Forum The eusociality continuum ing vertebrates can be regarded as eusocial, just as eu- social invertebrates are cooperative breeders. We believe Pud W. Sherman this integrated approach will foster potentially revealing Section of Ncurobiology and Behavior, cross-taxon comparisons, which are essential to under- Cornell University, standing social evolution in birds, mammals, and in- Ithaca, NY 14853. USA sects. Key words: avion eusociality, cooperative breed- ing, eusociality, mammalian eusociality, reproductive Eileen A. Lacey skews, social system convergence. [Behav Ecol 6:102- Animal Behavior Croup, 108 (1995)] University of California, Davis, CA 95616. USA The evolution of eusociality has been an important Hudson K. Reeve puzzle ever since Darwin (1859: 268) identified Section of Neurobiology and Behavior, worker ants as presenting "one special difficulty, Cornell University, which at first appeared to me insuperable, and ac- Ithaca, NY 14853, USA tually fatal to the whole theory." In 1966, Batra coined the term eusocial (meaning truly social) to Laurent Keller Zoological Institute, describe halictine bees in which "the nest founding Bern University, parent survives to cooperate with a group of her Ethologische Station Hasli, mature daughters, with division of labor" (p. 375). CH-3052 Hinterkappelen, Subsequently, Michener (1969: 305) referred to Switzerland bees as eusocial if they lived in "matrifilial family and groups consisting of.. mothers and daughters. .. Institute of Zoology and Animal Ecology, [showing] division of labor with more or less rec- University of Lausanne, ognizable castes (egg layers and workers)." Bitiment de Biologic, In 1971, Wilson broadened these criteria to in- CH-I0I5 Lausanne, Switzerland clude other insects. Following his lead, Holldobler and Wilson (1990: 638) denned eusocial species as those exhibiting "cooperation in caring for the Eusocial societies are traditionally characterized by a young; reproductive division of labor, with more reproductive division of labor, an overlap of generations, or less sterile individuals working on behalf of in- and cooperative care of the breeders' young. Eusociality dividuals engaged in reproduction; and overlap of | downloaded: 2.10.2021 was once thought to occur only in termites, ants, and at least two generations of life stages capable of some bee and wasp species, but striking evolutionary contributing to colony labor." Once thought to convergences have recently become apparent between the occur only in the orders Hymenoptera (ants, bees, societies of these insects and those of cooperatively breed- and wasps) and Isoptera (termites), eusociality has ing birds and mammals. These parallels have blurred now been reported in Japanese aphids (Homop- distinctions between cooperative breeding and eusocial- tera: Aoki, 1982; I to, 1989), Australian weevils (Co- ity, leading to calls for either drastically restricting or leoptera: Kent and Simpson, 1992), Australian expanding usage of these terms. We favor the latter thrips (Thysanoptera: Crespi, 1992), and African approach. Cooperative breeding and eusociality are not mole-rats (Rodentia: Burda and Kawalika, 1993; discrete phenomena, but rather form a continuum of Jarvis and Bennett, 1993;JarvisetaI., 1991,1994). fundamentally similar social systems whose main dif- As detailed information has accumulated on the ferences lie in the distribution of lifetime reproductive reproductive and social behavior of vertebrates and success among group members. Therefore we propose to invertebrates, distinctions between eusociality and array vertebrate and invertebrate cooperative breeders other social systems have become blurred. Indeed, along a common axis, representing a standardized mea- a number of authors have identified striking evo- sure of reproductive variance, and to drop such (loaded) lutionary parallels between the social systems of terms as "primitive" and "advanced" eusociality. The cooperatively breeding birds and mammals and terminology we propose unites all occurrences of olio- those of social insects (e.g., Alexander etal., 1991; parental helping of kin under a single theoretical um- Andersson, 1984; Emlen et al., 1991; Lacey and brella (e.g., Hamilton's rule). Thus, cooperatively breed- Sherman, 1991; Reeve and Sherman, 1991; Veh- https://doi.org/10.7892/boris.116135 102 Behavioral Ecology Vol. 6 No. 1 source: rencamp, 1979). Further, as Seger (1991: 346) not- the distinguishing attributes (e.g., morphological ed, detailed studies of insects have revealed a broad differentiation of colony members) is continuous, spectrum of social organizations among species tra- rather than discrete, both widiin and among taxa. ditionally characterized as eusocial (e.g., see Keller, Finally, the terms "primitive" and "advanced" are 1993; Michener, 1985; Ross and Matthews, 1991). both value-laden and ambiguous, as they may refer Not surprisingly, therefore, several authors (e.g., either to social complexity (sensu Michener, 1969) Crespi and Yanega, 1994; Tsuji, 1992) have re- or similarity to presumed ancestral forms (sensu cently questioned the adequacy of traditional def- Carpenter, 1991). initions of eusociality. Problems have arisen pri- To resolve these ambiguities, we propose using marily because the key denning characteristic— variation in lifetime reproductive success (LRS) "reproductive division of labor, with more or less among members of cooperatively breeding social sterile individuals working"—is vague and thus am- groups to quantify "reproductive division of la- biguous in its application. One solution is to define bor." Reproductive differences are central to all eusociality more narrowly. This approach has been definitions of eusociality, and they underlie much adopted by Tsuji (1992) and Crespi and Yanega of die diversity among vertebrate and invertebrate (1994), who argue that the term should be applied societies (see Bourke, 1991; Vehrencamp, 1979). to only a subset of the insects currendy recognized Such differences result from social competition and as eusocial. Alternatively, definitional problems suppression within groups as well as ecological fac- could be reduced by expanding the eusociality con- tors that preclude reproduction by some group cept to include all vertebrate and invertebrate so- members. Differences in LRS provide an evolu- cieties with helpers. tionarih/ relevant basis for interspecific compari- We favor the latter approach. It seems more pro- sons because it is through such differences that ductive to recognize that similar social systems oc- natural selection shapes the morphology, physiol- cur in birds, mammals, and insects than to debate ogy, and behavior of eusocial species. whedier particular insects are eusocial (e.g., Furey, One could standardize LRS variation in numer- 1992 versus Tsuji, 1992). Behavioral convergences ous ways. One possibility is the index of reproduc- between eusocial insects and cooperatively breed- tive skew (5) developed by Reeve and Ratnieks ing vertebrates should long ago have focused our (1993) and Keller and Vargo (1993): attention on common selective factors favoring so- ciality and alloparental care in these taxa (see Strassmann and Queller, 1989). Research on these N> + N. groups has proceeded largely independently, how- where Nm is the number of nonbreeding alloparents ever, and as a result there is currendy one set of (helpers) in a group, Nt is the number of breeders evolutionary explanations for cooperative breeding in the group (some of which may also behave as in birds and mammals (e.g.. Brown, 1987; Emlen, alloparents), and v is a measure of the variation in 1991; Jennions and Macdonald, 1994) and a par- reproductive success among breeders. In groups allel, but distinct, set of explanations for sociality containing a single breeder, v is defined as 1.0; in in insects (e.g., Seger, 1991; Trivers, 1985). We groups widi multiple breeders, v is the variance suggest that the evolution of sociality in both groups among breeders in their proportion of die summed will be best understood if these explanations are LRS of the group divided by the maximum possible merged. value for diis variance. Thus, v — N^s*, where J1 is As a first step toward this unification it would be the sample variance in die proportion of total off- useful to have a quantitative way to compare social spring produced by breeders: systems across diverse taxa. Current schemes for comparing insect societies are qualitative, however, and emphasize traits that result only secondarily from reproductive differences among colony mem- bers. For example, some authors (e.g., Cowan, 1991; (N»- 1) Eickwort, 1981; Michener, 1974) distinguish "ad- (in diis expression, p, is die proportion of offspring vanced" from "primitive" eusociality. Advanced produced by die idi breeder). eusocial species inhabit large, long-lived colonies Using S, one can begin to compare die degree containing worker* that typically are unable to mate of reproductive skew widiin and among social spe- and that are well-differentiated morphologically cies on a common scale diat ranges from 0 to 1. from queens, whereas primitively eusocial species When LRS is equal among group members, S — 0; live in small, often annual colonies containing work- when reproduction is restricted to a single individ- ers that are morphologically similar to queens and, ual and odier group members never breed, S =• 1. usually, capable of mating. If, as seems likely, skews vary considerably among The advanced-primitive dichotomy was