Data from: Multiple mating reveals complex patterns of assortative mating by personality and body size

1. Understanding patterns of non-random mating is central to predicting the consequences of sexual selection. Most studies quantifying assortative mating focus on testing for correlations among partners’ phenotypes in mated pairs. Few studies have distinguished between assortative mating arising from preferences for similar partners (expressed by all or a subset of the population), versus from phenotypic segregation in the environment. Also, few studies have assessed the robustness of assortative mating against temporal changes in social conditions. 2. We tracked multiple matings by stream water striders (Aquarius remigis) across variable social conditions to investigate mating patterns by both body size and behavioural type (personality). We documented temporal changes in partner availability and used a mixed model approach to analyse individual behaviours and changes in mating status recorded on an hourly basis. We assessed whether all or only a subset of individuals in the population expressed a tendency to mate with similar phenotypes. Our analyses took into account variation in the level of competition and in the phenotypes of available partners. 3. Males and females exhibited significant assortative mating by body size: the largest males and females, and the smallest males and females mated together more often than random. However, individuals of intermediate size were equally likely to mate with small, intermediate or large partners. Individuals also displayed two contrasting patterns of assortative mating by personality (activity level). Individuals generally mated preferentially with partners of similar activity level. However, beyond that general trend, individuals with more extreme personalities tended to exhibit disassortative mating: the most active males mated disproportionately with less active females, and the least active males tended to mate with more active females. 4. Our analyses thus revealed multiple, distinct patterns of non-random mating. These mating patterns did not arise from differences in partner availability among individuals and were robust to temporal changes in social conditions. Hence mating patterns likely reflect mate preferences or arise from male – male competition coupled with sexual conflict. Our study also stresses the importance of accounting for variation in partner availability and demonstrates the influence of behavioural variation on mating patterns.

1. 阐明非随机交配（non-random mating）的模式，是预测性选择（sexual selection）后果的核心学术议题。当前多数量化选型交配（assortative mating）的研究，均聚焦于检验交配配对中伴侣双方表型（phenotype）间的相关性。鲜有研究区分两种成因的选型交配：一种源于种群全体或部分个体对相似伴侣的偏好，另一种则源于环境中的表型分离现象。此外，极少有研究评估选型交配对社交环境随时间变化的鲁棒性（robustness）。
2. 本研究对溪流尺蝽（stream water striders，学名*Aquarius remigis*）在多变社交环境下的多次交配事件进行追踪，以探究基于体型与行为类型（即个性，personality）的交配模式。研究记录了伴侣可获得性随时间的动态变化，并采用混合模型（mixed model）方法，分析以小时为单位记录的个体行为与交配状态变化。我们评估了种群中究竟是全体个体，还是仅部分个体，表现出与相似表型个体交配的倾向。本研究的分析过程同时考量了竞争强度与可获得伴侣表型的变异情况。
3. 雌雄个体均表现出基于体型的显著选型交配：体型最大的雄性与雌性、体型最小的雄性与雌性，其交配频率显著高于随机配对的预期频率。但体型中等的个体，与小型、中型或大型个体交配的概率并无显著差异。此外，个体基于个性（即活动水平，activity level）的选型交配还呈现两种截然相反的模式：总体而言，个体倾向于优先选择活动水平相似的伴侣；但在这一总体趋势之外，个性更为极端的个体往往表现出反选型交配（disassortative mating）：活动水平最高的雄性，与活动水平较低的雌性的交配比例显著失衡，而活动水平最低的雄性则倾向于与活动水平更高的雌性交配。
4. 综上，本研究的分析揭示了多种截然不同的非随机交配模式。这些交配模式并非源于个体间伴侣可获得性的差异，且对社交环境的时间变化具有鲁棒性。因此，交配模式大概率反映了配偶偏好，或是雄性间竞争与性冲突（sexual conflict）共同作用的结果。本研究同时强调了考量伴侣可获得性变异的重要性，并证实了行为变异对交配模式的显著影响。