748 



THE AMERICAN NATURALIST [Vol. XLIV 



the reduced temperature the pure signaticollis perished, leaving 

 only the heterozygous individuals, giving the stated results. 

 Were this the case there would be no real modification of in- 

 heritance. So far good, but now we come to the next group of 

 experiments. Another cross was made, signaticollis 5 X diversa 

 under the same conditions as the first experiment, giving the 

 same results. Then the same beetles were used again, at re- 

 duced temperatures (as in the second experiment above), and 

 gave in F, only signaticollis, and this, inbred, continued true 

 for four generations. This experiment was repeated seven times 

 with uniform results. It seems astonishing that the cross in the 

 first set of experiments at a lower temperature should give only 

 the blend, and in the second only signaticollis. This, however, 

 is not only explicitly stated, but is illustrated by plates showing 

 the pedigrees of the beetles. In spite of all this there must be 

 something wrong, because on page 295 it is stated that the results 

 of the second group of experiments confirmed those of the first, 

 and on page 304 it is said that the results of the second were 

 "in every way the duplicate" of those of the first. On page 330 

 experiment H. 410 (the first group, at the lower temperature) 

 is said to have given beetles exactly like the female parent 



