232 MESSRS, T. A. SPRAGUE AND J. HUTCHINSON ON 
In 1824, A. P. De Candolle (Prodr. i. 506) recognized two sections of 
Triumfetta as follows :— | 
I. Lappura, DC. (= Triumfetta, Gaertn.).  Petals wanting ; fruit 
indivisible. 
П. Barrramea, DC. (= Bartramia, Gaertn.). Petals present; fruits 
divisible. 
Endlicher in 1840 (Gen. Plant. 1008) and Masters in 1868 (Fl. Trop. Afr. 
i. 254) adopted De Candolle’s sections, defining them entirely by the fruit 
characters: LaPPULA, with indehiscent fruit; and ВАВТВАМЕА, with de- 
hiscent fruit. Baillon in 1873 (Hist. Pl. iv. 195) replaced the name 
Lappula by Eutriumfetta, still distinguishing it from Bartramia (sic) by the 
indehiscent fruit. Finally, in 1886, К. Schumann (Mart. Fl. Bras. xii. тп. 
131) rejected the characters derived from the fruit, and defined the two 
sections as follows :— 
EUTRIUMFETTA. Petals absent ; gonophore very short, without glands. 
BARTRAMIA. Petals present ; gonophore with 5 glands. 
Schumann’s sections are clearly unnatural, since they involve the separation 
of two such closely allied species as T. Lappula and T. semitriloba, of which 
Linnæus remarked in 1707, * T. Гарри calyx nullus . . . ., at T. semitrilobe, 
Jacq., calyx perfectus ....; utrasque simillimas distinguere an specie liceat?” 
(Gen. Pl. ed. 6, 239). Apart from the great reduction in the flower, there 
is very little to distinguish 7. Lappula from T. semitriloba ; and when it is 
considered that in 7. pentandra, Rich., a series of intermediates has been 
found between flowers with 13 stamens and a well-developed ciliate dise and 
others with only 5 stamens and hardly a vestige of the dise, there can be no 
doubt that 7. semitriloba and T. Lappula are very nearly related. 
Nor does a primary division between the species with dehiscent and those 
with indehiscent fruits give satisfactory results. The texture of the fruit, 
nowever, and the nature of the prickles or bristles which it bears, yield 
excellent characters, which, associated with others derived from the inflores- 
cence and the indumentum of the sepals, have enabled us to divide the genus 
into four apparently natural sections. 
In the simplest type of inflorescence the cymes are solitary at the nodes 
and are opposite the leaves, the branching being sympodial. This and its 
modifications are characteristic of two of our proposed sections, Porpa and 
Lasiothriz. In а more complex type, described by K. Schumann (Engl. u. 
Prantl, Nat. Pflanz. iii. v1. 10), several (two or more) eymes are borne at 
éach node, the primary one being opposite the leaf, and the remainder arising 
on one side of the stem between the insertion of the leaf and the primary 
суше. This type is characteristic of the sections Lepidocalyx and Lappula. 
