Memoirs of the Museum of Victoria 56(2):411-419 (1997) 28 February1997 https://doi.org/10.24199/j.mmv.1997.56.32 POPULATION BIOLOGY OF SOCIAL CONSERVATION

PEKKA PAM1Lo1.2 AND Ross H. CROZIER' 1 School of Genetics and Human Variation. La Trobe University, Bundoora 3083, Australia, and Victorian Centre for Conservation Genetics 2 Program of Conservation Biology, Department of Genetics, Uppsala University. Box 7003, 750 07 Uppsala. Sweden

Abstract Pamilo, P. and Crozier. R.H., 1997. Population biology of social insect conservation. Memoirs of the Museum of Victoria 56: 411-4 I 9. Social , especially and termites, are keystone species in most non-arctic ter­ restrial ecosystems, but their special features have received little attention in conservation discussions. Both plants and other arc affected by social insects as keystone species. Despite the abundance of species it may be that rather fewin any one location are critical to understanding the ecosystem, because of the mosaic nature of the distributions of the dominant species. Social insects present a wide array of life patterns within the broad sway of sociality, varying in having colonies which are annual or perennial, multi- or single-nested, and multi- or single-queened. Populations of social insects may often be much more vul­ nerable than they appear on the basis of numbers of individuals: the reproductive division of labor means that the effective population size may be small for a species of apparently reasonable abundance. Factors increasing effective population size occur in many species, especially rare ones, but it is uncertain these represent adaptations to rareness. In eusocial , many species. probably most, show inbreeding depression of a special kind due to the production of sterile males caused by homozygosity at the sex locus. In many termites, on the other hand, colonies pass naturally through cycles of inbreeding and out­ breeding. Socially parasitic ants which habitually inbreed appear to have evolved a different means of sex-determination and do not show inbreeding depression. Differences in the mode of colony formation between species also lead to difference in the longevity of col­ onies. dispersal abilities, and the robustness of populations to disturbance. The greatest threats to social insects, apart from humans, are other social insects, and this applies also to invading pest species, which tend to have particularly strong impacts on native species. Social insects thus form a vital part of ecosystems, but also impel the need to accept long­ term studies because of the slow pace at which their populations change.

Introduction area can be small and comparable to that of many vertebrates. Yet, the non-reproductive Invertebrates generally differ from vertebrates workers constitute a large biomass which affects in the conservation problems they are prone to both the food requirements of the population because of the scale on which they live. Even a and the impact of the population on its environ­ small area can contain a large population, so that ment. Second, the life of social insects normally considerations such as the area of habitat centres around a nest, which makes them seden­ required and problems ofa small population size tary and can restrict dispersal. In that respect (leading, at least potentially, to inbreeding their population biology can resemble that of depression) can be less important and more plants rather than of other insects. Third, most easily solved. For such populations, habitat loss social insects (ants, bees and wasps) belong to can remove whole populations without an inter­ the order Hymenoptera and have a male­ vening stage of fragility. haploid sex-determination system. This genetic Social insects, however, differ in several mechanism affects some genetic population important aspects from other invertebrates. We characteristics, particularly the effective popu­ can recognize three major features that charac­ lation size and the genetic load resulting from terize social insects and are important in their the sex-determination mechanism. population biology and in the population bio­ logical aspects of their conservation. First, the Our aim here is to describe how these features, ratio of population size to biomass is small. Only which are characteristic of social insects, affect a small number ofindividuals reproduce and the their populations and in which ways they can effectivepopulation size maintained in a given influence their conservation.

411 412 P. PAMILO AND R.H. CROZIER

The role of social insects in ecosystems pletely on the host ants. There have been several national extinctions of Maculinea species in One criterion that can be used when setting western Europe (five species live in Europe). The priorities for conservation is to what extent conservation of the butterfly populations other species depend on the species in question requires simultaneous attention on the host ant (Soule, 1987). Without focusing on any single populations, and active management on these species, we can state that social insects in general lines has been carried out to reintroduce the form an important group both in ecosystems in large blue (M. arion) in Britain (Elmes and general and in having important mutual Thomas, 1992; Thomas, 1995). Similar situa- relationships with other organisms, plants as tions are expected in Australia, where the lycae- well as animals. nid butterflies include species depending on the The social life pattern has made social insects care provided by ants (New, 1993). evolutionarily very successful, not so much in As regards the plants, the main roles of social species diversity but in biomass. Social insects are in protecting them against herbivores Hymenoptera make only about 10% of all (which has led to the evolution of special doma- described hymenopteran species, but the esti- tia, extrafloral nectaries and food bodies by mates from South-American tropical rainforests some plants (Beattie, 1 985)), in spreading seeds, show that social insects make more than a quar- and as pollinators. For example, Handel et al. ter of all biomass, and c. 80% of all insect (1981) counted that 13 of 45 herbaceous plant biomass (reviewed by Wilson, 1990). It is species in a mesic forest had seeds dispersed by reasonable to assume that they have a key role in ants, and they comprised ca 40% of the above- the energy budgets of many other ecosystems, ground herbaceous biomass. Ants contribute although the ant biomass in temperate grassland significantly to the population dynamics of such areas is only 1-15% of all invertebrates (Pis- myrmechochores and shape the structure of the arski, 1978). The diversities of social insect whole plant community. They are also import- species arc highest in the tropics, except of ant for the viability of some rare and endangered bumblebees that exist mainly in temperate plant populations. Although bees are major pol- areas. Of the major groups of social insects, ter- linators of angiosperms, social bees have gener- mites and stingless bees occur mainly in tropics ally a less important role. They can, however be and subtropics. The major centres of social locally important. Some bees visit only a few wasps are partly in south-east Asia (vespine closely related species of plants, and the protec- wasps) and in neotropics (polybine wasps). tion of such oligolectic bees relies on the preser- Interspecific competition among ants has led vation of these plants. At the same time special- to dominance hierarchies between species, and ist pollinators can be important for the success the dominant species can largely influence the of the plant populations. Such a close relation- composition of the whole ant community. This ship exists between Aconitum septentrionaie has led to so called ant mosaics (Leston, 1978: and its pollinator, the bumblebee Bombus conso- Majer, 1993). found particularly in tropical brinus, both of them being rare and endangered forests. Replacement of the dominant species in northernmost Europe (Pekkarinen, 1979). can lead to a large change in the assemblage not The relatively low species diversity of of only ant species but also of other social insects, at least when compared to many other (perhaps even vertebrates) as well. In fact, insects, makes it feasible to use the phylogenetic Gilbert (1980) suggested as one of the research approach for defining priority areas priorities for neotropical conservation biology to protect the worldwide species diversity. 'autecology of link species, keystone mutualists Bumblebees of the Bombus sibiricus-gwup comprise one of the and dominant ant species'. Majer (1993) notes first taxonomic groups where this approach that ant mosaics become rarer as one moves has been applied, although away from the tropics. the authors note that conservation of bumblebees is not yet generally Social insects affect their environment also by considered to have a high priority building nests and by manipulating the foraging (Williams et al., 1993). There are, however, reports areas. These activities create niches for many indi- cating that bees are generally declining other organisms, and indeed, the nests of social (OToole. 1993). insects harbour many different guests — and parasites. Some of the best studied examples come from the Maculinea butterflies, the big Social insect life patterns blues, whose caterpillars live in colonies of the Social insect life patterns vary enormously, ants and w hose life cycles depend com- and their societies can be classified on several bases. POPULATION BIOLOGY OF SOCIAL INSECT CONSERVATION 413

We present here some basic features (Crozier nest and/or foraging area against competing and Pamilo, 1 996), all of these adding important colonies. Such single-nest colonies are termed aspects also to conservation, mainly because monodomous. It is, however, common particu- they affect either the effective population size or larly in some ants that neighbouring nests are dispersal. connected to each other and exchange individ- Annual vs perennial colonies. Bumble bees, uals, brood and food. Such multinest colonies

many primitively social bees, and many social are termed polydomous. It is evident that such wasps have annual colonies. They have a solitary polydomous colonies can effectively occupy and phase, with a female (or a group offemales) start- dominate a habitat patch. In an extreme form, ing a colony and taking care of the first brood, when the network of interconnected nests covers

after which the offspring will stay as helping a large area, it has been called a supercolony, and individuals raising additional broods. The col- if there is no hostility within the whole popu-

ony has first an ergonomic stage when it grows, lation, the population is termed unicolonial. and the sexual offspring are produced towards Unicolonial species appear to become pests the end of the season when the colony moves to a more often than other species, but this apparent reproductive stage (Oster and Wilson, 1978). trend may simply indicate that many unicolon- Honeybees, stingless bees, swarm-founding ial species remain to be detected. wasps, and all ants and termites have perennial colonies. A colony can have a short solitary Effective population size period if a single female establishes the colony, but after that the colony can survive several Vulnerability of populations can be caused by years and produce sexuals repeatedly. deterministic or stochastic factors. Stochastic Independent vs dependent colony foundation. factors depend on the population size and its There are two major modes of nest foundation: fluctuations, and on the environmental stochas- independent and dependent. Independently ticity. As mentioned above, the ratio of effective founding females establish new nests without population size to biomass is generally low, or any help from workers, either singly or in small very low, in social insects. The territory of a groups. Dependent nest founding relies on single ant nest can cover several hectares, yet workers helping the queen. Dependent nest there may be just one reproductive female in the foundation commonly takes place by budding or nest. The density of social insect populations can fission, in which a new nest is established by therefore be low, even though the worker ants workers and queens departing from the parental are abundant.

nest. New nests can also in some ants be estab- The concept of effective population size (N e ) lished by help of alien species, in form of tem- refers to the genetic effects of a finite population the porary social parasitism. The queen intrudes size. There are various concepts of N e , and into a host colony, and the host workers raise the one commonly used is the inbreeding effective is affected first brood, after which the colony gradually size. In diploid populations, this Ne turns into that of the intruding species. by biased sex ratios as follows /(N Monogynous vs. polygynous colonies. Many N c = 4N rN m r + Nm ) refer the numbers of social insects have a single reproductive female, where N f and Nm to a queen, in their nests and the colonies are called reproductive females and males, respectively. In monogynous. This is the case in the honeybees, male-haploid insects, e.g., in ants, bees and most other eusocial bees, many eusocial wasps, wasps, is, = /(2N the majority of termites and many ants. This N e 9N rN m r + 4N m ) however, not a rule and about half of the Euro- which for the same numbers of females and ie. (N and yields smaller effective sizes pean ant species have polygynous colonies, males f N m ) they have several to many reproductive females than in a diploid population, unless the sex ratio per nest (Buschinger, 1974). Some wood ants of is very female-biased. The sex ratios refer to the Formica rufa group have up to hundreds of those among breeding individuals, and the effec- queens in their mounds (Rosengren and Pamilo, tive population size therefore depends on the females. 1983). Even if a colony has many queens, they numbers of matings of males and Both may not share reproduction evenly and their sexes are known to mate multiply in many contributions have a non-zero variance, ie. there species, but the number of matings by females is species typically have exists a reproductive skew (Keller and Reeve, commonly low. Many 1994; Pamilo and Crozier, 1996) once-mating females, and the effective number rarely exceeds 2. Exceptions are the Monodomous vs. polydomous colonies. A of matings in cerana. single nest usually forms a colony and defends its honeybees with up to 30 matings Apis 414 P. PAMILO AND R.H. CROZIER

social wasps of the genus Vespula, and leaf- tation to rarity, with selection on local aggre- cutting ants (Crozier and Pamilo, 1996). gations occurring as for colony-level selection. Whereas multiple mating by females increases When estimating population sizes in rare the effective population size, multiple mating by species, it is important to not only count the males decreases it, because the effect is that sev- nests but to estimate the number of reproductive eral females in the population would carry females in them. Genetic studies of an sperm of the same male. Multiple mating by endangered ant Formica cinerea in northern males is known in several ants, but its general Europe gives an example. This is a ground- occurrence and genetic effectiveness are poorly nesting species living in open sandy habitats. known. Males of many species are known to Such habitats are rare in northern Europe and mate only once, as they either die or lose their constitute either of areas where the retreat of ice genitalia when mating (e.g., honeybee). stopped for a longer period at the end of last A further factor reducing effective population glaciation, or of coastal sand fields. The species size is worker reproduction in social Hymenop- is considered vulnerable in Sweden, and it has tera. Workers in many species are sterile, but in very fragmented populations in both Sweden others they can lay haploid eggs developing into and Finland, with some populations occupying males, sometimes workers are assumed to be the only tens of square metres of suitable habitat. major source of males. Even though worker Preliminary genetic studies indicate that in reproduction increases the number of repro- some populations the nests are polygynous, ductive individuals, it decreases (slightly) the whereas in other populations the nests have a effective population size, because it does not single queen (Lindstrom et al., 1996). Clearly, increase the gene pool but simply adds a further the effective size, and perhaps vulnerability, of step for stochastic fluctuations and increases the the populations depend also partly on their variance of male allele frequencies (Crozier, social organization. 1979). Patterns of population size fluctuations vary The effective size of a social insect population among different taxonomic groups. Bumblebees depends clearly not only on the nest density, but and most social wasps have annual colonies, and on the number of reproductive individuals in their populations fluctuate as a function of the nests, on the reproductive skew among these weather conditions. Honeybees, stingless bees, individuals, and on the genetic mechanisms. ants and termites have perennial colonies, and The genetically effective population size consequently their populations are more stable. depends on the effective levels of polygyny and Ant queens can live 1 0-20 years (Holldobler and polyandry, and when the contributions of differ- Wilson, 1990), and an estimate of the mean ent queens, or those of the males copulating with longevity of queens in the ant Formica exsecta in the same female, are unequal, the effective natural conditions was over 20 years (Pamilo, number of individuals is lower than the actual 1991). Evaluating the vulnerability of their col- number. When an individual / contributes a onies and estimating the effective population fraction v, to a colony's production, the effective size require information on the age structure of number of reproductives in the colonv is the population. l/tx]. Baroni Urbani et al. (1973) give some guide Wilson (1971) noted that polygyny tends to lines for estimating population densities of ants, occur commonly in rare ant species, habitat but very few data exist on the demographic par- specialists and social parasites. He argued that ameters needed for population viability analy- polygyny increases the survivorship of such sis. The mortality of incipient colonies, when colonies and populations and that this may have resulting from independent founding, is high in selected for polygyny in these species. H611- all groups of social insects. It can take a long time dobler and Wilson (1977) developed the idea for successful colonies to reach maturity. In that polygyny is particularly selected for in two annual colonies, this ergonomic stage takes place kinds of ants: those with frequently-fragmenting during a summer, but perennial colonies can colonies (tramp species) and habitat specialists grow several years before they start producing where the habitat patches are fragmented. sexuals. Although the pattern is far from clear, it does Scherba(1963), and Pamilo (1991) estimated suggest that rare species associated to specific annual death rates of nests in two related species habitat types can have locally large effective of Formica ants to be 5-9%, and Scherba esti- populations because of polygyny. Crozier (1979) mated the birth rate to be 5-13%. It is a general further suggested that unicoloniality is an adap- observation in many ants that mature colonies POPULATION BIOLOGY OF SOCIAL INSECT CONSERVATION 415

outnumber young colonies and populations look If one wants to estimate population size in very stable (Wilson, 1971: 445). However, order to evaluate genetic aspects, one should demographic data on the turnover rate of nests estimate the number of effectively reproducing may not tell about the turnover rate of repro- individuals. However, for estimating popu- ductive individuals if a nest has more than one lation viability, it may be sufficient to estimate queen. Wilson (1971) remarked that colonial the birth and death rates of colonies, if one organization serves as a homeostatic device in assumes that these variables do not interact with damping fluctuations in the numbers of individ- changing queen numbers when the population ual insects. It would be important to know if declines. Most life history models of social there is any correlation between the density of insects have focused on the growth of colonies nests and the number of reproductive individ- (Oster and Wilson, 1978: Bourke and Franks, uals per nest in populations where the number of 1995), and in consequence the dynamics of the nests shows temporal variation. Elmes (1987) whole population are still poorly understood. found in a long time series of an ant Myrmica sukinodis in a heathland habitat regenerating Dispersal and fragmentation of populations after a fire, that the number of queens in the nests fluctuated synchronously in the popu- The sedentary life style of social insects affects lation, and the fluctuations seemed to follow a both the turnover of individuals in local popu- cycle of 4-5 years, which probably coincides lations and dispersal and differentiation with the life-span of the queens. between populations. A clear indication of lim- Three categories of colonies can be extremely ited dispersal of females in social insects with stable, and in theory immortal (Wilson, 1971: polygynous colonies is that the coexisting other 444), namely polygynous colonies that recruit queens are genetically related to each new queens, colonies that reproduce by budding (Crazier and Pamilo, 1996), and the females or fission (resembling vegetative growth), and therefore seem to have a tendency to stay in their monogynous colonies that can raise replacement natal colony after mating. reproductives in the case the colony queen (or Dependent nest founding, particularly when includes either member of the royal pair in termites) dies. based on splitting of existing nests, least of females) and is It is an important task for population studies of limited dispersal (at of endangered social insect species to try to esti- expected to lead to genetic differentiation mate the relevant demographic parameters and local populations. A comparison of conspecific closely related species the social organization of colonies. Such infor- populations, or pairs of of colony foundation and mation is needed also for common species, as with different modes indeed is the data on population dynamics of social social organisation, suggest that this from such compari- insects, particularly those with perennial colon- the case. A pattern emerging suggests that populations with colony ies, are few. Yet, these can be keystone species sons differences among with a significant influence on other taxa in the budding have greater genetic with indepen- ecosystem (LaSalle and Gauld, 1993). Vulner- subpopulations than populations (Table 1). The observed pattern ability of even major ecosystem components can dent founding that occasional immigrating females in some cases be reduced to simple population indicates frequently eliminated by local ants. size effects in a few keystone species. have been

populations with different colony types. Differentiation is Table 1 Genetic differentiation of and the colony types are M-M: monogynous and monodomous colonies, measured as FST , and range from 1 to 20 km, except in F. P-P: polygynous and polydomous colonies. The study areas aquilonia. where it is 400 km.

Reference Species FST M-M P-P

Sundstrom(1993) Formica truncorum 0.04 0.19 Pamilo (unpubl.) Formica aquilonia 0.18 — Seppaand Pamilo (1995) Myrmica ruginodis 0.02 Seppaand Pamilo (1995) Myrmica rubra 0.20 416 1'. PAMILO AND R.H. CROZIER

It is interesting to note, as remarked by Wilson that individuals heterozygous for the locus (1971), that the species which reproduce by develop into females (workers or new queens), budding have traded dispersal capacity for while other genotypes develop into males. longevity of the existing nests. This pattern of Normal males are haploid (hemizygous), but colony foundation also means that a species can diploid individuals that are homozygous for a be slow in recolonizing and re-establishing areas sex allele develop also into males. Diploid males where it has gone extinct. This, however, does are shown to be sterile or inviable. It has been not apply to all species in this category, as some proposed that the same, or similar, sex-deter- of the quickly spreading pest ants have a highly mining mechanism exists in most social hyme- polygynous and polydomous colonial struc- nopterans (Cook and Crozier, 1995). There is ture. clearly strong frequency-dependent selection It is still unclear to what extent social insect maintaining allelic diversity at the sex locus. populations are subdivided due to restricted Decreasing variation increases the proportion of dispersal. Hill-topping behaviour of mating diploid males and causes a genetic load. When a swarms has been described in some species, and female mates once and the male and the female genetic differentiation within a homogeneous carry an identical sex allele (matched mating), habitat in a population of Formica transkauca- half of the diploid offspring are males. sica strongly suggests that dispersal of sexual In the introduced fire ant, Solenopsis invicta, individuals is restricted in spite of a lack of in northern America colonies started by a single physical barriers, even though the size of habitat female producing diploid males do not survive is only 200 X 1000 m. distances that could be to maturity (Ross and Fletcher, 1986). Some moved by a flying ant (Pamilo, 1983). Formica ants produce rather high frequencies of The low ratio of population size to biomass, diploid males, and monogynous colonies with and the restricted dispersal at least in many (but matched mating can survive the founding stage not all) social insect species, suggest that any and reach maturity (Pamilo et al., 1994). The analyses of population viability should pay Formica females have dependent nest founding, attention to the spatial scale even in apparently as they take over established nests of another continuous populations (Lande, 1987). species (subgenus Serviformica). The proportion of nests producing diploid males is particularly high in isolated island populations that have very few Inbreeding depression colonies. Diploid males occur in Formica nests only at a One of the on-going debates in conservation time of normal sexual production. During other biology is whether inbreeding depression and times, only workers are produced and diploid genetic erosion contribute significantly to males are apparently eliminated at an early extinction of populations, or whether ecological developmental stage, as also happens in the factors cause extinction before inbreeding has honeybee. This effectively reduces the load time to significantly influence demography. caused by diploid males, once the colony has sur- Hymenopteran species, because of their male- vived the founding stage. As the species have haploidy, have been both predicted (Crozier, evolved mechanisms to eliminate some of the 1985) and shown (e.g.. Pamilo et al., 1984) to load caused by diploid males, it is too early to have lower levels of allozyme variation than conclude how significantly diploid male pro- other insect orders, and variation is particularly duction contributes to the vulnerability of small low in social species. It is, however, difficult to and isolated populations (Table 2). think that such a lower level of heterozygosity In many termites colonies enter a cycle of would largely affect the viability of populations inbreeding, when the primary reproductives die (see also Unruh and Messing, 1993). Another and are replaced. The replacement repro- consequence of male-haploidy is that recessive, ductives originate from the same colony and deleterious mutations are effectively selected inbreed in it, and a cycle of inbreeding can con- against in haploid males, and this can reduce tinue several generations. Such a naturally high inbreeding depression (but see Crozier, 1985, level of inbreeding is effective in eliminating any and Werren, 1993, for some complications). harmful recessive alleles in the same way as There is one reason to predict that social male-haploidy in the Hymenoptera, and it can hymenopteran species can suffer seriously from be predicted that termites are unlikely to suffer inbreeding depression. The sex in the honeybee severely from inbreeding depression when the is determined by a single locus in such a way, population size decreases. POPULATION BIOLOGY OF SOCIAL INSECT CONSERVATION 417

Table 2. Possible genetic effects in sparse and fragmented ant populations as a function of

colony type (M-M and P-P as in Table 1).

M-M P-P

N c small large Genetic differentiation gradual distinct Sex alleles few affected by bottlenecks when colonizing Genetic load Diploid male Dependent nest production can foundation cause a load diminishes the load

Social parasites Wasmannia auropunctata are spreading in dif- ferent islands and continents. They are all Socially parasitic species are known in bees, characterized by highly polygynous and polydo- wasps and ants, but not in termites. Females of mous (or unicolonial) colonial structures, and socially parasitic species take over nests of other, they replace native ants of similar ecological commonly closely related species, and let the requirements. Whereas the Argentine ant occu- host species raise the larvae. The parasite popu- pies largely urban and other disturbed habitats lations are naturally smaller than those of the (at least in Australia), the other three species host species, and the parasites are often very have also penetrated native ecosystems. species specific. Honeybee is introduced and maintained by Social parasitism has according to the so humans, and bumblebees are also cultured and called Emery's Rule evolved between two closely introduced for pollination purposes. These related species, or the parasite has evolved introduced bees, particularly the honeybee, can directly from the host. Parasitism in ants is par- compete with the native bee fauna and can, at ticularly common in leptothoracine ants, and least potentially, have undesirable effects on the many parasitic species are found in isolated local bee diversity (see New (1994) for a populations in mountain areas (Buschinger, discussion on this problem). 1989). Many of the social parasites are One important conservation aspect is to extremely rare and have a restricted geographi- preserve the native faunas, because they play cal distribution. important roles in the native ecosystems. Because of low population densities, socially inbreed frequently parasitic ants are known to Conclusions and the sex-locus system of other ants has been evolutionarily replaced (Buschinger, 1989). It is We have shown that social insects, a major pan evident that they have often small and isolated of many ecosystems, have many differences populations, and parasitic ants belong to those from other invertebrates in characteristics of the rarest and most endangered species. important for conservation. Low dispersal rates Holldobler and Wilson (1990: 212-213) note and small effective population sizes (in relation that parasitic species have a tendency to be to the biomass) make many species liable to be polygynous (as many other rare ants do), which easily endangered (and associated species increases their effective population sizes. threatened). This potential vulnerability is increased for hymenopterans by their male- haploidy and the single-locus sex determination system (Table 2). The long life-span of colonies Introduced species means that population change is slow, necessi- the need for studies to be Some social insects that have been introduced to tating acceptance of is a significant need for new areas, accidentally or on purpose, have led long-term. There manipulation experiments to ascertain the to a need to protect the native fauna. The introduced ants Solenopsis invicta (the extent of ecosystem-wide effects of changes in of the social insect com- imported fire ant), Linepithema humile (the the composition Argentine ant), Pheidole megacephala, and munity. 418 P. PAMILO AND R.H. CROZIER

Acknowledgements disperse a majority of herbs in a mesic forest com- munity in New York State. Bulletin of the Torrey thank the We Environmental Protection Board Botanical ClubWS: 430-437. and the Natural Science Research Council of Hblldobler, B. and Wilson. E.O., 1 977. The number of Sweden for research support to PP, and the Aus- queens: an important trait in ant evolution. tralian Research Council for supporting RHC's Naturwissenschaften 64: 8-15. research; PP's visit to La Trobe University was Holldobler, B. and Wilson, E.O.. 1990. The ants. made possible by a CRA/La Trobe University Springer Verlag: Berlin. Distinguished Visitor award. Keller, L. and Reeve. H.K.. 1994. Partitioning of reproduction in animal societies. Trends in Ecology and Evolution 9: 98-102. References Lande, R.. 1987. Extinction thresholds in demograp- hic models of territorial populations. Baroni-Urbani, C, Josens, G. and Peakin, G.J.. 1 978. American Empirical data and demographic parameters. Pp. Naturalist 130: 624-635. 5-44 in: Brian. M.V. (ed.). Production ecology of LaSalle. J. and Gauld, I.D., 1993. Hymenoptera: their ants and termites. Cambridge Univ. Press: diversity, and their impact on the diversity of Cambridge. other organisms. Pp. 1-26 in: LaSalle, J. and Beattie, A.J.. 1985. The evolutionary ecology of ant- Gauld, I.D. (eds), Hymenoptera and biodiversity. plant mutualisms. Cambridge Univ. Press: CAB International: Oxon, U.K. Cambridge. Leston, D., 1978. A neotropical ant mosaic. Annals of Bourke, A.F.G. and Franks. N.R., 1995. Social the Entomological Society of America 71: 649- evolution of ants. Princeton Univ. Press: 653. Princeton. New Jersey. Lindstrom, K, Berglind, S-A, Pamilo, P. 1996. Vari- Buschinger. A.. 1974. Monogynie und Polygynie in ation of colony types in the ant Formica cinerea. Insektensozietaten. Pp. 862-897 in: Schmidt, Insectes Sociau.x 43: 329-332. GH. (ed.), Sozialpolymorphismus bei Insekten. Majer, J.D., 1993. Comparison of the arboreal ant Wiss. Verlagsges.: Stuttgart. mosaic in Ghana, Brazil, Papua New Guinea and Buschinger, A.. 1989. Evolution, speciation, and Australia — its structure and influence on arthro- inbreeding in the parasitic ant genus Epimyrma pod diversity. Pp. 115-141 in: LaSalle, J. and (Hymenoptera: Formicidae). Journal of Gauld, l.D. (eds). Hymenoptera and biodiversity. Evolutionry Biology!: 265-283. CAB International: Oxon. U.K. Cook. J.M. and Crozier, R.H., 1995. Sex determi- New, T.R. (ed.) 1993. Conservation biology of nation and population biology in the Hymenop- Lycaenidae (Butterflies).WCN: Gland, Switzer- tera. Trends in Ecology and Evolution 10: 281- land. 286. New, T. R., 1994. Exotic insects in Australia. Crozier, R.H. 1979. Genetics of sociality. Pp. 223-286 Glenagles Publ.: Glen Osmond, Australia. in: Hermann. H.R. (ed.), Social insects. Vol. 1. Oster. GF. and Wilson, E.O., 1978. Caste and ecology Academic Press: New York. ofsocial insects. Princeton Univ. Press: Princeton, Crozier, R.H., 1985. Adaptive consequences of male- New Jersey. haploidy. Pp. 201-222 in: Helle, W. and Sabelis, O'Toole, C, 1993. Diversity of native bees and agroe- M.W. (eds), Spider mites, their biology, natural cosystems. Pp. 169-196 in: LaSalle, J. and Gauld. enemies and control, Vol. la. Elsevier: I.D. (eds), Hymenoptera and biodiversity. CAB Amsterdam. International: Oxon, U.K. Crozier, R.H. and Pamilo, P., 1 996. Evolution ofsocial Pamilo, P., 1983. Genetic differentiation within sub- insect colonies, sex allocation and kin selection. divided populations of Formica ants. Evolution Oxford University Press: Oxford. 37: 1010-1022. Elmes. G.W., 1987. Temporal variation in colony Pamilo, P., 1991. Life span of queens in the ant For- populations of the ant Myrmica sulcinodis. I. mica exsecta. Insectes Sociaux. 38: 111-119. Changes in the queen number, worker number Pamilo, P. and Crozier. R.H.,1996. Reproductive and spring production. Journal ofAnimal Ecology skew simplified. Oikos 75: 533-535. 56:559-571. Pamilo, P., Sundstrom, L., Rosengren, R. and Forte- Elmes, G.W. and Thomas, J. A., 1992. Complexity of lius, W., 1994. Diploid males and colony-level species conservation in managed habitats: inter- selection in Formica ants. Ethology, Ecology Evol- actions between Maculinea butterflies and their ution 6: 221-235. ant hosts. Biodiversity and Conservation I: 155- Pamilo, P., Varvio-Aho, S-L. and Pekkarinen, A., 169. 1984. Genetic variation in bumblebees (Bombus, Gilbert, L.E., 1980. Food web organization and the Psithyrus) and putative sibling species of Bombus conservation of neotropical diversity. Pp. 1 1-33 lucorum. Hereditas 101: 245-251. in: Soule. M.E. and Wilcox, B.A. (eds). Conser- Pekkarinen, A.. 1979. Morphometric, colour and vation biology, an evolutionary-ecological perspec- enzyme variation in bumblebees (Hymenoptera, tive. Sinauer Ass. Apidae, Bombus) in Fennoscandia and Denmark. Handel, S.N., Fisch, S.B. and Schatz, G.E., 1981. Ants Acta Zoologica Fennica 158: 1-60. POPULATION BIOLOGY OF SOCIAL INSECT CONSERVATION 419

Pisarski, B., 1978. Comparison of various biomes. pp. Thomas, J.A., 1995. The ecology and conservation of 326-371 in: Brian, M.V. (ed.) Production ecology Maculinea arion and other European species of of ants and termites. Cambridge Univ. Press: large blue butterfly. Pp. 180-197 in: Pullin, A.S. Cambridge, U.K. (ed.), Ecology and conservation of butterflies. Rosengren, R. and Pamilo, P. 1983. The evolution of Chapman & Hall: London. polygyny and polydomy in mound-building For- Unruh, T.R. and Messing, R.H., 1993. Intraspecific mica ants. Acta Entomologica Fennica 42:65- biodiversity in Hymenoptera: implications for 77. conservation and biological control. Pp. 27-52 in: Ross, K.G. and Fletcher, D.J.C., 1986. Diploid male LaSalle, J. and Gauld, I.D. (eds), Hymenoptera production — a significant colony mortality fac- and biodiversity. CAB International: Oxon, tor in the fire ant Solenopsis invicta (Hymenop- U.K. tera: Formicidae). Behavioral Ecology and Socio- Werren, J. H„ 1993. The evolution of inbreeding in biology 19: 283-291. haplodiploid organisms. Pp. 42-59 in: Thornhill, Scherba, G., 1963. Population characteristics among N.W. (ed.), The natural history of inbreeding and colonies of the ant Formica opaciventris Emery outbreeding. University of Chicago Press: (Hymenoptera: Formicidae). Journal of the New Chicago. York Entomological Society 71: 219-232. Williams, P.H., Vane-Wright, R.I. and Humphries, Seppa, P. and Pamilo, P., 1995. Gene flow and popu- C.J., 1993. Measuring biodiversity for choosing lation viscosity in Myrmica ants. Heredity 74: conservation areas. Pp. 309-328 in: LaSalle, J. 200-209. and Gauld, I.D. (eds), Hymenoptera and biodiver- Soule, M.E., 1987. Introduction. Pp. 1-10 in: Soule, sity. CAB International: Oxon, U.K. Harvard M.E. (ed.), Viable populations for conservation. Wilson, E.O., 1971. The insect societies. Cambridge Univ. Press: Cambridge. Univ. Press: Mass., USA. in ecosys- Sundstrom, L., 1993. Genetic population structure Wilson, E.O., 1990. Success and dominance and sociogenetic organization in Formica trunco- tems: the case of the social insects. Ecology Inst.: rum. Behavioral Ecology and Sociobiology 33: Oldendorf/Luhe, Germany. 345-354.