CORRELATIONS OF RED (CERVUS-ELAPHUS) AND ROE (CAPREOLUS-CAPREOLUS) DEER DENSITIES IN SCOTTISH FORESTS WITH ENVIRONMENTAL VARIABLES

The relationship between densities of red and roe deer and a range of environmental variables were investigated. The aim was to assess which factor, or combination of factors, was most associated with, and henc e potentially responsible for, variations in densities of deer. Twenty plantation forests throughout Scotland were used as study sites and e stimates of deer densities within them were based on dung-group counts . Using ratios of cull records and local information, 10 forests were selected which were expected to give a wide range of deer densities to aid the correlation exercise. The other 10 were selected randomly fro m within the geographic range of both species and without expectation of density: this allowed deer densities to be used as predictors of ea ch other in the analyses without any bias that selection on the basis of ratio may have introduced. The variables used in analyses included forest habitat structure, climate, vegetation composition, soil type, culling pressure, and the density of other deer species. Red deer dens ities were positively correlated with mean daily temperature (P < 0.05 ) throughout the year and mean daily snow depth (P < 0.05) for October through to March, but negatively correlated with mean daily windspeed (P < 0.05) for the quarter April to June. Roe deer density was negati vely correlated with mean daily rainfall (P < 0.05) throughout the yea r. Red deer density was negatively correlated (P < 0.05) with an index of culling pressure, but roe deer density was uncorrelated with this parameter. The actual influence of culling pressure on red deer popula tion density is unclear. Proportions of structural habitats within for ests (defined by tree age classes and growth form) were uncorrelated w ith deer densities, but were secondary predictors of deer density in r egression analyses: red deer density was predicted by the proportion o f open-thicket (trees aged 15-28 years), and roe deer density by the p roportion of restocked areas (0-8 years) and pre-thicket (9-14 years). Stepwise multiple regression analyses suggested that high red deer de nsity has an important negative influence on roe deer density (P < 0.0 5), but that the reverse did not hold. Possible mechanisms by which si gnificantly correlated variables may influence deer densities are disc ussed. It is suggested that the differential variation of red and roe deer densities across Scotland may be a result of different climatic a daptations and, possibly, an interspecific influence of red deer on ro e deer.