166 NATURAL HISTORY OF PLANTS. 



character is valueless, varying from one species to another in the 

 same genus. Nor is the presence or absence of stipules of more 

 value. Still we must lay stress on the union of the two intrapetiolar 

 stipules into a single organ in Ropalocarpea ; and the presence of 

 little glandular bodies in place of stipules and stipels in Morinyea. 

 The leaves are simple in the former, decompound in the latter. 



The flowers are usually hermaphrodite ; but polygamo-dicecious 

 in Apophyllum. The regularity or irregularity of the perianth is of 

 no use as a generic characteristic ; for in Capparis, Cleome, Cratceva, 

 Cadaba, &c, we find both regular and irregular corollas. Apetaly 

 is sometimes considered of generic import ; thus it occurs in Boscia, 

 Thylachium, and Roydsia ; but their perianth has some other pecu- 

 liarity in the number of its parts or the mode of dehiscence. But 

 mere apetaly is insufficient, for Mcerua and Cadaba have some species 

 apetalous, others with corollas. Nor is the number of stamens, defi- 

 nite or indefinite, or of carpels and placentas more serviceable. The 

 genera Steriphoma and Thylachium have been defined by the dehiscence 

 of their calyx, which will alone separate them from Capparis. If we 

 look at this last genus, and see, indivisible as it now appears, the 

 considerable variations found in the number of parts of the floral 

 whorls, in the prsefloration, in the relative form and size of the parts 

 of a single verticil, and even the arrangement of the placentas, which 

 may be parietal, or divide the ovary up into complete cells, we shall 

 understand that all these characters must lack any real significance 

 in the group as a whole. 



The same remark may be applied to the form of the fruit. Since in 

 the genus Capparis it may assume every possible form, from a perfect 

 sphere to a narrow cylinder, thirty or forty times as long as it is 

 thick, we can nowhere in this order retain genera only founded on 

 the difference of the diameters of the pericarp. 



The chief affinities of Capparidacece are in no way doubtful. The 

 order is allied to Papaveracece, Resedacece, and CrucifercB. x Of the 

 first it lacks the double corolla, the copious albumen, the milky juice 



J. G. Agabdh says, moreover (Tkeor. Syst., pseolis collaterals, floribus 4-meris et axi intra 

 209) : " Capparidacece sunt Oxalideis et Tro- rlorem valde elongato diverste." 



