STOUT & Boas: STATISTICAL STUDIES IN CICHORIUM 445 
+0.62. The further performance of this race was observed in 
two series of Fy, the progeny of two plants, with values of a which 
were somewhat lower than the average of the series. In regard 
to the actual range for values of a and averages, one series was 
decidedly higher, more uniform and less variable; its progeny 
digressed toward the higher average of the generation to which 
the parent belonged. For the other series (-то0-8-14-), the 
average а was somewhat lower than the value of the immediate 
parent, and the range was extended to lower values than were 
seen in the parent series. АП plants of the Ез and Е, of this race 
gave minus values for rate of change. 
The two radically different races, semi-dwarf bush (race 5) 
and the tall-growing race (Race 6), descended from the same Е, 
parent (Е» X А) по. то. 
The six races noted above all descended from three sister Е, 
plants which were quite alike in general vegetative characters. 
The three main lines of descent split up into six rather widely dif- 
fering races. In regard to values of a, it may be noted (TABLE 48) 
that races that segregated from the same line were of the same 
general performance. Values for races 1 and 2 were relatively 
high, those for races 3 and 4 were relatively low and those of 
5 and 6 were more intermediate. 
The differences in values for the lines and the noticeable 
uniformity within races, especially in the F,, is proof that some- 
what slight differences in number per head are hereditary. Al- 
though self-sterility of many plants prevented a rigid testing of 
the effects of selection, there is decided evidence, as noted above, 
that selection for high or low values within a race is in some degree 
effective. 
А further point appears to be clear. The very marked vege- 
tative differences may affect very strongly the total number of 
heads produced, but such vegetative differences are only slightly 
concerned with changes in the performance as measured by such 
values as o and a. 
