436 
PHYTOGRAPHY. 
BOOK IV. 
be most essential to it ; but admits nothing unimportant or 
superfluous, or that is common to all the species of the same 
genus, or to all the genera of the same order, or to all the 
orders of the same class. It may be said to comprehend the 
chief differences and resemblances of bodies. In drawing up 
essential characters, much discretion requires to be exercised : 
they may be over short, or over long ; characters of import- 
ance may be omitted, and others of no importance intro- 
duced. Hence no better evidence need be desired of the 
merit of a botanist than his essential characters, — from which 
a practised eye will readily detect both how much the author 
knows, and what he does not know. As models of the man- 
ner in which these should be drawn up, no book can be 
consulted with more advantage than the Genera Plantarum 
of Jussieu, in which classical elegance of language, and as 
much rigid botanical precision as was supposed necessary at 
the time the work was written, are combined in a manner that 
has seldom been surpassed. The defects of that work were 
inseparable from the state of botany at the time it appeared ; 
the characters of the genera and orders not embracing all 
those points of structure which are now known to be essential. 
The following character, assigned by Brown to the order 
Proteace^ [Prodr. FL N. Holl. p. 363.), may be taken as a 
specimen of the manner in which an essential character of the 
briefest kind ought to be constructed : — 
" Perianthium tetraphyllum v. quadrifidum, aestivatione 
valvata. Stamina quatuor (altero nunc sterili), foliolis peri- 
anthii opposita. Ovarium unicum, liberum. Stylus simplex. 
Stigma subindivisum. Semen (pericarpii varii) exalbumino- 
sum. Embryo dicotyledonous (quandoque polycotyledoneus), 
rectus. Radicida infera." 
In this character enough is expressed to distinguish the 
order from all others ; and, at the same time, by a careful 
suppression of all superfluous terms, it is reduced within 
exceedingly narrow limits. Such a character as this leaves 
nothing to be desired, when the essence only of a mass of 
characters is the object in view. 
The following, from the same author, is a specimen of an 
essential character of Acanthaceji:, of a more extended 
kind : — 
