10 STATION BULLETIN 355 



Available phosphorus was more concentrated in the D and E 

 zones and was slightly higher in the B and C layers under the light- 

 dropping trees. Magnesium, although available in small amount, 

 tended to be more concentrated in the E layer. Available potassium 

 was also present in very small amount. Aluminum and calcium 

 tended to be more concentrated in the A zone. This soil was high 

 in soluble aluminum but there was less calcium available than any 

 other element. Except for phosphorus the differences in available 

 amounts of each element beneath heavy- and light-dropping trees 

 were not in any case large enough to account for differences in drop. 



Humus was low in the A zone and very low in the other zones*. 



RELATION OF YIELD TO DROP 



Southwick^*^' studied the relation between yield and per cent 

 drop with Mclnto'sh. and from the values obtained for coefficients of 

 determination concluded that not more than 15 to 28 per cent of the 

 variance in drop could be attributed to the effect of yield. Further- 

 more, of 6 plots only 3 showed a significant positive correlation be- 

 tween yield and per cent drop. Apparently his calculations were 

 based on the average yield and drop from 1927 to 1937. 



Correlation coefficients, determined for the relation between 

 yield and per cent drops in several of the New Hampshire Station 

 blocks, are presented in Table 8. The data confirm the observations 

 of Southwick^'' in that there is indicated a variable association be- 

 tween yield and per cent drop. None of the correlations are large. 

 Southwick^*^ also found that dropping increased with increase in 

 yield. Dickson'*, working with Mcintosh, found that the per cent 

 of drop increased with increasing yield due to improved cultural or 

 growth conditions, but not when increased yield was due to on-year 

 cropping. In the writer's experience this correlation was not con- 

 sistent with relation to on-year cropping. 



For Mcintosh and Baldwin correlations between yield and per 

 cent drop large enough to be significant are positive; for Northern 

 Spy, Delicious, and Starking they are negative. This simply means 

 that in some years the heavier-yielding Mcintosh trees dropped a 

 larger percentage of their fruit than did lighter-yielding trees. With 

 Northern Spy, the reverse is true ; heavier-yielding trees dropped a 

 smaller percentage of their fruit. 



In considering correlation between drop and yield one must not 

 confuse per cent drop with the actual weight of dropped fruit. In 

 fact, the correlation between per cent drops and yield of Northern 

 Spy in the BFP block in 1935 was — .407, whereas the correlation be- 

 tween weight of drops and yield was .760, a change from a negative 

 to a positive correlation. Although the percentage of drops was less 

 with the heavier-yielding trees, they actually dropped a greater 

 amount of fruit than the lighter-yielding trees. 



*Chemical determinations were ni.ide l)y (|uick test methods in the Agricultural Chemistry De- 

 partment by G. P. Percival. 



