266 



BREEDING CROP PLANTS 



probably a frequent cause for the variation in yield from different 

 trees of the same variety when grown in the same orchard. Self- 

 sterility, which is so prevalent among fruit crops, often prevents 

 the production of homozygous material ; while the use of heterozy- 

 gous material does not allow the breeder to make systematic 

 crosses with a knowledge of the genetic constitution of the parents. 

 In spite of these difficulties which the fruit breeder must face, 

 there has been a consistent attempt to use fundamental breeding 

 principles and at present methods are becoming somewhat 

 standardized. The advantage which comes to the breeder from 

 the fact that an improved variety may be propagated asexually 

 and need not be reduced to a homozygous condition, tends to 

 offset other difficulties Some of the more general problems will 

 be here illustrated. 



Overcoming Soil Heterogeneity. Batchelor and Reed (1918) 

 have made an interesting study of variability in orchard plots. 

 They used orange, lemon, walnut, and apple trees in the investi- 

 gation. From 224 to 1,000 trees of each of the different fruits 

 were studied and the coefficient of variability for yield of single 

 trees determined. The coefficient of variability of the clonal 

 varieties ranged from 29.72 to 41.23 per cent. Thirty-five per 

 cent, might be considered a fair average. Multiplying this by 

 0.6745 gives 23.6, the probable error in percentage of the mean. 



The effect on the coefficient of variability of increasing the 

 number of trees in a plot was studied ; a comparison of plots con- 

 taining 1, 2, 4, 8, 16 and 24 trees being made. Table LXIX gives 

 an average of tests with oranges, lemons, apples, and walnuts. 

 The results are based on a study of more than 2,000 individual 

 trees. 



TABLE LXIX. EFFECT OF INCREASING THE NUMBER OF TREES PER PLOT 



Number | 

 of trees \ 

 per plot '; 



Average coefficient of 

 variability 



Average reduction of coefficient of variability by 

 increasing number of adjacent trees per plot 



Increase from 



Average reduction 



