io6 Mr. Knight on the Direction of 
point perpendicularly downwards. To enable myself to 
answer this objection, I made many experiments on seeds of 
the horse chesnut, dnd of the bean, in the box I have already 
described ; and as the seeds there were suspended out of the 
earth, I could regularly watch the progress of every effort 
made by the radicle and germen to change their positions. 
The extremity of the radicle of the bean, when made to point 
perpendicularly upwards, generally formed a considerable 
curvature within three or four hours, when the weather was 
warm. The germen was more sluggish ; but it rarely or 
never failed to change its direction in the course of twenty- 
four hours ; and all my efforts to make it grow downwards, 
by slightly changing its direction, were invariably abortive. 
Another, and apparently a more weighty, objection to the 
preceding hypothesis, (if applied to the subsequent growth 
and forms of trees,) arises from the facts that few of their 
branches rise perpendicularly upwards, and that their roots 
always spread horizontally ; but this objection I think may 
be readily answered. 
The luxuriant shoots of trees, which abound in sap, in 
whatever direction they are first protruded, almost uniformly 
turn upwards, and endeavour to acquire a perpendicular di- 
rection ; and to this their points will immediately return, if 
they are bent downwards during any period of their growth ; 
their curvature upwards being occasioned by an increased 
extension of the fibres and vessels of their under sides, as in 
the elongated germens of seeds. The more feeble and slender 
shoots of the same trees will, on the contrary, grow in almost 
every direction, probably because their fibres, being more 
dry, and their vessels less amply supplied with sap, they are 
