CORRESPONDENCE Spiteful animals still to be discovered In a recent TREE article, Gadagkar 1 raised the interesting question of whether or not spi teful behaviour e xi sts in animals. The interaction between members of a species is termed spiteful when both the actor and the recipient suffer. Gadagkar1 also coined the term 'weak spite' for a behaviour that harms the recipient without benefit to the actor, and he claims that 'at least two clear-<:ut examples of weak spite behaviour have been documented in the animal kingdom'. We suggest. however, that explanations other than spite may a ccount for the results of the two studies cited by Gadagkar. The first example comes from a study oy Pierotti 2 on the western gull (Larus occirlentalis). Pierotti 2 and Gadagkar1 suggest that pirating food from neighbours and attacking others' chicks are examples of spite ( or weak spite) behavi our. Pirate males have been shown to fl edge fewer offspring that non-pi rate males - hence the hypothesis for spite behaviour. However, Pierotti2 notes that the males that became pirates ' had lost all or most of their offspring as eggs or in the first days after hatchin g'. Therefor e, it should come as no surprise that pirate males fledged fewer -offspring since they became pirates after having lost their offspring! Indeed, Waltz3 suggested that pirating' might be an alternative foraging tactic by individuals which had been less succe ssful in raising offspring and possibly obtaining food c onventionally. Regarding the killing of neighbour's chicks, no fitness loss to the perpetrator is demonstrated (or even suggested) in Pierotti's paper2. and the fact that the chicks are not eaten does not prove that the motive of the attacks is spite. One possible benefit of this behaviour has been suggested Reply from R. Gadagkar Keller et al. claim that spiteful animals are 'still to be discovered' because 'explanations other than spite may account for the results of the two studies' cited by me. I believe that Keller et al. ignore the spirit of my article; my purt>ose was not so much to claim that we are sure of th e spite in the cited examples, but to encourage people to investigate wi th particular care anything that even 'smells' of spite. I think I was prett y explicit about this when I said 'If one wanted to be absolutely certain ... 011e should not yet l abel the behaviour of t he gulls and sticklebacks as spite ... it is possible that ultimately they may be explained as selfish behaviour' 1 . As far as their criticism of Pierotti ' s study of gu11s2 3 is concerned, they say nothing really new compared to what Pierotti2· 3, Waltz4 and 11 have already said. As regards the points raised by them about FitzGerald's study of sticklebackss, one can very easily provide counter arguments against each of their points. For example, Keller et al. argue that FitzGerald's rati onale for the predicted preference for conspecific eggs over heterospecific eggs is TREE vol. 9, no. 3 March 1994 by Pierotti 2 himself, who states that decreasing t he breeding success of other indivi duals may y ield some benefits by decreasing the competition that one's offspring ' may face in the crucial first few weeks after becoming independent'. Interestingly, piracy was more common during the year when food supply was poor, suggesting that competitive interJCtions indeed increase under harsher ecological conditions. The second ex ample is a st1,1dy on threespine stickleback (Gasterosteus aculeatus) females , which frequently attack conspecific nests to eat eggs 4 . FitzGerald 4 argued that three predictions must be upheld for conspecific egg-eating to be spiteful: the fish must prefer (1) conspecific eggs over heterospecific eggs; (2) conspecific eggs from a sympatr ic population compared to tho se from an allopat ric population; and (3) older conspecific eggs over younger ones (older eggs being expected to be more valuable to their parent's fitness). Since the three predictions were verified in choice ex periments, FitzGerald4 and Gadagkar1 concluded that conspecific egg- eating by females is spiteful or weakly spiteful behaviour. This conclusion suffers, however , from at least two caveats. First. the rationa le for prediction (2) is unclear , since females will never have a chance to feed on non-sympatri c eggs (the two selected populati ons were 200km apart) . Furthemiore, a theoretical analysiss that weak spite might be selected for under some restricti ve conditions, but only when directed toward individuals having less-than- average relatedness (giving the opposite prediction to that of FitzGerald). Second, and most i mportantly, there is a simple alternative unclear, since females will never have a chance to feed on non-sympatr ic eggs. It is entirely possible that there is a continuous variation in, say, the smell of eggs with distance, and that this may be sufficient to drive the evolution of a pr eference for non-sympatric over sympatric eggs. There is no need for a previous direct experience with non-sympatric eggs. Secondly, they claim that Hamilton's m0del6 leads to a prediction which is the opposi te of what is by FitzGerald . This is not at all obvious. In spite of Hamilton's model. t he predicti on will depend on the detai ls of population genetic structure, about which we know nothing. Finally, they claim that their alternative explanation of familiarity is sufficient because the fish FitzGerald used must have 'been used to eat conspecific eggs' . The question, then, is why were fish accustomed to eati ng conspecific eggs when heterospecific eggs were available? Keller et al. know as well as I do that these hair-splitting arguments ( theirs and mi ne) are not going to take us very far, and that more controlled ex periments are needed. But I wish to emphasize that their arguments against spite are explanation to account for the observed egg preferences. In all three experiments, females may have preferred familiar food , a phenomenon documented in many fish speciess. Since FitzGerald caught sexually mature, gravid females which had been used to eat conspecific eggs, conspecific eggs must have been more famili ar than heterospecific ones, sympatric eggs more than allopatric ones, and older eggs more familiar than freshly spawned ones. Furthermore, a female might prefer older eggs because freshly spawned ones are more likely to be hers under natural conditions. In conclusion, although we agree with Gadagkart that apparently paradoxi cal behaviour should not be overlooked, we believe that 'clear- cut' ex amples of spi te, or weak spite. behaviour are still to be discovered. We thank the Swiss National Science Foundation for support . L. Keller, M. Milinski, M. Frischknecht, N. Perrin, H. Richner and F. Tripet Zoolog!sches lnstitut, Bern University, Ethologische Station Hasll, Wohlenstrasse 50a, 3032 Hlnterkappelen, Switzerland ' References 1 Gadagkar, R. (1993) Trends Ecol. Evol. 8, 232-234 2 Pi erotti , R. ( 1980) Am. Nat. 115. 290- 300 3 Waltz, E.C. (1982) Am. Nat. 118, 588-592 4 FitzGerald, G.J. (1992) Behav. Ecol. Sociobiol. 30,201-206 5 Hamilton, W.D. (1970) Nature 228, 1218-1220 6 lvlev, V.S. (1961) Experimental Ecology of the Feedin g of Fishes, Yale University Press not any stronger than those in favour of spite. If their conclusion is that we should therefore ignore spite on the grounds of parsimony, then I disagree, because I think that by admitting the possibility of spite we will pay more attention to these behaviours than if they are simply considered as more examples of selfish behaviour. Raghavendra Gadagkar Centre ror Ecological Sciences and Jawaharlal Nehru Centre for Advanced Scientific Research, Indian Institute of Science, Bangalore 560 012, India References 1 Gadagkar, R. (1993) Trends Ecol. Evol. 8. 232- 234 2 P ierotti, R. ( 1980 ) Am. Nat. 115, 290-300 3 Pierotti, R. (1982) Am. Nat. 119, 116- 120 4 Waltz, E.C. (1981} Am. Nat. 118, 588- 592 5 FrtzGerald, G.J. (1992) Behav. Ecol. Sociobiol. 30,201-206 6 Hamilton, W.D. (1970) Nature 228, 1218-1220 103