104 
UPWARD GROWTH OF THE HEAD, AND [Part II. 
overgrown their pots, the roots may be seen to 
grow straight up the side of the ball of earth, 
and in all directions around it, owing to the 
mechanical confinement of the pot. But I do 
not think that the roots ever return through 
the ball of earth towards the plant. The want 
of light may probably be considered a sufficient 
reason for branches never returning towards the 
stem. I think it more difficult to find a reason 
for this perpetually centrifugal (or, if nova rerum 
nomina be allowed, ipsifugal ) determination of 
the root. 
strong down- In all the cases in all these experiments the 
ward determi¬ 
nation of tap- plants grew permanently, and independently of 
bie distinct or- their seeds. In all the cases where the gem- 
ganisation of it. ° 
mules or stems came down into the air, they 
grew across the wires to the light at the win¬ 
dow. None came from under the flower-pot on 
the sides towards the room, even though they 
emerged from the earth close to those sides. 
And in all cases, whether the gemmule grew 
upward through the earth, or downward into 
the air, the first or tap-root showed itself from 
above below, and never turned upward. But 
the branch-roots fixed themselves, and grew 
from below upward ; and I therefore imagine 
that there must be something perfectly distinct 
