MR. BENTHAM ON LOaANIACEjE. 



G3 



i 



and a fifth is added by Zollinger and Moritzi in their Catalogue of 

 Java plants, but of these, M, paniculata, supposed to be eommoii 

 to Brazil and India, appears to have been founded on insufficient 

 materials. All Gardner's Brazilian specimens which I have seen 

 have the fruit and all other characters of M, petiolata, which has 

 a wide range from the Southern United States over the West 

 Indies to equatorial America, whilst Wallich's Indian specimens 

 do not appear to differ from his M, oldenlandioides*. It is to 

 that species also that I would refer Zollinger's plant. It has a 

 much larger capsule than M, petiolata^ narrowed at the base, with 

 longer horns, which, as the fruit enlarges, become very broadly 

 divergent at the base ; the pod of M. petiolata is smaller, more 

 globular, and the short horns, at first erect, diverge but slightly 

 as the fruit enlarges. In both species the horns often curve more 

 or less inwardly, but more so in M, petiolata than in the majority 

 01 specimens of M. oldenlandioides ^ and in neither is it a constant 

 character. The seeds of the American species are always much 

 broader than in the Indian one, although both vary in this respect, 

 lo these annuals I have to add a very distinct perennial species 

 gathered by Drs. Hooker and Thomson in Khasiya. 



7. MiTKASACME, Ldbill. 



^ -• V- 



::. £: .< 



■^ ^ 



, ■ 



MitroBacme is closely allied to Mitredla in flower^ and in fruit, 

 and partakes of its affinities. The technical distinction, consist- 



• • • 



ing in its tetramerous, not pentamerous flowers, is however accom- 

 panied by a difference in habit which approaches rather to that of 

 some slender Gratioloid genera than to Oldenlandia, which, in that 

 i-espect, is its Eubiaceous representative. The capsule is also more 



occasionally 



Mitreola, and the tube of the corolla 

 I find the aestivation of its lobes always 



Ivate, as suspected by Alph. DeCandolle. 



Nineteen Australian species distributed in four divisions were 



Brovv7i 



added 



Australian ones, of which, however, two only prove to be really 



^ight names. 



Brown 

 shall 



I find in our herbaria ; but as I propose to reduce to varieties three 



stands at twenty-four 



Brown's divisions have been adopted by subsequent botanists 



• In Wight's ' Icones,' t. 1601, a curious mistake of the artist has occurred 

 in the flowers of the gmieral fig^ of M. paniculata ; the analysis, however, 

 drawn by Dr. Wight himself, correctly repWf ents the true structure. 



V* 



■ '^ 



,s 



'r^ 



f. - 



^ ■■-. 





.r ^ 



-\V- :^ '.^ <," 



r.i: 



