Relatedness modulates reproductive competition among queens in ant societies with multiple queens

Abstract Reproductive sharing in animal groups with multiple breeders, insects and vertebrates alike, contains elements of both conflict and cooperation, and depends on both relatedness between co-breeders, as well as their internal and external conditions. We studied how queens of the ant Formica fusca adjust their reproductive efforts in response to experimental manipulations of the kin competition regime in their nest. Queens respond to the presence of competitors by increasing their egg laying efforts, but only if the competitors are highly fecund and distantly related. Such a mechanism is likely to decrease harmful competition among close relatives. We demonstrate that queens of Formica fusca fine-tune their cooperative breeding behaviors in response to kinship and fecundity of others in a remarkably precise and flexible manner.

In order to investigate the scenario that queen fecundity variation would be driven by average relatedness of the nest, and apply to queens independent of their treatment/control status, we analysed three linear models, using the lm command in R. Each of them had as the response variable the daily average number of eggs laid by the queen during the experiment (i.e. in the absence of workers, while being exposed to donor queen odours (treatment queens) or solvent control (control queens)). As explanatory variables we used treatment/control status, a relatedness measure and their interaction. The three models differed only in the relatedness measure used.
The first model had r(all) as the relatedness measure (i.e. average of pairwise relatedness estimates among the eight genotyped workers and the treatment and donor queens). None of the explanatory variables had significant effects on queen fecundity (Treatment: t = 0.4, p = 0.69, r(all): t = 1.4, p = 0.17, interaction: t = -0.19, p = 0.85).
The third model had r(w) as the relatedness measure (i.e. average of pairwise relatednesses among the eight genotyped workers). Relatedness r(w) had a positive effect on egg laying, but treatment and the interaction term did not (Treatment: t = 0.05, p = 0.96, r(w): t = 2.2, p = 0.04, interaction: t = 0.18, p = 0.86).
Thus, we find no support for the scenario that treatment queens would respond to nestmate odours based on the relatedness structure of their natal nest on average rather than their pairwise relatedness to the donor queen. In such a case we would have expected to observe an interaction where relatedness has a positive effect on the fecundity of treatment queens (exposed to nestmate odours) but not for control queens (exposed to solvent only, and thus likely to behave as in the absence of nestmate queens).
Supplementary Information: description of data files found at doi:10.5061/dryad.r2280gbh9

total_egg_laying_R1.txt
Variables: colony: ID of the original field collected colony where queens and workers originated from queen: running ID of the queen role: whether the queen was a treatment or a control queen workers: absent: fecundity in the absence of workers, during the experiment, present: fecundity in the presence of workers, before the experiment egg-laying: daily number of eggs laid r_all: average of pairwise relatedness estimates among the eight genotyped workers and the treatment and donor queens r_q: pairwise relatedness between treatment and donor queen r_w: average of pairwise relatedness estimates among the eight genotyped workers q1pre: total number of eggs laid by the treatment queen of the colony in the presence of workers, before the experiment q2pre: total number of eggs laid by the control queen of the colony in the presence of workers, before the experiment q3pre: total number of eggs laid by the donor queen of the colony in the presence of workers, before the experiment q1post: number of eggs laid by treatment queen under exposure to odours from donor queen q2post: number of eggs laid by treatment queen under exposure to solvent