142 VEGETABLE ORGANOGRAPHY. 



and is usually in proportion to the angle which the 

 branches form with the trunk. 



When the lower branches are not able to extend, as 

 happens in forests, or in certain trees by the natural 

 effect of their growth, they perish by degrees, and thus 

 arises the denudation of the trunk of trees. 



Generally, the lower branches of large trees are 

 parallel with the ground ; and this is true, not only when 

 they grow on horizontal ground, which explains itself, 

 but also when they are placed upon a hill ; in this case 

 the lower branches of the cyme remain parallel with the 

 surface. This parallelism with the ground is also found 

 in individuals of which the trunk itself is oblique to the 

 horizon. Dodart, who first insisted upon this popular 

 observation (Acad, Science, 1699, p. 60), remarks that 

 the roots almost always spread out parallel to the surface ; 

 it results then that the plane of the branches is parallel 

 to that of the roots : in order to explain this, he sup- 

 poses that the fibres of plants have a determined length, 

 and that being continuous from the extremity of the 

 root to that of the branch, it is necessary, to preserve 

 the same length, that they should form angles. But we 

 cannot maintain the principle of the fixed length of the 

 fibres, for it is sufficient to expose a branch to favour- 

 able circumstances, to make it grow indefinitely. This 

 parallelism is explained, on the contrary, very easily, by 

 the proportion which all the branches and roots generally 

 preserve during their growth. 



It is a pretty constant observation, that a large branch 

 corresponds to a large root, and vice versa ; and this is 

 equally true, whether a root, placed in favourable cir- 

 cumstances, causes the increase of the branch which is 

 above it, or the branch, similarly situated, makes the 

 root which corresponds to it to be developed. But in 

 trees which grow upon hills, the two sides of the root 



