354 NKW I'ALMS OF TDE VALLKY OF THE AMAZON'. 



could be determined from tlie description quot(>d, thoufjli the absence 

 of spathc and spadix prevents identification. 1 think myself justified, 

 for the above reasons, in regardin<? D. 2)umilus as an undescribed 

 species. It is called " Jacitara pui" in Lingoa geral. 



Description of Tab. 183, Rephesentinc. new Palms from the Amazon. 



I. GcoHoma oligoclonn, n.s. — 1. A leaf, ^'nat. size. 2. Spadix and spathes, ^ 

 nat. size. 3. Portion of spadix, nat. size. 4. Portion of branch to show alveole. 

 5. ^ flower, nat. size ; parts dissected, x 2j. 6. ^ flower ; "parts dissected, x 3. 



II. G. leptospadix, n.s. — 1. A leaf, J nat. size. 2. Spathe, and 4. Spadix, J nat. 

 size. 3. Spathes, nat. size. 5. Portions of branch showing alveole and arrange- 

 ment of flowers. 6. Spadix with pedunclo flexed and branches crowded, as fre- 

 quently luund under the mass of rubbish (see description). 7. ^ flower, 

 dissected, and androoceuin of $ flower, X 3. 8. Fruit and nucleus, nat. size. 



III. Calyptronoina? robusla, n.s. — 1. Portion of branch of spadix. 2. Fruit. 

 3. Ditt'erent views of mesocarp to show arrangement of fibres. 4. Nucleus. 

 5. Section of nucleus to show position of embryo. All are of the natural size. 



JIV. Dr.smoiicus pumilus, n.s. — 1. A leaf, ^ nat. size. 2. A leaflet, nat. size. 

 3. Spadix and spathes, J nat. size. 4. $ flower, x 6. 5. A fruit, nat. size. 



Genus Bactris, Jacq. 



This genus alike in number of species and in number of individuals 

 is by far the richest in the Amazon Valley, Geonoma alone approaching 

 it. Owing to the comparatively small size of the species, however, 

 they do not aid much in giving a marked character to the forests, by 

 far the most species being hidden among the undergrowth, or among 

 the dense jungle that springs up on deserted clearings. They must, 

 therefore, be looked for, but when looked for arc found to bo absent 

 from few localities save opeu campos, in whicli 1 do not remember 

 ever to have seen any species. 



1 use the generic name in the wide sense, as it seems to me that 

 the various genera proposed by Karsten, Oersted, and others are not 

 founded on good characters, though some of them may with advantage 

 be employed lor dividing the genus into sections. Were the possession 

 of a sterile androccium or its absence to be regarded as good grounds 

 for founding a new genus (as Karsten has done), then very closely- 

 allied species must be separated, so that the genus Bactris (as it now 

 stands) would be broken up into two genera containing series of 

 species more or less parallel. However, this is not the place to enter 

 into a discussion on the value of the various genera into which it has 

 been proposed to break up the genus. 1 subjoin a key, with short 

 diagnoses of the species and varieties described below. ^N'here no 

 author's name is (juoted after a species or variety described below, 

 such species or variety is described here for the first time. 



Analytical key of the genus Eactris, Jacq. 

 § 1. (= § Humiles and § SphcerocarpcE of Martius.)— Fl. ? andrcccei 

 vestigium nullum. 

 A. Microcarpcc. — Fructus globosi vel obovati, pisi magnitudinis, 

 rubri (demum purpureo-nigri) ; putamen globosum vel turbi- 

 uatum ; palma? humiles i^K. tomentosa mediocris), tenues, 

 inermes vel aculeatoe ; Iblia simplicia vel pinnata ; pinnis con- 

 cinnis ; spadix simplex vel in ramos 2-7 fissus. 

 B. Fl. ? calyce ct corolla ciipularibus ; spadicc siniplici vel bifido ; 

 liuctibus inermibus. 



