A. H. Trow 249 



results summarized in Table III would be impressed with the obvious 

 rays of the NR plants, and adopting the theory of dominance would 

 count the NR and RR types together, and regard R as dominant; but 

 in the analysis of the results summarized in Table V, he might quite 

 easily overlook the NR individuals altogether and count them as NN, 

 in which case he would be obliged to regard R as recessive. The 

 association of the factors for hair and rays probably effects a reduction 

 of the ray character. It appears then to be of little consequence 

 whether we regard N or R a.s dominant, or neither. Whatever signs 

 we adopt, the interpretation remains the same. The theory of domi- 

 nance has probably been often pushed too far. Under these circumstances 

 it will suffice to choose the most convenient system of nomenclature. 

 The simpler ratios associated with the theory of dominance often give 

 the clearer view of the experimental results. 



Neither is it quite clear in such cases as these whether the presence 

 and absence hypothesis strictly applies. The original type of flower in 

 the Compositae was doubtless actinomorphic. In the section Tubuli- 

 florae we recognise the introduction of the radiate character or zygo- 

 morphy into the outer florets, and in the Liguliflorae, possibly the same 

 character, certainly a similar character, into all the florets. At least 

 two factors would be necessary to effect this. In the genus Senecio 

 the presence of typical ray florets is the rule, but the common groundsel 

 is generally described as non-radiate, the radiate type being regarded 

 as a variety. The groundsel must have assumed its peculiar non-radiate 

 character either by the loss of the factor for rays (absence) or by the 

 acquisition of some new factor (presence) which made itself manifest by 

 suppressing the rays. We have at present no experimental means of 

 testing the relative value of these two hypotheses. 



It seems fairly certain however that the non-radiate type of the 

 common groundsel is more recent than the radiate types found in 

 the related species, Senecio viscosus, sqvalidus and sylvaticus. Ad- 

 mitting this, we are still unable to decide whether the radiate 

 groundsels lanuginosus and erectus, radiatus are newcomers or relicts 

 of the older original type : they are apparently aggressive invading 

 forms and are certainly, to a considerable extent, replacing the non- 

 radiate forms. 



This kind of mutability is not confined to the common groundsel. 

 In Senecio Jacobaea there is a comparatively rare non-radiate form. It 

 is abundant in one part of Ireland, according to Praeger. Exp. 30. 

 A single specimen was collected near Cardiff, and was cultivated in my 



