i 



188 NOTICES OF BOOKS. 



I 



previously given, but because these proportions depend upon different 

 laws, sometimes of one kind, sometimes of another, sometimes general, 

 sometimes local. This last objection appears to us to apply to every 

 branch of the study of Botanical Geography ; whilst the very fact 

 that Monocotyledons are, as compared with Dicotyledons, more widely 

 spread, more variable as species, more difficult of association in limit- 

 able genera, and more difficult of distribution into Orders characterized " 

 by structural or physiological characters, demands the closest investi- 

 gation of the local and general conditions of the countries in which 

 they appear in unduly great or small proportion, relatively to the 

 Dicotyledons. The first step to be made in such an inquiry is un- 

 doubtedly to ascertain their numerical proportions. It matters not 

 that the class^ of Monocotyledons is represented by very diiferent ge- 

 nera, or even Orders of plants, in the different countries which afford 

 data to start from, for the variable element is everywhere present 

 in a greater degree amongst Monocotyledons than Dicotyledons. 



Chapter 22. On the comparison of different countries with respect 

 to those Natural Orders which abound most in Species. — For this in- 

 vestigation M. de CandoUe has collected an invaluable series of tabu- 

 lated materials, at great labour, for which alone he merits the thanks 

 of his fellow-botanists. Upwards of 130 general and local floras Tiave 

 been submitted to analysis, and the proportions of the seven or eight 

 largest Natural Orders contained in each are given, which, on the aver- 

 age, include half the Phaenogamic plants in each Flora. The naturalized 

 plants are (we think unfortunately) included ; on the one hand, it would 

 have been difficult to have eliminated them perfectly ; but on the other, 

 their introduction has sometimes led to the most serious errors. Thus 

 the flora of Ascension, which numbers only 4 or 5 native flowering plants, 

 is represented aa containing 39, many of the additional species not being 

 even naturalized, but garden plants; whilst, instead of taking Roxburgh's 

 list of St. Helena plants (in Beateon's Tracts), which includes about 30 

 species and is a very near approach to the truth, that of Antomarchi is 

 preferred, as "la moins pitoyable des quatres Flores publiees jusqu'a 

 present,'* and of which all we can say is, the least contemptible is the 

 most ridiculous. 



There axe two sources of difficulty in the investigation of the Floras 

 analyzed in this Chapter, which are quite inseparable from the sub- 

 ject ui the present state of our knowledge, and which arc, firstly, 



