RANUNCULACE2E. 



29 



Stavesacre' (figs. 53-58) has nearly all the characters of the 

 preceding plants ; but the spur of the posterior sepiil is relatively 

 shorter and broader, and slightly bifid at the tip. The petal 

 opposite this is sessile, and is prolonged downwards and into the 

 spur to form a thick hollow glandular double spur (fig. 55), while 

 its limb is deeply divided into two erect halves, nearly symmetrical 

 with one another, and united in front by a short cross-piece, so that 

 the division of this organ into two half petals is not quite complete. 

 The lateral petals are represented by little wings of two kinds ; 

 while the anterior petals are quite wanting in some flowers (fio-. 53) 

 and exist in others,' 

 which have then 

 eight petals arran- 

 ged like those of 

 Aconitum, Napellus, 

 four of them beino- 

 in pairs opposed to 

 sepals 1 and 3 (fig. 

 56). 



The androceum 

 is that of the pre- 

 ceding plants (figs. 

 55 and 56) and 

 the gynasceuni is 

 usually formed of 

 three carpels,^ of 

 which one is nearly 

 posterior. The follicles are thick, and each encloses seeds closely 

 pressed together, so as to be more or less deformed. The 

 copious albumen contains the minute embryo near its apex ; and the 



Delphinium StapJiisagria. 

 Fig. 55. 

 Flower without its calyx. 



Fig. 56. 



Diagram. 



' JD. Staphisagria L., Spec, 750. S. macro- 

 carpa Spach, /. cit. 



^ On the Siime plant we may find flowers with 

 eight petals, and others with less. When there 

 are eight we see, as in the Aconites, a single 

 one opposite sepal 4 and sepal 5, and a pair op- 

 jwsite sepal 1, sepal 2, and sepal 3. The two petals 

 which are opposite sepals 4 and 5 form at the 

 base a sort of flattened spnr, glandular and nec- 

 tariferous within. The anterior petals, when 

 present, are reduced to small llatteued uuocpial 

 scales, the anterior one of each pair remaining 



much less developed than the other. See Hroxo- 

 NiAiiT, Ann. Sc. Nat., si'r. 3, v. 300, and 

 Pater, loc. cit., 2G1, note. 



•'' From two to four carpels may be counted, 

 rarely more. Their position lias not yet been 

 accurately decided (see Adatuwnia iv. 21), any 

 more than in most sections of I his genus. The 

 ovules of Stavesacre are few in nmuber. and in 

 the typical species there are only four in (wo 

 vertical rows. They arc placed back to hack, 

 and are somewhat ascending. 



