In social organisms cooperation is widespread, however social groups also provide excellent opportunities for individuals to exploit the cooperative efforts of others, thus creating conflict. Ants provide some of the most intriguing examples of social conflict in nature, for example when colony members differ in their interests concerning male parentage. Such conflicts are often mediated by kinship, and the ability of individuals to assess kinship within colonies plays an important role in conflict resolution. However, whilst much is known about conflict in adult ants, the social role of larvae has been largely neglected. We investigated a novel social conflict in ants by measuring competition among larvae in the form of egg cannibalism. We found that first instar larvae of the ant F. aquilonia eat eggs and that levels of cannibalism differed depending on the origin of eggs. While 10% of larvae ate sibling eggs, 24% of larvae ate eggs from a foreign population. Discriminatory behaviour matched population-specific egg odours, providing first evidence that larvae detect and react to chemical cues in a similar way as adult ants. Cannibalistic larvae showed prolonged survival compared to non-cannibals, indicating that cannibalism bears benefits for ant larvae. We also found that males cannibalised significantly more often than females, which points to sex-dependent selection for selfish behaviour. These results suggest that larvae are not the powerless individuals they have been seen as so far. Instead, developing offspring and male larvae in particular may be important players in social conflict in ants, and social animals in general.
One major question in evolutionary biology is how non-reproductive workers have evolved in eusocial organisms. The ‘haplodiploid hypothesis’ attempted to explain why eusociality seems to have evolved more times in haplodiploids than in non-clonal diplodiploids. That theory was proven inadequate by Trivers and Hare (1976) who showed, however, that haplodiploidy could promote the evolution of eusociality provided the existence of variation in sex ratio among different broods (‘split sex ratios’). Gardner et al. (2012) reviewed the empirical progress in the understanding of the evolution of eusociality, and suggested that the potentially important causes of split sex ratios at the origin of helping are limited to queen virginity (some queens not mating and producing all-male broods) and queen replacement. After Gardner et al’s correction on the role of queen replacement has been taken into account, only queen virginity seems to be a possible cause for split sex ratios at the origin of helping in haplodiploids. However, when comparing haplodiploid and diplodiploid systems, Gardner et al. neglect the fact that unmated diplodiploid females have zero reproductive success. The analysis we propose reveals that unmatedness promotes the evolution of helping more in diplodiploids than in haplodiploids. Considering altogether Gardner et al’s results and our own, we conclude that the factors causing split sex ratios in haplodiploids do not make eusociality more likely to evolve in haplodiploids than in diplodiploids.
Understanding how the interaction of genotypes and environment produces distinct phenotypes from similar sets of genes is a crucial topic in evolutionary biology. The queen and worker castes of social insects are a striking example of such polyphenism. In social Hymenoptera both the workers and reproductive queens are females, and develop from diploid eggs. In ants, caste polyphenism is characterized by extensive phenotypic differences, with queens that specialise in reproduction being usually much larger than workers who carry out brood care, foraging, nest maintenance and defence.
Because the phenotypically distinct individuals develop from the same set of genes, these morphological, physiological and behavioural differences found between worker and queen must be due to variation in gene expression. However, little is known about qualitative and quantative variation in caste specific gene expression patterns across life stages. Because expression patterns of genes partly determine the strength of natural selection on them, separation of gene expression differences in developing and adult individual is crucial for understanding both the causes and consequences of caste differences. Specifically, it is crucial for strength of natural selection whether there are genes that are exclusive to one caste, or whether the same genes are expressed in both castes, but at different points in time.
We used Formica exsecta, a monogynous ant species, as our model species for this study. We assessed queen and worker gene expression levels at three different life stages (pupae, emerging adult and adult) by sequencing their transcriptome. Samples were collected in Finland during Spring 2011 and two sets of cDNA library were made. As phenotypic plasticity mediated by gene expression differences is the foundation of social insect evolution, we believe that this novel study will bring further knowledge on understanding the molecular basics of phenotypic evolution.