FOOD, HIBERNATION AND MIGRATION 



107 



when the winds tended to be southerly. Tested by the usual methods, 

 the conclusion is very highly significant (Chapter 10, Appendix i (3). ' 

 The next step was to fnid out what relation, if any, existed between 

 temperature and rainfall, as weather factors, and migration. The 

 linear correlation coefficients in each case were just over 020. For 



Table 6 



THE RELATION OF MIGRATION TO WIND 

 DIRECTION AT TOTTERIDGE 



47 pairs of varieties, these are not significant. It is, however, possible 

 that this is due to the small number of observations. If this value 

 were to be maintained in a larger number of observations, as is quite 

 possible, it would still show that the influence of rain and temperature 

 could only be small, for the square of the correlation coefficient is a 

 measure of the proportion of the total variance accounted for by these 

 factors. We cannot expect that these factors are major influences 

 when each only accounts for some 4 per cent of all the influences at 

 work. Since it is often thought that warm rain influences amphibians 

 more than cold rain, there was a possibility that a joint function would 

 show a higher correlation, so that another correlation was computed 

 using the product of rain and temperature as the variable. This gave 

 a result very little better at 0-23. It is probably unwise to press this 

 result to the conclusion that, in migration, temperature and rainfall 

 are merely additive factors, and do not act jointly as is usually supposed, 

 but that is where the statistics lead. 



