542 THE PURE LINE HYPOTHESIS. 



This simple supposition ^e^ives some notion of a unifying^ principle 

 between the two kinds of inheritance ; for there is then nothing 

 antagonistic to the view that blending characters are represented 

 in the chromosomes like the non-blending characters ; the 

 only difference being that the factors of the blending characters 

 in the contiguous homologous chromosomes mutually influence 

 each other and assume a mean nature. This conception of mutu- 

 ally influencing factors gives an insight into the possible cause 

 of parent prepotency, as we may imagine that in the case of the 

 prepotent parent the factor of the character in question is of 

 such a nature that a reduction to a mean O'f the factors does not 

 occur when they are combined in the same nucleus and located 

 in homologous chromosomes. 



As already remarked, the factorial hypothesis does not 

 explain heredity : it simply transfers the problem to a system of 

 impulses, memories, centres of force, corpuscles or the like in the 

 germ-cells. There is, however, a real justification for thus shift- 

 ing the venue of the problem, for by its means, in the case of non- 

 blending inheritance, we can predict the approximate results that 

 may be expected from various crosses, and in this sense the 

 hypothesis is enlightening ; but it is incapable of aft'ording any 

 information as to how it is possible for characters to depend on 

 the presence of factors in the germ-cell, since it does not throw 

 real light on the nature of the hereditary influence. 



8. Relationship betzveen Character and Factor. 



We have seen, then, that the peculiar relations which have 

 been observed in the inheritance of certain characiters have lead 

 to the conception that characters are represented in the 

 chromosomes by factors or a group of factors. The important 

 question arises as to how far the variations observable in the 

 expression of the characters are to be attributed to a variable 

 environment or to changes in tlie factor. According to the Men- 

 delians the factors are relatively stable and fixed, and when they 

 do change in any way the character exhibits a mutation which is 

 inheritable. Experimentation with respect to the variability or 

 otherwise of factors is desirable, but much data in the breeding 

 of plants and animals have accumulated in recent years, and 

 doubtless some might be found suitable for the purpose. If the 

 factor or factors are relatively invariable, and the variability of 

 the expression of a character is to be attributed to a changeable 

 environment, Johannsen's pure lines are a necessary consequence. 



II. Breeding Experiments. 



I have conducted certain breeding experiments with self- 

 fertilised generations of garden varieties of nasturtiums and fox- 

 gloves with a view to further testing the pure-line hvpothesis. 

 The experiments on foxgloves are still in prop-ress. and the results 

 are proving consistent and intelligible, and are giving no support 

 to the hypothesis of pure lines as above defined. 



In the case of the nasturtiums, the characters studied have 



