490 GENETICS IN RELATION TO AGRICULTURE 



expected, and it should persist as long as the ponies are kept under 

 the more favorable conditions. It would then be merely an instance 

 of the reimpressment of a given environmental effect on each succeed- 

 ing generation, not a case of the transmission of an acquired character 

 at all. But if the modification increases in degree from generation 

 to generation, as we may be led to believe by the above statement, 

 then we have something very like the inheritance of an acquired 

 character. Unfortunately we have no actual concrete evidence on this 

 point, and we have no instance of the subjection of the case to the third 

 point of proof which has been outlined above. If such increased size 

 should persist when the ponies are transported to their original habitat, 

 rigidly excluding the possibility of any effects of selection, then the 

 evidence of transmission might be accepted. 



This, of course, is not an isolated instance of the supposed trans- 

 mission of acquired characters; on the contrary, the agricultural litera- 

 ture is full of statements which indicate a tacit acceptance on the part 

 of the authors of the inheritance of acquired characters. Large-sized 

 breeds come from regions of correspondingly abundant food supply, 

 small-sized ones from regions of scanty provender. The small size of 

 Alderney cattle has been favored by systematic underfeeding. Sheep 

 transported to a dry climate acquire with succeeding generations a more 

 and more marked harshness of wool. Instances like these may be 

 multiplied indefinitely, but they are all very much alike; they are state- 

 ments of opinion rather than of fact, and their interpretation is based 

 upon a non-critical treatment of uncertain data. Very often it is ex- 

 ceedingly difficult to separate and to evaluate accurately the particular 

 effects of different factors in a given instance. Moreover, cases which 

 very closely simulate the inheritance of acquired characters may be readily 

 imagined. Thus a Shetland pony dam by systematic underfeeding has 

 been severely stunted in its growth. Now it is hardly conceivable that 

 such a pony could provide for her young while in utero or during the 

 suckling period the necessary food for its most favorable development, 

 particularly if the systematic underfeeding of the dam continued during 

 this period. It is conceivable that such an effect might last through 

 several generations when the individuals in question were placed under 

 more favorable conditions, and there would be a simple physiological 

 reason for the fact. But this would not be in any critical sense, a trans- 

 mission of an acquired character, for the germinal material would remain 

 the same throughout all these changes. The proof of the transmission 

 of acquired characters requires along with it proof that the germinal 

 material has been affected in a fashion corresponding to that of the soma. 

 A case such as the above would be more properly an effect of propagated 

 environment, if we may use such a term, and this might well account 



