OF ARTS AND SCIENCES. 81 



acknowledgment of his in(lef:itij;:i1)le and fruitful explorations of tlio 

 botany of the south-western frontiers of the United States, from Ari- 

 zona to the islands off Lower California, in whiili region he has ac- 

 complished more than all his predeeessors. 



The genus differs from Lobelia in the remarkalde adnation of the 

 stamens, as well as in the integrity of the corolla tube, at least its upper 

 portion. It soon splits from the base upwards for a good distance, 

 and, indeed, before withering the lower part of the corolla i? much 

 disposed to separate into five claws (liberating also the lower part of 

 the filaments) ; but this occurs in many species of Lobelia. From 

 Rhizocephalum, of Weddell, this is distinguished by the habit, the very 

 dissimilar lobes of the corolla and its narrower throat, and the com- 

 pletely 2-celled ovary, which probably matui'es numerous miimte 

 seeds. 



SPECULARIA Heister. Although the two sections made by 

 Alph. De Candolle are untenable, since species of the Old World {S. 

 falc(ita,{oT instance) have lenticular seeds, and an American one, much 

 confounded with S. perfoliata^has the valvular openings of the capsule 

 near the summit, yet the American species may be well distinguished 

 from the Kuropean, and into two sections, by taking account of the 

 cleistogamous flowers. These are regulaily pi'oduced in our species, 

 and not in those of the Old "World. (They have recently been said to 

 occur sometimes in aS". ttybrida and S. falcata, but 1 have not found 

 any trace of them.) They were noticed by Linnfcus in the species 

 known to him, but were wrongly thought to be incomplete and imper- 

 fect, as Dr. Torrey remarked, when he accurately described them over 

 half a century ago. They were what we termed precociously fertilized 

 flowers, of necessity close-fertilized, — a term to which Mr. Darwin 

 took objection in the instance of Impatiens and Viola, and with reason 

 if the name stands in the way of recognition of their special adaptation 

 to close fertilization; and the proper name of "cleistogamous" is 

 now established. But one of our species of Specularia, as well as the 

 genus Lespedeza, lends support to our original view, by showing close 

 fertilized blossoms as if in various stages of arrest of development. 

 There is a mistake in Dr. Torrey's remark (in Flora of the State of 

 jS'evv York, 1, p. 429), that Ruiz and Pavon's plate of Campanula 

 bijiora represents the two kinds of blossoms. It is curious that this 

 species should have been confounded with our common northern one, 

 as by Alph. De Candolle in the Prodroraus, and by Mr. Bentham in 

 his recent notes upon that order in Jour. Linn, Soc. 15, p. 13. The 



VOL. XI. (n. 8. II.) ^ 



