NOS. 618-619] THE BOLE OF REPRODUCTION 



270 



strains is due wholly to this difficulty, I believe. It may 

 he impossible to obtain a race homozygous in all factor-. 

 There may be a physiological limit to homozygosis even 

 in hermaphroditic plants. The best one can do is to use 

 a species which is naturally self-fertilized, relying on con- 

 tinued self-fertilization for the elimination of all the 

 heterozygous characters possible. I have examined many 

 populations of this character in the genus Nicotiaua and 

 have been astounded at the extremely narrow variability 

 they exhibit. Even though one can not grow each mem- 

 ber of such a population under identical condition- as 

 to nutrition, the plants impress one as if each had been 

 cut out with the same die. Qualitative characters such as 

 color show no greater variation, as far as human vision 

 may determine, than descendants of the same mother 

 plant propagated by cuttings. Further, in certain char- 

 acters affected but slightly by external conditions, such 

 as flower size, the sexually produced population not only 

 shows no greater variability than the asexually produced 

 population, but it shows no more than is displayed by a 

 single plant. Yet one must remember that in such a test 

 the seeds necessarily contain but a small quantity of 

 nutrients, and for this reason the individual plants are 

 produced under somewhat more varied conditions than 

 those resulting from cuttings, hence it would not have 

 been unreasonable to have predicted a slightly greater 

 variabilitv for the sexually produced population even 

 though the coefficient of heredity of both were the same. 



I have made similar though less systematic observa- 

 tions on wheat— an autogamous plant almost as satis- 

 factory for such a test as Nicotima-with practically iden- 

 tical results. I do not know of any published data on the 

 subject, however, taken either from these or any other 

 plants. In fact, there are few other plants from which 

 data could be obtained with so little likelihood of experi- 

 mental error. 



On the other hand, zoology has furnished a consider- 

 able amount of such evidence (cf. Casteel and Phillips, 



