Chap. III.] DOWNWARD GROWTH OF THE ROOTS. 
95 
flower-pots full of earth, turned them over on 
wire-work, and hung the inverted pots from the 
wood-work at the upper part of my window, so 
as to have the lower sides of the seeds exposed 
to light and air from below, and their upper 
sides in contact with the moist earth above them. 
The immediate results showed a most remarkable 
determination of the first or tap-roots down¬ 
ward, and of the gemmules or stems upward. 
In all the experiments the first or tap-roots of 
all the seeds, without a single exception, came 
straight down into the air, and ceased to grow 
when the ends in the air were from a quarter of 
an inch to an inch long. None ever turned up 
again. The plants, however, threw out branch- 
roots from their necks, fixed them upward in 
the earth, and continued to grow. Those parts 
of the roots which remained alive exposed to air 
and light for six weeks and upwards turned 
green, as did the cotyledons themselves; that is> 
the two divisions of the seed. 
At the same time, I took some horse-chestnut 
seeds, whose first or tap-roots had already begun 
to grow, and placed them so that these tap-roots 
pointed upward into the earth in the inverted 
flower-pot, and the seeds touched the wire below. 
In all the cases the roots immediately turned 
