vi.] CONCLUSION. 147 



rosmarinifolius, coronopifolius, chenopodifolius, lavan- 

 deriaefolius, salicifolius, mesembryanthemoides, digit- 

 alifolius,abietinus,arbutifolius,malvaefolius,erodiifolius, 

 halimifolius, hakeaefolius, resedaefolius, hederaefolius, 

 acerifolius, plantigineus, castaniaefolius, spiraeifolius, 

 bryonisefolius, primulifolius, and many more. These 

 names, however, indicate similarities to over thirty 

 other perfectly distinct families. 



It seems clear, then, that these differences have 

 reference not to any inherent tendency, but to the 

 structure and organisation, the habits and require- 

 ments, of the plant. Of course it may be that the 

 present form has reference not to existing, but to 

 ancient, conditions, which renders the problem all the 

 more difficult. Nor do I at all intend to maintain 

 that every form of leaf is, or ever has been, necessarily 

 that best adapted to the circumstances, but only that 

 they are constantly tending to become so, just as 

 water always tends to find its own level. 



But, however this may be, if my main argument is 

 correct, it opens out a very wide and interesting field 

 of study, for every one of the almost infinite forms of 

 leaves must have some cause and explanation. 



THE END. 



