2l8 WHITMAN. 



There is no quite satisfactory name for the new plant required 

 for such work, and no one has suggested a practical method of 

 developing it. Biological Farm, broadly defined, is perhaps the 

 best we can do for a name, as the work would be, so far as pos- 

 sible, upon plants and animals under cultivation. A considerable 

 tract of land, of varied surface, including woods, fresh-water 

 ponds, some brackish water, and a stretch of sea-shore that could 

 be utilized in the cultivation and study of marine animals, would 

 represent the essentials of the farm headquarters. 



The best location for the farm would be in the immediate 

 vicinity of a laboratory holding the position of a national center 

 of biological research. The Marine Biological Laboratory has a 

 good prospect of becoming such a center, if it is not one already, 

 and its proximity to the United States Fish Commission Station 

 only insures additional advantages, of great importance to the 

 farm. The farm, the laboratory, and the station would be re- 

 ciprocally helpful and stimulating, and in many problems the 

 three could work hand in hand, each supplementing the work of 

 the others. These three establishments would form a most com- 

 prehensive biological center, such as has never been equalled. 



In dealing with the problems of heredity and variation, it is of 

 the highest importance to know the history of the material to be 

 investigated. It is this prime essential that is so conspicuously 

 missine in most of the work hitherto done in these lines. Curves 



o 



and formulae may be all right mathematically and yet all wrong 

 biologically. Even Galton, the father of the statistical school, 

 warns us that " no pursuit runs between so many pitfalls and un- 

 seen traps as that of statistics." (Biometrika, I., p. 8.) 



The farm will furnish material with exact records, and will 

 thus render a most important service to laboratory workers. A 

 single illustration will suffice. It has been discovered that the 



o 



paternal and maternal chromosomes in the cross-fertilized egg 

 remain distinct, at least in the earlier stages of development 

 This seems to account for the fact that hybrids of the first gen- 

 eration between distinct species are generally " intermediates." 

 When these hybrids breed inter sc or with the parent species, we 

 often get so-called "reversions." Hitherto we have not found 

 any explanation for these "reversions." The solution of this 



