CROSSINGS IN MELANIUM-VIOLETS 281 



as I know, I will give a relatively detailed account oï the three hy- 

 brids made. 



V. Munbyana (line A5) X V. tricolor (line 10). 

 The hybrid was very robust in the vegetative parts, and the 

 branches obtained a length of up to '■% of a metre. However, tliey 

 were not quite as thick as those of V. Munbymui. In autumn the old 

 branches died away, just as is the case in this species, and new 

 shoots developed in their place; these wintered. The flowers were of 

 about the same size as in V. Munbyana although the petals were a 

 little more narrow. The colour is also about the same as in this 

 species although the ash-grey tinge is a little fainter. The crenation 

 was not so distinct. The labellum was large and the pollen-magazine 

 closed. 



V. arvensis (line 5) X V. Munbyana (line A5). 



The arvensis parent was prostrate. The hybrid was half prost- 

 rate and had the same growth form as the bastard 2X4 (type E) 

 in the first year. The bastard was also perennial. It then assumed 

 the same growth form as that of V. Munbyana. The hybrid gave the 

 impression of being less robust than the first mentioned. The colour 

 of the corolla was about the same as that of the hybrid between 

 y. tricolor and V. arvensis (10 X 2) already treated. Thus, the 

 upper petals were violet; the three lower were light yellow or, in 

 spring and autumn, light yellow with a violet margin. The arvensis 

 parent had very small flowers like the lines 3 and 4. The hybrid 

 was very small flowering. The corolla was not more than about 

 twice that of line 5. Any crenation of the petals was not to be seen. 

 The pollen-magazine of the arvensis form was almost open; that of 

 the hybrid a little more closed. 



V. Munbyana (line 45) X V- arvensis (line 25). 



Line 25 was a rather large flowering type with rather broad 

 petals. The flowers of the hybrid were large, perhaps not quite as 

 long as those of V. Munbyana but proportionally broader. In other 

 characters the hybrid resembled the other cross with V. ar- 

 vensis (5 X 45). 



The behaviour in Fi of the characters already treated in the 

 description of the hybrids between V. tricolor and V. arvensis follows 

 the course to be expected. Thus, the very small flowers were do- 

 minant to the larger ones, in the group of the larger types the very 

 large type was dominant to the less large. The violet flower colour 



