.1,1", '■ 



into tin; | i<'!i the composer of that character bad m.t 



«M)ntcii)j)l;it<M |. 



The mum law applies to natural orders, which are merely nat 

 groups of genera, or large natural genera; but here, unfortunately, 

 practice ami fcheoi -y in often iridelj in opposition. Numerous instances 

 miiiht bfl quoted. Thus UnwiiMHiliiieffi, as defined in some of our 

 British Floras, is said fco be polyandrous, whereas Myosurus has never 

 more than five stamens. If Myosurus had not been ■ native of Britain, 

 I would not object to this, because if we can abridge the character of a genus 

 or order, by omitting all that has no reference to the species describe 1 in 

 the book, it is a great boon to the student, or to one who is not engaged 

 in general botany ; but such abridgment must not be at the expense of 

 accuracy. In the same way I do not object to European, or N 

 American, or even medical floras stating the Violaceae to have irregular 

 corollas, because every species found in Europe, in North America, or 

 used in medicine, has such ; but if we published the flora of Guiana or 

 East India, we cannot restrict the order in this way, because plants do 

 occur there with quite regular flowers, and differing in no other respect 

 from that order. In a general work, we must therefore have a general 

 or universal character, but it is there that we find numerous failures. 



In nature every thing is continuous, and thus there may be said 

 to be but one great natural order, perhaps only one genus: every 

 attempt to break it up, and form smaller orders, must be, in a certaiu 

 degree, artificial: and there can be but one object in breaking it up, 

 that of conveying information more easily to others, about the plants 

 that are already described. Now, in order that the affinities of such 

 plants may bo exhibited by their relative position, it is of importance 

 that such divisions be as natural as possible : but in order that such 

 divisions be also useful, each must be rigorously defined. There are, 

 therefore, two elements inseparable from each other, and it is the 

 judicious combination of these that must limit a natural order. 



On a former occasion I made some remarks to this Society on the 

 Chrysobalaneae. So long as this was retained in the great group of Rosa' 

 there was necessarily a very great latitude in the ordinal character ; but 

 tho almost impossibility of framing one applicable to all the genera, ami, 

 at tho same time, sufficiently definite so as to exclude other or d 

 induced botanists to divide it into several, the more as there were three or 

 four tolerably well marked groups. Perhaps no living botanist has 

 studied that order with more attention than Dr. Lindley ; and yet, whei 

 from the desire not to split it up into too many orders, or from trusting too 

 much to natural appearances, we find that some parts of the characters 

 are in the above position ; I allude particularly to his Sanguisorbeaa and 

 ftpmOPCP (proper). The former (I refer to the Vegetable Kingdom, as 

 tuin lining tho latest views on the subject), he defines as apetafous, \>itli 

 a solitary carpel inclosed in the hardened calyx tube and forming a false 

 up, ami with the ovnlc Bofitarj ( to 



