97th ESA Annual Meeting (August 5 -- 10, 2012)

PS 36-180 - Modeling a sustained cell allocation pattern in the comb of honey bees (Apis mellifera)

Tuesday, August 7, 2012
Exhibit Hall, Oregon Convention Center

ABSTRACT WITHDRAWN

Kathryn J. Montovan, Center for Applied Mathematics, Cornell University, Ithaca, NY, Nathan J. Karst, Math and Science, Babson College and Thomas D. Seeley, Neurobiology and Behavior, Cornell University
Kathryn J. Montovan, Cornell University; Nathan J. Karst, Babson College; Thomas D. Seeley, Cornell University

Background/Question/Methods

Storage of honey, pollen and developing brood each occur in predictable regions of comb in any hive of Western honey bees (apis mellifera), and this cell allocation pattern is sustained throughout the life of each hive. It may be the case that no member of the hive has a global or even local blueprint for the observed pattern, but instead each member is following some collection of simple rules that result in the emergent property. Previous studies have shown that such a set of rules could explain the formation of the allocation pattern starting from an empty comb and proceeding over the first larval gestational period. This work shows that these rules are not able to maintain the pattern over longer time frames and proposes new, biologically relevant models that better capture the observed behavior. 

Results/Conclusions

Four models were analyzed by performing hundreds of simulation runs over many gestational periods for each over a wide range of parameter values. A new metric of pattern formation was developed and  used to analyze the degree to which the pattern is retained over each simulation run. Applied to our simulation results, these metrics showed that relatively minor changes in the rules constraining the behavior of individual bees in simulation drastically changed the model’s ability to maintain a well- formed brood and pollen regions, and point to elements of behavior that are not yet completely understood. The new set of rules are able to maintain the allocation pattern over long time frames and a fairly wide range of parameter values.