211 
the Direction of the Growth of Roots . 
confined themselves to the trenches, in which the soil of the 
former surface must have been buried. 
A trench which was twenty feet long, six wide, and about 
two deep, was prepared in my garden, in the bottom of which 
trench was placed a layer, about six inches deep, of very rich 
mould, incorporated with much fresh vegetable matter. This 
was covered, eighteen inches deep, with light and poor loam, 
and upon the bed thus formed, seeds of the common carrot 
( Daucus carota ) and parsnip ( Pastinaca sativa ) were sowed. 
The plants grew feebly till near the end of the summer, when 
they assumed a very luxuriant growth, grew rapidly till late 
in the autumn, and till their leaves were injured by frost. 
The roots were then examined, and were found of an extra- 
ordinary length, and in form almost perfectly cylindrical, 
. having scarcely emitted any lateral fibrous roots into the poor 
soil, whilst the rich mould beneath was filled with them. 
In another experiment of the same season, the preceding 
process was reversed, the rich soil being placed upon the sur- 
face, and the poor beneath. The plants here grew very luxu- 
riantly, and acquired a considerable size early in the summer; 
and when the roots were taken up in the autumn, they 
were found to have assumed very different forms. The 
greater part had divided into two or more unequal rami- 
fications, very near the surface of the ground, and those 
which were not thus divided tapered rapidly to a point at the 
surface of the poor soil, into which few of their fibrous roots 
had entered. 
In other experiments seeds of almost all the common escu- 
lent plants of a garden were so placed that the young plants 
had an opportunity of selecting either rich, or poor soil ; 
E e 2 
