AUTOREGRESSIVE-TYPE MODELS FOR THE ANALYSIS OF BARK BEETLE TRACKS

When the progress of female bark beetles is followed in a small arena as they are attracted to a source emitting male pheromones, the variat e that is of interest to biologists is the heading angle, i.e., the an gle between the direction toward the source and the direction of forwa rd motion. In the present paper we model the resulting angular time se ries variate using autoregressive-type models with von Mises random er rors. Estimates of parameters for four data sets were obtained by maxi mizing a von Mises likelihood function. Plots of residuals developed a fter fitting first order models indicated some interesting behavioral differences between beetles in control groups and those in bioassays w ith strong pheromone attractants.