SOME COMPONENTS OF COMPETITION IN
PARASITOID WASPS
by
Laura Mary Ridout
October, 1978
A thesis submitted for the degree of Doctor of
Philosophy of the University of London and for
the Diploma of Imperial College
Department of Zoology and Applied Entomology,
Imperial College Field Station,
Silwood Park,
Ascot
Berkshire 2
ABSTRACT
Experiments within and between four species of parasitoid wasps were run in the laboratory. The species used were the Ichneumonid Nemeritis canescens, and the Braconid Bracon hebetor, two parasitoids of the Indian meal moth Plodia interpunctelZa, and two aphid parasitoids, Diaeretiella rapae and Aphidius matricarias, Braconids attacking the peach-potato aphid
Myzus persicae.
Diaeretiella and Nemeritis show a decline in searching efficiency as the density of conspecifics increase, as does Bracon when searching for healthy or paralysed- hosts. In Diaeretiella, the proportion of females in the offspring declines as the initial density rises. Diaeretiella,
Aphidius and Nemeritis larvae compete predominantly by contest competition.
In Bracon there is no overall trend in mortality with density, but the lengths of resulting adult wasps is reduced as the number of eggs developing on one host is increased, suggesting that some scramble competition is occurring. The distribution of Bracon pupae per host becomes signi- ficantly more clumped, and the mean pupae recovered per parasitized host increases, as the parasitoid density increases. This may be due to a decrease in egg predation by active host larvae.
The contribution of behavioural effects to intraspecific competition is examined for Nemeritis canescens. Under the conditions used here, the percentage of time spent searching does not decline with density. This is due to a complex nggA,j4ve feedback system operating principally via walking. It is suggested that the decline in Nemeritis searching efficiency is due to eggs lost between expulsion from the ovipositor and safe arrival in a healthy host, due to the defence reactions of the host larvae. 3
The overall effects of the presence of a competing species are
examined and discussed. An increasing density of Aphidius causes a decline in the proportion of females in the Diaeretiella offspring. The number of
Nemeritis offspring and its searching efficiency, decline with the density of Bracon. A Nemeritis larva cannot survive in a host also attacked by
Bracon. Intraspecific competition in Nemeritis is also affected by the presence of a constant number of Bracon. The number of offspring, and searching efficiency relationships are shifted downwards, and depressed disproportionately at high Nemeritis densities.
The number of hosts paralysed by Bracon is stimulated by the presence of an increasing number of Nemeritis, although the efficiency with which
Bracon finds paralysed hosts for oviposition declines with Nemeritis density. The proportion of females in the Bracon offspring declines simultaneously.
Encounters with Bracon individuals have a similar effect upon the behaviour of a Nemeritis as encounters with a conspecific (i.e. they can cause changes in behaviour, and affect the direction of change), but do not explain the decline in Nemeritis searching efficiency with Bracon density. 4
TABLE OF CONTENTS
Page
ABSTRACT ...... 2
INTRODUCTION 6
CHAPTER ONE. THE BIOLOGY OF THE PARASITOIDS STUDIED AND THEIR
HOSTS ...... 10
1.1. The Peach-potato aphid Myzus persicae (Sulz.) and its
parasitoids Diaeretiella rapae (M'Int.) and Aphidius
matricariae (Hal.) ...... 10 1.1.1. Myzus persicae (Sulz.) ... 10 1.1.2. Diaeretiella rapae (M'Int.) ... 16
1.1.3. Aphidius matricarias (Hal.) ... 20
1.2. The Indian meal moth Piodia interpunctella (Hubner), and
its parasitoids Nemeritis canescens (Gray.) and Bracon
hebetor (Say) ...... 22 1.2.1. PZodia interpunctella (Hubner) 22
1.2.2. Nemeritis canescens (Gray.) ... 26 1.2.3. Bracon hebetor (Say). ... 30 CHAPTER TWO. INTRASPECIFIC COMPETITION: AN INTRODUCTION ... 33
CHAPTER THREE. COMPETITION WITHIN SPECIES I ... 54
3.1. Introduction ...... 54
3.2. Aphid Parasitoids ...... 54
3.2.1. Materials and Methods 54
3.2.2. Analysis of Results and Discussion ... 58
3.3. Meal Moth Parasitoids ...... 71
3.3.1. Materials and Methods 71
3.3.2. Analysis of Results and Discussion ... 76
3.4. Summary ...... 106
5
Page
CHAPTER FOUR. COMPETITION WITHIN SPECIES II. BEHAVIOURAL
OBSERVATIONS ON NEMERITIS CANESCENS ... 109
4.1. Introduction ...... 109
4.2. Materials and Methods ...... 109
4.3. Analysis of Results and Discussion ... 111
4.4. Summary and Conclusions ... 154
CHAPTER FIVE. INTERSPECIFIC COMPETITION: AN INTRODUCTION ... 156
CHAPTER SIX. COMPETITION BETWEEN SPECIES I. 190
6.1. Introduction ...... 190
6.2. Materials and Methods ...... 190
6.3. Analysis of Results and Discussion ... 191
6.4. Summary ...... 221 CHAPTER SEVEN. COMPETITION BETWEEN SPECIES II. BEHAVIOURAL
OBSERVATIONS ON NEMERITIS CANESCENS IN THE PRESENCE
OF BRACON HEBETOR ... 223
7.1. Introduction ...... 223
7.2. Materials and Methods ...... 223
7.3. Analysis of Results and Discussion 225
7.4. Summary and Conclusions ...... 260
CHAPTER EIGHT. GENERAL DISCUSSION 262
... ACKNOWLEDGEMENTS ...... 272
REFERENCES ...... 273
APPENDICES ...... 311
...... PLATES ...... 399 6
INTRODUCTION
The primary objective of this study was to investigate competition at the third trophic level i.e. at the predator or parasitoid level, placing special emphasis on behavioural interactions between individuals from different species. Considerable knowledge of conspecific interactions was thought necessary for comparison, and to give an idea of the ways in which individuals of a species may react towards others. For the present, competition will be taken to mean the use of a common resource such that there is a detrimental effect upon one or both users (individuals or species) involved.
Investigations of interspecific competition of one sort or another, from strictly controlled and replicated laboratory systems, to simple field observations have been undertaken for a wide variety of plant and animal species. They range from studies on planktonic algae and Protozoa
(Sykes, 1974; Tilman, 1977; Gause, 1934; Vandermeer, 1969), yeast and bacteria (Gause, 1932; Levin et al., 1977) through triclads, molluscs, crayfish, copepods, crabs, spiders and mites, and starfish (Lock- and
Reynoldson, 1976; Branch, 1976; Bovbjerg, 1970; Hebert, 1977; Bach et al.,
1976; Uetz, 1977; Croft and Hoying, 1977; Menge and Menge, 1974) to many higher plants and animals. Fish, frogs, salamanders, lizards, rodents, birds and flowering plants (Robertson et al., 1976; Wilbur, 1972; Fraser,
1976a, b; Lister, 1976; Levin and Anderson, 1970) have also attracted study.
Among the insects, the most exhaustive competitive studies have been carried out on flour and grain-eating beetles, usually Tribolium species
(Coleoptera: Tenebrionidae) (e.g. Park, 1948; Birch, 1953; Crombie, 1945;
Nathanson, 1975) and on Drosophila species (Diptera: Drosop hilidae) by a 7 number of authors (e.g. Ayala, 1971; Budnik and Brncic, 1974; Futuyma,
1970; Merrell, 1951; Ranganath and Krishnamurthy, 1975; Richmond et al.,
1975; Wallace, 1974). A similar series of experiments have been done by
Fujii and Utida on bean-weevils, Callosobruchus species (Fujii, 1967,
1968; Utida, 1953, 1961).
It is striking that most of the entomological examples of controlled laboratory experiments have been performed on insects belonging chiefly to the second trophic level. (There are also a number of mostly field- based examples; the controversy over the Erythoneura leafhoppers (Ross,
1957, 1958; Savage, 1958; McClure and Price, 1976), Rathcke's Study on stemborers (1976) and some interesting recent work on bees by Heinrich
(1976) and Johnson and Hubbell (1974, 1975)).
At first these experiments were concerned with testing Cause's hypothesis that two species cannot coexist for any length of time sharing the same limited resource. Later it became clear that this approach was too simplistic; and the emphasis was shifted to a consideration of how organisms manage to avoid severe competitive effects when they appear to be sharing a limited major resource, and how much real sharing can be tolerated. The mechanism of the interaction (e.g. by exploitation or interference etc.), as far as it is known, is sometimes mentioned in passing. However, there is a great deal of scope for a closer look at such mechanisms. In many cases, more than one method may be operating, and, where one species is better in one way and another in a different way, both may persist. Zwglfer (1971) gives some neat examples of such
"counter-balanced" competition.
On the whole, very little has been done on the behavioural inter- actions between adult heterospecific insects, although' behaviour which 8
could be labelled "contest" has been demonstrated between conspecifics in
insects further up the food chain (Hassell, 1971b). This raises the
question of interspecific encounters at higher trophic levels: do insect
predators and parasitoids exhibit behaviour along the lines of vertebrate
territoriality (e.g. as in cichlid fish (McKaye, 1977), chipmunks
(Meredith, 1977) and blackbirds (Orians and Collier, 1963))? Is there a
significant change in competitive interactions as one passes up a food
chain (due perhaps to an increasingly sophisticated behavioural repertoire
necessary to locate and utilize prey)? How comparable are the mechanisms
that operate in inter- and intraspecific interactions?
Studies that are available on interspecific interactions between
insects at the third trophic level are mostly concerned with the wisdom of
introducing further parasitoids or predators into the field to control pest
populations. Flanders (1965, 1966), Zwōlfer (1971) and DeBach and Sundby
(1963) spring to mind most readily. These studies often included lab-
oratory work on the acceptability'of hosts after they had been attacked by
other parasitoids, and the outcome of an interspecific interaction between
larvae within a single host. Further examples of this kind of work are
Vinson's experiments on tobacco budworm parasitoids (1972) and Ables and
Shepard's paper on the parasitoids of housefly pupae (1974). In addition
to this kind of investigation, Fisher (1961a, b; 1962) ran long term experiments using Nemeritis canescens and Horogenes chrysostictos, ichneumonid parasitoids of Ephestia sericarium. Horogenes died out after about eleven weeks, leaving the other two species coexisting. Utida
(1952, 1961) has demonstrated that a bean weevil and two of its parasitoids may coexist, because one species of wasp is more efficient at high host densities and the other is better adapted to low densities. 9
It is clear that there is a shortage of information on the behavioural interactions between adults from different species of insect. The invest- igations reported here were an attempt to begin to fill in this gap.
The method adopted was to run a series of short term experiments to look at the main effects of competition within and between four species of wasp: Diaeretiella rapae and Aphidius matricarias parasitic on the
Peach-potato aphid Nhyzus persicae, and Nemeritis canescens and Bracon hebetor parasitic on the Indian meal moth Plodia interpunctella. In addition, fairly detailed observations were made on the behaviour of
Nemeritis and Bracon, in an attempt to compare the behavioural reactions to encounters with conspecifics with those following encounters with heterospecifics; and relate these to the overall effects of competition. 10
CHAPTER ONE
THE BIOLOGY OF THE PARASITOIDS
STUDIED AND THEIR HOSTS
1.1. The Peach-potato aphid Myzus persicae (Sulz.) and its parasitoids
Diaeretiella rapae (M'Int.) and Aphidius matricariae (Hal.)
1.1.1. Myzus persicae (Sulz.)
The Peach-potato aphid, Myzus persicae (Sulz.) (Hemiptera-Homoptera,
family Aphididae) is one of the most cosmopolitan and polyphagous of
aphids. It attacks a wide variety of field crops, and for this reason has
been extensively studied. Since it is the subject of such a vast quantity
of literature, it would be fruitless to attempt to review all of it here.
For the purposes of this study, a brief account of its life history, and
some aspects of its behaviour are useful. Its economic importance and the
control measures adopted are mentioned briefly. For those seeking more
detailed information, Van Emden et al. (1969), and, to a lesser extent,
Kennedy and Stroyan (1959), and the references therein, would be useful
starting points.
Line drawings of winged and wingless adult females (called alates and
apterates respectively) are shown in figure 1.1. The alates may be up to
2.5 mm long, while the apterous forms are smaller, 1.5 - 2.00 mm. They
are usually apple-green in colour, but may be found in a variety of other
shades such as bright pink, yellow, olive green or even grey. There are
darker markings on the dorsal surface of the abdomen in the alates. In
figure 1.2. a diagrammatic summary of the life cycle is given. All the
forms in this figure, except the sexuales, bear live young, and are hence sometimes called "viviparae". These same forms, (excluding the sexuparae) 11
a) Alate x13
b) Apterate x19
Figure 1.1. Line drawings of adult female hlyzus persicae.
(from drawings by Franklin in Edwards and Heath, 1964)
Figure 1.2. Summary of life cycle of Myzus persicae, the peach-potato aphid. (All of these forms are female aphids,
except where specified otherwise).
Overwintering as adults sexuparae (= gynoparae) or nymphs during mild winters alate and apterous
(alate or Migration apterous) Many Early start in Alienicolāe sexuales generations (alate or (winged (depending on apterous) Many Spring males and weather) generations. females) Dispersal over SECONDARY HOSTS secondary hosts (A wide variety)
WINTER SPRING SUMMER AUTUMN
Overwintering Give rise Migration eggs --*fundatrices to PRIMARY HOSTS (viviparous, (mostly Peach apterous, sexuales Prunus persica) parthenogenetic) winged and citrus? Migration wingless females (= oviparae) fundatrigeniae (viviparous, apterous Alate mating oviposition—+ Many migrantes generations (2nd-3rd generation from eggs) 13 are also known as virginoparae, Since their offspring never mate. Males appear only on the secondary hosts at the end of the summer.
The major primary host is the Peach tree Prunus persica, although
Myzus may utilize other, closely-related trees (Broadbent, 1949). An extremely wide variety of secondary hosts, from about 30 different families, may be attacked, including many major field crops, such as cereals, grasses, rape, kale, mustard, potatoes, mangolds, turnips, swedes, parsnips, spinach, cabbage, cauliflower, Brussels sprouts, lettuce, asparagus, beans, sugar beet, sugar cane and tobacco (Edwards and Heath,
1964; Van Emden et al., 1969). It may also attack chrysanthemums, and pansies (Scopes, 1970; Batra and Kumar, 1962). As it is not a highly gregarious species, it seldom reaches the high densities commonly found in aphids such as Brevicoryne brassicae. Consequently, it is of economic importance not so much because of the damage caused by feeding, but because it is an important vector of plant viruses. It has been known to transmit over a hundred such diseases, including potato viruses Y and A, potato leaf roll, and the viruses causing sugar beet yellows (Edwards and
Heath, 1964; Van Emden et al., 1969; Berry and Simpson, 1967; Heie and
Petersen, 1961; Watson, 1940; Heathcote et al., 1965).
Chemical control measures have met with partial success at their best
(Van Emden et al., 1969; Galley, 1974). The development of Myzus strains, which are resistant to organophosphorus insecticides, has been reported
(Baker, 1977; Dunn and Kempton, 1977). There have been attempts to breed strains of plant resistant to aphid attack. Where this has met with some success, the mechanism of such resistance is largely unknown (Lowe and Russell, 1974). In Brussels sprout varieties, waxy forms are more resistant to Myzus than are the glossy forms (Way and Murdie, 1965). 14
Irish elegance, the variety used in this work, is a glossy. Cauliflower
and turnip seedlings, which have been sprayed with a wetting agent are
more susceptible to Myzus attack (Heathcote and Ward, 1958). Some
varieties of tobacco are more resistant because of a toxin in the leaf
hairs (Thurston and Webster, 1962).
As indicated by Van Emden et al.. (1969), there are a large number of
insects which will eat or parasitize Myzus readily, given- the opportunity
to do so. However, since Myzus causes damage at low densities by
disease transmission, it is unlikely that control by natural enemies in
the field could ever be adequate to prevent economic damage. In glass-
house crops of chrysanthemums, where the sheltered conditions are very
favourable to Myzus population growth, the aphids have been successfully
controlled by introducing Aphidius matricarias mummies at the start of the
growing season (Scopes, 1970). According to Mackauer (1969), the para-
sitoid Ephedrus persicae can control Myzus in glasshouses, provided that
the temperature is above 18.3°C. He also suggested that DiaeretielZa rapae would also be capable of such control, but there is no evidence for
this so far. Meier (1966) suggested that aphids can be controlled by aphidophagous insects provided that there are large enough numbers at the start of the season to allow the numbers of the natural enemies to build up sufficiently. So far there are no reports of the successful control of
Myzus by natural enemies in the field. According to Dunn (1949), the combined effect of natural parasitoids and predators exerted on Myzus pop- ulations on potatoes occurred too late in the season to control aphid numbers. Hagen and Van den Bosch (1968) indicate that the impact of parasitoids on Myzus in the field is not clear, although Anthocoris spp.
(Heteroptera, family Anthocoridae) can control populations on sugar beet. 15
The behavioural response of Myzus to the host plant has also been the
subject of some investigation. It has been clearly shown that migrating
alates alight on host and non-host plants with equal frequency, so that
differential re-take-off is responsible for the redistribution of alates
over host plants. The migrants are more attracted to yellow, when
alighting, and will probe more readily in the presence of yellow light.
Myzus is a phloem feeder, the stylets reaching the seive tubes by intra- or intercellular routes. It takes about 15 minutes for an aphid to reach
the phloem tubes, and another 5 or 6 minutes before food uptake begins.
On crucifers and potatoes, Myzus feeds mostly on senescing leaves. In chrysanthemums the young leaves are preferentially attacked; and in strawberries, fodder beet and tobacco, the young leaves and flowers are preferred (Van Emden et al., 1969). In tobacco, aphid densities may be controlled by pinching out the youngest leaves, three times a year. In the cultures maintained for this project, it was noticeable that the aphids grew and multiplied more rapidly on the very young or very old
Brussels sprout leaves. At 15°C, a generation of Myzus takes 8 - 15 days to reach maturity, depending on leaf condition.
Edwards and Heath (1964), state that Myzus is very sluggish in its movements when disturbed. However, although not so prone to fall off the plant as some aphids, they do become fairly active when attacked by a parasitoid. The aphid under attack responds with a violent kicking of the back legs; a behaviour pattern which quickly spreads to other aphids on the same leaf. Some individuals may cease feeding and begin to walk away.
This may be due to mechanical vibration of the leaf, set in motion by the reaction of the first aphid. It seems more likely, however, that some kind of alarm pheromone is involved, as leaf movements due to air currents apparently do not produce the same response. 16
1.1.2. Diaeretiella rapae (M'Int.)
DiaeretieZla rapae (M'Int.) (Hymenoptera: Braconidae, family Aphidiinae)
is a small black parasitoid wasp. A line drawing of an adult female is
shown in figure 1.3. A fairly detailed description of the adult may be
found in Stary (1960). It is usually about 2 mm long from the head to the
end of the abdomen (not including the antennae). It has a very wide
distribution, probably occurring all over the world. Among aphid para-
sitoids, it is nearer the polyphagous end of the spectrum. It will
readily attack aphids of the genera Brevicoryne, Myzus and Hayhurstia,
and will occasionally take aphids of the genera Brachycaudus, Schizaphis
and Sitobium (Stary, 1964). Its preferred host is Brevicoryne brassicae
with Myzus persicae as second best. However, even when off erred these two
species in equal proportion, Brevicoryne is preferentially attacked
(George, 1957; Hafez, 1961; Sedlag, 1958). According to George (1957), it
occurs only on Brevicoryne in the field, and will not attack Myzus when
living on potato plants. However, Hafez (1961) recorded a few from Myzus
on Brussels sprouts in the Netherlands. It has been widely recorded from
Myzus in the field in Hawaii and the U.S.A., where it is also recorded from Macrosiphum soianifolii, another potato aphid.
Between the years 1954-7, Sedlag (1958) found that Diaeretiella was responsible for 66.2% of the parasitism on Brevicoryne, but only 6.3% of the parasitism on Myzus persicae. In India, however, Atwal et al. (1969) found that Diaeretiella caused 83-97% of Myzus mortality in 1966-7.
While Diaeretiella may be of minor importance in the suppression of
Brevicoryne in England, Holland, S. Africa and Australia, the impact of parasitoids on Myzus in the field is not clear (Hagen and Van den Bosch,
1968), although Mackauer (1969) suggests that Diaeretiella could be of use as a control agent of Myzus. Unfortunately, it has been shown that, in 17
A
B1
132
Figure 1.3. A. Dtaretiella rapae female (x 25). (After Spencer, 1926)
Bl. Forewing of Aphidius species (x 43) . Arrow shows the distinguishing intermedfan vein (from a specimen).
B2. Forewing of Diaeretielia rapae (x 43). (from a specimen) 18
some cases, hyperparasites prevent parasitoid control (Hafez, 1961). In
1960, Hafez (1961), found that, out of 8,000 mummies only 16% yielded
Diaeretielia, while 63% produced hyperparasites.
After emergence, the adults are ready to mate as soon as they are dry.
According to Read et aZ. (1970) the males are attracted to the females by
the odour they emit. However, Spencer (1926), and casual observations in
the laboratory suggest that the males do not notice the females until they
are within 7-10 mm of them. Probably, pheromones are responsible for long
range attraction, with visual stimuli taking over at short range. Having
recognised a female of his own species, the male then approaches her from
behind, vibrating his wings rapidly. He mounts dorsally, curling the tip of his abdomen round hers, beating her antennae with his all the while.
This appears to keep the female stationary. A male attempting to mate a female larger than himself cannot reach the female's antennae, and she moves off, dislodging him as she goes. Males will mate several times; females only once.
The adults feed readily on honey and sugar solutions, and honeydew
(Hagen and Van den Bosch, 1968). There is apparently no evidence of such feeding in the field, presumeably they take honeydew. Van Emden (1962) has reported a general increase in activity in all parasitic Hymenoptera in the presence of flowers, especially those of the family Umbelliferae.
Water trap catches were higher once the flowers had opened. This suggests that some of these insects, at least, may be taking nectar. Unfed females are still able to oviposit, but their longevity is greatly reduced, although this increased mortality may be due to dehydration rather than sugar shortage. Sedlag (1959) found that the maximum longevity at 20°C
(about 5 days) was obtained at a relative humidity of 86.5%, when the 19 parasites were fed with sugar solution. However, Hafez (1961) managed to keep his parasitoids alive for one to two weeks. a
Although newly emerged females are ready to oviposit almost immediately, they must often disperse before they can find hosts to attack.
Read et al. (1970), showed that both males and females are attracted by the odour of mustard oils from the leaves of brassicas. Once in the vicinity of a host population, the female finds individual hosts by searching the surface of a leaf with the tips of her antennae. Hagen and
Van den Bosch (1968) consider antennal contact to be necessary before ovi- position can occur. However, identification is not necessarily very accurate, as wasps often attempt to oviposit in the exuviae. The ovi- position response to aphids is unaffected by the presence of sperm in the spermatheca (Stary, 1964; Spencer, 1926), so unmated females parasitize as readily as those that have been mated. When offered a range of instars of Brevicoryne, Dzaeretiella prefers half-grown nymphs (2nd-3rd instar), although parasitism in a later stage produced larger, more fecund off- spring in a shorter time (Hafez, 1961). Advanced pre-alate nymphs were avoided.
While Hafez (1961) stated that oviposition takes place in parasitized and healthy Brevicoryne indiscriminately, Mackauer (1969) reported that
DiaeretielZa could distinguish between healthy and parasitized Myzus, and avoided superparasitism to a certain extent. It may be that discrimination against already parasitized hosts is not possible until the larva within has reached a certain age.
Hafez (1961) found that the maximum number of eggs laid by one female in her lifetime (one to two weeks) was 205. The minimum was 75, with an average of 83 per female. 10 to 12 eggs were laid, on average per female, 20
per day. A standard female, as used in the experiments here, contained
about 70 eggs when dissected. In Brevicoryne, under field conditions,
the eggs hatch in 2-3 days (Hafez, 1961). At 15°C mummies were formed in
17-21 days, the adults emerging after 8-11 days. Hagen and Van den Bosch
(1968) found that in Myzus, it took an average of 20.2 days from egg
deposition to adult emergence. At 20°C, the development times were
approximately the same in Myzus and Brevicoryne: 8-9 days until mummy
formation, then a further 5-6 days until emergence. At 25°C, the life
cycle is complete in approximately ten days (six days until mummy
formation). This agrees well with the times observed in this study.
The 4th instar larva may be visible through the integument of the
aphid as a bright yellow horse-shoe shape. Shortly after this the aphid
turns a golden yellow colour (= mummy), and is found to be firmly anchored
to the substrate. The parasitoid then pupates inside the mummy and
eventually emerges through a circular escape hatch cut in the dorsal
surface of the aphid's abdomen. Diapause (a state of suspended develop-
ment allowing the insect to avoid adverse conditions) occurs just before
pupation, so that the parasitoid overwinters as the last larval instar
within the mummy (Hafez, 1961; Hagen and Van den Bosch, 1968). This can
be broken at any time by exposure to room temperature and a long photo-
period (Mackauer, 1969). In Holland, in 1959-60, there were 5-11
generations per year (Hafez, 1961).
1.1.3. Aphidius matricarias (Hal.)
Aphidius matricariae (Hal.) (Hymenoptera: Braconidae, family Aphidiinae) is practically indistinguishable from Diaeretiella with the naked eye.
Under the low power of a binocular microscope, however, the two can be distinguished by the presence of an extra vein in the forewing of Aphidius.
This is shown in figure 1.3. 21
Aphidius matricariae has been commonly recorded from all the major
aphid pests of potato in England, which include Macrosiphum solanifolii
(Ashmead) and Aulacorthum solani (Katterbach), as well as Myzus persicae
(Dunn, 1949). The host distribution of Aphidius, compared to that of
Diaeretiella, clearly shows that the parasitoid is attracted primarily to
the host plant or plant family (potatoes for Aphidius, crucifers for
Diaeretiella) and then will parasitize any aphid found on them, although
preference for a particular species is found. Diaeretiella rarely attacks
Myzus on potatoes, and this study shows that Aphidius is very inefficient
when attacking Myzus on Brussels sprouts.
Aphidius has been reported to attack 40 different aphid species, comprising 21 genera, but prefers Myzus persicae in Rhodesia and the
U.S.A. and Macrosiphum in Britain (Dunn, 1949; Būnzli and Būttiker, 1959;
Schlinger and Mackauer, 1963). It has been shown to control Myzus
effectively in greenhouses, providing the temperature is higher than about
180C (Scopes, 1970; Hagen and Van den Bosch, 1968). In 1957, there was a sudden widespread increase of Aphidius in California, which led to heavy parasitism of Myzus in the field, but generally it appears to be unable to control the aphid outside greenhouses.
There is much less information on the behaviour and life history of
Aphidius in the literature. The mating procedure and development times are more or less identical with those of Diaeretiella, under the conditions used in this study. However, according to Vevai (1942), there is a time delay of at least two hours between emergence and mating, and also between emergence and oviposition, if unmated. Unmated females which have started laying, are very unlikely to mate later. The most favoured site for ovi- position is in the dorsal abdomen near the cornicles. More posterior 22
attempts often result in fouling with honeydew, which may prevent further oviposition, and at least certainly involves much time spent cleaning. A single attempt does not necessarily result in oviposition. Vevai (1942)
showed that 23.2% of aphids attacked once, 45.7% of those struck twice,
80% of those struck three times, and all of those attacked four times were parasitized. A single female produced 180-350 eggs, and oviposition activity reached a peak in days 2 and 3 after emergence. A standard female, as used for the experiments reported here, was found to contain about 100 eggs upon dissection.
Mackauer (1969) reports that Aphidius is heliotropic, and may live longer in diffuse light. In bright light the activity of the parasitoid apparently increases, but the number of attempts at oviposition is not correlated with light intensity (Vevai, 1942).
Further observations by Vevai (1942) show that active combat and cannibalism by the early larvae follows superparasitism, so that only one larva survives to exploit the host.
1.2. The Indian meal moth Plodia interpunctella (Hūbner), and its para-
sitoids Nemeritis canescens (Grav.) and Bracon hebetor (Say)
1.2.1. Plodia interpunctella (Hūbner)
The Indian meal moth Plodia interpunctella (Lepidoptera: Pyralidae, sub- family Phycitinae) attacks a very wide range of stored products throughout the world. Richards and Thomson (1932) have gathered references to Plodia from many countries in practically every continent. Although recorded from
Japan (Kuwara, 1919) there is a relative scarcity of records from the Far
East. This is probably due to lack of communication with China and Russia, rather than an absence of moths. There are records from further south 23
i.e. Australia and New Zealand. An extremely wide range of products is
attacked by the larvae. Practically every kind of cereal, dried fruit and nuts is acceptible, as well as old books and fur. However, there are very
few records of wild populations outside warehouses. Treherne (1921) recorded the larvae in apple orchards in British Columbia. Presumably,
Plodia originally lived on fruit and nuts in the wild, before they were accumulated in warehouses, whither the moths quickly followed.
A line drawing of the adult moth may be found in figure 1.4.
Like other Lepidoptera infesting stored products, such as Ephestia cautelZa, Walker, the Almond moth and Ephestia kdhniella the Mediterranean flour moth, Plodia spoils as much food by the silking up of the produce, and secondary mould infections, as is actually consumed by the larvae
(Mookhejee at al., 1969).
Control measures of such pests have concentrated as much on building insect proof warehouses, as applying insecticides. Biological control is rarely used. Nevertheless, some attempts have been made to assess the control potential of some natural enemies. Reinart and King (1971) attempted to find the optimum ratio of the parasitoid Bracon hebetor to host density for effective control. (When 250 female Bracon were introduced to a culture of 1800 Plodia larvae, 97% moth mortality followed). This would obviously require a great deal of work breeding the parasitoids, to say nothing of the difficulty of assessing the host pop- ulation in a given warehouse. No attempt was made to assess the para- sitoid's capacity to search for unevenly-distributed hosts in a large area. It seems very unlikely that this is a feasible control method.
The bacterium Bacillus thuringiensis (Berliner) is a common cause of 24
A
B
Figure 1.4. A. Plodia interpunctella female. Bar shows actual size (16 mm)
B. Fifth instar larva. Bar shows actual size (12 mm)
(from illustrations by K. Grossman on a poster of stored product pests. Deutsche Gessellshaft fur Schatilings bekampfung mbR D-6000 Frankfurt (main) 1, Postfach 2644) 25
mortality in cultures of stored-product Lepidoptera. The spores contain
protein crystals which are released into the gut when ingested with food.
This causes gut paralysis which spreads to other parts of the body when
the gut contents leak into the haemocoel. This is followed by the death
of the larva, and its subsequent melanization (Steinhaus and Marsh, 1962).
Vago and Kurstak (1965) have reported that Bacillus thuringiensis may be
transmitted on the ovipositor of a Nemeritis canescens female. McGaughey
(1976) has suggested that Bacillus could be used for controlling moths in
stored grain.
The sporozoan parasite Mattesia dispora (Naville), is also fairly
commonly reported from Plodia populations (Musgrave and Mackinnon, 1938).
The life history of Plodia is very similar to that of other Phycitid
moths, which has been sufficiently documented by many authors; for
example Richards and Thomson (1932), Rogers (1970), Benson (1973) and
Morēre (1970). A brief summary follows. After emergence, when the wings
have been inflated and dried, the female begins "calling". The abdomen is
flexed upwards, extruding the eighth and ninth segments. The inter- segmental fold between these two segments is glandular, producing a volatile pheromone which attracts the male. The male approaches the female, face to face, vibrating his wings and emitting a pheromone in his turn, from tufts of modified scales on a pair of small chitinous areas on either side of the eighth abdominal tergite. The male then curves his abdomen forward and over the pair to clasp the end of the female's abdomen. The two moths now face in opposite directions, and remain quietly so for about one to three hours. Copulation apparently takes place in the early morning, soon after emergence. The females emit pheromones until they die or are mated (Brady and Smithwick, 1968). Ripe 26
eggs are found in the ovaries some hours after emergence (Joubert, 1969),
and are laid as soon after mating as possible. The eggs hatch after
about 1-2 days, and the first instar burrow into the food, leaving tiny,
silk-lined tunnels behind them. There are five instars, when cultured
under the conditions used here. However, Takahashi (1961) found that in
Ephestia cautella, the number of instars increased with population
density, i.e. in times of food shortage, which may well be the case with
Plodia. At the end of the fifth instar, the larvae become much more
active, and show a marked tendency to migrate. In warehouses they would
seek out cracks and corners to pupate in. When overcrowded or underfed,
or when temperatures are too high, the late larvae enter diapause .(Tsuji,
1958, 1959a, b; Tzanakakis, 1959; Morere and Berre, 1967). Generally, the
larvae spin thin silken cocoons to pupate in.
A mass-rearing technique has been developed by Silhacek and Miller
(1972), but the much simpler method used here was found to be adequate.
1.2.2. Nemeritis canescens (Gray.)
Nemeritis canescens (Gray.) (Hymenoptera: Ichneumonidae, family
Opioninae) is a moderately large, solitary parasitic wasp. It is approx- imately one centimetre long from the head to the tip of the ovipositor, completely black except for the orange abdomen, which is darker in colour when reared at low temperatures. Males have only been reported very rarely, for example Beling (1932), as the females reproduced partheno- genetically, laying diploid female eggs without the intervention of meiosis and fertilization. This is called thelytoky. This method of reproduction makes it an extremely convenient animal to breed in the lab- oratory as there is never any problem with unbalanced sex ratios, or male sterility etc. 27
A line drawing of a Nemeritis female is shown in figure 1.5.
The parasitoid has been endowed with a variety of generic names,
including Idecthis and Exidecthis (Salt, 1976). In 1961 this was changed
to Devorgilia and in 1966 to Venturia, so references to this species may
be found under any of these names. However, most of the work which has
been done on this animal, which is extensive, has been published using the
'name Nemeritis canescens, which is used here.
Salt (1976) has collected together all the references to different
host species parasitized by Nemeritis_ He found that development has been
completed successfully on 23 species, of which 12 are parasitized
naturally, including Plodia interpunctella and 4 species of Ephestia.
While most of the natural hosts feed on stored products as caterpillars,
a few of those most recently observed feeed on fruit under more or less
wild conditions. Presumably, Nemeritis originally attacked the Lepidoptera
feeding on fallen nuts, fruit and seeds exclusively, moving into warehouses
in the wake of the caterpillars once man started storing things. Its
present scarcity out of doors may be because it has become less tolerant
of cooler conditions. Its range is easily extended by the transport of
grain etc. containing caterpillars and parasitoids. It has been recorded
from most of Europe, North America, North Africa, Asia Minor, Japan and
Australia.
Upon emergence the hungry females are attracted to light. Laboratory studies have shown that wasps fed on honey or sugar solution live longer
(Ahmad, 1936) and Beling (1932) has reported Nemeritis feeding on flowers outside flour mills. Once feds-.the wasps become negatively heliotropic and are attracted back into dark places, such as the inside of a warehouse
(Ahmad, 1936). It is highly likely that this behaviour has evolved since 28
Figure 1.6. Adult female Bracon hebetor x 10.
(from a photograph of a specimen)
Figure I.S. Adult Nemeritis canescens x 7.
(from a photograph of a specimen) 29 the wasps took to a warehouse existence.
Inside the warehouse, the wasps are attracted to silked-up areas where the hosts are to be found. Mudd and Corbet (1973) have shown that the attractive scent is due to a secretion of the mandibular glands of the host larvae, presumably produced during feeding, and laid down on the silk. The same molecule elicits the probing response. The ovipositor is curled under the body to point forwards, and inserted into the medium.
This is repeated until contact is made with a host larva, when an egg is usually laid within it. Rogers (1970, 1972) has shown that during probing a single egg is carried in a cavity at the tip of the ovipositor. Once this has been laid a characteristic `cocking`* movement is necessary to replace it. This is an extremely useful behaviour pattern, which can be used to count the number of eggs laid. At 25°C the eggs hatch in 2-3 days.
There are five larval instars, each of which lasts about two days. The mature larva then pupates within a transparent capsule, usually within the host remains. About eight to nine days later the adult emerges. The complete cycle takes about 20-25 days.
This wasp has been extensively used in the laboratory for many projects, including research into intraspecific competition between larvae within a single host (Salt, 1961; Fisher, 1963), host immunity and avoid- ence of superparasitism (Rogers, 1970, 1972), interspecific competition
(Fisher, 1961), and searching strategies (Cook and Hubbard, 1977; Waage,
1977). A great deal of work has also been done on interference between adults (see Chapters two and three), and aggregation, and their effects on the stability of both host and parasitoid populations (for example Hassell and Rogers, 1972; Hassell and May, 1973, 1974; Rogers and Hassell, 1974). 30
1.2.3. Bracon hebetor (Say)
Bracon hebetor (Say) (Hymenoptera: Braconidae) is a small, black,
gregarious, ectoparasitic wasp. A line drawing of the adult female is
shown in figure 1.6. There has been great confusion as to the correct
name to apply to this insect. A large number of very similar forms have
been described under different names, and it is extremely difficult to
distinguish between specific differences, and those between races or
varieties of the same species. Insects from what is probably the same
species have been placed in at least three genera: Bracon, Microbracon and
Habrobracon. Differences in colour, and the number of antennal segments
were responsible for the assignation of some of these wasps to different
species. However, it has since been shown (Payne, 1934, and observations
made in this study) that the colour of the wasp depends largely on the
temperature at which it was raised, being yellower when reared at higher
temperatures. The number of antennal segments is highly variable, even in
wasps from the same culture.
However, Narayanan et al (1958) have shown that the male genitalia
are quite distinct in Bracon hebetor and B. brevicornis, and Chawla and
Subba Rao (1963) have demonstrated amino acid differences between Bracon
hebetor, B. brevicornis and B. gelechiae.
The wasp has been commonly reared from stored product Lepidoptera,
particularly PZodia interpunctella and Ephestia species (Richards and
Thomson, 1932). In addition it has been found attacking Spodoptera
Zittoralis in Israel (Gerling, 1969), the cabbage web worm Hellula undalis
(Herakly, 1968), Plutella maculipennis (Ullyett, 1943, 1947), Corcyra cephalonica, Eublemrsu scitula, E. amabilis, Holcocera puZverea and,
possibly, Hapalia machaeraZis (Beeson and Chatterjee, 1935). 31
The structure of the animal has been studied in Soliman (1941) and
Bender (1943); Soliman (1940) has also fully described the life cycle.
Upon emergence, the adult female is about 3-4 mm long, from the head
to the base of the ovipositor. Mating takes place almost immediately
after emergence, the males displaying a vigourous vibration of the wings
just prior to copulation. The males mate several times, but there is no
corresponding information on the females. Unlike Nemeritis, Bracon has a
very small capacity for egg storage, and the eggs are produced gradually
throughout life. In times of host shortage the eggs are resorbed, and host feeding at higher host densities is necessary for egg maturation
(Benson, 1972, 1973b). Both males and females will take dilute honey
solution if offered. This increases the longevity of the males, but does not affect that of the females if hosts are provided. However, Clark and
Smith (1967) found that in some strains, Bracon females lived longer when kept with only honey or sucrose solution than when provided with four
Anagaster larvae daily. In the absence of hosts both male and female wasps will live three to four weeks when given honey or sugar solution.
Murr (1930) reported that the females were attracted to regions of silk and host larvae; and observations made during this study show that the females are attracted to silk, and will go through the motions of oviposition in the absence of host larvae where there is a good layer of silk over the medium. Upon first contact with a healthy host the female will attempt to inject a paralysing venom. However, she is easily intimidated if the larva reacts vigourously. The venom and its action has been studied by Beard (1952, 1972). Paralysis takes about ten minutes, or more if a smaller dose of venom is injected than usual. The heart and gut muscles continue to function some time after the body wall muscles are useless, and the insect may remain alive for some time after being stung, 32
but death is inevitable. When a wasp encounters a paralysed larva, it will
either feed from the haemolymph oozing from the wound inflicted by
stinging, or lay eggs on or very near the body of the host. Generally up
to five eggs are laid on a host at a time, although the female is easily disturbed. When hosts are in short supply re-encounter with a parasitized host will often result in more eggs being laid. However, Shiga and
Nakamishi (1968) have reported that gregarious parasitoids may avoid superparasitism and Ullyett (1945) considered that Bracon can distinguish between parasitized and unparasitized hosts. He and Richards and Thomson
(1932) both suggested that paralysis and oviposition occur in two distinct phases, but Benson (1972) disputed this, as he had observed that, when the larvae are allowed to spin feeding tubes, the wasp remains for about half an hour in the same place when she has stung a larva, and then burrows down to it and oviposits, forming a single sequence of behaviour patterns. Ullyett (1945) did, however, suggest that the wasps retrace their steps using a pheromone trail. Observations undertaken in the course of this study suggest that the behaviour of Bracon is somewhat more flexible than indicated by the previous authors (see Chapters three and four). In many cases, a larva once stung has been observed to leave its feeding tube and cover considerable distances before paralysis occurs.
The eggs hatch in 1-2 days at 25°C and the first instar may consume unhatched eggs or fellow larvae, There are four larval instars, which feed upon the paralysed host, and then the fully gorwn larva constructs a silken cocoon on or near the remains of the host larva. After a total of about 14 days at 23°C from oviposition, the young adults emerge from their pupae. 33
CHAPTER TWO
INTRASPECIFIC COMPETITION: AN INTRODUCTION
There have been few quibbles about the definition of intraspecific
competition, and its existence. As Darwin pointed out in 1859, members of
the same species have very nearly identical needs, and the resources to supply those needs are always bounded in a real world. This is not to say that resources are always in short supply (indeed it is likely that a generous supply in excess of demand is available more frequently than
Darwin assumed), but will often be so. To restate the definition used in the introduction; competition occurs when two or more individuals share a necessary resource such that they suffer a detrimental effect when that resource is in short supply. (But see comment on contest competition on page 49).
The resource in question can be anything a plant or animal needs; food, a mate, shelter, space, sunlight. Even avoidence of predators can be thought of as a subtle form of competition. If a predator takes a certain number of individuals, one could compete by being more efficient at escape (camouflage, running away etc.) than that number of one's fellows. The resource is an abstraction, although none the less real;
"a place among the survivors". It is always in short supply even if a predator only eats one of their number.
Competition within a species has two very important consequences; natural selection and density dependence. It is one of the main driving forces of evolutionary change within a species, and the most powerful negative feedback loop available, which can restrict population growth. 34
Despite the troublesome outbreaks of pest species, it is obvious that
the numbers in a population do not increase indefinately. An attempt to
identify the mechanisms that keep populations in check has been in progress
from 1911 onwards, when Howard and Fiske listed and classified all the
mortality factors affecting populations of the Gypsy moth and the brown-
tail moth in New England. They distinguished between "facultative"
factors, causing a proportionally higher mortality when the population
density was high; "catastrophic" agents, such as thunderstorms, which
acted regardless of population density, and a third category, responsible
for removing a constant number of the population, and hence a decreasing
proportion, as the population density rises. In 1935, Smith renamed the
first two categories "density-dependent" and "density independent".
These authors considered that only density-dependent factors, biotic
factors such as parasitism, predation and competition, could keep a pop-
ulation in check effectively. This line of thought was pursued and
developed by many others during the 1930's and onwards, notably Nicholson
(1933), Lack (1954), Milne (1958), Klomp (1962), Varley and Gradwell (1970)
and Huffaker and Messenger (1964).
Much wrangling occurred between the exponents of density-dependence
and those who followed Andrewartha and Birch (1954) in considering pop-
ulations to be kept at low densities primarily by climatic adversity. An
excellent account of this debate, which has now more or less petered out,
can be found in Krebs (1972).
The controversy has been something of a red herring. Either form of
population control may occur, depending on the circumstances. Certainly, no population can increase forever. If (because of favourable conditions) it increases much above the level at which its resource supply can support 35
it, then mass starvation must occur, drastically reducing density to a
supportable level, or extinction.
Bounded population growth was expressed mathematically as early as
1838 by Verhulst, attempting to describe the growth of human populations.
This is the now familiar logistic equation, dN dt = rN (- )
where N = population size
r = maximal growth rate
t = time
K = a stable maximum which can be achieved_ by the population
(K is also called the environmental carrying capacity)
Population increase is moderated by the potential expansion left to the
population (K-N), which decreases as the population density rises, as K
is a fixed upper limit determined by the environment.
This, being a differential equation, is based on a mathematical
system which necessitates taking infinitesimally small steps in time. It
is therefore only suitable for organisms with a very short generation
time or completely overlapping generations. For those with longer
generation times, wholly or partially separated in time, corresponding
difference equations are more suitable. A difference form of the logistic
equation is
K--Nt Nt+1 - Nt = r Nt ( K ) where Nt, Nt+1 are the sizes of the tth and (t+l)th generation. This is much more difficult to handle mathematically than the differential form.
There have been quite a number of mathematical models for density- 36
dependence in both the difference and the differential form. Perhaps the
simplest is Varley and Gradwell's linear model of 1968 where
the k-value for a mortality = Log Nt = Log a + b Log Nt NS
where Nt = original population density
NS = survivors
a, b are constants
A variety of later models can be found in Hassell (1975), May (1974),
May at al. (1974) and May and Oster (1976).
In the differential models any disturbance in the population density
will be adjusted immediately by the density-dependent part of the equation.
This hardly corresponds to a biological situation where there will always
be some delay in responding to a change of circumstances. Attempts have
been made to remedy this naivete by introducing time delays into the models. However, difference equations always contain a time lag of one generation, and their stability properties are very similar to those of differential models with time delays.
Perhaps the most important feature is the ratio between the time delay and the characteristic return time, which indicates how quickly the system will tend to return to its equilibrium once it has been displaced from it. As long as the time delays are short in comparison with the return time, the population is fairly stable. However, as the time delay increases with respect to the return time, the population becomes less stable; a displacement from equilibrium moves from an exponentially damped return through an oscillatory damped response towards regular self- sustaining fluctuations (for example a stable limit cycle fluctuating in 37
a regular manner between four constant points). When the time delays are
large in relation to the return time, populations would be expected to
exhibit chaotic behaviour; irregular cycles of no fixed period, apparently
random behaviour depending on the initial conditions. A more detailed
mathematical account of this work can be found in May (1976), (1973) and
May et at. (1974).
While density dependence is not necessarily a stabilizing interaction,
it certainly can be a powerful means of curtailing population growth.
There is evidence that when it does occur naturally it has a greater
tendency towards the stabilizing end of the spectrum i.e.. the time delay
is generally short enough in relation to the return time to allow the
relationship to be effective in population regulation (Hassell et al.,
1976).
However, it is obvious from studies of animal populations that
density-dependent mortality may be negligable or non-existent. Stubbs
(1977) gathered together a number of case histories in order to look at
b, of the Varley and Gradwell model (1968). There was no information on
time delays. However, thirty-two of forty-six values for b were less
than one (undercompensating), and ten had slopes of less than 0.2.
Populations in the field within the range of densities specified by these ten cases would be subject to such slowly acting density-dependent mortalities that the effect would be almost negligable. In 1966, Tanner reviewed 71 correlation coefficients between population density and growth rate. Seven of these showed no significant difference from a random series, and in sixteen cases the correlation coefficient was not significantly different from zero.
Huffaker and Messenger (1964) suggested that density-dependent 38
effects were relatively more important in the centre of the species range,
where density was likely to be higher due to the most favourable climatic
conditions for that species. Towards the edge of the range however,
conditions are more likely to be difficult and population density is kept
low, so that density-dependent mortality has very little effect.
Differences in population dynamics between species may also be
related to the severity and/or predictability of the environment, in a
similar way. Consideration of these differences has led to the develop-
ment and elaboration of the idea of r and K selection (so called after the
constants used in the logistic equation). Although these terms were
coined by MacArthur and Wilson (1967), the germs of the idea may be traced
further back. In 1958 Margalef distinguished between species found in
communities during the initial stages of succession and those found in the
climax community. During succession the trend is to minimize the
dissipation of energy. Gradually, selection shifts the advantage to
species with a lower intrinsic respiration (and bigger individual size), reduces the number of superfluous descendants so that fewer but better protected offspring are produced, and lengthens the life span, with a general slowing down of turnover. The entropy of the system is reduced, as is the rate of production. The community becomes more efficient. This corresponds to a gradual shift from r-selected species (the colonizers) to the K-selected species found in a mature community. Basically, r-selected species are those that maximize their rate of increase (r), while those that are K-selected maximize their equilibrium levels i.e. become adapted so that the carrying-capacity (K) of the environment for that species is pushed up. The idea sounds a simple one but in practice it is difficult to apply. There are a large number of interrelated features of the life history and population dynamics of a species which show consistent changes 39
when comparing those living in severe and/or unpredictable environmental
conditions with those found in equable and predictable climes. The "r-K"
spectrum is a shorthand way of referring to those trends, which have been
tabulated by Pianka (1970). At the r end of the spectrum (unpredictable),
mortality is highly variable, climatically caused and density independent.
Competitive interactions (of both the inter- and intraspecific varieties) are thought to be of very little importance; the animal is small, short lived and highly fecund. (There is an unfortunate tendency to equate high fecundity alone with a position on the continuum nearer to the r-end.
However, as noted above, the r-K spectrum is concerned with a whole complex of related biological characteristics which, are very difficult to disentangle. It would be wrong to allocate a position on the continuum based on only one of these).
It is extremely difficult to adapt to unpredictable conditions; selection pressure is not consistent enough to produce an animal that can really cope. When conditions are merely severe, adaptation (through individual morphology or physiology) is more likely to occur, but species are limited by their evolutionary histories, and their general level of genetic variability (rates of mutation etc.). When all else fails, and the mortality risk per individual is still high, the only way that a genotype will persist is if it is very fecund. This, however, is a wasteful process. Thus, selection will favour the production of large numbers of small young whenever there is high risk to each individual which cannot be easily overcome by straightforward adaptation i.e. highly variable density independent mortality selects for high r.
Where conditions are less severe, and more predictable, an animal which uses its time and energy efficiently, producing smaller numbers of 40
large, well-cared for or well protected young, will leave more offspring
in the long run than one which continues to produce many small offspring,
a large proportion of which will die (in competition with their larger
conspecifics).
When migration and dispersal are necessary to an animal, a high risk
is again attached to each individual. The importance of this aspect of
an animal's life cycle has been stressed in a number of recent papers by
Southwood and his colleagues (Southwood et al., 1974; Southwood, 1975,
1976; Southwood and Comins, 1976). At the r-end of the spectrum are many
animals of temporary habitats, colonizers and fugitives. For these species
the relationship T/ff is" important,-where t is the generation time and H is
the length of time a habitat is favourable. This ratio always lies be-
tween 0 and 1; approaching 0 at the K end of the spectrum and 1 at the
r-end.
Other interesting work on the r-K idea includes considerations of
the evolution of r and K (Charlesworth, 1971; King and Anderson, 1971; and
Roughgarden, 1971), and of the distributions of resources between
reproductive and non-reproductive activities, (Brockelman, 1975). Not
only will r-selected animals be expected to produce more but smaller off-
spring, but also to devote a greater proportion of their resources to off-
spring production. Extra food is better used to produce more young than
to prolong the life of the adult; the risk is better spread among many
individuals. Gadgil and Solbrig (1972) have shown that dandelions in disturbed areas devote a greater proportion of their aerial biomass to seed production.
Competition, then, has an important part to play in both the regulation of population density, and the adaptation of species to their 41
environment. The main emphasis in this study, however, is not to underline
the importance of competition, but to look at its mechanisms.
While much time and energy has been expended upon the existence,
measurement and consequences of competition, less attention has been
bestowed on its mechanisms. For example, while Broadhead and Cheke (1975)
made careful measurements of competition in the parasitoid Alaptus fusculus,
and went on to model the dynamics of the animal, they did not identify the
mechanisms involved, although they assumed some kind of interference was
taking place.
In 1954, Nicholson distinguished between two forms of competition.
These he called "scramble" and "contest". In scramble competition, which
is found, for example in Indian meal moth caterpillars, Plodia inter—
punctella, competing for the medium on which they feed, there is no
clearly defined method by which the "resource is divided among the
competitors (Snyman, 1949; Podoler, 1974). Those which fail to get a critical minimum amount of the resource die or fail to reproduce. In others, body weight, fecundity and longevity may be reduced. At very high densities all may die.
In contest competition, successful competitors get all they need; the failures get nothing, a much more efficient use of resources. Most territorial species, and those that have complicated agonistic behaviour patterns are found near the contest end of the scale.
Figure 2.1. provides a simple illustration of the two forms. Suppose that the level of resources, R, is constant but varying numbers (N) of competing individuals are present initially. Each individual requires r resources to survive. When r N > R, deaths occur. The number of 42
A CONTEST
il CANNIBALISM
SCRAMBLE
Figure 2.1. The number of survivors, Ns, plotted against the initial.
S .. number. N# far scramble and contest competition._. The plots
for these two diverge at A. The level of resources available
is kept constant at R (see text).
N
RESOURCES OBTAINED BY A SINGLE INDIVIDUAL
Figure Aar example of a frequency distribution of individuals obtaining a given quantity of resources. Sere a normal distribution is shown for a population of individuals at A in figure 2.1. The mean obtained is Rill. If r resources are required by an individual to survive, individuals obtaining less than r (shaded portion of graph) die. 43
survivors is denoted N (see figure 2.1). In the first part of the graph s r N < R; all survive, i.e. N = N. At A, r N = R. In complete contest s competition the number of survivors will now become constant, because the
resource only goes to those that survive. Survival remains at its
maximum because there is no waste. In scramble competition, however, the
number of survivors will fall as the initial density increases above A,
because there are more failures to waste resources. Here the simplest relationship, a linear form, is used for illustration. If food is shared
equally, so that each scrambler gets R/N resources, and the probability of death by starvation falls linearly as R/N approaches r, (line B figure
2.3), then a linear form would be expected. However, any kind of declining function might be appropriate, while it seems unlikely that each individual would get exactly R/N resources. Some kind of distri- bution with a mode of R/N would be a much more reasonable assumption
(figure 2.2). Here r is less than R/N so most individuals will survive.
However, resources taken by individuals lying to the left of r are wasted.
This suggests that some deaths will occur even when, theoretically, there is enough to go round. In this case the scramble plot in figure 2.1 would never rise as high as A.
Figure 2.1 could be replotted as a k-value graph as in figure 2.4.
The contest plot rises with a slope of 1.00. Cannibalism, such as is found in flour beetles, Tribolium spp. (Park et al., 1965) would have the effect of improving efficiency by recycling material, although some energy would be lost in the extra transformation steps involved in cycling it through another individual. The scramble plot would be shifted towards the contest graph, in the direction indicated in figure 2.1 and 2.4.
Fujii (1965) took a closer look at scramble and contest. He assumed 44
1.00
x w
Pi
0.00
R/N r
Figure 2.3.0 The expected decline in the probability of death by
starvation as the average resources, R/N, obtained by one
individual, rises to r, the resources necessary for the
survival of one competitor (see text).
SCRAMBLE
Log Initial Population Density
Figure 2.4. Figure 2.1 replotted as a k-value.
(After Varley, Gradwell and Hassell, 1973). 45
that the resource was divided into equal blocks, and that the frequency
distribution of eggs laid in each block followed the binomial. A further
assumption that he made was that when K eggs were placed in a block, it
yielded K adults when K a (the maximum number of adults each block was
capable of producing), but when K > a, then (a - a) adults were produced.
a is a measure of wasted resource, and hence the degree of scramble
involved; when a = 0, a adults will always emerge, a pure contest result.
When x eggs were distributed in this system, the number of adults emerging
A(x) were calculated from:
_a A(x) = N E K.B(K; x, 1/ N) + (a - a) E b (K; x, K=0 K=a l
where N = number of blocks and
b(K; x, 1/N) = the terms of a binomial distribution.
Varying a, and keeping the other parameters constant, gives the
family of curves shown in figure 2.5. This is similar in general form to
figure 2.1. In 1957, Birch introduced the terms "exploitation" and
"interference" which Miller (1967) equated with scramble and contest. The
definitions of these terms are slightly different, but punctionally they
amount to the same thing. When exploitation occurs, the only way that an
organism affects its competitor is by decreasing the availability of the
resources they share. They may never come into direct physical contact.
This has been termed "indirect" competition. The definition of inter-
ference includes the idea of individuals directly harming or hindering
their competitors, by, for example, fighting or the production of toxic
wastes, or simply causing them to waste time by being physically in the way (see table 2.1 for examples). This kind of passive obstruction must
have been present right from the very beginning., at least between con- 46
a = 0
5 10 15 x
Figure 2.5. The number of adults emerging, A(x), when x eggs are binomially
distributed among equal sized blocks. Each block can
produce a maximum of a adults. When K, the number of eggs
in a block, is greater than a, (a — a) adults are produced.
a is therefore a measure of scramble (see text).
(After Fujii, 1965). Table 2.1. Some examples of mechanisms of intraspecific competition. Numbers in brackets, e.g. (2), refer to trophic level. (1) - herbivore,
(2) - predator/parasitoid.
Competition Species Resource Authors Type Mechanism
1. Scolytua soolytue Wood Scramble Decreased survival, and body weight. Beaver, 1974 S. multietriatua (1) Tomicue piniperda
2.Drosophila meianoga8ter Yeast Exploitation Retarded rate of development; survival; Miller, 1964 D. simula ns (1/2) adult weight reduced; total biomass Gilpin, 1974 (fruit flies) reduced. Interference Oviposition reduced; due to adult Pearl, 1932 obstruction?
3. PZodia infsrpunDtaZZe Meal/flour Exploitation Increased larval mortality, shortage of Snyman, 1949 (Indian meal moth) pupation sites, late pupation, reduced Podoler, 1974 (1) weight: Interference Rats of oviposition lowered. Snyman, 1949
4. Anagaeta kuhniella Meal/flour Exploitation Survival, length of larvae, eggs/7 in Ullyett b Merwe, 1947 (Mediterranean flour ovary reduced. moth) (1) Interference Oviposition reduced. ' Ullyett, 1945
5. Callosobruchua ohineneie Beans Interference Young larvae bite older larvae, which BellOws pers. comm. (Beetles) (1) bleed to death. Egg hatching success reduced by trampling. Rate of oviposition reduced. Super-oviposition on beans reduced because Honda at ai., 1976 1) pheromone marker, 2) eggs already present. C. modulator) Beane Interference Avoidence super-oviposition. Mitchell, 1975
6. Naaonia vitripennie Housefly pupae: Mkeoa Exploitation When >25 eggs/pupa - hatched last die from Wylie, 1966, 1971 (gregarious) (2) doneetioa starvation. Sex ratio disturbed (i d ). Interference 2V lay fewer fertilized (7) eggs when Wylie, 1966 adult density high,,and when encountering parasitized hosts. 7. NusdMifurax raptor (2) Mueoa pupae Interference First hatchling attacks super numerary Wylie, 1971 eggs.
B. Spalangia oamerOni (2) Mseoa pupae Interference As above. Wylie, 1971 9. Aphidiva app. (2) Aphids Contest Combat between first instar. Stary, 1966 10. Neme'i.tie oanesoens (2) Moth larvae Exploitation Asphyxiation younger larvae. Fisher, 1963 Interference Combat between first faster. Salt, 1961 Interference Avoidance superparasitism; chemical cues? Rogers, 1970, 1972 Fisher & Ganesalingam, 1970 Interference Time. wasting, dispersal after adult Hassell, 1971a, b contacts.
11.Trtiohogramna evanesoens Sitotroga oereatella Interference Avoidance superparasitism - odour trails, Salt, 1936, 1937 (2) eggs chemical cues? 12. Cardioohiles nigrioepe Beliothia viraaQeWa Interference Avoidance due to marking pheromones. Vinson & Guillot, 1972 (2) larvae Mioropletis orQOeipee
13.Pleolophue Interference Avoidence trail odours, Price, 19700 Erdaeys (2) Nutria }
14.TeZononus sphingie (2) Manduoa sexta eggs Interference Avoidence marked eggs. Rabb & Bradley, 1970
15. Crgilue lepidus (2) Phthorimaea QpsrokLella Interference Avoidence pheromone marked sites; internal Greany & Oatman, 1972 chemical markers in host haemolymph.
16.Apantslee melanosoelus Porthetria diepar Interference Avoidence superparasitism, combat in 1st Weseloh, 1976 (2) larvae instar. Exploitation Physiological suppression of young larvae.
17.Bmoon hebetor (2) Cadra oautella larvae Cannibalism Early larvae eat eggs and other larvae. Benson, 1973 Scramble Size reduction, increased mortality, shift in sex ratio 4 d .
18.Desert grasshoppers (1) Females Contest Fighting and stridulation establishes Otte & Joern, 1975 territories in creosote bushes.
19. Trigona app. (1) Nectar/pollen Contest Aggression - reduction in foraging time Johnson & Hubbell, 1974 (stingless bees) form of territoriality.
20.Rhyssa persuasoria (2) Sawfly larvae Contest Territoriality on logs maintained by Spradbery, 1970 threat and aggression.
21.C000inella ?-punotata Mysue psreiQae Cannibalism (2) Contest Contacts may stop search; violent reaction Michaelakis, 1973 -may fall off plant 49
specifics, and may have been the raw material from which more complex
behavioural mechanisms evolved. Interference or contest mechanisms may
still operate even when the resource is not limiting; and will then
produce an adverse density effect when resources are super abundant. This
has led to much confusion as to the necessity of "insufficient resources"
as a condition for the existence of any kind of competition. It is
certainly necessary for scramble or exploitation competition, and an
evolutionary necessity for the emergence of interference or contest.
Most competitive relationships usually have both scramble and contest
components in varying amounts and can be fitted into a gradient between
the two extreme forms. From the literature, an impression emerges of
more-or-less systematic variation in other features of the animal and its
environment parallel to this gradient. The r-K continuum is apparently
aligned in this way as indicated by Brockelman's (1975) work on the birth
rate and parental investment per offspring. He related high birth rate
and low parental investment per offspring with the utilization of transient
resources, a migratory habit and exploitation competition. There is some
indication that the tendency towards contest mechanisms may increase
progressively up the food chain. Among the mammals, carnivores are more
strictly territorial than the other groups, although such systems are
found in purely herbivorous species. It is striking that very nearly all
the examples of behavioural interference in insects is found in parasitoids.
(A useful selection of examples of types of competition is found in table
2.1). While many first trophic level insects do show a preponderance of
exploitation competition there are those that do show contest; for example
the bean and pea beetles (T. Bellows pers. comm.; Honda et al., 1976; and
Mitchell, 1975). This suggests that the packaging of the resource may be
more important than its trophic level. Higher up the food chain, units 50
are generally larger, more mobile and less evenly distributed. Not only
are they harder to find and catch (i.e. require a high investment of time
and energy per unit), but each unit is more valuable. Also, predators
and parasitoids generally have more highly developed nervous systems and
more complicated patterns of behaviour, which have evolved to cope with a
more demanding method of gathering food. They are more likely to have
the capacity to develop complex interference patterns such as
territoriality.
Miller (1967) suggested that scramble competition was more character-
istic of simpler, more primitive Metazoa. (Again, the nervous systems of
such animals are less likely to be able to cope with complicated
behavioural patterns). This implies that exploitation is a primitive and
unstable form of interaction, as he indiciated. At low resource levels,
or high population densities, mortality is higher than absolutely necessary
producing violent overcompensating oscillations. Nicholson's experiments
with blowfly larvae (1954a) show this. Contest is more efficient. It is
advantageous to the individual, as it minimizes time and energy waste due
to competition. Apart from physiological methods, such as toxin
production, it is primarily a behavioural phenomenon, and so the time
delays in adjusting to a new population density are short. This increases
stability.
Let us now take a closer look at the kind of mechanisms that have been found in various species of parasitoid.
From 1910 onwards superparasitism was recognized as an important factor limiting the growth of parasitoid populations (Fiske, 1910; Salt,
1936). This is the deposition of extra eggs (of the same species) in one host, over and above the number which can develop normally in a single 51
host. As Salt (1961) pointed out, first instar larvae of many solitary
endoparasitoids bear large mandibles, which are lost in the later instars.
In Nemeritis canescens supernumerary larvae are usually killed within the
first day of hatching by combat. This is also the case with Aphidius
species attacking aphids (Stary, 1966). If an egg is laid in a host
containing a large parasitoid larva, the younger larva usually dies
although there is no overt aggression. Fisher (1963) showed that older
Nemeritis larvae utilize increasing amounts of oxygen, but are less
susceptible to asphyxiation, and concluded that the physiological
suppression of younger instars was due to oxygen shortage (a form of
exploitation).
There has been a great deal of work recently on the avoidence of
superparasitism, which, of course necessitates the ability to discriminate
between hosts already parasitized, and healthy ones. That parasitoids can
distinguish already parasitized hosts has been demonstrated since 1936
(Ullyett, 1936; Salt, 1936, 1937; Lloyd, 1940; Rogers, 1970, 1972;
Wiseloh, 1976; Bakker et at., 1972). Many authors have shown that this
is due in many cases to the recognition of a trail odour or a host marking pheromone (Salt, 1937; Price, 1970; Rabb and Bradley, 1970; Vinson and
Guillot, 1972; Greany and Oatman, 1972). Fisher and Ganesalingam (1970), showed that even if such external marking systems failed, there were chemical changes in the host's haemolymph, in the amino-acid and protein constituents, which could provide a cue for a probing Nemeritis that the larva was already parasitized. This may also have a hand in physiological suppression; a toxic effect perhaps, or exploitation of a vital amino- acid in short supply. Of course, the changes in oxygen tension as the parasitoid larva grows, could provide another cue. These mechanisms serve to prevent the adult female wasting her time going over areas that she or 52
a competitor has already searched (Price, 1970c) and wasting her eggs in
hosts with already established larvae. This is analogous with mammalian
habits of scent-marking territorial boundaries, as, for example, foxes
do.
Hassell and Varley (1969) introduced the term "mutual interference",
which they used to describe the decline in the logarithm of the areas of
discovery (a) as the population density increased, logarithmically. The
area of discovery is based upon the k-value mentioned earlier specifically;
a- = 1.P Log e S
where P is the population density of the parasitoids,
N is the original host population and
S is the number of hosts that survive.
Hassell (1971a, b) suggested that this was due to changes in behaviour
following encounters between adults, and with already parasitized hosts.
He showed that in Nemeritis, the number of encounters between adults
increased with parasitoid density; and that the percentage of the time spent
searching decreased so that at higher densities, a smaller proportion of
adults were searching at any one time (Hassell, 1971b). However, the
parameter he called "m" (the slope of the Log a vs. Log P plot) in fact
includes any factor which affects the rate of successful attack on healthy
hosts. As Free et al. (1977) has shown, non-random search for a patchily distributed host is sufficient to produce this kind of relationship with- out involving any other behavioural mechanism. This is because the host distribution and the parasitoid's response to it alters the overall probability of a parasitoid encountering a healthy host. They called the resulting decrease in searching efficiency "pseudo-interference". Even 53 where mechanisms have been shown to exist and some measure of their effect demonstrated, it is important to remember that other, perhaps more important mechanisms may be at work at the same time. 54
CHAPTER THREE
COMPETITION WITHIN SPECIES I
3.1. Introduction
In this chapter, simple experiments designed to show the effects of
competition between conspecifics are described and discussed. In each of
the four species of wasp studied, varying numbers of standardized
parasitoids were introduced into a cage containing a constant number of
hosts, and left for 24 hours. The experimental parasitoids were then removed and the hosts maintained until all the offspring had emerged.
Living and dead hosts and parasitoids were classified and scored.
The data obtained in this way were used to investigate changes in searching ability due to parasitoid density. A compound parameter b w
(the product of the rate of encounter between adults and the time wasted per encounter) was then calculated and examined for trends with density.
As might be expected, the competitive mechanisms of the larvae of solitary parasitoids are predominantly of the contest variety. This can be clearly displayed by plotting the number of offspring expected, assuming the two extreme forms of competition, against the actual number obtained.
3.2. Aphid Parasitoids
3.2.1. Materials and Methods
The parasitoids Diaeretielia rapae and Aphidius matricar2ae were maintained in culture with their host Myzus persicae, as described in
Appendix A (p. 311).
Young Brussels sprout plants, Brassica olereacea variety Irish elegance, (4 to 6 weeks old) were kept at 20°C in a 16 hour day until they 55
produced 5-6 usable experimental leaves, i.e. with a healthy blade and a
stalk length of at least 3 cm. They were then transferred to 15°C, to
prevent rapid ageing, and watered well until use. Suitable leaves were
cut from the plant, swept clean of any contaminating insects, and left to
stand in water for two days. When the leaf blade exceeded the area shown
in figure 1.1 it was trimmed with scissors. Strips of cotton wool were
then wrapped around the stems where they were inserted through holes in corks fitting 5 x 2.5 cm glass tubes. This prevents undue damage to the stems, holds the leaves steady, and stops aphids drowning themselves in
the water supply in the tubes below.
Experimental aphids were reared.._as follows. Large wingless parthen- ogenetic females were placed on clean, insect free Brussels sprout leaves and left for 24 hours at 22.5°C. The adults were then removed (and placed on a new batch of leaves), and the first instar larvae, up to 24 hours old,were reared separately at 15°C for about seven days, until they reached 4th instar and were approximately 1.3 mm long. (A sample of 20 experimental aphids had a mean length of 1.291 mm, standard deviation
= 0.045). Unfortunately the leaf quality was extremely variable and very difficult to control. This affects the rate at which the aphids grow and often necessitated leaving the experimental insects for an extra day or so. On the eighth day, or upon reaching the appropriate size, the aphids were counted onto clean leaves and used in experiments the following day.
Fre-alate nymphs, slimmer and darker, with visible wing pads, were ignored. In this experiment, 128 aphids were used for each replicate.
Initially, 32 of these were transferred to the under-surface of each of four leaves with a damp paintbrush or feather. This species is fairly mobile, particularly after parasitoid attack, and so movement between leaves during the course of the experiment was inevitable. However, there 56
Figure 3.1. Standard Brussels sprout leaf.
Lifesize. 57
is no tendency to aggregate and so the aphids maintained a more-or-less
equal distribution over the four leaves.
At first the four leaves carrying aphids were placed on a four inch
pot saucer, covered with a clear plastic propagator. However, preliminary
trials with Aphidius matricarias showed extremely low levels of parasitism,
frequently down to zero. An attempt was made to combat this by decreasing
the amount of dead space in the cage, in which the wasps could linger and
not encounter aphids. This was done by placing the leaves on top of an
inverted butterdish (maximum diameter 10.5 cm, height 4.4 cm), which
removed the volume of the dish from the experiment and raised the level
of the leaves towards the light (see Plate 1). Both species of wasp are
attracted to the light and spent a higher proportion of their time in the
upper half of the cage. It was hoped that this alteration would increase
the success rate of Aphidius by increasing the probability of encounter with host aphids. Unlike Diaeretiella, Aphidius does not naturally attack
Myzus on plants of the cabbage family (Cruciferae), and is probably not attracted to the smell of Brussels sprout leaves (see Chapter One). The change in experimental conditions met with partial success; parasitism rates were improved sufficiently to make it possible to do a standard mutual interference experiments.
Newly-emerged parasitoid females (up to 24 hours old) were standard- ized by being placed with males of the same size or larger and left for
24 hours at 15°C in a 5 x 2.5 cm tube. All the females for one replicate were standardized together with an equal number of males. Each set was provided with dilute honey solution (one part honey to nine parts water) on a small cotton wool pad. In the morning the males were removed and the females introduced into the experimental cage containing aphids. 58
Each Aphidius female was approximately 1.5-2 mm long, and contained about
100 eggs at use. Diaeretiella females, 2 mm long, standardized under the
same conditions, contained 70 mature eggs. At least five replicates of
each density were run, at 20°C. After 24 hours the wasps were removed
from the experiment and returned to culture.
The cage containing the experimental aphids was then transferred to
25°C (60-65% R.H.) to speed up development. After about 6 days ummies
were formed, the adults emerging 2-3 days later. When no adults had
emerged for 3-5 days the insects were sorted and counted.
Meanwhile the experimental aphids had reched maturity and begun to
reproduce, much encouraged by the higher temperature. It was found
necessary to remove the offspring of the original aphids, at least once
every four days, to prevent overcrowding and premature mortality. There was also the difficulty of distinguishing between the original aphids and
their older offspring. After the first four days only the original aphids were mature. These adults were transferred to fresh, clean leaves and maintained for a further four days. After the second four days all the parasitoid mummies had formed, and the healthy aphids remaining were counted. The mummies were kept until emergence, when the sex ratio was scored.
3.2.2. Analysis of Results and Discussion
In some cases, particularly at the higher densities, parasitoids died during the course of the experiment. These were scored as "half insects" and the results kept. They were later found to agree well with the rest of the data, and were included in the analysis.
More unfortunately, there were cases when experimental aphids died. 59
The remains were usually shrivelled up and dried, and not, therefore, in
good condition for dissection in search of parasitoid remains. When dead
aphids numbered ten or more the replicate was discarded and repeated.
The efficiency with which the parasitoids discover healthy hosts is
a convenient measure of their performance. Such a measure was formulated
by Nicholson in 1933. He called his parameter a, the area of discovery.
N a = F Loge [ç]
where P = number of parasitoids,
Nt = total hosts present and
NS = number of survivors.
Hassell (1969 and onwards) and others have shown that this declines with
parasitoid density in many species.
Figures 3.2 and 3.3 show the plots of Log a against Log P for
Diaeretielta and Aphidius respectively. In the Aphidius data mean values
were used instead of the individual replicates because there were many
cases when the searching efficiency was zero. In Diaeretiella the slope
of the regression is -0.375, significant at the 0.1% level. Chua (1975) obtained a value for m (the mutual interference constant, which is the slope of the Log a vs. Log P regression) of -0.65 for Diaeretiella, which is quite a lot higher than that obtained here. However, he used the aphid Brevicoryne brassicae, which is the preferred host of Diaeretiella.
In addition the hosts were distributed in patches of different densities, in a larger cage than the one used here. Aggregation of the wasps in areas of high host density would produce locally crowded conditions, and tend to increase interference. As Free et al. (1977) have shown, non- 60
A. Including first set of points. y = -0.94 - 0.38 x.
o B. Excluding first set of points. y = -0.85 - 0.49 x.
0
x
0
0
0
0.3 0.6 0.9 1.2 Log Diaeretiella present
Figure 3.2. The decline in log a (searching efficiency) as the number of competing conspecifics rises, in Diaeretiella rapae. A. b/seb = t = 3.99, P <0.001 with 57 degress of freedom. b = coefficient of x (0.38 here), seb = standard error of b. B. b/seb = t = 4.08, P <0.001 with 45 degrees of freedom. o = single point x = double point + = treble point
The data for this and all the other figures can be found in Appendix B.
61
0.0 0.3 0.6 0.9 1.2 Log P
Figure 3.3. Log a (searchins efficiency) plotted against log P (para-
sitoid density) for Aphidius matricariae. No significant
relationship found. ( b = -0.67) seb
Means and 95% confidence limits shown. 62
random search for hosts which are patchily distributed may produce an
interference-like effect in the absence of time wasting following an
encounter with a competitor. Any, or all of these factors could act to
increase m.
As figure 3.3 shows, the slope of the regression for Aphidius is
small and insignificant. This is probably because the rate of parasitism
is low in all cases, as the insects are not known to be attracted to
crucifer leaves and are rarely recovered from aphids infesting crucifers
in the field.
Despite the overall difference, there is a slight similarity between
these two plots. In both cases the searching efficiency rises slightly
at a density of two parasitoids and falls away after that. This suggests
that, at low densities, contact with other individuals can enhance
searching behaviour.
From the values of the parameter a, more information can be obtained about behaviour prior to oviposition. There are two behavioural models for mutual interference in the literature: Rogers and Hassell (1974) and
Beddington (1975). In both cases it is assumed that after an encounter with another parasitoid, a searching female wastes a certain amount of time (Tw). This reduces the time available for search and hence the searching efficiency. Rogers and Hassell (1974) also assume that the parasitoid in question is not available for further encounters during T.
This has the consequence of producing an asymptote of m = -0.5, compared with that of m = -1.0, obtained by Beddington's (1975) model, in the Log a vs. Log P plot. From the observations reported later (see Chapter Four, p. 109) it is evident that neither of the wasps studied in detail is not available for further encounters, during T, although either may leave the 63
host area upon occasion. In this case therefore, it is probably better to
use Beddington's model. From this we get:
a' = a (1) I + bTw(P-1) where a is the searching efficiency when P = 1 (i.e. no encounters with other wasps); a' is the new, lower searching efficiency, b is the rate of encounter with other wasps, of which there are (P-1). Rearranging (1) gives
bT = a - a' (2) — w — - a'(P -1)
If b is a constant, then a plot of bTw against (P-1) shows whether Tw is constant or changes with. frequency of encounter. Ideally, T should be w plotted against b(P-1) (the number of encounters) however bT cannot be w broken down using the limited information available here and b(P-1) is not known. When the number of encounters is known, then both b and Tw can be calculated (as in Chapter Four).
Figures 3.4 and 3.5 show the plots of bTw against (P-1) for
Diaeretieila and Aphidiuo.- In both cases a linear regression is not significant, suggesting. no relationship between bT and (P-1). However, w in the case of Diaeretieiia, the data fits a curve in (P-1) and (P-1)2, very well. Since bTw is a product of two independent variables, it is not obvious why there is a curved relationship. Detailed behavioural studies are clearly necessary to explain this.
Further information about the nature of the overall competitive interaction can be gleaned from this very basic set of data. It can show whether competition between the larvae is closer to scramble or contest.
It has been shown in some cases, and assumed in many others that solitary insect parasitoids compete by contest in the larval stage (see Chapter 64
1.0 22.5 r y = - 0.066 ; 0.062x - 0.003x2
0.
0.0
-0. 0 10 15 P - 1
Figure 3.4. The compound parameter bTw (where b_ = rate of encounter be- tween wasps, w = time wasted per encounter) plotted against P - 1 (where P = number of wasps present) in Diaeretiella rapae. The points suggest a curved relationship. Means and 95% confidence limits are shown (see table B3.4 in Appendix B).
t = tests: b/seb = t = 5.33, P <0.05.
c/sec = t = -4.36, P <0.05
with 2 degrees of freedom. Both b and c, the coefficients of x and x2, are significantly different from zero. 65
1.8 5.0
1.0
0.0
-1.0 0 5 10 15 P - 1
Figure 3.5. The compound parameter bTw (where b = rate of encounter be-
tween adult wasps, Tw = time wasted per encounter) plotted
against P - 1 (where P = number of wasps present) for
Aphidius matricariae. Means and 95% confidence limits are
shown (see table B3.5 in Appendix B). No simple significant
relationship is apparent. 66
Two). This can be clearly displayed as follows.
If we assume that, at every encounter with a healthy host, a success-
ful attack is made, then the number of encounters with hosts per
parasitoid is a N where N is the total number of hosts present. Provided
that attack rates per encounter, and success rates per attack do not
change with parasitoid density, aN can be used as a measure of parasitoid/
host encounters with increase in parasitoid density.
These encounters can be distributed randomly among the hosts present
using the Binomial expansion.
JX n-x Nx =N.n Cx r 1 L1 LN N~ where is the number of hosts encountered x times, N is the total number X of hosts present and n is the total number of encounters (Rogers, 1975).
Random encounter is a reasonable assumption to make since variation in host density in different parts of the cage did not occur. We assume that at each encounter an egg is laid.
Now the null hypotheses of complete scramble or complete contest competition can be applied. In the first we assume that scramble competition between the larvae occurs. The hosts are not large enough to support the development of more than one parasitoid, so hosts containing more than one egg die, along with the larval parasitoids within them.
Only those hosts containing exactly one egg (putting x = 1) produce parasitoid adults.
A second estimate of expected offspring can be made on the assumption of complete contest competition between the larvae. From hosts containing two or more eggs, one parasitoid always emerges. This 67
could occur by the elimination of competitors by combat or cannibalism,
or physiological suppression (see Chapter Two). Avoidence of super-
parasitism has the same effect. The extra eggs are eliminated by being
prevented from being laid. (This could be thought of as a form of
territoriality). The number of offspring is then given by:
Expected offspring = N - Hosts not encountered (putting x = 0)
If a model is substantially correct then the regression of the
expected values on those observed will have an intercept of zero and a slope of unity. A distribution of points lying above this line shows
consistent overestimation by the model, points lying below show under-
estimation.
Figures 3.6 and 3.7 show expected values plotted against the numbers observed for DiaeretielZa and Aphidius respectively. Clearly the scramble model in Diaeretiella increasingly underestimates the number of offspring produced as the density increases. In the first part of the graph (up to a density of about 35), there does not appear to be much to choose between the two models, presumably because, under these conditions the number of re-encounters with parasitized hosts is low, and there are fewer opportunities for the avoidence of superparasitism and larval competition.
The contest model slightly but consistently overestimates the number of surviving offspring. This would be expected of a low independent mortality were to act on the survivors of competition.
Thus, the outcomes of scramble and contest competition diverge increasingly as the initial density of parasitoid eggs rises. This is because more and more parasitized hosts die without producing adult wasps under the scramble model. At low densities, however, there is very little difference between the two. This is well illustrated by the Aphidius 68
Figure 3.6. Offspring expected on the basis of scramble and contest
competition between larvae in Uiaeretiella ra:pae plotted
against the actual offspring observed.
A = 45° line. Points lying all on this line would show
perfect prediction (y = 0.00 + 1.00x).
B = regression of contest values on observed. a/sea = t = 0.26. Not significantly different from zero. 56 degrees of freedom. b-1 = t = 3.56. P <0.001, significantly different from se b unity.
C = regression of scrample values on observed. a/sea = t = 8.52, P <0.001, significantly different from zero. 1-b - t = 19.49, P <0.001, significantly different from se b unity.
Note: Some of these points occur more than once. (see table B3.6 in Appendix B) n
69
90 x = contest points Line B is regression line y = 0.08 + 1.03x
60 ing r ffsp d o
te • c e 034 00
Exp 0
30 x o = scramble points Line C is regression line y = 7.75 + 0.57x
30 60 Observed offspring 70
40
x = Contest (Line B) y = -0.01 + 1.03 x
0 30
7.4 a, ail 20
o = Scramble (Line C) y = 0.32 + 0.91 x
10 20 30 40 Observed offspring
Figure 3.7. Offspring expected on the basis of scramble and contest competition between Aphidius matricariae larvae plotted against the actual offspring observed. A = 45°C. Points lying on this line would show perfect prediction (y = 0.0 + 1.0 x). I = Regression of contest values on observed offspring. a - t = -0.08. Not significantly different from 0 sea with 34 degrees of freedom. b-1 = t = 0.04. Not significantly different from 1. seb
C = Regression of scramble values on observed offspring. a = t = 2.20, P <0.05, significantly different from 0. se a 1-b = t = 7.47,P <0.001. Significantly different from 1. se 71 plots. As discussed earlier, the rates of parasitism for this species are low, due to the use of an unfamiliar plant. Nonetheless, a regression analysis does show that the scramble values differ significantly in both intercept and slope from those expected from a correct model.
The percentage of females in the offspring declines with parasitoid density in Diaeretielia (see figure 3.8), but is apparently unaffected in
Aphidius. It is possible that encounters between adult wasps could affect the reproductive physiology of the insect, so that more unfertilized eggs are laid. This would have the effect of reducing competition between females in the next generation, thus shortening the time lag between the rise in population density and the compensating response to bring it down again. The shorter the time lag, the more quickly a population will return to equilibrium, with a smaller tendency to overcompensate and cause oscillations. The stability of a population is therefore increased by a mechanism which shortens time lags.
3.3. Meal Moth Parasitoids
3.3.1. Materials and Methods
The parasitoids Nemeritis canescens and Bracon hebetor, and their host
Plodia interpunctelia were cultured as described in Appendix A (p. 311).
The experiments were performed in large polystyrene boxes, with an introduction hole and a large ventilation window in the lid (see Appendix
A, p. 311, and Plate 2). Standard experimental medium was made up in the proportions 1 kilogram Middlings wheat bran, 100 gm. yeast and 50 mis glycerol. For each replicate of the Nemeritis experiment 55 gms. Of standard medium was used, to give a depth of approximately 0.5 cm (the length of the ovipositor of a medium-large Nemeritis). To this were added 0
72
100 y = 63.43 - 0.98 x
r
0+
Parasitoid density
Figure 3.8. The decline of sex ratio with parasitoid density in Diaeretiella rapae.
Note: The regression equation for this data was calculated
using the test for a linear trend in proportions given iu
Snedecor and Cochran (1967).
The normal deviate S = b = -2.59, P <0.002 Seb 73
128 large 5th instar Plodia larvae. These were extracted from three week
old cultures, which were stirred gently and then placed on a warming tray
at full heat. After about half an hour most of the large larvae became
very active, and could be picked off the surface of the medium with a
pair of forceps. The larvae were selected for size by eye, as any other
method would have involved a prohibitive amount of time. A random sample
of 100 such selected larvae were 9.840 mm long with a standard deviation
of 0.861. The box containing the standard larvae was transferred to 25°C
and left overnight.
Newly emerged Nemeritis adults (up to 24 hours old, 8.56 ± 0.26 mm
long), were counted into small clear polystyrene boxes (14 x 8 x 16 cm),
containing a cotton wool pad soaked in dilute honey solution, and left
overnight at 25°C. All the parasitoids to be used in a single replicate
were standardized in a box together.
The following morning, silk was removed from the lid and sides of the
experimental cage, down to the level of the medium layer, with a paper
towel. This silk, deposited by the larvae overnight contains a secretion
of the mandibular glands and both attracts the Nemeritis females and
stimulates them to probe (Mudd and Corbet, 1973; Waage, 1977). By
removing this from the lid and sides of the cage it was hoped to avoid undue time wasting by the parasitoids, and concentrate their searching activity on the medium layer, where the hosts were to be found. The wasps were cooled, to reduce activity, and then removed from their standardization boxes with a pooter. They were then left at 25°C for 15 minutes in the collecting tube, to warm up, before being introduced to the experimental box. The honey pad, with which they had been standardized
(remoistened, if necessary, with fresh honey solution), was also 74
transferred. Without the honey pad, the mortality of the parasitoids
during the 24 hour period was unacceptably high. After the removal of
the principals, the box was kept at 25°C for 2-3 weeks, during which time
the adult Plodia and Nemeritis emerging were counted and noted down. When
no new Nemeritis had emerged for 3 days, the box was deep frozen.
When convenient the box was thawed and the remains of the Plodia
population were examined and counted. 'Excess silk was dissolved away by
a warm 1 in 4 bleach solution, left for at least half an hour. The
medium containing the insect remains was then drained and rinsed in cold
water using a domestic flour sieve. Once this was spread out on a large
white enamel tray in a very shallow depth of water, dead larvae and pupae
and empty pupal cases were fairly easily found. These were then sorted
and counted.
Dead Plodia pupae, and, when possible, dead Plodia larvae, were
dissected and searched for Nemeritis larvae.
Conditions used for Bracon were as similar as possible to those
initially employed in the case of Nemeritis, but the amount of medium was
drastically reduced to a very thin layer, such that, with the aid of a mirror propped at an angle below the box, all parasites (and Plodia larvae) above or below the layer of medium, could be clearly seen (see Plate 2).
Ten large pinches of standard medium scattered evenly over the bottom of . the box were sufficient. •
Bracon females were standardized in just the same way as Nemeritis, except that an equal number of males of the same size or larger, and five small Plodia larvae (second instar 4-5 mm long) were also added to the standardization boxes. The small larvae were present to allow host 75
feeding by the female (see Chapter One). Males were not introduced into
the experimental boxes.
Since Nemeritis is thelytokous, its reproductive capacity is not
potentially limited by mating efficiency as is the case for Bracon. It
was therefore necessary to exclude the effect of density upon mating
efficiency. Known virgin females were standardized as above, and then
placed with three large Plodia larvae each in a 5 x 2.5 cm tube, and left
at 25°C until all three larvae were paralysed. The parasitoids were then
removed and returned to culture, and their offspring reared to emergence
and scored for sex ratio. The sex ratio of wasps emerging on the same day
were noted. Among the offspring of mated females, a maximum of six males
were recorded to emerge on one day without females. To qualify as
unmated, therefore, a female had to leave seven or more male offspring.
In many cases less than seven eggs were laid. However, the maximum and
minimum proportions of females mated could be calculated by assigning the
unknowns first to one group and then to the other. There were no
significant differences in the proportions of females mated when the
numbers of pairs per box were varied from 1 to 5 pairs, or when the
proportion of males was varied from 1:9 to 9:1 (10 insects in all). The
mean minimum and maximum proportions mated were 0.825 and 0.935
respectively. Thus females standardized at low densities stood more or
less the same chance of being mated as those standardized at high densities.
When a replicate of the competition experiment was concluded, and the parasitoids removed, the boxes were gently heated on a warming tray for
15 minutes and moving larvae within reach were removed and reared separately, with excess food, in the small round dishes, as used for Bracon 76
cultures. Not only does this reduce the probability of confusing
paralysed larvae with those that die later from other causes, but it also
makes the box contents easier to classify and count.
When there were Bracon pupal cases in the boxes, bleach extraction
was not used as this dissolves them away completely. The medium layer
was thin enough for dead insects and their remains to be removed directly.
An attempt was made to record the distribution of parasitoid pupae per
Plodia larva. This is not particularly easy as the parasites may wander up to 2 cm before pupation.
However, dead pupae and empty pupal cases do not provide information
on egg and larval mortality. To fill in this gap, six replicates of a small-scale experiment were run. Three large pinches of standard medium were scattered evenly over the bottom of a small round dish, as used for
Bracon cultures. 30 large, standard Plodia larvae were added to this, and left overnight. Six pairs of Bracon were standardized as described above.
The next day the six females were transferred to the round dish and left for 24 hours at 25°C. When the parasitoids had been removed the Piodia larvae were examined under the low power of a microscope. Eggs were transferred using a damp paintbrush, so that larvae carried 1, 2, 4, 8 or
12 eggs. The number moved in each case was noted down, as a check against mortality imposed by handling. This was found to be insignificant. Each larva and its accompanying eggs were isolated, and reared until the Bracon adults emerged from their pupae. These were counted and scored for sex and mortality. Some were kept and their lengths measured.
3.3.2. Analysis of Results and Discussion
Figure 3.9 shows the Log a versus Log P plot for Nemeritis. This relationship produces a value for m (the mutual interference constant) of 77
0 0.5 1.0 1.5 Log P
Figure 3.9. The decline in log a (searching efficiency) with log P (parasitoid density) in Nemeritis canescena.
t-test: b = t = -9.97, P <0.001 with 53 degrees of freedom. seb
(Some of these points are double, see table B3.9 in Appendix B) 78
-0.727. This compares fairly well with the value obtained by Hassell
(1971b), (m = -0.69) with insects searching under less crowded conditions.
Bracon, however, is a much more difficult insect to deal with because
of the complications caused by its habit of paralysing its hosts. Benson
(1972) has assumed that paralysis, feeding and oviposition are all part
of a single behaviour pattern, and has observed Bracon females waiting
patiently for about 30 minutes for'the venom to take effect. Ullyett
(1945), and Richards and Thomson (1932) suggested that the wasp first
searches for healthy hosts and paralyses them; and then searches for
paralysed hosts and lays eggs. Benson ascribed the difference in their
observations to the presence of medium (in his own experiments), in which
the host larvae can build feeding tubes, in which they remain while the
venom takes effect. In the experiments reported here, although the
medium layer was very thin, it was sufficient to allow the construction of
tubes. Observations made during the course of these experiments revealed
that the wasp's behaviour is more flexible than suggested by either of the
two previous authors. There is a tendency to remain in one place after
stinging a host, but encounters with other wasps, or an unexpected nudge
from a passing healthy larva, may both stimulate the wasp to move on.
Rather, the behaviour adopted by the insect depends upon the larva
encountered: a moving larva elicits stinging, a motionless host provokes
feeding or oviposition. In addition, hosts stung when in their feeding
tubes often leave the tube and travel considerable distances (5 cm or
more) before paralysis overtakes them.
Thus, the two processes, paralysis and oviposition run concurrently.
Two searching efficiencies are involved, which may be significantly
dissimilar, and time spent ovipositing will decrease the number of hosts 79
paralysed and vice versa.
If a is the searching efficiency of Bracon searching for healthy
hosts, s is the efficiency of search for paralysed hosts, and stinging is
assumed to be instantaneous, then the simultaneous equations:
dX_ a P(N-X) dt 1 + s X Th
dE s P(X-E) dt 1+ a XTh
where N = total hosts.
X = hosts paralysed.
E = hosts with eggs.
Th = handling time for oviposition.
are appropriate. These are mathematically intractable.
As a rough approximation, s was calculated as usual, using healthy
hosts compared with paralysed hosts, and S was calculated on the basis of
paralysed hosts and hosts with eggs found at the end of the experiment.
This is not strictly accurate since the number of hosts paralysed is
changing throughout the course of the experiment, as shown by the above
equations.
Figures 3.10 and 3.11 show Log a and Log s plotted against Log P.
Means were used for the Log s plot because of an appreciable number of
zero points, where hosts were paralysed but no pupae were recovered.
When all the points are included, the slope of the regression is not
significant for Log s. Like the Aphidius and Diaeretielia Log a plots, the efficiency rises to a peak at a density of two wasps. The last four points show a significant negative trend with parasitoid density. The slope of this relationship (-0.413) is similar to that obtained in the 80
-0.2
-0.5
-1.0
(0I 00 0 1-1
-1.5
-1.9 0 0.3 0.6 0.9 1.2 Log P
Figure 3.10. The decline in Log a (efficiency of search for paralysed hosts) with Log P (parasitoid density). in Bracon kebetor.
t-test: b = t = -3.71, P <0.001 with 46 degrees of freedom. seb
(Some of these points are double, see table B3.10 in Appendix B) 81
-1.2 All points: y = -1.51 - 0.20x
Last 4 points: y = -1.32 - 0.41x
-1.5
-2.0
-2.3 0.3 0.6 0.9 1.2 Log P
Figure 3.11. The relationship between Log s (efficiency of search for paralysed hosts) and Log P (parasitoid density) in Bracon hebetor. Means and 95% confidence limits shown.
All points, t-test: b = t = -1.65, not significantly seb different from zero, with 3 degrees of freedom.
Last 4 points, t-tests: b = t = -8.43, P <0.02, with 2 seb degrees of freedom. 82
Log a plot (-0.359), and both are comparable with the value of -0.44
obtained by Benson (1973). Thus the presence of other wasps not only
reduces the efficiency, of search for healthy hosts, but decreases the
rate at which eggs are laid on new paralysed hosts. This could occur in
two ways: by waste of time due to avoidence of superparasitism or
encounters with other wasps, or by a tendency to oviposit on hosts with
eggs already attacked. If time wasting cuts down the number of eggs laid,
then the total number of pupae produced per parasitoid would be expected
to decline with density. Figure 3.12 shows that the relationship is not
a simple one. The data fits a curve in P and P2, rising to a density of
about 11 and falling away after that. Thus, time wasting may be reducing
the number of eggs laid, but it does not come into effect until a wasp
density of 8-16 is reached. As figure 3.13 shows, the number of hosts
parasitized per wasp may follow a shallower curve. While oviposition rate
on new hosts is reduced at a density of two or higher, as shown by the
Log s plot, the number of hosts with eggs, and the number of eggs laid
increases up to a density of eight wasps. This is because the total
number of hosts paralysed, and hence available for oviposition rises with
density.
To investigate this more thoroughly, the distributions of pupal cases
were examined. Benson (1972) reported that eggs were laid in a clumped
distribution (variance exceeding mean), with a mean of about four eggs
per host. As described previously eggs were collected for other experi- ments by confining six Bracon females with thirty host larvae in a small
round butterdish. Only about two or three hosts survived the 24 hours,
and the egg distribution obtained was clumped with a mean of 3.921
(variance = 7.171; pooled from six distributions). 83
8 y = 0.99 + 0.67x - 0.03x2
a 4.1 .4 ao 0 0a.. 4 s+ a.m b u 0 0
m as ā 0 24 4 8 12 16 Parasitoid density
Figure 3.12. The pupae produced per wasp plotted against wasp density in Bracon hebetor. Means and 95% confidence limits shown. t-tests: b = t = 12.12, P <0.01, with 2 degrees of freedom. seb
c = t = -9.74, P <0.02, with 2 degrees of freedom. se c
a .r at m c» ar ot C3* 2
0 N 4J a k 1 a.ce co a 0 0 4 8 12 16 Parasitoid density
Figure 3.13. The number of hosts parasitized per wasp plotted against wasp density in Bracon hebetor. Means and 95% confidence limits shown. No significant relationship is apparent. It is possible that variation may be masking a shallow curve of the same form as that shown in figure 3.12. 84
The variance: mean ratio (S2/m) is plotted against the number of
parasitoids present in figure 3.14. Although the distribution of points
at each wasp density, does not differ significantly from one (i.e. random)
when the density is two or more, there is a significant overall increase
in clumping. Thus, the resulting pupal cases are more aggregated when
the initial number of parents is higher. Cannibalism among the larvae on
the same host would produce a decrease in clumping. Avoidence of super-
parasitism, breaking down at high parasitoid density is a possible
explanation.
However, during observations healthy host larvae have been seen to
eat parasitoid eggs on several occasions, and the number of these around
to do damage decreases, due to paralysis, as the number of parasitoids
increases. It is therefore possible that egg-predation by healthy host
larvae is the cause of change in pupal distribution. If a larva eats all
the members of an egg clump when it finds one, and finds these clumps at
random (i.e. in proportion to their frequency), then the frequency
distribution of egg clumps of different sizes will be altered until each
size of egg clump is equally frequent i.e. random.
Figure 3.15 shows S2/m plotted against the number of healthy hosts
surviving at the end of the experiment, as a rough measure of egg predation. Despite a great deal of scatter, there is a significant decline in aggregation as the number of surviving hosts decreases. However, the slope of the line, and the significance level is lower than that for figure 3.14, the plot against wasp density. This suggests that, while egg predation may be part of the explanation, another factor dependent on wasp density, is contributing to pupal distribution change. Avoidence of superparasitism, as reported by Weseloh (1976) in Apanteles, seems most likely. 85
2.0 y•= 0.39 + 0.05x x x x x
x x
1.0
0.0
0 8 12 16 Parasitoid density
Figure 3.14. S2/ts (clumpiness of egg distribution) plotted against parasitoid density in Bracon hebetor, where S2 = variance and m = mean of the egg distributions.
t-test: b = t = 2.89, P <0.01 with 34 degrees of freedom. seb
(Some of these points are duplicated, see table B3.14-5 in Appendix B). 86
x
x . y = 1.225 - 0.008x
X
x x x a N ui
0 50 100 Hosts surviving
Figure 3.15. S2/m (clumpiness of egg distribution) plotted against the
number of hosts surviving at the end of the experiment,
after exposure to Bracon hebetor, [ s2 = variance,
m = mean of egg distribution
t-test: b/seb = t = -2.41, P <0.05, with 34 degrees of freedom'. 87
Values of bTw can be calculated from a for Nemeritis and Bracon, and
also from s for Bracon. Figure 3.16 shows bTw plotted against (P-1) for
Nemeritis. This relationship is not significant. Thus, either b and T w are independent of parasitoid density, or they vary with density, in
opposite directions such that their product remains the same.
Figures 3.17 and 3.18 show bTw plots for Bracon. While bTw
calculated from s shows no significant trend with parasitoid density,
Log bTw calculated from a shows a significant decreasing relationship with
Log (P-1). The rate of encounter with other wasps, b, must be the same
in both cases. However, the total time wasted per encounter must be
partitioned between the times available for stinging and oviposition. It
is the variation in these two components which is shown in figures 3.17
and 3.18. i.e. TwT=T wa + Tws
where TwT = total time wasted
Twa = time wasted decreasing search before stinging
Tws = time wasted decreasing search before oviposition
Therefore the graphs show that Tws remains more or less constant, while
Log Twa declines with Log (P-l).
The scramble and contest models were applied to the Nemeritis data, and the resulting graphs are shown in figure 3.19. In this case, dead single parasitoids (i.e. found in sole possession of a host) were included with the healthy offspring and used as observed values. As expected, the plots indicate predominantly contest competition. This occurs very early on in the life cycle, as shown by the k-value plots described below. 88
3.0►.
2.0
H .nI
1.0
I
0.0 0 8 16 24 32 P - 1
Figure 3.16. bTw (b = rate of encounter between wasps, T = time wasted y per encounter) plotted against P - 1 (P = number of wasps
present) in Nemeritis canescens.
Means and 95% confidence limits shown, (calculated from a
values, see table B3.16 in Appendix B). No significant
relationship is apparent. 89
0 0.3 0.6 0.9 1.2 Log (P-i)
Figure 3.17. The decline in Log bT( calculated from a) with log (P-1) in Bracon hebetor, where b = rate of encounter between adult wasps, Tw = time wasted per encounter, P = parasitoid density t-test: b = t = -3.73, P <0.0 with 32 degrees of seb freedom.
(Negative values of bTw have beea omitted. See table 83.17) 90
0 12 16 P-1
Figure 3.18. bTw (calculated from s, the efficiency of search for
paralysed hosts) plotted against P-1 for Bracon hebetor.
(b = rate of encounter between wasps, T = time wasted per w encounter, P = parasitoid density). Means and 95%
confidence limits shown, see table B3.18 in Appendix B. 91
Figure 3.19. Expected offspring (based on the assumptions of scramble and
contest competition between larvae) plotted against healthy
observed offspring plus dead fully formed pupae in sole
possession of a host, in Nemeritis canescens.
Some of these points are double (see table B3.19 in Appendix B)
A = 45° Line. Points lying on this line would show perfect
prediction (y = 0.0 + 1.0x).
B = Regression of contest values on observed offspring
a = t = 1.22. Not significantly different from 0 with 53 sea degrees of freedom.
b-1 = t = 1.43. Not significantly different from 1. seb
C = Regression of scramble values on observed offspring
a = t = 9.45, P <0.001, significantly different from 0. se a
1-b = t = 19.75, P <0.001, significantly different from 1. seb 92
90 x = contest points Line B is regression line x y = 1.14 + 1.O3x
x
60
ing r
ffsp O x d o te c e Exp 30
o = scramble points Line C is regression line y = 9.95 + 0.56x
0 0 30 60 90
Observed healthy offspring +:dead singles 93
As a base to work from, the maximum number of offspring per para-
sitoid obtained from the complete data set, was used as the maximum potential number of eggs that could be laid under these experimental conditions. This is not strictly accurate since neither density- independent egg mortality nor competition between the offspring of a single female is included. As mentioned in Chapter Two
k-value = Log C Initial Number L Survivors
Neneritis offspring recovered from the experiments fall into three groups: supernumerary, dead; single, dead; and healthy adults. Four sets of k-values can be calculated from these as shown in figure 3.20.
Figure 3.21 shows the overall mortality, while 3.22 - 3.24 show the egg and early larval mortality (including eggs that are not laid), late larval mortality due to superparasitism, and the mortality of solitary parasitoids, respectively. The last is probably not a competitive mortality. It is clear from these graphs that most mortality due to competition occurs during the very early stage (k1).
No information is available on eggs or larvae encapsulated and destroyed soon after oviposition. However, Salt (1975) reports that in
Plodia, there is no haemocytic reaction to eggs or healthy larvae, and only a local reaction to wounds on supernumerary larvae. Sometimes, melanized first or second instar larvae were recovered from dissected
Plodia pupae containing advanced dead Nemeritis larvae. These were assumed to be the offspring of the second generation, which start probing immediately after emergence. In no cases were there any traces of cellular encapsulation. In the small scale experiments described in Chapter Seven, healthy and paralysed hosts were dissected five days after one oviposition K = kl + k (total mortality 2 + k3 — fig. 3.21)
Healthy adult offspring
k3 (fig. 3.24)
Adults * Total eggs Total parasitoid Single parasitoids parasitoids available remains found (1 per host) i k2 (Fig. 3.23, 3.25) k1 (fig. 3.22)
a)Death of eggs, Supernumerary young larvae parasitoids (> one per host) b)Avoidence of All dead superparasitism
c)Reduction in the time available for search
Figure 3.20. The breakdown of mortality imposed by intraspecific competition, during the life cycle of Nemeritis
canescens. 95
1.2 2.0 3.0 3.4 Log (initial eggs)
Figure 3.21. The increase in K-value (total mortality) with increase in
initial density of eggs in Nemeritis canescens.
t - test: b = t = 9.93, P <0.001, significantly different seb from 0, with 53 degrees of freedom. 96
Figure 3.22. k1-values plotted against Log (initial egg density) in
Nemeritis canescens (see figure 3.20).
t-test: b = t = 12.62, P <0.001, with 53 degrees of seb freedom.
Figure 3.23. k2-values plotted against Log (population density) in
Nemeritis canescens (see figure 3.20).
Figure 3.24. k3-values plotted against Log (population density) in
Nemeritis canescens (see figure 3.20). 1.8 97
1.0
0.0 Log (initial eggs)
1.8
1.0
0.0
1.0
0 0
x x x Cn x xx X x xse ,e' x4 xx ,j~ i% 0.0 ,r x if '~ i 1 0.6 1.0 2.0 3.0 3.4 Log (Population density) 98
was allowed in each. No cases of encapsulation or melanization were
observed, so it is very unlikely that a host immunity reaction contributes
to k1.
As described in Chapter Four the intraspecific competition experi- ments were repeated to gain extra information on the behaviour of the wasps. These were terminated immediately after the last observation
period i.e. after 9i hours, and bred through. Values for K, Log a and
expected offspring based on the larval competition models, were plotted
and the regression equations calculated. However, since these are mostly of exactly the same form as the graphs obtained for the 24 hour runs, they are omitted when they provide no extra information.
Figure 3.25 shows the k-value plot for late larval mortality, due to superparasitism for these short experiments, equivalent to 3.22 for the
24 hour runs. This graph shows that significant, density-dependent scramble competition between parasitoid larvae does occur when both survive past the first two instars. It is interesting that this is not detectable in the 24 hour experiment. This may be because competing larvae in the short experiment are closer together in age, and therefore more evenly matched.
In Bracon, larval remains are not found since they tend to be cannibalized by the other larvae on the same host. In no case was a host found with a single dead larva. The k-values for pupal mortality did not show a significant density dependent relationship.
In the small scale experiments, when the number of Bracon eggs per host were varied, the k-values for total mortality, and those for the component egg and larval mortality, and pupal mortality were not signi-
99
0.2 1.0 2.0 Log (total remains)
Figure 3.25. k2-values (death of supernumerary larvae) plotted against
log (total remains) on which they act, in Nemeritis
canescens (short term experiments).
t-test: b = t = 2.35, P <0.05 with 36 degrees of freedom. seb 100
ficantly related to the number of eggs per host. Most mortality occurred
before pupation.
While mortality was not significantly affected by the number of eggs
per host, in the density range tested, the length of the adults was.
Figures 3.26 and 3.27 show the relationships between the length in milli-
metres of males and females and the number of eggs per host. Both these
relationships are significant. (Adults were measured from the top of the
head to the end of the abdomen, ovipositors excluded). Reduction in size
with increasing density is one of the characteristics of scramble
competition, indicating that resource limitation has begun to affect the
individual, although it has not actually died.
While there are a number of intermediate density effects on Bracon
reproduction and survival, there is no significant overall trend in
k-value against Bracon potential offspring (maximum recovered x number of
wasps present), shown in figure 3.28. This is probably due to a counter-
balance of opposing affects. While Bracon searching efficiencies (reflecting
the number of encounters between wasp and hosts) do decline with wasp
density (figures 3.10 and 3.11), the average number of resulting pupae
per parasitized host appears to rise with the number of wasps present (see
figure 3.29). It is suggested that this is due to the relaxation of egg
predation by healthy hosts, as more are paralysed with rising wasp density.
Some support for this is gen by figure 3.30, which suggests a decline in
the mean pupae recovered as the number of healthy hosts surviving increases.
However, there is so much variation shown in figures 3.29 and 3.30 that more replicates would be desirable to confirm the relationships.
Unlike the previous pair of wasps, there is a very large gap between the reproductive strategies of Nemeritie and Bracon. Not only is one 101
Figure 3.26. The decline in the length of male Bracon adults emerging as
the initial number of eggs per host rises.
t-test: b = t = -2.37, P <0.05 with 20 degrees of freedom. seb
* = multiple points, see table B3.26.
Lengths do not include antennae.
Figure 3.27. The decline in the length of female Bracon adults emerging
as the initial number of eggs per host rises.
t-test: b = t = -3.44, P <0.002, with 37 degrees of seb freedom.
* = multiple points, see table B3.27.
Lengths do not include antennae or ovipositor. Length of females (mm) Length of males (mm)
I- N • 0
x K X 103
1.0 1.5 2.0 2.2 Log (initial egg density)
Figure 3.28. Total mortality (K-values) plotted initial egg density for Bracon hebetor.
* = multiple points (see table 83.28). Figure 3.29.Theincrease Mean pupae per host recovered 0.0 1.0 2.0 3.0 3.6 4 0 t-test: b=5.50, P <0.02, 95% confidence host withthenumberof se b
in themean limits Log (populatiosdensity) 104 shown. Bracon pupae 0.6 with 3 recovered females degrees present. Meansand per parasitized of freedom. 1.2 Figure 3.30.Thedecline inthemeanpupaerecoveredper host withthe
Mean pupae recovered per paralysedhos ts 0
t healthy hostssurviving theexperiment. -test: b=t se b
= 3.32,P<0.002;with 46degreesof Healthy hostssurvivingexperiment 40 105 freedom. 80
120 106
solitary, the other gregarious, but Bracon requires the production of
both sexes to persist. This can have a very far-reaching consequences as
is shown by Fisher's (1961) study of Nemeritis and Horogenes chzrysostictos.
The necessity of producing males drastically reduced Horogenes'
competitive ability.
The sex ratio of adults emerging did not vary significantly with
either the number of original parasitoids, the total number of pupae from
each box (mean females/total = 0.439), or the number of eggs or pupae.
3.4. Summary
1. a) Plots of Log a (searching efficiency) for Diaeretiella and Nemeritis
show a declining relationship with Log P On = -0.375, -0.727,
respectively), as is already commonly found in the literature.
b)In Aphidius this relationship is insignificant, which may be because
of the low rates of parasitism produced by the experimental
conditions.
c) Bracon is markedly different from the other three in having two
phases of search i) to find healthy hosts for stinging and ii) to
find paralysed hosts for oviposition and host feeding, so there are
two searching efficiencies. Log a (searching for healthy hosts)
follows the usual declining relationship with Log P On = -0.359),
as do the last four points of Log s (searching for paralysed hosts)
versus Log P On = -0.413).
d)An increase in Log a for DiaeretieZla and Aphidius, and in Log s for
Bracon between the densities of one and two wasps per cage suggests
that the presence of other wasps may stimulate search at low
densities. 107
2. The compound parameter bTw (calculated using Beddington's (1975) model
for mutual interference) is more or less constant with wasp density in
Aphidius and Nrocritis. In DiaeretielZa the relationship is curved.
In Bracon it is constant when calculated from s, and declines with
(P-1) when calculated from a.
3. The expected values calculated from simple models of scramble and
contest competition among wasp larvae, show clearly, as expected, that
Diaeretiella, Aphidius and Neneritis larvae compete predominantly by
contest.
4. Plots of k-values for Neneritis show that most competitive mortality
occurs in the early stages; before or just after oviposition, although,
when the larvae are close together in age, significant scramble
mortality may occur in the late larval instars.
5. In Bracon, pupal deaths are not significantly related to initial
density; and egg/larval and pupal mortality are not significantly
related to the number of eggs per host, up to a density of 12. How-
ever, the lengths of the resulting males and females are significantly
reduced.
6. The distribution of Bracon pupae per host larva becomes significantly
more clumped as the initial density of parasitoids increases. This is
at least partly due to egg predation by active host larvae. Avoidence
of superparasitism, breaking down at higher densities, may also
contribute.
7. There is no overall effect of conspecific density on Bracon performance.
However, the searching efficiencies for healthy and paralysed hosts do
decline with density. It is suggested that these effects are counter- 108 balanced by a relaxation of egg predation as the number of active healthy hosts declines with Bracon density. 109
CHAPTER FOUR
COMPETITION WITHIN SPECIES II.
BEHAVIOURAL OBSERVATIONS ON NEMERITIS CANESCENS
4.1. Introduction
In this Chapter, detailed behavioural observations made during the
course of the experiments on Nemeritis, described in chapter three, are
analysed and examined for trends with parasitoid density.
Continuous records were made of the behaviour of a single individual,
which was classified into three main activities: probing, walking and
resting. The time spent on the medium layer was also recorded, as were
the two classes of point events: cocking, and encounters between wasps.
Previous work (Hassell, 1971b; Hassell and Rogers, 1972; Hassell and
May, 1973; Rogers and Hassell, 1974) has suggested that interference
relationships are the results of a decrease in available searching time,
due to time wasted after encounters between searching insects. Attention
is therefore concentrated on searching time, and the changes in behaviour
occurring immediately after an encounter.
4.2. Materials and Methods
Individual wasps, under the varying density regimes described in chapter three, were observed continuously for 30 minutes, beginning 15 minutes after the parasitoids were first introduced. Their behaviour was recorded using a four-channel Rustrak event recorder (see Plate 4). This is controlled by four switches, which can be off or on. Each switch controls a needle, which scratches a line on specially prepared chart paper, which is driven along at a rate of 30 inches (76.2 cm) per hour. 110
When the switch is turned off, the needle moves sideways to continue the trace in the "off" position. The resulting behaviour trace consists of four parallel tracks, each of which is a single continuous line in one of two positions across the paper (see Plate 5). Changes in behaviour are marked by a short horizontal stroke. The length of the line scratched when the switch is on gives the length of time the animal has been engaged in a particular behaviour. When there are three mutually exclusive occupations, the measurement of two lines gives the third when both the switches are at "off". Point events, such as an encounter between searching wasps, can be recorded by switching on and off again immediately.
One switch, therefore, can be used to record one occupation and one set of point events.
The periods of continuous behaviour patterns were measured, correct to three seconds, and recorded in sequence, with the exact time of point events.
Preliminary analysis did not reveal a relationship which could explain the strong mutual interference relationship found in chapter three, so further observations were made later in the day. These were less detailed, using only a stop clock, and were made three, six and nine hours after the original observations. Point events were recorded by a key letter in sequence, but their exact times were unknown. One occupation only (probing) was fully recorded. The stop clock was activated when the insect started to probe, and stopped when it discontinued, and the time shown (correct to the nearest second) noted.
The total time spent searching was shown on the stop clock at the end of the observation period, and the lengths of searching sessions obtained by subtracting successive readings. The adult female Nemeritis searches for 111
hosts by testing the host medium with her ovipositor. The abdomen is
curled under the thorax so that the ovipositor points forwards. This is
then inserted into the medium layer as far as the ovipositor can reach,
withdrawn and inserted again. The host is generally found by contact.
During search, the tips of the antennae test the surface of the medium.
The whole pattern is called "probing" and is very easily identified. The
term is used interchangeably with "searching" in the following chapters,
when referring to Nemeritis.
4.3. Analysis of Results and Discussion
Previous work has shown (Hassell, 1971b) a reduction in the
proportion of time spent searching with the log of the parasitoid density.
This, then, was one of the expected outcomes of the present series of
observations. However, the data collected did not show a significant
relationship of this kind. When repeated later in the day, the
observations of time spent probing showed the same pattern of rise and
fall found first thing in the morning, as illustrated in figure 4.1.
This suggests that the pattern is a real feature of the insect's
behaviour, not merely the result of random fluctuation.
With one interesting exception, the observation periods later in the
day produced very similar results to those obtained in the early morning,
but generally at lower levels of activity.
Cocking movements (shown by Rogers, 1970, 1972b; to demonstrate the replacement of an egg in the cavity at the tip of the ovipositor, and
hence that an egg has been recently expelled) also show no significant
relationship with wasp density (figure 4.2).
Figure 4.3 shows that the number of encounters between searching 112
Q 8 16 24 32
Figure 4.1. Percentage time spent probing plotted against parasitoid
density in Nemeritis canescens.
A. After 15 minutes
B. After 3 hrs. 15 min.
C. After 6 hrs. 15 min.
D. After 9 hrs. 15 min. 11.3
12
6
24
1 8 1 J. 16 Parasitoid density 32 Figure 4.2. Number of cocking movements plotted against wasp density in
Nemeritis canescens in 30 minutes observation.
200
• y = -0.25 + 3.56x E 0
as 0 100
as ar .1 aa a► aJ
Q ua
8 16 24 32 Parasitoid density - 1
Figure 4.3. The increase in the number of encounters between wasps as
wasp density increases in Nemeritis canescens.
t-test: b = t = 19.99, P <0.001 with 3 degrees of freedom. seb 114
wasps rises with density, with an encounter rate, b, of 3.56. ['In this
figure, and the last, means were calculated from all four observation
periods, and weighted accordingly when the variance due to time was
similar to the error variance, as suggested in Bliss, volume 1, 1967.2
Hassell (1971b) found a curvilinear relationship between encounters and
the number of parasitoids present (in x and x2). He found a coefficient
of x of 4.19, which compares reasonably well with the figure of 3.56.
The discrepancy between the two relationships may be accounted for by the
slightly different experimental conditions used: he watched for 15 minute
periods and the hosts (a different instar of a different species) were
distributed in patches of different density.
This would be expected to cause local differences in parasitoid
density, with areas of high wasp density (aggregation, see Hassell, 1976)
over the host patches, and relatively low parasitoid density in the rest
of the cage. The frequency of encounters between wasps would therefore
depend on where in the cage an individual is, which may also restrict what it is doing. For example, probing occurs only over the host patches.
There is some evidence that encounters between wasps can cause local
dispersion (see Hassell, 1976). This would have the effect of moving
them from an area of high density (encounter rate high) to an area of low wasp density (encounter rate low). It is possible that this kind of local variation in density might well produce a non-linear relationship between encounter rate and the number of parasitoids, compared with the simple system studied here. The effect of host patches and hence local variation in parasitoid density is discussed more fully below.
The mean length of a searching session declines with the log of parasitoid density, as shown in figure 4.4. This is probably due to the 115
A. y = 20.46 - 4.92x B.y = 7.71 - 1.30x
0
0 a
A ii
I O
a a 0 0 0 0 0 0
0
0 0.3 0.6 0.9 1.2 1.5 Log (parasitoids present)
Figure 4.4. The decline in the mean length of probing bout with parasitoid density in Nemeritis canescena.
A After 15 minutes o t-test: b = t = -3.17, P <0.01, with 46 degrees of seb freedom.
B } After 3 hrs, 6 hrs, 9 hrs 15 minutes x t-test: b = t = -2.97, P <0.01, with 96 degrees of seb freedom. 116
number of encounters between wasps, as the relationship between mean length
and encounters (figure 4.5) is of the same form, but apparently more
consistent. However, neither relationship is particularly convincing.
The number of searching periods is not significantly related to
density, but seems to show a curved relationship when plotted against the
number of encounters (figure 4.6).
These two parameters (mean length and number of search periods) are
potential sources of the Nemeritis interference relationship shown in the
previous chapter. If either were to decline significantly with the number
of encounters between wasps (and hence with the number of wasps present)
a corresponding decline in the total time spent searching would be
expected, provided the other parameter did not increase simultaneously.
This could then decrease the rate of encounter between a wasp and its
hosts, and produce an interference relationship. However, not only is
there a good deal of scatter in these graphs (figures 4.4 to 4.6), but the form of the relationship is not of the kind to produce a strong interference effect. The decline in the mean length of searching period is only very slight, quickly levelling off; while the number of such periods rises and falls, but only slightly. The interaction of two such effects would be expected to produce an irregular but consistent (i.e. repeatable) rise and fall in the total time spent searching, which is the result observed. Interference in Nemeritis under these conditions there- fore cannot be a "time waste" effect.
As suggested by the last few figures, the major effect of the time since the start of the experiment is to reduce activity. This is clearly shown by figures 4.7 to 10. Surprisingly, the number of encounters be- tween searching wasps appears to reach a peak after three hours and then
117
30
x
X X
y = 12.56 - 2.28x
x
) ec 20 x x (s x t x x X x x
h bou X X x rc x xx x x x xk h sea x
t x X x x x x x x x x
leng x x x 10 x
Mean x x X X x x x x x x x x x x x x xx x x x x x x x x x x x x x x x x x xx xxx xAx X x A x x xx x x xx x x x xx x x x x x x x x X
0
0 1.0 2.0 2.6
Log (Encounters + 1)
Figure 4.5. The decline in the mean length of search bout with the
number of encounters between Neweritis adults.
t-test: b = t = 3.37, P <0.001 with 145 degrees of freedom. seh
100
0 0 O 0 A. y = 53.6 + 0,54x - 0.003x2
• x o 0 B. y = 37.01 • + 0.35x - 0.001x2 x x • xx
4 e• • o 0 ds io r e x x p
h W 50 • x ° x x x 0 x
earc • x x • x x x 0 x f s o ber m Nu
50 100 150 250 230 Number of encounters between parasitoids
Figure 4.6. Number of search periods plotted against the number of encounters between wasps in Nemeritis oanescens.
A,o - After 15 min. F-ratio, with 2, and 46 degrees of freedom = 7.33, P <0.01.
B,x - All points. F-ratio, with 2 and 154 degrees of freedom = 6.64, P <0.01. 119
y = 44.24 - 39.50 logx
t-test: b = t = 288.10, P <0.001 with 2 seb degrees of freedom. ing b ro
p 30 t n e sp ime T
5 10 Hours after experiment start
Figure 4.7. The decline in time spent probing by Nemeritis canescens as the experiment progresses.
y = 6.39 - 6.50 logx
t-test: b = t = 12.28, P <0.01 with 2 in.
m seb degrees of freedom.
30 in
ing k
Coc 5 10 Hours after experiment start
Figure 4.8. The decline in cocking movements in Nemeritis canescens as the experiment progresses.
y = 69.57 - 5.40x t-test: b = t = 6.74, P <0.05 with 2 seb degrees of freedom
ds io er h p c ear f s o ber Num
0 5 10 Hours after experiment start Figure 4.9. The decline in the number of search periods in Nemeritis
120
20 log y = 1.10 - 0.34 logx
t-test: b = -5.89, P <0.05 with seb 2 degrees of freedom
) c e (s d
io er 10 h p arc e h s t ng le
Mean
0 0 5 10 Hours after experiment start
Figure 4.10. The decline in the mean length of search period in Nemeritis
canescens as the experiment progresses.
160 Using 3 points: y = 92.00 - 7.43x t-test: b =t=30.04, P<0.05 seb with 1 degree of
freedom s wasp een
tw 80 be
rs te coun En
5 10
Hours after experiment start
Figure 4.11. The decline in the number of encounters between wasps as the experiment progresses. When all four points are included the relationship is not significant. 121 decline (figure 4.11). This may be an artificial result, arising from the change in the method of observation. On the behaviour traces, encounters occurring within three seconds of each other, could not easily be distinguished.
The percentage time spent walking rises with the log of wasp density, while the time spent resting falls, as shown by figures 4.12 and 4.13.
The components of walking time also show consistent trends with density.
The mean length of a walk may be described by a parabolic relationship with the number of wasps, and the number of walks rises with log wasp density (figures 4.14 and 4.15).
The total time spent walking or resting is related linearly to the mean length and number of behaviour sessions, (figures 4.16-4.19). This shows that variation in walking and resting time is significantly affected by both components of walk and rest. Under other circumstances it would be possible for one component to vary and hence cause variation in the total time while the other remained more or less constant. In this case there would be no obvious relationship between the constant component and the total time, although, of course, if it were forced to change the total would be changed too.
However, an examination of the probing data in the same way show an interesting difference. Data from the last three observation periods show a straightforward linear relationships between the percentage time spent searching and the mean length and number of searching bouts. However, when the data from the first set of observations is included, the relation- ships are changed, to a parabolic one between time searching and mean length of search period, and a log dependence on the number of search bouts. This is shown in figures 4.20 and 4.21. This suggests that, at 122
70 y = 23.79 + 9.66 logx.
x x x x x ing k l
a 35 w t en
sp x ime T % 0 0 8 16 24 32 Parasitoid density.
Figure 4.12. The rise in time spent walking with parasitoid density in Neneritis canescens. t-test: b = t = 3.26, P <0.01 with 40 degrees of freedom. seb
40
y = 16.84 -- 5.49 logx
x x
ing 20 t res t n e sp e im
% T 0 L • 0 8 16 24 32 Parasitoid density
Figure 4.13. The decline in the time spent restive with parasitoid density in Nemeritis canescens. t-test: b = t = -2.25, P <0.05, with 40 degrees of freedom seb 123
Figure 4.14. The mean length of walk periods plotted against parasitoid
density, in 30 minutes observations of Nemeritis canescens.
t—tests: b = t = 3.35, P <0.05, with 3 degrees of seb freedom
c = t = 3.05, P <0.05 se c
Figure 4.15. The increase in mean number of walk periods with parasitoid
density.
t—tests: b = t = 2.87, P <0.01, with 40 degrees of seb freedom 124 13
F = 5.58 + 0.30x - 0.008x2
10
) c d (se io er p k l h wa t leng n Mea
0 16 24 32 Parasitoid density
110 y = 69.29 + 12.48 logx
100
0 8 16 24 32 Parasitoid density 125
Figure 4.16. The dependence of time spent walking on the mean length of
walking bout in Nemeritis canescens.
t-test: b = t = 10.59, P <0.001, with 41 degrees of seb freedom.
[The point (16.62, 61.33) has been omitted from the graph
Figure 4.17. The dependence of time spent walking on the number of
walking bouts in Nemeritis canescens.
t-test: b = t = 4.21, P <0.001, with 41 degrees of seb freedom.
Figure 4.18. The dependence of time spent resting on the mean length
of resting period in Nemeritis canescens.
t-test: b = t = 9.60, P <0.001, with 41 degrees of seb freedom. 126
60
3 0 30 0.
0
0 4.00 6.00 8.00 10.00 11.00 Mean walk period (secs)
x 60 y vP 0.88 + 0.38x x
x x
ing x x x lk x x x31 x wa x t x xx 30 x en x x x x x x x x x x sp X x
e x X x x x im x T x
0 a 40 60 80 100 110 Number of walk periods
y = -3.58 + I.88x x
ing t res t en sp ime % T
10 15 19 Mean rest period (secs) Figure 4.19.Thedependenceofpercentagetimespentrestingonthe
%T ime spent rest ing t-test: b=t6.10,P<0.001,with41degreesof Nemeritis canescens. number ofrestperiods,in30minutesobservation 0 se.
Number ofrestperiods 127 y = -2.58 +0.58x 25 freedom. 50 128
Figure 4.21. The dependence of time spent probing on the number of
search (= probing) periods, in Nemeritis canescens.
A. All points, including those recorded after 15 min (o).
t = b = 16.16, P <0.001 with 145 degrees of freedom. seb
B. Last 3 observation periods (x).
t = b = 19.87, P <0.001 with 106 degrees of freedom. seb
Figure 4.20. The dependence of time spent probing on the mean length
of probe (= search) period in Neneritis canescens.
o = observations after 15 minutes from experiment start.
X = tI " 3 hrs 15 min. 6 hrs 15 min. } from experiment start 9 hrs 15 min. JJ1 129
0 . Log y = 0.47 + 0.02x
0 00 O • o 00 O 0 O O O O O 0 0 0 0 0 •
ing
b 0
o x r p t
en . y = -0.90 + 0.40x sp ime T Z
50 100 Number of search periods
80 • • 0 0 0 O 0 • • •
ing - o b O 0 ro • •
p 40 0 t • y = -20.61 + 6.52x - 0.12x2 en b = t = 13.09, P <0.001 sp • seb ime x
T X x
Z x x c = t = 7.54, P <0.001, with x ,~ x x sec 144 degrees of freedom x
20 40 Mean length search periods (sec) 130 high levels of activity, a balance between number and mean length is set up, which maintains the percentage of time spent searching at roughly the same level.
This is clearly shown in figure 4.22. Late in the day, mean length and number of search periods are independent, but there is a strong inverse relationship first thing in the morning. This may be because, having been deprived of hosts for 24 hours, and then exposed to a relatively high host density, there is a strong drive to search. Thus, when the mean length of search is reduced (due to encounters) there is a tendency to increase the initiation of search i.e. increase the number of search periods (but see page 148 and chapter seven p. 231).
While there are several clearly defined relationships in behaviour patterns with parasitoid density, the effect of a single encounter between searching adults is not immediately obvious. The continuous behaviour records were therefore examined more closely.
Each encounter was classified according to whether or not a change in behaviour occurred within three seconds. Where changes did occur, the direction was recorded i.e. which of the two other behaviour patterns then followed. Thus, if P = probe, W = walk and R = rest, encounters were classified into nine different categories: PP, PW, PR; WW, WP, WR;
RR, RP and RW.
Initially, it was hoped to show that encounters caused change in behaviour. In order to ascertain this, a set of data giving the expected frequency of change when no encounter occurred was- required. This was generated as follows. Since the behaviour traces could be measured correct to three seconds, there were 1800 possible "moments" when an encounter, or 131
100 y = 94.21 - 1.85x
x 0 O 0 0 00
0 0 0 0 0 XOx 0 wasp density O 0 •
0 0 0 x xx ox a x xx x 0 x x 0 • • 0 xx x • Xx • 3FI X x • 0~ XX x 0 X Y X xx xx X X x x K ,. x X le • x x 0 X x x x x x
X x x X x x X x x x x x x x x xY x x 0 x x 0 8 16 24 32 Mean length of search periods (sec)
Figure 4.22. The relationship between the number of search periods and
their mean length in Nemeritis canescens. (The point
37.2, 28) has been omitted).
x = observations made 3 hrs 15 min. 6 hrs 15 min. } after experiment start 9 hrs 15 min.
no significant relationship
o = observations made 15 min. after experiment start.
t-test: b = t = -7.17, P <0.001 with 45 degrees of seb freedom 132
a change in behaviour could occur. A table of random numbers was then
used to give a set of random times. These were then superimposed on the
behaviour records obtained when only one wasp was present, and changes in
behaviour, at these random moments, analogous to the nine encounter
categories above, were recorded. These were used as control values.
Table 4.1 shows clearly that encounters cause change in behaviour,
over and above that expected on the basis of the normal behaviour of a
single parasitoid. [The test assumes that encounters occur randomly in
time, i.e. follow a Poisson distribution. The distribution of encounters
in 12 second blocks was tested against a Poisson distribution of the same
mean and total number of observations. There was no significant differ-
ence, i.e. no evidence against the random distribution of encounters in
time].
The percentage change following encounters was then examined for
trends with density, using the method described in Snedecor and Cochran,
1967. This is illustrated in figure 4.23. (Confidence limits were
calculated from Table W in Rohlf and Sokal, 1969). Thus the tendency to
cause change in behaviour due to encounters, declines with parasitoid
density.
Next, each behaviour pattern was examined separately to see if the
effect of an encounter was affected by. the occupation of the individual
at the time. Table 4.2 shows that while encounters during probing or resting tend to increase change, encounters during walk have the effect of decreasing the frequency of changes in behaviour.
Within each behaviour pattern there were no significant changes with parasitoid density, but the percentage change following encounters in rest 133
Table 4.1. Contingency table showing that encounters between Nemeritis
wasps cause change more often than occurs at random.
Change No change Data source Totals 0 E 0 E
Encounters 390 254.32 729 864.68 1119
X2 with 1 degree of freedom = 92«99 P <0.001
0 = observed frequencies.
E = expected frequencies, based on the null hypothesis that there is no
difference between the two groups of data. 134
Figure 4.23. The decline in the percentage of encounters between wasps
resulting in a change of behaviour, with parasitoid density
in Nemeritis canescens.
b = -1.9652, P <0.05, with 1117 degrees of freedom. seb
Figure 4.24. The decline in the percentage of encounters between wasps
resulting in a change of behaviour, with average encounters
per parasitoid in rest in Nemeritis canescens.
b = t = -2.43, P <0.02 with 96 degrres of freedom. seb 1.35 r iou hav be in e hang c by d e low l fo ters n cou % en
0 8 16 24 32 Parasitoid density
r 100
iou y = 83.87 - 6.78x hav be
in e ng ha c
by d 50 llowe fo ters n cou en Z
t 0 2 4 6 Average encounters per parasitoid in rest 136
Table 4.2. Contingency tables showing the effects of encounters between
searching Nemeritis on frequency of change within each
activity.
A. Changes in Probe
Change No change Data source Totals X2 o f E 0 E Encounters 203 > ,.9 3 391 507.07 594 179..99, p~ 0.001
B. Changes in Walk
Change No change Data source Totals X2 0 0 E Encounters 132 < 202.45 297 226.55 429 45.77 P < 0..001
C. Changes in Rest
Change No change Data source Totals X2 01 E 0 E
Encounters 55 > 23.441 41 72.56L 96 54.45 P <0.001
0 = observed frequencies. -
E = expected frequencies, based on null hypothesis that there is no
difference between the two groups of data. 137
declines with the encounters per wasp within rest, as shown by figure
4.24.
The data was then examined for shifts in the direction of behaviour
changes due to encounters. In this case, all the natural changes in
behaviour from five of the records of solitary wasps, were used as control
data, because the random points produced an unacceptably low level of changes to work with. Table 4.3 shows that encounters in probe causing
change tip the balance further towards walk than occurs at random.
Similarly encounters in walk favour probe, rather than rest. There is no
significant change in these effects with density or frequency of
encounters.
Figure 4.25 summarizes the effects of encounters, while figure 4.26
is an attempt to show how these effects are interrelated. The plus or minus figures are attached by considering the effect of one encounter compared with that of a non-encounter, using figure 4.25. For example, in any ten random "moments" in probe, (3 second intervals) 1.23 will show a behaviour change to walk (10 x 0.1463 x 0.8383). After any ten encounters in probe, 3.32 will show a shift to walk (10 x 0.3418 x 0.9704). Thus an encounter in probe should increase the number of walking period, and hence the percentage of time spent walking. A positive relationship is there- fore expected between the number of encounters during probing and the number of walking bouts.
The data was then examined for the relationships suggested by the diagram. The major effects discovered are shown in figures 4.27 to 4.30.
Surprisingly, the number of probing sessions was found to be positively related to the encounters in walk. However, the primary effect of encounters in walk is to resist change, thus increasing the proportion of 138
Table 4.3. Contingency tables showing the effect of encounters between
searching Neraeritis on the direction of changes in behaviour.
A. Changes in Probe
PW PR Data source Totals X2 0 E 0 E
Encounters 197 182.61 6 20.39 203 19.60 Random 197 211.39 38 23.61 235 P <0.001
Totals 394 44 438
B. Changes in Walk
WP WR Data source - Totals X2 0 E 0 E
Encounters 95 83.63 37 48.37 132 5.44 Random 192 203.37 129 117.63 321 P <0.05
Totals 287 166 453
C. Changes in Rest
RP RW Data source Totals X2 0 E 0 E
Encounters 18 15.93 37 39.07 55 0.29 Random 46 48.07 120 117.93 166 NS
Totals 64 157 221
0 = observed frequencies.
E = expected frequencies.
PW, WR = changes from probe to walk, walk to rest etc.
NS = not significant. 139
Figure 4.25. Summary of the effects of encounters between Nemeritis
adults on changes in behaviour. Vertical arrows indicate
the effects of encounters. The dotted line indicates that
the change is not significant.
R = after random points
E = after encounters
For example, random points in rest are followed by more
resting (no change) in 75.58% of cases i.e. 7.56 times out
of 10 occasions.
Encounters in rest are followed by more resting (no change)
in 42.71% of cases i.e. 4.27 times out of 10 occasions. 140
97.04% walk 83.83% 34.18% Change 14.63% Probe 2.96% 65.82% rest 16.17% Not change 85.37%
71.97% probe 59.81% 30.77% Change 47.19% Walk 28.03% 69.23% rest 40.19% Not change 52.81%,.
32.73% probe 28.96% 57.29% Change 24.42% NS 67.27% Effect decreases walk 71.04% Rest with f(encs) 42.71% Not change 75.58%
34.85% Change 22.732 Effect decreases At any time with a 65.15% Not change 77.27%
On 10 occasions: R E R E walk 1.23 3.32 probe 2.82 2.21 Probe rest 0.24 0.10 Walk >rest 1.90 0.86 probe 8.54 6.58 walk 5.28 6.92 R E
probe 0.71 1.88 Rest C )walk 1.73 3.85 rest 7.56 4.27
Figure 4.26. Summary of the overall effects of encounters between adult Nemeritis.
A + B indicates that A affects B, positively, [:= +ve ::, or negatively, C:= -ve ::, as shown. Figure numbers
show which figures display the relationships. Dotted lines and brackets, e.g. (+ve), indicate expected
relationships which have been found to be insignificant.
%Tp, %Tw, %TR = percentage of time spent probing, walking, and resting respectively.
(1) Encounters in probe increase the tendency to walk when a change in behaviour occurs.
(2) 11 " walk 11 " probe " 11 il 11 11 11
(3) il " rest 11 II It " probe "
* This relationship was found to be positive, rather than negative, as expected. See text. Number of Fig. 4.27 A / rests Fig. 4.19 -ve* (-ye) +ve x
i
Encounters \ +ve 1 (-ve) in (:(7-rest . r N Change 7I Changed (-Encounters in \ ~ //(3) 7 walk / ve Fig. 4.29____ Mean walk +ve Fig. Fig. 4.28 4.16 • +ve Fig. 4.17 Number +ve / (+ve of walks 143
Figure 4.27. The increase in the number of search periods with the
number of encounters between wasps in walk (Aemeritis
canescens).
t—test: b = t = 3.94, P <0.01 with 42 degrees of seb freedom
Figure 4.28. The increase in the number of walk periods with the number
of encounters between wasps in probe (Nemeritis canescens).
t—test: b = t = 3.21, P <0.01 with 40 degrees of freedom. seb 144
7 = 57.17 ; 0.58x
0 30 60 Encounters in walk ds io er lk p f wa o ber Num
0 30 60 Encounters in probe Figure 4.29.Theincreaseinthemeanlengthofawalkperiodwith
Mean length walk period (sec) walking, in30minutesobservationof t-test: increase inthenumberofencountersbetweenwaspswhen 0 se b =t3.50,P<0.001,with40degreesof y = b
6.12 +0.08x Encounters inwalk 145 30
freedom x Nemeritis canescens. 60 Figure 4.30.Thedegreaseinthemean length ofprobeperiodsasthe
Mean length of probe period (sec) t-test: b=t-2.60, P Nes r►itiscanescens. number ofencountersbetweenwaspswhen probing,rises,in se b
Encounters inprobe 146 30 <0.02
with 42degreesof freedom 60 147
the time spent walking. Now, even though an encounter in walk reduces the
tendency to shift to probe, an insect which is walking is still more
likely to begin probing than one which is resting, encounter or no
encounter.
The primary effects of encounters are therefore as follows:
1)Encounters in proble decrease the mean probe length, and increase the
number of walking periods (Figures 4.28, 4.30).
2)Encounters in walk increase the mean walk length (figure 4.29).
There is therefore an increase in the percentage of the time spent walking
(see figure 4.12). [Encounters in rest, the minority activity, are rare
and have no significant effect under these conditions :I. Changes in
behaviour occurring during walking tend to shift towards probe, whether
stimulated by an encounter or not. Figure 4.27, the increase in the
number of search periods with encounters in walk, is therefore the result
of two components.
1)Encounters in walk decrease change, increase mean walk length and
hence walking time. Changes in behaviour in walk tend to be changes into
probe.
2)Encounters in walk which are followed by change have a greater
tendency to shift to probe.
This closes a negative feedback loop and boosts the number of probing,
back up again. This may be the explanation of the inverse relationship found between the mean length and number of probing bouts, shown in figure 4.22. 148
However, since the inverse correlation of mean length and number of
probing sessions is not found later in the experiment, it is possible that
the negative feedback effect decays as activity does. Nonetheless some-
thing keeps the pattern of time spent probing intact, (see figure 4.1).
This suggests that the lack of a mean length/number correlation is due to
a lack of long search periods, i.e. it is there but the variance masks it
because there is not a wide enough range of mean length values to display
it. The question. arises, why, then, is the mean length of probing period
so short, and how would this be reflected in the relationships between
the behaviour loops shown in figure 4.26?
A possible explanation is that search periods are short because search "drive" is low. Activity has been shown to fall with time of day
(figures 4.7-11). The time spent probing is a lot lower. Casual observations have suggested that the time spent walking is also low.
Therefore, there must be an overall increase in the time spent resting and the importance of the rest loop. The proportion of encounters falling in rest would therefore increase. Figure 4.25 indicates that these would produce negative feedback to both walk and probe, stabilizing them at about the same level.
A further question remains. Why did Hassell (1971b) find a declining relationship between the time spent searching and parasitoid density, when the slightly different conditions used here produced no such evidence?
The explanation which springs most readily to mind is that his feedback loops (through walk or rest) were weak or non-existent. This seems highly likely when his experimental conditions are reconsidered. Hosts were highly clumped, individual host areas fairly small and aggregation of the searching wasps produced local variation in parasitoid density. In the 149
host areas, wasps would be expected to aggregate, and the encounter rate
between wasps would be very high. During probe therefore, encounter rates
would be very high and effects such as that shown in figure 4.30 (decline
in mean length searching time) would be very strong, leading to decrease
in searching time. Under these conditions, the wasps also tend to leave
the small host areas altogether (see Hassell, 1976) when encounters occur.
Therefore, much walking and resting from behaviour changes after encounters
in probe, must have occurred outside the host areas, where parasitoid
density was punctionally lower. The encounter rate between wasps in rest
or walk would have been much lower, leading to weak and ineffectual feed-
back loops.
However, while this provides a fairly clear picture of the effect of
encounters between searching parasitoids on the balance of wasp behaviour
patterns, it offers no explanation of the mutual interference relation-
ship found in chapter three. This suggests that the percentage of time
spent probing is not directly related to the searching efficiency, or the
number of eggs laid, in this case. As mentioned on page 29, cocking movements have been shown to indicate the replacement of an egg in the egg-cavity at the tip of the ovipositor (which holds only one egg) (Rogers,
1970, 1972b). This has been commonly used as a measure of the number of eggs laid under observation.
Figure 4.31 shows the relationship between the number of cocking movements and the percentage time spent probing. As might be expected, the overall relationship is a strong positive one, but, on their own, the data from the first observation period, follow no significant trend with time spent probing. This may well be the effect of very high variation in success rate, coupled with high search drive and the feedback effect 150
Figure 4.31. The increase in cocking movements as the percentage time
spent probing increases, in Nemeritis caneseens.
Regression line is shown for all points. -
t-test: b = t = 13.72, P <0.001 with 154 degrees of seb freedom
Data from first observation period (o), 15 min. after
experiment start show no significant relationship, while
those from the last three observation periods (x), do show
an increase in cocking as the time spent probing increases
(y = 0.02 + 0.09x, b = t = 8.79, P <0.001 with 106 seb degrees of freedom).
N.B. Some of these points are duplicated; see table B4.31
in Appendix B. 151 ion t a erv bs o
% Time spent probing 152 already mentioned, which keeps the time spent probing in the first hour or so of the day within fairly narrow limits.
However, referring back to figure 4.2, and comparing that with figure
4.32, derived from the data analysed in chapter three, suggests that somewhere between discharge from the ovipositor and safe arrival in a healthy host eggs are lost.
During all the experiments reported here, the host larvae were often extremely lively, and reacted strongly to attack. Many attacks have been seen to fail. Thus one of the basic assumptions of many models for parasitoid behaviour, that an egg is laid on encounter with a healthy host, is violated. However, if the failure rate of an attack remains constant, this does not matter. On the other hand, if the failure rate increases due to an increasing defence reaction on the part of the host larvae, the searching efficiency, a, will fall, as in an interference relationship. Eggs may be lost at the surface of the host, or dropped in the medium as the larva withdraws and the ovipositor is jerked away.
Cocking is then necessary to replace the lost egg. As the number of attacks on larvae increases with parasitoid density, their avoidence reactions might well become more violent, and increase the activity of the rest of the host population by encounters between host larvae. However, there is no evidence to support or refute this idea. Certainly the number of eggs recovered per female declines sharply with density, and the number of cocking movements does not. As already mentioned (in chapter three), at no time was any encapsulation ever observed in a parasitized host upon dissection, an observation supported by Salt (1975), so immunity is probably not responsible for the decrease in searching efficiencies. The distribution of empty pupal cases remaining after emergence on the floor Figure 4.32.Thedeclineinoffspringrecoveredperfemalewith Offspring recovered per female t-test: b=t-27.74,P<0.001,with 4degreesof parasitoid densityin se Log y=1.50-0.78logic b
Parasitoid density 153 Nemeritis caneseens. 16 freedom 24
32 154
of an experimental cage when checked using a grid of 2 x 2 cm squares,
showed a variance/mean ratio of slightly less than unity, so there is no
evidence for aggregation of host larvae, which could produce pseudo—inter-
ference (Free et al., 1977).
There is no evidence for change in the percentage avoidence of super-
parasitism. This was calculated for each replicate, using the method given
by Rogers (1972b). The overall avoidence was 73.91 ± 8.80 (957 confidence
limits), which agrees well with the avoidence shown by Nemeritis parasitizing Ephestia eautella.
The mutual interference models used in chapter three are not stLictly relevant to these data, because of the behavioural complexity shown here.
However, substituting b (the rate of encounter) into bTw calculated for use in chapter three, produces values of Tw which are not related to density or to the number of encounters between wasps.
Behavioural observations of Bracon, suggest that encounters between adult wasps may have the same "startle"/change behaviour effect, as in Nemeritis.
4.4. Summary and Conclusions
1. Adult Nemeritis individuals were observed under conditions of increasing parasitoid density. Their behaviour was classified into three main types: probing, walking and resting, and encounters between adults as they occurred.
2. The percentage of time spent probing does not decline with density as has been found by previous authors. This is due to a complex negative feedback system, operating principally via walking early in the morning. Later in the day, decreasing activity may increase the importance 155
of the rest loop, also producing negative feedback and hence stabilization
of the time spent probing.
3. Cocking does not decline with density at any time, while the
overall number of offspring recovered per adult does. No encapsulation
was observed.
4. This suggests that the interference relationship shown in chapter
three is due to eggs lost between expulsion from the ovipositor and safe arrival in a healthy host. Larvae react violently to attack and an
increase in attack rate (due to parasitoid density) may cause an increase
in defensive behaviour in the hosts. This would result in an increase in
the failure rate of parasitoid attack, and eggs being lost as the host larvae decamp. 156
CHAPTER FIVE
INTERSPECIFIC COMPETITION: AN INTRODUCTION
Competition between species has been a subject of general interest to
ecologists since the mid 1920s. The same definition of competition which
is used in chapter one is appropriate here. As before, there are two
slightly different cases.
1)Species can be said to compete if the supply of a common resource is
insufficient for the needs of both (exploitation).
2)Species can be said to compete if, during the acquisition of a super-
abundant common resource, they nevertheless impair one another's survival
and/or reproduction (interference). This may occur by mechanisms such as
territoriality, aggression, secretion of toxins etc. These effects may
well have evolved under exploitation conditions.
Simple experiments to determine the outcome of interspecific
competition were performed in the early 1930s by the Russian scientist
Cause (1932, 1934). His 1934 paper was concerned with competition between two species of Paramecium (a Protozoan), preying on a common food source
(see table 5.1). The Paramecium caudatum population always died out.
Since these early experiments were performed, there have been many similar trials, using a wide variety of species, with much the same results. A selection of these are shown in table 5.1. These suggest that competition results in the extinction of one species or the other. Evidence which has accumulated from the field indicates that this conclusion may also be valid for some populations under natural conditions. This information falls roughly into two classes. Table 5.1. Some examples of extinction, negative correlation and displacement due to interspecific competition.
Competing species Resource Outcome Author Comments
A. Laboratory Experiments
1. Paramecium oaudatum, Yeast Extinction of P. Cause, 1934 P. aurelia caudatum. 2.P, oaudatu+B, P. aursiia, P. Yeast Extinction of one Vandermeer, 1969 Coexistence found in some twits. burearia competitor in some cases. Blepharisma sp.
3. Tribolium confuswn Flour Extinction of T. . Park, 1948 In the presence of Adelina T. castaneum confunwn, 12/18 (sporozoan parasite); extinction replicates. of T. oaetaneum: 66/74 replicates.
4. T. canfueum Flour Extinction Nathanson, 1975 As interval between flour renewal T. oaetaneum increases, advantage shifts from T. oaetaneum to T. conf2usum.
5. Rhiaopertha domtinioa Grain (maise or wheat) Extinction of one Birch, 1953 Calandra orysae competitor depending on (small and large strains) conditions (food, temperature).
6. Bean weevils: Beine Extinction of C. Utidat, 1961 Coexistence for at least 6 Callosobruchus chinensie, maculatue. generations in the presence of a common parasite (see also tables 5.2, 5.3). C. maoulatus Extinction of C. Fujii, 1967 Temperature and relative humidity chineneis. may reverse results of competition.
7. Neneritis canescena Flour moth, Ephestia Extinction of Horogems. Fisher, 1962 See also table 5.4. Rorogenes DhSaoattietQ8 k hniella.
8. Hydra littoralis Brine shrimps, Artemis Extinction of Hydra. Slobodkin,.1964 See also table 5.2. Chlorohydra viridieoima spp.
9. Drosophila nasuta Yeast Extinction of one Ranganath 6 D. neonosuta competitor, depending. on Krishnamurthy, 1975 starting ratio.
B. Field Observations
1. Insect parasitoids Californian red scale Geographical displacement beBach & Sundby, 1963 Experimental studies show ---- — extinction. Aphytie spp. Aonidiella aurantii 2. Parasitoids Opius spacies Mediterranean fruitfly Extinction of O. humilis. Willard 4 Maaom, Cerititia oapitata. 1937 cited in DeBach, 1964
3. Carabid cave beetles Small arthropods Character displacement in Van tant at al., Pwudophtbalmus app. (Collembola, mites overlapping isotopes. 1978
4. Whirligig beetles Small invertebrates? Geographical displacement Lstock, 1967 D. nigrior starts breeding Dineutea nigrior of D. nigrior. earlier, but D. horni is a D. horn superior larval competitor.
5. Mites; ParlOrychus uVni, Apple trees Displacement of P. ulmi Croft & Hoying, 1977 Aoulus sohleotendali from orchards.
6.Flatworms; Polyoelie nigra Small invertebrates Negative correlation of Reynoldson & Bellamy, Total flatworm population in lake P. tenuie e.g. Tubifem. population densities. 1970 remained more or less constant.
7. Desert rodents and ants Seeds Reciprocal increase when Brown & Davidson, Extensive overlap in diet. one taxon experimentally 1977 Complementary patterns of excluded. diversity and biomass in geo- graphic gradients of productivity.
8. Microtue ta.neendii, Grassland seeds? Negative correlation of Redfield at at., Peromysous manioulatus population density. 1977
9. Mice; Paromysoua manioulatus, Habitat displacement. Crowell & Pism, 1976 Peromyscus and Ciethrionornys Ciethrionomye gapperi, introduced onto island, support- Miorotus pannsylvanious ing Miorotus populations. C. displaced both Peromyscus and Microtus from woods, P. displaced M. from shrubby areas. Suggested success of Miorotus due to ability as opportunist.
10.Takand, red deer Grazing vegetation Displacement of Takah6? Mills & Mark, 1977
11. Blackbirds; Agelaiue Space Local displacement of Orians & Collier, phoenioeua, A. tricolor redwings A. phoeniceus 1963
12. Birds in Andes Various Range expansion in Terborgh & Weske, absence of competitors. 1975,
13.Lizards; Anolis spp. Habitat space, prey, Expansion of resource Lister, 1976 (West Indies) perching sites range used in absence competitors.
14. Salamanders Soil animals Local displacement of Jaeger, 1972 Plethodon richnondi P. richnondi to drier P. cinereas areas. See also Grant (1972) for examples from the rodents, Pianka (1976), and Elton (1958). 159
1)Negative correlation between population densities of competing species,
and the extinction of one species following the introduction another.
One species may become extinct or suffer a drastic fall in population
density as another is introduced or increases.
2)Habitat or resource displacement. Many species of similar ecology come
into contact in only one part of their species range. When this occurs
the habitat or resource normally used by one or both species in the
absence of the other may contract or change in the presence of the
competitor. Shifts in resource or habitat utilization are often
accompanied by morphological changes, which are found only where the range
overlaps with that of the competitor species. This is known as character
displacement (Brown and Wilson, 1956). The common examples of this
include changes in body size related to shifts in the size of food items
taken, (Wilson, 1975) and changes in both size and shape in birds also
with shifts in the size and type of resources used (Lack, 1971). Bulmer
(1974) has shown that character displacement may be predicted on the
basis that competition is more severe between very similar species (or
individuals). This is confirmed by Reynoldson's (1975) work on flatworms.
He found that the more closely related the species of triclad, the greater
the overlap in diet, and hence the more severe the competition when food
is short.
Examples from these two groups are also shown in table 5.1. However,
the field information is, at best, merely circumstantial. The possibility that an unknown and so far unobserved variable causes the population changes, rather than the presence of a competing species, cannot be ruled out.
While there are many examples of competition between insect para- 160
sitoids, only a few show clear evidence of extinction or displacement.
Fisher's (1962) work on Nemeritis and Horogenes (showing the extinction of
Horogenes) in a laboratory ecosystem, and field observations by DeBach and
Sundby (1963) on the displacement of Aphytis chrysomphali by A.
lingnanensis (parasitoids of the Californian red scale Aonidiella aurantii)
are well known. DeBach (1964) cites another example from a paper by
Willard and Mason (1937), which shows the elimination of Opius humilis by
other species of Opius, parasitoids on the Mediterranean fruit fly
Ceratitis capitata on coffee beans in Hawaii.
A simple mathematical model predicting the outcome of competition was
derived by Lotka (1925) and Volterra (1926). This was based on the
logistic equation (see chapter two), but included a term for the detri-
mental effect of the presence of the other species. One form of the
equations is as follows:
K1 - N1 - aN 611 - r1 N1 ( 2) for species 1 dt K1
- dN2 r N (K 41) - 2 - N2 for species 2 dt 2 2 K2 where N1 is the population size of species 1.
t is time.
r1 is the intrinsic capacity for increase of species 1.
K1 is the carrying capacity of the environment for species 1.
a is the detrimental effect one member of species 2 has on one
member of species 1.
0 is the detrimental effect one member of species 1 has on one
member of species 2.
From these equations, lines of zero population growth (isoclines) can be 161
calculated for each species. These two lines may have any one of four
possible arrangements, depending on the relative values of a, S, K1 and
K2, when placed on a graph of N2 versus Ni. These four cases are shown
in figure 5.1. The arrows show the direction in which species mixtures
would be expected to change. In two of the four cases, the model predicts
that one or other species will win (a, b). In one case (c), there is an
unstable equilibrium. If the species mixture is started off at this
point, both species will persist. If the system is perturbed, however,
one species becomes extinct depending upon the direction of the
perturbation. If the mixture is started off from any other position, one
or other species will win, depending on the starting point. In case (d),
there is a stable equilibrium point. Any species mixture will progress
towards that point, and perturbations are also corrected back towards it.
The model therefore predicts that, under some circumstances, coexistence
between competing species is possible. Indeed Gause's (1932) first
experiments, with yeast populations, did not lead to extinction for either
species, although the population densities of both were reduced. The
results of both sets of experiments (with yeast, and with Paramecium), and
the Lotka-Volterra model, led him to suggest that competing species cannot
occupy the same "niche" (Gause, 1934). The term "niche" he used following
Elton (1927), who used it to indicate the role a species played in its
community, including its relations with all the other species-present.
Coexistence, as predicted from case (d), and as shown by Gause's (1932)
yeast experiments, would therefore result only when the two species in question occupy different niches. This idea has come to be known as
"Gause's hypothesis" or "the competitive exclusion principle". Perhaps
the best definition is that of Hardin (1960) - "complete competitors cannot coexist".
162
a) Extinction of species B b) Extinction of species A K2 E~
B K1/a ies ec f sp o K2 ity dens ion t la u Pop
c) Unstable equilibrium. d) Stable equilibrium Displacement causes K
2 extinction of A or B B
depending on direction. ies ec
f sp K o ity
s K2 den
ion t la
u K„/8 Pop
Population density of species A Population density of species A
Figure 5.1. The four possible outcomes of competition based on the Lotka-Volterra models.
K1 carrying capacity of species A
n n ►i K2 B
a effect of 1 B upon 1 A individual $ effect of 1 A upon 1 B individual
Arrows show direction of population change (After Krebs, 1972). 163
Since the Lotka-Volterra model was proposed, it has been shown to be
inadequate to describe a number of experimental systems. For example,
Neill (1974) has shown that competitive interactions in communities of
microcrustacea may be non-linear, interdependent and also change with the
species composition of the community. Complex density effects were also
found by Smith-Gill and Gill (1978) in competition within and between
two species of tadpole. Some studies have shown that species using the
same resource may actually stimulate each other. For example, Moiseyeva
(1974) showed that the immune reaction of Pieris brassicae caterpillars to
the parasitoid Hyposoter ebeniuus disappears in the presence of eggs of
another parasitoid ApanteZes glomeratus. The data given by Ayala et al
(1973) for some Drosophila systems also does not conform to the Lotka-
Volterra model. Not only are the lines of zero population growth (the
isoclines) curved, instead of straight, as assumed by the model, but also
the conditions necessary for coexistence predicted by the model (the
product of the competition coefficients, p. 160, a$ < 1) are not fulfilled, despite the fact that coexistence occurs. This led the authors to, present a variety of new models, some of which included the Lotka-Volterra equations as a special case. These models therefore have greater generality i.e. apply to a wider variety of systems. They are, however, purely descriptive models; they do not attempt to suggest the mechanisms which operate in a competitive system. Hassell and Comins (1976) have also produced a model for two species competition which generates curvi- linear zero growth curves. The equations they used were a difference form of the Lotka-Volterra equations, i.e. they included a generation time lag.
They also added a simple age structure to the model. It was this feature that was responsible for producing curves in the isoclines. This is an extremely useful modification of the basic model, since age structure is 164
important in many natural populations.
Levins and Culver (1971) developed a model for competition over an
area containing several similar habitable patches, thus bringing in spatial structure. They found that, although competition between species caused
local extinction, the separation of the usable habitat into patches allowed coexistence.
However, rather than modifying and expanding the Lotka-Volterra equations, research in competition has recently tended to concentrate on the niche concept. It is more or less implicitly agreed that if two species have exactly the same niche i.e. 100% overlap, with the same resources, same predators and the same response to temperature etc. the exclusion principle would operate and one species would die out. There have been dissenters from this: Ross (1957, 1958) contended that his study of six species of leafhopper showed that more than one species could coexist in the same niche. Savage (1958) argued that Ross was using
"niche", incorrectly, to mean "habitat" instead of "sum of interractions with the rest of the ecosystem". Finally, McClure and Price (1976) clinched the matter by demonstrating segregation between the species in geographical distribution, thus reducing overlap.
Many other biologists have investigated cases of coexistence in the field where severe competition was suspected. In most of these cases, ecological differences between the species were brought to light. A variety of examples is given in table 5.2. One of the more interesting cases of coexistence occurs when both competitor species have frequency- dependent fitnesses i.e. have a competitive advantage when the density of conspecifics is low. Ayala (1971) has demonstrated such a system in
Drosophila species, and DeBenedictis (1977) has discussed its meaning and Table 5.2. A selection of examples of coexistence through limited niche overlap.
Competitors Resource Niche separation Author Comment
1. Grain beetles • Wheat grains Rhiaopertha larvae live inside wheat Crombie (1945) Laboratory experiments. Rhiaopertha sp. grain; Oryzaephilua larvae feed Oryzaephilue sp. externally.
2.Drosophila funebris, Yeast D. melanogaster better at using fresh Merrel (1951) Laboratory experiments: coexistence in D. melanogaster yeast; D. funebria use ageing yeast + population bottles for nearly 2 years. micro-organisms more efficiently.
3. Green hydra Brine shrimps Darkness, or intense predation allows Slobodkin Laboratory experiments. In light, Chlorohydra viridissima; Artemia app. coexistence. (1964) absence of predation Chlorohydra (which brown hydra has green algae in endoderm) causes Hydra littaralis extinction of Hydra.
4. Paramecium aurelia, Yeast P. aurelia feeds on suspension in upper Cause (1935) Laboratory experiments. P. bursaria layers; P. bursaria feeds on bottom layers.
5. 2 species of freshwater Phosphates, 1 species more efficient at obtaining Tilman (1977) Long term laboratory experiments, using algae silicates phosphate, the other at obtaining resource gradients. Results agree with silicates. models based on resource utilization response of each species. Stable coexistence when each species limited by different resource.
6. Microcrustacea Microhabitat and food specializations. Neill (1975) Laboratory ecosystems.
7.Patella species Algae Specialization most important in Branch (1976) Predation also contributes to maintaining allowing diversity. diversity.
8.Daphnia carinata, Algae Seasonal changes in relative fitness. Hebert (1977) D. oephalata
9. Daphnia hyalinz, Algae Seasonal changes in fitness, apart from Jacobs (1977) D. cucuitata seasonal temperature effects, which is also important.
10. Crayfish Differences in microhabitat used. 0. Bovbjerg (1970) Field + laboratory experiments indicate Orooneetes virilis, virilis - streams, lake margins; 0. competitive exclusion of 0. imnunia by 0. immune immuns - ponds. 0. virilia, which is more aggressive, evicts 0. imnmtnis from crevices in āubā,rātē.
11. Whip spiders Crickets, moths Competition probably reduced by use of Weygoldt (1977) Heterophrynua spp. different hiding places.
12. Spider spp. Insects Reduced overlap in habitat used, and Māhlenberg et (Seychelles) ranges of structural diversity. at (1977) 13. Wandering spiders Insects Prey size differences (and hence body Uetz (1977) size) between smaller species. Temporal segregation of foraging and breeding in larger species.
14. Nocturnal wolf spiders Wide variety of Vertical stratification separates off Kuenzler (1958) Separation between remaining two species Lyaosa spp. insects. one species not known.
15. Drosophila ailvarentie, Yeast D. silvarentia oviposits on sticky tree Kaneshiro et al A single tree species MgopoWuzn sand- D. hsadi on Hawaii fluxes well above ground level. D. (1973) Ax:oense supports both Drosophila. heedi oviposits on caked soil below.
16. F'ythroneura app. Sycamore trees Segregation by latitudinal distribution, McClure & Price This is a more detailed study to refute (leafhoppers) local segregation. (1976) Rosa (1957) who argued that these 5 species coexist in the same niche,
17. Meaopsocus spp. Fungal spores Segregation by vertical distribution. Broadhead & (Psocoptera) on larch and • Wapshere Pleur00000ua (1966)
18.BaWtog app. Nectar Overlap avoided by differences in Heinrich (1976) tongue lengths.
19. Waterboaueten Seasonal switch of dominance. Istock (1977) (Corixidae, Hemiptera)
20. Desert seed eating ants Seeds Seed size preferences differ and are Davidson (1977) highly correlated with worker body size.
21. Veromeseor pergandei Seeds As above Davidson (1978) Worker size polymorphism is inversely (desert ant) + other ants related to intensity of interspecific competition. •
22.Dragonflies Small pond . Temporal segregation. Henke & Henke Suggest that "errors of exploitation" of invertebrates (1975) t$ dominant species allows coexistence.
23. Megarhyaaa spp. Sawfly larvae Adults select larvae at different depths Heatwole & (Ichneumonid parasitoids) Tremex in the log; different lengths of ovi- Davis (1965) positor.
24.Assorted parasitoids Swaine jack Separated by differences in efficiency Price (1970a, pine sawfly at different host densities (humidity, 1971) Neodiprion alternative hosts of some importance). swainei
25.Paraeitoida Bean weevils Differential efficiency at different Utida (1961) Fluctuations in host density allows Hateroapilus prosopidia, Callosobruohua host densities. coexistence. Naooatolacous sp. mamesophagous Table 5.2. Continued
Competitors Resource Niche separation Author Comment
26. Scelionid egg parasitoids Eggs of T. nagagawai has higher host finding Nakasuji et aZ Counterbalanced competition. Asoleus mitaulocrii Pentatomid bug ability; A. miteukurii is superior (1966) and Nesara viridula larval competitor. Hokyo et at Telsnowses nakagaaysi (1966)
27. Encyrtid parasitoids Scale insect Counter-balance: M. luteolue has faster Bartlett & Ball Metaphycus flavus Coccus reproductive rate; M. flavus is superior (1964) and hesperidum larval competitor, but is very rare in M. luteolue the field.
28. Parasitoids Peach moth eggs Ascogaster tends to attack eggs, and can Van Steenburgh Ascogaster may attack larvae when on Macrocentrue ancytivorus, and larvae produce 56% parasitism on its own & Boyce apples; Maorocentrus attacks young Aseogaeter carpooapsae Laepeyresia Matrocentrus - 37% parasitism on its (1937) larvae. Higher % parasitism caused by mo festa own, destroys Ascogaster larvae within Ascogaster suggests higher searching host. efficiency - example of counterbalanced competition?
29. Starfish Range of prey Some microhabitat separation at low Menge & Menge Leptasterias heractts, app. tide. Different combinations of prey (1974) Pieaster Dthraceue size and species used.
30. Freshwater fish, a) Insects a) Differential preference, differences Fryer (1959) especially rock- in vertical and horizontal distri- frequenting cichlidae bution. b) Algae b) Different feeding mechanisms in some cases, spatial isolation.
31. Tropical stream fish Differences in range of food taken. Zaret & Rand In the dry season, when food abundance (1971) is at a minimum, diet overlap is also at a minimum. When food is short, competition is more severe, diet overlap is restricted.
32. Sunfish (Centrachidae) Microhabitat differences presence of Werner & Hall Green sunfish (L. cyanelius) appears to Lepomis maeroohirus, competitor. (1977), restrict access of competitor to L. claneiZus vegetation and associated larger prey. Bluegill sunfish (L. macrochirue) is therefore driven to open water column where it feeds on smaller foods. (see table 5.4)
33. Salamanders Exclusion from preferred areas, but Jaeger (1972) Plethodon olchmondi inferior competitor has higher tolerance P. cinereue for dryness, survives suboptimal conditions.
34. Ptethodon hofflnani, Differential use of food and surface Fraser (1976a, P. punotatus habitat in adults, staggered feeding b) schedules. 35.Frog tadpoles Limited overlap of microhabitat. Heyer (1974) Space more important than food in niche partitioning.
36.Lizards Habitat differences? Huey & Pianka Narrow zone of overlap partially nabuya app. (Scincidae) (1977) correlated with changes in habitat type.
37. Seabirds of family Similar diets Differences in foraging zoned at sea. Cody (1973a) Alcides
38. Chipmunks Differences in elevation, forest type Brown (1971) Eutaniaa dorsaZia, used. E. wmbrinue
39. Chipmunks Differences in elevation, forest types Heller (1971) Distribution of two species restricted Eutemiaa alpinise., used. Sheppard (1971) by aggressive behaviour of two others. E. epeeioaua, Small size of E. minima enables it to E. amoamse, survive in alpine situation, as it can E. minimise subsist on reduced food supply.
40. 5 species of desert Seeds Differential harvest of different seed Brown & In productive habitats, overlap is rodents sizes (correlated with body size). Lieberman greater and hence species diversity is Forage in areas of different plant (1973) higher. cover. Differ in annual activity.
41.6 species of successional Soil Differences in use of soil resources. Parrish & Mean overlap in successional communities annuals Temporal displacement in absorption Bazzaz was greater than that in mature prairie. of water and nutrients important to one (1976) species.
42. Guild of fugitive prairie Badger Small differences along many different Platt & Weis Probability of colonization is plants disturbances niche dimensions. (1977) • important.
See also Lack (1971) for Grant (1972) further Schoener (1974) examples Zwolfer (1971) 169
measurement. As yet, however, there is no evidence to suggest the mechanism behind such changes in fitness.
Coexistence may also occur solely on account of the variability of the environment. Jacobs (1977) explained the coexistence of several similar zooplankton species on this basis. Although one species may have a competitive advantage under one set of conditions, these conditions change before there has been time for any of the other species can be wiped out, and the advantage passes to another species. Sale (1977) has shown that suitable living space is the resource most likely to be in short supply among coral reef fish. He argues that fish species diversity is maintained by the unpredictable environment, which prevents the develop- ment of an equilibrium community. The most stable, persistent examples of coexistence caused by environmental fluctuation are those due to seasonal switches in advantage, for example in waterboatmen (Istock, 1977) and
Aaphnia species (Jacobs, 1977). Huffaker and Kennett (1966) have shown that, in the spring, the parasitoid Aphytis maculicornis is dominant, while Coccophagoides utilis dominates summer activity. These species both attack the olive scale, Parlatoria of ae. Stewart and Levin (1973) modelled the effects of competition when resources are seasonally renewed and confirmed that these conditions were sufficient to allow the coexistence of two species using a single resource.
The conditions allowing coexistence are of special relevance to the problem of complexes of coexisting parasitoids. There are fairly detailed studies of species complexes consisting of one host species and a surprisingly high number of coexisting parasitoid species (e.g. Price,
1970a, 1971, 1973; Force, 1970, 1972, 1974). The niche differences allowing coexistence appear to be fairly small but numerous. Flanders 170
(1965) has tabulated a large number of parasitoids which attack the black scale Saissetia oleae in Africa. The ecological differences between these species have not been clearly worked out, although it is known that only a few parasitoids attack any one host stage.
Zwōlfer (1971) drew attention to a particular kind of coexistence mechanism in parasitoids, which he called "counter-balanced" competition.
Briefly, this entails the superiority of one parasitoid when in "direct contact", i.e. within the host (intrinsic competition), while the other species is more efficient in the exploitation of the host population
(extrinsic competition). Thus, two parasitoids competing for the same host are said to be in "counter-balanced" competition when the inferior larval competitor (within the host) as a superior searching efficiency i.e. a greater ability to find host individuals. Three interesting examples of this kind of competition are given in Zwōlfer's (1971) paper.
Along the same lines, Miller (1977) has indicated that "intrinsic competitive ability and an index of potential reproductive capacity are inversely correlated among the three species" of parasitoids of the army- worm Spodoptera praefica. Bartlett and Ball (1974) have shown that
Metaphycus flavus, anencyrtid parasitoid of the scale insect Coccus hesperidum is a superior competitor within the host, compared with M. luteolus, but it has a lower reproductive rate, and is rare in the field. Further examples of parasitoid interactions which appear to be cases of counter-balanced competition, may be found in table 5.2.
An interesting interaction which is difficult to classify is described by Arthur et al (1964). They found that, of the parasitoids of the pine shoot moth, Orgilus obscurantor has a higher searching ability than Temelucha interruptor. However, Temelucha is a sup erior larval 171
competitor, and apparently finds its host by smelling out the pheromone
trail left by the footsteps of OrgiZus females.
Evidence such as this has tended to shift emphasis away from "what is
the outcome of competition?" to "how much niche overlap is permissible?"
How closely may species be packed? How does resource utilization affect
species diversity?
Before questions such as these can be answered, a standard definition
of the niche is necessary. The dispute ove'.- this has more or less died
down now; most biologists are content to work with Hutchinson's (1958)
definition as a basis. He envisaged the niche as an n-dimensional hyper—
volume. Each of the n dimensions corresponds to an environmental variable
such as prey size taken, microhabitat used, humidity etc., along which
the species in question can be placed. Consideration of two dimensions
produces an "area" of conditions under which the species can survive.
When more variables are investigated a niche volume (three factors) or hypervolume (more than three factors) is produced. This is well explained by Pianka (1976) and Krebs (1972). In a natural community there are so many variables that it is practically impossible to define a niche completely. In most competition studies the emphasis is placed on one or two factors only, usually resource utilization. Even so, it is extremely difficult to determine how best to measure niche breadth and overlap so that comparisons can be made across different communities. Colwell and
Futuyma (1971) proposed that the species composition of communities utilizing different resources should be used as weighting factors when calculating niche breadth and overlap. It is still not completely clear what is the best way to cope with this problem, which continues to be discussed in depth (Hanski, 1978; Hurlbert, 1978). 172
Many ecologists have produced niche overlap measurements using resource utilization curves. A clear account of this method and relevant references can be found in May (1976a). Briefly, the amount of any part- icular resource used by a species is plotted against the resource spectrum (see figure 5.2), and the overlap in area under the curves is taken to be the niche overlap in that particular dimension. This is also one way of estimating the competition coefficients used in the Lotka-
Volterra equations.
There is general agreement that, as MacArthur (1965) suggests, there is a limiting similarity to species which coexist within the same habitat, i.e. there is a limit to the amount of niche overlap which can be tolerated.
Species which are more alike in their ecology, and generally in their appearance as well must occupy different habitats. MacArthur (1965) also thought that this limiting value should be lower when productivity is high, family size is low, and the seasons are relatively uniform. Some confirmation of the first part was provided by observations by Ulfstrand
(1977) on passerines living in coniferous woodland in South Sweden which show that niche overlap along three niche dimensions was much lower in the summer, when presumably, production is higher. The exploitation curves moved along the resource spectrum axis away from the zones of interspecific overlap, so that the niche space of the whole guild was enlarged. EThe term "guild" is used to mean a complex of species living together which make their living in very roughly the same way. They are often closely related].
The critical similarity should, in addition, be lower for hunters which stalk or chase their prey, than for species which track down stationary prey (MacArthur, 1965). 173
A U
B,
A, ies c e sp by
ion t iza il t u rce ou Res
Resource spectrum e.g. prey size,humidity etc.
Figure 5.2. Illustrating the concept of niche overlap, using 1 dimension
only. Shaded areas = conditions of niche overlap (After
May, 1976b). 174
The niche overlap restriction implies that there is also a limit to
the number of species which can be fitted in to a particular habitat
(MacArthur and Levins, 1967). Consideration of a second niche variable
shows (as explained in Krebs, 1972; and Pianka, 1976) that a small niche
difference in each of twa dimensions can reduce overall niche overlap
considerably. Many small differences between two species might therefore
provide ample ecological segregation for coexistence. The large number of
potential niche dimensions available in a natural community provides scope for fitting in many different species. As MacArthur and Levins (1967)
noted, the potential species diversity increases as more and more niche dimensions are considered. Put another way, coexisting species must diverge along more and more niche dimensions as the number of species
accumulates; otherwise critical niche overlap is exceeded and the weaker competitor is reduced to extinction. Schoener (1974) reviewed a large number of examples of ecological differences in an attempt to relate diverge along one niche, dimension to divergence along another. He thought that similarity along one dimension should imply dissimilarity along another, and so examined the data for complementarity between pairs of dimensions. He found examples of nearly all possible combinations.
Surveying the relative importance of particular dimensions he came across several interesting trends. He discovered that habitat segregation was more often important than food type segregation, which, in turn, was more often important than temporal segregation. He found that predators, and terrestrial poikilotherms were more often segregated by time of day than were other groups. Vertebrates were found to segregate less by season than other groups. Lastly, segregation by food type was found to be more important for groups feeding on resource units which are large comparable to the consumer's body size. As Schoener suggests, attention could be 175 concentrated on clarifying and explaining interesting trends such as these
instead of merely cataloging ecological differences.
Werner (1977) attempted to develop a method to relate foraging theory
to species packing. He produced cost curves which rank prey by the cost/
benefit ratio to the predator. These were used to estimate the extremes of ranges in food size taken, given the size and species of fish. He applied this method to three species of sunfish and found that two species were separated by using food of different sizes, and the third occupied a different habitat. Here, then is an example of complementarity of habitat and food sizes.
MacArthur and Levins (1967) considered the evolutionary limit to the similarity between two coexisting species. When two species are more similar than this critical limit, a third intermediate species will converge to the nearer of the pair. If the original two species are more different than the critical value, the third species will diverge from both towards a phenotype intermediate between the two. An evolutionary approach such as this, in which the properties of the competing species are not regarded as fixed, but subject to selection pressures, is perhaps the most useful and interesting (although difficult). Roughgarden (e.g.
1976) has been at the forefront of this line of attack.
Where competition causes a shift in the resources used, the niche breadth may be reduced or expanded, i.e. one or other species may become relatively more specialized (narrower niche) or more generalized (broader niche). If we consider the specialization of both species such that the original range of resources is unchanged, then, obviously, the supply of resources within each of the new ranges must be sufficient and predictable enough to support the minimum populations necessary for survival. Schoener 176
(1969) looked at energy gain and handling time of different types of
predator over a range of different food classes. He concluded that:
1)at high food levels and/or low metabolic rates, a specialist would be
at an advantage.
2)at low food levels and/or high metabolic rates, a generalist would be
at an advantage.
3)for energy maximisers (i.e. those with lots of time) generalists would
be favoured.
4)for symmetrical fluctuations in food density about the equal advantage
point, generalist time minimisers and specialist energy maximisers
would be selected.
5)large predators are more like specialists and therefore are more
favoured at high food densities. Hence, if an invading competitor
reduces food abundance equally over the whole range of food classes
(i.e. a generalist), selection will favour a decrease in size all
round; while the invasion of a specialist will favour a divergence in
size.
Although these ideas are interesting and stimulating, the assumptions on which they are based are too simple for the conclusions to be applied directly to natural ecosystems. For example, Schoener made no allowance for time and energy spent searching for food, nor did he include the
effects of differential exploitation.
There have been a number of other attempts to investigate the conditions required for the evolution of specialization. MacArthur and
Pianka (1966) included the effect of patchiness and the components of 177 hunting time on specialization. They drew up a list of the features of the diet, and patch size and distribution which could act to increase specialization. So far, there are only limited data of the kind applicable to these ideas. Post and Richert (1977) have examined spider community structure in this way. They found that there were increases in the number of non-equilibrium species, and decreases in the number of specialists, as the homogeneity of the habitat increased. Fretwell (1978) examined guilds of bird species and compared them with the results of the models by
MacArthur and Levins (1964, 1967). According to theory, pressure to broaden a feeding niche is always followed by reduction in species density.
However, when the resource spectrum is discrete, the reduction occurs because generalists replace specialists (MacArthur and Levins, 1964).
Fretwell found this in sparrows and flycatchers. When the resource base is continuous, intermediate species are replaced by the expanding niches of the species on either side (MacArthur and Levins, 1967). Woodpeckers and finches show this.
This work clearly shows us the implications of a few limited starting assumptions for the evolution of community structure. If we observe the results predicted, this does not show us that the assumptions are correct, merely that they may be so. What is needed now is a methodical testing of the initial assumptions; how often are these conditions to be met with in the field? There are also other features of the environment which need to be considered. For example, how does the presence of a top predator influence the evolution of specialization?
Disregarding evolutionary change for the moment (by shortening the time scale i.e. taking an instantaneous view), we return to a consideration of the theoretical limit to niche overlap. Recent work along a single 178
niche dimension has concentrated on the ratio of the distance between adjacent niches, and their width (d/w; d, w as shown in figure 5.2). As explained by May (1976b), it is suggested that d/w must be greater than or approximately equal to one, if the species occupying the two adjacent niches are to coexist. However, it must be remembered that we so not know what shape a niche is (the normal curve assumed here may be incorrect), nor is it clear how to incorporate further dimensions.
If the ratio d/w limits niche overlap, how many species may be found in a community? Stewart and Levin (1973) showed that, when resources are continuously supplied, two species can only coexist if they have at least two resources. Levin et aZ (1977) extended this to include the action of predators. They concluded that the number of coexisting prey species cannot exceed the number of resources plus the number of predator species.
However, these two could be regarded as special cases of the conclusion reached by Levin (1970). He suggested that niche overlap was important in coexistence and displacement only when it concerned the limiting factor. Thus, he concluded that a stable equilibrium cannot be obtained in an ecological community with r components when these are limited by less than r limiting factors. A limiting factor could be a resource, a predator, environmental heterogeneity or seasonal climatic change.
Many studies have shown that the number of species in a community may be influenced by predation. A selection of these can be found in table 5.3. For example, Paine's (1966) work on intertidal communities shows clearly that the removal of a predator (a limiting factor) can decrease the species diversity of the level below.
Connell (1970) has discussed the conditions under which predation may increase species diversity. On the other hand, there is some evidence Table 5.3. Some examples of competition influenced by predation/parasitism pressure.
Predator/Parasitoid etc. Competitors Resource.., Authors Comment
1.Parasitic wasp Bean weevils Beans Utida (1953, 1961) Laboratory experiments. No equilibrium possible with Neocatolaocue mcsnesophague caiZDsobruchua only the two species of weevil. Parasitism allows (or Neteroepilue chinensie, C. quad - coexistence. . proeopidis) rimaculatue
2.Protozoan predator Bacteria Glucose Jost at al (1973) In absence of predation, no equilibrium possible. Tetrahpmena pyriformis Eecheriechia coli minimal medium Predation allows coexistence. Aaotobactor vinelandi
3.Starfish predator Intertidal Space, Paine (1966) Removal of Pieaater reduces species diversity from 15 Pfeaater sp. barnacles, bivalves, Grazing? species to 8. MNtilua (a bivalve) tends to crowd out limpets, chitons the others.
4. Starfish predator Limpets Grazing areas Branch (1976) Differential predation may allow coexistence. Marthasteriae glacialis Patella longicoeta, of the algae P. tabularis Ralfeia expanaa
5.Limpet grazing Species of algae Rock substrate Dayton (1971) Up to 21 app. of algae present in absence of limpets, reduced to a maximum of 6 when they were present.
6. Seed and seedling predators Tropical tree Space Jansen (1970) Seed and seedling predation may lead to a reduction in species adult density and/or increased distances between new adult trees and their parents, leaving more room for other species of tree, and therefore a higher tree species diversity.
7.Herbivores Grassland flora Space Harper (1969) • Overgrazing by generalists leads to higher species diversity, undergrazing leads to lower species diversity. Grazing by specialists, and generalists with palatability preferences (or frequency dependent generalists) will be expected to increase species diversity.
8.Predatory copepod Daphnia minnehaha Algae Dodson (19741 Predator is responsible for absence of the smaller D. Diaptomus ahoehone D. middendorfiana minnehaha in the field. In laboratory experiments, this species was excluded within one month. Predation reduces diversity.
9.Grazing herbivores Algae Rock substrate Lubchenko (1978) Consumer effect on species diversity depends on 1) food preferences; 2) competitive ability of the food species; 3) intensity of grazing or predation pressure. 180
(e.g. Dodson, 1974) that predation may reduce species diversity by eating
out favoured prey species. Predator preference would therefore be expected
to influence the impact of predation on competitive interactions, and
hence species diversity. Kerfoot (1977) has shown an inverse relationship
between fecundity and predator resistance in cladoceran communities. This
could lead to an increase or a decrease in species diversity depending on
whether
a) the prey species were sufficiently distinct ecologically to coexist in
the absence of predation.
b) predation actually removed the most fecund species or merely depressed it,
A graphical model along similar lines was developed by Allan (1974) to
investigate the optimum size strategy of the prey species. He found that
an increase in species diversity was to be expected only under some conditions of predation and competition.
Caswell (1978) modelled competition in open systems of habitable
patches connected by migration, in which transient non—equilibrium coexistence could persist for a surprisingly long time. He found that predation had a significantly positive effect on coexistence.
Roughgarden and Feldman (1975) analysed a model in which a predator population was added to MacArthur and Levins' (1967) model of three competing prey species. They found that predation pressure permits increased niche overlap. Indeed, if it is severe enough, complete niche overlap should be possible.
Coming and Hassell (1976) also considered the effect of a top 181
predator on the stability of competing prey species. Their model showed
that switching in predators could enhance prey coexistence, as could
interference between predators.
Thus, the effects of competition on the structure and function of
communities are far more complex than had been previously thought. ..Quite
apart from contributing to a general understanding of ecological systems,
competition studies may also be of direct practical use.
Experiments on competition between parasitoids are of particular
relevance to the introduction of predators and parasitoids into the field
for biological control. It has been suggested that a parasitoid with a
high searching efficiency may be adversely affected by a superior larval
competitor, such that the target pest population reaches higher pop-
ulation densities than it would if only the first parasitoid were present.
If this is true, then it would be inadvisable to introduce more than one
parasitoid to control a pest population, in case this kind of relation-
ship is created. A discussion of this hypothesis may be found in, for
example, Smith (1929), DeBach (1964) and ZwOlfer (1971). However, the
data used in such discussions is, in many cases, not sufficiently clear
cut to confirm or refute the idea, as it is often in the form of percent-
age parasitism, which does not provide information on searching efficiency
or larval competition.
One difficulty that can arise is if one parasitoid has the effect of
synchronizing the host species so that all the stages are no longer available at the same time. This could be the cause of the extinction of other species of predator or parasitoid with shorter life cycles than the
host. However, this effect is more relevant to parasitoid/host relations
than to competitive interactions. 182
Another difficulty that can arise is that some parasitoids are facultative hyperparasitoids i.e. may parasitize primary parasitoids rather than or as well as the pest (Askew, 1971; Price, 1973). This should be tested for in the laboratory prior to introduction, as a hyperparasitoid may depress parasitoid number, thus allowing pest population densities to rise.
Some interesting examples of the importance of interspecific competition in biological control are shown by studies on mites. Croft and Hoying (1977) showed that the Apple rust mite, Acutus schiectendali, can displace the European red mite, Panonychus Mini, in apple orchards.
This is useful as the former can be controlled by chemical means far more easily and cheaply. Thus, a superior pest competitor may nonetheless be more susceptible to control, and hence it is well worth instigating competitive displacement. In many cases, attempts have been made to use biological control agents that are resistant to pesticides, so that both biological and chemical lines of attack may be employed (e.g. Downing and
Moilliet, 1974). Such a resistant parasitoid or predator must also be able to survive in competition with endemic natural enemies. In either case (susceptible pest or resistant enemy) a thorough knowledge of competitive interactions is necessary before introductions are made.
From a more general point of view, deliberate or accidental intro- ductions may have a variety of repercussions on their new communities, including undesirable competitive effects. For example, it is thought that competition with red deer may be partially responsible for the decline of the takahe (a rare New Zealand flightless bird) in areas west of the Murchison mountains (Mills and Mark, 1977). Similar examples abound in Elton (1958). Whenever an introduction is contemplated, or, 183 having accidentally occurred, is observed, careful laboratory studies
(and isolated field trials, where possible) are necessary to pick up and prevent undesirable side effects. Without a wide variety of studies on all kinds of organisms, it would be extremely difficult to attempt to solve applied problems involving interspecific competition.
To date, there have been few attempts to control Myzus persicae and
Plodia interpunctella using biological or integrated control' methods. The reasons for this have been more fully discussed in chapter one. Briefly,
Mzyzus causes economic damage at very low population levels, and it is unlikely that any natural enemy could reduce it sufficiently to prevent this. Plodia interpunctella and a number of related stored product
Lepidoptera, have been the target of a slightly larger number of attempts at biological control. Recently, Press et al (1977) have shown that
Nemeritis and Bracon, used together, reduce small populations of Ephestia cautella marginally more efficiently than either used separately. How- ever, this would be difficult to implement on large scale because it would require a substantial effort to breed sufficient parasitoids to start the process off, which presumably would have to be repeated for every infestation. It is hard to see residual pest populations being tolerated in warehouses to maintain parasitoids against further pest outbreaks. There has been a tendency to concentrate on preventing access of pests such as PZodia to warehouses, rather than keep them down once they have arrived there.
Turning now to the mechanisms of competition between species, we find the same types of mechanism operate in interspecific competition as in intraspecific competition (see chapter one). Fairly extensive inform- ation has accumulated on competitive mechanisms in parasitoids, some of 184
which can be found among the examples shown in table 5.4. Avoidence of
multiparasitism and combat to the death between first instar larvae within the host are fairly common. Advanced interference mechanisms such
as these are analagous to the vertebrate habit of territoriality, and would be expected to have similar consequences. While many of these
examples have accumulated, there has been little attempt to find systematic trends in the kinds of mechanisms used. For example, is there a greater
tendency towards behavioural mechanisms in the higher vertebrates? In each case one might expect several mechanisms to operate. How can we evaluate the relative contributions of these? Do the most important mechanisms change with trophic level? How do the mechanisms used in - interspecific competition compare with those used in intraspecific interactions? Presumably, a conspecific is more of a threat than a heterospecific, so a more intense reaction might be expected against a conspecific.
Case and Gilpin (1974) have explicitly included interference mechanisms in a competition model, since, as they point out, such mechanisms are known to be widespread, and can be assumed to be of some importance. They suggest that selection pressure drives a species towards either a) high exploitation efficiency or b) the development of interference mechanisms (but not both); and that interference competition may be another possible explanation for the existence of the "prudent predator". EThis problem can be briefly expressed by the question "what restrains natural enemies from overexploiting their prey, and hence driving both the prey species, and themselves, to extinction?" :1
Miller (1969) has discussed the population characteristics of species which compete primarily by exploitation or interference. He links small Table 5.4. Examples of some mechanisms of interspecific competition.
Competitors Trophic level Resource Type of mechanism Authors Comment
1. Yeast 0/1 Sugars Interference Gause (1932) Coexistence; limitation of populations Saccharomyces by alcohol production, also other waste eervisiae, products. Forerunner of competition by Schuosacchoromytes toxin production. kephir
2. Drosophila pavani, 1 Yeast Interference Budnik & Brncic Metabolic wastes of D. pavani interfere D. wiblistoni (and exploitation?) (1974) with the development of D. willistoni. (and others)
3. Herbivorous stem 1 Plant juices Interference Rathcke (1976) Competition detected between only two borers species; resulting from aggression rather than exploitation.
4. 4 species of Bumble- 1 Nectar and pollen Exploitation Heinrich (1976) No interference observed. bees, Bombus
5. Stingless bees 1 Nectar and pollen Interference Johnson & Hubbell Observations of interspecific aggression. Trigona app. (1974, 1975) Suggested that T. ,.Mviuentris is excluded from better plants in clumps by T. flcscipennie.
.. 6. Parasitoids 2 Flour moth a) Interference Fisher; (19614,1)) Physical combat between larvae within Nemeritie canesoena, Ephestia aericarium the host. Avoidence of multiparasitism. Norogenee b) Exploitation Phyaiological suppression (exploitation chrysoetictos of oxygen supplies) of younger larvae by older ones.
7. Encyrtid parasitoids 2 Scale insect Interference Bartlett & Ball Internal combat between larvae. Metaphycus luteolue, Coccus hesperidum (1964) • M. flavus •
8. Parasitoids 2 Tobacco budworm a) Interference Vinson (1972) Some avoidence of multiparasitism. Cardiochelis Neliothis vireecena b) Exploitation Combat between larvae. Oxygen nigriceps, starvation? (exploitation) of younger by crnpoletis older larvae. perdistinetus
9. Parasitoids 2 Housefly pupae a) Interference Wylie (1972) Predation of competitor eggs and larvae Nasonia vitripennis, Musca danestica b) Exploitation by M. zoraptor (also S. cameroni). _f_ Rapid development and exploitation of aoraptor, host by N. vitripennis -+ competitive Spalangia cxvneroni advantage. _
10. Parasitoids 2 Greater wax moth pupae Interference Ryan (1971) Avoidente of multiparasitism. ' Apecthis ontario, Galleria melonella •.. - ` Itoplectis quadricingulatue • 11. Parasitoids 2 Pine shoot moth a) Interference Zv'lfer (1971) a) Superior territorial behaviour by T. TaneZuoha Rhyaoionia buoliana interruptor larvae within host. interrupter, b) Exploitation b) Superior searching efficiency of ' Orgilus obecurantor adult 0. obecurantor.
12. Itopleatis maoulator, 2 European fir budworm a) Interference (7) Zw8lfer (1971) a)Itopleotis parasitizes CephaZoglypta Cephaloglypta Choristoneura murinana within host. murinanae b) Exploitation b) CephaZoglypta better synchronized with host.
13, Aptssie abdaainarr, 2 Winter moth g*fteitetion 2w3149r (197» Fester larval development of Aptveia, albiam.► Operapht ra brwmata higher reproductive capacity (7 and greater searching range) of Cyaenis.
14. Aeoogastar 2 Peach moth a) Interference Van Steenburgh 6 a) More active Macrocentrus larvae carpeeapsae, Laspeyresia molesta Boyce'(1937) destroy Ascogaater larvae in the host. Maarooentrue b) Exploitation b) Aseogaster can cause 561 parasitism anoylivorue compared to 37% Maerocentrus can produce on its own. This suggests that Asoogaater has a higher searching ability.
15. Crayfish 2 Interference ' Bovjerg (1970) Aggressive eviction of 0. immuns by 0. Orconectes virilis, virilia from crevices in substrate. 0. inerunis
16. Hermit crabs 2 Shells Interference Bach at al (1976) Aggressive interactions over suitable Clibanarius triooZor, shells. C. antillensis, C. tibicien
17. Stream fish Interference Cadwallader' Change in microhabitat distribution of Gobiomorphue (1975) G. breviceps in stream in presence of brevioeps, G. vulgaris. Galaxies vulgaris
18. Fish 1 Browsing territories Interference Robertson at ai Schooling can circumvent territoriality. Soarus oroioeneia, (1976) FoparxrQentrus pianifrons
19. Freshwater cichlids 2 Breeding territories Interference McKaye (1977) Competition very intense. Over.901_a11_ territories lost prior to completion of breeding cycle.
20. Reef fish 1? Browsing territories Interference Ebersole (1977) Greater tendency to show aggression EYrpcasaoentrus against 9 egg eating species, and against leuoostictus, those which have high diet overlap with and competitors territory holder. Table 5.4. Continued.
Competitors Trophic level Resource Type of mechanism Authors Comment
21. Tadpoles of 1/2 Food Exploitation De Benedictis Rana pipiena, (1974) R. aylvatica e 22. Chipmunks 1 Nuts Interference Heller (1971) Aggressive interactions. Habitat Eutamias app. Meredith (1977) structure influences outcome. • Sheppard (1971)
23. Pocket gophers 1 Interference Baker (1974) Aggressive interactions. Thomomye bottae, , T. talpoidee
24. Blue winged and (Interference) Murray & Gill Some interspecific aggressive behaviour. golden winged (1976) warblers •
25. Blackbirds 2 Breeding territories Interference Orians & Collier Strong interspecific aggression Agelaiue phseniceue, (1963) A. tricolor 1
NB: It is very difficult to show that competition occurs primarily by exploitation. (This is sometimes assumed when no interference mechanisms
have been identified; a most unsatisfactory method of detection).
Examples such as these may therefore be biassed towards interference.
Further examples of interference may be found in Case and Gilpin (1974).
Cin trophic level column 0 - plant, 1 ∎ herbivore, 2 - parasitoid/predator 3 188
body size, short generation`tiMe, rapid replacement rate and relative susceptibility to physical factors with exploitation competition; and large body size, long generation time, slow replacement rate and relative independence of physical factors with interference competition. At first sight, these two groups of characteristics are reminiscent of features of r and K selected populations, (see chapter two), although Miller does not use these labels.
Gill (1974), however, argued that "r-selection is not relevant to competition, but K-selection is a consequence of exploitative competition," and that "under exploitative competition K and competitive ability are inversely related". This serves to emphasize that r and K selection are not merely concerned with size and reproductive rate, but are primarily the responses to different kinds of environment.
Price (1973) and Force (1974) have discussed succession in parasitoid complexes with special reference to r and K selection, and concluded that
K-selected species tend to replace r-selected species as the community evolves, but they do not comment on the relative preponderance of exploitation and interference. Since so much information is already available on the parasitoid guilds they have studied, it would be very useful to attempt to measure the relative contribution of interference mechanisms to competition in these species.
It has been shown (Peters, 1976) that much competition theory is tautologous; that is, the predictions are implicit in the initial assumptions. However, this does not necessarily detract from its useful- ness. As with mathematical modelling, a thoroughly worked-out tautology shows us the implications of our premises. If x occurs, then y must follow. If we find condition x in nature, then, immediately, we also 189 know many other intersting and useful facts about the system. The link between x and y may be a long and tortuous one, which is not immediately obvious. The rigorous application of logic to find such links is a useful service rendered to other scientists. The question before us now is not "what is the consequence of these conditions?" but "how often are these conditions found naturally?"
Competition between species evolves away. It may do this by extinction or by causing shifts in resource utilization, thus playing an important part in the evolution of community structure and diversity.
However, environmental fluctuation and frequency dependent fitness may interrupt this process so that competitive interactions cannot be resolved. The smaller the advantage enjoyed by the superior competitor, the longer it will take for extinction to occur, and the greater the opportunity for divergence in resource use to appear. 190
CHAPTER SIX
COMPETITION BETWEEN SPECIES I.
6.1. Introduction
In this chapter, the results of experiments conducted with two species
of parasitoid are described, and examined for trends with the density of
the competing species, in an attempt to identify the mechanisms involved.
As discussed in the previous chapter, we now have a general idea of the
result of a competitive interaction (which varies depending on the
conditions). However, our knowledge of the mechanisms involved and their
relative importance is limited and disorganised. The information reported
here may help to begin to fill in the gaps.
Recent work has suggested (e.g. Neill, 1974) that the form of some
competitive interactions may be affected by the existence of others. In
this case, the presence of competitors from another species may well affect
competition between conspecifics. Intraspecific relationships in the
presence of a constant number of competitors are therefore briefly referred
to, and compared with the relationships obtained in their absence.
6.2. Materials and Methods
Parasitoids were standardized and experiments set up as described in
chapter three. In the case of the meal moth parasitoids, the conditions used were those described for the Bracon experiments. Conditions for the
aphid parasitoids Diaeretiella and Aphidius were exactly as before. In
each system both parasitoids were used together, holding the density of one species constant at 2 females, and varying the density of the other from 1 to 16 (or 32 in Nemeritis). 191
Small scale pilot experiments were set up for the Plodia parasitoids.
Six pairs of Bracon, or six Nemeritis females were standardized as usual,
and 30 Plodia larvae were left overnight in a small plastic butterdish (as
used for Bracon cultures). One large pinch of standard medium was
sprinkled evenly over the bottom of the dish to provide a thin layer. The
following morning the parasitoids were introduced, left for 24 hours and
then removed. The experiment was repeated several times, using Bracon and
Nemeritis on their own, and then introducing first one species, and then
the other the following day. This was done both ways round i.e. Bracon
followed by Nemeritis, and Nemeritis followed by Bracon.
In all cases, the offspring were allowed to emerge, and the remains
of the host larvae classified and counted as described in chapter three.
This experiment cuts out the effect of behavioural interactions of adult wasps from different species. In particular, it was hoped that this would show whether Nemeritis or Bracon offspring could develop on hosts previously parasitized by the other species. It does not, however, distinguish between an inability of the eggs to hatch and develop in the presence of a competitor, and a refusal to oviposit in hosts already parasitized by the other species.
6.3. Analysis of Results and Discussion
In the presence of a varying number of Aphidius females, Diaeretiella produces a highly variable number of offspring. However, there is no significant trend in the number of offspring produced per female with
Aphidius density.
Where there are two species of parasitoid present, attacking the host population simultaneously, there is no simple way of calculating a, the searching efficiency, for example, an Aphidius adult, originally also
192
contained a Diaeretiella egg. However, there are two methods of approx-
imation available. In the first method, the original host population is
taken as the number of hosts available and competitor offspring are
counted as healthy surviving hosts (for solitary parasitoids). This is
equivalent to allowing one parasitoid to act first, with the additional
assumption that the competitor completely avoids those already parasitized
by the first wasp, i.e. hosts parasitized by the first parasitoid are
effectively removed from the host population. In the second method the
parasitoid under consideration is taken to be the one that acts second,
i.e. the number of hosts available is the total minus the number of
competitor offspring, and the survivors are the healthy hosts recovered.
Original hosts Estimate 1. a = 1 . Log P e Survivors + Competitor offspring
1 [original hosts - Competitor offspring] Estimate 2. a = . Log Survivors
where competitor offspring E. hosts attacked by competitor species. For
gregarious parasitoids, substitute hosts attacked by competitor species.
Where discrimination is incomplete and eggs or larvae are lost in internal
competition within the host estimate 1 is too low and estimate 2 is too
high. When a was calculated for Diaeretiella using these two methods,
estimate 2 was always higher than estimate 1. It is therefore very likely
that the true value of a will lie somewhere between those calculated
using estimates 1 and 2.
In Diaeretiella, log a showed no significant trend with Aphidius
density. This is probably due to the fact that Aphidius is not attracted
to Brussels sprouts. Presumably they do not spend much time searching on
the plants. This would explain their low searching efficiency (see
chapter three) due to low rates of encounter with healthy hosts, and their 193 low effect on DiaeretielZa, due to low rates of encounter with Diaeretiella females.
However, as figure 6.1 shows, the percentage of females in the
Diaeretiella offspring falls significantly with Aphidius density. It is possible that encounters with other wasps affects the reproductive physiology so that fewer eggs are fertilized before they are laid. This might well be the case whether encounters occur in the host area (i.e. on the plant) or elsewhere (e.g. on the sides of the cage). Thus Aphidius individuals may affect Diaeretiella sex ratios via encounters away from the host area, but have little or no effect on the total numbers laid, since encounters on the host area are rare. However, there is no evidence for this explanation. As discussed in chapter three (see figure 3.8),
Diaeretiella sex ratios also fall with an increasing number of conspecifics, suggesting that a similar process occurs within species.
The pilot experiments on the PZodia parasitoids, show a very clearly marked difference between the action of the two wasps. While there was no significant change in the numbers of Bracon offspring produced when
Nerneritis adults also had access to the hosts, there were no surviving
Nemeritis offspring at all, when Bracon females also had access. In the absence of Bracon attack, about 13 Nemeritis adults were produced from 30 hosts, with approximately three hosts surviving, and the rest super- parasitized. It is possible that, in this experiment, when presented with paralysed hosts, Nemeritis completely avoided oviposition. However, no healthy hosts were provided, so it seems unlikely that no eggs were laid at all. In the small experiments with individual wasps reported in chapter seven, avoidence of paralysed hosts did occur, but was nowhere near 100%. Paralysed hosts known to contain Nemeritis eggs, never 1 44 00 14 0 %fema les in Diaeretiella 0. Figure 6.1.Thedeclineinthepercentage offemalesin 100 50 0
rapae t-test: b=t-4.49,with788degrees offreedom. present. offspring withthenumberof se b 4
Number of 194 Aphidius 8
present Aphidius matricariae Diaeretiella 12
16 195
produced Nemeritis adults, when kept for that purpose. This experiment
also shows that hosts containing Nemeritis eggs do not produce adults when
subsequently paralysed by Bracon. It suggests that Nemeritis eggs or
larvae do not survive in hosts which have been previously paralysed by
Bracon, which is confirmed by the experiments described in chapter seven.
The experiment also indicates that the presence of Nemeritis adults is
immaterial to Bracon; however, the full scale experiments discussed below
do show some significant trends in Bracon performance with Nemeritis
density.
In this case, estimate 1 was used for Bracon searching efficiency, as
hosts attacked by Bracon can always be distinguished. Neither method is
strictly correct for Nemeritis, so both are shown.
The number of offspring per Nemeritis declines logarthmically with
Bracon density, as shown in figure 6.2. Estimate 1 for a follows a similar relationship with log Bracon density, but estimate 2 seems to follow a concave curve (figure 6.3). This is probably because a is
increasingly overestimated as the number attacked by Bracon increases.
The true relationship lies somewhere between the two, but is unlikely to rise at high Bracon densities. The presence of Bracon wasps therefore affects Nemeritis reproduction in two ways: by decreasing the searching efficiency of the adults, and by the death of eggs and larvae in the hosts attacked by Bracon.
Log a and log s were calculated for Bracon as described in chapter three. Surprisingly, log a for Bracon also shows a curved relationship with log competitor density (figure 6.4). This is not an effect of the method of estimation since a host attacked by Bracon is always distinguishable as such. It has been suggested (chapter four) that 196
0 0.3 0.6 0.9 1.2 1.4 Log (Bracon + 1)
Figure 6.2. The decline in Neweritis offspring per head with increasing
Bracon density. [Since a log scale is appropriate for the
x axis, 1.0 has been added to each value so that the control
conditions, with no Bracon competitors present, may be
included]
t-test: b = t = -18.76, P <0.001 with 4 degrees of freedom. seh 197
1.0
0.5 a
0.3 0.6 0.9 1.2 Log (Bracon density + 1)
Figure 6.3. Tbe relationship between a (searching efficiency) of Admeritis canescens (of two individuals) with Log (Bracon density present). Cmn order to include the control points (se Bracon), 1.0 has been added to each Bracon density before taking logs]
A. Estimate 2 (x) y = 0.93 - 1.47x + 1.03x2 t-tests: b = t = -6.87, P <0.01 with 3 degrees of seb freedom c = t = 6.27, P <0.01 se c
B. Estimate 1 (o) Log y = -0.13 - 0.97x b = t = -8.08, P <0.002, with 4 degrees of freedom seb
198
-0.1 y = -0.70 - 0.57x + 0.47x2 x
x x
x
—0.5 x x
x
x x
-1.5
-1.8 0 0.5 1.0 1.5 Log (Nemeritis + 1)
Figure 6.4. Log a (efficiency of search for healthy hosts) in Bracon
plotted against Log (Nemeritis density). E In order to include control points, 1.0 is added to the number of
Nemeritis present before taking logs
t-tests: b = t = -2.04, P <0.05 1 with 61 degrees of seb 1 freedom c = t = 2.68, P <0.01 se C 199
Nemeritis adults increase host activity. If this is the case here, then
the number of encounters between Bracon and healthy hosts may well increase
when Nemeritis density is high. The initial fall in efficiency could be due to time wasted by Bracon following encounters with Nemeritis.
Observations on Bracon have suggested that they are easily disturbed when
encountering other parasitoids (of the same species or not), and will also "start and run" when touched from behind by a host larva. Figure
6.5 shows the decline in log s observed with log Nemeritis density. This could be due either to time wasting following encounters with Nemeritis, or to egg predation by the hosts. If Nemeritis females stimulate host
activity, then encounters between healthy hosts, and those with egg batches will increase and more egg clumps will be eaten, so the hosts
are classified as paralysed instead of parasitized, thus increasing the apparent value of a, and decreasing the apparent value of s. However, as discussed in chapter three, egg predation of this kind would be expected to depress the clumpiness of the egg distribution (i.e. reduce S2/m).
While there is a slight decline in S2/m with Nemeritis density, it is not significant.
The number of hosts paralysed per Bracon female increases with
Nemeritis density, as shown in figure 6.6. [It is surprising that there is no initial decrease, as might be expected from the log a plot (figure
6.4). However, there is a great deal of scatter in the data, and it may be that the initial decrease in log a is not really there, and is a spurious result of variation between replicates.] If, as already suggested, and discussed later, Nemeritis attack increases host mobility, then the encounter rate between healthy hosts and Bracon females would rise as Nemeritis density rises. The number of hosts paralysed per Bracon wasp would therefore also be expected to rise with Nemeritis density, and 200
-1.0
-1.5
-2.5
-2.6
0 0.3 0.6 0.9 1.2 1.5 Log (Nemeritis 4. 1)
Figure 6.5. The decline in Log s (efficiency of search for paralysed
hosts) in Bracon hebetor, with Nemeritis density.
t-test: b = t = -5.88, P <0.01 with 5 degrees of freedom. seb 201
50 y = 18.11 + 0.39x x x x
8 16 24 32 Nemeritis density
Figure 6.6. The increase in the hosts paralysed per Bracon female with
Nemeritis density.
t—test: b = t = 2.94, P <0.01, with 61 degrees of freedom. seb 202
produce a relationship like that Shown in figure 6.6.
The mean Bracon pupae recovered per host is plotted against log density of Nemeritis in figure 6.7. On the whole, there is no significant difference between the mean Bracon recovered per host, in the presence of any number of Nemeritis individuals and in their absence. The presence of a Nemeritis population would therefore be expected to have no overall effect on this parameter. However, there is a suggestion that, despite the very high variation between replicates, the mean Bracon pupae per host declines slightly as the density of Nemeritis increases. There might therefore be a slight difference between the performance of Bracon when low numbers of Nemeritis are present, and performance when the
Nemeritis population is high. It is possible that encounters with competitors could cut short oviposition sessions, thus lowering the mean number of eggs laid per host.
The percentage of females found in the offspring of Bracon declines with Nemeritis density, as shown in figure 6.8. This is not found when the density of conspecifics rises. Benson (1973b) has shown that males are better larval competitors than females, so a decrease in the proportion of females would be expected when the mean number of eggs per host rises.
However, a simple increase in the mean number of eggs is not found
(figure 6.7). The change in sex ratio is therefore probably not due to larval competition. It may be that, as in Diaeretiella, encounters affect the reproductive physiology of the wasp so that fewer fertilized eggs are laid. However, it is surprising, in that case, that a change in sex ratio is not obtained when the Bracon density rises. On the other hand these wasps are fairly slow moving, so the rate of encounter between conspecifics is very low. A low rate of encounter is unlikely to produce 203
ts hos
er p s egg
on Brac
Mean Control points
Q 0.5 1.0 1.5 Log (Nemeritis density + 1)
Figure 6.7. Mean Bracon eggs recovered per parasitized host plotted
against Log (Nemeritis density + 1). females in Brawn offspring Figure 6.8.Thedeclineintheproportionoffemalesoffspring 100 50 0 se Brawn asNemeritis b =-2.08,P<0.05,with384degreesof freedom b
8 Nemeritis
204 density increases. density 16
24
of 32 205
a significant effect of any kind, unless each encounter produce a very
strong reaction.
The number of hosts parasitized, and the number of pupae produced per
Bracon showed no significant trend with Nemeritis density, so the number
of Bracon offspring produced is unaffected by the presence of Nemeritis,
although the sex ratio declines. Press at al (1977) also found that
Bracon was unaffected by the presence of Nemeritis when parasitizing
Ephestia cautella. In the experiments reported here, the absence of a
numerical effect is due to a counter balance between two components of
competition. The number of hosts paralysed rises with Nemeritis density
(figure 6.6), but the searching efficiency for finding these, falls
(figure 6.5).
It was thought likely that the presence of a constant number of competitors might affect the intraspecific relationships discussed in chapter three, and so a few of these are discussed briefly below.
Turning first to the aphid parasitoids, the number of offspring produced per Aphidius females was examined for trends with conspecific density, in the presence and absence of two Diaeretiella females. In both cases the number of offspring was low and very variable, showing no significant trend with the density of conspecific parasitoids. A slight increase in sex ratio was observed in each case, but this was found to be insignificant. The overall sex ratio was low (28.5% in the presence of
Diaeretiella, 34.3% in its absence, with no significant difference between the two) .
From the information contained in this chapter, on the effects of heterospecific competitors, it should be possible to surmise what will be 206
the outcome of competition, if the species used here were to come into
contact naturally.
It seems highly unlikely the Diaeretiella rapae and Aphidius
matricariae would ever come into severe competition, on Brassica species
at least. Aphidius would be expected to die out very rapidly if this did
occur, due to its extremely low searching efficiency on these plants.
The effect on Diaeretiella would be a slight depression of the sex ratio,
and probably, of total population size in the next generation, quickly
made up when the competitor became extinct. If the Diaeretiella pop-
ulation was already at a relatively high level, then the decline in sex
ratio would not be expected to affect the size of the following generation,
since a decline in the number of females would be counter balanced by a
rise in searching efficiency due to a lower level of intraspecific
competition.
In the meal moth parasitoids, the data was examined to see whether
the presence of a constant number of heterospecific competitors affects
the relationships between conspecific density and the number of offspring
produced (in both species), and between conspecific density and the
numbers of hosts paralysed and parasitized by Bracon females. Graphs such
as these show the overall effect of competition i.e. they include the
effects of exploitation as well as interference.
Examination of plots of searching efficiency demonstrate the effects of competition on the rate of encounter between searching parasitoid and
host caterpillars, and exclude the effects of exploitation. Comparison of plots of offspring production and searching efficiency should therefore
provide an indication of the relative importance of exploitation and interference. 207
The decline in the log offspring produced per Nemeritis with the log
Nemeritis density in the presence and absence of Bracon females is shown
in figure 6.9. Since the conditions were changed slightly after experi- ments on intraspecific competition in Nemeritis, a series of replicates were run under the new conditions, with no Bracon individuals, to give a
correction point, which was used to estimate the new position of the
intraspecific relationship. The figure indicates that the presence of
two Bracon females both shifts the position and changes the shape of the
curve, depressing it disproportionately at high Nemeritis densities. A
very similar result is obtained when log a is used instead of log off-
spring (figure 6.10). This suggests that the change in the number of off-
spring produced is mostly due to a change in searching efficiency i.e. the most important contribution to the overall effect is made by interference mechanisms. This is probably due to the paralysis of hosts by the Bracon females. As discussed in chapter seven, Nemeritis adults tend not to recognise paralysed larvae as potential hosts. However, eggs are laid in them sometimes (by accident?), but these never reach maturity. When eggs are laid, the usual handling time (time for cleaning, resting, cocking etc.) must elapse before the wasp is ready to lay again. This would reduce the time available to search for healthy hosts. Non-recognition of paralysed hosts by Nemeritis means that paralysis has the effect of restricting Nemeritis access to the population of host larvae, although many such paralysed hosts are not actually used for rearing Bracon young.
At the same time eggs may be lost and time wasted when the ovipositor of a Nemeritis (accidentally?) jabs a paralysed larva. It has already been suggested that Nemeritis attacks stimulate defensive reactions in the caterpillars, and increase host mobility. At high Nemeritis density this would raise the encounter rate between Bracon wasps and healthy cater- 208
Figure 6.9. The decline in Log (offspring produced per Nemeritis) with
Log (Nemeritis density), in the presence and absence of Bracon
wasps.
A. Relationship obtained as described in chapter three, with
0.5 cm depth of medium (x).
y = 1.33 - 0.81x
b = -18.33 with 4 degrees of freedom, P <001. seb
B. Corrected relationship using a set of replicates at a
density of 2 Nemeritis (+), with a thin layer of medium.
(y = 1.93 - 0.81x)
It is assumed that the relationship is of the same shape
under the changed conditions. This may not necessarily
be true.
C.Relationship obtained in the presence of two Bracon wasps,
with a thin layer of medium (o).
y = 1.49 - 0.73x2
b = t = -19.32, P <0.001 with 4 degrees of freedom. seb Log (offspring produced per Ndmarittis) ō u, •0 UI 0 o 8
) 0 aN zaur 2. 824 rsuap Lj (
0
• 210
Figure 6.10. The decline in searching efficiency (mutual interference) of
Nemeritis canescens with log (conspecific density) in the
presence and absence of Bracon hebetor individuals.
A. Nemeritis alone, with 0.5 cm medium depth, as in
figure 3.9. Regression line only.
y= -0.62 - 0.73.
B. Nemeritis alone, with a thin layer of medium. This is
calculated using a correction based on a set of points
at 2 Nemeritis (+).
(y = 0.18 - 0.73x)
[It is assumed that the new relationship, B, is of
the same shape as the old, A, which may not necessarily
be the case
C. The searching efficiency of Nemeritis in the presence
of two Bracon females.
Cl. Estimate 1. (o)
y = -0.37 - 0.61x2
b = t = -15.26, P <0.001 with 4 degrees of freedom seb
C2. Estimate 2. (o)
y = -0.10 - 0.48x2
b = t = -8.43, P <0.002, with 4 degrees of freedom seb
Although it doesn't fit well, the x2 form is used for
C2 for the sake of comparison. 0 0.5 1.0 1.5 Log (Nemeritis) 212
pillars, which would result in an increase in the number of paralysed
hosts. This would then be expected to decrease the number of hosts
potentially available to Nemeritis adults, and hence the number of off-
spring produced, thus depressing the relationships shown in figures 6.9
and 6.10 at high Nemeritis densities.
We now turn to the Bracon data. The effects of intraspecific
competition in Bracon in the presence of two Nemeritis are shown in figures
6.11 to 6.15. The efficiency of search for healthy hosts, and hence the
total number of hosts paralysed, both follow log declines with the log
parasitoid density, as shown in figures 6.11 and 6.12. (The line B2 is
discussed below). These relationships are extremely close to those
obtained in the absence of Nemeritis wasps, but lie slightly above the
single species plots. This suggests that the presence of Nemeritis
competitors may slightly increase the efficiency of search for healthy
hosts, either by stimulating the defence reactions of the hosts and hence
their activity, or by stimulating the Bracon females to move about more
after encounters.
The number of hosts parasitized and pupae produced per Bracon female
are shown in figures 6.13 and 6.14. In neither case is there a signi-
ficant relationship of any kind with Bracon density in the presence of two
Nemeritis, compared with the humped curves obtained from Bracon females on their own. However, there are indications that the presence of two
Nemeritis increases both the number of hosts parasitized and the pupae
produced at low Bracon densities (one or two), and at very high Braeon densities (sixteen), although the differences are not statistically significant. This is probably due to the indirect effect of Nemeritis via host activity. At all Bracon densities, the presence of Nemeritis females 213
Figure 6.11. The decline in the log (hosts paralysed per Brawn) with log (Bracon density) in the presence and absence of Nemeritis canescens.
A. Without Nemeritis (x)
y = 1.51 - 0.59x
t = b = -20.08, P <0.001 with 3 degrees of freedom. seb
B. With 2 Nemeritis wasps (o)
B1 y = 1.56 - 0.62x, t = b = -5.69, P <0.02, 5 seb
points - 3 degrees of freedom.
B2 y = 1.41 - 0.45x, t = b = -5.76, P <0.02, 4 seb
points - 2 degrees of freedom. 214
) n oma Bra ?
er d p e s ly a r a p ts hos ( Zog
0 0.6 1.2
Log (Bracon density) 215
0
-0.5
of -1.0 OS
-1.5
-1.8 0 0.3 0.6 0.9 1.2 Log P (Bracon)
ī►igure 6.12. The decline in log a (efficiency of search for healthy hosts) in Bracon, with log Bracon density in the presence and absence of two Nemeritis.
A. Line only, no Nemeritis present. y = -0.63 - 0.36x
B. Regression line and points. Two Nemeritie present. y = -0.62 - 0.34x b = t = -2.92, P <0.01, with 44 degrees of freedom. seb 216
A (x) Hosts parasitized per Bracon in the absence of Nemeritis
B (o) It N t! in the presence of two Nemeritis.
B
A
0 8 12 16 Bracon density
Figure 6.13. The relationship between host parasitized per &coon female with Bracon density, in the presence and absence of Nemeritis wasps.
12 A (x) Nemeritis absent y = 0.99 + 0.67x - 0.3x2
B (o) Two Nemeritis present
41.
8 12 16 Bracon density
Figure 6.14. Pupae produced per Brawn female plotted against Bracon density, in the presence and absence of Nemeritia wasps. 217
may increase the activity of the healthy hosts. This increases the
probability of an encounter between a healthy host and a Bracon female,
and hence the number paralysed. This effect would be greater at low
Bracon densities, when the proportion of hosts left healthy is at its
highest. Nemeritis activity therefore cuts down the proportion of healthy
hosts, and hence the predation of these larvae on the Bracon eggs. The
production of Bracon pupae would therefore be higher in the presence of
Nemeritis, as would the number of hosts found paralysed. Figure 6.15 shows
that, unlike the results obtained in the absence of Nemeritis, there is no
apparent fall in searching efficiency for paralysed hosts at very low
Bracon densities. This tentatively supports the idea that oviposition by
Bracon is masked by egg predation at these low densities. Stimulate-
paralysis (by increasing the encounters between host and Bracon) and egg
predation is reduced, allowing a truer measure of oviposition to appear.
One might expect, in that case, that the number of hosts paralysed would
be somewhat higher than expected, at low Bracon densities, in the presence
of Nemeritis. A re-examination of figure 6.11 suggests that the data is
fitted equally well by a shallower relationship through the last four
points, with an abrupt rise to the lowest Bracon density (regression B2),
thus adding a little more circumstantial evidence. Experimental studies
designed to separate the two kinds of search and attack in Bracon are
clearly necessary for full understanding of the inter-relations between
the two phases. This could be very easily done by "sit and watch"
experiments using healthy hosts, replacing hosts with fresh healthy hosts
as soon as they become paralysed, for paralysis, and by providing a host
population consisting solely of paralysed hosts, for oviposition studies.
Lastly, systematic observations on egg predation by healthy hosts would also be of use. 218
Figure 6.15. The decline in log s (efficiency of search for paralysed
hosts) in Bracon hebetor, in the presence and absence of
Nemeritis canescens.
A. No Nemeritis present (see figure 3.11)
Regression line only
y = -1.31 - 0.43x
B. 2 Nemeritis present
y = -1.91 - 0.46x
t = b = -4.12, P <0.001 with 41 degrees of freedom. seb 219
0 0.3 0.6 0.9 1.2 Log P (Bracon) 220
At high Bracon densities, the number of pupae produced and hosts
paralysed also apparently rises in the presence of a constant number of
Nemeritis. Under these conditions the number of healthy hosts is very
low, and this effect is very unlikely to be due to egg predation.
The co-occurrence of Bracon and Nemeritis in the "wild" (warehouses
etc.) is fairly likely. Bracon will attack a wide variety of stored
product caterpillars (see chapter two) which includes many of the hosts
readily parasitized by Nemeritis. While the searching efficiency of
Nemeritis is much higher than that of Bracon (a = 13.48 m2/day for Nenieritis;
a = 0.69 m2/day, s = 0.18 m2/day for Bracon, for one individual), a
Nemeritis egg or larva does not stand a chance when its host is attacked
by a Bracon female. Thus, providing the density of host larvae is high relative to the density of Bracon, Nemeritis populations could survive on the hosts that the Bracon females miss. When Bracon densities are high however, Nemeritis competitors are bound to become extinct. When both parasitoids are at relatively low levels, Nemeritis stands a better chance of surviving. Bracon females require males, or the eggs they lay are unfertilized, producing only males in the next generation, which therefore becomes extinct. In addition, Bracon females have low searching efficiencies and lay relatively few eggs during their lifetime. On the other hand, a single Nemeritis female is quite capable of setting up a flourishing colony all on her own.
The major effects of interspecific competition between the two pairs of wasps studied are summarized in the next section. The inter-relations between Nemeritis and Bracon, when attacking the same host population are surprisingly complex, and appear to centre around the ability of Bracon to paralyse hosts, thus preventing Nemeritis from using them. 221
In the next chapter, observations on the effects of encounters with
Bracon adults on Nemeritis behaviour are examined, to see whether they make a significant contribution to the competitive interaction.
6.4. Summary
1. An increasing density of Aphidius females depresses the proportion of females in the offspring of Diaeretiella. This may be the result of encounters with the competitor species, away from the host area, affecting the reproductive physiology of the wasp. This also occurs when the density of conspecifics rises. Otherwise, Aphidius has no significant effect upon the performance of Diaeretiella.
2. The number of Nemeritis offspring, and its searching efficiency, decline sharply with the log density of Bracon present. A Nemeritis larva cannot survive in a host also attacked by Bracon.
3. Intraspecific competition in Nemeritis is also affected by the presence of a constant number of Bracon. The number of offspring, and searching efficiency relationships are shifted downwards, and depressed disproportion- ately at high Nemeritis densities, so that the new relationships are curved.
4. The number of hosts paralysed by a single Bracon female is stimulated by the presence of an increasing number of Nemeritis. However, the efficiency with which Bracon finds paralysed hosts for oviposition declines with the log Nemeritis density. There is therefore no signi- ficant change in the number of hosts parasitized or pupae produced as the density of Nemeritis increases. However, the mean number of eggs per host seems to follow a curved relationship with Nemeritis density, and the proportion of females in the offspring declines. This may be another 222 example of encounters affecting the reproductive system of a wasp.
5. The presence of a constant number of Nemeritis appears to have a slightly stimulating effect upon the performance of Bracon individuals in the presence of an increasing number of conspecifics. 223
CHAPTER SEVEN
COMPETITION BETWEEN SPECIES II.
BEHAVIOURAL OBSERVATIONS ON NEMERITIS CANESCENS
IN THE PRESENCE OF BRACON HEBETOR
7.1. Introduction
In this chapter, the behaviour of Nemeritis in the presence of a variable number of Bracon females is investigated, in an attempt to decide whether this is an important component of the overall effects of competition discussed in chapter six. Special reference is made to the effects of encounters with a parasitoid of a different species. Very little work of this kind has been attempted in insects, with the exception of some interesting studies on solitary bees by Johnson and Hubbell (1975).
It is possible that behavioural mechanisms could have an important effect upon the outcome of competition between insect parasitoids and predators and hence contribute to the success or failure of biological control _ programmes.
Behaviour was recorded as described in chapter four, noting inter— specific encounters in addition to the other events.
Individual Nemeritis were presented with a series of host larvae, half of which were paralysed previously by Bracon, to see if the response to a paralysed larva was the same as that to. a healthy one.
The effects of interspecific encounters were compared with those of intraspecific contacts.
7.2. Materials and Methods
Experiments were run with two Nemeritis and varying numbers of Bracon, 224
attacking a constant, high population of 5th instar Plodia larvae, as described in chapter six. Observations on the behaviour of one of the
Nemeritis wasps were recorded on a Rustrak event recorder as described in chapter four, except that observation periods were an hour long instead of
30 minutes, because of the very low rate of encounter with Bracon females.
Similarly, it was not felt worthwhile to watch replicates made at very low
Bracon densities (i.e. one or two females).
In addition, the response of a Nemeritis female to a paralysed host was determined as follows.
Newly emerged Nemeritis females were standardized as described in chapter three. At the same time, about 20 standard Piodia larvae were placed in a butterdish with five or six Bracon females and left overnight.
In addition, a silked-up surface was prepared for use the following day.
A new butterdish lid (with no holes or windows) was sprinkled with three pinches of standard medium. About 20 healthy host larvae were placed in an empty butterdish, which was then quickly inverted over the lid. A thick layer of silk and mandibular secretion was then deposited over the entire container, overnight. The next day these healthy hosts were removed (and used in the experiment) and the lid used as a base. The paralysed larvae were removed from the dish containing Bracon, and examined for eggs, which were removed. Sets of ten random numbers were extracted from random number tables, and five clean and five paralysed hosts presented to the standard Nemeritis in an order dictated by whether the random numbers were odd or even. The standard wasps were held in 5 x 2.5 cm glass vials, which were inverted over the prepared lid, providing a floor similar to that obtained in the larger experiments. The medium and silk layer also stimulated searching behaviour, so that observation time was kept to a 225 minimum. Clean or paralysed larvae were introduced singly under the base of the vial with a pair of forceps. The host was left in the vial until jabbed with the ovipositor of the wasp. It was then removed, and the
Nemeritis observed closely for cocking. It was found to be unnecessary to present another larva to stimulate cocking, as the presence of the prepared surface stimulated probing. If an egg had been laid, cocking occurred before probing was re-started. After a few preliminary trials with healthy hosts, this was checked by dissection. Under these conditions, eggs were always found in the host larvae when cocking had occurred.
This is presumably because the larvae were fairly closely confined, and did not have sufficient space to withdraw so that eggs were lost (although they tried).
Whenever a physical contact between wasp and host occurred, this was classed as an encounter and noted down. Jabs which did not result in oviposition were also noted. A given larva was presented, and re-presented to the parasitoid until an egg was laid in it. The hosts were then kept, singly in glass vials, with a little medium. A few of them were dissected after five days but most were kept at 25°C until the offspring emerged.
7.3. Analysis of Results and Discussion
In the presentation experiments, host/parasitoid encounters were classified as "rejected" and "accepted" depending on whether or not they were followed by an oviposition. Table 7.1 shows that encounters with paralysed hosts are rejected significantly more often than encounters with healthy hosts (76%). Observation suggests that this is because the wasp does not recognise the caterpillar as a potential host because it does not move. An accidental contact often results in host movement, which attracts the wasp's attention and frequently results in pursuit. 226
Table 7.1. Contingency table showing the number of encounters followed
by successful attack when Neweritis is presented with
healthy and paralysed hosts.
Rejected Accepted Larvae Totals X2 0 E 0 E
Healthy 92 118.37 82 55.63 174 26.76
Paralysed 257 230.63 82 108.37 339 P <0.001
Totals 349 164 513
0 = observed frequencies.
E = expected frequencies, based on the null hypothesis that there is no
difference between the two groups of data. 227
However, in 53% of encounters, a healthy host manages to evade a jab. (A
defeat rather than a rejection). When the wasp does manage to insert the
ovipositor (which can occur by accident in the case of paralysed hosts),
there is no significant difference in the number of successful ovipositions
in healthy and paralysed hosts. This suggests that there is not a
discernable chemical difference in the haemolymph of a paralysed host,
which could act to prevent oviposition. Thus, Nemeritis do avoid
ovipositing in paralysed hosts, but this is probably not due to chemical
o s but because they do not recognise a motionless larva as a potential
host. As far as a Nemeritis wasp is concerned, then, the presence of
Bracon females decreases the available host population. Accidental jabs
into paralysed hosts may result in oviposition, but the resulting eggs
are unable to mature (see chapter six). Dissections, after five days,
showed that eggs laid in paralysed hosts were able to hatch, as small
larvae were found. However, the development of the larvae was unable to
proceed as far as pupation, since pupae were never produced under these
conditions.
Once the behaviour traces had been transcribed, (in three second
units) the data was examined for trends with Bracon density. Figure 7.1
shows a surprising relationship, the increase in mean probe length as the
density of Bracon rises. It is not easy to understand why this occurs.
An increase in the number of Bracon present may have several effects upon
the conditions acting on a Nemeritis. Primarily one would expect the number of physical contacts between the two to rise. Indirectly, the
Braconid might produce a pheromone in increasing concentrations as wasp density rises, or the Nemeritis wasps might detect a change in the host population due to paralysis or parasitism. However, since the experiment was observed first thing in the morning, the latter effect would be at its 228
40 y = 6.20 + 0.53x x
x X
X x
0 8 16 Bracon density
Figure 7.1. The increase in mean length of probe session in Nemeritis, with Bracon density. t-test: b = t = 2.15, P <0.05, with 22 degrees of freedom seb
x y = 0.69 + 0.35x
be
ro x
p x in
s ter
n x l encou ta
To x x
0 8 16 Bracon density
Figure 7.2. The increase in the total encounters is probe of Nemeritis canescens (with Bracon and other Nemeritis) with Bracon density. t-test: b = t = 2.24, P <0.05, with 22 degrees of freedom. seb 229 minimum.
Encounters between wasps from the two different species were rare and could not be shown to follow a relationship with Bracon density.
Encounters between the two Nemeritis individuals were also recorded, and did not show a significant trend with the number of Bracon. However, when the total number of encounters (both within and between species) was considered, the total encounters in probe did follow a rising relationship with Bracon density, (figure 7.2) although the overall total (in any behaviour) did not. The low and variable rate of encounter between the two species may be masking trends, of which only one has been detected.
The number of cocking movements is very significantly related to the percentage of time spent probing (figure 7.3). This is in agreement with the data discussed in chapter four (figure 4.31), which shows a strong positive relationship between cocking and time spent searching when the percentage of time spent searching lies below 50% (as in the data presented here). However, this data was collected at the beginning of the experiment. The corresponding data from chapter four falls in the second half of figure 4.31 (when the time spent searching exceeds about 50%), and shows such wide variation that there is no overall relationship between cocking and time spent searching (discussed in chapter four). It is interesting that searching time early in the experiment is much higher in the previous experiments than those reported here. This is not due to the presence of Bracon wasps since a set of replicates run with two
Nemeritis only, shows an average time spent searching of 36%, which is much lower than the 60% time spent searching at a density of two Nemeritis in the experiments of chapter four. The difference in activity is there- fore probably due to the change in conditions between the two sets of Figure 7.3.Theincreaseincockingmovements
Number of cocking movements 0 t-test: b=4.63,P<0.001,with presence ofvaryingnumbers in onehour'sobservationof se b
% Timeprobing 230 25
Nemeritia B
with 22 degrees hebetor. canescens, the timespentprobing of freedom. 50 in the 231
experiments, and indicates that the presence of a deeper medium layer
encourages searching.
Figures 7.4-6 show that the times spent probing, walking and resting
are significantly related to the number of probes and walks, and the mean
rest period, respectively. Obviously, the amount of time spent on a
behaviour depends equally on the number and mean length of each behaviour
session. However, if one component varies much more than the other, this
will cause variation in the total time spent in that behaviour, which may
obscure the relationship with the other component. An examination of the
data did show that the number of probes and walks, and the mean rest were
more variable than the mean walk and probe, and the number of rests.
Figures 7.7 and 7.8 show that the mean walk and mean rest period are
inversely related to the number of rest and walk bouts. This would produce
a feedback effect, stabilizing the total time spent walking or resting at roughly the same level. One possible explanation is that there is a maximum amount of time which can be spent on any one activity; and when
this maximum is approached the animal compensates by balancing mean and number against one another. If this is so, the relationship between the
two components would be expected to fade away when the time spent on that
activity is well below the maximum. In chapter four, the time spent walking and resting is much lower, and there is no significant relation- ship between the components. Here, there is more time spent resting than walking, and the relationship between the rest components is more significant. In figure 4.22, the significant relationship between probe components (early in the experiment) is found when the time spent probing is high (about 70%); at lower levels of activity the relationship is lost.
In this set of experiments the time spent probing is low, and no relation- % Time spent probing 50 25 Figure 7.5. 0 Figure 7.4.Thedependenceofthetotaltimespent 0
% Time spent walking y = 3.79 +0.22x canescens The dependenceofthe totaltime of walking sessionsinone t-test: b=t = 4.15, number ofprobesessionsinonehr's t-test: b=t6.90,P<0.001,with Bracon Nemeritis hebetor. se se b in canescens b 50
the presence of Number ofprobesessions Number ofwalkingsessions 232 in thepresenceof 100 P <0.001 with22 degreesof freedom.
hour's observation of varying spent 100 •
numbers of walking onthenunber observation of varying numbersof 22 degreesoffreedom. probing withthe 200 Bracon Nemeritis
x 240 hebetor. 150
x 170 233
20 40
Mean length of rest (secs)
Figure 7.6. The dependence of total time spent resting on the mean
length of rest (secs) in one hour's observation of Nemeritis
canescens in the presence of varying numbers of Bracon
hebetor.
t-test: b = t = 7.68, P <0.001 with 22 degrees of freedom. seb 234
Figure 7.7. The inverse relationship between mean walk length (secs)
with the number of walk sessions in one hour's observation
of Nemeritis canescens in the presence of varying numbers
of Bracon hebetor.
t-test: b = t = -3.08, P <0.01, with 22 degrees of freedom. seb
Figure 7.8. The inverse relationship between mean rest length (secs)
with the number of walk sessions in one hour's observation
of Nemeritis canescens in the presence of varying numbers
of Bracon hebetor.
t-test: b = t = -3.89, P <0.001, with 22 degrees of freedom. seb 235 ) ecs (s h t leng lk wa n Mea
0 100 200 240 Number of walk sessions
) cs h (se t ng le t es r Mean
0 50 100 140 Number of rest sessions 236
ship between the components was discerned.
Encounters between the species were examined and classified as
described in chapter four. Random points were also generated and
similarly classified. Table 7.2 shows that encounters with a Bracon wasp
cause change in the behaviour of a Nemeritis, over and above that
expected due to random change. Table 7.3 shows the effect of encounters
when they occur in the different behaviour patterns. Unfortunately,
section 3, encounters during rest, is not valid as one of the expected
values is less than 5. However, the first two sections show that
encounters in probe or rest do not significantly increase change. The conclusion from the third section, that encounters in rest cause change is likely to be correct, since this is the only remaining source of the
significance found in table 7.2. The direction of change was not found
to be significantly affected by an encounter, whatever the insect was
engaged in doing.
The data was then examined for relationships with the three sets of encounters (in probe, walk and rest) but none was found, probably because the numbers of encounters obtained were low.
Encounters with another Nemeritis were investigated in the same way.
Tables 7.4 and 7.5 show that these encounters also cause change, overall, and this overall change is due to encounters in probe and rest.
Encounters in walk do not have a significant effect on the frequency of change. Table 7.6, however, shows that encounters in walk do affect the direction of a change, shifting it more towards probe. Encounters in probe and rest do not have a significant affect on the direction of a change.
Encounters between conspecifics did produce a significant effect on 237
Table 7.2. Contingency table showing the effect of encounters with Bracon
on the frequency of changes in the behaviour of Nemeritis.
Change No change Data source T X2 0 I E 0 E
34.74 Encounters 39 16.82. 47 69.18) 86 P <0.001
T = totals
0 = observed frequencies
E = expected frequencies 238
Table 7.3. Contingency tables showing the effect of interspecific
encounters on the frequency of changes within each activity
in Nemeritis.
A. Changes in Probe
Change No change Data source T X2 i s E 0 ! E 6.53 Encounters 14 7.28 2 P' B. Changes in Walk Change No change Data source I T X2 0I E 0 1 E 0.03 Encounters 13 109 10 9.91 23 NS C. Changes in Rest Change No change Data source T X2 0 ( E E 46.41- Encounters 12 2.:09* 12 21.91 24 1 (P <0.001) T = totals 0 = observed frequencies E = expected frequencies * Part C is not valid because this expected value is less than five (see text). 239 Table 7.4. Contingency table showing the effect of encounters with conspecifics on the frequency of changes in behaviour of Ne11lNrZtZs . Change No change Data source T X2 0 E 0 E 125.51 Encounters 85 29.72 67 122.28 152 P <0.001 T = totals 0 = observed frequencies E = expected frequencies 240 Table 7.5. Contingency tables showing the effects of encounters with conspecifics on the frequency of changes in Nemeritis within each activity. A. Changes in Probe Change No change Data source T X2 0 I E 0 I E 29.07 Encounters 27 10.64 30 46.36 57 P <0.001 B. Changes in Walk Change No change Data source T X2 E 0 0 I 0.001 Encounters 29 29.60J 2 3 22.40 52 NS C. Changes in Rest Change No change Data source T X2 0 E 0 E 178.96 Encounters 3.75* I 14 43 p <0.001 T = totals 0 = observed frequencies E = expected frequencies *uThis value is< and y a 5, -- therefore part C of the table is not strictly valid. 241 Table 7.6. Contingency table showing the effect of encounters with conspecifics on the direction of change during walk in Nemeritis. Walk -} Probe Walk -> Rest Data source T X2 0 E 0 E Natural changes 145 151.08 210 204.84 355 4.27 Encounters 18 11.98 11 16.16 28 P <0.05 Totals 163 221 383 T = totals 0 = observed frequencies E = expected frequencies 242 the overall behaviour of a wasp. Figures 7.9 and 7.10 show that the number of probing sessions is significantly related to the encounters in walk and rest. Since time spent probing is related to the number of probes, these effects should increase probing time, and possibly, searching efficiency. This would oppose, and hence cannot explain, the decline in Nemeritis searching efficiency found with Bracon density in chapter six. Figure 7.11 shows that the number of walks is related to the number of encounters in probe. The effects found when considering interspecific encounters are also found associated with intraspecific contacts, plus a few more. This may be either because intraspecific encounters have more effect, or because, in this set of data, there are fewer interspecific contacts. So far, the data indicates that intraspecific encounters have more effect (which may be because they are more frequent) but that this effect would be expected to be positive rather than negative, i.e. stimulate search. The two sets of encounters were compared in the same way that each set was compared with the random points. There was no significant difference between the two in the frequency with which they caused change. The two sets of data were therefore pooled and retested against the random points. Table 7.7 and 7.8 show that pooled encounters cause change, due to the encounters in probe and rest. The X2 value for change in rest is higher than that obtained from the intraspecific data alone, thus supporting the previous conclusion, that interspecific encounters cause change in rest. Table 7.9 shows that pooled encounters in walk shift change significantly towards probe. This is more significant than the value obtained from encounters with Nemeritis alone, suggesting that encounters with Bracon also have this effect, but that the data set is too 243 Figure 7.9. The increase in the number of probe bouts with the number of encounters with conspecifics while walking in one hour's observation of Nemeritis canescens in the presence of varying numbers of Bracon hebetor. t-test: b = t = 2.64, P <0.02 with 22 degrees of freedom. seb Figure 7.10. The increase in the number of probe bouts with the number of encounters with conspecifics while resting in one hour's observation of Nemeritis canescens in the presence of varying numbers of Bracon hebetor. t-test: b = t = 2.29, P <0.05, with 22 degrees of freedom. seb Number o f probe bou ts 0 0 Encounters withconspecificswhilewalking Encounters withconspecifics whileresting 244 5 5 y = 69.85 + 10.68x 10 10 245 y = 133.91 + 8.74x 0 5 10 Encounters with conspecifics while probing Figure 7.11. The increase in the number of walk periods with the number of encounters with eo specifics while probing, in one hour's observation of Nemeritis canescens, in the presence of a varying number of Bracon hebetor. t-test: b = t = 3.41„ P <0.01, with 32 degrees of freedom. seb 246 Table 7.7. Contingency table showing the effect of all encounters (be- tween and within species) on the frequency of changes in the behaviour Nemeritis. Change No change Data source T X 2 0 E 0 E 158.20 Encounters 124 46.54 114 191.46' 238 P <0.001 T = totals 0 = observed frequencies E = expected frequencies 247 Table 7.8. Contingency tables showing the effect of all encounters (be- tween and within species) on the frequency of changes in Nemeritis within each activity. A. Changes in Probe Change No change Data source T X2 0 E 0 E Encounters 41 17.92 55 70.08 96 31.49 P <0.001,' B. Changes in Walk Change No change Data source T X2 0 E 0 E Encounters 42 42.69 33 32.31 75 0.001 NS C. Changes in Rest Change No change Data source T X2 0 0 Encounters 41 5.84 26 61.16 67 225.35 P.<0.001 i T = totals 0 = observed frequencies E = expected frequencies 248 Table 7.9. Contingency table showing the effect of all encounters (be- tween and within species) on the direction of change in walk in Neweritis. Walk - Probe Walk -} Rest Data source T X2 0 E 0 E Encounters 27 18.20 15 23.80 42 7.47 Natural changes 145 153.80 210 201.20 355 P <0.01 Totals 172 225 295 T = observed frequencies 0 = observed frequencies E = expected frequencies 249 small to show it. Figures 7.12 and 7.13 summarize the effects of the pooled encounters in the same way as figures 4.25 and 4.26 in chapter four. For example, figure 7.12 shows that, in probe, a random point in time (generated from random number tables) is followed by a change of behaviour within the next three seconds in 18.67% of cases. An encounter in probe is followed by a change of behaviour in 42.71% of cases. After a random point, this change is to walk in 89.42% of cases, compared with 92.68% after an encounter, although this difference is not significant. On ten occasions, a random point in probe is followed by more probing 8.13 times, compared with 5.73 after an encounter. In figure 7.13, encounters in walk (centre) are shown to have an effect upon the direction in which changes occur, shifting the tendency more towards probe. This increases the number of probing bouts (as shown in figures 7.9 and 7.14), which in turn increases the percentage of time spent probing. This will affect the number of encounters in probe, and so on. The relationship between the number of walks and the total encounters in probe is shown in figure 7.14. This is slightly less significant (lower t value) than the corresponding relationship, shown in figure 7.11, with the intraspecific encounters only. This is concsistent with the decrease in X2 when changes in probe are considered, thus confirming that interspecific encounters in probe do not have the same disrupting effect on searching as encounters with a conspecific. Figures 7.15 and 7.16 show that the number of probing bouts is again related to the number of encounters in walk and rest. The relationships, based on both inter- and intraspecific encounters are apparently more significant than those shown in figures 7.9 and 7.10. If the increase in 250 Figure 7.12. Summary of the effects of heterospecific and conspecific encounters on changes in Nemeritis behaviour. Vertical arrows indicate the effects of encounters. Dotted lines indicate that the change is not significant. R = after random points E = after encounters For example, random points in rest are followed by more resting (no change) in 91.28% of cases i.e. 9.13 times out of 10 occasions. Encounters in rest are followed by more resting (no change) in 38.81% of cases i.e. 3.88 times out of 10 occasions. 251 52.10% hange 19.55% At any time 47.90% No change 80.45% walk 89.42% 92.68% Change 18.67% NS 42.71% Probe rest 10.58% o change 81.33% 7.32% 57.29% robe 40.85% 64.29% hange.56.92% 56.00% rest 59.15% Walk ; NS 35.71% No change 43.08% 44.00% probe 16.07% 17.07% hange 8.72% 61.19% NS Rest walk 83.93% No change 91.28% 82.93% 38.81% On 10 occasions: R E R E robe 8.13 5.73 probe 2.33 3.60 Probe walk 1.67 3.96 Walk< >walk 4.31 4.40 rest 0.20 0.31 rest 3.37 2.00 R E probe 0.14 1.04 Restes >walk 0.73 5.07 rest 9.13 3.88 Figure 7.13. Summary of the overall effects of inter- and intraspecific encounters on the behaviour of adult Nemeritis. A B indicates that A affects B, positively, Ei. +ve J , or negatively, [:= -ve ], as shown. Figure numbers show which figures display the relationships. Dotted lines and brackets, e.g. (+ve), indicate expected relationships which have been found to be insignificant. %Tp, %Tw, %TR, = percentage of time spent probing, walking and resting respectively. (1) Encounters in probe increase the tendency to walk when a change in behaviour occurs. (2) " " walk rr rr rr " probe " rr 11 rt rt rr (3) " " rest It rr rr " probe " rr ►r rr n rr +ve Number Fig. 7.10, 7.15 of i +ve probes (-ve) Fig. 7.9, 7.14 (-ve) Fig. 7.4 +ve / (+ve) Mean probe %Tp / N'14-ve) +ve (1) Encounters in \ /\ probe Change_—> 1 //(-"e) Encounters +ve in rest n3 (2) Encounters Change in walk +ve Mean (+ve) walk / -ve +ve Fig. 7.7 Fig. 7.5 Fig. 7.11, 7.16 Number +ve of walks Number of walk periods Figure 7.14.Theincreaseinthenumber ofwalkperiodswithtotal 0 one hour'sobservationof t-test: b presence ofvaryingnumbers encounters inprobe(with se Total encountersinprobe b t =2.95,P<0.01with22 254 Nemeritis canescensin Bracon 8 Bracon hebetor. and other degrees Nemeritis), of freedom. the in 16 255 Figure 7.15. The increase in the number of probe sessions with encounters (with Bracon and other Nemeritis) during walking, in one hour's observation of Nemeritis canescens. t-test: b = t = 3.03, P <0.01, with 22 degrees of seb freedom Figure 7.16. The increase in the number of probe sessions with the encounters (with Bracon and other Nemeritis) during rest, in one hour's observation of Nemeritis canescens. t-test: b = t = 2.40, P <0.05, with 22 degrees of seb freedom. Number of probe sessions Number of probe sessions 0 Total Total encounters 256 encounters inrest 5 5 in. walk 10 10 257 significance is real, this indicates that heterospecific encounters do affect the number of probing bouts. Pooling these two relationships produces figure 7.17, where the number of probes is plotted against the total number of encounters in rest and walk. This is highly significant. The percentage time spent probing would therefore be expected to be related to the total encounters in rest and walk, which it is (figure 7.18). However, the total encounters in rest and walk are not significantly related to Bracon density, so there is no discernable association between the time spent probing and Bracon density. This, therefore, is not the explanation for the interspecific interference relationship found in chapter six. [ In any case, the trend is in the opposite direction. As figure 7.2 shows, total encounters in probe follow a rising relationship with Bracon density. This could contribute towards an interference effect if encounters in probe shortened mean probe length, and this in turn had a significant effect on the time spent probing. However, the last two conditions are not fulfilled, probably because encounters are so rare. What, then, is the cause of the interference, found in the last chapter, inflicted on Nemeritis by an increasing density of Bracon? There are at least three possibilities: a) Over the whole course of the experiment interspecific encounters might decrease searching time, although this is not detectable during the first hour. This seems improbable since such encounters are so rare (a Nemeritis encounters its single conspecific more often than any Bracon, even when there are 16 Bracon present). Also, previous experiments have 258 X x x x ds 100 y = 60.93 + 5.42x io r e p x x be ro p f o ber Num x 0 10 20 Total encounters in rest and walk Figure 7.17. The increase in the number of probe periods with encounters, with Bracon and other Nemeritis, while resting or walking, in one hour's observation of Nemeritis canescens. t—test: b = t = 2.72, P <0.02 with 22 degrees of freedom. seb 259 0 10 20 Total encounters in rest and walk Figure 7.18. The increase in time spent probing by a Nemeritis individual, with the total number of encounters (with other Nemeritis and Bracon adults) in rest and walk. t—test: b = t = 3.33, P <0.01 with 22 degrees of freedom. seb 260 shown that activity falls considerably as the experiment proceeds. To test this would require watching for 24 hours, or running another set of experiments, which would be terminated as soon as the observation period was over. b)Eggs accidently lost in paralysed hosts could waste handling time (a recovery period of washing, walking and resting), thus reducing the time available for probing. This could not be'expected to be detectable at the beginning of the experiment, when the proportion of hosts paralysed is still very low. Again, very long observation periods, or observations near the end of the experiment would be necesary to detect this. c)A decrease in the density of perceivable hosts during the course of the experiment, due to paralysis, might reduce the tendency to search as the density of Bracon females increases. Certainly, this would produce a decline in the oviposition rate. Waage (1977) has shown that oviposition affects the decline of the response shown by Nemeritis to an area of host secretion. It does not seem unreasonable to suggest that the time spent probing may also be related to oviposition rate. This could be tested by presenting a single Nemeritis with varying proportions of healthy and paralysed hosts, and noting the rate of oviposition, and the time spent probing. 7.4. Summary and Conclusions a)Encounters with paralysed hosts are followed by attacks in Nemeritis significantly less frequently than encounters with healthy hosts. This may affect the time spent probing and hence the searching efficiency of a Nemeritis, relative to the number of Bracon present. b) Encounters with a conspecific causes change of behaviour in probe 261 and rest in Nemeritis. Encounters in walk cause a shift in the direction of changes towards probe, although the frequency of such changes is unaffected. c)Encounters with a Bracon female produce similar results, except that encounters in probe do not significantly affect the frequency of changes in behaviour. They therefore appear to be less important than conspecific encounters. d) All encounters in rest and walk increase the number of probing sessions, and encounters between conspecifics in probe increase the number of walks. However, encounters are rare, and these trends do not produce a significant link between time spent searching and Bracon density, or explain the interference relationship found in chapter six. 262 CHAPTER EIGHT GENERAL DISCUSSION In this final chapter an attempt will be made to bring out the features of general interest which emerge from this study, and thus place it in perspective. The paramount importance of competition within and between species to population dynamics, pest control, and the evolution of both species and community structure and composition has already been discussed in chapters two and five. This study has been an attempt to look at some of the components of competition, in order to try to decide which are of the greatest importance to the competitive interaction. The significant decline in searching efficiency (a) in Nemeritis with conspecific density shows that the decline in the number of offspring per head, is substantially caused by an apparent decline in the number of encounters between a single parasitoid and its hosts. The behavioural observations reported in chapter four, however, show that, while some of the consequences of encounters between conspecific wasps could reduce the searching efficiency via the time spent searching (encounters in probe cause changes in behaviour), others boost the search time. Thus the effect of encounters during search (decrease search time) is counter balanced by the effect of encounters when the animal is not searching. Loss of eggs from the ovipositor (as shown by cocking) remains more or less constant with parasitoid density, although the number of offspring declines sharply. This suggests that an increasing proportion of eggs are lost. Since no evidence of encapsulation of young larvae or eggs was obtained, thus supporting Salt (1975), it is suggested that eggs are lost 263 before they can be lodged safely within the host, due to the defensive behaviour of the larvae. If the population of host larvae is increasingly stimulated as the number of attacks rises, this would produce an increase in the failure rate of parasitoid attacks, and hence a decline in searching efficiency i.e. an increasing proportion of encounters between wasp and hosts become "invisible". Increasing Bracon density produces no overall effect on the number of healthy Bracon offspring surviving, despite the fall in searching efficiencies as the number of wasps increases. This is probably due to the concurrent decrease in the number of healthy hosts left, which have a tendency to eat Bracon eggs. As Bracon wasp density increases, more hosts are paralysed, and less hosts are left healthy, to cause mortality by egg predation. Bracon wasps affect the performance of Nemeritis primarily through their habit of paralysing hosts. Nemeritis eggs laid in paralysed hosts will hatch, but cannot complete their development. Individual Nemeritis wasps seem to have difficulty in recognising a paralysed larva as a potential host. The presence of Nemeritis individuals apparently marginally raises the searching efficiencies of a Bracon female, perhaps by increasing the mobility of the host population (via stimulation due to Nemeritis attack). Thus the two most influential factors in the competitive interaction of Nemeritis and Bracon are probably a) the reactions of the host larvae and b) the habit of Bracon wasps of paralysing their hosts. Most studies of defence mechanisms in the larvae of Lepidoptera such as Plodia have been concerned with internal immunity reactions to 264 parasitoid eggs and larvae (for example; Rogers, 1970; Salt, 1975; Moiseyeva, 1974). Behavioural avoidence mechanisms have received little attention so far. The work reported here suggests that this might well be a profitable area of study. Rogers (1970) has shown that immune reactions to Nemeritis eggs and larvae in Ephestia cautelZa may decrease the apparent searching efficiency of a Nemeritis adult. Behavioural avoidence may have the same effect. Not only may such defence systems decrease the efficiency of an attacking parasitoid (as suggested here by avoiding Nemeritis attack and concurrently causing parasitoid egg loss; and by predation of Bracon eggs), but they may have an indirect effect on the competitive interaction between parasitoid species. The reaction td Nemeritis attack, increased host mobility, may make them more susceptible to attack from Bracon, by increasing the rate of encounter between Bracon and the host larvae. There are a number of other possible effects which may result from increased host mobility (see figure 8.1). Increased encounter rate with Bracon adults may have the effect of disturbing them, causing changes in behaviour or wasting search or oviposition time. This could reduce both kinds of searching efficiency. Increased encounter rate with Bracon eggs would increase egg predation and reduce apparent s. The overall result will depend upon a balance of all these i.e. whether, for example, a fixed increase in host mobility increases encounters with Bracon adults more than it increases encounters with eggs. On the other hand host movement appears to be an important feature used by the parasitoids for recognition. It seems to be one of the cues which aid Nemeritis in identifying a potential host, and must also be important to a Bracon female in determining whether an attack will result in paralysis, or oviposition. 265 frequency of Nemeritis attack +ve +ve mobility of PZodia larvae tendency to react violently to attack; defensive with- drawal coincidentally +ve +ve causing Nemeritis egg loss. Change in a for Nemeritis. encounters encounters with Bracon with Bracon adults eggs +ve egg ? -ve predation Time waste number of Bracon by Bracon? +ve eggs (change in s) Disturbance //i+ve -ve number of +ve encounters hosts paralysed >between Bracon (change in a and paralysed for Bracon) hosts Figure 8.1. Showing the possible effects of Nemeritis attack on healthy Plodia larvae, in the presence of Bracon females (+ve indicates that a positive relationship would be expected be- tween the parameters linked). 266 This serves to emphasize that interactions between trophic levels are not just one-way. Where there is the potential, a host/prey/plant pop- ulation will be subject to selection pressures to evolve and perfect defence mechanisms. These will affect the population dynamics of the parasitoid/predator or herbivore species in the short term, and exert selection pressure which will influence their evolution in the long term. Recently there has been accelerating interest in the coevolutiotu of plants and animals (e.g. Gilbert and Raven, 1975). It is to be hoped that the research stimulated by this will spread to evolutionary changes generated by interactions between trophic levels higher up the food chain, which also merit investigation. It has been seen that paralysis is a useful and powerful trait for Bracon. Since this is a relatively slow-moving, easily disturbed insect, with a fairly low searching efficiency, it is doubtful if the species could survive without the ability to paralyse. Certainly it could not hold its own in competition with Nemeritis. It seems improbable that it would be able to use Plodia larvae as hosts at all. Oviposition is a slow process; it is unlikely that eggs could be placed on a moving target. Even if this were possible, eggs would be dislodged as the caterpillar moves through the medium. A host carrying eggs is also likely to eat them. Host feeding by the parasitoid would be very difficult. Paralysis is therefore a necessity for the survival of parasitoids with these features. Presumably the ability to paralyse allows the evolution of such traits, which would be hopelessly impractical without it. The incidence of the use of paralysing venom in the parasitoid insects would therefore be expected to coincide with ectoparasitism of active larval hosts. In the Braconidae, ectoparasitism, together with an ability to inflict paralysis is characteristic of the subfamilies Braconinae and Doryctinae, and is 267 thought to represent the primitive condition of the family (Matthews, 1974). Paralysis and. ectoparasitism is also found among the chalcids and the ichneumonids, for example as practised by Melittobia acasta, which attacks bee larvae, and by Rhyssa, which parasitizes sawfly larvae (Askew, 1971). In the Vespoidea, such as the spider hunting wasps (family Pompilidae), paralysis of the host or prey regularly occurs. Eggs are laid on the host, or beside it in a cell constructed by the adult. E In these insects, it becomes difficult to distinguish between predation and parasitism. Sometimes an egg is laid on one paralysed spider or insect, sometimes beside or near many.] An ability to induce temporary paralysis has also been found in an Apocephalus species, a fly of the family Phoridae, which lays its eggs singly on the necks of leaf cutting ants (Askew, 1971). It is possible that an ability to paralyse might well confer upon these insects, as in Bracon, an advantage over competing parasitoids wt}Q,h normally attack their hosts without paralysing them, simply because they are not optimally adapted to using paralysed hosts, but healthy ones. In this case, Bracon induced paralysis is a powerful counter balance to the higher searching efficiency and fecundity of a Nemeritis wasp. Not only does a Nemeritis lay far more eggs in one day, but since males are not necessary, no time and energy are wasted on mating or on laying male eggs. The potential female offspring of a single Nemeritis is therefore much higher than that of a single Bracon. Also, a single healthy Nemeritis is sufficient to ensure the survival of the population, while a Bracon female must find a mate. When competing with Horogenes chrysostictos, a similar, but bisexual ichaeumonid, for Ephestia kuhnielia larvae, the lack of necessity to produce males in Nemeritis is an important factor influencing the outcome of competition (Fisher, 1962). 268 In this case, however, since a Nemeritis individual cannot emerge from a paralysed host, the advantages of efficiency and fecundity are out- weighed whenever the Bracon population is high enough relative to the host population to paralyse all or most of them. However, as long as there are marginal host populations, not attacked by Bracon, which some Nemeritis can reach, the two species can coexist. This is a very good example of counter balanced competition, as described by Zw8lfer (1971). He distinguishes between intrinsic competition (the elimination of a competitor in direct contact over a single host) and extrinsic competition (the relative ability to exploit the host population). In this study, Bracon is the superior intrinsic competitor, due to paralysis, and Nemeritis is the superior extrinsic competitor, due to its higher searching efficiency. Since paralysis produces absolute exclusion of Nemeritis from those particular hosts, it could be looked upon as a form of territoriality (or sequestering). Some of the advantages which are enjoyed by the higher territorial vertebrates might therefore be expected to accrue to an insect which employs host or prey paralysis. In times of food shortage, a species using this method of interference competition would be at a great advantage compared with a simple exploiter. Apart from a consideration of which is the most significant component, perhaps the most important observations arising from this study are as follows: 1) An apparently simple response may in fact be composed of a number of contributary components, each of which may change in a different way as conditions change. There is a danger that a simple mathematical model, while adequate for the range of conditions investigated, will break down 269 when more broadly applied. There is no guarantee that a complex model, describing each component, will be any more useful; but if it is based on real observable mechanisms, then this is more likely. 2) A given response which has been found in a number of insects may result from entirely different mechanisms in different insects, or even in the same insect, under different conditions. For example, a large number of interference relations have been found in the literature (Hassell, 1976). Several different mechanisms have now been invoked to explain them. Hassell (1971b) suggested that they were due to time wasting after encounters between adults. Rogers and Hassell (1974) have shown that this, and also time wasting after the avoidence of super- parasitism can produce such a relationship, as has Beddington (1975) working from slightly different assumptions. An apparent decline in searching efficiency can be produced by an immunity reaction within the host (Rogers, 1970) and possibly by behavioural defence mechanisms (suggested here). Free et al (1977) have shown that this could also be produced by presenting an unevenly distributed host population to a parasitoid or predator which employs non-random search, i.e. aggregates on areas of high host density. Another interesting feature of these results is the occurrence of both positive and curved competitive relationships, as have been found by other ecologists such as Neill (1974) and Moseyeva (1974). Things are not always as simple as they seem, or we should like them to be. On the other hand it is encouraging to find that, as expected (see chapter five), interspecific interactions may have a very similar form to intraspecific interactions. As shown in chapter seven, an encounter with a Brecon female has a very similar effect upon a Nemeritis wasp as an 270 encounter with another Neneritis, but is apparently somewhat less severe. In the aphid parasitoids studied, it is interesting to note that both an increasing density of conspecifics and of wasps of another species, have the effect of decreasing the proportion of female offspring left by D aeretiella wasps. There is plenty of scope for more studies of this kind, preferably broader and in greater depth. These are more interesting if they can be run side by side with observations from the field. Laboratory experiments can give a fuller picture of the capacity of an animal. They can test its reaction to a whole range of possible conditions, instead of merely taking the samples of conditions that are available from the field. They can test the effect of one factor at a time, holding all other conditions constant, which is very difficult to do in the field. It is, of course, vitally important to consider how many of these conditions the animal is actually subject to in the field. These natural conditions are important for they determine evolutionary pressures and direct population change. However, it may be that the mechanism of a process cannot be clearly understood without subjecting an animal to a wider range of conditions than are found in the field over a reasonable length of time. Although in the experiments reported here the behaviour of the insects does not contribute significantly to the dynamics of competition, this does not mean that such observations are not relevant. It may be useful to know that in a given situation, behavioural components are of little or no importance. As has already been shown, (Hassell, 1971b), under slightly different conditions (a clumped host population), behavioural reactions may contribute towards an interference relationship between conspecifics (see chapter four). Ideally, experiments such as these should 271 be conducted over at least the whole range of conditions, that the insects would be subject to, naturally. In the results reported here, as in many studies of biological systems, there is much variation between replicates and between individuals. To the statistician, this is unfortunate; how does he know what is really going on? Variation may mask a real phenomenot, so there is a tendency to standardize the individuals used more and more; using insects of the same genetic strain, of the same sex, of the same age (correct to days or even hours), discarding those that behave "abnormally" (e.g. not searching at all). While this produces convincing graphs with replicate points tightly clustered together, they are, of necessity, not very representative of the behaviour of the species as a whole, but only of the subsection comprised, for example, of 24 hour old females between 5 and 6 cm long. The obvious answer to this is to repeat the experiment until all possible ages, sexes and sizes have been covered. This is impractical in cases where time is limited (as it usually is). It is probably more satisfactory to keep standardization to a minimum, and increase the number of replicates, where possible. Variation is not "bad" or "wrong", as is often implied by the more statistically minded. It is an essential feature of nearly all biolocial systems. Without it, there would be no evolution; and hence no ecological communities. 272 ACKNOWLEDGEMENTS I would like to thank my supervisor, Dr. M.P. Hassell, for his help, in particular for his meticulous attention to the text of this thesis. I am very grateful to the Natural Environment Research Council, for a research studentship which financed this work; and to Professor T.R.E. Southwood for providing laboratory space and other facilities at Imperial College Field Station. A host of people provided me with insects to start cultures, including Dr. R.D. Dransfield, Dr. T.H. Chua, Dr. N. Scopes, Dr. D.S. Grosch, and Mr. D. Batt. Frank Wright took the photographs for plates 2 and 5, and those on which figures 1.5 and 1.6 are based, and printed copies of the other photographs, when he had a great deal of other work to cope with. I am very grateful to Carole Collins, who typed this long and wearisome thesis so beautifully. Lastly, but by far the most important, I would like to thank my family and friends for their constant support and encouragement. 273 REFERENCES ABLES, J.R. & SHEPARD, M. (1974). Responses and competition of the parasitoids Spalangia endius and Muscidifurax raptor (Hymenoptera: Pteromalidae) at different densities of house fly pupae. Can. Ent. 106(8): 825-830. ARMAD, T. (1936). The influence of ecological factors on the Med. flour moth Ephestia and its parasite Nemeritis canescens. J. Anim. Ecol. 5: 67-93. ALLAN, J.D. (1974). Balancing predation and competition in cladocerans. Ecology 55: 622-629. ANDREWARTHA, H.G. & BIRCH, L.C. (1954). The Distribution and Abundance of Animals. University of Chicago Press, Chicago. ARTHUR, A.P.; STAINER, J.E.R. & TURNBULL, A.L. (1964). The interaction between Orgilus obscurantor (Nees) ['Hymen. Bracon. J and Temelucha interruptor (Grav.) Hymen. Ichneumonidae]; parasites of the Pine Shoot Moth, Rhyacionia buoliana (Schiff) (Lepidopt: Olethreutidae). Can. Ent. 96: 1030-4. ASKEW, R.R. (1971). Parasitic Insects. Heinemann Educational Books, London. ATWAL, A.S.; CHOUDHARY, J.P. & RAMZAN, M. (1969). Some preliminary studies in India on bionomics and rate of parasitism of Diaeretiella rapae Curtis (Braconidae: Hymenoptera) a parasite of aphids. Rev. Appl. Ent. 58: 290. 274 AYALA, F.J. (1971). Competition between species: frequency dependence. Science 171: 820-824. AYALA, F.J.; GILPIN, M.E. & EHREINFELD, J.G. (1973). Competition between species: theoretical models and experimental tests. Theoret. Pop. Biol. 4: 331-356. BACH, C.; HAZLETT, B. & RITTSCHOF, D. (1976). Effects of interspecific competition on fitness of the hermit crab, Clibanarius tricolor. Ecology 57(3): 579-586. BAKER, A.E.M. (1974). Interspecific aggressive behaviour of pocket gophers Thomomys bottae and T. talpoides (Geomyidae: Rodatia). Ecology 55: 671-673. BAKER, J.P. (1977). Assessment of the potential for and the development of organophosphorus resistance in field populations of Myzus persicae. Ann. AppZ. Biol. 86(1): 1-10. BAKKER, K.; EIJSACHERS, H.J.P.; VAN LENTEREN, J.C. & MEELIS, E. (1972). Some models describing the distribution of eggs of the parasite Pseudocoila bochei (Hym., Cynip.) over its hosts, larvae of Drosophila melanogaster. Oecologia 10: 29-57. BARTLETT, B.R. & BALL, J.C. (1964). The developmental biologies of 2 encyrtid parasites of Coccus hesperidum and their intrinsic competition. Ann. Ent. Soc. Amer: 57: 496-503. BATRA, H.N. & KUMAR, K. (1962). Occurrence of green peach aphid, Myzus persicae (Sulz.) as a serious pest of pansey flower and its control. Indian J. Entomol. 23: 149-51. 275 BEARD, R.I. (1972). Effectiveness of paralysing venom and its relation to host discrimination by Braconid wasps. Ann. Ent. Soc. Am. 65: 90-93. BEARD, R.L. (1952). The toxicology of Habrobracon venom: a study of a natural insecticide. Conn. Agric. Exp. Sta. Bull. 562 BEAVER, R.A. (1974). Intraspecific competition among bark beetle larvae (Coleoptera: Scolytidae) . J. Anim. Ecol. 43(2): 455-468. BEDDINGTON, J.R. (1975). Mutual interference between parasites or predators and its effect on searching efficiency. J. Anim. Ecol. 44: 331-40. BEESON, C.F.C. & CHATTERJEE, S.N. (1935). Biology of the Braconidae (Hymenoptera). Ind. Forest Records (NS) Entomology 1: 105-38. BELING, I. (1932). Zur Biologie von Nemeritis canescens Gray. Z. Angew. Ent. 19: 223-49. BENDER, J.C. (1943). Anatomy and histology of female reproductive organs of Habrobracon juglandis Ashmead (Hym. Braconidae) . Ann. ent. Soc. Am. 86: 537-45. BENKE, A.C. & BENKE, S.S. (1975). Comparative dynamics and life histories of coexisting dragonfly populations. Ecology 56(2): 302-317. BENSON, J.F. (1972). Laboratory studies of insect parasite behaviour in relation to population models. D.Phil. Thesis. BENSON, J.F. (1973a). The biology of Lepidoptera infesting stored products, with special reference to population dynamics. Biol. Rev. 48: 1-26. 276 BENSON, J.F. (1973b). Intraspecific competition in the population dynamics of Bracon hebetor Say (Hym. Brac.). J. Anim. Ecol. 42: 105-24. BERRY, R.E. & SIMPSON, R.G. (1967). Flight activities of the green peach aphid Myzus persicae, a natural vector of potato leafroll virus in Colorado. Colorado Agr. Expt.'Sta. Tech. Bull. 92. BIRCH, L.C. (1953). Experimental background to study of distribution and abundance of insects. III Relation between innate capacity for increase and survival of different species of beetle living together on the same food. Evolution 7: 136-144. BOVBJERG, R.V. (1970). Ecological isolation and competitive exclusion in 2 crayfish (Orconectes virilis and 0. inmunis). Ecology 51: 225-236. BRADY, U.E. & SMITHWICK, E.B. (1968). Production and release of sex attractant by the female Indian meal moth, Plodia interpunctella. Ann. ent. Soc. Am. 61: 1260-5. BRANCH, G.M. (1976). Interspecific competition experienced by South Africa Patella spp. J. Anim. Ecol. 45: 507-530. BROADBENT, L. (1949). The grouping and overwintering of Myzus persicae on Prunus spp. Ann. Appl. Biol. 36: 334-40. BROADHEAD, E. & CHEKE, R.A. (1975). Host spatial pattern, parasitoid interference and the modelling of the dynamics of Alaptus fusculus (Hymenoptera: Mymaridae), a parasitoid of two Mesopsocus species (Psocoptera). J. Anim. Ecol. 44: 767-789. 277 BROADHEAD, E. & WAPSHERE, A.J. (1966). Mesopsocus populations on larch in England - the distribution and dynamics of two closely-related coexisting species of Psocoptera sharing the same food resource. Ecol. Monogr. 36: 327-88. BROCKELMAN, W.Y. (1975). Competition, the fitness of offspring and optimal clutch size. Am. Nat. 109: 677-699. BROWN, J.H. (1971). Mechanisms of competitive exclusion between two species of chipmunk. Ecology 52: 305-311. BROWN, J.H. & DAVIDSON, D.W. (1977). Competition between seed eating rodents and ants in desert ecosystems. Science 196(4292): 880-882. BROWN, J.H. & LIEBERMAN, G.A. (1973). Resource utilization and coexistence of seed-eating desert rodents in sand dune habitats. Ecology 54: 788-97. BROWN, W.L. & WILSON, E.O. (1956). Character displacement. Syst. Zool. 5: 49-64. BUDNIK, M. & BRNCIC, D. (1974). Preadult competition between Drosophila pavani and D. Melanogaster, D. simians and U. willistoni. Ecology 55: 657-661. BULMER, M.G. (1974). Density dependent selection and character displacement. Am. Nat. 108: 45-58. BUNZLI, G.H. & BUTTIKER, W. (1959). Host plants of Myaua persicae Sulz. with a list of aphids of common occurrence in the tobacco growing districts of S. Rhodesia. J. Entomol. Soc. S. Africa 22: 35-50. 278 CADWALLADER, P.L. (1975). A laboratory study of interactive segregation between two New Zealand stream dwelling fish. J. Anim. Ecol. 44: 865-875. CASE, T.J. & GILPIN, M.E. (1974). Interference competition and niche theory. Proc. Natn. Acad. Sei. U.S.A. 71: 3073-3077. CASWELL, H. (1978). Predator-mediated coexistence: a non equilibrium model. Am. Nat. 112(983): 127-154. CHARLESWORTH, B. (1971). Selection in density-regulated populations. Ecology 52: 469-474. CHAWLA, S.S. & SUBBA RAO, B.R. (1963). Further studies on the Bracon hebetor - brevicornis complex (Hymenoptera, Braconidae) by paper chromatography. Beitr. Ent. 15: 83-86. CHUA, T.C. (1975). Population studies on the cabbage aphid, Brevicoryne brassicae (L.), and its parasites, with special reference to synchronization. Unpublished Ph.D thesis, University of London, Imperial College. CLARK, A.M. & SMITH, R.E. (1967). Egg production and adult life-span in two species of Bracon (Hymenoptera, Braconidae). Ann. ent. Soc. Am. 60: 903-05. CODY, M.L. (1973a). Coexistence, coevolution and convergent evolution in seabird communities. Ecology 54: 31-44. COLWELL, R.K. & FATUYMA, D.J. (1971). On the measurement of niche breadth and overlap. Ecology 52: 567-576. 279 COMINS, H.N. & HASSELL, M.P. (1976). Predation in multi-prey communities. J. T heoret. Bio 1. 62: 93-114. CONNELL, J.H. (1970). On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. In: Dynamics of Populations. See Royama (1970) (Predator-Prey). COOK, R.M. & HUBBARD, S.F. (1977). Adaptive strategies in insect parasites. J. An2m. Ecol. 46: 115-25. CROFT, B.A. & HOYING, S.A. (1977). Competitive displacement of Panonychus ulmi (Acarina: Tetranydidae) by Aculus schlectendali._(Acarinā: Eriophyidae) in apple orchards. Can. Ent. 109(8): 1025-1034. CROMBIE, A.C. (1945). On competition between different species of graminiverous insects. Proc. Royal Soc. Lond. Ser. B 132: 362-395. CROWELL, K.L. & PIMM, S.L. (1976). Competition and niche shifts of mice introduced onto small islands. Oikos 27: 251-258. DAVIDSON, D.W. (1977). Species diversity and community organisation in desert seed-eating ants. Ecology 58(4): 711-724. DAVIDSON, D.W. (1978). Size variability in the worker caste of a social insect (Veromessor pergandei Mayr) as a function of the competitive environment. Ann. Nat. 112(985): 523-532. DAYTON, R.K. (1971). Competition, disturbance and community organization: the provision and subsequent utilization of space in a rocky intertidal community. Ecol. Monogr. 41: 351-89. 280 DEBACH, P. (1964) Ed. Biological Control of Insect Pests and Weeds. London, Chapman and Hall. DEBACH, P. & SUNDBY, R.A. (1963). Competitive displacement between ecological homologues. Hilgardia 34(5): 105-166. DEBENEDICTIS, P.A. (1974). Interspecific competition between tadpoles of Rana pipiens and Rana sylvatica: and experimental field study. Ecological Monographs 44(2): 129-151. DEBENEDICTIS, P.A. (1977). The meaning and measurement of frequency- dependent competition. Ecology 58(1): 158-166. DODSON, S.I. (1974). Zooplankton competition and predation: an experi- mental test of the size efficiency hypothesis. Ecology 55: 605-13. DOWNING, R.S. & MOILLIET, T.K. (1974). Comparisons between an exotic and an endemic phytoseiid mite predator (Acarina: Phytoseiid ae) in British Columbia. Can. Entomol. 106(7): 773-776. DUNN, J.A. (1949). The parasites and predators of potato aphids. Bull. Entomol. Res. 40: 97-122. DUNN, J.A. & KEMPTON, D.P.H. (1977). The development on potatoes of a glasshouse strain of Myzus persicae resistant to organophosphorus insecticides. Annals AppZ. Biol. 85(2): 175-180. EBERSOLE, J.P. (1977). The adaptive significance of interspecific territoriality in the reef fish Eupomacentrus leucostictus. Ecology 58(4): 914-920. 281 EDWARDS, C.A. & HEATH, G.W. (1964). Principles of Agricultural Entomology London, Chapman & Hall. ELTON, C. (1927). Animal Ecology. Sidgwick and Jackson, London. ELTON, C. (1958). The Ecology of Invasions by Animals and Plants. Methuen, London. FISCHER, R.C. & GANESALINGAM, V.A. (1970). Host hemolymph changes com- position after insect parasitoid attacks. Nature 227: 191-192. FISHER, R.C. (1961a). A study in insect multiparasitism: I Host selection and oviposition. J. Exp. Biol. 38: 267-76. FISHER, R.C. (1961b). A study in insect mutliparasitism: II The mechanism and control of competition for possession of the host. J. Exp. Biol. 38: 605-628. FISHER, R.C. (1962). The effect of multiparasitism on populations of two parasites and their host. Ecology 43(2): 314-6. FISHER, R.C. (1963). Oxygen requirements and the physiological suppression of supernumerary insect parasitoids. J. exp. Biol. 40: 531-540. FISKE, W.F. (1910). Superparasitism: An important factor in the natural control of insects. J. Econ. Ent. 3: 88-97. FLANDERS, S.E. (1965). On the sexuality and sex ratios of Hymenopterous populations. Am. Nat. 99: 489-94. FLANDERS, S.E. (1965). Competition and cooperation among parasitic Hymenoptera related to biological control. Can. Ent. 97(1-6): 409-422. 282 FLANDERS, S.E. (1966). The circumstances of species replacement among parasitic Hymenoptera. Canad. Ent. 98(10): 1009-1024. FORCE, D.C. (1970). Competition among four Hymenopterous parasites of an endemic insect host. Ann. Entomol. Soc. Am. 63: 1675. FORCE, D.C. (1972). r and K strategists in endemic host-parasitoid communities. BUZZ. Ent. Soc. Am. 18: 135-137. FORCE, D.C. (1974). Ecology of insect host-parasitoid communities. Science 184: 624-632. FRASER, D.F. (1976a). Coexistence of salamanders in the genus Plethodon: A variation of the Santa Rosalia theme. Ecology 57(2): 238-251. FRASER, D.F. (1976b). Empirical evaluation of the hypothesis of food competition in salamanders of genus PZethodon. Ecology 57(3): 459-471. FREE, C.A.; BEDDINGTON, J.R. & LAWTON, J.H. (1977). On the inadequacy of simple models of mutual interference for parasitism and predation. J. Anim. Ecol. 46(2): 543-554. FRETWELL, S. (1978). Competition for discrete versus continuous resources: tests for predictions from the MacArthur-Levins models. Am. Nat. 112(983): 73-81. FRYER, G. (1959). The trophic interrelationships and ecology of some littoral communities of Lake Nyasa with especial reference to the fishes and a discussion of the evolution of rock frequenting cichlidae. Proc. Zool. Soc. Lond. 132: 153-280. 283 FUJII, K. (1965). A statistical model of the competition curve. Res. on Pop. Ecol. 7: 118-125. FUJII, K. (1967). Studies on interspecies competition between the azuki bean weevil Callosobruchus chimensis and the southern cowpea weevil C. maculatus. II Competition under different environmental conditions. Res. PopuZ. Ecol. IX: 192-200. FUJII, K. (1968).. Studies on interspecies competition between the azuki bean weevil and the southern cowpea weevil. III Some characteristics of strains of two species. Res. PopuZ. Ecol. 10: 87-98. FUTUYMA, D.J. (1970). Variation in genetic response to interspecific competition in laboratory populations of Drosophila. Am. Nat. 104: 239-252. GADGIL, M. & SOLBRIG, 0.T. (1972). The concept of r- and K-selection: Evidence from wild flowers and some theoretical considerations. Am. Nat. 106(947): 14-31. GALLEY, D.J. (1974). Relative feeding rates of Brevicoryne brassicae and Myzus persicae on Brussels sprout plants in relation to their susceptibility to systemic, insecticides. Ann. AppZ. Biol. 76(2): 171-178. GAUSE, G.F. (1932). Experimental studies on the struggle for existence. I Mixed populations of two species of yeast. J. exp. Biol. 9: 389-402. GAUSE, G.F. (1934). The Struggle for Existence. Hafner, New York. (Reprinted 1964). 284 GAUSE, G.F. (1935). Experimental demonstration of Volterra's periodic oscillation in the numbers of animals. J. exp. Biol. 12: 44-48. GEORGE, K.S. (1957). Preliminary investigations on the biology and ecology of the parasites and predators of Brevicoryne brassicae. Bull. Ent. Res. 48: 619-629. GERLING, D. (1969). The parasites of Spodoptera littoralis Bois. (Lepidoptera, Noctuidae) eggs and larvae in Israel. Israel J. Ent. IV: 73-81. GILBERT, L.E. & RAVEN, P.H. (1975) Ed. Coevolution of Animals and Plants. University of Texas Press, Austin and London. GILL, D.E. (1974). Intrinsic rate of increase, saturation density and competitive ability. II The evolution of competitive ability. Am. Nat. 108(959): 103-116. GILPIN, M.E. (1974). Intraspecific competition between Drosophila larvae in serial trasfer systems. Ecology 55: 1154-1159. GRANT, P.R. (1972). Interspecific competition among rodents. Ann. Rev. Ecol. & Syst. 3: 79-106. GREANY, P.D. & OATMAN, E.R. (1972). Analysis of host discrimination in the parasite Orgilus Zepidus (Hymenoptera: Braconidae). Ann. ent. Soc. Am. 65: 377-383. HAFEZ, M. (1961). Seasonal fluctuations of population density of the cabbage aphid, Brevicoryne brassicae (L.), in the Netherlands, and the role of its parasite, Aphidius (Diaeretielia) rapae (Curtis). Tijdschr. Plantenziekten 67: 445-548. 285 HAGEN, K.S. & VAN DEN BOSCH, R. (1968). Impact of pathogens, parasites and predators on aphids. Ann. Rev. Ent. 13: 325-84. HANSKI, I. (1978). Some comments on the measurement of niche metrics. Ecology 59(1): 168-174. HARDIN, G. (1960). The competitive exclusion principle. Science 131(3409): 1292-8. HARPER, J.L. (1969). The role of predation in vegetational diversity. In: Diversity and Stability in Ecological Systems. Brookhaven Symp. in Biol. No. 22. HASSELL, M.P. (1971a). Mutual interference between searching insect parasites. J. Anim. Ecol. 40: 473-86. HASSELL, M.P. (1971b). Parasite behaviour as a factor contributing to the stability of insect host-parasite interactions. In: Dynamics of Populations. (Ed. by P.J. den Boer & G.R. Gradwell) pp. 366-79. Proc. Adv. Study Inst. (Oosterbeek, 1970). HASSELL, M.P. (1975). Density-dependence in single species populations. J. Anim. Ecol. 44: 283-295. HASSELL, M.P. (1976). Arthropod predator-prey systems. In: Theoretical Ecology Principles and Applications. Ed.. by R.M. May. Chapter Five, pp. 71-93. Oxford, Blackwell Scientific Publications. HASSELL, M.P. & COMINS, H.N. (1976). Discrete time models for two species competition. Theor. Pop. Biol. 9(2): 202-221. HASSELL, M.P.; LAWTON, J.H. & MAY, R.M. (1976). Patterns of dynamical behaviour in single species populations. J. Anim. Ecol. 45: 471-486. 286 HASSELL, M.P. & ROGERS, D.J. (1972). Insect parasite responses in the development of population models. J. Anim. Ecol. 41: 661-676. HASSELL, M.P. & MAY, R.M. (1973). Stability in insect host-parasite models. J. Anim. Ecol. 42: 693-726. HASSELL, M.P. & MAY, R.M. (1974). Aggregation of predators and insect parasites and its effect on stability. J. Anim. Ecol. 43: 567-594. HEATHCOTE, G.D.; DUNNING, R.A. & WOLFE, M.D. (1965). Aphids on sugar beet and some weeds in England, and notes on weeds as a source of beet viruses. Plant Pathol. 14(1-10). HEATHCOTE, G.D. & WARD, J. (1958). The preference shown by Myzus persicae (Sulz.) for Brassica Plants sprayed with wetting agents. Bull. Ent. Res. 49: 235-37. HEATWOLE, H. & DAVIS, D.M. (1965). Ecology of three sympatric species of parasitic insects of genus Megarhyssa (Hymenoptera: Ichneumonidae). Ecology 46(1-2): 140-50. HEBERT, P.D.N. (1977). Niche overlap among species in the Daphnia carinata complex. J.A.E. 46: 399-410. HEIE, 0. & PETERSEN, B. (1961). Investigations on Myzus persicae Sulz., Aphis fabxe Scop. and virus yellows of beet (Beta virus 4 in Denmark. Condensed reports from the virus committee. (Danish Acad. Tech. Sci. Copenhagen). HEINRICH, B. (1976). Resource partitioning among some eusocial insects: Bumblebees. Ecology 57(5): 874-889. 287 HELLER, H.C. (1971). Attitudinal zonation of chipmunks (Entanias): interspecific aggression. Ecology 52: 312-319. HERAKLY, F.A. (1968). Biological studies on the cabbage web-worm, Hellula undalis Fabr. (Lepidoptera: Crambidae - Pyraustinae). Bunt. Soc. ent. Egypte 52: 191-211. HEYER, W.R. (1974). Niche measurements of frog larvae from a seasonal tropical location in Thailand. Ecology 55: 651-656. HOKYO, N.; SHIGA, M. & NAKASUJI, F. (1966). The effect of intra and interspecific conditioning of host eggs on the ovipositional behaviour of two scelionid egg parasites. Jap. J. Ecology 16: 67-71. HONDA, H.; OSHIMA, K. & VAMAMOTO, I. (1976). Oviposition marker of azuki bean weevil, Callosobruchus dimensis L. Proceedings of the joint United States - Japan Seminar on Stored Product Insects. pp. 116-128. HUEY, R.B. & PIANKA, E.R. (1977). Patterns of niche overlap among broadly sympatric versus narrowly sympatric Kalahari lizards (Scincidae: Mabuya). Ecology 58(1): 119-128. HUFFAKER, C.B. & KEANETT, C.E. (1966). Studies of two parasites of olive scale, Parlatoria oleae (Colvee). IV Biological control of Parlatoria oleae (Colvee) through the compensatory action of two introduced parasites. Hilgardia 37: 283-335. HUFFAKER, C.B. & MESSENGER, P.S. (1964). The concept and significance of natural control. In: Biological Oontrol of Insect Pests and Weeds. Ed. by P. DeBach. Chapman Hall, London. Chap. 4. pp. 74-117. 288 HURLBERT, S.H. (1978). The measurement of niche overlap and some relatives. Ecology 59(1): 67-78. HUTCHINSON, G.E. (1958). Concluding remarks. Cold Spring Harbor Symp. Quant. BioZ. 22: 415-27. ISTOCK, C.A. (1967). Transient competitive displacement in natural pop- ulations of whirligig beetles. Ecology 48: 929-937. ISTOCK, C.A. (1977). Logistic interaction of natural populations of two species of waterboatmen. Am. Nat. 111(978): 279-287. JACOBS, J. (1977). Coexistence of similar zooplankton species by differential adaptation to reproduction and escape in an environment with fluctuating food and enemy densities. II Field data analysis of Daphnia. Oecologia 30: 313-329. JAEGER, R.G. (1972). Food as a limited resource in competition between two species of terrestrial salamanders. Ecology 53: 535-546. JANZEN, D.H. (1970). Herbivores and the number of tree species in tropical forests. Am. Nat. 104: 501-528. JOHNSON, L.K. & HUBBELL, S.P. (1974). Aggression and competition among stingless bees: Field studies. Ecology 55: 120-127. JOHNSON, L.K. & HUBBELL, S.P. (1975). Contrasting foraging strategies and coexistence of two bee species on a single resource. Ecology 56(6): 1398. JOST, J.L.; DRAKE, J.F.; FREDERICKSON, A.G. & TSUCHIYA, H.M. (1973). Interactions of Tetrahymena pyriformis, Escherischia coli and Azotobactor vinelandi and glucose minimal medium. J. Bact. 113: 834-840. 289 JOUBERT, P.C. (1969). Reproduction in the Pyralidae (Lepidoptera). II Post-embryonic development and structure of the female reproductive systems of Cadra cautella (Walker) and Plodia interpunctella (Hubner). Phytophylactica 1: 209-16. KANESHIRO, K.Y.; CARSON, H.L.; CLAYTON, F.E. & HEED, W.B. (1973). Niche separation in a pair of homosequential Drosophila spp. from the island of Hawaii. Am. Nat. 107(958): 766-774. KENNEDY, J.S. & STROYAN, H.L.G. (1959). Biology of aphids. Ann. Rev. Ent. 4: 139-60. KERFOOT, W.C. (1977). Competition in cladoceran communities: The cost of evolving defenses against copepod predation. Ecology 58(2): 303-313. KING, C.E. & ANDERSON, W.W. (1971). Age specific selection. II The interaction between r and K during population growth. Am. Nat. 105: 137-156. KLOMP, H. (1962). The influence of climate and weather on the mean density level, the fluctuations and the regulation of animal pop- ulations. Arch. NeerZ. Zool. 15: 68-109. KREBS, C.J. (1972). Ecology: The Experimental Analysis of Distribution and Abundance. Harper and Row, New York. KUENZLER, E.J. (1958). Niche relations of three species of Lycosid spiders. Ecol. 39(3): 494-500. KUWARA, I. (1919). Observations on injurious insects of 1918. Rev. AppZ. Ent. 7(A): 100. 290 LACK, D. (1954). The Natural Regulation of Animal Numbers. Oxford University Press, New York. LACK, D. (1971). Ecological Isolation in Birds. Blackwell, Oxford. LEVIN, B.R.; STEWART, F.M. & LINCHAO, (1977). Resource limited growth, competition, and predation: a model and experimental studies with bacteria and bacteriophage. Am. Nat. 111(977): 3-24. LEVIN, D.A. & ANDERSON, W.W. (1970). Competition for pollinators between simultaneously flowering species. Am. Nat. 104: 455-467. LEVIN, S.A. (1970). Community equilibria and stability and an extension of the competitive exclusion principle. Am. Nat. 104: 413. LEVINS, R & CULVER, D. (1971). Regional coexistence of species and competition between rare species. Proc. Nat. Acad. Sci. 68: 1246-1248. LISTER, B.C. (1976). The nature of niche expansion in West Indian AnoZis lizards. I Ecological consequences of reduced competition. Evolution 30(4): 659-676. LLOYD, D.C. (1940). Host selection by hymenopterous parasites of the moth Plutella maculipennis Curtis. Proc. Roy. Soc. B 128: 451. LOCK, M.A. & REYNOLDSON, T.B. (1976). The role of interspecific competition in the distribution of two stream-dwelling triclads, Crenobia alpina (Dana) and Polycelis felina (Dalyall) in North Wales. J. Anim. Ecol. 45: 581-592. LOWE, H.J.B. & RUSSELL, G. (1974). Probing by aphids in the leaf tissues of resistant and susceptible sugar beet. Ent. exp. & App Z. 17: 468-476. 291 LOTKA, A.J. (1925). Elements of Physical Biology. Baltimore, Williams and Wilkins. LUBCHENKO, J. (1978). Plant species diversity in a marine intertidal community: importance of herbivore food preference and algal competitive abilities. Am. Nat. 112(983): 23-39. MACARTHUR, R.H. (1965). Patterns of species diversity. BioZ. Rev. 40: 510-533. MACARTHUR, R.H. & LEVINS, R. (1964). Competition, habitat selection and character displacement in a patchy environment. Proc. Nat. Acad. Sci . 51: 1207-1210. MACARTHUR, R.H. & LEVINS, R. (1967). The limiting similarity, convergence and divergence of coexisting species. Am. Nat. 101: 377-385. MACARTHUR, R.H. & PIANKA, E. (1966). On optimal use of a patchy environ- ment. Am. Nat. 100: 603-609. MACARTHUR, R.H. & WILSON, E.O. (1967). The Theory of Island Biogeography. Princeton University Press, Princeton, New Jersey. MACKAUER, M. (1969). Insect parasites of the green peach aphid Myzus persicae Sulz. and their control potential. Rev. AppZ. Ent. 57: 580. Entomophaga 132: pp. 91-106. MATTHEWS, R.W. (1974). Biology of Braconidae. Ann. Rev. Ent. 19: 15-32. MAY, R.M. (1973). Stability in randomly fluctuating versus deterministic environments. Am. Nat. 107(957): 621-650. 292 MAY, R.M. (1974). Biological populations with non-overlapping generations: stable points, stable cycles and chaos. Science N.Y. 186: 645-7. MAY, R.M. (1976). Models for single population. In: Theoretical Ecology: Principles and Applications. Ed. by R.M. May. Blackwell Scientific Populations, Oxford. MAY, R.M. (1976a) Ed. Theoretical Ecology, Principles and Applications. Blackwell Scientific Publications. MAY, R.M. (1976b). Models for two interacting populations. In: Theoretical Ecology, Principles and Applications. Ed. by R.M. May. Blackwell Scientific Populations, Oxford. MAY, R.M.; CONWAY, G.R.; HASSELL, M.P. & SOUTHWOOD, T.R.E. (1974). Time delays, density dependence, and single species oscillations. J. Anim. Ecol. 43: 747-70. MAY, R.M. & OSTER, G.F. (1976). Bifurcations and dynamic complexity in simple ecological models. Am. Nat. 110: 573-599. MCCLURE, M.S. & PRICE, P.W. (1976). Ecotope characteristics of coexisting Erythroneura leafhoppers (Homoptera: Cicadellidae) on sycamore. Ecology 57(5): 928. MCGAUGHEY, W.H. (1976). Bacillus thuringiensis for controlling three species of moths in stored grain. Can. alt. 108: 105-112. MCKAYE, K.R. (1977). Competition for breeding sites between the cichlid fishes of Lake Jiloa, Nicaragua. Ecology 58(2): 291-302. 293 MEIER, W. (1966). Importance of aphidophagous insects in the regulation of potato aphid populations. In: Ecology of Aphidophagous Insects. Ed. I. Hodek. MENGE, J.L. & MENGE, B.A. (1974). Role of resource allocation, aggression and spatial heterogeneity in coexistence of two competing intertidal starfish. Ecological Monographs 44(2): 189-209. MEREDITH, D.H. (1977). Interspecific agonism in two parapatric species of chipmunks (Eutamias). Ecology"58(2): 423-430. MERRELL, D.J. (1951). Interspecific competition between Drosophila funebris and Drosophila melanogaster. Am. Nat. 85: 159-169. MICHELAKIS, S.E. (1973). A study of the laboratory interaction between Coccinella septempunctata L. larvae and its prey Myzus persicae Sulz. M.Sc. DIC, University of London. MILLER, J.C. (1977). Ecological relationships among parasites of Spodoptera praefica. Environ. Ent. 6: 581-585. MILLER, R.S. (1964). Larval competition in Drosophila melanogaster and D. simulans. Ecology 451..132-48. MILLER, R.S. (1967). Pattern and process in competition. Adv. EcoZ. Res. 4: 1-69. MILLER, R.S. (1969). Competition and species diversity. In: Diversity and Stability in Ecological Systems. Brookhaven Symp. in Biology 22: 63-70. 294 MILLS, J.A. & MARK, A.F. (1977). Food preferences of takahe in Fiordland National Park, New Zealand, and the effects of competition from introduced red deer. J. Anim. Ecol. 46(3): 939-974. MILNE, A. (1958). Theories of natural control of insect populations. Cold Spring Hab. Symp. Quant. Biol. 22: 253-271. MITCHELL, R. (1975). The evolution of oviposition tactics in the bean weevil, Callosobruchus maculatus (F.). Ecology 56: 696-702. MOISEYEVA, T.S. (1974). On the weakening of the defense reaction in larval Lepidoptera when simultaneously parasitized by two species of parasite. Zool. Zh. 53(1): 51-57. MOOKHEJEE, P.B.; BOSE, B.N. & SINGH, S. (1969). Some observations on the damage potential of the almond moth, Cadra cautella (Walker) in eight different stored grains. Indian J. Ent. 31: 1-6. MORERE, J.L. (1970). Developpement et reproduction de Plodia inter- punctella (Hbn) (Lepidoptere: Pyralidae) sur un regime meridique (= semi-synthetique) ā base d'acides amines. C.r. hebd. Sēanc. Acad. Sci., Paris 271: 1019-22. MORRRE, J.L. & BERRE, J-R. (1967). etude au laboratoire du développement de la Pyrale Plodia interpunctella. C.r. hebd. Sēanc. Acad. Sci., Paris 272: 133-6. MtJDD, A. & CORBET, S.E. (1973). Mandibular gland secretion of larvae of the stored product pests Anagasta kuehnielZa, Ephestia cautella, PZodia interpunctella and Ephestia elutella. Ent. Exp. Appl. 16: 291-292. 295 MUHLENBERG, M.; LEIPOLD, D.; MADER, H.J. & STEINHAUER, B. (1977a). Island ecology of arthropods.. I Diversity, niches and resources on some Seychelles Islands. Oecologia 29: 117-134. MURR, L. (1930). Uber den Gerushsinn der Mehlmottenschhepfwespe Habrobracon juglandis Ahmead, zugleich ein Beitrag zum Orient- ierungsproblem. Z. vergi. Physiol. (Z. ftr. wiss. Biologie), 11: 210-70. MURRAY, B.G. & GILL, F.B. (1976). Behavioural interactions of blue- winged and golden-winged warblers. Wilson Bulletin 88(2): 231-254. MUSGRAVE, A.J. & MACKINNON, D.L. (1938). Infection of Plodia inter- punctella Hb (Lepidoptera: Phycitidae) with a Schizogregarine Maltesia dispora Naville. Proc. R. ent. Soc. Lond. (A) 13: 89.90. NEKASUJI, F.; HOKYO, N. & KIRITANI, K. (1966). Assessment of the potential efficiency of parasitism in two competitive scelionic parasites of Nezara viridula L. (Hemiptera: Pentatomidae). AppZ. Ent. Zool. 1: 113-19. NARAYANAN, E.S.; SUBBA RAO, B.R. & SHARMA, A.K. (1958). Studies on the Bracon hebetor - brevicornis complex (Hymenoptera, Braconidae). Proc. Ind. Acad. Sci (B) 48: 1-13. NATHANSON, M. (1975). The effect of resource limitation on competing populations of flour beetles, Tribolium spp. (Coleoptera, Tenebrionidae). BUZZ. Ent. Res. 65(1): 1-12. 296 NEILL, W.E. (1974). The community matrix and interdependence of the competition coefficients. Am. Nat. 108: 399-408. NEILL, W.E. (1975). Experimental studies of Microcrustacean competition, community composition, and efficiency of resource utilization. Ecology 56: 809-826. NICHOLSON, A.J. (1933). The balance of animal populations. J. Anim. Ecol. 2: 132-178. NICHOLSON, A.J. (1954a). Compensatory reaction of populations to stresses and their evolutionary significance. Aust. J. Zool. 2: 1-8. ORIANS, G.H. & COLLIER, G. (1963). Competition and blackbird social systems. Evolution 17: 449-459. OTTE, D. & JOERN, A. (1975). Insect territoriality and its evolution: population studies of desert grasshoppers on creosote bushes. J. Anim. Ecol. 44: 29-54. PARK, T. (1948). Experimental studies of interspecies competition. I. Competition between populations of the flour beetles Triboliu m confusum Duval and Tribolium castaneum Herbst. Ecol. Monogr. 18: 265-307. PARK, T.; MERTZ, D.B.; GRODZINSKI, W. & PRUS, T. (1965). Cannibalistic predation in populations of flour beetles. Physiol. Zool. 38: 289-321. PAINE, R.T. (1966). Food web complexity and species diversity. Am. Nat. 100: 65-75. 297 PARRISH, J.A.D. & BAZZAZ, F.A. (1976). Underground niche separation in successional plants. Ecology 57(6): 1281. PAYNE, N.M. (1934). The differential effect of environmental factors upon Microbracon hebetor (Hymenoptera: Braconidae) and its host Ephestia kuhniella (Lepidoptera: Pyralidae) II. Ecological Monographs 4: 1-46. PEARL, R. (1932). The influence of density of population upon egg - production in Drosophila melanagaster. J. exp. Zool. 63: 57-84. PETERS, R.H. (1976). Tautology in evolution and ecology. Am. Nat. 110: 1-12. PIANKA, E.R. (1970). On r and K selection. Am. Nat. 104: 592-97. PIANKA, E.R. (1976). Competition and niche theory. In: Theoretical Ecology, Principles and Applications. Ed. by R.M. May.. Blackwell Scientific Populations, Oxford. PLATT, W.J. & WEIS, I.M. (1977). Resource partitioning and competition within a guild of fugitive prairie plants. Ann. Nat. 111(979): 479-513. PODOLER, H. (1974). Analysis of life tables for a host and parasite (Plodia - Nemeritis) ecosystem. J. Anim. Ecol. 43(3): 653-670. POST, W.M. & RICHERT, S.E. (1977). Initial investigation into the structure of spider communities. I Competitive effects. J. Anim. Ecol. 46(3): 729-750. 298 PRESS, J.W.; FLAHERTY, B.R. & ARBOGAST, R.T. (1977). Interactions among Nemeritis canescens (Hymenoptera: Ichneumonidae), Bracon hebetor (Hymenoptera: Braconidae) and Ephestia cautella (Lepidoptera: Pyralidae). J. Kansas Ent. Soc. 50: 259-262. PRICE, P.W. (1970a). Characteristics permitting coexistence among parasitoids of a sawfly in Quebec. Ecology 51: 445-454. PRICE, P.W. (1970b). Biology of and host exploitation by Pteolophus- indistinctup (Provancher) (Hymenoptera: Ichneumonidae). Ann. Am. Ent. Soc. 63: 1502-1509. PRICE, P.W. (1970c). Trail odors: recognition by insects parasitic on cocoons. Science 170: 546-547. PRICE, P.W. (1971). Niche breadth and dominance of parasitic insects sharing the same host species. Ecology 52: 587-596. PRICE, P.W. (1973). Reproductive strategies in parasitoid wasps. Am. Nat. 107(957): 684-693. RABB, R.L. & BRADLEY, J.R. (1970). Marking host eggs by Telenomus sphingis. Ann. ent. Soc. Am. 63: 1053-1056. RANGANATH, H.A. & KRISHNAMARTHY, N.B. (1975). Reversal of dominance in the competition between Drosophila nasuta and D. neonasuta (Diptera: Drosophilidae). Experientia 31(3): 288-289. RATHCKE, B.J. (1976). Competition and coexistence within a guild of herbivorous insects. Ecology 57(1): 76-87. 299 READ, D.P.; FEENY, P.P. & ROOT, R.B. (1970). Habitat selection by the aphid parasite Diaeretiella rapae (Hymenoptera: Braconidae) and hyperparasite Charips brassicae (Hymenoptera: Cynipidae). Can. Ent. 102: 1567-1578. REDFIELD, J.A.; KREBS, C.J. & TAITT, M.J. (1977). Competition between Peromyscus maniculatus and Microtus townsendii in grasslands of coastal British Columbia. J. Anim. Ecol. 46: 607-616. REINERT, J.A. & KING, E.W. (1971). Action of Bracon hebetor Say as a parasite of Plodia interpunctella at controlled densities. Ann. Ent. Soc. Am. 64: 1335-1340. REYNOLDSON, T.B. (1975). Food overlap of lake dwelling triclads in the field. J. An-rn. Ecol. 44: 245-250. REYNOLDSON, T.B. & BELLAMY, L.S. (1970). The establishment of inter- specific competition in field populations, with an e.g. of competition in action between Polycelis nigra (Mull) and Polycelis tenuis (Turbellaria: Tricladida). In: Dynamics of Populations. Proc. Adv. Study Inst. on "Dynamics of numbers in populations". Ed. P.J. den Boer and G.R. Gradwell. RICHARDS, 0.W. & THOMSON, W.S. (1932). A contribution to the study of the genera Ephestia Gn. (including Strymax, Dyer) and Plodia Gn. (Lepidoptera, Phycitidae), with notes on parasites of the larvae. Trans. Roy. Ent. Soc. Land. 80: 169-250. RICHMOND, R.C.; GILPIN, M.E.; SALAS, S.P. & AYALA, F.J. (1975). A search for emergent competitive phenomena: The dynamics of multispecies Drosophila systems. Ecology 56(3): 709-714. 300 ROBERTSON, D.R.; SWEATMAN, H.P.A.; FLETCHER, E.A. & CLELAND, M.G. (1976). Schooling as a mechanism for circumventing the territoriality of competitors. Ecology 57(6): 1208-1220. ROGERS, D.J. (1970). Aspects of host parasite interaction in laboratory populations of insect. Unpublished D.Phil. Thesis, University of Oxford. ROGERS, D.J. (1972a). Random search and insect population models. J. Anim. Ecol. 41: 369-383. ROGERS, D.J. (1972b). The ichneumon wasp Venturia canescens: oviposition and avoidance of superparasitism. Ent. exp. & appl. 15: 190-4. ROGERS, D.J. (1975). A model for avoidance of superparasitism by solitary insect parasitoids. J. Anim. Ecol. 44: 623-38. ROGERS, D.J. & HASSELL, M.P. (1974). General models for insect parasite and predator searching behaviour: Interference. J. Anim. Ecol. 43(1): 239-253. ROHLF, F.J. & SOKAL, R.R. (1969). Statistical Tables. Freeman, London. ROSS, H.H. (1957). Principles of natural coexistence indicated by leaf- hopper populations. Evolution 11(2): 113-129. ROSS, H.H. (1958). Further comments on niches and natural coexistence. Evolution 12: 112-113. ROUGHGARDEN, J. (1971). Density dependent natural selection. Ecology 52: 453-468. 301 ROUGHGARDEN, J. & FELDMAN, M. (1975). Species packing and predation pressure.- Ecology 56: 489-492. RYAN, R.B. (1971). Interaction between two parasites Apecthis ontario and Hoplectis quadricingulatus. I. Survival in singly attacked, super- and multiparasitized greater wax moth pupae. Ann. Ent. Soc. Am. 64: 205-8. SALE, P.F. (1977). Maintenance of high diversity in coral reef fish communities. Am. Nat. 111(978): 337-359. SALT, G. (1936). Experimental studies in insect parasitism. IV: The effect of superparasitism on populations of Trichogramma evanescens. J. exp. Biol. 13: 363-375. SALT, G. (1937). Sense used by Trichogramma to distinguish between parasitized and unparasitized hosts. Proc. Roy. Soc. Lond. Ser. B. 122: 57-75. SALT, G. (1961). Competition among insect parasitoids. S.E.B. Symposium XV: Mechanisms in Biological Competition. 96-119. SALT, G. (1975). The fate of an internal parasitoid, Nemeritis canescens, in a variety of insects. Trans. Roy. Ent. Soc. 127: 141-161. SALT, G. (1976). The hosts of Nemeritis.canescens, a problem in the host specificity of insect parasitoids. Ecol. Ent. 1: 63-67. SAVAGE, J.M. (1958). The concept of ecological niche with respect to the theory of natural coexistence. Evolution 12: 111-112. 302 SCHLINGER, E.I. & MACKAUER, M.J.P. (1963). Identity, distribution and hosts of Aphidius matricariae (Haliday), an important parasite of the green peach aphid Myzus persicae. Ann. Ent. Soc. Am. 56: 648-53. SCHOENER, T.W. (1969). Optimal size and specialization in constant and fluctuating environments: An energy-time approach. In: Diversity and Stability in Ecological Systems. SCHOENER, T.W. (1974). Some methods for calculating competition coefficients from resource-utilization spectra. Am. Nat. 108(961): 332-340. SCOPES, N.E.A. (1970). Control of Myzus persicae on year round Chrysanthemums by introducing aphids parasitized by Aphidius matricariae into boxes of rooted cuttings. Ann. Appl. Biol. '70 66: 323-327. SEDLAG, U. (1959). Untersuchungen ūber Bionomie und Massenweaksel von D2aeretus rapae (Curt.) (Hymenoptera: Aphidiidae). Trans. I. Int. Cont. Insect Pathology and Biol. Control, Praha 367-373. SHEPPARD, D.H. (171). Competition between two chipmunk species (Eutamias). Ecology 52: 320-329. SHIGA, M. & NAKANISHI, A. (1968a). Intraspecies competition in a field population of Gregopimpla himalayensis (Hymenoptera, Ichneumonidae) parasitic on Malocosoma neustria testacea (Lepidoptera, Lasiocamp). Res. Pop. Ecol. Kyoto Univ. X: 69-86. 303 SILHACEK, D.L. & MILLER, G.L. (1972). Growth and development of the Indian Meal Moth, Plodia interpunctella (Lepidoptera: Phycitidae), under laboratory mass-rearing conditions. Ann. Ent. Soc. Am. 63: 1084-7. SLOBODKIN, L.B. (1964). Experimental populations of Hydriola. J. Anim. Ecol. 33(Suppl.): 131-148. SMITH, H.S. (1929). Multiple parasitism: its relation to the biological control of insect pests. Bull. Ent. Res. 20: 141-9. SMITH-GILL, S.J. & GILL, D.E. (1978). Curvilinearities in the competition equations: an experiment with ranid tadpoles. Am. Nat. 112(985): 557-570. SNEDECOR, G.W. & COCHRAN, W.G. (1967). Statistical Methods. 6th edn., Iowa State University Press. SNYMAN, A. (1949). The influence of population densities on the develop- ment and oviposition of Plodia interpunctella (Hubn) (Lepidoptera). J. ent. Soc. Sth. Aft. 12: 137-71. SOLIMAN, H.S. (1940). Studies in the biology of Microbracon hebetor Say. Bull. Soc. Fouad. 1er Ent. 24: 215-47. SOLIMAN, H.S. (1941). Studies in the structure of Microbracon hebetor Say. Bull Soc. Fouad. ter Ent.'25: 1-96. SOUTHWOOD, T.R.E. (1975). The dynamics of insect populations. In: Insects, Science and Society. Ed. D. Pimentel. 151-199. SOUTHWOOD, T.R.E. & COMINS, H.N. (1976). A synoptic population model. J. Anim. Ecol. 45: 949-65. 304 SOUTHWOOD, T.R.E.; MAY, R.M.; CONWAY, G.R. & HASSELL, M.P. (1974). Ecological strategies and population parameters. Am. Nat. 108: 791-804. SPENCER, H. (1926). Biology of parasites and hyperparasites of aphids. Ann. Ent. Soc. Am. 19: 119-157. SPRADBERY, J.P. (1970). Host finding by Rhyssa persuasoria (L.), an Ichneumonid parasite of siricid wood wasps. Anim. Behan. 18: 103-114. STARY, P. (1960). The generic classification of the family Aphiidae (Hymenoptera). Acta. Soc. Ent. Cechosi. 57: 238-252. STARY, P. (1964). Food specificity in the Aphidiidae (Hymenoptera). Entomophaga 9: 91-99. STARY, P. (1966). Aphid parasites of Czechoslovakia. A Review of the Czechoslovak Aphidiidae (Hymenoptera). The Hague: Junk. STEINHAUS, E.A. & MARSH, G.A. (1962). Report of diagnosis of diseased insects 1951-1961. HiZgardia 33: 349-490. STEWART, F.M. & LEVIN, B.R. (1973). Partitioning of resources and the outcome of interspecific competition: a model and some general considerations. Am. Nat. 107(954): 171-198. STUBBS, M. (1977). Density-dependence in the life cycles of animals and its importance in K- and r-strategies. J. Anim. Ecol. 46: 677-689. 305 SYKES, R.M. (1974). Competition for nutrients among the plankton. Am. Nat. 108(962): 578. TAKAHASHI, F. (1961). Studies on the fluctuation in the experimental population of the Almond moth, Ephestia cautelia Walker. Jap. J. Ecol. 11: 239-45. TERBORGH, J. & WESKE, J.S. (1975). The role of competition in the distribution of Andean birds. Ecology 56: 562-576. THURSTON, R. & WEBSTER, J.A. (1962). Toxicity of Nicotiana gossei Domin. to Myzus persicae (Sulzer). Ent. Exp. Appl. 5: 233-38. TILMAN, D. (1977). Resource competition between planktonic algae: An experimental and theoretical approach. Ecology 58(2): 338-348. TREHERNE, R.C. (1921). Some notes on the fruit worms of British Columbia. Scient. Agric. 1: 116-9. TSUJI, H. (1958). Study on the diapause of the Indian meal moth, Plodia interpunctella (Hb). The influence of temperature on the diapause and the type of diapause. Jap. J. appl. Ent. Zool. 2: 17-23. TSUJI, H. (1959a). Studies on the diapause of the Indian meal moth Plodia interpunctella (Hb). II The effect of population density on the induction of diapause. Jap. J. appZ. Ent. Zool. 3: 33-40. TSUJI, H. (1959b). Study on the diapause of the Indian meal moth Plodia interpunctella (Hb). III Influence of high temperature on the inception of diapause in the 20°C non-diapause stock. Jap. J. app l . Ent. Zool. 3: 250-4. 306 TZANAKAKIS, M.E. (1959). An ecological study of the Indian meal moth Plodia interpunctella (Hb) with emphasis on diapause. Hilgardia 29: 205-46. UETZ, G.W. (1977). Coexistence in a guild of wandering spiders. J. Anim. Ecol. 46: 531-542. ULFSTRAND, S. (1977). Foraging niche dynamics and overlap in a guild of passerine birds in South Swedish coniferous woodland. Oecologia 27(1): 23-45. ULLYETT, G.C. (1936). Host selection by Microplectron fuscipennis Zett. (Hymenoptera: Chalcididae). Proc. Roy. Soc. Lond. Ser. B. 120: 253-291. ULLYETT, G.C. (1943). Some aspects of parasitism in field populations of Plutella maculipennis. J. ent. Soc. S. Africa VI: 65-80. ULLYETT, G.C. (1945). Distribution of progeny by Microbracon hebetor Say. J. ent. Soc. S. Africa VIII: 123-31. ULLYETT, G.C. (1945). Oviposition by Ephestia kuhniella Zell. J. Ent. Soc. S. Afr. 8: 53-59. ULLYETT, G.C. (1947). Mortality factors in populations of PZutella maculipennis Curtis (Tineidae: Lepidoptera) and their relation to the problems of control. Union S. Afr. Dept. Agr. For. Ent. Memoirs 2: 77-198. ULLYETT, G.C. & MERWE, J.S.V.D. (1947). Some factors influencing pop- ulation growth of Ephestia kuhnielZa Zell (Lepidoptera: Phycitidae). J. Ent. Soc. Sth. Afric 10(1): 46-63. 307 UTIDA, S. (1952). Density effects observed in the interacting pop- ulations of a host and its two parasites. Experimental studies on synparasitism. 1st report 0yo Kontyū 8(1): 1-7. UTIDA, S. (1953). Interspecific competition between two species of bean weevil. Ecology 34: 301-307. UTIDA, S. (1961). Experimental studies on the interactions between bean weevils and their parasitic wasps. XI int. Kongr. Ent. Wien (1960) 1: 731-734. VAGO, C. & KURSTAK, E. (1965). Mēcanisme d'action de Bacillus thuringiensis introduit par un parasite Hymenoptēre dans l'hēmolymphe d'un Ldpidoptēre. Antonie Van Leeuwenhoek 31: 282-294. VANDERMEER, J.H. (1969). The competitive structure of communities: an experimental approach with Protozoa. Ecology 50: 362-371. VAN EMDEN, H.F. (1962). Observations of the effect of flowers on the activity of parasitic Hymenoptera. Ent. Mon. Mag. 98: 265-70. VAN EMDEN, H.F.; EASTOP, N.F.; HUGHES, R.D. & WAY, M.J. (1969). The ecology of Myzus persicae. A. Rev. Ent. 14: 197-270. VAN STEENBURGH, W.E. & BOYCE, H.R. (1937). The simultaneous propagation of Macrocentrus ancylivorus Roh. and Ascogaster carpocapsae Vier. on the peach moth (Easpeyresia molesta Busch.), a study in multiple parasitism. Rep. Ent. Soc. Ont. 68: 24-26. VAN ZANT, T.; POULSON, T.L. & KANE, T.C. (1978). Body-size differences in carabit cave beetles. Am. Nat. 112(983): 229-234. 308 VARLEY, G.C. & GRADWELL, G.R. (1968). Population models for the winter moth. In: Insect Abundance. Ed. T.R.E. Southwood. Symp. R. ent. Soc. Lond. 4: 132-142. VARLEY, G.C. & GRADWELL, G.R. (1970). Recent advances in insect pop- ulation dynamics. A. Rev. Ent. 15: 1-24. VEVAI, E.J. (1942). On the bionomics of Aphidius etatricariae Hal. A braconid parasite of Myzus persicae Sulz. Parasitology 34: 141-151. VINSON, S.B. (1972). Competition and host discrimination between two species of tobacco budworm parasitoids. Ann. ent. Soc. Am. 65: 229-36. VINSON, S.B. & GUILLOT, F.S. (1972). Host marking-source of a substance that results in host discrimination in insect parasitoids. Entomophaga 17: 241-245. VOLTERRA, V. (1926). Fluctuations in the abundance of a species considered mathematically. Nature 118: 558-560. WAAGE, J.K. (1977). Behavioural aspects of foraging in the parasitoid Nemeritis canescens. Unpublished Ph.D thesis, University of London, Imperial College, Silwood Park. WALLACE, B. (1974). Studies on intra- and interspecific competition in Drosophila. Ecology 55: 227-244. WATSON, M.A. (1940). Studies on the transmission of sugar beet yellows virus by the aphis Myzus persicae (Sulz.). Proc. Roy. Soc. (B) 128: 535-552. 309 WAY, M.J. & MURDIE, G. (1965). An example of varietal variations in resistance of Brussels sprouts. Ann. Appl. Bio Z. 56: 326-28. WERNER, E.E. (1977). Species packing and niche complementarity in three sunfishes. Am. Nat. 111(979): 553-578. WERNER, E.E. & HALL, D.J. (1977). Competition and habitat shift in two sunfishes (Centrachidae). Ecology 58(4): 869-876. WESELOH, R.M. (1976). Behavioural responses of the parasite, Apanteles melanoscelus to Gypsy moth silk. Envir. Ent. 5: 1128-32. WEYGOLDT, P. (1977). Coexistence of two species of Whip spiders (Genus Heterophxynus) in the Neotropical rainforest (Arachnida, Amblypygi). Oecologia 27: 363-370. WHITING, P.W. (1935). Sex determination in bees and wasps. J. Hered.. 26: 263. WILLARD, H.F. & MASON, A.C. (1937). Parasitization of the Mediterranean fruit fly in Hawaii in 1921. J. Agric. Res. 33: 9-15. WILBUR, H.M. (1972). Competition, predation and the structure of the Ambystoma - Rana sylvatica community. Ecology 53: 3-21. WILSON, D.S. (1975). The adequacy of body size as a niche difference. Am. Nat. 109: 769-784. WYLIE, H.G. (1966). Some mechanisms that effect the sex ratio of Nasonia vitripennis (Walk) (Hymenoptera: Pteromalidae) reared from superparasitized housefly pupae. Can. Ent. 98: 645-653. 310 WYLIE, H.G. (1971). Observations on intraspecific larval competition in three hymenopterous parasites of fly puparia. Can. Ent. 103: 137-142. WYLIE, H.G. (1972). Larval competition among three hymenopterous parasite species on multiparasitized housefly (Diptera) pupae. Can. Ent. 104: 1181-1190. ZARET, T.M. & RAND, A.S. (1971). Competition in tropical stream fishes: support for the competitive exclusion principle. Ecology 52: 336-42. ZWOLFER, H. (1971). The structure and effect of parasite complexes attacking phytophagous host insects. Proc. Adv. Study Inst. Dynamics Numbers Popul. (Oosterbeek, 1970) 405-418. 311 APPENDIX A CULTURING METHODS A 1. The Peach-Potato aphid Myzus persicae (Sulz.) and its parasitoids Diaeretiella rapae and Aphidius matricariae Initially, large populations of Myzus were maintained by placing infested leaves on clean Brussels sprout plants, Brassica oleraceae (c.v. Gemmifera var. Irish Elegance). It was found to be extremely important that the plants were not under water stress, when they produce tough, waxy leaves. Under these conditions, and when overcrowded, the aphids tend to remain very small, with relatively low fecundity. Overcrowded aphids also produce large quantities of honeydew, which becomes infested with a sooty mould if left too long. Aphids reach maximum size, and fecundity on very young or very old, yellowing leaves. Once large numbers of aphids were being standardized from birth (see Chapter Three), there were always sufficient left over, after experiments had been set up, to maintain both the parasitoid and the standard aphid populations. The large, whole plant cultures were therefore dispensed with. The small scale cultures were maintained in a small cage consisting of a plastic pot saucer, diameter 12.7 cm, over which fitted a clear polystyrene propagator, height 19.0 cm. A circle of diameter 5.7 cm was cut from the top of the propagator with a heated scalpel and replaced by a piece of fine-meshed muslin, glued down with chloroform, for ventilation. Oval introduction holes (3.5 x 3.3 cm) were cut in the wall of the cage; one 4 cm from the top, and the other 5 cm from the base directly opposite. These were covered by removeable muslin patches, held down with masking tape. The propagator was also sealed to the pot saucer with masking tape. The propagator was also sealed to the pot saucer with masking tape (see Plate 1). 312 Small Diaeretiella cultures were set up twice weekly, using large 4th instar or young adult aphids (7-12 days old) left over from the standardized insects after experiments had been set up. 30-60 newly emerged adult wasps were introduced into a cage containing 2 or 3 potted leaves covered with aphids, and 1 or 2 clean leaves. The cages in which the standard aphids had matured were re-used for parasitoid cultures because they contained honeydew deposits, on which the adult wasps feed. A fresh batch of adult parasitoids was added every other day for a week. Cultures were kept at 20°C, and moved to 25°C or 15°C, if speeding up or slowing down of the rate of development was necessary, depending on the demand for fresh adult wasps. These emerged after about 20 days at 20°C. Clean leaves were added as required. Myzus has a useful tendency to desert overcrowded leaves and wander up the walls of the container, which indicates that a new leaf is necessary. Aphidius cultures were set up in exactly the same way. A 2. The Indian meal moth Plodia interpunctella and its parasitoids Nemeritis canescens and Bracon hebetor Plodia interpunctella cultures were made up twice weekly in medium or large plastic boxes, two-thirds full of culturing medium. The boxes were of clear polystyrene (28.1 x 17.1 x 10.3 cm or 22.8 x 12.1 x 8.9 cm), ventilated by an oval introduction hole (3.5 x 3.3 cm) and a square window (11.5 x 10.3 cm) cut into the lid and covered with muslin (as in Plate 2) . Culture medium was made up in the proportions: (roughly) 3 Kgm Middlings wheat feed, 150 gm dried powdered yeast and 200 mls glycerol, rubbed in with the fingers to give a granular texture. A 5 x 2.5 cm tube, filled with water and corked was added to provide drinking water. A strip 313 of filter paper, projecting beyond the cork, was used as a wick. 140 adult moths, 1-2 days old, were added to each box, and the cultures reared at 25°C with a 16 hour day. (All rearing and experi- mental work was subjected to a 16 hour day). The relative humidity was 60-65%. After 20 days most of the larvae had reached the fifth instar and were suitable for use in experiments or for rearing parasitoids. Fresh Nemeritis cultures were made up weekly (or twice weekly if demand was high). A layer of 20-30 day old Plodia culture (used prior to the emergence of any adults), roughly 0.5 cm deep, was placed in a medium-sized culture box (see above). A plastic top (diameter 2.5 cm), containing a pad of cotton wool soaked in one in ten honey in water solution, was added. 20 Nemeritis adults (1-2 days old) were added to start with, and 12 more added 4 days later. After about 20 days at 25°C, the adult wasps began to emerge. These were removed daily. Three small Bracon cultures were set up twice weekly, in small round polystyrene dishes (maximum diameter 10.5 cm, height 4.4 cm). A very thin layer of fresh culturing medium (three large pinches) was scattered over the bottom. To this were added 25 large fifth instar Plodia larvae, a 2.5 cm plastic top with a honey pad, and eight pairs of Bracon, 1-2 days old (see Plate 3). The host larvae were hand picked to minimize the spread of disease. A population of Bracon had been wiped out by the use of Plodia cultures containing diseased larvae. When especially high numbers of Bracon were required, ten extra 314 larvae and four or eight pairs of parasitoids were added four days after the culture had been started. The cultures were then kept for one week on a warm shelf (22.5 ± 3.5°C), at a relative humidity of about 40%. They were then transferred to a cooler corner (18.3 ± 1.5°C) for a further week before the adult wasps were removed to avoid confusion between the generations. The offspring began to emerge one or two days later. They were removed daily. Males recovered from one culture were mated with females from another, and vice versa, in an attempt to maintain genetic diversity, as inbreeding has a tendency to lower fecundity and reduce egg viability (Whiting, 1935). 315 APPENDIX B TABLES OF DATA FOR FIGURES These tables are numbered to coincide with the figures plotted from them. Thus, table B3.2 gives the data which produced figure 3.2. Gaps in the numbering will therefore occur when a figure consists, for example, of a flow diagram with no raw data. In the figure legends, a = constant; b = coefficient of x (or log x); c = coefficient of x2; se a b, c standard error of the estimates of a, b, or c respectively (in e.g. y = a + bx + cx2). Table B3.2. Diaeretiella rapae. P = parasitoid density; a = searching efficiency. Log P Log a Log P Log a Log P Log a 0.000 -0.644 0.602 -1.572 0.845 -1.341 -0.747 -1.143 -1.341 -0.970 -1.175 -0.541 -1.143 0.477 -1.567 -1.625 -1.143 -1.296 -1.625 -0.546 -0.904 0.544 -1.188 -0.683 0.903 -1.839 -1.268 -0.557 -1.014 -0.984 -1.090 -1.627 -1.314 -1.400 -1.476 -1.250 -1.528 1.190 -1.292 -0.865 -1.311 0.301 -1.347 -1.219 -1.048 -1.175 1.161 -1.391 -0.891 -1.197 -0.664 -0.858 1.204 -1.498 -0.441 -1.848 -1.071 -1.277 -1.025 -1.363 -0.926 -1.101 -1.237 -0.664 0.875 -1.094 -1.465 -1.191 -1.250 -1.180 316 Table B3.3. Aphidius matricarias. P = parasitoids present; a = searching efficiency. 95% Confidence 95% Confidence limits Log P a Mean a Log a limits Upper Lower Upper Lower 0.000 0.090 0.012 0.028 -0.004 -1.915 -1.554 Log -0.004 0.000 0.000 0.000 0.000 (0.008 x7) 0.301 0.028 0.043 0.085 0.002 -1.363 -1.070 -2.811 0.016 0.062 0.036 0.114 0.004 0.602 0.000 0.011 0.020 0.003 -1.944 -1.699 -2.558 0.000 0.004 0.020 0.006 0.027 0.020 0.014 0.903 0.000 0.012 0.025 -0.001 -1.928 -1.600 Log -0.001 0.017 0.001 0.017 0.024 1.204 0.017 0.012 0.021 0.003 -1.921 -1.680 -2.510 0.006 0.004 0.021 0.012 317 Table B3.4. Diaeretie fla rapae. P = parasitoid density; a = searching efficiency; bTw = compound parameter of encounter rate be- tween adult parasitoids and the time wasted per encounter. 95% Confidence 95% Confidence limits P a Mean a bT limits - - - w Upper Lower Upper Lower 1 0.227 0.137 0.205 0.069 - - - 0.179 0.107 0.288 0.024 0.024 0.207 0.277 0.081 0.040 0.056 2 0.045 0.141 0.203 0.079 -0.028 0.734 -0.325 0.090 0.129 0.217 0.139 0.362 0.085 0.094 0.119 0.058 0.217 4 0.027 0.099 0.196 0.002 0.128 22.500 -0.100 0.072 0.067 0.072 0.072 0.284 8 0.014 0.058 0.088 0.028 0.195 0.556 0.080 0.097 0.024 0.033 0.030 0.136 0.060 0.067 0.064 318 Table B3.4. Continued. 95% Confidence 95% Confidence P a Mean a limits limits bTw Upper Lower Upper Lower 16 0.032 0.044 0.074 0.014 0.141 0.586 0.057 0.014 0.053 0.043 0.079 7.5 0.081 0.060 0.081 0.039 0.197 0.387 0.106 0.034 0.064 0.056 0.066 bTw = —a - —a where a = searching efficiency when P = 1 a' (P-1) a' = new searching efficiency. 319 Table B3.5. Aphidius matricariae. See table B3.3 for individual measure- ments of a. a = searching efficiency; P = parasitoids present; b = encounter rate between wasps; Tw = time wasted per encounter. 95% Confidence limits 95% Confidence limits P Mean a bTw Upper Lower Upper Lower 1 0.012 0.028 -0.004 - - - 2 0.043 0.085 0.002 -0.721 5.000 -0.859 4 0.011 0.020 0.003 0.030 1.000 -0.133 8 0.012 0.025 -0.001 0.000 * -0.074 16 0.012 0.021 0.003 0.000 0.200 -0.029 * A negative searching efficiency is meaningless. If the lowest realistic value (a = 0.0) is used, this value for bTw is co. bTw = a - a' as given in table B3.4. a'(P-1) 320 Table B3.6. Diaeretiella rapae. The observed offspring are uncorrected surviving adults or dead mummies. The expected values are based on a, calculated from the corrected number of healthy surviving hosts, i.e. when a number of aphids (up to ten) died from unknown causes, the corrected number of unparasitized hosts was calculated using healthy aphids healthy aphids x original aphid density parasitized aphids Parasitoids Expected Expected Offspring present offspring scramble contest 1 20 19.29 21.05 12 12.40 13.04 32 27.83 32.14 3 3.02 3.05 26 23.32 26.09 3 3.02 3.05 24 21.69 24.02 31 27.06 31.07 10 9.63 10.00 5 4.96 5.04 7 6.83 7.00 2 11 10.61 11.06 20 19.38 21.16 27 25.69 29.21 45 36.21 45.21 29 27.13 31.17 65 45.15 66.12 19 18.49 20.08 22 20.08 22.02 25 24.18 27.20 14 13.32 14.07 43 36.21 45.21 4 13 12.50 13.15 32 27.83 32.14 29 26.42 30.19 32 27.83 32.14 32 27.83 32.14 87 46.85 87.08 8 14 12.91 13.61 68 45.93 69.27 21 20.43 22.44 28 26.13 29.80 321 Table 83.6. Continued. Parasitoids Expected Expected Offspring present offspring scramble contest 25 24.33 27.41 85 47.08 85.06 49 38.25 48.94 51 40.37 53.27 50 39.51 51.44 16 51 39.51 51.44 26 23.08 25.78 68 46.70 73.36 64 44.49 63.84 87 45.84 92.02 7.5 54 42.59 58.44 29 25.47 28.91 49 38.85 50.13 44 35.54 44.04 48 38.25 48.94 7 34 30.07 35.36 35 30.07 35.36 3 10 9.63 10.00 18 16.93 18.23 3.5 26 23.36 26.14 21 20.17 22.12 36 32.58 39.18 20 18.63 20.25 15.5 67 46.12 70.12 67 45.71 68.29 14.5 56 42.22 57.53 322 Table B3.7. Aphtidius matricariae. Observed Expected Offspring P Offspring Scramble Contest 1 11 10.61 11.06 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 0 0.00 0.00 1 1.02 1.02 x7 2 7 6.83 7.00 4 4.00 4.05 15 14.12 14.98 8 8.64 8.93 26 23.40 26.19 1 1.02 1.02 4 0 0.00 0.00 0 0.00 0.00 2 2.03 2.04 10 9.52 9.88 3 3.02 3.05 13 12.50 13.15 10 9.52 9.88 7 6.83 7.00 8 0 0.00 0.00 16 15.31 16.34 1 1.02 1.02 16 15.31 16.34 20 20.43 22.44 16 28 26.71 30.59 11 11.25 11.76 7 7.74 7.97 36 30.94 36.65 23 20.43 22.44 323 Table B3.8. Diaeretielia rapae. Sex ratio changes with parasitoid density. 957 Confidence-- levels Parasitoid Females Total % Females density Upper Lower 1 67 110 60.91 70.99 52.22 2 104 165 63.03 70.97 55.33 3 4 8 50.00 62.18 17.75 3.5 66 103 64.08 73.19 54.07 4 67 133 50.38 59.01 41.70 7 28 47 59.57 73.52 44.31 7.5 107 194 55.15. 62,23 47.88 8 100 158 63.29 70.65 55.40 14.5 8 17 47.06 69.45 23.49 15.5 14 23 60.87 79.96 40.45 16 23 61 37.70 50.85 25.86 Note: The 95% confidence limits were calculated using table W in Rohlf and Sokal (1969). 324 Table B3.9. The decline in log a (searching efficiency) with log P (parasitoid density) in Nemeritis canescens. Log P Log a Log P Log a 0.000 -0.818 1.204 -1.638 -0.496 -1.921 -1.092 -1.569 -0.352 -1.432 -0.607 -1.444 -0.526 -1.538 -0.426 -1.527 -0.558 -1.168 -1.252 0.301 -1.237 -0.480 1.505 -1.678 -0.490 -1.678 -1.310 -1.585 -0.963 -1.585 -0.842 -1.921 -0.664 -1.469 -0.714 -1.620 -1.699 0.602 -1.420 -1.056 -0.928 -0.791 -0.676 -1.114 -1.174 -0.896 -0.873 0.903 -1.301 -1.854 -1.569 -1.409 -1.051 -1.131 -1.495 -2.222 -0.947 -1.102 -1.367 -1.229 -1.229 325 Table B3.10. The decline in log a (searching efficiency for healthy hosts) with log P (parasitoid density) in Bracon hebetor. Log P Log a Log P Log a 0.000 -0.305 1.204 -1.000 -0.724 -1.268 -0.328 -1.328 -0.440 -1.260 -0.510 -1.000 -0.400 -1.201 -0.426 -0.719 -0.703 -0.936 -0.971 -0.664 0.301 -0.285 -0.824 -0.796 -0.870 -0.511 -0.845 -0.688 -1.168 -1.824 -0.648 0.602 -0.893 -0.604 -0.921 -0.833 -0.967 -0.574 -0.712 -0.963 -0.527 -1.056 -0.925 -0.666 0.903 -1.061 -0.559 -0.658 -1.102 -0.996 -1.268 -0.623 -0.757 326 Table 33.11. The decline in log s (efficiency of search for paralysed hosts) with log P (parasitoid density) in Bracon hebetor. 95% Confidence 95% Confidence limits limits Log P s Mean s Log S Upper Lower Upper Lower 0.000 0.020 0.023 0.039 0.007 -1.638 -1.404 -2.186 0.047 0.000 0.027 0.080 0.030 0.000 0.000 0.024 0.000 0.025 0.301 0.000 0.037 0.062 0.012 -1.432 -1.205 -1.932 0.075 0.038 0.065 0.094 0.053 0.000 0.000 0.045 0.000 0.602 0.000 0.027 0.042 0.012 -1.565 -1.374 -1.915 0.072 0.033 0.022 0.028 0.015 0.021 0.021 0.033 0.903 0.011 0.018 0.026 0.011 -1.739 -1.588 -1.972 0.015 0.009 0.018 0.024 0.009 0.026 0.034 327 Table B3.12. The relationship between the pupae produced per parasitoid and P, wasp density, in Bracon hebetor. 957. Confidence limits P Pupae/parasitoid Mean Upper Lower 1 0.0000 1.545 2.684 0.407 0.0000 0.0000 0.0000 1.0000 1.0000 1.0000 4.0000 4.0000 4.0000 2.0000 2 0.0000 2.364 4.252 0.475 0.0000 0.0000 0.0000 0.0000 0.0000 4.0000 4.5000 7.0000 5.0000 5.5000 4 0.0000 3.063 4.313 1.812 6.0000 3.2500 2.2500 0.2500 1.7500 5.7500 2.0000 3.7500 2.5000 5.0000 4.2500 8 1.5000 4.422 6.990 1.854 10.1250 4.3750 2.1250 2.6250 2.0000 5.0000 7.6250 328 Table B3.12. Continued. 95% Confidence limits P Pupae/parasitoid Mean Upper Lower 16 2.1875 3.914 6.218 1.610 1.3750 2.8750 2.8125 6.0625 0.8125 7.8125 7.3750 329 Table 83.13. The relationship between hosts parasitized per wasp and wasp density (P) in Bracon hebetor. 95% Confidence limits P Hosts parasitized/wasp Mean Upper Lower 1 0.000 0.818 1.405 0.231 0.000 0.000 0.000 1.000 1.000 1.000 1.000 3.000 1.000 1.000 2 0.000 1.136 2.063 0.209 0.000 0.000 0.000 0.000 0.000 3.000 2.000 3.000 1.500 3.000 4 0.000 1.313 1.892 0.733 2.500 3.000 0.750 0.250 1.000 2.250 1.000 1.000 0.750 1.500 1.750 8 0.625 1.453 2.225 0.681 2.250 1.500 0.625 0.625 0.750 2.500 2.750 330 Table B3.13. Continued. 95% Confidence P Hosts parasitized/wasp Mean limits Upper Lower 16 0.688 1.336 2.043 0.629 0.688 1.000 0.813 2.125 0.500 2.563 2.313 331 Table B3.14-15. Data for the relationships between the clumpiness (S2/m) of the egg distributions of Bracon hebetor and parasitoid density (P), and between clumpiness and hosts surviving the experiment. Parasitoid density S2/m Hosts surviving 1 0.000 78 0.000 106 0.000 90 0.000 94 0.000 86 0.000 88 0.250 89 2 1.733 85 0.200 45 0.333 94 0.440 93 0.100 97 1.455 68 4 1.533 47 1.924 78 0.115 71 0.721 44 0.524 59 0.282 79 0.000 40 0.937 79 8 0.524 61 0.909 23 1.833 67 1.749 22 0.500 56 0.631 83 0.632 19 0.428 30 16 1.000 26 0.904 54 1.200 59 0.854 53 1.107 26 1.286 46 1.049 0 332 Table B3.16. The relationship between bTw b = rate of encounter between adults, Tw = time wasted per encounter :I and P-1, (P = parasitoid density) in Nemeritis canescens. 95% Confidence 95% Confidence P a Mean a limits limits - - bTw Upper Lower Upper Lower 0.152 0.274 0.372 0.177 - - - 0.319 0.081 0.445 0.247 0.298 0.375 0.277 2 0.058 0.173 0.280 0.065 0.584 3.215 -0.021 0.331 0.324 0.049 0.109 0.144 0.217 0.193 4 0.038 0.114 0.154 0.073 0.468 0.918 0.260 0.088 0.118 0.162 0.211 0.077 0.067 0.127 0.134 8 0.050 0.053, 0.071 0.034 0.596 1.008 0.408 0.014 0.027 0.039 0.089 0.074 0.032 0.006 0.113 0.079 0.043 0.059 0.059 333 Table B3.16. Continued. 95% Confidence 95% Confidence P a Mean a limits bT limits — — -- _ w Upper Lower Upper Lower 16 0.023 0.035 0.048 0.022 0.455 0.764 0.314 0.012 0.027 0.037 0.036 0.029 0.030 0.068 0.056 32 0.021 0.023 0.028 0.018 0.331 0.431 0.266 0.021 0.026 0.026 0.012 0.034 0.024 0.020 334 Table 33.17. The decline in Log bTw (calculated from a) with Log (P-1) in Bracon hebetor, where b = rate of encounter between adult wasps, Tw = time wasted per encounter, P = parasitoid density. P Log (P-1) bTw Log bTw 2 0.000 -0.434 - 0.960 -0.018 0.838 -0.077 1.178 0.071 -0.045 - 1.056 0.024 0.434 -0.362 3.324 0.522 18.600 1.270 0.307 -0.513 4 0.477 0.432 -0.364 0.060 -1.220 0.483 -0.316 0.333 -0.477 0.574 -0.241 0.034 -1.472 0.172 -0.765 0.566 -0.247 -0.003 - 0.780 -0.108 0.490 -0.310 0.120 -0.919 8 0.845 0.340 -0.469 0.009 -2.031 0.048 -1.318 0.389 -0.410 0.273 -0.564 0.635 -0.197 0.034 -1.473 0.097 -1.013 16 1.176 0.129 -0.888 0.296 -0.528 0.350 -0.455 0.290 -0.538 0.129 -0.888 0.244 -0.612 0.036 -1.444 335 Table B3.18. bTw (calculated from s) changes with P-1 for Bracon hebetor (where b = rate of encounter between wasps, Tw = time wasted per encounter, s = searching efficiency when seeking paralysed hosts, P = parasitoid density). •957 Confidence limits P-1 Mean bT w Upper Lower 1 -0.378 0.917 -0.629 -0.049 0.306 -0.151 7 0.040 0.156 -0.016 15 0.029 0.104 0.003 -wbT = —s - —s' s' (P-1) where s = s when P = 1 s' = new efficiency as P increases 336 Table B3.19. Expected offspring (calculated on the basis of scramble and contest competition between larvae), with the observed healthy offspring produced plus dead, fully-formed adults in sole possession of a host, for Nemeritis canescens. Expected Expected Observed healthy Observed healthy offspring offspring adults plus adults plus dead singles dead singles Contest Scramble Contest Scramble 18 18.12 16.83 49 57.35 42.15 35 35.08 29.88 54 56.21 41.68 9 10.00 9.63 41 47.66 37.57 44 46.12 36.72 46 48.94 38.25 28 28.11 24.87 72 85.06 47.08 28 33.10 28.50 75 75.93 47.02 40 40.16 33.20 62 62.80 44.16 29 31.07 27.06 62 62.80 44.16 13 14.07 13.32 70 72.48 46.56 57 62.15 43.94 71 72.48 46.56 59 61.21 43.62 40 40.95 33.69 12 11.99 11.46 84 85.06 47.08 23 25.16 22.60 58 68.79 45.83 32 32.14 27.83 59 60.68 43.43 44 45.21 36.21 40 41.12 33.80 17 18.12 16.83 34 38.10 31.89 47 48.31 37.91 58 61.21 43.62 73 73.14 46.66 32 34.04 29.17 30 30.19 26.42 51 51.14 39.35 53 53.27 40.37 40 42.33 34.54 9 13.61 12.91 24 24.95 22.43 33 34.42 29.43 63 65.37 44.94 55 57.35 42.15 28 29.01 25.54 6 6.02 5.90 75 76.35 47.06 55 60.13 43.23 36 37.38 31.42 48 48.31 37.91 47 48.31 37.91 33 39.54 32.81 23 22.44 20.43 36 45.04 36.12 337 Table B3.21. Total mortality imposed by competition (K-value) and the log (initial eggs present) for Neweritis canescens. Log (initial eggs) K-value Log (initial eggs) K-value 1.681 0.602 2.885 1.487 0.204 1.584 0.982 1.655 0.190 1.454 0.380 1.394 0.266 1.394 0.176 1.524 0.380 1.242 1.161 1.982 0.982 0.391 3.186 1.454 0.391 1.505 0.982 1.446 0.903 1.438 0.621 1.709 0.505 1.446 0.505 1.668 1.479 2.283 1.137 0.792 0.903 0.621 0.727 0.903 0.982 0.704 0.715 2.584 1.169 2.584 1.306 1.186 0.912 0.931 1.283 1.806 0.792 0.993 1.186 1.053 1.028 338 Table B3.22. k1-values and Log (initial egg density) for Nemeritis canescens. Log (egg density) - k1-value Log (egg density) k1-value 1.681 0.426 2.885 1.232 0.137 1.524 0.640 1.153 0.000 1.072 0.234 1.122 0.204 1.161 0.079 1.169 0.163 0.894 0.999 1.982 0.806 0.183 3.186 1.380 0.903 1.373 0.551 1.306 0.477 1.311 0.150 1.563 0.320 1.252 0.359 1.317 1.401 2.283 1.005 0.727 0.584 0.477 0.420 0.727 0.806 0.576 0.559 2.584 0.931 1.306 1.169 1.040 0.771 0.813 1.107 1.806 0.698 0.771 1.004 0.903 0.894 339 Table B3.23. k2-values and the log (population density - total Nemeritis remains) upon which they act, for Nemeritis canescens. Log (population k2 values Log (population k2-values density) density) 1.255 0.000 1.806 0.014 1.544 0.000 1.813 0.021 1.041 0.087 1.881 0.036 1.681 0.038 1.875 0.024 1.447 0.000 1.623 0.021 1.477 0.030 1.935 0.010 1.602 0.000 1.869 0.106 1.519 0.056 1.785 0.014 1.176 0.062 1.799 0.028 1.079 0.000 1.431 0.070 1.505 0.000 1.833 0.077 1.663 0.019 1.623 0.021 1.279 0.048 1.556 0.025 1.699 0.027 1.806 0.043 1.863 0.000 1.556 0.051 1.477 0.000 1.708 0.000 1.724 0.000 1.653 0.051 1.279 0.325 1.415 0.035 1.544 0.026 1.813 0.014 1.771 0.030 1.477 0.030 0.778 0.000 1.887 0.011 1.813 0.073 1.580 0.023 1.681 0.000 1.690 1.672 1.653 0.135 1.362 0.000 1.732 0.176 1.813 0.123 1.763 0.031 1.724 0.111 1.716 0.053 1.991 0.134 1.887 0.011 340 Table B3.24. k3-values and the log (population density - solitary remains) upon which they act, in Nemeritis canescens. Log (population Log (population k3-values density) k3-values density) 1.255 0.176 1.792 0.060 1.544 0.067 1.792 0.111 0.954 0.255 1.845 0.105 1.644 0.152 1.851 0.103 1.447 0.146 1.602 0.125 1.447 0.032 1.924 0.184 1.602 0.097 1.763 0.245 1.462 0.161 1.771 0.063 1.114 0.114 1.771 0.180 1.079 0.079 1.362 0.283 1.505 0.143 1.756 0.165 1.644 0.166 1.602 0.125 1.230 0.084 1.532 0.040 1.672 0.292 1.763 0.101 1.863 0.307 1.505 0.125 1.477 0.176 1.708 0.128 1.724 0.156 1.602 0.187 0.954 0.954 1.380 0.101 1.519 0.121 1.799 0.127 1.740 0.087 1.447 0.146 0.778 0.000 1.875 0.083 1.740 0.149 1.556 0.158 1.681 0.150 1.672 0.116 1.519 0.121 1.362 0.061 1.556 0.326 1.690 0.259 1.732 0.241 1.613 0.121 1.663 0.301 1.857 0.214 1.875 0.151 341 Table B3.25. k2-values (death of supernumerary larvae) with the log (total remains) on which they act, in Nemeritie canescens, (short term experiments). Log (total remains) k2 0.778 0.000 0.301 0.000 1.255 0.000 0.778 0.000 1.176 0.000 1.519 0.000 1.415 0.000 1.279 0.000 1.000 0.000 1.230 0.000 1.505 0.028 1.041 0.000 0.903 0.000 1.447 0.000 1.477 0.000 1.477 0.273 1.362 0.040 1.613 0.000 1.362 0.000 1.230 0.054 1.477 0.135 1.732 0.051 1.362 0.040 1.653 0.020 1.204 0.125 1.785 0.014 1.398 0.076 1.778 0.030 1.663 0.050 1.568 0.024 1.591 0.000 1.778 0.062 1.959 0.219 1.716 0.211 1.690 0.146 1.826 0.063 1.799 0.000 1.477 0.030 342 Table 83.26. Length of males emerging with varying numbers of eggs developing per host in Bracon hebetor. Eggs/host Length of males (mm) 1 2.846 2.846 2.846 2.385 2.615 2.923 2.923 3.000 2 2.769 2.692 3.000 3.000 2.846 4 3.077 3.000 2.923 7 2.538 2.538 2.538 2.538 2.615 2.692 343 Table B3.27. Length of females emerging with varying numbers of eggs developing per host in Bracon hebetor. Eggs/host Length of females (mm) 1 3.154 3.000 3.308 3.231 3.308 3.154 3.154 3.154 3.308 2.692 2 3.154 3.308 3.077 3.231 3.385 4 3.154 3.231 3.231 2.923 3.308 7 2.923 8 3.154 3.154 3.154 2.385 2.385 3.077 2.615 2.615 2.692 1.923 1.692 3.308 12 2.615 2.615 3.077 2.846 2.846 2.923 344 Table B3.28. Total mortality (K-value) as initial density of eggs rises, in Bracon hebetor. Log (initial eggs) K-value Log (initial eggs) K-value 1.005 1.005 1.909 0.829 1.005 0.033 0.403 0.390 0.528 1.005 1.005 0.586 0.403 0.704 0.704 0.306 4.000 0.246 4.000 4.000 2.210 0.678 4.000 0.887 0.556 1.306 0.403 0.566 0.352 0.255 0.160 1.130 0.306 0.113 0.265 0.138 + co * +c + m + co 1.608 + co 0.227 0.494 0.653 0.829 0.306 0.704 0.431 1.607 0.607 0.306 0.377 * When there are no survivors, the "killing power" i.e. the K-value becomes + co, and is not plotted. 345 Table B3.29. Mean pupae recovered per parasitized host with number of Bracon females present. 957 Confidence limits Density Pupae/host Mean Upper Lower 1 1.00 2.05 3.32 0.77 1.00 4.00 1.33 1.00 4.00 2.00 2 1.33 2.29 3.17 1.41 2.25 2.33 3.33 2.20 4 2.40 2.37 2.99 1.74 2.43 1.08 3.00 1.00 1.50 2.22 2.00 3.75 3.33 3.33 8 2.40 2.74 3.54 1.93 4.17 2.25 1.60 4.20 2.67 2.00 2.09 16 3.09 2.70 3.25 2.15 1.91 2.81 3.38 2.65 1.50 3.05 3.19 NB. Replicates with no Bracon offspring produced have been omitted. 346 Table B3.30. Mean number of pupae recovered per host with the number of healthy hosts surviving the experiment, for Brecon hebetor. Healthy hosts Mean pupae surviving per host recovered 78 1.00 106 1.00 90 4.00 89 1.33 '94 1.00 86 4.00 88 2.00 46 1.33 95 2.25 93 2.33 98 3.33 70 2.20 46 2.40 56 2.43 80 1.08 71 3.00 39 1.00 82 1.50 44 2.22 59 2.00 79 3.75 90 3.33 80 3.33 67 2.40 23 4.17 22 2.75 69 1.60 58 4.20 83 2.67 20 2.00 36 2.09 26 3.09 54 1.91 61 2.81 53 3.38 26 2.65 47 1.50 7 3.05 2 3.19 105 0.00 114 0.00 115 0.00 103 0.00 97 0.00 85 0.00 113 0.00 64 0.00 77 0.00 347 Table B4.1. Percentage time spent searching by Nemeritis canescens. After 15 min. After After After Wasp 3 hrs 15 min. 6 hrs 15 min. 9 hrs 15 min. density Means, Means, Means, Means, Tp 95% C.L. % Tp 95% C.L. % Tp 95% C.L. % Tp 95% C.L. 1 43.33 57.04 37.06 23.32 25.39 12.40 2.17 4.81 54.67 67.01 19.94 45.77 12.78 20.94 7.94 12.57 63.00 47.07 0.00 0.88 12.17 3.86 0.17 -2.95 54.38 0.83 14.06 18.56 72.20 55.33 0.00 0.00 43.83 26.78 10.00 0.00 74.62 50.26 2 60.27 59.77 16.17 28.81 11.39 15.11 10.78 8.60 45.98 69.11 31.67 46.89 12.33 21.19 12.78 15.42 50.42 50.42 24.06 10.72 21.98 9.03 1.33 1.78 60.73 5.83 7.33 0.00 66.21 53.28 16.44 10.44 50.38 41.83 21.39 16.28 81.12 63.02 4 62.04 57.08 33.44 21.86 24.67 11.35 4.72 2.20 58.73 64.04 15.39 30.04 8.00 19.57 4.00 4.35 41.86 50.10 29.83 13.69 14.78 3.13 1.33 0.05 67.12 17.67 1.83 0.00 56.62 19.22 7.39 0.00 61.03 15.61 11.44 3.17 61.68 47.56 8 40.17 57.88 27.44 24.41 22.39 10.68 11.39 7.36 24.91 69.73 27.94 28.33 12.39 17.73 10.22 12.09 58.60 46.04 18.56 20.49 7.78 3.63 11.28 2.63 67.92 21.78 2.06 0.00 65.99 26.94 9.06 4.72 67.34 23.78 10.39 6.56 62.76 60.89 72.37 16 44.17 52.17 16.44 17.06 4.44 9.34 4.06 4.07 69.76 63.70 17.22 18.07 11.56 12.82 3.61 7.84 58.16 40.68 17.56 16.04 6.78 5.86 0.72 0.30 33.39 15.56 10.28 0.00 50.67 19.00 9.33 9.61 45.83 17.28 13.67 6.44 63.35 348 Table B4.1. Continued. After After After After 15 min. 3 hrs 15 min. 6 hrs 15 min. 9 hrs 15 min. Wasp density Means, Means, Means, Means, % Tp 95% C.L. % Tp 95% C.L. % Tp 95% C.L. % Tp 95% C.L. 32 43.45 53.66 18.22 21.90 6.56 13.39 1.17 5.65 45.59 62.37 16.33 27.65 6.17 7.39 12.55 46.94 44.94 31.33 16.14 9.17 1.89 -1.25 65.06 24.83 21.33 17.78 53.55 21.61 19.94 5.67 61.97 19.06 17.17 0.00 36.94 57.97 C.L. = confidence limits 349 Table B4.2. Number of cocking movements in 30 minutes observation of Nemeritis canescens. After After 3 hrs After 6 hrs After 9 hrs Wasp 15 min. 15 min. 15 min. 15 min. Overall density means, 95% C.L. Cocks Means Cocks Means Cocks Means. Cocks Means 15 9.50 3 2.33 1.33 0 0.50 3.42 10 1 r-+ 9.98 13 0 r-I - 3.15 10 1 r-I0 3 7 0 13 2 5 7 2 3 8.13 4 3.17 1.17 0 0.50 3.24 7 2 1-1 8.73 10 5 00 - 2.25 16 0 7 8 1-+ 1-I 11 0 0 11 4 11 10.25 2 2.17 2.33 r-+0 0.17 3.57 11 2 10.78 15 4 00 - 3.65 4 0 13 3 00 6 2 11 11 (" 8 7 6.22 2 3.00 1.00 I 1.17 2.85 3 3 r-i 6.70 12 5 N - 1.01 8 3 0 6 3 0 rI 8 2 4 5 3 16 10 7.43 4 1.71 0.33 O 0.50 2.52 8 1 O 7.84 9 4 r-10N - 2.79 4 ' 1 14 1 0 3 0 4 , 350 Table B4.2. Continued. After After 3 hrs After 6 hrs After 9 hrs Wasp 15 min. 15 min. 15 min. 15 min. Overall density means, 95% C.L. Cocks Means Cocks Means Cocks Means Cocks Means 32 10 9.89 1 1.33 1 0.67 00 0.00 2.97 5 1 2 10.35 4 2 0 00 - 4.41 13 2 1 - 16 1 0 00 8 1 0 9 14 C.L. = confidence limits Overall means were calculated (weighted or unweighted as necessary) as described in Bliss (1967). Table B4.3. Number of encounters between adult Nemeritis in 30 minutes observation. After 15 min. After 3 hrs 15 min. After 6 hrs 15 min. After 9 hrs 15 min. Wasp Overall Means, 95% C.L. density Encounters Means Encounters Means Encounters Means Encounters Means 1 ------2 0 3.00 3 8.50 1 3.83 1 2.69 4.45 0 10 4 0 8.67 0 16 3 0 0.22 2 0 0 1 14 6 5 7 3 16 10 7 1 4 4 21 16.63 19 18.17 10 8.83 7 1.83 11.37 17 22 6 2 23.39 13 29 12 2 - 0.66 11 18 11 0 1 12 7 0 13 9 7 0 27 30 8 13 21.67 42 46.50 26 25.67 16 12.67 26.63 21 60 34 9 49.41 27 53 41 31 3.84 16 30 5 1 15 51 32 14 27 43 16 5 25 21 30 16 63 40.57 69 76.00 76 51.50 35 18.17 46.61 23 68 60 13 85.10 37 66 32 12 8.19 49 56 46 3 29 101 37 15 43 98 58 31 40 32 29 80.89 202 178.33 133 133.00 31 69.67 112.91 48 145 75 82 196.80 97 203 124 95 29.02 82 166 138 110 95 132 144 77 97 122 184 23 82 35 C.L. = confidence limits. Overall means calculated as described in Bliss (1967) 353 Table B4.4. Mean lengths of search periods in 30 minutes observations of Nemeritis canescens (in seconds). Wasp After 3 hrs After 6 hrs After 9 hrs density After 15 min. 15 min. 15 min. 15 min. 1 15.92 12.83 8.96 7.80 13.30 5.89 4.89 6.22 31.27 - 5.09 3.00 37.29 5.00 6.84 4.99 22.24 13.64 - - 13.50 8.03 16.36 - 24.33 13.23 2 19.53 8.82 6.83 9.70 15.47 7.22 5.29 8.52 14.33 7.47 7.13 12.00 22.40 6.56 7.76 - 17.28 12.96 6.30 6.71 16.80 8.18 5.23 6.98 28.62 18.78 4 13.09 9.71 8.54 6.54 18.39 4.69 4.80 6.00 10.74 7.46 6.19 4.80 26.40 6.49 3.00 - 12.99 5.67 6.65 - 12.53 6.85 5.42 11.40 13.99 10.61 8 13.94 6.68 8.40 5.54 12.34 6.62 6.03 5.11 14.99 4.91 3.68 4.41 19.16 5.44 5.29 - 18.77 6.22 5.43 5.31 19.77 5.63 4.92 6.56 13.82 14.40 22.47 16 8.56 5.92 3.64 6.08 17.75 7.21 5.78 5.00 15.14 7.02 6.10 13.00 9.80 5.71 8.41 - 13.30 4.96 5.25 5.97 16.18 6.62 5.59 6.83 26.65 354 Table B4.4. Continued. After 9 hrs Wasp After 15 min. After 3 hrs After 6 hrs density 15 min. 15 min. 15 min. 32 15.17 5.56 5.13 3.50 16.44 5.88 6.94 4.75 10.62 8.29 6.11 4.86 13.29 6.88 6.86 6.04 10.69 7.34 8.35 5.67 12.03 5.91 5.83 - 9.14 17.10 Table B4.5. Mean length of search periods in 30 minutes observation on Nemeritis canescens, with the number of encounters between wasps in the same time. After 15 min. After 3 hrs 15 min. After 6 hrs 15 min. After 9 hrs 15 min. Wasp density Mean length Mean length Mean length Encounters Encounters Encounters Encounters Mean length search period search period search period search period O 1 0 see table 0 see table 0 see table see table B4.4 B4.4 B4.4 B4.4 2 0 19.53 3 8.82 1 6.83 1-10 9.70 0 15.47 10 7.22 4 5.29 8.52 0 14.33 16 7.47 3 0 7.13 1- 12.00 2 22.40 0 6.56 0 7.76 - r 14 17.28 6 12.96 5 6.30 -r 6.71 3 16.80 16 8.18 10 5.23 6.98 1 28.60 4 18.78 1 4 21 13.09 19 9.71 10 8.54 -N 6.54 17 18.36 22 4.69 6 4.80 6.00 13 10.74 29 7.46 12 6.19 N0 4.80 11 26.40 18 6.49 11 3.00 - 1 12.99 12 5.67 7 6.65 00 - 13 12.53 9 6.85 7 5.42 11.4 27 13.99 30 10.61 OV'LT S£ hr 6 Z8 £Z £5'S h8T T6'S ZZZ £0'ZT L6 L9'S LL S£'8 hin h£' L Z£T 69'0T S6 h0'9 OTT 98'9 8£T 88'9 991 6Z'£I Z8 98'h S6 II'9 hZI 6Z'8 £OZ Z9'OT L6 SL' ti Z8 h6.9 SL 88'5 ShI hh'9T 8h OS' £ T£ £T'S £8T 95-S ZOZ LT'SI 6Z Z£ 59'9Z Oh £86 9 T£ 6565 85 Z9'9 96 81'9T £h L6'S ST SZ'S LE 96'h TOT £0'£i 6Z ih'8 9h IL'S 95 08'6 6h 00'£T ZI OT'9 ZE Z0'L 99 hI' ST L£ 00'S £I 8L'S 09 IZ'L 89 SL'LI £Z 80'.9 S£ h9'£ 9L Z6'5 69 95'8 £9 91 Lh'ZZ 0£ 0h'hi TZ Z8'£T SZ 95'9 S Z6'h 91 £9'5 £h LL'6I LZ I£'S 171 £h'S ZE ZZ'9 TS LUST SI 6Z'S S hh'S 0£ 91'61 9T. Th'h T£ 89' £ 16'h £S 66'hi LZ TT'S 6 £0' 9 h£ Z9'9 09 h£'ZI TZ VS'S 91 h'8 9Z 89'9 Zh h6'£T £T 8 Table 84.6. Number of search periods and the number of encounters in 30 minutes of observation of Nemeritis canescens. After 15 min. After 3 hrs 15 min. After 6 hrs 15 min. After 9 hrs 15 min. Wasp density Number of Number of Number of Number of Encounters Encounters Encounters Encounters search periods search periods search periods search periods 1 0 49 0 52 0 51 0 5 0 74 0 61 0 47 0000 23 0 33 0 0 0 43 1 0 28 0 3 0 37 67 0 55 0 73 0 0 0 0 0 56 0 60 0 11 0 0 54 0 66 '-a 2 0 55 3 33 1 30 00 20 0 51 10 79 4 42 27 0 63 16 58 3 55 2 2 47 0 16 0 17 r4 0 N 14 67 6 74 5 47 28 N 3 55 16 92 10 94 42 1 50 4 58 r 4 21 85 19 62 10 52 . 13 N 17 56 22 59 6 30 CN 12 I 13 69 29 72 12 43 00 5 11 45 18 49 11 11 0 1 79 12 61 7 20 0 13 85 9 41 7 38 0 5 27 77 30 80 09 S£ 47L Z9 0 £Z ES 4781 8S ZZZ T6 L6 81 LL £17 7171 £5 Z£T 68 56 £S OTI 95 8£T S9 991 L8 Z8 L 56 LZ 7ZI 89 coZ 8L L6 8Z Z8 9T SL 05 5171 OS 817 9 I£ £Z £E1 65 ZOZ £S 6Z Z£ EV 017 LI IC *711 85 1,17 86 IS 6Z SI Z£ L£ 69 TOT OL 6Z 0 ZZ 917 617 95 09 67 ZT OZ Z£ S17 99 59 L£ ET £T 9£ 09 £17 89 69 £Z ZI S£ zz 9L OS 69 68 £9 91 LS 0£ £L TZ 6L SZ 81 S 8£ 91 9L £17 09 LZ 91 171 0£ Z£ 8L TS Z9 SI 0 I L S ZL 0£ Z9 9T 917 I£ 8£ 117 89 £5 99 LZ 9£ 6 L£ 47£ 9L 09 S£ TZ LE 91 817 9Z 17L Z17 OS £1 8 359 Table B4.7-11. Overall means of time spent probing, cocking, number of search periods, mean length search periods, encounters between wasps, in Nemeritis canescens, as experiment progresses. Individual replicates can be found in tables B4.1-6. Overall means were calculated as described in Bliss (1967), weighted when necessary. Overall means, 95% confidence limits Time after Number experiment start Time Mean length Cocks Encounters probing probe bout probe bouts After 15 min. 30.50 56.09 16.74 62.82 8.55 62.98 59.52 19.94 71.70 10.24 -.1.98 52.65 13.54 53.93 6.86 After 3 hrs 15 min. 65.57 22.91 7.18 56.80 2.30 150.45 26.90 8.51 68.09 3.03 -19.31 18.93 5.86 45.52 1.58 After 6 hrs 15 min. 44.57 12.04 6.38 34.44 1.03 110.15 14.19 7.42 39.99 1.53 -21.02 9.90 5.34 28.90 0.53 After 9 hrs 15 min. 21.00 5.59 6.54 16.31 0.47 55.84 7.92 8.13 23.44 0.89 -13.83 3.25 4.94 9 .17 0.05 Results from all parasitoid densities. 360 Table B4.12-19. Percentage time occuppied by walking and resting, mean length and number of walking and resting periods, in 30 minutes observation of Nemeritis canescens, 15 minutes after experiment start. Wasp 7 Time Mean Number % Time Mean Number Mean, 95% density walking walk (sec) walks resting rest (sec) rests C.L. walk bouts 1 • 24.17 6.30 69 31.50 15.32 37 5.45 33.50 6.22 97 13.17 7.18 33 6.24 15.54 5.10 50 21.43 7.98 44 4.66 10.16 4.76 41 35.47 19.46 35 15.23 4.03 64 12.57 10.65 20 21.57 5.39 69 34.61 14.21 42 20.44 6.32 57 4.94 6.69 13 31.09 6.14 88 18.65 8.10 40 2 35.14 9.28 65 18.88 18.00 18 6.87 23.44 5.06 80 13.54 8.07 29 8.70 28.72 7.32 68 10.55 4.69 39 5.05 26.07 5.85 78 7.72 8.44 16 35.71 8.79 75 13.90 11.59 22 15.65 4.93 56 3.23 9.50 6 4 26.88 5.61 84 14.38 5.86 43 6.66 45.25 8.17 98 12.88 5.70 40 7.89 20.34 6.10 59 12.54 11.68 19 5.44 35.10 6.17 103 8.28 4.84 31 33.01 6.00 96 6.02 6.18 17 32.47 5.97 95 5.84 4.64 22 45.55 8.93 91 6.89 6.83 18 8 43.30 10.96 69 31.96 12.98 43 6.97 35.02 8.23 74 6.38 6.17 18 8.89 27.21 6.49 74 6.80 7.06 17 5.04 25.13 6.08 74 7.54 4.66 29 33.10 6.94 83 4.14 5.14 14 27.29 5.18 91 11.82 6.80 30 19.83 4.88 72 7.80 5.31 26 16 46.61 7.66 105 9.23 6.91 23 8.61 23.06 5.47 74 7.18 7.00 18 12.26 37.59 8.48 75 4.26 4.50 16 4.96 61.33 . 16.62 65 5.28 8.46 11 33.79 6.42 92 14.07 7.94 31 44.00 10.29 77 12.33 7.66 29 15.58 5.32 53 21.06 15.88 24 361 Table B4.12-19. Continued. Wasp % Time Mean Number % Time Mean Number Mean, 195% density walking walk (sec) walks resting rest (sec) rests C.L. walk bouts • 32 25.46 6.04 76 28.95 15.82 33 .6.88 48.81 10.50 82 4.25 7.50 10 8.88 26.33 4.78 98 8.61 6.96 22 4.88 36.66 6.20 105 9.80 5.80 30 32.72 5.62 102 5.31 4.90 19 54.38 10.83 92 8.68 5.89 27 29.32 8.11 85 12.71 7.26 31 C.L. = confidence limits Table B.4.20-21. Percentage time spent probing, mean length probe session, number of probe sessions in 30 minutes of observation of Nemeritis canescens. After 15 min. After 3 hrs 15 min. After 6 hrs 15 min. After 9 hrs 15 min. Wasp density Number of Mean Number of Number of Mean p % Tp Mean Mean Number of 7 T search periods probe search periods % Tp Tp search probe search periods probe search periods period 1 43.33 49 15.92 37.06 52 12.83 25.39 51 8.96 2.17 5 7.80 54.67 74 13.30 19.94 61 5.89 12.78 47 4.89 7.94 23 6.22 63.00 33 31.27 0.00 0 - 12.17 43 5.09 0.17 1 3.00 54.38 28 37.29 0.83 3 5.00 14.06 37 6.84 18.56 67 4.99 72.20 55 22.24 55.33 73 13.64 0.00 0 - 0.00 0 - 43.83 56 13.50 26.78 60 8.03 10.00 11 16.36 0.00 0 - 74.62 54 24.33 50.26 66 13.23 2 60.27 55 19.53 16.17 33 8.82 11.39 30 6.83 10.78 20 9.70 45.98. 51 15.47 31.67 79 7.22 12.33 42 5.29 12.78 27 8.52 50.42 63 14.33 24.06 58 7.47 21.98 55 7.13 1.33 2 12.00 60.73 47 22.40 5.83 16 6.56 7.33 17 7.76 0.00 0 - 66.21 67 17.28 53.28 74 12.96 16.44 47 6.30 10.44 28 6.71 50.38 55 16.80 41.83 92 8.18 21.39 74 5.23 16.28 42 6.98 81.12 50 28.60 63.02 58 18.73 4 62.04 85 13.09 33.44 62 9.71 24.67 52 8.54 4.72 13 6.54 58.73 56 18.36 15.39 59 4.69 8.00 30 4.80 4.00 12 6.00 41.86 69 10.74 29.83 72 7.46 14.98 43 6.19 1.33 5 4.80 67.12 45 26.40 17.67 49 6.49 1.83 1.1 3.00 0.00 0 56.62 79 12.99 19.22 61 5.67 7.39 20 6.65 0.00 0 61.03 85 12.53 15.61 41 6.85 11.44 38 5.42 3.17 5 11.4 61.68 77 13.99 47.56 80 10.61 8 40.17 50 13.94 27.44 74 6.68 22.39 48 8.4 11.39 37 5.54 24.91 35 12.34 27.94 76 6.62 12.39 37 6.03 10.22 36 5.11 58.60 68 14.99 18.56 68 4.91 7.78 38 3.68 11.28 46 4.41 67.92 62 19.16 21.78 72 5.44 2.06 7 5.29 0.00 0 65.99 62 18.77 26.94 78 6.22 9.06 30 5.43 4.72 16 5.31 67.34 60 19.77 23.78 76 5.63 10.39 38 4.92 6.56 18 6.56 62.76 79 13.82 60.89 73 14.40 72.37 57 22.47 16 44.17 89 8.56 16.44 50 5.92 4.44 22 3.64 4.06 12 6.08 69.76 69 17.75 17.22 43 7.21 11.56 36 5.78 3.61 13 5.00 58.16 65 15.14 17.56 45 7.02 6.78 20 6.10 0.72 1 13.00 33.39 60 9.80 15.56 49 5.71 10.28 22 8.41 0.00 0 50.67 70 13.03 19.00 69 4.96 9.33 32 5.25 9.61 29 5.97 45.83 51 16.18 17.28 47 6.62 13.67 44 5.59 6.44 17 6.83 63.35 43 26.65 32 43.45 53 15.17 18.22 59 5.56 6.56 23 5.13 1.17 6 3.50 45.59 50 16.44 16.33 50 5.88 6.17 16 6.94 7.39 28 4.75 46.94 78 10.62 31.33 68 8.29 9.17 27 6.11 1.89 7 4.86 65.06 87 13.29 24.83 65 6.88 21.33 56 6.86 17.78 53 6.04 53.55 89 10.69 21.61 53 7.34 19.94 43 8.35 5.67 18 5.67 61.97 91 12.03 19.06 58 5.91 17.17 53 5.83 0.00 0 36.94 74 9.14 57.97 60 17.10 364 Table B4.23. The decline in the percentage of changes following encounters as parasitoid density increases in Nemeritis canescens. 95% Confidence Number Wasp Total limits followed 7 Change density encounters by change Upper Lower 2 24 14 58.33 94.49 22.17 4 109 31 28.44 50.57 6.31 8 166 72 43.37 50.08 35.46 16 283 101 35.69 44.67 30.55 32 537 k 172 32.02 36.01 28.09 Confidence limits calculated using table w in Rohlf and Sokal (1969). Table B4.24. The decline in the percentage of changes following encounters during rest as the average encounters per wasp in rest increase, in Neneritis canescens. Average 95% Confidence Number Wasp encounters Total limits followed % Change density per wasp encounters by change in rest Upper Lower 2 1.20 6 3 50.00 78.93 12.22 4 1.40 10 8 80.00 97.48 44.40 8 3.43 24 14 58.33 77.33 39.33 16 2.14 15 12 80.00 95.67 51.93 32 5.86 41 18 43.90 72.29 21.56 Confidence limits calculated using table w in Rohlf and Sokal (1969). Average encounters per wasp in rest calculated only from replicates in which encounters occurred. 365 Table B4.27-30. Number of encounters in walk, rest and probe with the number of walk and probe sessions, and the mean lengths of walk and probe sessions in 30 minutes observations of Nemeritis canescens. Number Number Mean Mean Wasp Encounters Encounters Encounters of of lengths lengths density in walk in probe in rest walk probe walk probe sessions sessions sessions sessions 2 0 1 1 68 47 7.32 22.40 6 5 3. 78 67 5.85 17.28 0 2 1 75 55 8.79 16.80 0 1 0 56 50 4.93 28.62 0 3 1 80 58 5.06 18.78 0 0 0 65 51 9.28 15.47 0 0 0 - 55 - 19.53 0 0 0 - 63 - 14.33 4 3 11 3 84 56 5.61 18.36 6 4 0 98 69 8.14 10.74 8 21 1 91 80 8.93 10.61 5 3 3 59 45 6.10 26.40 1 0 0 103 79 6.17 12.99 4 8 1 96 85 6.00 12.53 9 16 2 95 77 5.97 13.99 8 9 6 6 69 35 10.96 12.34 2 12 1 74 62 6.49 18.77 5 21 2 74 61 6.08 19.77 11 12 2 83 79 6.94 13.82 7 8 6 91 73 5.18 14.40 2 22 5 72 57 4.88 22.47 10 15 2 74 68 8.23 14.99 16 28 33 2 105 89 7.66 8.56 6 14 3 74 69 5.47 17.75 13 23 1 75 65 8.48 15.14 36 11 1 65 60 16.62 9.80 10 17 2 92 70 6.42 13.03 15 24 4 77 51 10.29 16.18 7 31 2 53 43 5.32 26.65 32 10 29 9 76 50 6.04 16.44 52 41 4 82 78 10.50 10.62 20 57 6 98 87 4.78 13.29 39 51 5 105 89 6.20 10.69 45 47 5 102 91 5.62 12.03 46 29 6 92 74 10.83 9.14 14 16 6 85 60 6.11 17.10 366 Table B4.27-30. Continued. Number Number Mean Mean Wasp Encounters Encounters Encounters of of lengths lengths density in walk in probe in rest walk probe walk probe sessions sessions sessions sessions 1 0 0 0 69 49 6.30 15.92 0 0 0 97 74 6.22 13.30 0 0 0 50 33 5.10 31.27 0 0 0 41 28 4.76 37.29 0 0 0 64 55 4.03 22.24 0 0 0 69 56 5.39 13.50 0 0 0 57 54 6.32 24.33 0 0 0 88 66 6.14 13.23 367 Table B4.31. Number of cocking movements and the percentage time spent probing in 30 minutes of observation of Nemeritis canescens. After After After After 15 wins 15 min. 6 hrs 15 min. 9 hrs 15 min. Wasp 3 hrs density % Time % Time % Time % Time Cocking Cocking Cocking Cocking probing probing probing probing 1 43.33 15 37.06 3 25.39 2 2.17 0 54.67 10 19.94 1 12.78 1• 7.94 1 63.00 13 0.00 0 12.17 2 0.17 1 54.38 10 0.83 1 14.06 2 18.56 1 72.20 3 55.33 7 0.00 0 0.00 0 43.83 13 26.78 2 10.00 1 0.00 0 74.62 5 50.26 7 2 60.27 3 16.17 4 11.39 0 1.33 0 45.98 7 31.67 2 12.33 0 0.00 0 50.42 10 24.06 5 21.78 0 10.44 1 60.73 16 5.83 0 7.33 0 16.28 1 66.21 7 53.28 8 16.44 5 10.78 0 50.38 11 41.83 0 21.39 2 12.78 1 81.12 0 63.02 .11 4 62.04 11 33.44 2 24.67 1 4.72 1 58.73 11 15.39 2 8.00 1 4.00 0 41.86 15 29.83 4 14.78 5 1.33 0 67.12 4 17.67 0 1.83 1 0.00 0 56.62 13 19.22 3 7.39 1 0.00 0 61.03 6 15.61 2 11.44 1 3.17 0 61.68 11 47.57 11 8 40.17 7 27.44 2 22.39 2 11.39 3 24.91 3 27.94 3 12.39 1 10.22 1 58.60 12 18.56 5 7.78 0 11.28 2 67.92 8 21.78 3 2.06 1 0.00 0 65.99 6 26.94 3 9.06 0 4.72 0 67.34 8 23.78 2 10.39 2 6.56 1 62.76 4 60.89 5 72.37 3 16 44.17 10 16.44 4 4.44 0 4.06 0 69.76 8 17.22 1 11.56 0 3.61 0 58.16 9 17.56 4 6.78 1 0.72 1 33.39 4 15.56 1 10.28 0 0.00 0 50.67 14 19.00 1 9.33 0 9.61 2 45.83 3 17.28 0 13.67 1 6.44 0 63.35 4 368 Table 84.31. Continued. After After After After 15 mins 3 hrs 15 min. 6 hrs 15 min. 9 hrs 15 min. Wasp density % Time % Time % Time % Time Cocking Cocking Cocking Cocking probing probing probing probing 32 43.45 10 18.22 1 6.56 1 1.17 0 45.59 5 16.33 1 6.17 2 7.39 0 46.94 4 31.33 2 9.17 0 1.89 65.06 13 24.83 2 21.33 1 17.78 0 53.55 16 21.61 1 19.94 0 5.67 61.97 8 19.06 1 17.17 0 0.00 0 36.94 9 57.97 14 369 Table B4.32. Total offspring recovered per parasitoid, for Nemeritis canescens (24 hours). Wasp Total Mean 95% Confidence limits density offspring offspring recovered per wasp Upper Lower 1 18 30.38 40.20 20.55 35 11 48 28 30 40 33 2 15 19.06 27.62 10.46 63 12 27 32 68 46 42 4 19 11.44 14.72 8.17 36 50 64 73 36 30 51 53 8 45 5.40 6.94 3.87 26 35 65 59 30 6 77 65 38 48 49 19 . 370 Table B4.32. Continued. Total Mean 95% Confidence limits Wasp offspring offspring density recovered per wasp Upper Lower 16 45 3.65 4.65 2.64 23 54 65 58 53 52 98 77 32 64 2.12 2.47 1.78 65 76 75 42 86 74 61 371 Table B6.1. The decline of the percentage females in Diaeretiella off- spring as the density of Aphidius matricariae present increases. 957 Confidence Diaeretiella % females Aphidius Female limits density offspring record d offspring Upper Lower 0 166 104 62.65 69.90 54.93 I 98 45 45.92 56.28 35.87 2 170 107 62.94 70.11 55.31 4 99 53 53.54 68.17 43.29 8 147 69 46.94 55.22 38.82 16 110 40 36.36 45.99 27.50 The confidence limits for percentages were calculated from table W in Rohlf and Sokal (1969). 372 Table B6.2. Log offspring per Nemeritis with varying Bracon hebetor density. 95% Confidence limits Log (Bracon density) Log offspring/Nemeritis + 1 Upper Lower 0.000 1.672 1.697 1.645 0.301 1.377 1.470 1.258 0.477 1.331 1.421 1.217 0.699 1.008 1.168 0.>±3 0.954 0.809 1.045 0.250 1.230 0.537 0.758 0.064 373 Table B6.3. The searching efficiency (a) of Nemeritis canescens, as Bracon hebetor density rises. Estimate 1 Estimate 2 Bracon 95% Confidence 95% Confidence density limits limits a Mean a a Mean a Upper Lower Upper Lower 0 0.764 0.912 1.142 0.683 as 0.912 1.142 0.683 1.006 column 0.881 2 1.172 0.574 1.378 0.920 0.603 1 0.347 0.278 0.345 0.212 0.521 0.569 0.692 0.446 0.259 0.321 0.283 0.623 0.196 0.656 0.141 0.453 0.351 0.600 0.278 0.819 0.370 0.559 2 0.234 0.316 0.389 0.243 0.602 0.548 0.701 0.395 0.367 0.433 0.350 0.424 0.124 0.197 0.299 0.419 0.117 0.788 0.259 0.424 0.283 0.422 0.308 0.464 0.443 0.599 0.558 0.644 0.405 0.460 0.358 1.243 4 0.169 0.127 0.186 0.069 0.509 0.354 0.603 0.106 0.134 0.987 0.094 0.103 0.234 0.420 0.201 0.305 0.024 0.048 0.076 0.242 0.086 0.221 374 Table B6.3. Continued. Estimate 1 Estimate 2 Bracon 95% Confidence 95% Confidence limits limits density a Mean a a Mean a Upper Lower Upper Lower 8 0.173 0.073 0.130 0.017 0.684 0.437 0.727 0.146 0.012 0.037 0.119 1.040 0.122 0.519 0.012 0.032 0.049 0.549 0.041 0.163 0.107 0.471 16 0.051 0.059 0.078 0.041 1.283 0.709 0.984 0.434 0.071 0.441 0.020 0.896 0.041 0.733 0.068 0.424 0.091 0.433 0.054 0.347 0.043 1.243 0.093 0.582 See text for methods of estimation. 375 Table B6.4. Efficiency of search for healthy hosts (a) in Brawn hebetor, with the increasing density of Naneritis canescens. Nerneritis Log a Nemeritis Log a density (Bracōn) density (Bracōn) 0 0.143 8 0.217 0.205 0.193 0.519 0.233 0.151 0.176 0.160 0.016 0.135 0.147 0.068 0.232 0.363 0.180 0.308 0.191 0.432 16 0.278 1 0.141 0.132 0.178 0.248 0.253 0.283 0.041 0.147 0.071 0.142 0.044 0.126 0.534 0.088 0.336 2 0.313 0.054 32 0.251 0.063 0.273 0.197 0.576 0.115 0.291 0.667 0.410 0.174 0.637 0.139 0.321 0.136 0.086 0.037 0.041 0.292 4 0.099 0.060 0.195 0.146 0.149 0.199 0.707 0.316 0.282 376 Table B6.5. The efficiency Of search for paralysed hosts (s) in Bracon hebetor as Nemeritis canescens density rises. Nemeritis 95% Confidence 95% Confidence density s Mean s levels Log s levels Upper Lower Upper Lower 0 0.038 0.037 0.062 0.012 -1.432 -1.206 -1.933 0.065 0.094 0.053 0.000 0.000 0.045 0.000 1 0.032 0.041 0.083 -0.002 -1.390 -1.080 - 0.045 0.021 0.000 0.134 0.053 0.000 2 0.080 0.071 0.106 0.036 -1.147 -0.973 -1.443 0.243 0.032 0.012 0.080 0.118 0.049 0.049 0.072 0.053 0.059 0.053 0.027 4 0.000 0.032 0.062 0.003 -1.491 -1.209 -2.561 0.121 0.025 0.030 0.030 0.024 0.000 0.000 0.060 377 Table B6.5. Continued. 95% Confidence 95% Confidence levels levels Nemeritis Mean s Log s density -s - - Upper Upper Lower 8 0.012 0.038 0.065 0.012 -1.416 -1.189 -1.922 0.038 0.067 0.027 0.000 0.016 0.012 0.102 0.072 16 0.038 0.020 0.036 0.004 -1.702 -1.443 -2.438 0.018 0.020 0.011 0.067 0.018 0.000 0.000 0.008 32 0.023 0.029 0.048 0.009 -1.540 -1.316 -2.030 0.009 0.033 0.051 0.027 0.000 0.059 378 Table B6.6. Hosts paralysed per Bracon hebetor, as Nemeritis canescens density increases. Hosts Hosts Nemeritis Nemeritis paralysed paralysed density density per Bracon per Bracon 0 15.87 8 22.53 21.50 20.50 41.32 23.81 16.63 19.00 17.50 2.00 15.12 16.26 8.14 23.79 33.04 20.28 29.46 27.00 16 27.29 14.81 1 15.75 25.00 19.15 27.68 25.39 16.25 5.04 15.86 8.43 14.22 5.38 10.29 41.98 31.29 2 29.80 32 25.25 6.55 26.95 7.59 43.79 20.85 28.20 13.19 35.82 47.13 46.08 18.79 30.32 15.52 15.24 10.16 4.61 5.00 28.28 4 11.50 7.23 20.66 16.24 16.48 21.00 48.43 29.98 27.58 379 Table B6.7. Average Bracon pupae recovered per parasitized host with Nerneritis canescens density. Bracon Nemeritis Mean Bracon Nemeritis Mean Mean, C.L. Mean, C.L. density eggs/host density eggs/host CONTROL 3.00 2.44 16 2.00 2.04 (No 1.33 3.17 4.00 2.91 Nemeritis) 2.25 1.71 1.50 1.17 2.33 2.00 3.33 1.75 2.40 2.00 1.00 1 7.00 4.12 2.33 6.96 32 4.50 2.52 3.50 1.27 2.40 4.03 1.75 2.60 1.00 6.00 1.75 1.33 2 2.00 2.55 2.25 3.41 1.50 1.69 1.00 1.00 3.00 2.44 3.80 3.00 1.00 2.00 3.35 2.67 6.67 4 2.00 2.08 1.50 2.48 2.00 1.69 2.50 2.50 2.00 8 2.00 1.98 2.00 2.63 2.17 1.33 2.50 1.00 3.00 1.17 C.L. = 95% confidence limits 380 Table B6.8. The decline of the percentage females in the offspring of Bracon hebetor as the density of Nemeritis canescens rises. 95% Confidence Nemeritis Bracon Female limits density offspring offspring Females Upper Lower. 0 52 33 63.46 76.24 49.01 1 40 19 47.50 63.75 31.71 2 149 75 50.34 58.48 42.16 4 31 14 45.16 63.54 27.44 8 48 22 45.83 60.82 31.48 16 27 15 55.56 74.25 36.40 32 39 13 33.33 1 1 95% confidence limits calculated from table W in Rohlf and Sokal (1969). 381 Table 86.9. Offspring produced per Nemeritis as Nemeritis density increases in the presence of two Bracon hebetor females. Offspring Nemeritis per Mean, 95% C.L. Log mean, offspring/Nemeritis 95% C.L. density Nemeritis 1 13.75 35.68 1.55 42.34 55.00 1.74 41.47 16.36 1.21 31.24 41.67 70.99 8.19 2 21.01 21.61 1.33 28.22 26.33 1.42 29.39 16.88 1.23 8.73 15.63 13.38 19.38 25.21 27.43 32.51 9.22 23.50 27.29 4 22.00 15.53 1.19 25.29 20.96 1.32 17.13 10.10 1.00 16.96 21.09 16.25 4.61 8.36 8.07 8 7.68 8.54 0.93 7.00 9.93 1.00 8.31 7.14 0.85 8.00 12.63 6.43 8.07 9.02 9.69 382 Table B6.9. Continued. ring Nemeritis Offsp Mean, 95% C.L. Log mean, per density offspring/Nerneritis 95% C.L. Nemeritis 16 2.85 3.22 0.51 3.83 3.78 0.58 2.88 2.65 2.31 1.59 3.43 3.90 3.56 3.07 3.85 32 0.56 0.62 -0.21 0.48 0.91 -0.04 0.14 0.34 -0.47 0.88 0.54 0.64 1.12 2 46.64 48.34 1.68 (absence 50.39 51.65 1.71 Bracon 48.50 45.04 1.65 correction 52.22 factor) 41.65 53.53 49.27 44.54 C.L. = confidence limits N.B. In many cases a few host larvae (generally not more than five) were missing at the final count; a linear correction was used to estimate the offspring from 128 larvae. 383 Table 86.10. Searching efficiency (a) of Nemeritis canescens, in the presence of two Brawn hebetor. Estimate 1 I Estimate 2 Neuritis density Mean, Log mean, Mean, a a Log mean, C.L. C.L. C.L. C.L. 1 0.437 0.466 -0.331 0.634 0.633 -0.165 0.445 0.743 -0.129 0.718 0.956 -0.019 0.330 0.190 -0.722 0.627 0.410 -0.387 0.312 0.343 0.503 0.668 1.090 1.290 0.146 0.503 2 0.234 0.316 -0.501 0.602 0.548 -0.262 0.367 0.389 -0.410 0.433 0.701 -0.155 0.350 0.243 -0.615 0.424 0.394 -0.404 0.124 0.197 0.299 0.419 0.117 0.788 0.259 0.424 0.283 0.422 0.308 0.463 0.443 0.599 0.557 0.644 0.405 0.460 0.358 1.243 4 0.355 0.260 -0.585 0.644 0.683 -0.166 0.497 0.359 -0.445 0.901 0.833 -0.080 0.271 0.161 -0.793 0.940 0.533 -0.274 0.315 0.805 0.280 0.597 0.256 0.767 0.058 0.477 0.171 0.673 0.137 0.340 8 0.125 0.146 -0.835 0.463 0.438 -0.359 0.131 0.175 -0.757 0.385 0.512 -0.291 0.116 0.117 -0.931 0.412 0.363 -0.440 0.136 0.361 0.238 0.264 0.163 0.477 0.118 0.446 0.147 0.560 0.141 0.572 384 Table B6.10. Continued. Estimate 1 Estimate 2 Nemeritis density Mean, Log mean, Mean, a a Log mean, C.L. C.L. C.L. C.L. 16 0.042 0.063 -1.201 0.119 0.151 -0.820 0.083 0.080 -1.098 0.197 0.181 -0.743 0.048 0.046 -1.336 0.135 0.122 -0.915 0.047 0.172 0.078 0.197 0.075 0.168 0.075 0.142 0.091 0.154 0.028 0.077 32 0.021 0.016 -1.788 0.052 0.059 -1.232 0.019 0.022 -1.661 0.047 0.082 0.006 0.011 -1.969 0.027 0.035 -1.454 0.022 0.070 0.017 0.093 0.010 0.019 0.063 2 0.764 0.912 1.142 0.683 (absence 1.006 Bracon 0.881 correction 1.172 point) 0.574 1.378 0.920 0.603 C.L. = 95% confidence limits. 385 Table B6.11. The hosts paralysed per Bracon hebetor, in the presence and absence of Nemeritis canescens, as Bracon density increases. In the absence of Nemeritis In the presence of Nemeritis Bracon density Hosts Log mean, Hosts Log mean, paralysed Mean, C.L. C.L. paralysed Mean, C.L. C.L. per 2 per 2 1 50.00 30.91 1.49 30.00 46.25 1.67 22.00 39.08 1.59 20.00 63.05 1.80 23.00 22.74 1.36 52.00 29.45 1.47 38.00 65.00 39.00 74.00 34.00 36.00 14.00 64.00 13.00 42.00 25.00 40.00 2 29.80 17.13 1.23 15.50 22.25 1.35 6.55 24.49 1.39 21.50 29.35 1.47 7.59 9.77 0.99 41.00 15.15 1.18 20.85 16.50 13.19 17.50 47.13 15.00 18.79 7.50 15.52 32.00 15.24 29.00 10.16 27.00 4.61 5.00 28.28 4 18.50 14.78 1.17 12.75 14.39 1.16 24.00 20.55 1.31 12.00 17.62 1.25 2.50 9.01 0.95 20.50 11.16 1.05 11.50 11.75 8.75 21.25 15.25 17.25 20.00 12.50 17.75 9.50 12.00 8 9.88 10.94 1.04 7.63 10.11 1.00 7.00 12.68 1.10 13.25 12.73 1.10 12.50 9.20 0.96 8.75 7.49 0.87 11.88 5.63 9.75 13.13 13.88 7.38 11.50 13.50 11.13 11.50 386 Table B6.11. Continued. In the absence of Nemeritis In the presence of Neneritis Bracon density . Hosts Log mean, Hosts Log mean, paralysed Mean, C.L. paralysed Mean, C.L. C.L. C.L. per per + 16 6.88 6.71 0.83 6.38 5.86 0.77 6.19 7.21 0.86 4.63 7.06 0.85 7.56 6.21 0.79 0.44 4.66 0.67 7.13 4.69 6.13 6.38 5.88 5.06 6.94 7.56 7.56 8.00 6.13 C.L. = 95% confidence limits 387 Table B6.12. The efficiency of search for healthy hosts (a) in Bracon hebetor, in the presence of Nemeritis canescens, as Bracon density increases. Bracon Bracon density Log a density Log a 1 -0.588 16 -0.854 -0.799 -1.032 -0.300 -0.719 -0.091 -0.848 -0.053 -1.027 -0.485 -1.108 -0.197 -1.004 -0.599 -0.824 -1.538 2 -0.504 -1.268 -1.201 -0.706 -0.939 -0.176 -0.759 -0.857 -0.866 -1.066 -1.432 -1.387 -0.535 4 -0.699 -0.488 -1.721 -0.979 -1.092 -0.780 -0.604 -0.688 8 -0.932 -0.863 -0.752 -0.866 -0.924 -0.609 -0.799 -0.842 388 Table B6.13-14. Hosts parasitized per Bracon female, and offspring per female with Bracon hebetor density, in the presence of two Nemeritis canescens. Hosts Bracon Pupae produced Mean, C.L. parasitized Mean, C.L. per female density per female 1 7.00 2.88 18.00 6.13 1.00 5.38 2.00 11.42 2.00 0.37 7.00 0.83 8.00 13.00 3.00 5.00 1.00 1.00 1.00 3.00 0.00 0.00 2 1.00 2.14 2.00 6.68 2.00 3.49 4.50 11.51 1.00 0.80 1.50 1.85 0.50 0.50 0.50 0.50 1.50 4.50 1.50 10.00 1.50 4.00 8.50 28.50 2.00 4.00 0.50 0.50 0.50 1.50 3.00 11.00 6.00 20.50 4 1.75 1.19 6.25 3.34 1.00 1.73 2.50 5.24 2.00 0.64 6.25 1.45 0.75 3.00 1.75 5.00 1.25 1.75 1.00 2.00 0.00 0.00 8 2.63 1.78 3.88 4.95 1.75 2.71 3.75 8.34 2.75 0.85 10.38 1.56 3.13 9.25 1.88 9.00 1.75 2.75 0.38 0.63 0.00 0.00 38 9 Table B6.13-14. Continued. Brawn Hosts Pupae produced parasitized Mean, C.L. Mean, C.L. density per female per female 16 1.13 2.01 4.63 6.33 2.06 2.83 4.44 8.92- 1.31 1.20 3.00 3.73 3.75 10.25 1.81 6.56 1.69 6.00 0.5 1.88 3.56 12.31 2.25 7.88 390 Table 86.15. The efficiency of search for paralysed hosts (s) in Bracon hebetor, with Bracon density, in the presence of two Nemeritis canescens. Bracon Bracon Log s Log s density density 1 -0.575 16 -1.959 -1.293 -1.602 -1.420 -1.921 -0.883 -1.328 -1.387 -1.658 -1.553 -1.678 -1.456 -2.222 -1.398 2 -1.097 -1.538 -0.614 -1.495 -1.921 -1.097 -0.928 -1.310 -1.310 -1.143 -1.276 -1.229 -1.276 -1.569 4 -1.585 -1.959 -1.319 -1.658 -1.523 -1.796 -1.398 8 -1.409 -1.444 -1.509 -1.420 -1.569 -1.770 -2.398 391 Table B7.1. Mean probe length shown by Nemeritis canescens with Bracon hebetor density. Nemeritis Bracon density mean probe length (sec) 4 10.71 6.43 15.99 8.88 6.69 7.77 8.51 8.86 8 9.92 9.02 6.92 16.50 6.04 7.47 9.38 7.52 16 38.11 12.82 13.36 10.10 9.09 13.98 15.05 8.75 392 Table B7.2. Total encounters in probe (with Bracon and other Nemeritis) of 1 Nemeritis, with Bracon density. Bracon density Total encounters in probe 4 1 5 0 1 1 6 1 2 8 4 0 2 1 3 7 11 0 16 8 1 3 3 10 1 16 9 393 Table B7.3. Number of cocking movements, and percentage time spent probing in 1 hr. observation of Nemeritis canescens in the presence of Bracon hebetor. Number Time Bracon density of cocking probing movements 2 10.50 12 32.94 49 31.55 8 19.31 35 6.47 15 28.24 40 9.32 14 10.67 23 '4 3.52 2 34.60 41 26.68 35 15.06 18 23.78 23 39.98 53 44.25 44 15.66 21 8 10.44 53 32.36 44 11.11 6 38.88 42 36.17 50 11.01 12 49.30 43 30.85 30 3 94 Table B7.4-6. Percentage time spent probing, walking and resting, with the number of probe and walk bouts, and the mean length of rest sessions, in one hour's observation of Neweritis canescens in the presence of Brawn hebetor. 7 Time spent Number of % Time spent Number of % Time spent Mean length probing probes walking walks resting rests (sec) 10.50 35 15.69 97 70.25 34.83 32.94 148 22.64 205. 44.42 13.46 31.55 73 24.98 124 43.47 29.24 19.31 78 25.17 149 55.52 19.92 6.47 35 21.91 127 f 71.62 25.63 28.24 122 27.88 186 43.88 15.19 9.32 37 19.34 117 71.34 25.39 10.67 42 13.08 99 76.25 34.52 3.52 13 21.19 95 75.29 31.01 34.60 134 21.68 160 45.52 16.80 26.68 131 20.58 192 52.74 14.33 15.06 32 14.71 84 70.23 33.74 23.78 135 29.72 209 46.50 13.76 39.98 174 33.06 231 26.96 7.82 44.25 165 25.73 200 30.02 10.10 15.66 71 35.27 145 49.08 20.17 10.44 133 22.14 187 39.75 13.29 32.36 88 20.40 124 47.25 23.20 11.11 33 21.01 158. 67.88 19.38 38.88. 134 31.38 187 ., 29.74 11.63 36.17 136 29.41 200 34.42 12.38 11.01 28 13.21 105 75.78 30.17 49.30 133 17.96 169 32.74 12.54 30.85 122 19.15 162. 50.00 17.85 395 Table B7.7-8. Mean lengths of walk and rest bouts with the numbers of walk and rest bouts, in one hour's observation of Nemeritis canescens in the presence of Brawn hebetor. Mean walk Number of Mean rest Number of (sec) walks (sec) rests 7.08 97 34.83 72 4.19 205 13.46 125 7.45 124 29.24 55 6.06 149. 19.92 100 6.24 127 25.63 101 5.03 186 15.19 97 5.59 117 25.39 95 4.61 99 34.52 77 8.01 95 31.01 89 4.71 160 16.80 90 3.64 192 14.33 125 6.14 84 33.74 73 4.88 209 13.76 116 4.65 231 7.82 112 4.50 200 10.10 104 8.30 145 20.17 83 4.32 187 13.29 109 5.73 124 23.20 71 5.23 158 19.38 139 5.81 187 11.63 89 5.03 200 12.38 95 4.46 105 30.17 89 4.31 169 12.54 106 4.01 162 17.85 97 396 Table B7.9-11. Number of probe and walk bouts with the number of encounters with conspecifics in walk, rest and probe in one hour's observation of Nemeritis canescens in the presence of Bracon hebetor. Conspecific Conspecific Conspecific Number of Number of encounters encounters encounters Probes walks in walk in rest in probe 1 35 4 1 97 0 148 0 0 205 1 73 1 0 124 2 78 0 0 149 2 35 0 1 127 4 122 3 6 186 0 37 1 1 117 2 42 1 0 99 3 13 0 3 95 0 134 0 0 160 1 131 3 2 192 0 32 2 1 84 0 135 1 3 209 9 174 8 7 231 3 165 3 4 200 0 71 0 0 145 4 133 1 6 187 1 88 0 1 124 1 33 1 0 158 6 134 2 3 187 6 136 2 6 200 0 28 0 1 105 3 133 8 9 169 3 122 2 2 162 397 Table B7.14-16. Number of probe and walk bouts with all encounters (with Nemeritis and Bracon) in probe, walk and rest, in one hour's observation of Nerneritis canēscens in the presence of Brawn hebetor. Total Total Number of Number of Total encounters encounters encounters walks p robes in probe in walk in rest C` 1 97 1 35 4 „ 5 205 .4 148 1 . 0 124 -~ 73 1 ~ 1 149 t 78 3 s .t 1 127 35 0 '6 186 122 4 1 117 ON 37 1 2 99 42 5 r 4 95 10 13 1 0 160 134 1 2 192 131 4 1 84 0 32 2 3 209 0 135 1 7 231 9 174 8 0 11 200 165 5 O 0 145 71 0 . . .1* 8 187 C 133 1 NI 1 124 88 2 3 158 33 2 3 187 r 134 2 10 200 o 136 4 0 28 0 1 105 I.r 16 169 1o 133 8 9 162 \ 122 7 398 Table B7.17-18. The number of probe bouts, and the percentage of time spent probing, with the total encounters (with Nemeritis and Bracon) in rest and walk. Total encounters Number % Time spent in rest and walk of probes probing 6 35 10.50 2 148 32.94 2 73 31.55 7 78 19.31 4 35 6.47 8 122 28.24 1 37 9.32 7 42 10.67 4 13 3.52 1 134 34.60 5 131 26.68 2 32 15.06 1 135 23.78 18 174 39.98 14 165 44.25 0 71 15.66 5 133 10.44 4 88 32.36 3 33 11.11 9 134 38.88 10 136 36.17 0 28 11.01 13 133 49.30 16 122 30.85 399 Plate 1. Experimental cage used for competition between parasitoids of Myzus persicae. Plate 2. Experimental cage, and support housing a mirror, used for competition between parasitoids of PZodia interpuncteila. 400 i Plate 3. Butterdish container used to culture Bracon hebetor wasps. C Plate 4. Event recorder used to record the behaviour of Nemeritis canescens. 402 T f_ -r Plate 5. An example of a behaviour trace produced by the event recorder.