November Dead Leaves. 13 



To retrace the path we came by might have been pos- 

 sible, though difficult; but I felt an invincible repug- 

 nance to a mere retreat. So partly in reliance upon 

 chance, and partly trusting an instinct of locality that 

 I cannot account for or explain (an animal instinct, like 

 that of the salmon or the housemarten), I determined to 

 push on through the dense wood till the topography of 

 the country became somewhat more intelligible to us. 



Before quitting the great chestnuts, I made an obser- 

 vation which confirmed what I had observed before with 

 reference to the adherence of dead leaves. These trees, 

 as a rule, were entirely denuded of their foliage ; but 

 two or three branches, on the contrary, retained almost 

 every leaf that had adorned them in the glory of summer 

 changed indeed, in color, from rich dark green to a 

 lovely pale gold, far more delicate than the winter color- 

 ing of beech or oak, yet scarcely altered in form, and pre- 

 serving great purity of curve. Now the question which 

 interested me was, how it happened that these branches 

 retained their foliage whilst all the others had lost it ? 

 The answer is, that a branch which retains its foliage has 

 always been virtually severed from the stem by fract- 

 ure before the fall of the leaf^ Why the leaves fall from 

 a branch that shares the life of the tree, and adhere to 

 one that is separated from it, I am not scientific enough 

 to decide quite positively, but naturally conclude that it is 

 due to the continuance of circulation in the one case and 

 its stoppage in the other, the leaves adhering when the 

 sap has not been able to descend, but detaching them- 

 selves easily when the course of the descending sap has 



