AbstractCommunity ecology and sociobiology share a number of problems (e.g. understanding part-to-whole relationships), so sharing methods may be beneficial. In this paper, we re-analyze a large social network database for Camponotus fellah ants, from the perspective of keystone individuals, based on methods developed in community ecology. We study different network types for six colonies, over a time series of 10 days. These include the total network of weighted interactions, the subnetworks for the castes and the subnetworks for Queen-linked and Queen-independent individuals. We perform time-independent (i.e. average values) and temporal (i.e. trends) analyses. The most commonly used global network metrics showed high variability in time but no clear trends. Yet, the variability of a network centrality index (topological importance, TI), developed earlier in community ecology, shows clear increase in each colony over time, suggesting the emergence of keystone individuals as a general tendency in all studied ant colonies.