Dulce M. Bustamante, dbusta1@lsu.edu1, Luis Escobar, luis@lsu.edu1, and José D. Villa, jvilla@ars.usda.gov2. (1) Louisiana State University, Experimental Statistics, 161 Ag. Admin. Bldg., LSU, Baton Rouge, LA, (2) USDA-ARS, Honey-Bee Breeding, Genetics & Physiology, 1157 Ben Hur Rd, Baton Rouge, LA
The death rates of 104 honey bee colonies from southeastern Louisiana were estimated using a recurrence data approach. The colonies were established during nine swarming seasons, from 1990 to 2000. Varroa destructor was first detected in the area in the Fall of 1992.
Using a non parametric estimate of the mean cumulative number of deaths as a function of time, lower death rates were observed for the colonies established after 1997.
The non parametric estimate was then fitted with a parametric model, a Homogeneous Poisson Process (HPP). This fit was compared to fitting the curve with two HPP, one for the data from colonies founded previous to 1997 and the other one for the colonies founded after 1997.
Using maximum likelihood methods it was determined that two HPP described the curve better than a single HPP (LR 14.75, 1 df, p=0.00012). The death rate estimates for these two periods are: Previous-to-1997=0.087 deaths per month per colony, and After-1997=0.037 deaths per month per colony. These rates can also be interpreted as approximate probabilities that the colony will die in the next interval of time.
The analysis suggests that about five years after the arrival of the varroa mites, the mortality rates are lower than when the mites just arrived. This result adds more evidence to the increasing literature suggesting that Apis mellifera could be adapting to this parasite, or that the parasite is becoming less virulent.
Hymenoptera Apidae
Apis mellifera (honey bee)
Keywords: honey bee parasite, Poisson process
Poster (.ppt format, 463.0 kb)
