Anonacea^ flora indica. 89 



The ovaries of Anonacea are generally very numerous and small, and closely 

 packed together. la JJvaria they are columnar, and quite straight, and grooved 

 along the inner face ; but generally they are rounded on the back, and oblong in shape. 

 They are usually very hairy, but sometimes perfectly glabrous. This character, though 

 constant in each species, is of no avail for the distinction of genera. The style is in- 

 variably terminal, and is either continuous with the ovary, and uudistingiiishable from 

 it except by the absence of a cavity, or separates by a joint. In the latter case the 

 mass of styles often coheres together by means of a viscid or gelatinous fluid. The 

 style is usually grooved on the inner face, and is stigmatic over its whole surface, and 

 often covered with papilli. Sometimes it is short and capitate, more generally ob- 

 long, and occasionally elongate and subulate. With occasional exceptions, which will 

 be noticed under the genera in which they occur, these characters seem constant. 

 The number of ovaries is of less value. In Xylopia, a very natural genus, they vary 

 from one to ten, and in Orop/iea from three to fifteen. In Asimina and other genera 



their number is equally uncertain. 



The number and position of the ovules are of great importance as generic charac- 

 ters. When solitary, the ovule is either erect from the base of the cell, as in JJnona 

 and (hiatteria, or attached to the ventral suture, as in Etlipeia and in some Mili- 

 uste. In Artabotrys and Anaxagorea there are always two collateral ovules, erect 

 from the base of the cell. When the ovules are definite, and attached to the ventral 

 suture, their number seems less constant. Thus, in JJnona they vary from two to 

 eight, but are nearly constant in each species. In Miliusa they vary from one to 

 two, and in Xylopia from two to six. In Polyalthia and Phaanthus there are two 

 superposed ovules inserted very near the base of the cell, one of which seems occa- 

 sionally absent, in which case Polyalthia is with difficulty distinguishable from 

 Guatteria. When the ovules are numerous they are arranged more or less distinctly 

 iu two rows, and are closely packed together : they are then occasionally subdefinite, 

 especially where the ovary is very short, but this is in no case a character of generic 

 value. The section Kentia of Melodorum, where they are reduced to two, is the only 

 very marked exception to the importance of the difference between definite and inde- 

 finite ovules in the Order. 



The shape of the stamens forms a very important character in Anonacea, when- 

 ever it deviates from the ordinary type. This type, which depends mainly upon the 

 great compression of the anther, is nearly sessile, cuneate, tetragonal, with two 

 dorsal cells almost in contact with one another, and the connective produced be- 

 yond the anthers into a depressed rounded head. More rarely the cells are distant, 

 and almost lateral. The process of the connective is, however, in some genera elon- 

 gated, and not at all depressed or truncate. In one section of JJvaria the anthers 

 are flat and almost foliaceous ; and in the whole group of Saccopetalea? they are ovoid, 

 with a scarcely conspicuous process of the connectivum. When the stamens are 

 definite in number they are very irregular in shape, but usually trapezoidal, with a 

 thick fleshy connective and small dorsal anther-cells. 



The torus varies remarkably in amount of development. Whore the number of 

 ovaries and stamens is definite, it is very small; but in general it is large aud con- 

 spicuous, being sometimes cylindrical and elongated, as in Pa. ncuhts or Magnolia, 

 but more generally conical, somewhat after the fashion of R*6us 7 or broadly cylin- 

 drical and truncated. It is not unfrequently slightly concave iu the centre ; and this 

 concavity becomes extreme in Xylopia, where the stamens are borne on the outside 

 of the torus, which completely encloses the ovaries. The modifications of this organ 

 are very constant, but not always sufiScieutly capable of definition to render them 



available to the systematist. 



The shape of the petals has been much neglected in the formation of genera. 

 Rlume, however, has employed it as a sectional character in JJvaria, under which 

 genus he has united most of the many-ovuled Anonacea, and also in Polyalthia, 

 in which he includes many of those with two ovules. The sections thus formed, 

 are highly natural, as the species included in them agree very closely in habit ; 

 and we have accordingly raised them to the rank of genera, following an indica- 



N 



