Experiments under controlled laboratory conditions can produce decisive evidence for testing biological hypotheses, provided they are representative of the more complex natural conditions. However, whether this requirement is fulfilled is seldom tested explicitly. Here we provide a lab/field comparison to investigate the identity of an egg-marking signal of ant queens. Our study was based on ant workers resolving conflict over male production by destroying each other's eggs, but leaving queen eggs unharmed. For this, the workers need a proximate cue to discriminate between the two egg types. Earlier correlative evidence indicated that, in the ant Pachycondyla inversa, the hydrocarbon 3,11-dimethylheptacosane (3,11-diMeC₂₇) is more abundant on the surface of queen-laid eggs.

We first tested the hypothesis that 3,11-diMeC₂₇ functions as a queen egg-marking pheromone using laboratory-maintained colonies. We treated worker-laid eggs with synthetic 3,11-diMeC₂₇ and found that they were significantly more accepted than sham-treated worker-laid eggs. However, we repeated the experiment with freshly collected field colonies and observed no effect of treating worker-laid eggs with 3,11-diMeC₂₇, showing that this compound by itself is not the natural queen egg-marking pheromone. We subsequently investigated the overall differences of entire chemical profiles of eggs, and found that queen-laid eggs in field colonies are more distinct from worker-laid eggs than in lab colonies, have more variation in profiles, and have an excess of longer-chain hydrocarbons.

Our results suggest that queen egg-marking signals are significantly affected by transfer to the laboratory, and that this change is possibly connected to reduced queen fertility as predicted by honest signaling theory. This change is reflected in the worker egg policing response under field and laboratory conditions.