No. 588] EXPEBIMENTS IN MASS SELECTION 719 



1907. The initial stock consisted of less than a dozen indi- 

 viduals all ' 'pure recessives," which produced only "re- 

 cessive" hooded young, in accordance with Mendelian 

 expectation. But though all the young were recessive 

 (hooded), all were not exactly alike, and to assist in their 

 classification we devised arbitrary "grades" of increased 

 (plus) or decreased (minus) pigmentation as compared 

 with the modal (zero) condition in our hooded race. The 

 scale of "grades" is shown in part in Fig. 1. It has 



been found necessary to extend it in both directions, 

 beyond the range shown in the figure, in order to 

 admit the new grades of rats which have made their ap- 

 pearance as the experiment progressed. The first plus- 

 selected parents produced 150 offspring ranging in grade 

 from + 1 to +3, mean + 2.51. The first minus-selected 

 parents produced 55 offspring ranging in grade from — 2 

 to + 1, mean —1.46. It will be observed that the ranges 

 of the young produced in the two selections were prac- 

 tically continuous with each other, though they did not 

 actually overlap. But actual overlapping did occur in the 

 following generation, in which no advance was made in 

 the mean grade of the parents, practically all the available 

 females being used as parents in an effort to increase the 

 stock. The grade of the offspring also remained prac- 

 tically stationary in this second generation (see Tables I 



