SAFFORD—CLASSIFICATION OF ANNONA. 5 
author has been unable to secure specimens of these African species 
of Annona, he can not with certainty assign them to sections." 
SPECIES INCORRECTLY REFERRED TO ANNONA, 
Several African plants originally described as Annonas have been 
found to be generically distinct; as in the case of Annona mannii 
Oliver and A. laurentii Engl. & Diels, which have been placed by 
Engler and Diels in the genus Anonidium. In this genus, the flowers 
are more or less unisexual, as in Raimondia, but the stamens differ 
radically from those of the latter genus in having the connective 
thickened and dilated at the extremity over the pollen sacs. 
Several American plants were removed from the genus at an early 
date on account of peculiarities of their flowers or fruits, as in the 
case of Rollinia mucosa and R. silvatica and several species of 
Duguetia. A few others are undoubtedly distinct and should be 
raised to generic rank, as has been done in the case of Anonidium; 
among them, Annona longifolia Aubl. and A. rhizantha Kichl., the 
latter of which bears the same relation to the genus Duguetia that 
Raimondia bears to Annona. Included in material recently received 
from M. Augustin De Candolle the author finds a species, described 
by Kunth under the name of Annona quinduensis, which, on account 
of its peculiar stamens and unisexual flowers, must be assigned to 
the recently described genus Raimondia. Of this species an amended 
description is given hereafter, together with descriptions of two new 
genera: Geanthemum, already recognized as a distinct section by 
Robert E. Fries, and Fusaea, based upon Baillon’s section of the 
1 Since writing the above the author has received, through the kindness of 
Professor Urban, a series of fine specimens of African Annonas from the 
Berlin Herbarium, including original specimens, or type material, of Annona 
klainit Pierre. He was surprised to find this species very closely allied to 
A. glabra L. (A. palustris L.), which also occurs on the east coast of Africa, in the 
same region as that in which A. klainti was collected. It is possible that the 
type of the latter species is an abnormal specimen of A. glabra, which not 
infrequently has a 4-parted calyx and 8 petals in 2 series of 4, both the inner 
and outer series valvate in westivation. Annona senegalensis Pers. is remarkably 
variable, occurring sometimes as a small shrub less than a meter high, and 
sometimes as a large tree, as in the form occurring in the vicinity of Lindi, on 
the east coast of Africa (Busse, nos. 2823, 2990°). The flowers are 6-petaled, 
with the inner petals narrow, connivent, their tips meeting above the center 
of the gynecium. The seeds are small, oblong, hard, smooth, and glossy, with 
relatively large caruncles at the base, somewhat like those of A. cornifolia and 
A. nutans of southern Brazil and Paraguay. It is possible that the species 
A. senegalensis as now understood may be found to be composed of several 
species. Certainly several of the forms show more decided differences than 
those which separate A. klainii from A. glabra or some of the South American 
species allied to A. tomentosa R. E. Fries from one another. Annona glauca 
Schum. & Thonn, is a well-defined species quite distinct from A. senegalensis, 
