14 WILTING COEFFICIENT FOE DIFFERENT PLANTS. 



to be most satisfactory for use at ordinary temperatures, the exact 

 proportions being unimportant. This mixture melts at so low a tem- 

 perature and has such a low heat conductivity that it can be poured 

 into a pot around the most delicate seedlings without injury. This 

 feature is important in experimenting with delicate dicotyledons. On 

 cooling, this wax adheres well to the glass and to the soil, forming a 

 perfect seal. Other mixtures, notably those containing beeswax, 

 show a decided tendency to separate from the glass on cooling, neces- 

 sitating the resealing of the edges with a hot iron. 



Xone of the substances mentioned give good results when used 

 alone. Soft paraffin (45°) stretches, petrolatum creeps, and beeswax, 

 tallow, and the higher paraffins crack. Except when soft paraffin is 

 used, the plants show no difficulty in penetrating the wax cover, even 

 when a wax as hard as that with which the ordinary commercial 

 phonograph records are made is employed as a seal. 



During the winter in the greenhouse the paraffin-petrolatum mix- 

 ture gives excellent results even when left in direct sunlight. During 

 the warm portion of the year, however, direct sunlight is likely to 

 melt this wax and in this way break the seal and cause damage to the 

 plants, due to the wax creeping over the plant surface. This creeping 

 does not occur if the pots are kept in the temperature-control tank. 

 A mixture of 10 to 30 per cent of beef tallow with beeswax or of 8 to 

 12 per cent of petrolatum with beeswax has been found to be an 

 excellent material for use in warm weather both in the greenhouse 

 and for out-of-door work. 



Modeling clay has also been used to seal the pots, but it is not so 

 easily applied as the wax and is not suitable for use with delicate 

 seedlings. 



EXPERIMENTAL ERROR. 



In determining the wilting coefficient of a given soil for a particular 

 plant some variation will be found in the results obtained from the 

 individual pots. This variation appears to be due in part to the lack 

 of uniformity of the soil in the different pots, but mainly to the fact 

 that in some pots the roots are distributed through the soil mass much 

 more uniformly than in others. When the root distribution is defec- 

 tive, the mean distance through which the soil moisture must move 

 through capillary action is greater. Since capillary movement be- 

 comes very slow as the moisture content approaches the wilting 

 coefficient, these portions of the soil not penetrated by roots will have 

 a somewhat higher moisture content. The wilting coefficient deter- 

 mination in the case of an imperfect root distribution will conse- 

 quently be somewhat too high. On the other hand, errors arising 

 from the distillation of water on the walls of the pot will result in 

 giving a wilting coefficient below the true value. 



Some uncertainty also arises in connection with the determination 

 of the wilting point. Plants often wilt during the day and recover 



230 



