C. Pellew and F. M. Durham 169 



Variation in Kewensis on self-feHilization. 



As alrc'ruly stated, P. Kewensis on self-fertilization generally breeds 

 true to type, but varies in some particulars, such as size, degi'ee of 

 hairiness, amount of meal, and Hovver colour. 



Size. The variations in size are considerable, small plants occurring 

 in the same families with large plants. The chromosomes of a small 

 plant were counted by Miss Thomas, and were found to be 36 diploid 

 and 18 haploid. Hence the number of chromosomes is not directly 

 connected with the size of the plant. 



Hairs. The F^ hybrids possess short stalked glands visible with a 

 lens, resembling those seen in verticillatu when the meal is dissolved in 

 alcohol. In succeeding generations occasional plants of Kewensis occur 

 which have hairs on the leaves visible to the naked eye, and on micro- 

 scopical examination these are seen to be glandular, the stalk much 

 longer than in verticillata, but still generally not so long as the long 

 hairs of floribunda, although an occasional hair may attain this length. 

 (See p. 163, description of types.) 



Mealiness. PI. XXVIII, fig. 10. The inheritance of the different 

 degrees of mealiness has been studied in some detail. Mealiness is not 

 an easy character to analyse. The distribution of meal on the plant 

 may vary, as well as the quantity, and there is further the difficulty 

 occasioned by the fact that a plant which at the beginning of the season 

 is only slightly mealy may become mealy all over at the end of the 

 season. Therefore the degrees of development of this character only 

 admit of a rough classification as follows : 



(1) Meal distributed over the entire surface of the plant = il/'. 



(2) Meal on the corolla tube, calyx and bracts, and slightly on the 

 peduncles ; or restricted to the nodes and calyx, and other intermediate 

 forms = M'\ 



(3) Meal only present inside calyx and on corolla tube = ilP. 

 (Occasionally the meal of such plants is very slight in quantity.) 



We find that i/' throws if' only. One exception has been met 

 with in a family of four plants ; one of these was less mealy. This 

 exceptional plant was not kept, and we do nut know if it would have 

 developed more meal later. 



Plants classified as M^ have sometimes thrown only M^, but our 

 numbers are too small to prove that M' can be fixed. They may throw 

 all other forms. 



