1902.] I. H. Burkill— Flower of Ranunculus arvensis. 97 



four, fourteen had five and two had six. So much for the free valuation 

 above the badly fitting line in Parnassia. In the garden Gloxinia on 

 which I have made, when at Kew, some unpublished observations, it is 

 the same. Gardeners have selected and raised beautiful races with more 

 than the normal number of petals ; the selection was never for the sepals 

 or stamens, but these two sets of organs have varied hand in hand 

 with the petals while the ovary which normally has two carpels hesi- 

 tates in the improved race between two and three. 



A table which I gave in my note on Parnassia shewed that when 

 the sepals were 4, the carpels were generally 3 ; and when the sepals 

 were 6, the carpels were generally 5. Herein we see a correlative 

 increase or decrease in both. Now it is easier by -^ of the unit 

 to squeeze five than to expand three into the space of four and it 

 happened in Parnassia, as I showed in a table on page 13 of the Journal, 

 that five carpels were more common in 6-merous than three in the 

 4-merous flowers, — an observation in accord with ideas of pressure but 

 of a ring on a confined area ; and not of organs compelling others to 

 fall into the niches between them. Towards satisfying myself in this 

 matter, I devised a little machine for measuring divergences and 

 succeeded in demonstrating (see Annals of Botany, XV, 1901, pp. 

 187-192) that, at least when near fruit-ripening, the carpels in Par- 

 nassia have no very exact relationship in position to the sepals. 



After examining Parnassia I sought for a flower with worse fit- 

 ting joints or better with no joints at all and took Ranunculus arvensis 

 for my purpose. 



Ranunculus arvensis is a little cornfield weed of Europe and Tem- 

 perate Asia, an annual and easily grown. It is very variable in the 

 flower and in all parts of it ; it has not got that concentration of the 

 moods for the formation of the various floral organs which occurs in 

 all regularly whorled flowers, its moods for the formation of petals 

 and stamens being particularly ill-defined. These irregularities seemed 

 to me qualifications suiting it particularly to my purpose. The sepals are 

 commonly 5 with a divergence of f , the petals are 5 or fewer alternating 

 with the sepals and repeating their divergence; but the stamens and 

 carpels have a completely different arrangement ; the former are very 

 variable in number and the latter generally 4-7. 



I grew my plants in 1895 in the University Botanic Garden, 

 Cambridge, from seed which had ripened in the Botanic Gardens of 

 Bonn and Hiedelberg, Paris, Stockholm and Bordeaux, and in 1898 

 in a window box at Kew from seed which had ripened in the years 

 1896 and 1897 in the Royal Botanic Gardens, Kew. I made a point 

 of examining every flower produced, counting and recording the 



