186 NOTICES OF BOOKS. 



also of statistical tables, showing the proportions of forest, tilled, marsh, 

 grass, etc. land, we also entirely concur in, but cannot regard it as 

 strictly, and in detail, a branch of Botanical Geography. With regard 

 to the classifications of plants under certain forms, proposed by Hum- 

 boldt and carried out further by Meyen, they are of little service ; for, 

 however useful they appear to those who confine their attention to in- 

 dividuals in herbaria and gai'dens, they are found to be useless in prac- 

 tice to a much greater degree than is commonly supposed. 



3. Characters relating to Natural Orders, — The calculations usually 

 made to determine the proportion of species in a Family, necessarily 

 supposes, we are told, that the species of different Families are equally 

 abundant in individuals in the same country.* In this and in the three 

 following sections, devoted to characters relating to the genera, to the 

 species, and to the uniformity or variety of the vegetation of a coun- 

 try, M. de Candolle dwells almost wholly upon the exceeding vague- 

 • ness of the inquiry, the multitudes of sources of error, and the small 

 value of the results. lu the main he is no doubt right ; but when 

 he says, under the article, " On the relative value of characters of 

 vegetatioa," that he dissents from the opinion of certain botanical geo- 

 graphers, who deem it expedient to commence by giving numerical 

 data, because "exact methods" only satisfy him, and because exacti- 

 tude docs not always consist in employing figures instead of words, 

 but in giving to every fact and every point of view its true value, we 

 are tempted to ask, where is the exactness of the methods employed 

 by M. de Candolle in determining climatic conditions ? and above all, 

 what are the values of the figures employed in the articles devoted to 

 the areas occupied by species, genera, families, and classes? what in- 

 deed does any branch of the subject of Botanical Geography consist 

 of, but vague hypotheses, and a collection of facts of unknown value 

 and application ? Moreover M. de Candolle appears here to confound 

 unnecessarily two very different subjects of inquiry, which are never sup- 

 posed, by inquirers of ordinary intelligence, to have even a relative 



♦ This does not appear to us to be the ease; all these methods of ascertaining re- 

 lative proportions are confessedly extremely imperfect, nor have authors regarded them 

 as affording anything but rude approximations to truths. That laws regulating the 

 proportioua of the Families, etc., do exist, no one will deny ; and the reason of these 

 attempts to ascertain them by laborious calculations being of less absolute value than 

 would be wished, hes in the fact that Botany is not an "exact science," and that 

 J^otumeal Geography is one of the loosest branches of it 



