And God created ants

When the great unsung English natural historian the Revd William Gould wrote his book, English Ants, in 1747, he penned the following words: "There is no point of philosophy more difficult to resolve than final causes, or the particular ends designed by providence in the various part of the creation." Even so, Gould was not deterred from speculating on the reasons ordained by the Almighty when He created the variety of ants, each with specific behaviours.

Gould's words ring equally true for post-Darwinian scientists when natural selection is substituted for "providence" and evolution for "creation". Similarly, the difficulties have not stood in the way of modern biologists speculating on the origins of insect societies. In particular, the resolution of how social insects could evolve sterile castes (workers) by natural selection, has proved to be an especially difficult problem for evolutionary biologists. The concept of kin selection and inclusive fitness developed by William Hamilton in the 1960s provided a solution and heralded a new more analytical approach to insect social biology, outlined in 1971 by E.O. Wilson in The Insect Societies, and in 1975 in Sociobiology: the New Synthesis. Together, these books were a major influence on the approach and work of the next generation of invertebrate social biologists.

The majority of the new generation concentrated on describing and quantifying the selective pressures that have led to the complex societies of the true (eusocial) social insects - all ant species, the social bees and wasps, and the termites. The first three belong to the order Hymenoptera and although termites are often called "white ants" they belong to a different insect order, the Isoptera, that is taxonomically much closer to cockroaches than to ants. However, these groups are the pinnacles of social evolution among the insects. There is a plethora of lesser peaks, represented by taxa which show some of the attributes of social behaviour, many of which were first observed and documented by naturalists working in the 19th century.

Scientists generally feel that nature is more easily understood when it is categorised and labelled, and social behaviour was no exception to this rule. Various schemes were in use until Wilson promoted and popularised a four-part hierarchical classification that had been devised by his contemporary, the American social biologist Charles Michener. This was based on possession of three social traits - co-operative brood care, specialised reproductive castes and overlap between generations. Eusocial insects have all three traits while subsocial and non-social species have none. Semi-social insects have no overlap of generations whereas quasi-social insects show only co-operative brood care. However, the implied exclusion of many sub-social species was problematic; for example, many species live communally and show parental care of brood, traits which need certain social skills.

Some of the new generation of social biologists revisited the non-eusocial groups with an evolutionist's eye, in an attempt to understand more fully the evolution of eusociality. The fruits of these labours are synthesised in this new book, Social Behaviour in Insects and Arachnids. It is a multi-authored text that focuses firmly upon the overall evolutionary message. Despite this, the essential enthusiasm of the individual scientists for their organisms shines through many of the chapters. The editors provide a short introduction, and they write the final chapter which is an evolutionary synthesis. Following several recent reviews of insect sociality, they adopt a modified classification of sociality from that of Michener. The need for overlap of generations is discarded from eusociality and three new traits are introduced: brood care, shared breeding site and alloparental or shared brood care. The first chapter is a short review of the new classification and its associated terminology, in which the author cautions against becoming straitjacketed by terms and classifications. The result is that the editors have cast their net broadly and drawn together a wide variety of social behaviours from a considerable range of species.

On average each chapter is 22 pages and is a self-contained mini-review of a specialised topic. Nevertheless, the basic biology outlined by many of the authors is fascinating. Take, for example, the aphids (Homoptera). Aphids are clones, that is they are produced asexually and offspring are identical to their parents. Given the importance of kin selection in the evolution of eusociality, aphids should in theory evolve high levels of altruism and reproductive specialisation. Yet they do not. Why? Chapter 7, written by David Stern and William Foster, gives us some of the answers. The problem is that aphid lineages are not permanently physically associated, there is a considerable probability that any aphid colony will encounter individuals of different lineage. A highly altruistic clone would immediately be open to exploitation by a less altruistic one. Consequently, high levels of altruism have not evolved. Aphid colonies are generally static and exposed, so that even with low levels of altruism, there has been considerable selection pressure to evolve individual specialists which defend their clones that attack.

It is impossible here to give a detailed review of such a wide-ranging book. Excluding the first and last, there are 22 chapters arranged in a more or less traditional taxonomic sequence. Representative species from more than a quarter of the orders of insects are discussed. The Hymenoptera (seven chapters) get the most coverage, although the emphasis is not upon the highly eusocial taxa. For similar reasons there is only one article on sociality in termites (Isoptera) which is complemented by one on cockroaches (Dictyoptera). Four chapters are devoted to the Coleoptera - beetles including weevils - while Hemiptera are represented by aphids and giant water bugs. The Lepidoptera, Thysanoptera (thrips) and Embioptera each get a single chapter. Arachnids are included in the title more for technical accuracy than for the quantity of content, there being one chapter on mites (Acari) and two on spiders (Araneae).

Although the editors have successfully imposed some sort of common style upon the individual contributors and persuaded them to emphasise the evolutionary aspects of their subject, each chapter must be judged on its own merits:some are more technical and harder to read than others. The book is packed with information which will make it a valuable reference source. It will be especially useful as an introduction to "new groups" for professionals such as myself, who specialise in the highly eusocial taxa. But, with the paperback edition priced at over Pounds 30, I doubt whether the book will be purchased by a wider audience despite the fascinating natural history it encompasses. It is no light reading and it requires considerable determination to read the book in its entirety. The detailed evolutionary arguments that are presented in most chapters are probably beyond the requirements of the average student, school teacher or naturalist.

Graham W. Elmes is seniorprincipal scientific officer, Furzebrook Research Station, Natural Environment Research Council.