The mysteries of ants and peacocks raised by Darwin have been the major driving forces for the advance of sociobiology and behavioural ecology of recent decades (The Ant and The Peacock by Cronin, Cambridge University Press, 1991). The role of parasites has already been extensively discussed in the evolution of sex and sexual selection. This ‘red queen’ paradigm has come into effect on ‘the ant issue’, i.e. social evolution. Paul Schmid-Hempel’s book Parasites in Social Insects might be reviewed in such a context of the history of science. However, I am afraid that such a historical view may conceal the tremendous importance of this book, which is definitely instructive to future directions of social insect biology. While some chapters and the huge list of parasites with social insects in Appendix 2 are good reviews of empirical knowledge, the book is mostly conceptual, addressing many new questions and proposing many new hypotheses. Its main asset is that it provokes future empirical and theoretical studies in parasite – social insect interaction rather than presenting conclusive remarks based on what is already known. In social insect evolutionary biology, it can be compared with Oster and Wilson’s Caste and Ecology in the Social Insects (Princeton University Press, 1978), which was also very conceptual and has played a similar instructive role.

While Schmid-Hempel wrote, factors in all modesty that parasites would be clearly just one of many, he has also correctly justified his stress on the potential import- ance of parasites in ecology and evolution of social insects’ life; despite the logical possibilities we generally lack empirical knowledge about this issue. In fact, reading this book one will realize that almost all sociobiological issues in social insects, e.g. division of labour, the number of queens, queen mating frequency, colony size, life span of workers, sex ratio and kin recognition, etc., can be re-interpreted along the lines of the dynamics of host–parasite interaction. For example, one of the most intriguing discussions in this book is on the age-related division of labour (age polyethism) which is a general phenomenon among workers of social Hymenoptera and some termites. The usual order of events is that youngsters stay in the nest caring for brood and maintaining nests, and older individuals defend the nests and finally forage. This pattern has been interpreted as the colony level adaptation; risky outdoor tasks are better taken on at the end of and individual worker’s life in order to maximize both colony efficiency and workers’ inclusive fitness (e.g. J. Theor. Biol.93: 153, 1998).

However, the labour in European honeybee colonies is divided not only into within-nest and outside-nest tasks but also more finely to some extent, e.g. youngest workers clean cells, and second youngests care for broods and the queen and so on. While proximate explanations may be able to explain this pattern (Social Evolution in Ants by Bourke & Franks, Princeton University Press, 1995) and the generality of this pattern is still in dispute (The Wisdom of the Hive by Seeley, Harvard University Press, 1995), risk difference does not seem to explain the phenomena fully.

Schmid-Hempel proposes a new hypothesis that age polyethism is a colony-level adaptation to reduce the load of epidemic disease. If parasites invade a colony by first infecting a forager working outside the nest, limited interactions among workers, i.e. more contacts among members of the same task group but fewer contacts among different task groups, should impede disease transmission within the colony. Schmid-Hempel explains this using a conveyer belt analogy, i.e. older infected workers are replaced by young uninfected workers time after time. Often larvae of the honeybee are more sensitive to epidemic disease than adults, and furthermore infection of the queen is fatal to the entire colony. Larval care and queen attendance by the youngest workers, least likely to be infected, can be an effective antidisease strategy. Variation of lifespan of sterile workers, which had almost no logically convincing evolutionary explanation so far, can also be discussed using the conveyer belt analogy. A quick conveyer belt, i.e. short-lived workers, is an effective counter strategy for a high disease load in some situations, because a colony can more quickly ‘metabolize’ working force.

Importantly, many of these concepts have been analysed by explicit mathematical models (though most of them were basic and do not cover each special case of biological complexity) which have provided a number of testable predictions. The extremely wide scope of this book, however, may make it difficult for readers to organize their own studies to focus on a few relatively important issues. Therefore, I will try to suggest a few topics of general interest in social evolution. The first issue is colony size and infection risk. There has been no general consensus on the evolutionary explanation for inter- and intraspecific variation of colony size with an explicit biological mechanism. It is well known that larger colonies tend to have larger reproductive output, whereas per capita (per worker) reproductive output often decreases as colony size increases (reproductivity effect). Schmid-Hempel sets out a possible mechanism for this, namely that larger colonies will have a higher infection risk. His rationale seems logically robust; if there is a given fixed probability of new infection equal for all forager workers, there is a higher probability that a disease is imported into larger colonies. Although interspecific comparisons presented in the book did not fully support the above idea, Schmid-Hempel uses the assumption of positive correlation between colony size and disease load in order to deduce many other hypotheses. I feel that empirical and theoretical studies should pay careful attention to this topic, since colony size is one of the most important characteristics that affect many other traits of social insects (J. Evol. Biol.12: 245, 1999).

Secondly, disease and genetic variability particularly related to queen multiple mating. Since this topic has already attracted much attention, the current status is not only exploring theoretical models but also empirically testing of models predictions. While the genetic bet-hedging hypothesis as the cause of multiple mating is recently at a little discount for solitary organisms (Tren. Ecol. Evol.13: 246, 1998), in social insects genetic variation among worker progenies caused by queen multiple mating may enhance the mother queen’s fitness. The key factor is the special population structures where related individuals frequently interact with each other, under which genetic variability among offspring can reduce the disease load by ‘half sib-co-operation’ (Tren. Ecol. Evol.13: 246, 1998). I believe interdisciplinary discussions on this theme among students of sexual selection, social evolution and life history strategy will be fruitful.

Finally, the evolution of virulence and resistance and the host–parasite population dynamics, discussed in Chapters 7 and 6, respectively, are also very fundamental topics. To advance further we particularly need empirical data, as Schmid-Hempel states. We should seriously seek for model organisms, like Drosophila in the biology of solitary animals, for experimental studies in laboratories and seminatural conditions. Honeybee colonies, the best known domesticated systems for disease issues, seem too bulky for this aim. Some ants and bumblebees could be the candidates for the host materials. We also have to choose corresponding parasites which are easily tractable.