r 



LARGE GENERA AND VARIABLE SPECIES 456 



or a few well marked unvarying species, especially if its 

 generic name is a very familiar one, hence Amygdalus, 

 Prunus, Cerasus, are kept up, though certainly not good 

 genera in a scientific view of Eosaceae. Few plants are more 

 variable than Hawthorn — it is a small genus dismembered 

 from Pyrus, but no British author makes varieties of it. 

 Genera in short are almost purely artificial as established 

 in Botany : some are objective like Salix and Bosa, i.e. every 

 ignoramus recognises them and they are called natural 

 genera, good genera, &c., &c. Others are subjective, they 

 require a special knowledge of the Order to which they 

 belong to know them — ignorami do not recognise them : 

 such are genera of Grasses, Crucif erae, Umbelliferae, &c. 

 But between what the ignoramus does recognise and does 

 not there is no limit ; and the first rate Botanist, working 

 upon a partial knowledge of a group, is only in the position 

 of an ignoramus after all. His two very distinct groups of 

 an Order are to him two genera ; had he the whole species 

 of the Order he would never have recognised the groups at 

 all, as growps. This is a terrific screed. 



[Darwin repHed on February 28 (M.L. i. 105) and March 11 

 (CD. ii. 102), and Hooker responded] : 



March 14, 1858. 



I quite see in what respects local Floras are much the 

 best suited to your purpose ; or rather, how they would be so, 

 if they were worked out upon the same principle as the general 

 Floras, but the fact that they are not so, and that they are 

 hotbeds of bad big genera, is a very serious objection to the 

 use of them. 



I shall be however most curious to see the results of 

 Bentham's British Flora. He reduces the Rubi to 6 species, 

 I think (and about 11 varieties, I suppose), which gives 

 you a small very variable genus, whilst Babington has 

 28 species or so, besides varieties — so Callitriche, of which 

 Babington has several species but which Bentham reduces 

 to 1 with 2 ? varieties. You must however take care not 

 to get entete with your results. I shall certainly go over 

 the Tasmanian Flora for your sake, and see whether or no I 

 should not have noticed varieties to many small genera, to 

 make their species consistently worked with the big. I am 



