M0NIMIAGE2E. 319 



III. The fruits are also drupaceous in the TAMBOURissEiE ; but the 

 common receptacle instead of freeing them, becomes hypertrophied, 

 and rises around them and in their intervals so as to enframe each 

 in a sort of complete chamber, surrounding the whole in a common 

 mass partitioned ofi' into as many compartments as there are drupes. 



IV. In the Atherosperme.e, on the contrary, the carpels finally 

 become free, as in the Hortoniece, but the pericarp is dry, and each fruit 

 is an achene or caryopsis, covered with numerous accrescent hairs to 

 promote dissemination. 



V. Finally, the Gtomortege^ have their carpels in contact, and 

 forming a thick stone with several cells, which is closely united 

 with the receptacle, and finally becomes with it a single drupe, 

 crowned with the scar of the perianth. 



The characters that we have not employed to distinguish these 

 five series may be ranged in three categories. 



1.. Some are constant, and hence cannot serve to subdivide this 

 group, but only to separate it from certain other orders more or less 

 closely allied. These are as follows : The concavity of the floral 

 receptacle, and its direct consequence in the perigyny of the perianth 

 and androceum ; the imbrication of the pieces of the perianth and an- 

 droceum ; the primitive existence of two cells in the anthers ; the com- 

 plete or nearly complete anatropy of the ovules ; the direction of the 

 micropyle, which is always introrse when the ovule is descending, or 

 what amounts to the same thing, extrorse when it is ascending; 

 the absence of stipules ; and finally, the consistency of the stem, all 

 known Monimiacea being trees or shrubs, never herbaceous plants. 



2. Other characters are nearl}^ constant ; we only find very rare 

 exceptions, usually occurring in tribes or genera, all the other 

 features of which are found in those species in which one of these 

 nearly absolute characters is wanting. Such exceptions are — the 

 alternation of the leaves, found in a single Tamhourissa^ and two 

 other species of the order which are still doubtful or little known ;- 



' See p. 305, note 2. ceptacle bearing; on its edges a variable number 



"^ Tlie one has been noticed by Asa Gray as (6-15) of unequal imbricate leaves, and in its 



probably belonging to ^^Aero.^i^ei-wea' (Jo«>'«. q/* concavity an equally variable number (8-12) of 



liol., iv. 83). The other is a plant from New- stamens, like those of Hedi/carya, cacli consisting 



Caledonia, whose male flowers alone are known of a sessile erect introrse anther, dehiscing by two 



as yet, and which we provisionally name Am- longitudinal clefts. These flowers are solitary 



borella trichopoda. Each flower is borne on a or fascicled on the wood of the branches or in 



long pedicel, and may be described as a very the axils of the leaves, which are irn'gularly 



small flower of Hedycarya, with a concave re- elliptical or oval, and coarsely crenulate. 



