on the Geography/ of Plants. 



247 



The second, and indeed the most difficult, method, which 

 considers all the known plants, has also its defects, and amongst 

 them particularly this, that we are not equally well acquainted 

 with all the zones, nor with all the parts of each zone. How- 

 ever, this imperfection is of less importance than might, per- 

 haps, be imagined ; for, with the exception of some families, 

 which by botanists, particularly in botanical travels, are neg- 

 lected, (the Crj/ptogamia, and in a great measure likewise the 

 Glumacete), we may, with a tolerable degree of probability, 

 draw conclusions from those which are known respecting such 

 as are not known, since it cannot be admitted as a general law, 

 that a greater proportion of the plants of some families than of 

 others has been discovered. We may, therefore, reasonably 

 suppose, that the proportions of families amongst themselves 

 will be very little altered by new discoveries. In order to il- 

 lustrate this by example, I have compared Michaux's Flora of 

 North America with the new one by Pursh, and Linne*s Flora 

 Lapponica* with that lately published by Wahlenberg. The fol- 

 lowing are the results : — 



Micbaux. 



Pursh. 



Linne. 



iWahlen 

 berg. 



Leguminosoe 

 Umbelliferas 

 Compositse . 

 Cruciferse , , 

 Juncese , , . , 

 Gramineoe ,, 



■tiu 



or 

 A* 



1-^3" 





-h 





Now since the conformity between the formerly known and 

 lately discovered plants is so striking, even in small numbers, 

 this conformity would be even greater, were we to comprehend 

 in our reckoning all the plants of a certain zone. 



« Edit. Smithii. 



t The very small number of North American plants of this family is the 

 reason why some newly-discovered species alter so considerably the pro- 

 portion. 



