96 1= H. Bui'kill — Flower of Ranunculus arvensis, [No. 2, 



asserts the individual distinctly if we allow the possible formation of 

 sexual organs by order according to nutrition available, and the fixing 

 of the number by the need of packing. I shall show later, at least in 

 Ranunculus arvensis^ another assertion of the individual — a setting aside 

 of the claims of the race by allowing a kind of right of primogeniture 

 to the moods in the flower. This right of primogeniture is the more 

 interesting when we consider it in connection with the view that sepals 

 and petals are sterilised stamens ; for it gives preference to the mood 

 which by origin is then supposed secondary. 



The above remarks are to be taken as embodying some notion of 

 the foundations of the Phanerogamic flower. Working upon them we 

 may make a study of a particular species of plant in order to seek how 

 far the fixed and definite relationships of the organs in number to one 

 another, which we can observe in most Phanerogams, may be due 

 to the compelling influence of pressure in the bud acting inwards from 

 the outermost organs (sepals), or to the way in which nutrition becomes 

 available in the developing axis, or to nutrition and the influence of 

 pressure combined, or to the attempt of the plant to produce an effective 

 and economical assemblage of reproductive members. I have proposed 

 to approach the question by comparing the variation in adjacent sets 

 of floral organs, and seeing how far in different types of flower any 

 one set is free to deviate from pattern. 



There are flowers where the jointing of set on set may be consi- 

 dered to be loose, where adjacent rings of organs are not isomerous 

 and such flowers seemed best for my purpose. One such is Parnassia 

 palnstris where a 4-merous ovary tops an otherwise 5-merous flower ; 

 another is the garden Gloxinia where 2 carpels top a similarly 5-merous 

 (potentifilly in stamens) flower. It is to be asked if, as a rule, variation 

 from normal is more easily accomplished on the upper side of the badly 

 fitting joint than elsewhere. If so, then the inference is obvious that 

 pressure is playing a large part in keeping to type the moods of that 

 flower which are well jointed. 



This I found to be the case with Parnassia pahistris. In 1894 

 and J 895 I examined over 5,000 flowers and I recorded my observations 

 in the Journal of Botany, 1896, pp. 12-15. 



I had appi-oximately 5,152 flowers normal in the number of sepals 

 and in only two of them did the petals, stamens and staminodes fail to 

 keep true to symmetry ; but the carpels diverged from the normal 

 four in 450 cases. I had 36 flowers abnormal in the number of sepals, 

 15 with only four, 21 with six, and in all but three of those flowers 

 petals, stamens and staminodes followed the lead and varied with the 

 sepals ; but in them eleven flowers had three carpels, nine had the usual 



