744 THE AMERICAN NATURALIST [Vol. XLIV 







Iv'rTvnV. 





Percent. 



PercLt. 



Percent. 





9.40 



3.27 



3.57 



4.54 



0.61 



1.76 

 375 



2.46 

 3^28 



Hybrid 

 Fillbasket 

 l...nu'i.'ll..w 

 Mid-parent 



7.75 

 9.19 

 7.93 



1.11 



t : >7 



3.63 

 2.84 



496 

 1.69 



4^54 



1.79 

 339 



4.3o 

 L7ti 



3 85 

 2^46 



6.09 

 7.01 



3.25 



3.57 

 2.89 



4.16 



2^00 

 3.35 



2,55 



3.16 



F, Hvbrid 

 F 2 Hybrid 



7.63 

 17.43 



3.53 



lo'.33 



9^45 



7.61 



12.88 



1310 



Fillbasket 

 Snowflake 



7.93 



3.22 

 2.89 



2.20 

 3.(18 



3.78 

 5.38 



3.39 



6.00 

 5.74 



3.85 

 4.29 



i'ii' i'";!r 



F, Hybrid 



8.67 



3.06 



2.94 



3^83 



3J8 

 1.29 



5.87 

 5.04 



4.07 

 7.06 



2448 



3.47 

 8.32 



2.70 

 5.96 



8.f5 



7.26 



39.74 





parents. The most noticeable feature of the records, 

 however, is the coefficient of variation in F 2 as compared 

 with the parents and with F x . The coefficient of varia- 

 tion is not, on the whole, materially greater for F ± than 

 for the parents. In F 2 , on the contrary, it is noticeably 

 greater than in F t . It is usually twice and in some cases 

 six or seven times as great as in F 1# This is merely a 

 mathematical way of expressing the fact that the F 2 in- 

 dividuals exhibit marked segregation of size and shape 

 characters. If the intermediates seen in F, were tend- 

 ing to breed true as blends, the coefficients of variation 

 for F 2 would not be appreciably greater than for the 

 parents and for F r This segregation in F 2 is so pro- 

 nounced as not to need statistical treatment for its 

 proper appreciation. Even a casual examination of the 

 material can not fail to impress one with the fact that 

 about all grades from one parent to the other are repre- 

 sented in F 2 . As a matter of fact, some individuals 

 among the F 2 gourds are decidedly larger than either 



If such results as those reported here are to be "ex- 

 plained" by assuming that variation is increased (in 

 some mysterious way?) by hybridization, we can doubt- 

 less also explain, in the same way, why this increase in 



