L. Forró szerk.: Miscellanea Zoologica Hungarica 12. 1998 (Budapest, 1998)

Horváth, Gy.: Population dynamics and trappability of four rodent species in a forest habitat

The coefficient of variation values in the case of the other three species rose steeply when abundance dropped (in spring and summer 1996 as mentioned above), which indicates that the small number of captures did not allow the Jolly-Seber estimator to be applied successfully for this period. Values estimated by the programme were higher in this period than the MNA figures, meaning that the Jolly-Seber estimator overestimated population sizes in comparison with MNA. If the two parameters are compared monthly using the Wilcoxon test in the case of all four species, the greatest difference between the two data series was obtained for A. sylvaticus (z = 3.29, p < 0.001). Values of the D model estimation were higher than MNA in A. flavicollis and A. agrárius at z = 3.17-3.23, p < 0.002, and in C. glareolus at z = 3.05, p < 0.003 significance level. Many more captures occurred in 1997 than in the preceding years, which can only partly be attributed to the modified sampling regime. Because of the five trap-nights instead of four, and the higher number of traps, population sizes in 1997 cannot, of course, be compared with the results from the previous years. Yet, it can be stated that the higher abundance, which is indicated both by MNA and by D model estimations, is partly a result of the more favourable weather conditions. Population estimations by the D model gave lower values for this year than MNA, except for A. sylvaticus. According to these results of the Wilcoxon test, MNA values were higher at the significance level z = 2.66, p < 0.01 in A. flavicollis and C. glareolus, while at z = 2.54, p < 0.02 in A. agrárius. The estimated values also demonstrate the repeated autumn population buildup of A. agrárius (Fig. 4). A. flavicollis and C. glareolus were constantly present in the area at a relatively high abundance, and the asynchrony of their population maxima is reflected in the estimated population sizes, too. The presence of A. sylvaticus in much smaller numbers Fig. 4. Population size estimates of four rodent species in 1997, as estimated by the D-model of the JOLLY computer programme (with 95 % confidence interval values shown as bars)