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)

Table 7. Constants added to the original data to standardize to age class IV from lowland locality. The upper rows of figures refer to lowland, the lower ones to highland localities (values 0 and 1 in the dummy regression) Character Tooth-wear class Character I II III IV + 1,09 -0.54 + 0.06 Lefi + 0.51 -1.09 + 0.61 -0.55 + 3.30 + 0.75 + 0.35 Dim3 + 4.07 + 1.50 + 1.11 T0.76 + 1.18 + 1.18 + 0.21 Wzyg + 1.37 + 0.58 + 0.39 + 0.19 + 1.13 + 0.28 + 0.14 Hnac + 1.27 + 0.42 + 0.28 + 0.14 + 1.63 + 0.03 + 0.05 Glies + 1.93 + 0.32 + 0.35 + 0.30 + 0.23 -0.26 -0.19 Hefm + 0.22 -0.24 -0.18 -0.01 + 0.40 -0.19 -0.06 Wifm + 0.48 -0.12 -0.11 + 0.07 -0.10 + 0.13 -0.09 Wich + 0.01 -0.13 + 0.08 + 0.00 + 0.95 + 0.26 + 0.17 Wpml + 1.04 + 0.34 + 0.26 + 0.09 + 0.67 + 0.04 + 0.05 Lioc + 0.73 + 0.11 + 0.12 + 0.08 + 1.04 + 0.37 + 0.18 Dcoc + 1.47 + 0.78 + 0.59 + 0.43 + 2.05 + 0.50 + 0.27 Dcna + 2.48 + 0.95 + 0.70 + 0.43 + 3.65 + 0.86 + 0.44 Daic + 4.37 + 1.58 + 1.15 +0.72 + 11.12 + 3.81 -0.28 Venv + 13.93 + 7.60 + 3.55 + 3.83 -0.43 -0.65 -0.03 Dorv -1.40 -1.64 -1.02 -0.99 The relationship between cranial and mandible measurements Since obviously were quite disappointed to find that the variable with the largest specific variance is not the one most suitable for discrimination, we further sought to clarify the relationship between the cranial and the mandibular measurements. We employed canonical correlation analysis to determine the correlation between the two sets of characters. Initially all the relevant variables were included, but the multicollinearity of the characters resulted in spurious results. Therefore the four cranial measurements ranked highest by the SS method were analysed with the three mandible measurements found to be discriminating (Table 8). Dim3 was highly correlated with the first canonical variate of the mandible set, and all of the three mandible variables were highly correlated with the first canonical variate of the cranial set. Fig 11 shows the relationship between the variables Dim3 and the first canonical variate of the mandible set representing the three measurements. It is obvious that variation in Dim3 is highly correlated with variation in the mandible, and though directly with less discriminatory power than the latter ones, some light has been thrown on the relationship between them.