THE I'UKIi LINE II VPOTliESlS. 56/ 



Oil takino- the families in which both kinds of offspring: occurred, 

 the numbers were no red and ji,^ non-red, which h^ures are very 

 close to the 3 : i ratio. 



20. Owing to the circumstance that there occurred a wide 

 range of colour intensity in the 20 hybrids, the character appeared 

 suitable for testing the pure-line hypothesis. 



In the families raised by self -fertilising the 20 hybrids it was 

 found that the mean colour-intensity of the dilierent families 

 varied considerably, and it became obvious that a pale red parent 

 tended to produce pale red offspring, and a dark red parent dark 

 red oft'spring. This occurred notwithstanding the fact that the 

 grandparents were the same in the two cases, and owing to self- 

 fertilisation, the factors handed down were all derived from the 

 same grandparental factors. Thus, selection within a self- 

 fertilised family was effective, wdiile according to the pure-line 

 hypothesis there should be either no result from such selection, or 

 the utmost that could be expected would be a sorting out of two 

 strains of redness, one corresponding to the factor of the male 

 grandparent and one to that of the female grandparent. 



Karl Pearson's co-efficient of correlation which measures 

 the strength of inheritance between the red parents and red off- 

 spring was calculated ; it amounted to .435 ; wdiile if parents and 

 offspring of similar gametic constitution (i.e., individuals which 

 are capable of producing both red and yellow oft'spring) only are 

 taken the co-efficient is raised to .720. 



21. It may be said that the general results of the experiments 

 are not in conformity with the pure-line hypothesis which may 

 be regarded as a necessary consequence of the strict Mendelian 

 standpoint ; but they favour the view that a small variation in a 

 character of an individual (B) can be due to a definite small 

 change which had occurred in the particular factor which had 

 been handed down to that individual by the parent (A), and, 

 owing to the occurrence of self-fertilisation, only this slightly 

 modified factor can be passed on to the offspring of B, and these 

 offspring will therefore tend to inherit the slight variation 

 exhibited by B. In other words, slight variations in a character 

 of an individual can be due to minute variations in the factors of 

 the germ-cells from which the individual arose, and are not 

 necessarily due to the action of a varying environment on a rela- 

 tivelv fixed and unchangeable factor. 



The contention is that factors are to be regarded as variable 

 and capable of being gradually modified, in place of being in- 

 herently stable, as supposed bv many of the Mendelians. Thus 

 the smallest variations may be inheritable, and can be utilized in 

 the course of evolution. 



N.\TAL University Collecik, 



MARITZnuRG. 



