NATURAL ORDERS. 139
Prodromus Florrc Novai Hollandiae, I had it particularly in view to exclude Phytolacca, Pivina, Microtea, and Pctivcria, which I even then considered as forming the separate faniilv now for the first time proposed.
In Chenopodem the stamina never exceed in number the divisions of the perianthium, to which they are opposite. In Phjtolace(B they are either indefinite, or wdien equal in num- ber to the divisions of the perianthinm, alternate with them. This disposition of stamina in Phytolacca?, however, uniting genera with fruits so different as those of Phytolacca and Petiveria, it would be satisfactory to find in the same order a structure intermediate between the multilocular ovarium of the former and the monospermous ovarium, with lateral stigma, of the latter.
Two plants in the herbarium from Congo assist in estab- lishing this connection.
The/;'.5^ is a species of PhfjtoJacca, related to P. abys- sinica, whose quinquelocular fruit is so deeply divided, that its lobes cohere merely by their inner angles, and I believe ultimately separate.
The second is a species of Giseliia, a genus in which the five ovaria are entirely distinct. This genus is placed by :4.->.:i M. de Jussieu in Portulacacea? ; but the alternation of its stamina with the segments of the perianthium, a part of its structure never before adverted to, as well as their insertion, seem to prove its nearer affinity to Phytolacca.^
Stiil, however, the lateral stigma, the spiral cotyledons, and want of albumen in Petiveria, remove it to some distance from the other genera of Phytolacca?, and at the same time connect it with Segidcria, with which also it agrees in the alliaceous odour of the wdiole plant.
The affinity of Sefjuicria has hitherto remained unde- termined, and is here proposed from the examination of three species lately discovered in Brazil, one of which has
^ Ai}clst,'ocarpn of ]\r. Kunlli (Nov. Gen. et Sp. PI. Orb. Nov. 2, p. 18G) belongs to Phytolaccre, though its stamina are described to be opposite to the segments of the calyx: and it is not improbable that ^filtiis of Loureiro (Flor. Cochin, p. 30:2) whose habit, according to the description, is that of Gisekia, from which it differs nearly as Ancistrocarpus does from Microtea, or Jlivina octandra from the other species of its genns, may also belong to this order.
�� �
