Computational methods for the characterization of Apis mellifera comb architecture.

ABSTRACT

The architecture of honey bee combs embodies a range of expressions associated with swarm intelligence, emergent behaviors, and social organization, which has drawn scientists to study them as a model of collective construction processes. Until recently, however, the development of models to characterize comb-building behavior has relied heavily on laborious manual observations and measurements. The use of high-throughput multi-scale analyses to investigate the geometric features of Apis mellifera comb therefore has the potential to vastly expand our understanding of comb-building processes. Inspired by this potential, here we explore connections between geometry and behavior by utilizing computational methods for the detailed examination of hives constructed within environments designed to observe how natural building rule sets respond to environmental perturbations. Using combs reconstructed from X-ray micro-computed tomography source data, we introduce a set of tools to analyze geometry and material distributions from these scans, spanning from individual cells to whole-hive-level length scales. Our results reveal relationships between cell geometry and comb morphology, enable the generalization of prior research on build direction, demonstrate the viability of our methods for isolating specific features of comb architecture, and illustrate how these results may be employed to investigate hive-level behaviors related to build-order and material distributions.