220 BRITISH PLANTS 



and Hooker regard all the forms of the water-buttercup as 

 inconstant varieties of one species, which they name 

 Ranunculus aquatilis ; in Babington's Flora, fifteen of 

 these varieties are raised to specific rank. Bentham and 

 Hooker divide the hawkweeds (Hieracium) into four 

 species. This variable genus has been split up by others 

 into a bewildering multitude of species. In Babington, 

 the number of critical forms described as true species is 

 210, besides numerous sub-species and varieties. Other 

 genera with poljnnorphic species are the bramble {Bubus), 

 Rosa, Euphrasia (eye^bright), and Salix (wUlow). 



What, then, is there to guide us in this matter ? In 

 the first place, we must know exactly what a variation is, 

 and, fortunately, upon this considerable light has been 

 thrown during the last few years. 



Variations are of two kinds : 



1. Inconstant Variations, due to the environment, such 

 as accidents or inequalities in light, soil, climate, or 

 nutrition. These affect, within certain limits, the form, 

 size, hairiness, and duration of the leaves and shoots, the 

 general habit of the plant, and, in some cases, the colours 

 of the flowers. These qualities, however, are not con- 

 stant. They are characters acquired by the individual 

 plant as adaptations to its environment. If the environ- 

 ment changes, these characters change, too. They are 

 not transmitted, but every plant gradually assumes them 

 during its own individual life, providing the conditions 

 favourable for their production are present. Such 

 varieties have no right to rank as critical species ; they 

 are merely forms, not distinct and constant races. Thus 

 the prostrate maritime variety of broom found in Corn- 

 wall {Gytisus vulgaris, var. prostratus) is a mere form, 

 with no claim to specific or even varietal rank, any more 

 than luxuriant garden specimens have. The seeds of this 

 plant, when sown in the garden, grow into the ordinary 

 form of broom. 



2. Constant Variations. — These are not dependent upon 

 accidents of environment, and always breed true to seed. 

 They are true races, and have a right to rank as species, 

 if the species is to be regarded as the unit of classi- 

 fication. 



In practice, however, the difficulty is to decide what 

 variations are constant and what inconstant, and the 



