Selection on the joint actions of pairs leads to divergent adaptation and coadaptation of care-giving parents during pre-hatching care

Abstract

The joint actions of animals in partnerships or social groups evolve under both natural selection from the wider environment and social selection imposed by other members of the pair or group. We used experimental evolution to investigate how jointly expressed actions evolve upon exposure to a new environmental challenge. Our work focused on the evolution of carrion nest preparation by pairs of burying beetles Nicrophorus vespilloides, a joint activity undertaken by the pair but typically led by the male. In previous work, we found that carrion nest preparation evolved to be faster in experimental populations without post-hatching care (No Care: NC lines) than with post-hatching care (Full Care: FC lines). Here, we investigate how this joint activity evolved. After 15 generations of experimental evolution, we created heterotypic pairs (NC females with FC males and NC males with FC females) and compared their carrion nest making with homotypic NC and FC pairs. We found that pairs with NC males prepared the nest more rapidly than pairs with FC males, regardless of the female's line of origin. We discuss how social coadaptations within pairs or groups could act as a post-mating barrier to gene flow.

Keywords: division of labour; nest-building; reproductive isolation.

Figure 1.

(a) Experimental design, showing all possible crosses between lines, including all replicate lines within parental care treatments. Blue boxes indicate the crosses that were made, and red boxes indicate the crosses that were not made. FC = Full Care, NC = No Care; numbers refer to replicate lines. (b) The experimental design is in detail. Four genetically similar lines were generated from a single source population that was split into two replicate FC (in red) and two replicate NC (in blue) lines [34]. The FC lines evolved in an FC environment (red solid lines) and the NC lines evolved in an NC environment (blue dashed lines) for 15 generations (refer [34] for more information about the experimental set-up). After 15 generations of breeding in these contrasting environments, individuals from each population were passed through a common garden regime, in which all broods received FC to minimize variation between lines owing to parental effects. The populations were then crossed. The two control groups involved crossing the populations that had evolved under the same social environment during development: FC pairs were formed by males and females from FC1 and FC2; and NC pairs were formed by males and females from NC1 and NC2. The two experimental groups involved crossing populations that have evolved under different social environments during development. We only crossed one FC line with its NC replicate (FC1 with NC1, and FC2 with NC2) for both sets of crosses. The offspring from each pair were then randomly allocated one of the social environments (FC or NC) as a treatment during larval development.

Figure 2.

The proportion of successful broods when the feeding incision in the carcass was either present or absent prior to larval hatching, in an NC and FC post-hatching environment. In the NC environment, there were n = 138 broods without a feeding incision, and n = 60 where the feeding incision was present. In the FC environment, there were n = 104 broods with an absent feeding incision, and n = 33 with a feeding incision. Note, one pair from the NN pairing in an FC environment was excluded from any analysis involving the presence or absence of the feeding incision as that data were absent. Predicted values (and 95% credible intervals) are shown derived from the model with carcass mass and female and male standardized size as other covariates.

Figure 3.

The predicted brood size (and 95% credible intervals) for all pairs across both NC and FC environments. Treatments are listed with female first and male second (e.g. FN is an FC female paired with an NC male), with red letters indicating an FC parent and blue letters indicating an NC parent. Predicted means are shown and are derived from the model containing the lines of origin of the male and the female and their interaction, as well as all other covariates, even if the model with the interaction is not the favoured model (see table 1).

Figure 4.

The proportion (and 95% credible intervals that describe the posterior distribution of the estimate) of prepared carrion nests found with an incision hole at 53 h after pairing, in relation to both the male and female’s line of origin. Treatments are listed with female first and male second (e.g. FN is an FC female paired with an NC male), with red letters indicating an FC parent and blue letters indicating an NC parent. Predicted means are shown and are derived from the model containing the lines of origin of the male and the female and their interaction, as well as all other covariates.