of the Family of Bufinanniacese. 377 



dried state in Thismla, although it has 6 almost imperceptible nervures ; the 

 corona is circular, not hexagonal: it differs also in the shape and disposition 

 of the filaments, which are quite distinct and unconnected, while in Thismia, 

 although at first separate in their origin, they soon become united into a 

 long and completely monadelphous pendulous tube with 12 distinct ner- 

 vures*: the anther-cells are here arranged in 6 pairs, near the commence- 

 ment of the line of tubular junction, each pair consisting of 2 distinct and 

 unconnected lobes, separated by the nerval lines that spring from the corona; 

 alternate with these are 6 other nervures, arising from each sinus formed by 

 the junction of the filaments, and terminating in as many coloured glands, at 

 the points where the tube again becomes cleft into a double line of 6 equal 

 pairs of thin parallel truncated laminse, one overlapping the other, each being 

 fringed at its extremity by 3 or more teeth. In Ophiomeris, on the con- 

 trary, the filaments are in no way monadelphous, and are considerably nar- 

 rower at their origin ; the anthers are placed, not towards the middle, but 

 close to the apex of the filament, which is remarkable for its 2 obtuse termi- 

 nating lobes and winged lateral appendages ; the 2 cells of the anther, too, 

 are laterally conjoined, not in separate lobes as in the Asiatic genus. 



The alliance of Thismia to Burmanniacece did not escape the notice of Mr. 

 Griffith, although on the whole he was somewhat more disposed to compare it 

 with Tacca, on account of the analogy in the number and direction of the sta- 

 mens fixed on a 6-partite perianthium, and the parietal placentation of the 

 seeds : but it is clear that this eminent botanist was not then aware of the 

 structure of the ovarium in the section Apteriece, which I have proposed in the 

 last-mentioned natural order. In fine, I have no hesitation in referring both 

 Thismia and Ophiomeris to Burmanniacece, notwithstanding that both these 

 genera differ from all others of this order in the greater number of stamens, 

 their reflected position, the annular corona, the circumscissure of the perian- 

 thium, and the opercular opening of the capsule ; but the character of the order 



* This is rather indistinctly shown in vol. xix. pi. 39, illustrative of Mr. Griffith's memoir above-cited, 

 •where the separation of the filaments at their origin, and their subsequent junction into a tubular form, 

 are not made evident ; the former circumstance is however distinctly stated in the text, but the latter 

 is too ambiguously worded in page 341 to enable a reader to comprehend the author's true meaning; 

 it is difficult indeed to say which part really constitutes the filament, and which the connective, for the 

 whole appears one homogeneous petaloid membrane. 



3 D 2 



