OX THE FLORA OF SHETLAND 53 



to it the capacity to produce two kinds of leaves ; the par- 

 ticular environment merely determines which particular 

 one of the two kinds shall be produced. This capacity for 

 adaptability reaches its climax in Britain in some of our 

 amphibious plants, such as BatracJiium and Potamogeton. 

 In the Hieracia, though present, it is developed to a com- 

 paratively small extent only. 



In connection with the foregoing observations it may now 

 be pointed out that some of the more critical Hieracia, although 

 differing in comparatively slight ways from other known 

 types, are not the result of this last-named kind of variation. 

 Mr. F. J. Hanbury, for example, has cultivated these plants 

 for years, and has now, I believe, some two hundred British 

 forms growing in his garden. He finds, when transplanted to 

 this new environment, so different from their native sur- 

 roundings, that, although now growing under practically 

 identical conditions, they do not tend either to converge 

 towards each other, or to revert to other known types. 

 Hence their variation is of another order, and they (the 

 British endemic forms), only differ from some of the most 

 marked endemic forms of other islands in degree, and not in 

 kind, of variation, which is precisely what might be expected. 

 Whether they be classed as species or varieties is a question 

 chiefly of convenience ; no amount of lumping, however, will 

 make the groups in critical genera correspond to the isolated 

 " species " in genera which are now stationary, and in which 

 the tendency to vary is quiescent. Xo doubt a very distinct 

 endemic species from some distant island cuts a very im- 

 posing figure ; but it may be questioned whether its value 

 does not decrease in a degree corresponding to the increase 

 in its distinctness. It may be here suggested that some 

 oceanic islands, with their strongly-marked endemic forms, 

 have perhaps, to a considerable extent, taught their lesson 

 already; and that more is likely to be learnt from the 

 critical genera (the wild analogues of variable florists' flowers) 

 of one's own country, with their numerous intermediates and 

 almost certainly present endemic forms, than by describing 



those more distinct ones which we cannot study in connecti 



with their nearest relations, partly because they may already 



be too far evolved to have any. and also because they are 



