This week

[Thomas Daniel <danielt@u.washington.edu>]

Come one come all to mathbio seminars
(http://faculty.washington.edu/danielt/Mathbio).

We meet in 114 Kincaid Hall at 1230 on Wed and Friday.

This week is

Steven Viscido
Title: Analysis of fiddler crab flock movements: Evidence for selfish
herd behavior

When attacked by predators, flocks of sand fiddler crabs (Uca pugilator)

appear to bunch up tightly. We videotaped fiddler crab flocks in the
field being attacked by shorebirds and by human beings. In the lab,
video segments were digitized using computer motion analysis. Plots of
mean velocity over time show a sharp increase in crab velocity from 4
cm/s to 22 cm/s after the onset of an attack. However, this acceleration

was not simply flight away from the predator; rather, the crabs fled
quickly toward the center of the flock, even when that also meant
running toward the predator. Geometric analysis using Voronoi
tessellations showed a 20% decrease in the mean median area occupied by
individual crabs following the flight response, which is a predicted
characteristic of selfish herd behavior. We then used the movement
parameters of the real crabs to help build an individual-based
simulation model of fiddler crab movements. We tried a series of
different movement rules and compared the results to real crab behavior.

 The one that best reduced predation risk (as measured by the Voronoi
tessellation) and reproduced crab behavior the best was the one that
caused crabs to take all flock-mates into account, with corrections for
distance.