Lack of seasonal variation in the hippocampus: statistics is not the issue

Abstract

Correspondence: Dr P. Lavenex, Department of Psychiatry and Behavioral Sciences, Center for Neuroscience, University of California, Davis, 1544 Newton Court, Davis, CA 95616, U.S.A. (email: plavenex@ucdavis.edu). M. A. Steele is at the Department of Biology, Wilkes University, Wilkes-Barre, PA 18766, U.S.A. L. F. Jacobs is at the Department of Psychology, University of California, Berkeley, 3210 Tolman Hall, Berkeley, CA 94720, U.S.A. Two recent commentaries, one in Animal Behaviour (Darlington & Smulders 2001) and one in the Journal of Animal Ecology (Garcia-Berthou 2001), point to potential problems with the use of residual analysis. In one of these papers (Darlington & Smulders 2001), the authors focused on the use of residual analysis to control for differences in overall brain size between samples collected at different times of the year to identify seasonal changes affecting specifically and selectively the hippocampus. This article followed a recent paper published in the Journal of Comparative Neurology (Lavenex et al. 2000b), in which we reported the absence of seasonal variation in volume or neuron number of any subdivision of the hippocampal complex (dentate gyrus+Ammon’s horn) of eastern grey squirrels, Sciurus carolinensis. Darlington & Smulders (2001) indicated that we had criticized Smulders and colleagues for their use of general linear models (GLM) to analyse their data (Smulders et al. 1995). They further questioned the results and the conclusions drawn from our squirrel study (Lavenex et al. 2000b) because we relied on residual analysis, rather than GLM or analyses of covariance (ANCOVAs) to control for differences in brain size between samples collected at different times of the year. In the present forum, we would like to clarify two points. First, we wish to clarify that it was not the statistical procedure itself (i.e. the use of GLM analysis) that we had primarily questioned, but rather the inclusion of data from juveniles in Smulders et al.’s analyses (1995; see Lavenex et al. 2000b, pp. 162–163). Second, we present results from GLM