248 Observations on llnmbo\dVs Works 



But although we may in this manner, with seeming accuracy, 

 conclude from the known to the unknown plants, we should 

 nevertheless always, in our investigation of the proportions of 

 families in any country, in order to attain the truth as nearly as 

 possible, consider all the known plants of this country, and not 

 a paH only of the same. It is therefore surprising, that M. 

 Humboldt, in fixing the distribution of the families of plants in 

 South "America, p. xvi., employs only 3,880 phanerogamous 

 plants, although he himself, p. viii., estimates the number of 

 known plants in this portion of the world at 13,000, of which 

 he gives 4,500 as already known in catalogues; and what is still 

 more astonishing, p. vii., 5,500 as found by himself. In re- 

 spect to the families formerly neglected, this makes a remark- 

 able difference. So p. xvi. 333 Glumacese are given, which 

 number to 3,880, stands nearly in the proportion of -^ : 

 whereas p. 239, it is stated that M. M. Humboldt and Bon- 

 pland found in their travels 343 Glumaceae, which number, com- 

 pared with the whole number discovered, viz., 5,500, presents 

 the very different proportion of -j^. 



Whilst therefore I cannot approve of the Humboldtian me- 

 thod, I will by no means deny that a comparison of the Floras 

 of particular countries may be very interesting, for the purpose 

 of ascertaining the different proportions of the families in any 

 narrow district. But, in such a comparison, we ought to pro- 

 ceed with all possible circumspection, and fail not to take into 

 consideration the various objects and views, by which the 

 authors were actuated in the construction of their Floras. Ac- 

 cording to M. Humboldt, the ferns in Germany amount to ^\, in 

 France to -^^ *, but this difference will, in a great measure, 

 vanish, if we examine the sources of these results. Hoffmann's 

 Flora, made use of by M. Humboldt, is, in respect to the whole 

 number, very inaccurate ; but, in the family of the ferns, he has 

 hitroduced ten species, which more lately have justly been 

 omitted f. If these 10 be deducted from the 42 given, there 



• P. xiii. 



t Polypodium daitatunit incisum, trifidunij anthrisifolium, cynapifolium^ te- 

 nere,fumarmdes,p€dicularifoliumftanacetifoliumy moUe. Ofi. Wildenow's 

 Species Phtntarum. 



