A. H. Trow 245 



distinguished at a very early stage. When two forms of the types NN 

 aud RR are used for crossing, as in the cases under consideration, the 

 ray florets in the hybrids are of the NR type and the previous recogni- 

 tion is amply confirmed. 



Hybrids are formed quite readily in tliis way as Table II shews. 



TABLE II. 



Total No. No. of Percent, 



of plants Fi hybrids of 



Exp. Date Cross raised secured hybrids 



Exp. 9 190G Praecox x erectus, lailiatiis 16 9 56 



Exp. 10 1906 Latifolius x erectus, radiatus 19 1 5 



Exp. 11 1907 Multicaulis x erectus, radiatus 172 40 23 - 



Exp. 12 1909 Genevensis x erectus, radiatus 36 12 33 



The number of hybrids produced averages 2.5 °l^. 



The F, plants derived from any cross are always very much alike 

 and so are the pure-bred plants. In the beds the two types have 

 invariably been very obvious, so much so that inexperienced visitors 

 to the garden easily recognise them. Hence provided tliat both parents 

 have been adequately protected there is no difficulty either (1) in raising 

 hybrids or (2) being certain of their parentage. It is scarcely possible 

 to secure absolute certainty as to the effectiveness of the isolation and 

 protection, but generally there would be no great difficulty in recognising 

 the introduction of some foreign strain by the agency of a misplaced 

 pollen grain. There can only be doubt as to the paternal parentage. 

 One must remember that the ordinary method of selfiug, involving 

 castration, is at least as fallible as the one outlined above. Hence the 

 labour of crossing individual flowers is not only tedious and unprofitable 

 but unnecessary. 



Analysis of the F., generation. The second generation, so far as 

 concerns the characters RR and NN, in the four crosses under con- 

 sideration, is very easily analysed. In each case the three types of 

 plant — RR, NN, and NR — are recognisable at the first glance. We 

 have before us a very simple case of segregation without dominance, 

 as the ray florets of the heterozygotes, A^R, are typical intermediates. 

 With respect to the other characters, of which there are several, one 

 finds, as a rule, even after days of analysis, a variety which seems at 

 first to defy classification. For the present it will be well to confine 

 our attention to the ray character. The results of the analysis are 

 given in Table III. 



The only remarkable feature of this analysis is the constant tendency 

 to a slight excess of NN forms. This may have some importance and is 

 at any rate interesting, as the typical groundsel is also NN. 



