PROF. T. H. HUXLEY ON THE GENTIANS. 109 
but, in the great majority, the presence or absence of nectarial 
surfaces has yet to be determined. Many are said to have 
“rotate” corolle ; but in all such which have come under my 
observation the corolle are really hypocrateriform. Usually, 
the corolla varies from campanulate to long-tubed infundibulate 
forms; and the cup or tube is thin-walled, elastic, and pretty 
closely embraces the ovary. The filaments are inserted, some- 
times very low down, and sometimes very high up in the tube. 
In some cases the anthers are united round the style, and the 
stigmas form a circular disk above the synantherous staminal 
ring (Voyria). 
a. Corolle short-tubed (hypocrateriform). 
Lapithea, Exacum, Chironia, Sabbatia, Dejanira. 
ß. Corolle long-tubed (infundibulate). 
Chlora, Erythrea, Canscora, Coutoubea, Prepusa, Lisianthus, 
Tachiadenus, Belmontia, Voyria. 
Here, as in preceding cases, the principle of the observed 
modification is obviously the interposition of obstacles in the way 
of access to the deep-seated portions of the cavity of the 
flower. In this respect, the flowers of Woyria and those of the 
long-tubed Lisianthus and Tuchiadenus are models of obstruc- 
tiveness. 
I have examined species of the following genera:— 
. A. B. C. D. E. 
Erythrza. Erythrza. Erythrzea. Erythraa. Erythrza. 
Chiora. - 
Sabbatia. Sabbatia. Exacum. 
Dejanira. Seba. Sebza. Sebrea. 
Coutoubea. ^ Belmontia. 
Prepusa. Tachiadenus. 
Tachia. Chironia. 
Lisianthus. — Canscora. Canscora. 
‘Voyria. T 
The nature ofthe cup-like “ hypogynous disk " of some of the 
“ Lisyantheæ ” of Grisebach requires investigation. In the 
only species presenting this structure which I have examined, it 
appears to be formed by the coalescence of expansions of the 
| bases of the filaments. 
Ay 
Ad 
