1902.] I. H. Burkill — Floiver of Ranunculus arvensis. 119 



borrowing as may be noticed in Table XII, (see p. 102) probably is a 

 more common occurrence in the German races than in the Kew race. 



The last two Tables (Nos. XXVIII and XXIX) show the relative 

 abundance of abnormal petals and staminodal petals at different times 

 in the plants flowering'. 



The net result of the investigation is that we have in Ranunculus 

 arvensis just a little of what (for want of a better term) may be called 

 foresight in the formation of the flower. We find the flower com- 

 pleted however scanty the nutrition for it may be; and, when the 

 nutrition is adequate, provision is, it seems, made in good time for 

 the carpellary mood. The next problem will be to show how far in 

 such a flower as that of Parnassia or of any Phanerogam, the cons- 

 tancy of the carpels is due to provision made for them when the bud 

 first begins to be formed. Can the sepaline mood lead the carpellary 

 by the nose, or is the carpellary not too important to the race to be 

 without an assertiveness of its own ? 



It is interesting to observe that the staminal mood forms a sort of 

 residuary legatee to the three early moods of the flower; interesting 

 because we not uncommonly find that mood to disappear under condi- 

 tions which have generally been ascribed to something disadvantageous 

 to the plant (see Willis, On Gynodicecism, 3rd paper, Proc. Cambridge 

 Phil. Soc, viii., 1893, p. 129). » 



We have sought in passing for any indication in the flower which 

 might suggest that pressure of organ on organ exercises an influence in 

 shaping the flower ; and we found that flowers of 20 organs did come 

 near to having the formula K 5 C B A B G 5 : and in Table V we saw K^C^ 

 and K 3 3 to be commoner combinations than K 4 3 r 5 (especially 5) 

 and K 3 C 4 or indeed any other number, and in Tables VI and VIII ten 

 stamens to be commoner than nine or eleven in association with five 

 sepals or with five petals. These observations do not suffice for building 

 up any very definite statement. 



It is equally advisable at present from these tabulutions to make 

 no statement regarding the possibility of female organs demanding per 

 unit for their inception more nutriment than male organs. 



One notices in regard to the variation of the flower of Ranuncu- 

 lus arvensis that it is always hungry, i.e., always capable to taking in 

 more organs ; the hungriest of its moods is that for the formation of 

 stamens, next that for the formation of carpels, thirdly that for petals 

 and least hungry that for sepals. 



Just as we find sepals to tend to be constant in number throughout 

 our larger groups such as the Dicotyledons and Monocotyledons ; petals 

 to be constant in number in lesser groups ; carpels to serve by their 



