248 NATURAL HISTORY OF PLANTS. 



they united edge to edge to form a one-celled ovary with parietal 

 placentas. All other AiionacecB are what have been termed /^o/y- 

 car/jict'fP. 



IV. The aspect of the anthers. — These are introrse in Eupomatia 

 only, lateral or extrorse in all the other genera. 



In the third rank come characters that are undoubtedly of less 

 importance than the preceding, as different ones are absent in 

 several genera more or less closely allied. They can only serve 

 to distinguish these genera, or at most to separate them into sub- 

 series, such as we have had to establish in the immense series 

 Anonece. We will especially call attention to the following : — 



The t3'pe of the verticils of the perianth. — These are almost 

 always trimeroas. But the binary type is found in and characterizes 

 Tetrapetalam among the Uvariece, and Disepaliim among the Unonece. 



The presence of dorsal appendages on the petals. — These only 

 exist in the secondary group, RoUmece, and their form may serve to 

 distinguish its four genera. 



The absence of the outer corolla has only been ascertained in the 

 genus Enantia. 



The consistency and dehiscence of the pericarp. — In Anaxagorea 

 alone does the fruit consist of true follicles. This one point cha- 

 racterizes the genus. The fruits of Xylojna and Cymhopetahm, if they 

 open at all, do not open in so marked and complete a way. In 

 all other AnonacecB the pericarp is indehiscent, and the fruit con- 

 sists of more or less fleshy berries. 



There are, finally, characters which, though we cannot refuse them 

 a great value in particular cases, must be relegated to the last place ; 

 for, as shown above, they never possess that absolute significance 

 which was often accorded to them at a time when the Anonacece 

 studied were relatively few in number. On enumerating these 

 characters in succession, we shall see under what exceptional 

 circumstances they may acquire sufficient importance to become 

 the stamp of a genus, or even a sub-tribe of the order.' 



1 . The prailloration. — It has no value as regards the calyx, for 

 in one single genus we may find the sepals imbricate, valvate, and 



' Most of these have been already discussed by we must refer the reader for a developmeut of 

 us in the sj.erial memoir on tlie Ancnacect we the subject unsuited to the present work, 

 published ni AJa^utonia (viii. 162, 295), to which 



