Parasites and social insects P Schmid-Hempel

To cite this version:

P Schmid-Hempel. Parasites and social insects. Apidologie, Springer Verlag, 1995, 26 (3), pp.255-271. ￿hal-00891263￿

HAL Id: hal-00891263 https://hal.archives-ouvertes.fr/hal-00891263 Submitted on 1 Jan 1995

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Review article

Parasites and social insects

P Schmid-Hempel

ETH Zürich, Experimental Ecology, ETH-Zentrum, NW, CH-8092 Zürich, Switzerland

(Received 23 December 1994; accepted 8 March 1995)

Summary — A short review shows that social insects have many parasites on larvae and adults but few on eggs. Social organisation simultaneously affects parasite transmission within and between colonies. Overall, the biology of social insects may be favourable to an epidemic, but such a process may quickly die out in the host population. social insects / parasite / polyandry / social organisation / epidemiology

INTRODUCTION the final host of the parasite (Moore, 1983). Such changes of host behaviour which increase the chances of transmission for Under natural conditions, organisms are the parasite seem not to be uncommon and exposed to a wide range of parasites that may have consequences for may threaten their survival and reproduc- larger-scale (see Dobson, 1988; tive success. Research over the past 2 population dynamics for discussion). decades has broadened our knowledge and demonstrated that the effects of parasitism Social insects as hosts are no different may often be quite subtle rather than obvi- from other organisms, but the investigation ous. For example, a number of studies have of their parasites has been largely neglected, documented subtle behavioural changes of with the exceptions of commercially impor- the infected hosts that often seem to favour tant species (eg, honey bees; Bailey and the parasite. Many species of acanto- Ball, 1991), pest species (eg, fire ants; Jou- cephalan parasites seem to rely on inducing venaz, 1983), and studies of social para- changes in host behaviour to facilitate trans- sitism and inquilines (mostly in ants and mission to the next host (Moore, 1984). An bees; eg, Lin, 1964; Fisher, 1988). For infected isopod which serves as an inter- example, similar observations on mediate host will become positively photo- behavioural changes have been made in tactic and stay in exposed areas more fre- the classic case of a parasite infecting a quently than their uninfected counterparts. social insect. The liver fluke (Dicrocoelium) As a consequence, the infected host is more infects wood ants as intermediate hosts. likely to be preyed upon by a bird which is The infected ant changes its behaviour so as to climb onto exposed parts of its habitat infection intensity often show a positive cor- (eg, grass tips), where the ant will normally relation with group size (Davies et al, 1991; stay until eaten by the final host of the fluke, Keymer and Read, 1991). Negative corre- a herbivorous mammal such as a sheep lations have been reported for some vector- (Hohorst and Graefe, 1961; Schneider and transmitted diseases (Poulin and Fitzgerald, Hohorst, 1971). 1989). A possible cause for this difference rate con- Parasites can have additional effects on may be related to the increased of hosts. Snails infected by the trematode tacts in large groups that facilitates the trans- Schistosoma (the causative agent of human fer of directly transmitted parasites, and, at bilharzia) change their life history by increas- the same time a dilution effect that reduces the rate of vector attacks ing fecundity early in life at the expense of per capita (’the later reproduction. Minchella and Loverde selfish herd’). Besides this, the spread of (1981) demonstrated that such altered an infection within a group is likely to pro- fecundity schedules benefit the host ceed more rapidly than the spread of the because parasitism leads to castration of infection between groups. Such sugges- tions have been made for human the snail at an older age. Parasites have smallpox in African households and also been implicated in the pattern of geo- (Becker Angulo, graphical distribution of species because 1981), a population that for many purposes is structured in similar to social insect parasites could act as weapons of compe- ways tition among their hosts (Price et al, 1986, populations. Furthermore, fitness effects of have been for animals 1988). In fact, some drastic cases of range parasitism reported reduction in birds and ungulates have been living in groups, eg, cliff swallows that have traced back to the action of parasites (Dob- high levels of ectoparasitism (Brown and son and Hudson, 1986). More recent work Brown, 1986). Such evidence strongly sug- has concentrated on the hypothesis that gests that we should be concerned about entire species communities may be affected parasitism in social insects, and that in fact by the presence of parasites (Dobson and social insects may be highly suitable study Crawley, 1994). Finally, parasites have also objects for investigating the interactions been assigned a prominent role for the evo- between parasitism and social organisation. lution of sexual selection and mate choice This paper discusses a somewhat arbitrary (Andersson, 1994), an area of interest which selection of issues on host-parasite inter- has grown explosively over the last years. actions in social insects. Parasites should be especially important for social this has species, although aspect PARASITES OF SOCIAL INSECTS often been neglected in discussions on the evolution of sociality. In social animals, par- asites can be contracted by many individ- Colonies of social insects should be a highly uals and subsequently may threaten many rewarding target for parasites, because they members of the social group. For this rea- offer both a locally high density of potential son, it has been suggested that character- hosts and, at the same time, an assemblage istics which at first sight seem quite dis- of hosts that (on a crude level) are geneti- parate, such as group aggression towards cally more similar to each other than the newcomers, territory maintenance, or group population average. In discussions of mod- size, may all have evolved to prevent novel ern epidemiology and evolutionary ecology infections from being introduced to the group it is usual to use the category of micropara- (Alexander, 1974; Freeland, 1976). Indeed, sites which refers to parasites whose the number of parasite species per host and dynamics can be described by referring to infected, uninfected (susceptible), or resis- whether this lack of egg parasites is real or tant hosts (’SIR models’, Anderson and May, reflects a lack of studies. Given the many 1981). Viruses, bacteria, and protozoa would taxa of parasitic wasps that specialise on all qualify for this category despite the many eggs of lepidoptera (Strand, 1986), this defi- differences in their biology. On the other ciency is remarkable and is possibly the hand, macroparasites are those parasites result of the widespread mode of brood care where the dynamics can be better described and the use of sheltered nest sites that may by looking at individual parasites and their make eggs difficult to utilise for most para- fate within hosts (’Sir model", Anderson and sitoids. Consequently, if present at all, the May, 1981). Helminths, such as trematodes most likely targets of egg parasites must be and nematodes, fall into this category. surface nesters, a nesting habit found in Finally, social insects are hosts to para- many wasps and some bees. sitoids and social I will not dis- parasites. Larval parasites, on the other hand, are cuss the case of subsocial which arthropods, quite common, but these often include infec- shed an extra on the of the may light origin tious parasites. This may reflect the fact that association between social insects and their infectious diseases can readily spread within parasites. a nest once infected. It is unclear, however, A brief survey of the literature will show why eggs are not attacked at the same time, that a wide variety of parasites of all cate- but essentially a similar absence of infec- gories are associated with social insects tious diseases of eggs is also true for non- (table I). As in other organisms, parasites social insects. Possibly, it is simply more can attack different life cycle stages of the difficult to penetrate the egg and its chorion. host (table II). In social insects, this would Another reason may be the fact that infec- include eggs, larvae, pupae and adults, and tions cannot be easily transmitted by the we may, for practical reasons, add the nest eggs since they are immobile and will not as such, for example, its supporting structure easily get in contact with new hosts, unless (eg, wax) or debris. In termites, which have the mode of brood care ensures transmis- a hemimetabolous development, different sion. In polydomous social insects eggs and nymphal stages could be the target of para- brood are often carried from one nest to the sites. Furthermore, parasites may infect a other and thus infections may easily spread particular stage, eg, the larva, but be trans- to different nests within the colony. mitted to the next host or a vector by A large number of parasites attack work- from the or adult. An another, eg, pupa ers. These include viruses, bacteria, fungi, example is the microsporidian Burenella protozoa, helminths, and parasitoids (table the fire ant dimorpha infecting Solenopsis. II). Workers, together with larvae, are thus Infection occurs in the fourth larval instar by the major target of parasitic infections. On the attendant nurse that workers provide the other hand, few cases seem to exist food, but parasitic spores are released and where only the sexuals, particularly the transmitted to these workers which act as queen but not the workers, are attacked. vectors in the et pupal stage (Jouvenaz al, The case of the nematode Sphaerularia 1981). bombi, which infects hibernating queens of In social insects, there seem to be few Bombus spp while they are staying in the cases in which the egg is attacked (table ground (Lundberg and Svensson, 1975), is II). In particular, only those parasites that special because it relies on the solitary are transmitted vertically, ie from mother to phase of the species. Such parasitism must offspring, via ovaria and eggs, seem to be be expected in similar cases of primitively described (eg, some viruses). It is not known eusocial species where the queens (or males) pass some time as solitary animals. GENETIC ASPECTS Although little is known about such taxa, the general pattern appears to be that parasites Variability in the susceptibility of hosts to will be able to attack all castes in the social infection and in the ability of parasites to phase of the species, as is the case for infect and establish in the host is based on Bombus (Schmid-Hempel et al, 1990). host-genotype vs parasite-genotype inter- actions in almost all cases analysed so far inheritance is polygenic and resistance or (eg, Wakelin, 1985). In a narrow sense, such infectibility is a quantitative trait (eg, Thomp- interactions are based on gene-for-gene son and Burdon, 1992; Schmid-Hempel and interactions (with host resistance genes Koella, 1994). Although less well investi- exactly matched by parasite virulence gated than other groups, this is also the genes; see Thompson and Burdon, 1992, case for social insects. Breeders of honey for discussion). However, more commonly bees have long taken advantage of this fact. Selection for increased resistance is, for powerful sting (Starr, 1985), or for life cycles instance, possible against the microspori- that favour biased sex ratios (Seger 1983) dian Nosema (Rinderer et al, 1983), acarine have all been identified as important fac- disease, hairless-black syndrome, Ameri- tors. can foulbrood and (Rothenbuhler, 1964b) Because of the apparent violation of pre- various other diseases (Kulincevic, 1986; dictions from kin selection theory, a num- Tanada and Kaya, 1993). Moreover, honey ber of hypotheses have been suggested to bees show natural variation in resistance explain the adaptive significance of (which is of course a prerequisite for selec- polyandry and polygyny. For example, tive breeding) against the mite Acarapis Crozier and Page (1985) hypothesised that woodi, foulbrood and microsporidia (Bailey increased genetic variability within the colony and Ball, 1991). may allow the expression of a better colony Genotypic variation as a determinant of phenotype. Genetic variability for task atten- host-parasite interactions has ramifications dance of workers has indeed been demon- for the discussion of the genetic make-up strated (eg, Robinson and Page, 1988). of colonies. Consider the widely accepted Polyandry would thus allow workers to hypothesis that the unusually close genetic respond to a wider and more variable set relatedness of female offspring in social of task-specific stimuli (Robinson, 1992). hymenoptera, as generated by the haplo- Polyandry could also reduce the amount of diploid sex determination system, facilitates genetic load that is associated with the pro- the evolution of social behaviour and altru- duction of less viable or sterile diploid males ism (Hamilton, 1964). Hence, it is expected (Page, 1980; Page and Metcalf, 1982). It that extant colonies of social insects are may also have evolved as a response to characterised by high degrees of average sperm limitation in the males (Cole, 1983), relatedness among workers. A large number although it remains unclear why under such of studies, however, have demonstrated circumstances males do not evolve larger that this is not universally the case and that sperm loads, as, for instance, in Apis where colonies are genetically quite variable (eg, single males are capable of filling the sper- Pamilo, 1981; Ward, 1983; Schwarz, 1986; matheca of a queen (Koeniger, 1991). Queller and Strassmann, 1988). The major Polyandry may also reduce the conflict over reason for this observation is that queens sex ratio of offspring between the queen often mate multiply with a number of males (that prefers an even sex ratio) and the (’polyandry’) as in the honey bee (Laidlaw workers (preferring a female bias) and Page, 1984), or that several functional (Boomsma and Grafen, 1991; Queller, queens are contributing to the worker force 1993). Polygyny has also been explained of the colony (’polygyny’), as in many ants by a number of adaptive hypotheses. These (Keller, 1993). In either case, the genetic include an unspecified advantageous variability within colonies is high by com- degree of genotypic variation at the colony parison to the total available variation in the level (Crozier and Consul, 1976), low prob- population and the average relatedness ability of independent colony foundation drops below the theoretically possible value (Pamilo, 1991), or queens being compara- of 3/4 for full-sibs of female hymenoptera. tively short-lived with respect to the colony, There are alternative views on the relevance such that supernumerary queens can act of genetic relatedness and kin selection for as replacements (Nonacs, 1988). the evolution of sociality. In fact, the need for On the other hand, and more generally, cooperative construction of nests (West- parasites have been repeatedly invoked as Eberhard, 1975), for nest defence by a selective agents that can maintain genetic variability. In particular, the production of Here, I briefly address the problem of genetically diverse offspring would allow the multiple mating or polyandry in social host to evade the constantly coevolving hymenoptera. In the bumble bee Bombus parasites (Hamilton, 1980; Hamilton et al, terrestris L, Shykoff and Schmid-Hempel 1990; Ladle, 1992). These scenarios are (1991b) have found that the transmission associated with genotype-genotype inter- of the trypanosome Crithidia bombi is eas- actions and negative frequency-dependent ier between relatives, ie full-sibs, than selection. In this case, rare host genotypes between unrelated individuals (fig 1). Fur- are favoured because the parasites have ther studies have shown that this variation not yet been able to adapt to them. Com- is likely to be due to the match between strain and host mon host genotypes, on the other hand, are parasite genotype (Shykoff, is a at a disadvantage, since they are rapidly 1991). This finding necessary precon- dition that within colonies targeted by the more rapidly evolving and genetic diversity would reduce the of transmis- more numerous parasites. It is obvious that probability sion and to render an such a selection regime is likely to apply in polyandry adaptive in the face of the interaction with parasites such as pro- mating strategy parasite pres- sure. results have been tozoa that have short generation times and Opposite recently for the bee occur in large numbers. It should be less reported honey (Woyciechowski et al, where the of a disease in relevant for other groups, eg, parasitoids, 1994) spread mated colonies was that have similar generation times and pop- experimentally multiply ulation sizes to their hosts. These hypotheses mainly addressed the problem of sexual reproduction whose cru- cial element, genetic recombination, gen- erates a large variety of genotypes among offspring. They have nevertheless impor- tant implications for social insect biology. Consider a case where offspring dispersal is high. Not only could hosts in such a situation escape from their parasites in time, but also in space, by occupying new sites. In con- trast, when dispersal of offspring is limited, the generation of variability among offspring is much more important. This should be par- ticularly relevant for social insects, where, by definition, most offspring do not disperse but stay at home to help (ie barring the sex- uals leaving their natal colony). Conse- quently, several authors have suggested that the maintenance of increased degrees of within-colony genotypic variation, as a result of polyandry and polygyny, could be selected for by parasites (Tooby, 1982; Hamilton, 1987; Sherman et al, 1988). Although of far-reaching importance, this hypothesis, unfortunately, has not been widely tested. not reduced as compared with those mated with fewer males. It could be argued that polyandry would not only increase the possibility for a queen to garner good resistance alleles to transmit to offspring, but may also increase her risk of mating with ’bad’ males, ie those that do not provide useful alleles, or increase the risk of becoming infected during mating (a cost of mating which seems relevant mainly for sexually transmitted parasites). This sit- uation is analysed in more detail elsewhere (Schmid-Hempel, 1994), where a simple system of gene-by-gene interactions is assumed, together with multiply mated queens, sufficient degrees of horizontal transmission (ie among colonies), and the potential for multiple infections. The calcu- lations show that polyandry is favoured in some regions of the parameter space (fig 2). At present, however, it is not clear whether parasites could act as selective fac- tors for the maintenance of high degrees of within-colony genetic variability and thus for different mating systems.

SOCIAL ORGANISATION

Parasites have to infect and establish them- selves in a host. In addition, parasites have to be transmitted to new susceptible hosts. Such should be an easy task in social insects, because many similar individuals are found in dense aggregations of the nests. The chances of a parasite being transmitted to new hosts is additionally increased in cases where it is able to manip- ulate its host’s behaviour in its own favour. Such a situation is found in the liver fluke wind over larger areas. A further situation that infects wood ants as intermediate hosts, is demonstrated by the example of the which was mentioned above. Interestingly, brown or yellowish colouration of worker a parallel case has been reported for ants ants infected by a cestode (eg, Crosland, infected by the fungus Entomophtora (Loos- 1988). Their normal colour is black and Frank and Zimmermann, 1976) which also hence parasitisation should make workers causes the ant to climb onto grass tips from more visible to the final host, which is where the fungal spores can be spread by assumed to be a bird. When considering parasite transmission, natively, mites may assume this role and it is necessary to distinguish vertical (ie they have indeed been identified as carri- from parent to offspring) from horizontal ers of entomopathogenic fungi (Schabel, transmission. For social insects, it is also 1982). Particularly those mites that live on necessary to distinguish within- from haemolymph, such as tracheal mites in the between-colony transmission. honey bee (Bailey and Ball, 1991), should be able to vector bacteria Within-colony transmission is vertical protozoa, viruses, and other in the haemo- when the parasite transfers from queens or parasites residing coel. The best evidence comes from the males to their worker offspring and to the mite Varroa brood of sexuals. To the extent that within-colony jacobsoni infecting bees. Mites nurse bees over transmission between workers keeps the honey prefer other classes of workers (Kraus et al, 1986) parasite within the same breeding unit, this and are thereby transported to the larvae may be considered ’vertical’ in an extended of the Several studies have shown sense. It therefore seems rather obvious colony. that infection by different viruses (acute bee how vertical transmission can be achieved, paralysis viruses APV, Egypt bee virus EBV, namely through the spread of an infection sacbrood virus SBV, black queen cell virus within the colony and then onto the sexu- BQCV) is associated with infection by Var- als that will carry the parasite to the next roa . From experiments (Koch and Ritter, and to own Between- generation offspring. and from such correlations it is clear transmission is horizontal to the 1989) colony that the mites act as vectors for viral dis- extent that different colonies are not nec- eases 1985, 1989; Allen et al, 1986; of each other. Parasitoids (Ball, essarily offspring Ball and Allen, 1988; Chastel et al, 1990). or social parasites, where the parasitic Actual infection seems to occur via trophal- females seek out their hosts, have actively laxis, faeces and transovarially, by nectar no difficulty in different colonies in affecting and pollen, and by contacts among mites the and act as population they may thereby and between bees and mites. Certainly, horizontal vectors for other parasites. The exciting bee-mite associations with the impli- available evidence does not allow yet judg- cation of parasite transmission remain to ment of whether vector-transmitted disease be discovered (Eickwort, 1994). is common in social insects. Of course, The workers themselves may act as vec- insects in general are vectors for disease tors if enter nests. This seems of vertebrates rather than hosts to vector- they foreign relevant for kleptoparasitic, honey- or nest- transmitted parasites, perhaps because no material robbing species of bees, for blood-sucking vectors prey on insects in instance in the eusocial meliponine genera similar as do on vertebrates. ways they Lestrimelitta in the Americas and Clepto- However, social insects could as a qualify trigona in Africa (Roubik, 1989). One can target for a vectoring species, because they easily imagine that such habitual robbery form and host ’bodies’ large long-lived (ie would facilitate the spread of spores or other the colonies) which remain available as a infective forms to and from the robbed rich source for potential vectors to utilise. colonies. The occasional drifting and entry of In this light, it is possible that some inquilines foreign workers or drones into a nest (Free, may act as vectors of parasites, eg, for 1958) could also lead to between-colony viruses, bacteria, fungal infections or pro- transmission. Given the well-documented tozoa (Jouvenaz, 1983). Vectoring by social phenomenon of parasite-induced manipu- parasites may be important, because social lations of host behaviour mentioned above, parasitism has arisen in all major groups of such drifting among colonies may perhaps social insects, including the termites. Alter- be in the genuine interest of the parasite and not reflect an ’erroneous’ behaviour on tacts among individuals of different castes is the part of the host. Indeed, in an experi- likely to be more structured than in a society ment the workers from honey bee colonies where division of labour is more diffuse. infected by V jacobsoni were more likely to Moreover, the highly sophisticated systems drift to other colonies than those from unin- of division of labour are often accompanied fested nests (Sakofski, 1990). This transfer by a corresponding spatial separation of the occurs mainly in late summer. However, tasks (eg, Seeley, 1982) which adds to the horizontal transmission among colonies in a within-colony structure of possible contacts population could be achieved in still other among individuals. To the extent that con- ways. In a recent study, Durrer and Schmid- tacts also determine parasite transfer (as Hempel (1994) demonstrated that the try- in directly transmitted pathogens such as panosome C bombi can be transmitted microsporidia or fungi; Wang and Moeller, between workers within and between dif- 1970; Kramm et al, 1982), or when certain ferent species of bumble bees through the groups of colony members must be infected shared use of flowers. Hence, infected work- (eg, last instar larvae; Jouvenaz et al, 1981), ers seem to deposit infective cells that are the routes of parasite transmission depend later picked up by uninfected workers which on colony organisation in important ways. happen to visit the same flower. Similar infer- Using the bumble bee, B terrestris and ences might be drawn from the case of the its intestinal parasite C bombi, Schmid- bacterium Spiroplasma picked up by honey Hempel and Schmid-Hempel (1993) tested bee foragers in the spring (Clark, 1977), the hypothesis that transmission routes may although it is not known precisely where the depend on the age-dependent polyethism parasite is contracted in this case. This typical for social insects. In particular, hori- transmission pathway has several ramifi- zontal transmission of this parasite to other cations for the relationship between plants colonies may be accomplished the for- and their In the context, by pollinators. present older for via the fact that transmission agers (the workers), example, between-colony flowers is established by foraging activity and the (Durrer and Schmid-Hempel, 1994), while transmission is common use of food resources, also has within-colony possi- ble those workers that inside the implications for the kind of parasites that by stay nest, ie the workers. The results food specialists among social insects might by younger in be able to sustain. suggested variation among colonies their propensity to transmit the parasite either by The problem of within- and between- young or old workers but could not directly transmission has some on colony bearing confirm the social and the division of labour hypothesis (Schmid-Hempel organisation and Schmid-Hempel, 1993). Nevertheless, in social insect societies. Imagine a species these findings lend at least some credibil- with a strict division of labor such that only ity to the idea that the social certain morphological and/or temporal organisation may affect transmission. castes attend a set of tasks with little over- parasite lap. This extreme system is found in no In addition to the effects of multiple mat- social insect, but cases like the leaf-cutting ings discussed above, social organisation ants, Atta, with their impressive array of dif- may be affected by parasitism. Some know- ferent worker morphologies, or Pheidole ledge comes from scattered findings of how and Apis with pronounced age-dependent parasite transmission within colony occurs. polyethisms may come close (Wilson, 1971; For example, worker honey bees not only Hölldobler and Wilson, 1990). In these soci- spread American foulbrood (a larval dis- eties, not only will work be rather strictly ease) among each other but also get it from partitioned, but also the frequency of con- the queen when attending her (Bitner et al, 1972). Acute bee paralysis virus (ABPV) is of host-parasite interactions in the sense presumably transmitted via trophallaxis (Bai- that colonies assume the role of ’individu- ley and Gibbs, 1964), and many viruses per- als’ in epidemiological theory. Colonies, of haps in additional ways (Chastel et al, 1990). course, are not biological individuals but a Grooming is involved in the transmission of colony of individuals. Consequently, the a fungal parasite in termites (Kramm et al, macroscopic properties of colonies vary con- 1982). Queen and brood attendance, and siderably among species, as there are dif- allogrooming, therefore gain additional sig- ferent colony sizes, structures, degrees of nificance in the light of infection risks. In polyandry, polygyny and so forth. Further- particular, workers may infect their (single) more, we find semelparous annual species such as in bumble bees and queen and thus doom the entire colony. iteroparous, such as in ants. In Therefore, if novel infections are primarily perennial species gen- within-host are brought into the colony by the foraging or eral, processes important guarding workers, then the pattern observed for the dynamics of host and parasite (eg, Antia and so the actual in social insects, namely that the young et al, 1994), colony structure and the fact that colonies consist of workers attend the queen, would be advan- individuals will have to be accounted for in tageous, as the probability that a worker framework. However, carries an infection should generally any comprehensive for the time these will be increase with age (Schmid-Hempel and being complications aside and the as a whole is Schmid-Hempel, 1993). Whether age- put colony to have the model’s dependent polyethism has evolved mainly assumed properties. for such reasons or because of its As an example, the case of infectious ergonomic advantages given the high for- diseases and the well-known SIR-type of ager mortality rates (Jeanne, 1986) could model is used (eg, Anderson and May, be investigated further. On the other hand, 1979). In this theory hosts are assumed to direct effects of parasitic infection on divi- be susceptible (ie not yet infected but avail- sion of labour have been found. For able to the parasite), infected, or recovered instance, infection by Nosema apis in the and immune. The dynamics of the parasite are this honey bee is associated with an earlier start then considered against background in a host that is assumed of the foraging career of the affected work- large population not to be in numbers the ers (Wang and Moeller, 1970) which thus regulated by par- shifts the entire work profile of the colony. asites. Hence, for this purpose parasites are assumed to have small effects on host Moreover, certain parasites are more spe- ie on the rate of colonies cific in preferentially attacking certain age mortality, mortality as a whole. The fitness of the is groups of hosts (Kraus et al, 1986). This parasite should give the ergonomics of colony then given by its net reproductive rate, R0, demography (cf, Oster and Wilson, 1978) which describes the number of secondary infections per infected host an additional dimension. For the time being, generated The if > 1. A a lot of this discussion must remain specu- (colony). parasite spreads R0 characteristic result of epidemiological lative but it is certainly an aspect worthy of is that further attention. dynamics

EPIDEMIOLOGY

Instead of focusing on single colonies in all where N = the number of hosts in the pop- detail, one could adopt a macroscopic view ulation, β = the transmission rate to new hosts, a = the parasite-induced mortality demic, where parasites can realise their full rate of hosts, b = the background host mor- potential of R0, the disease should rapidly tality rate, and v = the recovery rate from spread. Therefore, in principle, social insects parasitic infection (Anderson and May, should be quite prone to such invasions. 1991). This equation describes the repro- However, if rates of recovery and recruit- ductive rate of parasites in a non-equilibrium ment of new colonies into the population situation, ie during the build up of an epi- are low (colonies being long-lived), which demic (epizootic) infection. seems likely for social insects, then the epi- How does this equation apply to social demic will soon die out, leaving a propor- insects? With the macroscopic view the unit tion of colonies infected. In this equilibrium of interest is the colony as a whole, N refers state, following the epidemic build-up, each infected will on one to the number of colonies in a population, colony produce average while β is the rate of parasite transmission secondary infection, therefore standard the- between colonies (ie the horizontal trans- ory predicts that R0 • x* ≈ 1, where x* the mission component). The background mor- equilibrium proportion of susceptible (unin- tality rate, b, of established colonies is typi- fected) colonies in the population. Since R0 cally small (this is of course not true for is large, x* must be small, hence, virtually all incipient colonies). But also the values of a colonies are left infected once the epidemic to a halt due to the lack of and v should typically be small for social has ground newly insect colonies. The latter quantity, ie the rate available host colonies. of recovery from infection, is closely related to It is hardly known whether such epi- immunity and clearance of infection in indi- demics occur often, with the exception of vidual hosts. Insects are capable of mounting economically interesting cases, such as Var- immune defences and recovering (Gupta, roa mites in honey bees which have rapidly 1986). In colonies of social insects, which in spread throughout Europe. The lack of more this view count as the ’individual’, infections reports, which after all would describe quite can probably persist for extended periods of spectacular cases of infestation of colonies, time, as workers could constantly become suggests that epidemics are perhaps not either reinfected or the newly emerging work- often observed. Perhaps this relates to low ers contract the disease which is thus sus- values of β or low values of N in many tained within the nest. On the other hand, cases. As already stated, the assumptions hygienic behaviour, such as removal of dead of the standard epidemiological models may individuals (Robinson and Page, 1988) or not be met in many cases. In particular, removing brood, as documented for honey within-colony dynamics could be very impor- bees infected by Bacillus larvae (Rothen- tant and lead to different outcomes. Never- buhler, 1964a), would help to eliminate the theless, the considerations of epidemiology infection and lead to recovery of the colony. should lead to new insights about the evo- Swarming also could lead to the elimination lution and maintenance of parasites in pop- of a parasite, as implicated for tracheal mites ulations of social insects. These and other in honey bees (Royce et al, 1991). questions about parasitism in social insects Barring these uncertainties about the fre- have only just begun to be addressed. quency and effectiveness of infection elim- ination in entire colonies, evaluating equation [1] with values for the parameters as dis- ACKNOWLEDGMENTS cussed above would result in a high value for R0 , depending of course on N and β. I thank R Schmid-Hempel for discussions of these This means that at the beginning of an epi- issues and for help during many experiments and field work. Comments by an anonymous reviewer insectes sociaux qui jusqu’à présent sem- improved many pages of an earlier version of the blent sans rapport les uns avec les autres. manuscript. Financial support came from grants of the Swiss National Science Foundation (No insectes sociaux / 31-32193 91) and the Swiss Priority Programme / parasite / polyandrie ’Environment’ (No 5001- 035217). division du travail / épidémiologie

Résumé — Parasites et insectes Zusammenfassung — Parasiten und soziale Insekten. Parasiten sind sociaux. Bien que les parasites soient allge- omniprésents, ils ont été jusqu’à présent genwärtig und haben oft subtile Wirkungen auf ihre Wirte. Insekten peu étudiés chez les insectes sociaux, à Parasiten sozialer l’exception du parasitisme social et de sind bis heute nicht gut untersucht worden, mit der Ausnahme der und quelques espèces hôtes ayant une impor- Sozialparasiten, tance commerciale. Tous les grands derjenigen kommerziell wichtiger oder groupes de parasites utilisent les insectes schädlicher Arten. Parasiten sind in allen sociaux comme hôte : virus, bactéries, wichtigen Gruppen zu finden und umfas- champignons, protozoaires, helminthes sen Viren, Bakterien, Pilze, Protozoen, ver- ainsi que des parasitoïdes. Bien que tous schiedene Helminthen und Parasitoiden. les stades soient attaqués, il semble qu’il y Sie attackieren alle Stadien des Wirtes, ait peu de parasites d’œufs (par exemple doch sind bemerkenswert wenige Eipara- quelques virus), il y a en revanche une mul- siten (zB einige Viren) bekannt. Dagegen titude de parasites des stades larvaires et gibt es eine Vielzahl von Parasiten der Lar- adulte. Diverses caractéristiques des ven und Adulten. Verschiedene Charakte- insectes sociaux peuvent être en relation ristika von sozialen Insekten könnten im avec le parasitisme. Par exemple, les Zusammenhang mit Parasitismus stehen. accouplements multiples (polyandrie) Mehrfachpaarung durch die Königin accroissent comparativement la variabilité (Polyandrie), zum Beispiel, erhöht ver- génétique des ouvrières au sein d’une colo- gleichsweise die genetische Variabilität nie et sont ainsi susceptibles de diminuer la unter den Arbeiterinnen einer Kolonie und transmission des parasites infectieux entre würde dadurch eine reduzierte Transmis- les ouvrières. L’organisation sociale et la sion von Parasiten innerhalb der Kolonie division du travail déterminent les interac- zur Folge haben. Auch die Sozialorganisa- tions entre les membres de la colonie et tion und Arbeitsteilung kann durch Parasi- devraient donc influencer les voies de trans- ten beeinflusst sein. So wird zum Beispiel mission des parasites dans et entre les die horizontale Transmission zu anderen colonies. Par exemple, la transmission hori- Kolonien in der Population durch die Sam- zontale entre colonies pourrait être liée à meltätigkeit der Arbeiterinnen beeinflußt, l’activité des butineuses qui disséminent welche Parasiten verbreiten oder neu in die les parasites ou en rapportent de nouveaux Kolonie eintragen. Schließlich legt die Stan- dans la colonie. Finalement la théorie de dardtheorie der Epidemiologie den Schluss l’épidémiologie moyenne prédit que la bio- nahe, daß soziale Insekten im Prinzip gün- logie des insectes sociaux peut être favo- stige Bedingungen für das Entstehen einer rable à une épidémie, bien que celle-ci Epidemie bieten, diese aber wahrschein- puisse rapidement disparaître. Des études lich nicht lange aufrechterhalten werden plus détaillées sur le parasitisme social sont kann. Mehr und detaillierte Studien von nécessaires et semblent prometteuses pour Parasiten sozialer Insekten sind nötig. Sie expliquer des faits de la biologie des bergen die Hoffnung, daß vorher schein- bar zusammenhanglose Aspekte neu ver- Bedding RA (1985) Nematode parasites of Hymenoptera. standen werden können. In: Plant and Insect Nematodes (WR Nickle, ed), Marcel Dekker, New York, USA, 755-795 Bitner RA, Wilson Hitchcock JD Soziale Insekten / Parasite / / WT, (1972) Passage Polyandrie of Bacillus larvae spores from adult honeybees to Arbeitsteilung / Epidemiologie attendant workers (Apis mellifera). Ann Entomol Soc Am 65, 899-901 Boomsma JJ, Grafen A (1991) Colony-level sex ratio REFERENCES selection in the eusocial Hymenoptera. J Evol Biol 3, 383-407 Broome JR, Sikorowski PP, Norment BR (1976) A mech- Alexander RD (1974) The evolution of social behaviour. anism of pathenogenicity of Beauveria bassiana on Annu Rev Ecol Syst 5, 325-383 larvae of the red imported fire ant, Solenopsis rich- Allen MF, Ball BV, White, RF, Antoniw JF (1986) The teri. J Invertebr Pathol 28, 87-91 detection of Acute Paralysis Virus in Varroa jacobsoni Brown CR, Brown MB (1986) Ectoparasitism as a cost of by the use of a simple indirect ELISA. J Apic Res coloniality in cliff swallows. Ecology 67, 1206-1218 25, 100-105 Chastel C, Robaux P, Le Goff F (1990) New virus from Anderson D, Giacon H (1992) Reduced pollen collec- honey bee colonies. Is Varroa a vector or does it tion by honey bee (Hymenoptera: Apidae) colonies amplify viruses? In: Proceedings of the International infected with Nosema apis and sacbrood virus. Symposium on Recent Research on Bee Pathology J Econ Entomol85, 47-51 (W Ritter, ed), International Federation of Beekepers Anderson RM, May RM (1979) Population biology of Association, Gent, Belgium, 150-154 infectious diseases. I. Nature (Lond) 280, 361-367 Clark TB (1977) Spiroplasma as a new pathogen in the Anderson RM, May RM (1981) The population dynam- honeybees. J Invertebr Pathol 29, 112-113 ics of microparasites and their invertebrate hosts. Cole BJ (1983) Multiple mating and the evolution of Philos Trans R Soc London, B 291, 451-524 social behavior in the Hymenoptera. Behav Ecol Anderson RM, May RM (1991) Infectious Diseases of Sociobiol 12, 191-291 Humans. Oxford University Press, Oxford, UK Crawley WC, Baylis HA (1921) Mermis parasitic on ants Andersson M (1994) Sexual Selection. Princeton Uni- of the genus Lasius. J R Microsc Soc, 353-364 vesity Press, Princeton, NJ, USA Crosland MWJ (1988) Effect of a gregarine parasite on Antia R, Levin BR, May RM (1994) Within-host popula- the color of Myrmecia pilulosa (Hymenoptera: Formi- tion dynamics and the evolution and maintenance cidae). Ann Entomol Soc Am 81, 481-484 of microparasite virulence. Am Nat 144, 457-472 Crozier RH, Consul PC (1976) Conditions for genetic Bailey L, Ball BV (1991) Honey Bee Pathology (2nd ed). polymorphism in social hymenoptera under selec- Academic Press, London, UK tion at the colony level. Theor Popul Biol 10, 1-9 Bailey L, Ball BV, Perry JN (1981) The prevalence of Crozier RH, Page RE (1985) On being the right size: viruses of honey bees in Britain. Ann Appl Biol 97, male contributions and multiple mating in social 109-118 Hymenoptera. Behav Ecol Sociobiol 18, 105-116 Bailey L, Gibbs AJ (1964) Acute infection of bees with Davies CR, Ayres JM, Dye C, Deane LM (1991) Malaria paralysis virus. J Insect Pathol 6, 395-407 infection rate of Amazonian primates increases with body weight and group size. Funct Ecol5, 655-662 Ball BV (1985) Acute Paralysis Virus isolates from honey- bee colonies infested with Varroa jacobsoni. J Apic Davis LR Jr, Jouvenaz DP (1990) Obeza floridana, a Res 24, 115-119 parasitoid of Camponotus abdominalis floridanus from Florida (Hymenoptera: Ecuharitidae, Formici- Ball BV (1989) Varroa as a virus vector. In: jacobsoni dae). FI Entomol 73, 335-337 Proceedings of a Meeting of the EC-Experts’ Group, Udine (R Cavalloro, ed), Commission of the Euro- Dobson A, Crawley M (1994) Pathogens and the struc- pean Communities, Brussels, Belgium, 241-244 ture of plant communities. Trends Ecol Evol 9, 393- 398 Ball BV, Allen MF (1988) The prevalence of pathogens in honeybee (Apis mellifera) colonies infested by Dobson AP (1988) The population dynamics of para- with the mite Varroa jacobsoni. Ann Appl Biol 113, site-induced changes in host behavior. Q Rev Biol 63, 237-244 139-165 Becker N, Angulo J (1981) On estimating contagious- Dobson AP, Hudson PJ (1986) Parasites, diseases and ness of a disease transmitted from person to per- the structure of ecological communities. Trends Ecol son. Math Biosci 54, 137-154 Evol 1, 11-14 Doull KM, Cellier KM (1961) A survey of the incidence of Keymer AE, Read AF (1991) Behavioural ecology: the Nosema disease (Nosema apis Zander) in honey impact of parasitism. In: Parasite-Host Associations: bees in South Australia. J Insect Pathol 3, 280-288 Coexistence or Conflict? (CA Toft, A Aeschlimann, L Bolis, eds), Oxford Press, Oxford, UK, Durrer S, Schmid-Hempel P (1994) Shared use of flow- University 37-61 ers leads to horizontal pathogen transmission. Proc R Soc London, B 258, 299-302 Koch W, Ritter W (1989) Examination of artificially infested brood with Varroa mites for infec- Eickwort GC (1994) Evolution and life-history patterns of secondary tions. In: of a of the mites associated with bees. In: Mites (MA Houck, Proceedings Meeting EC-Experts’ Udine Commission of the ed), Chapman and Hall, New York, USA, 218-251 Group, (R Cavalloro, ed), European Communities, Brussels, Belgium, 245- Fisher RM (1988) Observations on the behaviours of 251 three Euopean cuckoo bumble bee species in In: (Psithyrus). Insectes Soc 35, 341-354 Koeniger G (1991) Diversity Apis mating systems. Diversity in the Genus Apis (DR Smith, ed), Westview Free JB (1958) The drifting of honey bees. J Agric Sci 51, Press, Boulder CO, USA, 265 294-306 Kramm KR, West DF, Rockenbach PG (1982) Termite Freeland WJ (1976) Pathogens and the evolution of pri- pathogens: transfer of the entomophagous Metarhiz- 12-24 mate sociality. Biotropica 8, ium anisoplia between Reticulitermes sp termites. Gupta AP (ed) (1986) Hemocytic and Humoral Immunity J Invertebr Pathol 40, 1-6 in Arthropods. John Wiley & Sons, New York, USA Kraus B, Koeniger N, Fuchs S (1986) Unterscheidung Hamilton WD (1964) The genetical evolution of social zwischen Bienen verschiedenen Alters durch Var- behavior. J Theor Biol 7, 1-52 roa jacobsoni Oud und Bevorzugung von Ammen- bienen im Sommervolk. Apidologie 17, 257-266 Hamilton WD (1980) Sex vs non-sex vs parasite. Oikos 35, 282-290 Kulincevic JM (1986) Breeding accomplishments with bees. In: Bee Genetics and Hamilton WD disease, and honey Breeding (TE (1987) Kinship, recognition, Academic constraints of social evolution. In: Animal Rinderer, ed), Press, Orlando, FL, USA, intelligence: 391-414 Societies: Theory and Facts (Y Ito, JL Brown, J Kikkawa, eds), Japanese Scientific Society Press, Ladle RJ (1992) Parasites and sex: catching the red Tokyo, Japan, 81-102 queen. Trends Ecol Evol 7, 405-408 Hamilton WD, Axelrod A, Tanese R (1990) Sexual repro- Laidlaw HHJ, Page RE (1984) Polyandry in honeybees duction as an adaptation to resist parasites (a (Apis mellifera): sperm utilization and intracolony review). Proc Natl Acad Sci USA 87, 3566-3573 genetic relationships. Genetics 108, 985-997 Hohorst W, Graefe G (1961) Ameisen - obligate Zwi- Lin N (1964) Increased parasite pressure as a major schenwirte des Lanzettegels (Dicrocoelium den- factor in the evolution of social behavior in halictine driticum). Naturwissenschaften 48, 229-230 bees. Insectes Soc 11, 187-192 Hölldobler B, Wilson EO (1990) The Ants. Springer, Loos-Frank B (1978) Zum Verhalten von Ameisen der Berlin, Germany Gattung Formica (Hymenoptera; Formicidae) gegenüber Schleimballen des Kleinen Leberegels Jeanne RL (1986) The evolution of the organization of Dicrocoelium dendriticum (Digenea: Dicrocoeliidae) work in social insects. Monitore Zool Ital 20, (NS) und über ihrer 119-133 infektionsbedingte Veränderungen Hämolymphe. Entomol Germ 4, 12-23 Jouvenaz DP (1983) Natural enemies of fire ants. FI Loos-Frank B, Zimmermann G (1976) Über eine dem Entomol 66, 111-121 Dicrocoelium-Befall analoge Verhaltensänderung Jouvenaz DP, Ellis EA (1986) Vairimorpha invictae nsp bei Ameisen der Gattung Formica durch einen Pilz (Microspora: Microsporida), a parasite of the red der Gattung Entomophthora. ZParasitenkd49, 281- imported fire ant, Solenopsis invicta Buren 289 Formicidae). J Protozool 33, 457-461 (Hymenoptera: Lundberg H, Svensson BG (1975) Studies on the Jouvenaz DP, Lofgren CS, Allen GE (1981) Transmission behaviour of Bombus Latr species (Hymenoptera: and infectivity of spores of Burenella dimorpha Apidae) parasitized by Sphaerularia bombi Dufour (Microsporidae, Burenellidae). J Invertebr Pathol 37, (Nematoda) in an alpine area. Norw J Entomol 22, 265-268 129-134 Kathirithamby J, Johnston JS (1992) Stylopization of Minchella DJ, Loverde PT (1981) A cost of increased Solenopsis invicta (Hymenoptera: Formicidae) by early reproductive effort in the snail Biomphalaria Caenochloax fenyesi (Strepsiptera: Myrmecolacidae) glabrata. Am Nat 118, 876-881 in Texas. Ann Entomol Soc Am 85, 293-297 Moore J (1983) Responses of an avian predator and its Keller L (ed) (1993) Queen Number and Sociality in isopod prey to an acanthocephalan parasite. Ecology Insects. Oxford University Press, Oxford, UK 64, 1000-1015 Moore J (1984) Altered behavioral responses in inter- Rothenbuhler WC (1964a) Behavior genetics of nest mediate hosts - an acantocephalan parasite strategy. cleaning honeybees. IV. Responses of F1 and back- Am Nat 123, 572-577 cross generations to disease killed brood. Am Zool Nonacs P (1988) Queen number in colonies of social 4, 111-123 Hymenoptera. Evolution 42, 566-580 Rothenbuhler WC (1964b) Behavior genetics of nest in bees. I. of four inbred Ogloblin-AA (1939) The Strepsiptera parasites of ants. cleaning honey Response lines to disease-killed brood. Anim Behav 578- des VII. Internationalen 112, Verhandlungen Kongresses 583 für Entomologie 2, 1277-1285 Roubik DW and Natural of Oster GF, Wilson Caste and the (1989) Ecology History Trop- EO (1978) Ecology in ical Bees. Press, Social Insects. Princeton Press, Cambridge University Cambridge, University Princeton, UK NJ, USA LA, PA, DM, BA Otis CD Effects of Royce Rossignol Burgett Stringer (1991) GW, Scott-Dupree (1992) Acarapis Reduction of tracheal mite parasitism of bees woodi on overwintered colonies of bees honey honey by swarming. Philos Trans R Soc London, Ser B in New York. J Econ Entomol (Hymenoptera: Apidae) 331, 123-129 85, 40-46 Rupp L (1987) The genus Volucella (Diptera: Syrphi- RE The evolution of behav- Page (1980) multiple mating dae) as commensals and parasites in bumblebee ior bee mellifera Genetics by honey queens (Apis L). and wasp nests. In: Chemistry and Biology of Social 96, 263-273 Insects (J Ederand, DH Rembold, eds), Peperny Page RE, Metcalf RA (1982) Multiple mating, sperm uti- Verlag, Munich, Germany, 642-643 and social evolution. Am Nat lization, 119, 263-281 Sakofski F (1990) Quantitative investigations on transfer Pamilo P (1981) Genetic organization of Formica san- of Varroa jacobsoni. In: Proceedings of the Interna- guinea populations. Behav Ecol Sociobiol 9, 45-50 tional Symposium on Recent Research on Bee Pathology (W Ritter, ed), International Federation of Pamilo P (1991) Evolution of colony characteristics in Beekepers Association, Gent, Belgium, 70-72 social insects. 2. Number of reproductive individu- als. Am Nat 137, 83-107 Schabel HG (1982) Phoretic mites as carriers of ento- mopathogenic fungi. J Invertebr Pathol 39, 410-412 Pesquero MA, Campiolo S, Fowler HG (1993) Phorids P Infection and variabil- (Diptera: Phoridae) associated with mating swarms Schmid-Hempel (1994) colony of Solenopsis saevissima (Hymenoptera, Formici- ity in social insects. Philos Trans R Soc London, Ser dae). FI Entomol 76, 179-181 B 346, 313-321 Koella JC and its Poulin R, Fitzgerald GJ (1989) Shoaling as an anti- Schmid-Hempel P, (1994) Variability ectoparasite mechanism in sticklebacks. Behav Ecol implications for host-parasite interactions. Parasitol Sociobiol 24, 251-255 Today 10, 98-102 Price PW, Westoby M, Rice B et al (1986) Parasite medi- Schmid-Hempel P, Müller C, Schmid-Hempel R, Shykoff JA and correlates of ation in interactions. Annu Rev Ecol Syst (1990) Frequency ecological ecological flies in 17, 487-505 parasitism by conopid (Conopidae, Diptera) populations of bumblebees. Insectes Soc 37, 14-30 Price PW, Westoby M, Rice B (1988) Parasite-medi- Schmid-Hempel P, Schmid-Hempel R (1993) Trans- ated competition: some predictions and tests. Am mission of a pathogen in Bombus terrestris, with a Nat 13, 544-555 note on division of labour in social insects. Behav Queller DC (1993) Worker control of sex ratios and selec- Ecol Sociobiol 33, 319-327 tion for extreme Am Nat multiple mating by queens. Schneider G, Hohorst W (1971) Wanderungen der 142, 346-351 Metacercarien des Lanzett-Egels in Ameisen. Natur- Queller DC, Strassmann JE (1988) Reproductive suc- wissenschaften 58, 327-328 cess and in the Polistes group nesting paper wasp, Schwarz MP (1986) Persistent multi-female nests in an annularis. In: Reproductive Success (TH Clutton- Australian allodapine bee, Exoneura bicolor Brock, ed), The University of Chicago Press, (Hymenoptera: Anthophoridae). Insectes Soc 33, Chicago, USA, 76-98 258-277 Rinderer TE, Collins AM, Brown MA (1983) Heritabili- Seeley TD (1982) Adaptive significance of the age ties and correlations of the honey bee: response to polyethism schedule in honeybee colonies. Behav Nosema apis, longevity, and alarm response to Ecol Sociobiol 11, 287-293 isopentyl acetate. Apidologie 14, 79-85 Seger J (1983) Partial bivoltinism may cause alternating Robinson GE (1992) Regulation of division of labor in sex-ratio biases that favour eusociality. Nature (Lond) insect societies. Annu Rev Entomol 37, 637-665 301, 59-62 Robinson GE, Page RE (1988) Genetic determination of Sherman PW, Seeley TD, Reeve HK (1988) Parasites, guarding and undertaking in honey-bee colonies. pathogens, and polyandry in social Hymenoptera. Nature (Lond) 333, 356-358 Am Nat 131, 602-610 Shykoff JA (1991) On genetic diversity and parasite Thompson JN, Burdon JJ (1992) Gene-for-gene coevo- transmission in socially structured populations. PhD lution between plants and parasites. Nature (Lond) Thesis Phil-Naturw Fakultät, University of Basel, 360, 121-125 Basel, Switzerland Tooby J (1982) Pathogens, polymorphism, and the evo- Shykoff JA, Schmid-Hempel P (1991 a) Genetic relat- lution of sex. J Theor Biol 97, 557-576 edness and selection eusociality: parasite-mediated Wakelih D (1985) Genetic control of immunity to helminth on the of Behav Ecol genetic composition groups. infections. 1, 17-23 Sociobiol 28, 371-376 Parasitol Today DI, Moeller FE The division of labor and JA, P Parasites and Wang (1970) Shykoff Schmid-Hempel (1991b) attendance behavior of Nosema-infected the advantage of genetic variability within social insect queen worker bees. J Econ Entomol 63, 1539-1541 colonies. Proc R Soc London, Ser B 243, 55-58 honey Ward PS Genetic relatedness and Skou JP, Holm SN, Haas H (1963) Preliminary investi- (1983) colony orga- nization in a species complex of Ponerine ants. 1. gations on diseases in bumble bees (Bombus Latr). and of colonies. Kgl Vet-og Lanbohojsk Aarsskr, 27-41 Phenotypic genotypic composition Behav Ecol Sociobiol 12, 285-299 Starr CK (1985) Enabling mechanisms in the origin of sociality in the Hymenoptera - the sting is the thing. West-Eberhard MJ (1975) The evolution of social behav- Ann Entomol Soc Am 78, 836-840 ior by kin selection. Q Rev Biol 50, 1-33 Wheeler Wheeler EW A new of Strand MR (1986) The physiological interactions of par- GC, (1924) species with notes on a asitoids with their hosts and their influence on repro- Schizaspidia (Eucharidae), eulophid ductive strategies. In: Insect Parasitoids (J Waage, ant parasite. Psyche 31, 49-56 D Greathead, eds), Academic Press, London, UK, Wilson EO (1971) The Insect Societies. Harvard Uni- 97-136 versity Press, Cambridge, MA, USA Stuart RJ, Alloway TM (1988) Aberrant yellow ants: Woyciechowski M, Krol E, Figurny E, Stachowicz M, North American Leptothorax species as intermediate Tracz M (1994) Genetic diversity of workers and hosts of cestodes. In: Advances in Myrmecology infection by the parasite Nosema apis in honey bee (JC Trager, ed), EJ Brill, Leiden, The Netherlands, colonies (Apis mellifera). In: 12th Congress of the 537-545 International Union for the Study of Social Insects Tanada Y, Kaya HK (1993) Insect Pathology. Academic (A Lenoir, G Arnold, M Lepage, eds), Université Press, San Diego, CA, USA Paris-Nord, Paris, France, 347