Correlation and Fariatioii ni the Toad 



599 



ponderate and would afford a sufficient basis for the action of a 

 reproductive selection in this direction. Among monogamous 

 forms the excess of males would similarly be an advantageous 

 condition, insuring the fertilization of every female and might 

 similarly result from a reproductive selection in this direction. 



One farther point should be mentioned in this connection though 

 not strictly apropos. The collection of toads giving the ratio of 

 658 males to 1000 females involved, as we have seen, practically 

 the entire population. Of the first half of the population collected 

 /'. ^., of the first 220 specimens, 108 were males and 112 females, 

 a ratio of 964 : 1000. These were collected purely at random, 

 being taken as they were found and no rejections whatever made. 

 The first 220 individuals represented what might with perfect 

 right be called a "random sample" and yet collection of the entire 

 population showed how far this was from being a true sample 

 of the whole. The discrepancy was utterly beyond detection by 

 any statistical method and yet the usual precautions in the col- 

 lection of the material were fully observed. In the second half 

 collected the ratio of males to females fell to 438 : 1000. 



A fairly close study of the habits of the toad during the entire 

 summer had not disclosed the fact that a true sample was not being 

 obtained but the complete data show that this was the case. The 

 reason for this can not now be determined certainly. It might 

 have been the result of such a fact as that the males appeared 

 earlier in the evening than the females so that collections made 

 early in the season when material was abundant and enough for 

 the next day's measurement could be obtained in a few minutes 

 would consist of a larger proportion of males than later in the 

 season when the search for material was necessarily continued 

 frequently until midnight. Whether this was the case or not can 

 not now be said. In itself the point is of no particular importance 

 but it serves to emphasize the thoroughness with which the pre- 

 liminary study of an organism and its habits must be made when 

 samples of a population are to be used as representative of the 

 whole. Furthermore it affords a clear illustration of a fact occa- 

 sionally overlooked in the collection of data, namely that the statis- 

 tical method gives no exact indication as to the truly representa- 



