The Production of New and Improved Varieties of Timothy 357 



from which any one of the numerous types could have been determined, 

 and this has been found to be an almost hopeless task with the races or 

 varieties of many of our cultivated plants. The great range of variation 

 is well illustrated by the character of yield. Taking the average yield 

 for the second and third years of all the 16 plants of one kind, the lightest 

 yielder averaged 1.005 ounces per plant and the heaviest 20.274 ounces 

 per plant. Among the 200 plants, 5 gave for the two years an average 

 yield per plant of one pound or over, while 39 gave average yields of less 

 than one half pound per plant. 



It should be clearly recognized that in many cases such characters as 

 yield, height, niimber of ctilms, and the like, shown by the original plant, 

 may not be exhibited in equal degree when the plant is propagated as 

 a clonal variety and a number of plants are grown. The one original 

 plant may have been grown in a particularly favorable or unfavorable 

 place and the rows grown from it may thus be either poorer or better. 

 New York soils are very variable and it is almost impossible to distribute 

 manure uniformly; and there is thus ever-^ reason to expect that many 

 of the size and yield characters exhibited by the original selections will 

 be found to be due in some measure to the local environment in which 

 the plant is grown. Some original plants, that were heavy yielders, 

 when tested as clonal varieties gave only ordinary yields ; while other original 

 plants that were medium or even light yielders proved to be heavy yielders. 

 The test in rows as clonal varieties is easily made and brings out the 

 intrinsic characters of the individual as does no other method. When, 

 in a test of this sort on fairly uniform soil, all the plants of a row exhibit 

 the same characters and show differences between different rows or 

 types, such as are illustrated in Plate VI, it may be taken as 

 conclusive evidence that these differences are not due to environment. 

 The test of individual selections in clonal rows is thought by the writer 

 to be one of the most important steps in judging their comparative 

 value. 



On the other hand, a plant that gives very promising indications cf 

 value by this test may be of hybrid or mixed parentage and may not 

 transmit its good characters by seed propagation. In attempting to 

 produce good types, the experimenter would select the most promising 

 sorts grown in clonal test plats and would then test the transmission of 

 their qualities in seed propagation, using self -fertilized seed. This was 

 the next step taken in the Cornell experiments. The various clonal plats 

 were carefully studied and compared and the most promising types were 

 selected for tests of their transmission by seed propagation. Many types 

 of no apparent value, but showing interesting variations, were also selected 

 for further testing and inbred seed was obtained. 



