INHERITANCE THROUGH VEGETATIVE PROPAGATION. 15 



approximately 0.5. The probable error of the acidity results vary 

 from 0.01 to 0.11, the average being approximately 0.04. 



Any differences in uniformity can be accounted for by the fact that 

 the Lisbon variety afforded a better opportunity to get uniform 

 samples. The average number of fruits to a sample was 33 for the 

 Lisbon, against 26 for the Eureka; but the number of samples con- 

 taining less than 25 fruits was 29 in the case of the Eureka, with only 

 10 in the case of the Lisbon. 



Taken as a whole, the variability in fruit from the same trees is 

 not great in either variety; neither is the variability of the fruit from 

 different trees- of the same strain. The probable errors of the three 

 determinations are about equal when comparisons between strains 

 are made. Isolated cases of high average probable errors are usually 

 due to the fact that a single instance of a high probable error has 

 affected the average. 



When the differences in composition of fruit from trees belonging 

 to the same strain are considered, only a few trees have fruit which 

 varies greatly from the general average. Both the monthly samples 

 and those taken later have instances of variation, however. 



In the case of the Eureka trees, the results on the monthly samples 

 (Table 1) show that the fruit from tree 34-57-5 has a lower specific 

 gravity and a higher percentage of rind than the other two trees of 

 the Eureka strain; also the percentage of sugar in the juice is some- 

 what lower than that shown by the others. The later results 

 (Table 3) show that the lemons from tree 34-73-7 had a higher 

 specific gravity and a lower percentage of rind than any of the other 

 samples of the group taken at the same time. The fruit of the Shade 

 Tree strain (tree 34-74-13) (Table 1) is high in acid, but only five 

 monthly samples were available for consideration. Among the trees 

 sampled later, tree 34-76-56 (Table 3) of this strain shows one 

 peculiarity; each of the three times it was sampled the fruit from 

 this tree had a higher acidity than that from any of the other trees 

 of the group sampled at the same time. 



There are similar instances in the Lisbon variety. Tree 1-27-10 

 of the Lisbon strain has a higher acidity than most of the other trees 

 of the strain. The two samples taken from tree 1-27-11 have low 

 specific gravity and high percentage of rind. In the Dense Unpro- 

 ductive strain the four samples from tree 1-30-9 have a high per- 

 centage of rind and low specific gravity. Tree 1-29-8 apparently 

 produces fruit with a tendency in the opposite direction. The fruit 

 of tree 1-29-12 has a high percentage of acidity. In the Bull strain, 

 tree 1-56-17 has fruit with the highest acidity of the group, but there 

 seem to be no other instances of consistent tendencies. 



Considering the matter as a whole, differences in composition of 

 fruit from different strains of trees are not greatly affected by the 

 unusual composition of fruit from the few trees consistently above 

 or below the average. 



DIFFERENCES IN COMPOSITION OF FRUIT FROM TREES OF DIFFERENT STRAINS. 



Some differences in the composition of lemons from different 

 strains of trees exist. Are these differences significant? What con- 

 stitutes a significant difference for this purpose % 



When the methods of comparison of Wood (9) and Pearl and 

 Miner (4) were used, practically all of the strains showed great 



