of specimens which had been collected in the autumn, principally in September. Prior to 1954, adult S. autumnalis had been observed only in early summer, i.e. in May and at the beginning of June.

Throughout the late summer and autumn of 1954, therefore, a watch was kept for the reappearance of Systoechus on the locust sites. On 20th August adults were again seen, and soon afterwards larvae in early stages of development were found in locust egg-pods. A re-examination of preserved material collected in earlier years showed that there had in fact been young larvae, in the samples collected in September and October, whose presence could not be explained otherwise than by autumn infestation; they were few and their significance had previously been overlooked.

This second period of attack offered an explanation of the greatly increased Systoechus predation rate in the winter samples. It did not, however, explain the smaller increase that occurred during the winter months or the superpredation. This was still believed to be due to a second attack by Systoechus larvae after they had left the primarily infested pod. However, this explanation was not completely satisfactory. If one assumed that a larva, after leaving the primarily infested pod, might attack either a Trichodes-infested or a healthy pod at random, i.e. that the chances of superpredation of Trichodes-infested pods were the same as the chances of Systoechus attack on the available healthy ones, it was possible to calculate approximately the proportion of primary infestations, assuming that a single Systoechus larva leaves each primarily infested pod (that is, ignoring multiple infestations). In the Mia Milea sample for December, 1953, the calculated figure was 103, exactly the number of the observed BX pods. However, in the Akhna sample for December, 1953, the calculated figure was 49, while the BX total was only 13. The results from other samples were equally confusing.

In June, 1955, egg-pod samples were collected from Mia Milea shortly after oviposition. In these, a number of BX pods were found, although Systoechus attack was only just beginning and the larvae were all young. A re-examination of them suggested that a beetle larva might have been the predator and not Systoechus as was previously thought.

Samples were collected from the upper site at Mia Milea at fortnightly intervals from the middle of July, 1955 until the beginning of February, 1956, and were examined by Dempster with this possibility in mind. On two occasions in July and August, beetle larvae were found within the pod, in one case with a few eggs remaining uneaten and in the other with none. These larvae were subsequently identified as the Tenebrionid Apontanodes globosus Reich. They were associated with the type of damage that had been classified as BX, i.e. all of the eggs were completely eaten (rarely one or two remained) and there were one or more holes in the basal wall of the pod. Free-living larvae were also found in the soil. It appeared possible that all similar damage was due to Apontanodes and in the subsequent samples such pods were reclassified as A.

The analysis of the 1955-1956 Mia Milea samples is given in Table 26, and a summary showing the percentage attack by various predators in Table 27. It will be seen that the percentage infestation by Apontanodes rose gradually to a maximum in mid-winter.

Table 26 Analysis of egg-pod samples collected fortnightly at Mia Milea in 1955-1956. (Symbols as in Table 25. A = Apontanodes)

Table 27 Summary of analysis of egg-pod samples from Mia Milea in 1955-1956 (see Table 26)

With the recognition of a separate Apontanodes attack, the development of Systoechus predation becomes clear. It will be seen from Table 26 that the spring attack occurred soon after oviposition and the larvae remained in the pods until the beginning of October. Meanwhile, the autumn attack started at the beginning of September, by which time the larvae of the spring attack were large and easily distinguishable from those newly entered. All the larvae left the pods in November, leaving pods showing the type of damage classified as BR and BS. This followed immediately after the first heavy rain of the winter, 0.41 in. being recorded at Nicosia on 31st October.

In Table 27 it will be seen that after the peak predation rate at the end of October the percentage infestation dropped slightly. This drop was paralleled by a similar slow fall in the percentage Trichodes infestation, beginning after the death of the adult Trichodes. This fall cannot be due to predator mortality, after the early summer months, for the damage is characteristic and unlikely to be overlooked. There are two possible explanations: the inclusion of pre-1955 pods in the Apontanodes totals, which by increasing the total number of pods would cause an apparent decrease in the percentage infestation by other predators, which is unlikely; or, a change in the population actually sampled, in the same way as has been suggested in discussing the number of eggs per pod (p. 73). During the course of sampling it became increasingly difficult to find pods in the limited area of the upper site at Mia Milea. Nearly 1,500 pods were collected in all. Although the whole site was sampled each time, the most likely places for pods were dug over first. These were the well-defined patches of bare ground. In the later sampling, smaller and less well-defined patches of ground were sampled to a greater extent.

This is little more than a suggestion, but a similar fall in Trichodes infestation rate compared with the earlier samples was observed at Akhna in the samples for February 1954, when great difficulty was experienced in finding pods, and again in February 1955 at Mia Milea (Table 23). In the latter case, the whole site was sampled (Dempster, 1957) and the ground in the sampling squares dug over whether vegetated or not.

It is noteworthy that the Systoechus predation rate shows no rise after the autumn attack. This suggests that there is no second attack by the larvae after they leave the primary pod, and we must assume that the Trichodes pods which suffered superpredation were attacked by a predator other than Systoechus, possibly Apontanodes.

With the discovery of Apontanodes attack, the analyses of samples from earlier years were re-examined. Fortunately, the data were sufficiently full to allow the reclassification of certain samples, including a reassessment of anomalous pods, on the assumption that all BX pods resulted from Apontanodes attack (which may not be the case). Those for Akhna of November-December, 1953 and February, 1954 and of Mia Milea for December, 1953 are given in Table 28. The data for the samples from Mia Milea in February, 1954 and February, 1955 were not full enough to allow reclassification from the original descriptions; a redistribution was made, however, on the assumption that all BX pods were attacked by Apontanodes, and the figures are given for comparison. At Akhna, the Systoechus infestation rate remained substantially the same, while that for Apontanodes rose. At Mia Milea in 1953-1954, on the other hand, the percentage Apontanodes attack remained unaltered while the percentage Systoechus attack increased. In 1954-1955, the September sample was taken before the full effect of the autumn attack could be estimated, so the increase in the February samples cannot be used for comparison with other years. The Apontanodes attack rose. It is noteworthy that in the only one of the three sets of samples in which the percentage Systoechus infestation rose markedly during the winter months, the Trichodes infestation rate remained steady. Here, the February samples were taken from well-defined bare ground only. It is clear that the problem of Systoechus secondary attack cannot be settled without further sampling designed to elucidate it, although on the basis of the data presented here, if attack occurs, it is unlikely to be heavy.

Table 28 Analysis of egg-pod samples collected at Akhna in 1953-54 and at Mia Milea in 1953-54 and 1954-55 after reclassification (cf. Table 25)
The figures for the samples for Mia Milea in February, 1954 and February, 1955 are estimated on the assumption that all BX pods were attacked by Apontanodes. The symbols are as in Table 25, except that A% is percentage attack by Apontanodes on healthy pods and A'% is percentage attack by Apontanodes including Trichodes-infested pods which have suffered superpredation. P% is the total percentage of pods suffering predation.

A calculation was made of the expected number of Trichodes-infested pods suffering superpredation in those samples for which the data were adequate on the assumption that the chances of a third predator attacking a pod already infested by Trichodes or a pod not infested by Trichodes were equal. The 19551956 Mia Milea samples were not used, Apontanodes damage, i.e. BX pods, having been found in early summer samples whereas, in other years and other sites, they were not found until the late autumn. The expected and observed numbers of Td pods are as follows:

This is not a very significant difference (t = 1.094 for d. f. 3, P = 0-3) and the data are consistent with the theory that there was a third predator attacking pods at random.

The 1955-56 Mia Milea samples differ from the others in that Apontanodes damage was found as early as June, and that the number of Td pods is very much lower than would be expected in view of the heavy Apontanodes attack.

It is probable therefore that more than one predator was involved. These are most likely to be casual predators rather than specific ones.

There remain only those pods attacked by unknown predators, which occurred in samples collected at all times of the year. These are usually empty pods without a lid or any hole in the wall (see Fig. 19). The simplest and most likely explanation of their destruction is that they were pods the lids of which had been damaged or removed and whose contents had been eaten by casual predators such as ants. Their numbers were never large. Occasionally they contained egg-shells packed together at the base suggesting that the pod had previously been attacked by Systoechus.

Briefly, then, the situation appears to be as follows: Trichodes attack follows immediately after locust oviposition and reaches a peak a month or so after the death of the adult locusts (Ts pods). The larvae remain in the pod until the following spring unless attacked by Apontanodes or some other predator (Td pods). There are two periods of attack by Systoechus, one in May or June, shortly after locust oviposition, and another in September (BB pods). After the first rains, the larvae leave the pod and remain as larvae in the soil until they pupate and become adult in June or September of the following year. They leave pods showing BR- or BS-type damage. Apontanodes attack begins in June and continues at least until mid-winter and possibly until the hatching of locusts in March. This attack may be made either on healthy pods (leaving BX, i.e. A-type damage) or on pods previously attacked by other predators (Td pods, etc.).

The identification of Trichodes and Systoechus damage is certainly correct; there is doubt in attributing all the BX type of damage to Apontanodes, because of the small number of larvae found and for the other reasons given above. The BX pods are consistent and for the most part readily identifiable, and only the assumption of some third predator makes the data intelligible. No other predator which attacks the pods below ground can be suggested. Nevertheless, damage of this type in the earlier samples cannot be attributed to Apontanodes with certainty and may be the result of attack by more than one predator.

go to next section of this publication