REFORM OF LINNAEUS. 357 



Giseke's difficulties. But afterwards, in 1771, he 

 had the good fortune to spend some time at Upsal ; 

 and he narrates a conversation which he held with 

 the great teacher on this subject, and which I think 

 may serve to show the nature of the difficulty ; 

 one by no means easily removed, and by the general 

 reader, not even readily comprehended with dis- 

 tinctness. Giseke began by conceiving that an Order 

 must have that attribute from which its name is 

 derived ; that the Umbellatce must have their flower 

 disposed in an umbel. The "mighty master" 

 smiled 18 , and told him not to look at names, but at 

 nature. "But" (said the pupil) "what is the use 

 of the name, if it does not mean what it professes 

 to mean ?" " It is of small import" (replied Lin- 

 nseus) " what you call the Order, if you take a pro- 

 per series of plants and give it some name, which 

 is clearly understood to apply to the plants which 

 you have associated. In such cases as you refer to, 

 I followed the logical rule, of borrowing a name 

 a potiori, from the principal member. Can you" 

 (he added) "give me the character of any single 

 Order ?" Giseke. " Surely, the character of the Um- 

 bellate is, that they have an umbel?" Linncem. 

 " Good; but there are plants which have an umbel, 

 and are not of the Umbellatce." G. " I remember. 

 We must therefore add, that they have two naked 

 seeds." L. "Then, Echinophora, which has only 

 one seed, and Eryngium, which has not an umbel, 



18 " Subrisit o 



