266 



Gates. 



former died before reaching maturity. The appearance of two such 

 plants is crucial, and is in itself amply sufficient to prove that muta- 

 tion is a process independen' of hybrid recombinations. The plant 

 (No. I. 20) which survived was first recognised as a lata rosette on 

 May iS"". A month ater stem-formation had begun, and the leaves 

 were now recognised as very similar to certain pollen-producing lata 

 plants from Lancashire (see Gates 1913 a, p. 45, fig. 44), but unfor- 

 tunately a photograph was not 

 secured. Wlien fully developed, the 

 main features were as follows : Stem 

 more or less weak and decumbent 

 (lata); leaf midribs pink (rubri- 

 calyx), lower leaves nearly smooth 

 (semilata) with obtuse rounded 

 points, upper leaves much crinkled 

 (lata), broad pointed; buds R (ni- 

 hricalyx) but pale red, stout but 

 conical, anthers producing pollen 

 f semilata); capsules very large and 

 stout and filled with seeds. The 

 relatively huge size of the capsules 

 (50 mm in length and 10 mm in 

 width at base), is the most striking 

 feature distinguishing this plant 

 from semilata. In lata the capsules 

 are very short (19 — 24 mm) owing 

 to a great amount of sterility in the 

 ovules (see Gates 1913 b, p. 126), 

 but in this plant the capsules were 

 both longer and stouter even than 

 in Lamarckiana, showing that the 

 ovular sterility has in some way 

 been overcome. 



As regards the origin of this plant, the structure of the nuclei 

 furnishes very definite evidence. In a paper now in the press, Miss 

 Nesta Thomas and myself will show that in a number of plants which 

 ■combine lata or semilata foliage with various pigment or flower cha- 

 racters, the number of chromosomes is 15. We have found this to be 

 true of the lata rubricalyx individual above-described. Obviously, 

 -then, the extra chromosome determines, or at any rate is concomitant 



Fig. 24. Mutant from {rubricalyx 

 grandiflora) X grandißora. 



