Experimental support for partial compensation, not matched, partner response rules in blue tits

Outcomes of theoretical models on conflict resolution between investing partners in bi-parental care systems typically predict "partial compensation" or "matched" response rules, depending on underlying assumptions. Further, although experimental tests in birds suggest that care levels by pair members are largely associated with partial compensation responses, this outcome partly depends on the type of experiment used and its effects on model assumptions. To elucidate support for both the underlying assumptions and predictions of models predicting partner compensation versus matching, we performed temporary, bi-directional brood size manipulations during late nestling provisioning in blue tits (<i>Cyanistes caeruleus</i>) in the French Pyrenees. We found support for assumptions of both compensation and matching models. For example, females spent more time with the brood than males, leading to potential information asymmetries on brood demand as expected by matching models. Nevertheless, both pair members changed their provisioning comparably in response to brood size manipulations, suggesting that pair members have comparable cost-benefit functions in response to variation in brood demand, as assumed under partial compensation models. Despite support for the underlying assumptions of both models, we primarily found support for predictions of partial compensation models in provisioning responses. Notably, modest changes by one member of the pair on experimental days were met with larger changes by the other, after controlling for brood size and provisioning rates on control days. Our results corroborate previous findings in blue tits using alternative experimental approaches. We suggest that future studies could benefit from understanding when partial compensation responses dominate matched ones, despite apparent asymmetries in information over brood demand.