Szekessy Vilmos (szerk.): A Magyar Természettudományi Múzeum évkönyve 60. (Budapest 1968)
Berinkey, L.: The variance of Phoxinus populations
(«•«-»)' .3.799 (a5-? 5)2 =5.385 6.45 4.55 (15.5-10.55) 2 „ OM (7.45-2.S) 2 „ : 2.322 — = 3.289 10.55 3.799 5.385 2.322 3.289 X 2 = 14.795 The result is 0.1% smaller than 1 degree of freedom, showing a highly significant value, hence the two populations cannot be considered homogeneous. This is still to be regarded as such even if the value of % 2 is not a rate of homogeneity, as I have shown it in details in a previous paper. In the present calculations, the number of elements was small in any case. Next I wanted to see whether further differences might be found in comparing the Durmitor specimens with other populations. To this end, I used 32 exemplars from the river Tapoly in Slovakia, 35 from the river Bélbor in Transylvania, and 32 individuals from the river Morgó in the Mts. Börzsöny, Hungary. Studying the data of body-measurements, I found four measures in which the Yugoslavian specimens showed considerable differences from those of the other three populations. These were the length of the head, the predorsal distance, the preventral distance, and the length of the caudal peduncle. The basic data of these measurements are given in Tables 1-4. Even a casual glance at the Tables will show that there is but a negligible difference between the populations deriving from the rivers Tapoly, Bélbor, and Morgó, whereas the means of the exemplars from the lake Crno Jezero differ from all three. Eor a more thorough study, I applied significance calculations. The mean values of the length of the head (Table I) completely agree (22.1) in two cases (Tapoly and Bélbor), that of the Morgó population is higher (23.7), while the mean of the Crno Jezero specimens is the highest (25.1). I applied the twosample test first to the Tapoly and the Crno Jezero exemplars. For the sake of uniformity, I worked with 18 specimens in all cases. Xl = 25.1 x 2 = 22.1 the sum of squares = SS SS 1 = 23.6 SSji=12.1 the standard error (s d ) of difference is available by the following formula: SSJL +SS 2 n(n— 1) 236 +12 1 =0.3414 18.17 while the t-value is t .!äJ5Ä = 8 .787 SH 0.3414 accordingly, the difference between the mean values of the two populations is significant even at the 0.1 per cent level.
