362 Bulletin 313 



There are two factors acting here that make this regression appear more 

 marked than it really is. In the first place, plants piopagated by bulbs, 

 as clons, start more rapidly than do seedlings and produce larger plants 

 as a result; so that the yields in the second and third years, which are 

 taken in all cases in these experiments in comparing plants, are rather 

 larger for clonal plants than for seedlings of the same age (Table i). In 

 the second place, in this diagram of comparative yields the seed propagation 

 was all from self-fertilized seed and there is a considerable loss of vigor due 

 to this cause. (Seethe preceding section, p. 358, Table 2.) If the proper 

 correction for these two factors could be made, it is probable that the 

 increasing yield in the original plant would be followed, on the average, 

 by about an equal increase in the yield of the progeny. 



It is not necessary for our purpose in this bulletin to discuss the trans- 

 mission of all characters in detail. It is necessary only to state that such 

 characters as height and diameter of plant, number of culms, and other 

 characters influencing yield, were transmitted in very much the same 

 degree as was yield. In many instances such characters as long leaves, 

 broad leaves, large heads, small heads, branched heads, and the like, were 

 exhibited by all the seedlings, showing that the parent plants were pure 

 with reference to these characters. 



In many cases the careful examination of the seedlings grown from 

 different types showed that all the main characteristics had been inherited 

 and the types could be considered as practically pure races. As an illus- 

 tration, one of the original plants (Plate II, Fig. 4) selected as valuable 

 formed very large and dense, strictly erect plants, which were leafy well 

 to the top and very late in season. The whole appearance of this plant 

 was very characteristic and every plant in the clonal test row showed the 

 same characteristics and was uniformly good. Inbred seed of this 

 plant produced progeny that were exceedingly uniform in all their 

 characters and were so nearly like the original plant that they might 

 easily have been mistaken for a clonal row. This uniformity of type 

 transmission can be seen by an examination of Plate VII, Fig. 2, right- 

 hand row. We may safely conclude that we have here a type which 

 reproduces its characters true, and which, if valuable, could be sent 

 out immediately as a distinct new race or variety. 



A large number of plants differing widely in their characters were found 

 to reproduce similarly true to type, and thus these m.ay be considered as 

 so many distinct new races or varieties. The further problem in con- 

 nection with such races is to determine whether the high-yielding, good- 

 quality sorts selected from the above tests will continue to show the same 

 characters when they are sown broadcast and grown as is timothy 

 ordinarily. The experiments outlined ha\'c shown that timothy, as 



