1918.] HORTICULTURE. 745 



" This result shows a marked correlation, indicating a pronounced hetero- 

 geneity in the soil of this grove influencing fniit production. However, when 

 the correlaton between the 8-tree plat as the ultimate unit and the yield of the 

 combination of four such systematically scattered plats was calculated it was 

 found that 



r =+0.137±0.120. 



" This coefficient is practically equal to its probable error and can be regarded 

 as significantly zero. 



" In the computations made by the writers emphasis is also laid upon the 

 nature and magnitude of the probable error. It is shown in several cases that 

 the probable error of comparison between plats may be so large that relatively 

 large differences must be evident between treated and untreated plats for a 

 reasonable assurance that It is due to the factors being experimented upon. 

 With the plats of 16 to 32 adjacent trees which were studied, a difference of 

 from 62.94 to 81.97 per cent of the mean production would be necessaiy in order 

 to obtain chances of 10 to 1 that the results were due to differential treatment 

 and not to casual variation in the productivity of the trees. With the same 

 number of trees in scattered units, a difference of 28.42 to 50.02 per cent would 

 be necessary for the same odds. It seems probable, therefore, that a difference 

 between two tree plats of less than 50 per cent ®f the mean production should 

 be considered with caution before attributing it to differential treatment 



" The relation between the shape of a plat and its variability was investi- 

 gated by making comparisons between square plats and linear plats containing 

 the same number of trees. Except in the case of large plats, the difference 

 in the variability of plats of different shapes was insignificant. 



" In anj- method of field experimentation where a standard of comparison is 

 desired the theoretical or ' normal ' yield of a plat is a question of importance. 

 By the use of certain formulas the ' normal ' yield may be computed from control 

 plats. As a standard, one may use the average yields of the control plats of the 

 entire area, or of the nearest control plats, or a combination of the two. In 

 cases studied, the coeflicient of variability was reduced 50 per cent by calculat- 

 ing the normal yield from the nearest controls in place of using the mean of the 

 entire area. The employment of every alternate row as a control plat was not 

 suflicient to offset the variability due to soil heterogeneity. 



" Computations made on the yields of orange, walnut, and apple trees for 

 several consecutive years showed little annual fluctuation in their variability. 

 One or two crops may not show greater variability than the average of six or 

 seven crops." 



Fa.ctors influencing th.e abscission of flowers and partially developed 

 fruits of the apple (Pynis m.alus), A. J. Heinicice {New York Cornell Sta. 

 Bui. S9S (1911), pp. 45-114, figS: 8).— This bulletin contains the results of ob- 

 servations and experiments made during the three seasons 1914-1916 with the 

 view of determining the factors influencing the abscission of flowers and of 

 partially developed fruits of the apple during the so-called June drop. The 

 literature of the subject is briefly surveyed and a bibliography is appended. 

 A popular summary of the work had been previously noted (E. S. R., 38, p. 541). 



The data collected during the course of the study and here recorded show 

 the percentage of flowers developing fruits and of flower spurs retaining fruits 

 after the first drop and after the June drop ; the relation between amount of 

 bloom and set of fruit ; set of fruit on limbs with large leaves and on limbs 

 with small leaves ; set of fruit as influenced by the location of the spur on the 

 twig growth of different years ; set of fruit on spurs formed on different parts 

 of a given year's growth; relation between the number of flowers to the spur 



