SYSTEMATIC BOTANY. 477 



marginal union of floral organs of the same whorl, when occurring, must be 

 considered merely as an exception, in opposition to the universality of this 

 process. Puchartre's most important discovery relates to the position of the 

 stamens, and serves to confirm the supposed affinity of the Malvaceae to 

 the Rhamnaceae. After the formation of the calyx, the stamens are de- 

 veloped somewhat before the corolla (as in several families with the stamens 

 opposite), as five rudiments of leaves (mamelons) which alternate with the 

 segments of the calyx. These divide while their formation is yet scarcely 

 completed, at first into two segments (dedoublement collateral), in the same 

 manner as a divided leaf (their development proceeding more rapidly at the 

 two sides than in the median line, the five primitive eminences become five 

 pairs of minute rounded tubercles). Almost simultaneously with the division 

 of those stamens which are first formed, the petals appear; they are opposite 

 the former, and a considerable distance apart. The polyandrous character 

 is produced by the frequent repetition of the same formation in front of the 

 above ten stamens, united in pairs, that is, on their inner side (parallel de- 

 duplication : five new pairs of tubercles are formed in a circle, which is 

 situated more internal and opposite to the first). This multiplication of the 

 stamens is not considered by Duchartre as arising from the formation of 

 new and opposite whorls upon the torus ; but he appears to regard them, 

 and certainly with truth, as formed from the expansion of the substance of 

 the primary leaf toward the interior. The Polyandrous character is often 

 increased by a second collateral division of the individual stamens. In fact, 

 in Malope trifida y and some other species, Duchartre has even finally observed 

 a third collateral division, both of the anthers and of the stamens, so that 

 here, and perhaps generally, the unilocular anthers ought to be considered as 

 the halves of a truly dimidiated stamen. Five teeth to the tube of the fila- 

 ments, which alternate with the petals, appear to be universal in the buds, 

 and are considered as forming a second circle of stamens, without any con- 

 vincing argument being brought forward. Regarding the pistil of the Mal- 

 vaceae, Duchartre assumes the existence of five primary forms, the two 

 former of which agree in the circumstance, that at first a pentagonal collar-like 

 protuberance (bourrelet pentagonal) arises from the torus at the circum- 

 ference of the apex of the axis (mammelon central), the angles of which are 

 opposite the getals (at least this position is -mentioned as occurring in 

 Malope} ; either numerous carpels then shoot out from the margin of this 

 protuberance (Malopeae), or five only from its angles (Hibisceae). The 

 formation of the carpels is also preceded in the Malveae and Sideae 

 by a protuberance, which is not, however, pentagonal, but annular. The 

 number of carpels growing from its margin is undetermined. Lastly, the 

 greatest deviations occur in Pavonia, and some allied genera, in which, upon 

 an annular protuberance, the rudiments of ten styles arc said to appear first, 

 and subsequently to be fused into five ovaries. 



