PITTIER—-PLANTS FROM COLOMBIA AND CENTRAL AMERICA. 177 
Costa Rica: Along roads on the peninsula of Nicoya, Tonduz, flowers, April, 
1900 (Instituto fis.-geog. Costa Rica no. 13891; type U. 8. National Herbarium no. 
472375). 
This species differs from H. setulifera (Miers) Hemsl., with which it was associated 
by Mr. Donnell Smith, in having its floral branchlets generally opposite, the leaves 
much narrower, and the peduncles longer, and in other minor characters. 
Carpotroche Endl. 
There are here to be described three new Costa Rican species of this genus, the 
hitherto known species of which are confined to the eastern part of tropical South 
America. Of the new species one ((. platyptera), at least, is rather common in the 
shady, damp forests of the Atlantic plains and lower hills of Costa Rica, and will doubt- 
less be found also in Panama and in the zone of permanent rains in the northern part 
of Colombia. No reason can be assigned why these interesting trees have so long 
escaped the attention of botanists, unless it be the difficulty of collecting them, on 
account of their large, bunched leaves, all crowded toward the ends of the branchlets. 
The genus Carpotroche was based by Endlicher¢ on Mayna_ brasiliensis Raddi.® 
Mayna itself is an obscure genus, incompletely defined by Aublet, and Zuccarini ¢ 
had already stated that Raddi’s plant, which he carefully described, had been wrongly 
included in it and should form a genus by itself, mainly differing by its winged fruits, 
but also by a few floral characters. 
Endlicher’s elaborate description of the new genus founded by him at Zuccarini’s 
suggestion was modified by Bentham & Hooker on account of the introduction of 
several new species, also detached from Mayna. In 1845, Poeppig and Endlicher ¢ 
described and figured their C. ( Mayna) longifolia (Poepp.) Benth., the first known of 
the truly dieecious species of the genus. In his revision of the genus in 1861, Bentham 
transferred this last species from Mayna into its right place, included the new C. gran- 
diflora Spruce, but ignored C. amazonica Mart., the description of which was published 
ior the first time in 1871.¢ These additions caused important changes to be made in 
Bentham & Hooker’s/ definition of the genus, and this was further modified, although 
not in its essential parts, by Warburg.g The further addition of the three new Costa 
Rican species makes it necessary to introduce again a few minor changes. The follow- 
ing definition including these amendments will cover every one of the 7 species at 
present known: 
Flowers unisexual, seldom polygamous. Prefloration imbricate. Sepals 2 or 3, 
persistent. Petals 4 to 12, in two rows. Staminate flowers numerous. lacking even 
the rudiments of a pistil. Stamens hairy, inserted on a scarcely thickened torus; 
filaments short; anthers linear, 2-celled, free or connected at base, opening in a longi- 
tudinal slit. Pistillate flowers solitary, often larger than the staminate, lacking any 
rudiments of stamens. Ovary superior, l-celled, with 4 to 8 carpels and an equal 
number of parietal placentas, provided outside with twice as many longitudinal 
wings. Styles 4 to 8, short, persistent, with scarcely capitellate stigmas. Ovules 
numerous, anatropous. Fruit a large coriaceous or ligneous capsule, indehiscent, 
ovate or rounded, provided with large, slightly undulated wings and crowned at the 
end with the persistent styles. Seeds numerous, irregularly obovate, immersed in a 
fleshy pulp that originates in the aril-like outer layer of the seed envelope, smooth 
aGen. 918. no. 5066. 1839. 
b Mem. Soc. Ital. delle Sci. 18: Fisica, 402. 1820. 
ce Abh. Miinch. Akad. 2: 363. 1837-40. 
d Nov. Gen. ac Sp. Pl. 8: 64. pl. 271. 1845. 
eSpruce; Eichler, in Mart. Fl. Bras. 13'; 437. 1871. 
JGen. Pl. 1: 125. 1862-67. 
gIn Engl. & Prantl, Pflanzenfam. 3°*: 19. 1893. 
