212 THE ZOOLOGIST. 



extent account for an occasional reversion to the original type ; 

 but under quite normal conditions the reversion to an ancestral 

 type takes place without any apparent reason for it, and seemingly 

 without any influence of external conditions. It must, therefore, 

 be attributed to a latent heredity which from time to time asserts 

 itself by an unprovoked impulse. 



This tendency towards reversion is one of the most remark- 

 able phenomena of heredity, which, although very erratic in its 

 appearance, is one of the most valuable evidences for the tracing 

 of species back to an otherwise obscure origin. Did we, for 

 instance, possess nothing but polled cattle or sheep, and occa- 

 sionally a specimen was born which developed horns, we should 

 be justified in concluding that these animals, by selection under 

 domestication, had lost their organs of defence, but that their 

 ancestors at one time possessed them. In this case we are certain 

 of it, for though the origin of some of the polled races of sheep 

 is not known, yet the occasional birth of a lamb that more or less 

 distinctly develops horns leaves no doubt of their original descent 

 from a horned race. 



It will scarcely have escaped the most casual observer that 

 children sometimes do not resemble either of their parents, but 

 are much more like one or the other of their grandparents. In 

 most of such cases it will be found that the blood of the father 

 predominates. Such instances are the simplest presentations of 

 reversion, but occasionally a characteristic makes its appearance 

 which is not found in either parent or grandparent, but can be 

 traced back to a more remote ancestor. In such cases the 

 characteristic in question has lain dormant in the intervening 

 generation. It may also be frequently noticed that a male off- 

 spring resembles the maternal grandfather more in some attri- 

 butes than those of his father. Such interesting occurrence is 

 especially noteworthy, as it is also analogous, and constantly 

 observable, with the procreation of the drone. By reversion, 

 however, must not be understood an entire reversion to the 

 ancestral type. This could scarcely take place under the 

 simplest change of condition, as, for instance, a return to the 

 same climate after the removal to another has been productive 

 of a distinct variation. 



Even if the original conditions were most completely repro- 

 duced and were to assert themselves in all their complex effects, 



