XIV] Numerical Abeynitions 253 



masked ; for though they may be really Indicative of 

 physiological disturbance, these disturbances must be rather 

 of the nature of sporadic events than the consequence 

 of predominating large causes affecting long series of 

 records. 



Occasionally however the effects of such definite causes 

 can be traced In our numbers. One excellent example Is 

 provided by the work of E. Baur (24), who Investigated 

 the Inheritance of the colour of the leaves In Antirrhinum. 

 He found that the green type breeds true, but that the 

 golden-leaved or '' aurea'' form gives, on self-fertilisation, 

 a mixture of goldens and greens. The resulting numbers 

 were 573 golden and 286 green, showing very clearly the 

 proportion 2:1, the golden being a majority. The question 

 why are the goldens not 3 to i of the greens is, as he points 

 out, readily answered. The missing group of plants are 

 those which would have been pure to the yellow character. 

 Such plants would contain no chlorophyll and consequently 

 would perish. A study of the germinating seeds sub- 

 sequently proved that this was the true account, for such 

 an examination gave "]"] green, 160 golden, 51 almost or 

 quite without green. The last group all died. Such an 

 observation throws a clear llorht on the meanino- of some at 

 least of those " ever-sporting " types to which de Vries drew 

 attention"^. As stated In a note on p. 163 Miss Durham 

 has now proved the condition of yellow mice to be exactly 

 comparable with that of these varietates aureae. It would 

 be a matter of very great interest to determine what Is the 

 exact cause of the non-vlablllty of the pure yellows, and 

 what the physiological action of the " yellow factor " may 

 be. 



In the light of this experiment, and Indeed from a priori 

 consideration. It is clear that the non-viability of zygotes or 

 even of gametes bearing special combinations of factors 

 may play a large part In genetic phenomena. Nevertheless 

 with the exception of some cases in Pri))iu!a and the 

 peculiar phenomenon recorded in regard to maize (p. 161) 

 I know no larofe series of numbers which show a consistent 

 departure from expectation. There are however suggestions 



* Mutationstheorie, i- P- 597- 



