S. Mahunka szerk.: Folia Entomologica Hungarica 44/1. (Budapest, 1983)

species. Furthermore, it was intended to estimate the range of variability within the populations of the mite and to evaluate the usefulness of some morphological structures as taxonomic characters. Particular emphasis was put on characters commonly used by various authors for species identification. MATERIAL AND METHODS Material for this study consisted of 30 females and 30 males respectively laboratory reared and field collected populations. Mites were hand-collected at various dates during the period of 1965-1967 and mounted in HOYER'

s medium. In the mouting procedure living specimens were placed in the medium on microscope slides, dried for 24 h at about 80°C and framed. Until the beginning of this research (e.i. autumn 1979) they were stored in boxes at room temperature. Mites were examined in phase contrast microscope using xlOO oil immersion objective and xlO ocular. Measurements were taken with 0^1 mm microscopic scale with an exactness of ± 1 urn. To avoid gross error each measurement was taken at least twice. 85 characters were measured in each male and 67 characters in each female. Legs I, II, III (character No. 51, 52, 53 in males and 46, 47, 48 in females) were measured without trochanter, for other details see Fig. 1, through 4 and Appendix I, II. Besides 57 numerical ratios between measurable characters were calculated for each male and 42 ratios for each female (Appendix III, IV). In measurements and ratio calculations all morphological structures which were considered as potentially valid taxonomic characters were included (EWING, 1939, BEER 1954, CROMROY, 1958, SCHAARSCHMIDT, 1959, NUCIFORA, 1965). All the measurements and their numerical ratios were statistically analyzed using Student' s t-test to compare averages between the two investigated populations and coefficient of variability (CV) was used to study character variability within each population. The relation between the size of a specimen (y) and measurements of its morphological characters (x^) was estimated using values of correlation coefficient (r), where the independent variable (y) was calculated as a product of multiplication of length and width of the propodosomal plate (characters Nos 34 and 35 in males and 29 and 30 in females).. RESULTS The comparison between the two populations show a great divergence of characters (Table 1). It was considerably larger in males than in females. Among the total of 132 characters investigated in males, significant differences were found in 57.6% cases at confidence level 0.95 while among 109 characters investigated in females significant differences existed only in 19.5% cases. In general direct measurements diverged to a greater extend than their ratios. Out of 152 measurements taken in both sexes significant differences were found in 51.4% cases at confidence level 0.95 and in 36.3% cases at confidence level 0.99. Among 99 ratios calculated significant differences were found only in 32.3% and 22.2% cases respect!velly. In further analyses the characters were grouped according to their properties which might possibly facilitate errors in measurement taking, distortion in mounting procedure, etc. (Table 2). It was found that measurements taken on the surface of the body show fewer significant differences than measurements involving flexible seta. This was particularly apparent in females. In males characters located on ventral plates show fewer significant differences than those located on dorsal shields. But it was just the opposite in females. Particularly large proportion of significant differences show characters on legs IV, however, in females it was not as large as in males. It is worthy noticing that general measurements pertaining to the size of a mite were also in a great proportion significantly different between the two populations. All the averages which were significantly different between the two populations are as a rule larger in field collected mites than in laboratory ones. Only characters Nos. 5, 50, 56 in males and No. 57 in females show averages larger in the laboratory population. Further attempt was made to describe the variability of characters in statistical terms. The most appropriate for this purpose seemed the coefficient of variability CV. Results are summarized in Table 3 showing no great differences between the populations. CV distribution was nearly identical between field and laboratory females. The pair by pair comparison demonstrated that only in the case of three characters (Nos. 3, 5, 63) the difference in CV values between the