134 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 



sight the two arc easily mistaken for each other. Both have trunks 

 which, after creeping along the ground a few yards, ascend and 

 attain about an equal height. Their leaves also resemble each other; 

 and their fruit grows in a similar way, attached to comparatively 

 short peduncles. The habit, however, is nearly the only link which 

 connects Phytelephas with the order of Palms: its simple spadix, its 

 imperfect flower, its indefinite number of stamens, and its embyro 

 situated in the axis of a fleshy albumen, separate it from Palms, and 

 proclaim it (in conjunction with other characters which it presents) 

 a member of Endlicher's class Spadiciflorae, and Lindley's alliance 

 Arales.' ?n 



The first of Seemann's diagnostic differences, the simple spadix, 

 is certainly not peculiar to Phytelephas. Simple spadices are found 

 in several other groups of American palms, belonging to such unre- 

 lated genera as Bactris, Geonoma, Wettinia, Chamaedorea, and 

 Malortiea. 



Imperfect flowers are the rule among palms rather than the ex- 

 ception. Only the primitive fan palms have perfect flowers. The 

 completely dioecious condition found in Phytelephas is paralleled 

 in the date palms of the Old World and in the large group of 

 American palms belonging to Chamaedorea and related genera. 



The third peculiarity alleged for Phytelephas is the indefinite 

 number of the stamens, which here become much more numerous than 

 in any other palm. Spruce's Pltytelephas equatorialis is described 

 as having more than 1.000 stamens. Most other palms have only 3, 

 0, or 9 stamens, but a few genera show larger multiples of 3, as Car- 

 yota, Arenga, Iriartea, Jubaea, Attalea, and Manicaria. 



The considerable range of variation shown among the species of 

 Phytelephas forbids the assignment of any very great weight to this 

 numerical feature. Seemann described the Phytelephas of Panama 

 as having only 30 stamens, a smaller number than is found in some 

 of the undoubted palms. The Peruvian Pltytdephas macrocarpa 

 is credited by Spruce with from 150 to 280 stamens, four to eight 

 times as many as the Panama species. Another multiplication by 

 four would cany us from 288 to 1,152, which would approximate the 

 number found in the species from Ecuador. Though the numbers 

 are large they may prove not to be altogether indefinite. 



Seemann's statements regarding the seed of Phytelephas certainly 

 give no clue to a differential character. Many other palms have the 

 embryo in a similar basal position, and thus in the axis of the 

 albumen. Why the ivory -hard albumen should be described as 



« Seemann. It., Botany of the voyage of H. M. S. Herald . . . during the years 

 1845-1851, p. 210. (1852-1857.) 



