236 Dr. 0. Masson. On the Wetting of [Apr. 25, 



the influence of a known amount of previous moisture is observed, the 

 procedure being similar to that already described. In Series II (water 

 immersion tests) previous moisture was given to the cotton in roughly 

 predetermined quantity by exposing it for suitable times in moist air 

 at the ordinary temperature. 



To obtain the curve which shows the progress of absorption of 

 moisture during immersion in air saturated at any fixed temperature, 

 it is necessary not only to accurately maintain the bath at that 

 temperature but to guard against any interference with the temperature 

 of the cotton. It is an essential feature of the process that the cotton 

 shall become hotter than the bath, and the velocity of absorption at 

 any moment must be affected by the magnitude of this difference of 

 temperature. It is, therefore, impossible to trust an absorption curve 

 obtained from one long experiment, interrupted from time to time for 

 the purpose of weighing. After each such interruption the cotton 

 resumes operations at a lower temperature than that at which it left 

 off, and moreover it never again gets back exactly on to its original 

 temperature curve. There are other errors also which result from 

 such a method, but they are of less importance. Therefore the plan 

 was adopted of determining each point on the absorption curve by 

 means of a separate experiment, starting always with dry cotton, and 

 immersing it for different times, but keeping all other conditions the 

 same. The temperature curve was taken simultaneously in each test. 



Unfortunately it has not been found possible to so arrange things 

 that each temperature curve follows the exact course of the last, 

 which would be the best proof of the trustworthiness of the absorp- 

 tion curve. The differences, though small, are unmistakable even 

 in the case of Series V, which was the best in several respects. 

 The extent of these discrepancies will be shown in the tables 

 summarising results. The cause of them is not to be found in the 

 drying process, as is proved by the uniformity of dry weight; nor 

 in premature absorption through the corks after weighing, as this 

 was disproved by special tests; nor in incompleteness of saturation 

 of the air, nor in any other circumstance peculiar to the air immer- 

 sion experiments, for similar small variations occurred in the 

 temperature curves got by water immersion. ' It is probable there- 

 fore that the errors arise at the moment the thermometer is with- 

 drawn from its tube at the start of each experiment, and that this 

 exposure to air of variable temperature and moistness, brief though it is, 

 initiates a disturbance that shows itself later in the manner described. 

 It is not easy to devise a remedy that would not introduce troubles of 

 its own. Finally it may be pointed out that the constancy of the 

 bath temperature, and the fact that the cotton is always hotter than the 

 bath, afford security against any error such as might otherwise be caused 

 by deposition of dew. 



