448 ^^ws^\<^Mft!^^Cla»^"^vnnRe/a^»Pos»^^o«,• \M\\\\\\\'\n<. 



NycTAGiNEiE. In the ovary of Mirabilis a flattened tuberosity 

 with a depression in its centre, situated at the base of the style, 

 shows the ventral side of the carpel, the stigma curving down on 

 the opposite side (PI. XIV. figs. 11, 12 & 13). The relative po- 

 sition of the ovary is more evident if the flower is examined in an 

 early stage some time before it expands, as the style at the time 

 of expansion becomes much elongated and loses its curvature. 

 The carpel is always latei-al and internal (PI. XIII. fig. 3), i. e. its 

 dorsum is next the internal branch of the dichotomy, between 

 the forking of which the flower is situated ; and Cuphea may be 

 considered as analogous to Mirabilis in the position of its larger 

 cell *. From this peculiarity Nyctaginese are placed in connec- 

 tion with the Silenal instead of with the Urtical Orders. 



SalvadoracejE. The affinities of Salvadora have by most 

 writers been considered as uncertain, but supposing the raphe to 

 be next the placenta, the position of the fertile carpel would then 

 nearly agree with that of Plumbago, being however not so fre- 

 quently posterior. Its foliage also agrees with that of some spe- 

 cies of Statice in having a granulated appearance, and two or 

 three species have dotted leaves resembhng those occasionally 

 occui'ring in Myrsinacese. 



Plumbagine^. In Plumbago and Statice the position of the 

 funiculus is variable, being anterioi-, less frequently lateral, and 

 rarely posterior ; the mode of growth in Statice is however com- 

 plicated, which occasions its position in that genus to be less 

 obvious. 



Brunoniace^. This genus, usually considered as belonging 

 to the epigynous series, if not of uncertain affinity, should rather, 

 I would suggest, be arranged with Primulacese and its allies for 

 the following reasons. The position of the raphe in Brunonia 

 agrees nearly with that of the funiculus in Plumbago in being 

 anterior or lateral, rarely posterior, from wliich it may be in- 

 ferred, that in both these genera the fertile carpel is variable 

 from anterior to posterior, the latter position predominating 

 (PI. XIII. fig. 5). Brunonia may also be regarded as approaching 

 Dipsacese in having the same mode of growth as Morina and 

 Valerianacese, i. e. regularly dichotomous, each capitulum in Bru- 

 nonia consisting of six or eight fascicles, each of which contains 

 a centre of growth, the ramifications of which are formed by the 

 pedicel of each flower in succession becoming the axis of the two 

 succeeding ; the scales also surrounding the calyx may be com- 

 pared to the involucel of Dipsacese, and in the adhesion of the 

 anthei's an analogy may be traced with Calyceracese. 



* The irregularity of the flower in Cuphea is considered to be from side 

 to side, and the mode of gron th irregularly dichotomous, each flower being 

 terminal. 



I 



