1916] ROBERTS—EPIDERMAL CELLS OF ROOTS 503 
zation of swelling takes place, which immediately follows the 
initial swelling of wall z. | 
The position of the nucleus can have nothing to do with the 
initial swelling, as the swelling is over the entire length of the 
wall r. Since the weaker places bear no relation to the position 
of the nucleus, the nuclear position can in no way affect the forma- 
tion of the root hair. 
In this discussion the second type given by LEAVITT is not con- 
sidered, the type in which only specialized epidermal cells form 
root hairs. Preliminary observations indicate that the osmotic 
pressure of the short cells varies from that of the other cells, but 
there are doubtless other factors determining the hair formation. 
Investigations so far indicate that there are two factors of 
importance in the initial formation of root hairs. One is the 
unequal pressure acting upon either side of wall 7; the other is 
the variation in the physical character of the wall. 
The difference in pressures on the two sides of walls 2, 4, 6 is so 
much less than the difference on the two sides of wall z that it is 
negligible. The osmotic pressure of the root hairs of the plants 
investigated when grown in moist air shows slight variation. In 
the-plants examined the osmotic pressure approximates 5 atmos- 
pheres. The walls 2, 4, 6 have an opposing pressure equal to or 
greater than 5 atmospheres, but wall z must sustain a pressure of 
4 atmospheres, for on this wall the internal pressure is opposed by 
only one atmosphere, when the root is grown in moist air. This 
pressure is sufficient to account for the initial swelling of wall r. 
The result of increasing the osmotic pressure on the outside of 
wall z by growing radish seedlings in sucrose solutions of increasing 
osmotic value proved that the opposing pressures on either side of 
wall z still maintained a balance of at least 4 atmospheres in favor 
of an outward pressure. 
Root hair formation is retarded when the moisture content of 
the air is decreased. This is shown by the experiments with seed- 
lings grown in Petri dishes in alternating dry and moist air, and 
in those grown over sulphuric acid solutions. Reduced moisture 
affects the membranes by decreasing the extensibility of the walls 
due to an increase in the elasticity of the colloids. The osmotic 
