230 THE JOUR^fAL OF BOTANY 



its neighbour. In September 1921, it was again found that no 

 distinction could be made between the resultant plants. All were 

 large and t^^pical P. lanceolata without exception. 



In addition, seeds of the robust form from Reading were sown 

 broadcast in a shallow earthenware dish. The dish contained about 

 five or six millimetres of soil only, and it was liable to be waterlogged 

 or rather deficient in moisture, according to the weather. Bj Sep- 

 tember, the plants that grew were all small, narrow-leaved specimens, 

 closely resembling var. splicBrosiacliya in habit. The inflorescences, 

 however, did not appear that season, neither have they yet appeared. 



The above experiments seem to indicate that the varieties of 

 P. lanceolata are largely dependent on habitat, and that the variety 

 sph(B)'ostacliya can be produced at will by putting the plant under 

 unfavourable conditions of growth. 



P. MAJOR. 



Two fairly well-marked forms of this plant are found. The normal 

 form has a small, somewhat rounded leaf with a blunt apex, the petiole 

 with a rather shallow channel, and the inflorescence onl^-- a few inches 

 lono". It is widely distributed, and occurs wherever the soil is 

 surticiently moist, and even on walls when overshadowed by trees. 

 The other form is larger, the leaf prominently veined and apex 

 acute, the petiole deeply channelled, and the inflorescence over a foot 

 lono-. It is rather sporadic, occurring in corn fields frequently, but 

 also elsewhere. 



Seeds were collected from a normal plant growing by the side of a 

 road, and from a large-type growing among grass beside a canal. The 

 collections were sown broadcast in two plots at Reading in May 1920. 

 In September 1920 the small-type seeds had produced nothing but 

 small-type plants, but the large-type seeds gave four large-type plants 

 and about sixty or seventy small-type. Seeds of the small-type thus 

 grown Avere sown at Armstrong College in April 1921 ; they produced 

 nothing but small-t3T3e. Seeds of one of the above four large-type 

 plants were sown similarly, but no large-type resulted : all were 

 small-type. 



It would appear likeh^ therefore, that these two forms of P. major 

 are not varieties caused by differences in habitat, but are more in 

 the nature of pure lines, the large-type being possibly a recessive to 

 the small -type. The determination of these would of course require 

 careful experimentation on the usual genetic lines. 



MERISTIC FLORAL VARIATION IN GALIE.^. 



By L. a. M. Riley, B.A. 



Specific descriptions in floras are, as a rule, simply diagnostic, 

 little attention being paid to the extent of variation of species be^^ond 

 recording some well-marked variety. This is particularly the ease 

 with meristic floral variation. It is often forgotten tliat the purpose 

 of Systematic Botany is not to provide a convenient method by which 



