ON GKNEl'.A AND SPECIES. 71 



law of nature to mark wliere a genus begins or where it 

 ends. On this point, in uiy "Arrangement of the Genera 

 of Ferns," published in 18-11, I have said — "A gradual 

 transition of form is seen to jjer-\'adc through the whole, 

 not in a lineal, but in a complex reticulated manner, it 

 becomes very difficult to determine in which gemis the 

 extreme or transition species of gi-oups should bo placed." 



On considering that in many cases habit is not excluded 

 in assisting- to define the genera of flowering plants, and 

 in numerous instances it is allowed bj' eminent botanists 

 to constitute the chief distinction between allied genera. 

 For example, the order Liliacew is represented by nearly 

 1,700 species, arranged under about 150 genera, which all 

 agree as regards their parts of fructification, the slight 

 variations of which ai'O made use of for defining genera ; 

 and, indeed, if it was not for the difference in habit, the 

 number of genera are capable of being greatly reduced. 

 This and other instances that might bo quoted favour my 

 view, that in oixler to constitute natural g'-enera of Ferns, 

 habit must play an important part ; and so long as any 

 number of species ag-ree in habit, and possessing other 

 characters in common, it matters not whether such groups 

 are called genei-a, sub-genei'a, or sections of genera. For 

 my own piart, T prefer regarding- them as genera, as it 

 saves a great deal of unnecessary^ trouble in speaking- or 

 writing about Ferns, it being much easier to say Drynaria 

 r-oronans than Pohjpodium (^Drynaria) e.oronans, or Elaplw- 

 fjlossum conforme tha.n AcrosticJiniu. {Jilaplioglossuni) con- 

 forme, or Qyinnogramma caloiiuslanos than Oymnogramma 

 (I'Jariymnogramina) calonielaiws ; also by studj'ing the cha- 

 racter of the smaller groups individually, and treating 

 them as genera, their nature at once becomes fixed upon 

 the mind, without having to think of them being- asso- 



