ABSTRACT
In multiparasitism of larvae of the moth Ephestia serecarium by the solitary ichneumonid endoparasitoids Horogenes chrysostictos and Nemeritis canescens, neither has an intrinsic superiority over the other and free competition occurs between their immature stages for possession of the host.
There is no interaction between the eggs of these species before they hatch.
When larvae of these two species are present in the same host at the same time they compete for possession of the host by one of two mechanisms, physical attack or physiological suppression.
When the larvae hatch from the eggs at the same time, neither has a clear advantage over the other and either species may win by physically attacking the other with its mandibles.
If they are of different ages, the older larva usually wins by physical attack when the difference in age is less than 40 hr. at 25° C.
If the difference in age is 50 hr. or more at 25° C. the older larva always wins in competition by physiological suppression of the younger one.
The factors which are important in determining the result of competition are: difference in age between the 1st-instar larvae at eclosion, environmental temperature and the species of host attacked, which determine the parasites’ rate of development.
A hypothesis of asphyxiation is presented with experimental evidence of its validity as an explanation of the physiological suppression of supernumerary parasitoids. It is suggested that asphyxiation by oxygen-lack can better account for physiological suppression than the theories of toxic secretion, specific inhibitor and starvation.
INTRODUCTION
Many endophagous insect parasitoids require an entire host insect for their own complete development. When such a host is attacked by two species of parasitoids simultaneously it follows that a state of competition occurs between them. The occurrence of this type of multiparasitism depends primarily upon the oviposition behaviour of the two parasitoids in response to hosts that are already parasitized; the results of such multiparasitism must depend upon the outcome of interspecific competition between the immature stages of the parasitoids within the host’s body.
In many cases of multiparasitism one species has an intrinsic superiority over its opponent and invariably destroys it, either by use of the mandibles (Pemberton & Willard, 1918a; van Steenburgh & Boyce, 1937; Ullyett, 1943; Simmonds, 1953a), or by an unspecified means of physiological suppression (Muesebeck, 1918 ; Pemberton &Willard, 1918b; Webber, 1932; Muesebeck & Parker, 1933;Parker, 1933; Willard & Mason, 1937; Graham, 1948; van den Bosch, Bess & Haramoto, 1951; van den Bosch, 1951a, b; Jenni, 1951).
More commonly, there is no intrinsic superiority on the part of either parasitoid, and free competition occurs between them, the victor completing its development and the loser dying, either as an egg or a young larva. Several suggestions have been made in the literature as to the possible mechanism of competition between such solitary endoparasitic species. In the first place, the older parasitoid is presumed to survive by eliminating the younger through starvation (Fiske & Thompson, 1909; Tothill, 1922), thus emphasizing the importance of time of oviposition as the determining factor in competition. Secondly, cases of direct physical attack by one parasitoid on another using the mandibles for fighting have frequently been recorded (Howard & Fiske, 1911; Thompson, 1923; Wheeler, 1923; Willard, 1927; Willard & Bissel, 1930; Bess, 1936; Eliot Hardy, 1938; Simmonds, 1944, 1953 b). In these cases neither competitor has an intrinsic advantage over the other and the result of competition is apparently decided by the time of oviposition.
A third suggestion is that one parasite eliminates the other by physiological suppression, either by conditioning the haemolymph of the host so that it becomes unsuitable for the development of any successor (Balduf, 1926a, b; Compère & Smith, 1927, 1932; Daniel, 1932; Cendaña, 1937; Taylor, 1937; Bess, 1939; Jenni, 1951, van den Bosch & Haramoto, 1953 ; Johnson, 1959), or by the postulated secretion of a toxic substance which kills the opponent (Timberlake, 1910, 1912; Spencer, 1926; Thompson & Parker, 1930).
Most of the published information on multiparasitism exists in papers that are primarily concerned with population studies on parasitoids and their use in the biological control of pests. Apart from a study by Smith (1929) the factors influencing the result of multiparasitism and the mechanism by which it is achieved have received little attention. This experimental study of the subject is divided into two parts: the behaviour of the adult female parasitoids in host selection and oviposition (Fisher, 1961) and the present paper, which is devoted to an experimental analysis of the factors which determine the result of multiparasitic competition when neither parasite has intrinsic superiority and of the exact mechanism by which the competition takes place.
MATERIAL AND METHODS
The hymenopterous parasitoids used in this study were ichneumon wasps belonging to the Ophioninae: Nemeritis canescens Grav. and Horogenes chrysostictos Gmelin. Both species attack mature larvae of the moth Ephestia sericarium Scott (Phycitidae), and all three species were maintained in continuous laboratory culture by methods described elsewhere (Fisher, 1961).
In the study of multiparasitism on a quantitative basis it is necessary to obtain numbers of hosts each containing one parasitoid of known age of each species. The technique of exposing a set of hosts first to Horogenes and then to Nemeritis, or of simultaneous exposure, proved to be unsuitable for quantitative work owing to erratic oviposition. Even the most favourable host-parasite ratio left some of the hosts unused by the parasites, while others received more than one egg of each species. Since anything but a 1:1 ratio between Horogenes and Nemeritis in any host would weigh heavily in favour of the superparasitizing species, measures had to be taken to ensure that there was only one egg of each species in each host used. Since preliminary experiments showed that there is no apparent interaction between the parasitoid eggs before they hatch, it was possible to attain multiparasitism of Ephestia larvae by the artificial injection of eggs of known age without endangering the survival of the eggs or altering the outcome of competition. The artificial injection of Nemeritis eggs by means of a glass micropipette has been used extensively by Salt (1955, 1956, 1957) in a series of papers devoted to the defence reactions of insects to this parasitoid. Since the eggs of Horogenes have a smaller diameter than those of Nemeritis the micropipette designed by Salt could easily be used for the simultaneous injection of eggs of both species.
Both Horogenes and Nemeritis can be induced to superparasitize Ephestia larvae very easily, so the problem of obtaining large numbers of their eggs raises no difficulty. For this, a standard exposure of three females to fifteen hosts for 4 hr. at 25°C. was used throughout the experimental work. For Horogenes, this gave a mean superparasitism of 5·12 eggs/host and for Nemeritis, 3·59 eggs/host. The superparasitized larvae were dissected 48 hr. later to obtain the eggs. After washing in two changes of insect Ringer, multiparasitism of a fresh series of hosts was attained by injecting each with an egg of Horogenes and an egg of Nemeritis. The technique of injection followed that which is described by Salt (1955).
In order to reduce the risk of bacterial infection from the host cuticle each host was washed in a 0·5% solution of sodium hypochlorite before injection, and the insect Ringer, instruments and glassware were sterilized before each experiment. Following each injection the wound was sealed with a small quantity of the commercial collodion preparation ‘New Skin’ and the injected Ephestia larvae were isolated in 3×1m. glass vials plugged with cotton wool and incubated at constant temperature until competition had taken place.
The results could be obtained in two ways. Either the host was left in the incubator until the victorious parasitoid emerged, or it was dissected after several days when the parasitoid larvae had reached the end of the first instar, by which time the outcome of competition had been decided. The eliminated larva could then be removed and mounted for examination. The second method was generally used, for it allowed a detailed examination to be made of the losing larva in each case of competition.
The advantage of artificial injection is that the age of the eggs can be varied at will, so as to vary the stage at which the parasitoids meet in multiparasitism. The age of each egg can be arranged so that a larva of Horogenes, on hatching from the egg, will meet an egg or a larva of any given age within the same host. Using this technique the effects of the relative ages of the competitors, the environmental temperature and the species of host attacked were examined individually as well as in relation to one another.
THE MECHANISM OF COMPETITION
Both Horogenes and Nemeritis readily lay their eggs in previously parasitized hosts (Fisher, 1961). Competition must therefore occur between their immature stages within a multiparasitized host, since both are solitary internal parasitoids and only one of them can develop to maturity in each host attacked.
In preliminary experiments to determine the way in which the parasite larvae competed with each other, sets of ten Ephestia larvae were exposed first to ovipositing females of Horogenes and then to those of Nemeritis, or to both simultaneously for a total parasitization period of 4 hr. They were then incubated for 1, 2, 4 or 6 days in separate test series before being dissected for observation of the result of competition. In each experiment only those hosts which contained a single parasite of each species were taken for examination. All superparasitized hosts were rejected.
Those examined 24 and 48 hr. after oviposition contained unhatched eggs of both species. Both were developing normally, containing embryos at the stages relevant to 24 and 48 hr. at 25°C. and there was no indication that either egg had exerted any influence over the other.
No effect was observed until the 1st-instar larvae of both species had successfully hatched from the egg. In the hosts dissected 4 and 6 days after oviposition only one parasite remained actively in possession of the host. The other parasite had been eliminated by being physically attacked by its opponent. Some of these losing larvae were of Horogenes and the others were of Nemeritis, thus demonstrating that neither species has an intrinsic superiority over the other.
In a second series of preliminary tests in which the attacks by the two parasites were separated by 3 days, the older larva was invariably found to be the survivor, and examination showed that it had won the competition by physiologically suppressing the development of the younger parasite.
Dissection of all the Ephestia larvae multiparasitized in the experiments, which are summarized in Tables 4–6, showed that when 1st-instar larvae of Horogenes and Nemeritis hatch from the egg at the same time, neither has a clear advantage over the other and either species may win. However, if the two are of different ages the older larva usually wins when the difference in age is less than 40 hr. at 25°C. Furthermore, if the difference in age is 50 hr. or more at this temperature the older larva always wins. This can be attributed to the fact that two mechanisms of competition occur between the same two parasite species. In the first case one larva makes a physical attack upon the other, whereas in the second it survives by physiologically suppressing the development of its opponent.
(1) Physical attack
On hatching from the egg in a multiparasitized host, the young larvae of Horogenes and Nemeritis move about actively in the haemocoele of the host by vigorous movements of the abdomen and tail. When they meet, one embeds its mandibles into the body of the other and remains in this position from several minutes to half an hour (Pl. 1 a). After they have separated there may be some leakage of body fluid from the punctured cuticle and, in some cases, the whole of the bitten larva becomes considerably shrunken. Following this the wounded larva becomes susceptible to the blood reactions of the host, encapsulation by haemocytes and deposition of melanin at the points of wounding. Parasite larvae that have been attacked in this way show patches of melanin in one or more places on the cuticle (Pl. 1b, Text-fig. 1A) that contain pairs of distinctly darker spots. The distance between the members of each pair of spots corresponds to the distance between the mandibles of its attacker. In eliminated parasite larvae examined after 4 days at 25°C. there is an aggregation of host phagocytes around the wound, and by 6 days this enlarges considerably and either encloses the whole larva in a loose capsule or forms a tubular capsule around it (Pl. 1 c).
Normally one of the competitors is eliminated within 24 hr. of eclosion, but occasionally the parasites do not meet until the second day of larval life. Then the biting reaction is the same and similar host reactions of melanin deposition and phagocytosis follow, although the phagocyte capsule does not cover the whole of the now enlarged larva (Pl. 1d, Text-fig. 1B).
When the two larvae separate, the victor resumes its active life of feeding and growth while the victim ceases to feed and hardly moves at all. Melanization and encapsulation by the phagocytes of the host follow, sealing the parasite off from the internal environment of the host and probably killing it by asphyxiation (Wigglesworth, 1959a; Salt, 1960). From the point of view of competition the actual time of death is unimportant, for from the hundreds of multiparasited host larvae dissected it was invariably found that once a parasite had been bitten by another it had lost in competition.
(2) Physiological suppression
When a parasitoid attacks a host that already contains a parasitoid larva more than 50 hr. old in the 1st instar, the younger larva ceases development on hatching from the egg, does not feed and soon becomes shrunken in size. Subsequently it may be enclosed by a subspherical capsule of host phagocytes and death follows rapidly. Compact capsules of regular outline may form about Horogenes larvae eliminated in this way (Text-fig. 1C, Pl. 1e), whereas shrinking and ‘stunting’ without feeding is more typical of Nemeritis. In some cases the younger parasitoids become suppressed as soon as they break out of the egg shell and are discovered in dissection still half enclosed by the broken chorion and surrounded by phagocytes. In no case did the larval cuticle show the pairs of melanin spots that result from physical attack by the opponent’s mandibles.
That the competing parasitoid larvae interact through the medium of the host haemolymph was confirmed by two experiments in which physical contact between the parasites was prevented by ligaturing the body of the host with fine cotton thread. This prevented the two from making direct contact with each other, but was loose enough to allow free circulation of the host blood. The injections of eggs were so timed that the older larva was at the 1st-2nd-instar moult by the time the younger one reached eclosion. Two experiments were carried out, Horogenes being the younger larva in the first and Nemeritis the younger in the second (Table 1).
In all cases the development of the younger larva had been suppressed and each was found to be dead and partially encapsulated in the manner described. In the first experiment the ligaturing of six hosts was not successful and both larvae were recovered from the posterior half of the host. Nevertheless, the Horogenes had all been filled by physiological suppression and no melanized mandible marks were found on their bodies. Thus it is apparent that the means by which the older parasitoid arrests the development of the younger is carried in the blood of the host.
Several possible explanations for this kind of inhibition have already been put forward. The older parasite may secrete a substance which is toxic to others in the same host (Timberlake, 1910; Spencer, 1926), or it may have removed from the host haemolymph some substance which results in the starvation of the younger larva (Simmonds, 1943), or it may stimulate the host blood to produce more phagocytes which, though they cannot affect the older larva, are able to encapsulate and eliminate the younger ones within the same host. Another possibility is that the older parasite renders the host blood unsuitable for subsequent parasites by its own metabolic activities; that is to say, the younger larva is affected by the salivary, excretory or respiratory products of its opponent.
One of these can be ruled out without experiment. In Horogenes and Nemeritis the inhibitory factor precedes any attack by host phagocytes, for the young larva is arrested in its growth before any phagocytes begin to collect around it. Encapsulation is an effect of suppression and not its cause.
Experimental work
The first two proposed mechanisms, the secretion of a toxic substance and the removal of a growth substance, are derived from observations that younger parasites cannot survive in the same host with an older parasite but are able to do so when removed to a hitherto unparasitized host (Simmonds, 1943). To test such a possibility the converse of Simmonds’s experiment of transplantation was performed—namely the observation of the effect of injecting blood from hosts containing 2nd- or 3rd- instar parasites into hosts containing eggs or very young 1st-instar larvae.
Two sets of fifteen hosts each containing a young 1st-instar larva were injected with approximately 0·0015 0f blood taken from hosts containing 2nd- or 3rd-instar parasites. At the same time two further sets of ten hosts were injected with blood from healthy hosts and set aside as controls. Altogether, eight of the injected hosts died after the operation, and of the surviving twenty-two, young larvae were recovered from seventeen hosts. These were alive and developing well and showed no difference in size from the control larvae.
The difficulties of transferring blood from one Ephestia to another are considerable and possibly the conditions of these experiments were too artificial. Avoiding these, a further experiment was carried out by injecting eggs into hosts containing killed 2nd-instar larvae. This was done by locating the 2nd-instar Nemeritis within its living host by strong transmitted light, manoeuvring it away from the fat body of the host and killing it by pinching it with forceps without penetrating the host’s cuticle. Single eggs of Horogenes, 52 hr. old, were then injected into fifteen hosts and left for 3 days before dissection. In the second replicate Nemeritis eggs were injected into fifteen hosts containing killed 2nd-instar Horogenes larvae. In both replicates all younger parasites were found alive and developing normally after 3 days.
These two experiments show first, that the haemolymph of the host had not been made permanently unsuitable by the older parasite, and second, that the arresting factor is connected with the presence of a living parasite. A theory of suppression by a toxic secretion is therefore untenable for Horogenes and Nemeritis. Yet a reversible change could be brought about as a result of the metabolic activities of the older larva.
(a) Salivary secretion
Tothill (1922) suggested that the elimination of supernumeraries of Campoplex pilosulus Prov. (Ichneumonidae) might be associated with salivary secretion, and Simmonds (1943) suggested that the same mechanism might occur in Nemeritis.
A histological examination of the paired labial (salivary) glands of Horogenes and Nemeritis showed that they are modified for silk production. Although they are well developed from the 1st larval instar onwards, they have no true salivary function but are solely concerned with production of silk which is stored in the lumen of each gland until it is used by the mature larva in spinning the cocoon.
In experimental work two sets of fifteen Ephestia larvae, one containing 48 hr. eggs of Horogenes, the other eggs of Nemeritis, were injected with an extract made of crushed salivary glands from 3rd-instar larvae in insect Ringer solution. No effect was observed on either group ; all parasites developed normally.
(b) Excretory products
Horogenes and Nemeritis are among those endoparasites whose larvae have no continuous lumen between the mid and hind gut until the mature 5th-instar larva emerges from the body of its host. Serial sectioning of 1st- and 2nd-instar larvae showed that the occlusion exists at the posterior end of the sac-like mid gut. The Malpighian tubules therefore open normally into the hind gut and may be excreting through the anus, though there is reason to doubt this, since larvae of Hymenoptera are known to accumulate excretory products in mate cells of the fat body (Wigglesworth, 1953). The young larvae of these parasites also have prominent rectal glands, of unknown function, which may be associated with resorption of water and uric acid. Extracts of Malpighian tubules and of rectal glands, crushed in insect Ringer and then centrifuged, of 1st-, 2nd- and 3rd-instar parasites were injected into hosts containing eggs of the competing species. Twenty-five hosts were injected in each experiment. In some of these the parasites’ development was slightly retarded in relation to the control series, but none became arrested in its growth. No pattern of inhibition and subsequent encapsulation was found in any of the experimental parasite larvae.
(c) Specific inhibitor
Since neither salivary nor excretory products appeared to have any effect on young parasitoids within Ephestia, if the older larva is secreting any kind of inhibiting subs-stance, be it an ordinary product of its own metabolism or an unknown specific inhibitor as some authors have postulated, the effect should be observable in hanging drop cultures of Ephestia blood containing both young and old parasite larvae.
In an extensive series of tests, eggs of both species which were about to hatch were placed in drops of Ephestia blood on coverslips, inverted over deep cavity slides, and incubated at 25°C. To each, a minute drop of penicillin sulphate was added (diluted to 100 units per million) to prevent bacterial growth. The first series of eggs were placed singly in blood drawn from a host containing 1st-, 2nd- or 4th-instar parasites of the other species ; in the second series the eggs were placed with five parasites of the other species in the 1st, 2nd or 4th instar. The results showed that out of sixty replicates, of all instars and blood conditions, no younger parasite became inhibited, either in blood which had previously contained a parasite, or in the actual presence of an older one, even though both lived for as long as 10 days in the same drop. The younger larvae not only hatched successfully but often developed to the 2nd instar before dying in the slowly congealing blood. These observations rule out the possibility of a specific inhibitor and salivary or excretory inhibition, all of which, were they effective, would have accumulated in the blood drop and affected the younger parasite.
Normally the younger parasite in the haemolymph of a host containing an older parasite is first inhibited in development and later attacked by the host phagocytes. In these tests where it was contained in haemolymph outside the host it is able to survive equally well with an older parasite until the drop dries up. What then are the relevant differences between the in vivo and in vitro situations? In the first instance, the activity of the haemocytes is altered in a hanging drop. After removal from the host they tend to aggregate into loose clumps within a few hours, and though remaining alive for about 48 hr., they do not attack the younger parasite and, in the few instances where they do collect round a parasite, the latter is unaffected by them. An experiment was therefore set up in which thirty young 1st-instar larvae of Nemeritis which had been subjected to 5 days of physiological suppression in already parasitized hosts were divided into three groups. The first set of ten were removed from them and put into hanging-drop cultures of fresh Ephestia blood; the second set were transferred by injection to fresh, hitherto unparasitized hosts, and the third set were left in their original hosts as controls. The first set developed normally through the 1st instar for 6 days until the drop of blood congealed. The second set regained activity and seven of them completed development and emerged as adults in 28 days, while three died in the 1st instar and were later recovered from the bodies of the emerging moths. Those left in their original hosts as controls remained suppressed and died. Clearly the host haemocytes are not responsible for the initiation of inhibition since the second set were able to complete their development and emerge as adults.
In the second instance, the hanging-drop culture method would upset any respiratory relationship between the parasite and its immediate environment, the host blood, for it would allow free exchange of oxygen and carbon dioxide between the blood and the air in the slide cavity.
(d) Respiratory inhibition
Since none of the foregoing theories of the suppression of younger parasites can be confirmed by planned experiments, and since a theory of suppression through lack of oxygen would account for the existing facts, it may well be that the older larva Utilizes all the available oxygen in the host blood, leaving the younger one retarded by partial asphyxiation, after which it is attacked by the host phagocytes. If this is so, then a decrease in the oxygen content of the host blood should suppress parasite development and an increase should lead to the survival of supernumeraries and to the removal of the suppression.
To test this hypothesis an apparatus was constructed which allowed known proportions of oxygen and air to be passed over parasitized larvae at a rate of 0·5 l./hr. at constant temperature and humidity. The gas mixture was equilibrated by being bubbled through a wash bottle of water inside the incubator before passing through 6×1 in. glass tubes containing the parasitized larvae. Gas exit from the tubes was provided by capillary valves. By this means the Ephestia larvae could be maintained in atmospheres of differing oxygen content throughout the period of each experiment.
A series of tests were carried out in which the two competing parasites in each hos differed in age at eclosion by 3, 4, 5 and 6 days in atmospheres containing 5, 10, 20 (controls in air) and 50 % oxygen.
In five experiments, each of twenty replicates, in an atmosphere of 50 % oxygen no suppression of any younger larvae was observed. On the contrary, not only did the younger larvae survive, but in those tests where the difference in age at eclosion was more than 4 days, they used the well-developed mandibles of the 1st instar to attack the older larvae which had lost this armature by moulting to the 2nd instar. In other tests where the two larvae were only 3 days apart in age, some of the younger parasites were bitten by the older ones and yet remained alive and often free from haemocytes on the 7th day after eclosion. Table 2 shows a summary of the course of multiparasitism in an atmosphere containing 50 % oxygen. Similarly in an atmosphere of 33 % oxygen the younger larvae attacked the older ones with their mandibles and thus effectively won in competition. The development of control larvae bred singly in Ephestia in normal air is shown for both the younger and older competitors. It is noticeable that the rate of development of both competitors in the experimental hosts is slower than in the controls. This is apparently not due to the increased oxygen content of the atmosphere, as separate controls have shown, but, for the older parasite, due to physical attack by the younger, and for the younger, probably due to nutritional causes as the older larvae had already converted some proportion of the host into its own tissues.
Finally, in fifteen hosts ligatured to prevent physical contact between the competing parasites, in 50% oxygen both competitors continued to develop satisfactorily for 7 days after eclosion of the younger. No younger larva became suppressed in any host.
Conversely, the effect of lowering the oxygen content of the air to 10 or 5 %, by the addition of nitrogen, was to retard the development of all the parasites. The eggs did not hatch until the 4th or 5th day after oviposition at 25°C. (normal time is 59 ± 2 hr. for Horogenes’). Many of the 1st-instar larvae did not develop and later became invested by host phagocytes. Others grew to a length of 1·0 mm., equivalent to the normal mid-1st instar, but did not feed properly and then ceased to develop. The development of such larvae compared with those reared in air is shown in Table 3. Both in super- and multiparasitized hosts the development of all eggs and larvae was retarded and they showed all the features of physiological suppression.
THE FACTORS CONTROLLING THE RESULT OF COMPETITION
Preliminary tests of multiparasitism of Ephestia by Horogenes and Nemeritis showed that neither species has an intrinsic superiority over the other in this host. The phenomenon can therefore be subjected to experimental analysis to ascertain the external conditions which control its result. In this, three factors were found to be of importance: the time of oviposition by each species, environmental temperature and the species of host attacked by both parasitoids.
(1) Time of oviposition
Both species readily lay eggs in previously parasitized hosts (Fisher, 1961). Multi-parasitism can therefore occur when a host contains two eggs, an egg and a larva, or two larvae of different species, depending upon the time of oviposition by each parasite. Since there is no mutual interaction between eggs, the experimental work is divided into two parts: competition between 1st-instar larvae, by physical combat, and between larvae in later instars and eggs or 1st-instar larvae, by physiological suppression.
(a) Competition between 1st-instar larvae
In a preliminary test fifteen healthy mature larvae of Ephestia were exposed simultaneously to one female of Horogenes and one of Nemeritis for 4 hr. Six replicates were set up and incubated at 25·C. for 4 days. On dissection of all hosts, seventyseven were found to contain larvae of both species in differing numbers, but only thirty-six of these contained a single larva of each species and could be counted for the experiment. The results showed that, although the eggs were laid at the same time (±2 hr.), Horogenes defeated Nemeritis in twenty-four out of thirty-six cases of multiparasitism. If both species had an equal chance of surviving competition a numerical result in the ratio of 1:1 would be expected. However, Horogenes won on twenty-four occasions out of thirty-six, which is equal to twice the standard error of the expected ratio (S.E. = 3·0, x2 = 4·0, P = 0·05). It therefore has a significant advantage over Nemeritis.
Horogenes’ rate of pre-imaginal development is, however, higher than that of Nemeritis. Its egg hatches in 59 ± 2 hr. at 25°C., whereas the egg of Nemeritis takes 69 ± 2 hr. (Text-fig. 2). It follows that with simultaneous oviposition the parasites are not of the same age in the 1st instar when competition takes place. If the apparent advantage possessed by Horogenes is due not to any intrinsic advantage other than its more rapid development, then adjustment of the time of egg laying so that both eggs hatch at the same time would make the result of competition attributable to chance. Each species would therefore be expected to win in about 50% of the competitions.
In six experiments, ninety-four hosts were injected simultaneously with one egg of Horogenes and one of Nemeritis 10 hr. older. Six days after injection the hosts were dissected for recovery of the parasite larvae. In eight hosts, the injection had been unsuccessful through the loss of one or both eggs in the bleeding from the wounded host after injection. Sixteen hosts had died before conclusion of the experiment. Of the seventy cases of multiparasitism, in twenty both parasites were dead and encapsulated with host haemocytes. Out of the fifty remaining in which one larva had survived, Horogenes was successful in thirty-two and Nemeritis in eighteen. The number of wins by Horogenes was not quite equal to twice the standard error from an expected ratio of 1:1 in fifty hosts (2×S.E.+25 = 32·072). Contingency table and analysis showed no significant divergence from equal survival in a 1:1 ratio. The tendency within this ratio towards survival by Horogenes can be attributed to the duration of the oviposition period.
The victor of each competition was found to be active and undamaged, showing no sign of host reaction and of length varying between 1·0 and 1·3 mm., equal to the size normally attained after 6 days at 25° C. The loser had been eliminated in each case by being bitten by the victor, sometimes in several places. Melanin deposits were found at the points of wounding and each had been surrounded to a variable extent by a capsule of phagocytes of the host blood. In no case was a dead larva found without the melanin spots marking the attack by its opponent’s mandibles. Their small size (0·55−0·75 mm.) indicated that fighting had taken place shortly after hatching, a conclusion which has been confirmed by the frequency with which newly hatched larvae of Horogenes have been dissected from their hosts still attached to their opponents by the mandibles.
To ascertain whether these results were due solely to a difference in age between the competitors in the 1st instar, a series of experiments was designed which took account of the discrepancy in duration of the egg stage between Horogenes and Nemeritis and caused one to have an excess age of 10, 20 or 30 hr. over the other in the 1st instar.
Horogenes the older parasite in the 1st instar. Four series of experiments were carried out in which the relative ages of the injected eggs were such that the Horogenes eggs hatched 10, 20, 30 or 40 hr. before those of Nemeritis. The results are shown in Table 5 and in Text-fig. 3. From these it is clear that the older Horogenes is, in the ist instar, the greater is its chance of eliminating Nemeritis and becoming the survivor of multiparasitism. When the age difference between them is nil or only 10 hr., the result of competition is complicated by a third possibility, the death of both parasites. In such cases both larvae were found to have been bitten and by the time of dissection had both become melanin spotted and partly encapsulated. That the number of such occasions fell to zero as the age difference was increased suggests that the ability to attack, and thereby to eliminate the younger larva, increases rapidly with age in the 1st instar.
When Horogenes is only 10 ± 2 hr. old in the 1st instar by the time the Nemeritis egg hatches, this age difference is sufficient for it to survive multiparasitism in 77’7% occasions of competition (x2 = 16·67, P = 0·001) (Table 5).
When the age difference was increased in favour of Horogenes as the older parasité the percentage of occasions on which it was successful rose accordingly, reaching a value of 96% when the age difference was 40 hr. When 50 or more horns older than Nemeritis in the 1st instar, Horogenes wins on all occasions of competition by physiological suppression of its opponent.
Nemeritis the older parasite in the 1st instar. By adjusting the times of oviposition of both parasites, series of Nemeritis eggs were obtained which were older than those of Horogenes by 10−50 hr. Four series of experiments, the results of which are presented in Table 6, show the effect of increasing age advantage of Nemeritis in multiparasitism. In each case the results give a highly significant divergence from the 1:1 ratio expected in equal competition. These results are in contrast with the equivalent table for Horogenes in that Nemeritis attains the 80−90 % level of dominance in competition at smaller age differences. An excess age of 10 hr. over Horogenes gives a result of 82·5 % for Nemeritis’ success, and by 20 hr. this has risen to 95 %. Nevertheless, at 25°C. the 100 % level is not reached until the age difference is about 50 hr. in the 1st instar (the same as in Horogenes).
It is therefore evident that the time of hatching from the egg is critical in deciding the outcome of competition. When one competitor is from 10 to 40 hr. older than the other, it has a statistically significant advantage but it does not have absolute superiority. However, when it has a 50 hr. advantage it wins outright in all cases of competition. Examination of the dead larvae in these latter experiments (Table 6) revealed that elimination had taken place, not by physical attack, but by physiological suppression.
(b) Competition between larvae of different instars
Observations on the oviposition behaviour of Horogenes and Nemeritis have shown that neither species completely avoids laying eggs in hosts containing advanced larval stages of the other (Fisher, 1961). The occasion may arise when a parasitoid lays an egg in an already parasitized host, and since the proportion of the pre-imaginal life spent as an egg is very small, the probability is that the second-comer will meet an advanced parasitoid larva in competition.
By using artificial parasitization the time interval between each injection was arranged so that, on hatching, the younger larva met a late 1st-, 3rd-, 4th-, or 5th-instar larva of the opposing species. In all cases the development of the younger larva was arrested at eclosion and, if they suceeded in breaking free from the egg shell, the young larvae were rapidly enclosed by haemocytes. This result was obtained regardless of the species of the older or younger competitor.
For multiparasitism between larvae of different instars the older parasitoid invariably wins in competition, since by the time it has reached the 2nd instar it is able to suppress the survival of any subsequent parasite.
(2) Environmental temperature
The larval stages of an endophagous parasitoid are necessarily affected by the external physical conditions through the medium of the hosts’ blood in which it Eves. In this period of the life history, environmental temperature is the important factor ; humidity can be disregarded when the larva Eves within the body cavity of its host. Since the relative ages of the competitors are important in deciding the outcome of multiparasitism, the rate of development, regulated by environmental temperature, may be expected to affect both the result of competition and the time after hatching from the egg at which it occurs.
Observations made on the duration of the egg stage at 15°, 18°, 25°, 28° and 30° C. for both parasites showed that Horogenes develops more quickly than Nemeritis throughout this temperature range. Time-temperature and temperature-rate curves for both species are given in Text-fig. 2. The exact time (± 2 hr.) for egg development at each temperature and its reciprocal are recorded in Table 7.
The results of experiments in which the two eggs hatched in the same host at the same time (±2 hr.) at 15°, 18°, 25° and 28°C. are shown in Table 4 and summarized graphicaEy in Text-fig. 3. It may be noted that Horogenes wins more cases of competition than does Nemeritis on a simple numerical basis. A contingency table analysis for x2 applied to the results of Table 4 showed no significant divergence from en expected ratio of 1:1, from which it is concluded that, over the viable temperature range of both species, neither has an intrinsic superiority over the other when the egg of both hatch at the same time.
The combined effects of difference in age at hatching and temperature were assessed in a series of tests in which the hatching larva of each species encountered a 1st-instar larva of the opposite species from 10 to 40 hr. older than itself. The results of these, together with the x2 test for their significance, are given in Tables 4-6. In this way it was established that the older parasite need hatch only 10 hr. before its rival for it to become the winner of the competition. At this stage elimination takes place through direct physical attack by the older parasite, but if the older parasite hatches from the egg 50 hr. or more before its successor, it achieves absolute superiority by physiological suppression. The age difference of 50 hr. was established as the minimum necessary for this effect to appear at 25° C. With the lower rate of development at 15° and 18° C. this state of dominance is not achieved until a greater age, corresponding to that reached by the 1st-instar larva at 25°C. in 50 hr. At 28°C. the rate of larval development is increased and consequently the stage of complete superiority is achieved only 30 hr. after eclosion.
The results of the combined effects of temperature and age difference in the first instar are presented together in Text-fig. 3. The outcome of competition is not fundamentally altered by environmental temperature, but the greater percentage of winnings are accorded to the older parasite at higher temperatures.
It is concluded, since at all temperatures throughout the viable range of both species Horogenes develops more quickly than Nemeritis and hatches from the egg in Advance of the latter by times varying from 4 hr. at 28°C. to 62 hr. at 15°C., that it is able to win the majority of competitions when oviposition by both occurs simultaneously.
(3) Host species
Smith (1929) suggested that the outcome of competition between two species in multiparasitism is influenced by the number of ecological niches which each species can occupy in any given environment. For endoparasitic species, the ecological niches can be considered to be the various species of host insects which each parasite attacks. Of the moths known to be parasitized by Horogenes (table I in Fisher, 1959) the following are recorded for Nemeritis’. Ephestia elutella (Richards & Waloff, 1946), E. cautella (Chittenden, 1897; Richards & Thompson, 1932), Galleria mellonella (Richmond, 1925), Achroia grisella (Thorpe & Jones, 1937). Preliminary tests carried out on the four host species showed that both parasites will attack E. elutella and E. cautella and complete their development in them at the same rate as in E. sericarium. In the Galleriinae, however, the rate of development differs for the two species. Both will oviposit in G. mellonella, but neither will develop to maturity in it. Yet both species will successfully parasitize the small wax moth Achroia grisella though with markedly different rates of development. When parasitizing Achroia the development of Nemeritis is complete but very much slower than in Ephestia. The 1st-instar moult does not take place until the nth day after oviposition, whereas in Ephestia it occurs on the 6th or 7th day. Similarly, emergence is delayed from 22−33 days in Ephestia to 32−33 days in Achroia. Horogenes, on the other hand, is not delayed in development and, as in Ephestia, emerges after 21−23 days at 25°C.
The suggestion that the rate of larval development is important in establishing the result of competition was confirmed by experiments using Achroia as host in which the difference in age between the larvae was altered. When there was no age difference at eclosion, Horogenes won twenty-three of thirty competitions (Table 8 a). When it hatched 10 hr. before Nemeritis, it won twenty-nine out of thirty-one (Table 8,a). Examination of the losing larvae from each host showed that competition took place by physiological suppression. In each case Nemeritis as the losing parasite had remained very small (0·6−0·65 mm.), had turned a deep dark-brown colour and was surrounded by a dense haemocyte capsule (Pl. 1f). The rapid establishment of physiological dominance of Achroia by Horogenes was confirmed in a third test in which its eggs hatched 34 ± 2 hr. before those of Nemeritis (Table 8 b). In all hosts but one Horogenes had attained dominance by the time the Nemeritis eggs hatched, so that all the larvae of the latter died and were encapsulated soon after hatching.
To observe the effect of difference in age in Nemeritis’ favour, eggs of Horogenes were injected into Achroia larvae at 1, 3 and 6 days after a first injection of single Nemeritis eggs. In the first test twenty hosts were injected with single eggs of Nemeritis. Twenty-four hours later they were injected with 48 hr. Horogenes eggs, so that the Nemeritis larvae were 83 hr. old when the Horogenes eggs hatched. Four days after the second injection all the hosts were dissected and it was found that Horogenes, though the younger parasite, had defeated Nemeritis in a significant majority of hosts by fighting (Table 8 c).
It appears that the normal advantage of age in the 1st instar, held in this experiment by Nemeritis, is nullified by its slow rate of development in Achroia. When the time taken by each species to complete the 1st instar is considered, it is found that, at the time of competition, Nemeritis has 8 days to develop to the end of the 1st instar while Horogenes takes only 4 days after hatching to do so. Horogenes can therefore be considered as being effectively 4 days older than Nemeritis although it hatched from the egg 24 hr. after it. This difference is sufficient to account for the dominance of Horogenes in this experiment.
In later experiments, when Horogenes was injected into Achroia larvae containing Nemeritis 3 days old, the latter achieved partial domination over the subsequent parasites (Table 8 c), though the x2 test showed no particular advantage for either species in the small number of hosts injected.
However, when Nemeritis is injected into Achroia 6 days before Horogenes it gains absolute superiority in the host and the younger Horogenes larvae are physiologically suppressed (Table 8 c).
In summarizing multiparasitism of Achroia it is clear that, as in the natural host, the outcome of competition is decided by the time of oviposition. However, the effect of the host on the result of multiparasitism depends upon its suitability for the growth of the competing larvae, since it is not only age, but also rate of development in the combatant instar of the parasites, which is important in determining which of two parasites shall win.
DISCUSSION
In this experimental work to ascertain the factors which determine the outcome of multiparasitic competition it has been found that the time of oviposition is of prime importance. However, since the actual competition does not occur until the eggs hatch, it follows that the result is also dependent on the rate of development in each species of host. In Ephestia, Horogenes wins in competition because its period of egg| development is shorter than that of Nemeritis throughout the viable temperature range of both species. This effect is emphasized in another host, Achroia grisella, to which they are differently adapted in respect of rate of development. Therefore, on strict application of Smith’s (1929) classification of multiparasitism, Horogenes is intrinsically superior to Nemeritis when their eggs are laid at the same time. But since simultaneous oviposition is unlikely to occur in the field and the difference in the duration of the egg stage is very small, for practical purposes neither has an intrinsic superiority, and the result of competition is decided solely by the time of oviposition.
In relation to their larval environment, both species develop equally well in the lepidopterous hosts present in flour mills. Any extrinsic superiority therefore depends upon the fecundity and maximum population density attained by each. Although the fecundity of both species is approximately the same, the potential capacity for increase is very much greater for Nemeritis since it is a wholly parthenogenetic species. In consequence it is superior to Horogenes to an extent which, in field populations, far outweighs any superiority the latter may show in multiparasitism.
The elimination of supernumerary larvae of solitary insect parasitoids has been known ever since work on the biological control of insect pests began at the end of the last century. Yet, although frequently mentioned in the literature, the mechanism of competition has not been fully investigated. Using the solitary parasitoids Horogenes and Nemeritis it has been demonstrated here that competition by both physical attack and physiological suppression occurs between them. The two mechanisms occur in superparasitism and multiparasitism by both species and their appearance depends on the difference in age between the competing larvae in the 1st instar.
In the past competition has been reported to occur either by physical attack or by inhibition but not, so far as I am aware, by both reactions in any one species. The fact that they both occur in Horogenes and Nemeritis leads me to suggest that this possibility may have been overlooked by former authors, whose deductions were based on the dissection of field-collected material of unknown age. An example of this is Simmonds’s (1943) statement that physical combat in Nemeritis is wholly insignificant in comparison with physiological suppression. Because, in his experiments, the oviposition period was not regulated, he was unable to distinguish competition between larvae of the same age from that which occurs when they are of widely different ages.
Physical attack is well known in the competition of supernumerary larvae of solitary internal parasitoids. So also is the subsequent phagocytic reaction of the host blood cells which surround the eliminated larva. Physiological suppression, however, has received very little attention, usually because it is mentioned only as a cursory observation in papers primarily concerned with the biological control of the host insect. The experimental approach taken here has shown that several of the theories by which inhibition or suppression are supposed to take place are untenable for Horogenes and Nemeritis. However, of all the possible effects of the metabolic activities of the first parasite investigated, only a gross change in the oxygen content of the host has shown a definite effect on the outcome of competition. The striking way in which the normal result of competition can be reversed by raising the oxygen in the air surrounding the host to 33 and 50 % indicates that this is very probably the key to the whole mechanism of physiological suppression.
The majority of internal parasitoid larvae make no direct contact with the tracheal system of their hosts, and rely on obtaining oxygen direct from the haemolymph in which they lie. Within the first 48 hr. of hatching from the egg the growth of the first instar larva is comparatively slow (see table II in Fisher, 1959), but thereafter growth is very rapid and the larva passes through one instar each day. It may be assumed then that the oxygen requirements of the larva increase accordingly, and on these grounds it is postulated that by the time a 1st-instar larva has lived for 50 hr. in a host at 25° C. it has reached a stage where it is utilizing all the available oxygen in its host’s blood. The following hypothesis is therefore put forward : that the physiological suppression can be directly attributed to an alteration in the oxygen or carbon dioxide content of the host haemolymph, due to the increased respiratory metabolism of the oldest parasitoid larva present.
There are many references to physiological inhibition of supernumerary parasite larvae in the literature which have never been satisfactorily explained. It is suggested here that the theory of toxic secretion originally postulated by Timberlake (1910,1912) and subsequently used by Spencer (1926) and Thompson & Parker (1930) can better be explained by lack of oxygen in the host’s blood. Similarly, it can better account for suppression than the unproven theory of inhibition through starvation postulated by Tothill (1922) and later used by Taylor (1937). The hanging-drop culture experiments described here have demonstrated that, for Horogenes and Nemeritis at any rate, the younger larva does not suffer from starvation or a toxic secretion since it can develop normally and moult to the 2nd instar in the presence of an older larva. The subsequent development of suppressed larvae when they are transferred to unparasitized hosts has also shown that physiological suppression is not irreversible, and that this is strong evidence that the inhibition is metabolic and not the result of poisoning by a toxic secretion.
Many authors have observed that physiologically suppressed eggs and larvae show signs of disorganization such as stained and spotted eggs (Labeyrie, 1959), or a granulated appearance of eggs and larvae (van den Bosch & Haramoto, 1953). These have been attributed to the cytolysing action of the specific inhibitor produced by the older larva. But it is more probable that this is merely the ordinary process of cytolysis of the parasite after it has been killed by prolonged oxygen deficiency and accompanying inability to move and feed normally. Van den Bosch & Haramoto also observed that the supernumerary eggs were not killed immediately. Often the larvae hatched from the eggs before becoming suppressed, though in some cases the eggs did not hatch even when their contained embryos were fully developed. This, too, points to a metabolic inhibition rather than a specifically toxic secretion which would be expected to kill the supernumerary eggs immediately they were laid.
Oxygen-lack could also account for the suppression of gregarious larvae which has been observed when large numbers of eggs are laid in the same host (Parker, 1931; Hamilton, 1935). Supporting evidence for this theory of suppression by lack of oxygen may be gained from a number of seemingly diverse observations. Any increase in the demand made upon the host blood for oxygen would be expected to affect both younger and older parasites proportionately. Since this results in the total suppression of growth of the younger larva it will also be expected to have at least a slight retarding action on the development of the older. Compère & Smith (1927) noted that the younger chalcid larvae they were investigating were eliminated by a factor which was also inhibiting to the survivor. Retardation of development has been noted as a feature of increasing superparasitism by Simmonds (1943) and also in the present work (cf. Table 2).
The possibility of respiratory inhibition was at least mentioned by Pemberton & Willard (1918a) when they wrote ‘the death of the Opius or Diachasma larvae results usually from starvation, or suffocation, or possibly by the absorption of toxic excretions of the Tetrastichus larvae’. Compère & Smith (1932) noticed that in the elimination of supernumerary parasites of Pseudococcus gahani ‘the phagocytic action is a secondary process acting upon organisms that have been killed by some obscure defensive host reaction’. They had observed correctly that elimination preceded phagocytosis, but failed to see that this was a result of parasitic competition and not, as they suggest, a defensive host reaction. Bess (1939) recognized that the two effects were separate, that the host reaction of phagocytosis of living parasites occurred only to those which were supernumerary, and that this phagocytosis did not occur until the eggs were ready to hatch. Both of these examples separate the actual suppression from attack by host phagocytes and the latter rightly associates them with supernumerary parasitism in these hosts which do not normally show reactions to single parasitoids of these species.
A recent observation by Lewis (1960) of a correlation between the difference in age of gregarious parasitoids in multiparasitism of spruce budworm larvae can easily be explained by respiratory suppression. He found that the longer the time interval between first attack by Apanteles fumferanae and the second attack by Glypta fumi-feranae, the greater was the percentage of Glypta which became suppressed. On the theory advanced here, the older the Apanteles larvae were, the greater their oxygen requirements became, and consequently an increased percentage of the younger Glypta larvae became suppressed.
There have been other suggestions of respiratory inhibition in parasites which may have some bearing on this explanation of physiological suppression. Keilin (1944) observed that dipterous parasites are considerably modified by their need for oxygen and stated that they do not grow or feed actively until contact with air is established. It has also been suggested that parasitoid larvae surrounded by phagocyte capsules in immune hosts are ultimately killed by asphyxiation (Muldrew, 1953; Wigglesworth, 1959 a ; Salt, 1960). There have also been indications that extremely localized sensitivity to oxygen lack may occur in trecheóles (Thorpe, 1936; Locke, 1958) and in epidermal cells (Wigglesworth, 1959b).
Clearly the sensitivity to oxygen-lack in fluid environments is important at the cellular and supra-cellular levels and there is sufficient evidence presented here to suppose that deficiencies in this respect, due to the metabolism of a well-developed parasite, are sufficient to suppress the development of supernumerary parasite larvae in the same host.
ACKNOWLEDGMENT
I should like to thank Dr George Salt, F.R.S., for his encouragement and supervision of much of this work and for reading the manuscript. I am also grateful to the Director and staff of the Entomology Research Institute for Biological Control at Belleville, Ontario, where part of this work was done, for their hospitality, and to the Agricultural Research Council of Canada for their financial support.
References
EXPLANATION OF PLATE
Competition by physical attack in which one Horogenes larva embeds its mandibles into the cuticle of the other (× 60).
First-instar Nemeritis larva eliminated by physical attack and showing melanin at points of wounding and partial encapsulation (× 60).
First-instar Horogena larva eliminated by physical attack, showing melanization and a tubular haemocyte capsule (× 40).
First-instar Nemeritis larva, 6 days old, melanized but only partly encapsulated after physical attack 48 hr. after eclosion (× 30).
Physiologically suppressed ist-instar Horogena larva 3 days after eclosion (× 50).
First-instar Nemeritis larva physiologically suppressed by Horogenes in Achroia grisella (× 80).