1914] /• Arthur Harris and Ross Aiken Gortner 199 



of a small series of apples, affirms that there is. The Condensed^ 

 correlation table (i) gives the relationship between the number of 

 seeds per fruit and the weight in two coliections of pears. Table 

 2 presents the data for apples. To complete the calculation of r 

 from these tables the distributions of the weight of the fruits are 

 necessary. The data are given in Table 3. 



From these data we deduce for pears, Lot i : 



_£= 6.646, as = 2.384, 



w = 71.298, (rto= 16.204, 



whence 



rsw = +0.368 



or, expressing in terms of linear regression of weight in grams on 

 seed number according to the usual equation. 



w 



For pears, Lot 2 : 



/ _ (Tu, _ \ aio 



= [ w — r-- s ) -\- r s, 



ZV = 54.684 + 2. 500 ^. 



^= 6.173, as= 2.534, 



ZV = 84.289, o-w = 17.640, 



Tsiv = + 0.398. 



For pears, both series : 



J=z 6.401, o-s= 2.474, 



^=78.027, CT«; =18.162, 



rsw = + 0.323, ZV = 62.829 + 2.374 s. 



For apples : 



J= 7.676, 0-5=1.746, 



W= 51.271, (r,„ = 9.754, 



rsw ^+0.217, ^ = 41.982 + I.2I0i'. 



Thus there is in all these cases a material correlation between 

 the number of seeds matured and the weight attained by the fruit. 

 These results confirm the conclusions of Ewart for apples and of 

 Harris for Cercis and Staphylea. 



5 Harris, J. Arthur: Amer. Nat., 43, 693-699; 1910. 



