500 T&Jl. G. BENTHAM ON CLASSIFICATION AND 



cence or pedicel ; but frequently also there are two or more closely 

 imbricate at the base of the peduncle or scape, or one, two, or 

 more smaller ones scattered along it. These are usually termed 

 sheaths, sheathing scales, or simply scales, according to the degree 

 in which they embrace or envelope the stem. In the inflorescence 

 there is usually, but not always, one subtending each branch or 

 pedicel, and specially termed bract in contradistinction to the 

 scales of the peduncle or scape. In some cases (as in most 

 Aroidese and Palms) the lowest bract is much larger than the 

 others, enclosing the whole inflorescence when young ; and then 

 it receives the special name of spatha. Often, however, it is but 

 little more developed or even smaller than those next above it ; 

 and it is then termed a bract or a spatha according to its degree 

 of development in the majority of allied genera. In Pandanus, 

 in Gymnostachys, in Typha, &c, where two or more bracts under 

 some of the upper branches are as much developed as the lowest, 

 they are either all or none, or the lowest only, designated as 

 spathas with almost equal fairness. Where the spatha is well 

 developed, especially in Aroidese and often in Palms, the enclosed 

 inflorescence is usually termed a spadix ; but there is really no 

 more definite distinction between a spadix and a panicle or spike 

 than there is between a spatha and a bract. The spatha, how- 

 ever, enclosing the young spadix is very general in Palms and in 

 Aroide® ; and as, being called upon to exercise a more important 

 function than ordinary bracts, it is also fairly constant in its modi- 

 fications, it becomes an important though not an absolute cha- 

 racter, and there is great convenience in adopting the special ter- 

 minology in use. But in several other orders, where the passage 

 from the ordinary bract to the specially enlarged one is gradual 

 or uncertain, the vague use of the term spatha in technical cha- 

 racters often tends to mislead rather than to give clearness to a 

 definition. 



The Nudiflorae proper include a group of five orders univer- 

 sally acknowledged to be closely allied to each other, although 

 very variable in the structure of their seed. All are without 

 perianth or with only a small scale under each stamen. Bracts 

 are rare except a spatha at the base of the inflorescence, and some- 

 times under its principal branches. The anthers are usually ses- 

 sile, or on short thick filaments with the cells opening outwards, 

 or in terminal pores. The ovary is, with few exceptions, apo- 

 carpous or monocarpellary. The seeds are albuminous with a small 



