CAUSES OF VARIATION IN DETERMINATIONS. 55 



CAUSES FOR DETERMINATIONS HAVING A VALUE GREATER THAN 



NORMAL. 



The errors under this heading are far more serious than those under 

 the heading just discussed and are more difficult to' eliminate entirely. 

 One source of error is in assuming that a temporarily wilted condition 

 is the permanently wilted condition. A temporarily wilted condition 

 generally results from a rapid increase in the saturation deficit com- 

 bined with an increase in temperature and light intensity, which 

 result in greatly increasing the transpiration of the plant. Many 

 plants, such as tobacco and corn, normally show temporary wilting 

 during the hot part of the day, and the same is observed in many 

 different plants if the conditions are sufficiently severe. This source 

 of error may be avoided by placing the wilted plants under a bell jar 

 in a comparatively saturated atmosphere. If they fail to recover, 

 the wilted condition may be regarded as permanent. If, however, 

 they soon resume their normal condition the wilting has been merely 

 temporary. Such temporary wilting has not been encountered very 

 generally in our experiments. Plants grown under ordinary green- 

 house conditions seldom show temporary wilting, and by a little care 

 in maintaining equable conditions the difficulty can be entirely 

 avoided. As we have shown above, the degree of humidity or of light 

 intensity does not appear to affect the wilting coefficient so long as 

 conditions are uniform, but a sudden increase in temperature, fight 

 intensity, and saturation deficit often brings about a sudden temporary 

 wilting which may lead to error. 



Occasionally a pathological condition is encountered, the plant 

 roots or even the top being affected by some parasite, causing the 

 plant to wilt as if the water supply were inadequate. Such abnormal 

 conditions can be detected easily by an examination of the plants at 

 the time the sample is taken. If it is evident that the plant is in a 

 diseased condition moisture samples should not be taken. 



The third and greatest cause of uncertainty in the determination of 

 the wilting coefficient is found in imperfect root distribution. The 

 results of the experiments herein recorded indicate that to a very 

 great extent the variations in the wilting coefficient for different 

 plants are dependent upon the degree to which the roots of the plants 

 penetrate every portion of the soil mass. An ideal determination of 

 the wilting coefficient would be one in which every portion of the soil 

 was brought into contact with the plant roots. This ideal condition 

 is never reached. Some plants, particularly the grasses, when planted 

 in a pot approach this condition more nearly than any others tried, 

 and with these forms a lower wilting-coefficient determination was 

 obtained. Other plants, especially those having large, coarse roots 

 and which develop few rootlets, give a correspondingly high wilting 

 coefficient. 



230 



