The adaptive significance of male egg carrying in the golden egg bug: defining research avenues. A reply to Härdling et al.

© 2007 The Authors 578 Journal compilation © 2007 The Royal Entomological Society Härdling et al. (2007) present a commentary on our recent paper ( García-González et al. , 2005 ) in which we analysed the paternity of the eggs carried by male golden egg bugs ( Phyllomorpha laciniata Villers; Heteroptera, Coreidae) and determined whether males enclosed with females carried true genetic offspring with a higher frequency than that expected from random oviposition on conspecifics. Phyllomorpha laciniata females exhibit a very flexible pattern of oviposition behaviour: they can lay eggs on host plants ( Paronychia argentea ), where they develop unattended, or on the body of conspecific males and females, where they are carried until hatching ( Mineo, 1984; Kaitala, 1996; Reguera, 1999; Gomendio & Reguera, 2001; García-González, 2002 ). Because most of the eggs carried by conspecifics are carried by males ( García-González & Gomendio, 2003a ), and males carry genetic offspring as well as unrelated eggs ( Tay et al. , 2003; García-González et al. , 2005 ), the adaptive significance of male egg-carrying behaviour has been the subject of controversy ( Gomendio & Reguera, 2001; Kaitala et al. , 2001 ). Härdling et al. (2007) challenged our conclusions based largely on a re-analysis of the data presented in our original paper ( García-González et al. , 2005 ) and made some further comments on this model system that warrant discussion. Härdling et al. (2007) made three major propositions: (a) the data cannot be used as evidence that females preferentially lay eggs on males with whom they have mated or as evidence of nonrandom acceptance of eggs by males with respect to paternity, (b) the data cannot be used as evidence that females preferentially lay eggs on males, and (c) the experimental set-up does not strengthen the null hypothesis of no paternal care. We will discuss each of these in turn before presenting some counterarguments to the general picture drawn by Härdling et al. (2007) for male egg-carrying behaviour in P. laciniata . (a) Acceptance of eggs by males with respect to paternity. Härdling et al . (2007) main concern arises following a re-analysis of the simulations based on random allocation of eggs that we performed in our original paper. They conclude that the rates at which males carry genetic offspring does not differ significantly from the random expectation. In other words, that the results in the experiment can be explained by females laying eggs randomly and males accepting eggs randomly regardless of whether they are the sires or not. However, we believe that Härdling et al. ’s (2007) re-analysis does not serve the purpose of calculating the probability of carrying fertilised eggs under the real (observed) egg-carrying and egg-laying scenarios. In order to do so one must take into account not only the number of eggs that the males sired and the total number of males in the group, but also the number of eggs carried by each male and the number of eggs not fertilised by males in the experimental group. Härdling et al. (2007) seem to ignore this fact in their simulations of random egg allocation as they simply dispensed the eggs carried by males in the group ‘ taking into account that the males had sired different numbers of eggs ’. In the simulations it is crucial to bear in mind that the question to address is whether or not the rates at which males carry true genetic offspring is the result of random egg allocation, and for this reason the number of eggs ‘carried’ by each male has to be considered. In addition, one must consider that females also laid eggs fertilised by males outside the experimental groups, and consequently that males were carrying not only eggs sired by them and other experimental males but also by non-experimental males. The re-analysis by Härdling et al. (2007) neglects this information, and consequently the probabilities arising from their re-analysis are not relevant for the events occurring in the experiment described in GarcíaGonzález et al. (2005) . The question to be addressed with the simulations is not to find out the probability of N m males carrying their own offspring when N E eggs sired by these males are randomly dispersed between them. The question is: What is the probability of male i carrying X genetic offspring when this male has sired Y eggs Correspondence: Montserrat Gomendio, Reproductive Ecology and Biology Group, Department of Evolutionary Ecology, Museo Nacional de Ciencias Naturales (CSIC), José Gutiérrez Abascal 2, 28006 Madrid, Spain. E-mail: montseg@mncn.csic.es The adaptive signifi cance of male egg carrying in the golden egg bug: defi ning research avenues. A reply to Härdling et al.