Kaszab Zoltán (szerk.): A Magyar Természettudományi Múzeum évkönyve 76. (Budapest 1984)

Demeter, A. ; Lázár, P.: Morphometric analysis of field mice Apodemus: character selection for routine identification (Mammalia)

Fig. 8. Distribution of the original variables, in the 3 dimensional factor space obtained in the principal factor analysis. Only a few variables of special interest are indicated by their mnemonics percentage of correct classification was somewhat less than in the stepwie procedure, when only 12 variables met the entry criteria. When the 22 variables ranked highest by the SS method were stepwise entered in the order of their rank, 18 including the pelage characters and 13 without them were included, respectively. However, it is from the mandible that we can achieve the most parsimonious identification, for including three variables only yields classification with about 95°/

0 certainty, which meets the reasonable objective of 5% misidentification. The inclusion of more variables improved the effectiveness of identification only slightly. The classification function coefficients are given in Table 6 ; Table 7 contains the constants that are to be added to the original data, since the analysis was carried out on standardized measures. The general conclusion from the discriminant analysis is that using three variables only, identification may be achieved efficiently. Using the 18 characters selected by the SS method, the level of correct identification may be somewhat increased.