vii Relations Between Plants and Animals 125 



so to speak, "in the blood." Their occurrence is wide- 

 spread, and in some cases their primary meaning in the 

 internal economy of the plant is well known. 



The idea that tannin has been developed as a protection 

 against snails and other animal enemies is no doubt at 

 first sight attractive, while we are ignorant of its real 

 nature, and also while we ignore the fact that a plant rich 

 in tannin, like the oak, may be peculiarly a prey to animal 

 enemies, or meet it by inventing a fresh hypothesis of the 

 adaptation of some animal enemies to withstand the 

 defence. But when it, like the protective crystals, becomes 

 viewed as essentially a waste product, and as therefore 

 necessarily turned out in definite chemical proportion by 

 the life-processes of the plant, the natural selectionist argu- 

 ment recedes as far into the background as it would have 

 to do in explaining the origin of the crystalline form or 

 taste of any chemical product, the colour of any precipitate, 

 the lustre or specific gravity of a mineral. 



We laugh at those who said, " So are fleas black that 

 they may be caught more readily upon a white ground," 

 but are we becoming wiser now, if it be true, as Professor 

 Stahl thinks, and we fear only with too much justice, that 

 the majority of modern naturalists would corroborate the 

 opinion that the protective characters of plants stand in 

 direct causal connection with the appetite of animals ? To 

 give snails credit for evolving plants with crystals, sourness, 

 and poison, to make cattle and the like responsible for the 

 thorns on plants, is like giving snakes the credit of evolving 

 boots which protect our heels. In all these cases alike the 

 possibility of some defensive utility is undenied, nor even 

 of some improvement through selective agency; what is 

 contended for is, however, a change in our evolutionary 

 perspective, laying increased importance upon the definite- 

 ness and cumulativeness of the internal variation and con- 



