CHAP. III. FUNCTION OF NUTRITION. 223 



a few days this water will be found strongly im- 

 pregnated by a matter excreted from the roots. A 

 fresh plant should be placed daily in the water, to avoid 

 the effects which might otherwise be produced by an 

 incipient decomposition. It is also found that the 

 matters thus procured from plants of different families 

 are dissimilar. Thus that which is excreted by the 

 Leguminosae contains an abundance of mucilage, whilst 

 that which exudes from the Gramineae has very little. 

 The Chicoraceae excrete a bitter matter analogous to 

 opium; the Euphorbiaceae a gum-resinous matter, &c. 



(218.) Rotation of Crops. So far as observations 

 have hitherto been made, it appears probable that 

 the excretions given out by plants of different fami- 

 lies possess very different qualities, and act differently 

 upon other plants. It had been long known to gar- 

 deners that flowers and fruit-trees will not prosper so 

 well when they have been planted in a situation where 

 others of the same kind had previously grown, as if 

 they were planted in situations where they succeeded 

 to others of a different kind. It is also a well-esta- 

 blished fact in forestry, that when a wood principally 

 composed of one species of timber trees has been 

 cleared, the trees which then spring up spontaneously 

 and supply the place of the former growth are for 

 the most part of a different species. And lastly, 

 the agriculturist has established a rotation of crops 

 upon experimental proof that grain of one kind suc- 

 ceeds better when it follows certain other kinds, than 

 when it is sown immediately after a crop of the 

 same plant. The various theories which had formerly 

 been proposed to account for these facts were all liable 

 to serious objections ; but M. De Candolle has suggested 

 the probability, that the excretions of any one plant 

 although they may be noxious to others of the same 

 species, genus, or family, may nevertheless be per- 

 fectly harmless or even beneficial to plants of other 

 families. In this manner he would account for the 

 fact, that plants of the natural order Leguminosae (as 



