THE TRIUMFETTAS OF AFRICA. 231 
The TRICMFETTAS of Africa. ` 
By T. A. SPRAGUE, B.Sc., F.L.S., and J. HUTCHINSON. 
(PLATE 17.) 
[Read 17th June, 1909.] 
Our attention was directed to the genus Triumfetta by the great difficulty 
experienced in determining the new African material whieh was received 
from time to time at the Kew Herbarium; and in order to overcome this 
difficulty, a provisional arrangement of the African species was drawn up 
some years ago. The nature of the fruit was taken as the basis of classifica- 
tion, as affording easily ascertainable and, in the majority of the species, 
highly constant characters ; and further study has shown the choice to have 
been a happy one. The importance of the characters derived from the fruit 
was appreciated to some extent by Masters in 1868, when he described the 
15 species then known from Tropical Africa (Fl. Trop. Afr. i. 254), but it 
has not always been realized: thus Sehumann wrote that it was extremely 
difficult to distinguish the species, owing to the polymorphy of the leaves and 
fruits (Mart. Fl. Bras, xii. 11. 181). In dealing with the African Triumfettas, 
however, we have found none except T. heterocarpa of which it could be 
said that the fruits are polymorphic, although in a few, such as 7. cordifolia, 
the indumentum and size of the fruit appear to be subject to considerable 
variation. 
The scope of the paper as originally planned was limited to the species 
found on the continent of Africa, but it was afterwards found desirable for 
purposes of comparison to include those of the African islands in the widest 
sense. In our area, as thus extended, all the more important groups of 
Triumfetta are represented, and the classification we propose, while designed 
primarily for the African species, is therefore applicable to the genus as a 
whole. 
Previous authors, from A. P. De Candolle to Baillon, have recognized two 
sections of the genus, based respectively on Triumfetta, L., and Bartramia, 
L., as redefined by Gaertner in 1791 (Fruct. i. 137, t. 111). Linneus 
(Gen. Pl. ed. 5, 203) distinguished Triumfetta from Bartramia by the absence 
of a calyx [corolla] ; and Gaertner added that the fruit of Triumfetta was 
indivisible, with 1-seeded cells, whereas that of Bartramia was divisible into 
2-seeded cocci. 
