Drying oj Timber, 81 



less the time should be decreased as the square of the fractional part 

 of an inch." Again, he says, for thickness above an inch, " The time 

 ordinate should be increased in proportion to the thickness up to 

 three inches, and about one and a-half times the thickness for thick- 

 nesses over three inches." 



Humidity and Temperature. — The fifth factor with which drying 

 is concerned is that of humidity, and with this we may conveniently 

 consider the sixth factor, temperature. Humidity is not a satisfac- 

 tory variable to use, since it is only a relative quantity. When we 

 speak of a humidity we must always specify a temperature. The 

 same humidity at different temperatures gives very different drying 

 conditions. Drying depends on the difference between the vapor 

 pressure of the air surrounding the wood and the vapor pressure at 

 saturation. At a temperature of 50 °C. and a relative humidity of 90 

 per cent., there is a difference in pressure between that actually exist- 

 ing and that at saturation point of 9.25 mm. of mercury. At a tem- 

 perature of 20°C. and the same humidity, namely, 90 per cent., there 

 is only a difference of 1"75 mm. Hence there is a very great differ- 

 ence in the rate of drying of timber, when placed under these twa 

 sets of conditions. A high humidity at a high temperature may cause 

 timber to dry at a much faster rate than a low humidity at a low 

 temperature. Thus blocks drying at 80 per cent, humidity at 50°C.^ 

 were losing moisture faster than those at 10 per cent, humidity at 

 20 °C. This point is frequently lost sight of in kiln drying, when it 

 is recommended to commence the seasoning at a high temperature 

 and a high humidity. 



Experiments were carried out at different humidities, at the tem- 

 peratures 20°, 30°, 40° and 50°C. For the experimental work blocks 

 approximately 2'' x 2"' x 1" were used. For any one experiment the 

 blocks were cut from the same length of timber. Only straight 

 grained, freshly felled material was used. The sides of the specimens 

 were sealed, leaving the 2" x 2'' faces exposed. The exposed faces 

 were always radial surfaces. To obtain the various humidities, solu- 

 tions of sulphuric acid were used. After each weighing, the solution 

 in each desiccator was brought up to strength by adding the amount 

 of acid corresponding to the loss of weight of the specimen. For des- 

 iccators, tall gas cylinders were used, and the specimens were sup- 

 ported on glass rods. The humidities used ranged from 95 per cent, 

 down to 2 per cent. There were two humidities below 10 per cent., 

 namely 7 per cent, and 2 per cent. 



In the experiment at 20°C. losses increased with decreasing^ 

 humidity. The greatest loss occurred at the lowest humidity, that is 

 at 2 per cent. At the other temperatures, however, the maximum loss 

 occurred at 7 per cent., and the curve of loss bent downwards from 

 the 7 per cent, to the 2 per cent, humidity. This is shown in Fig. 8, 

 which gives a typical series of losses. Similar graphs were obtained 

 with beech, elm, and chestnut. A duplicate experiment with oak at 

 50°C. gave precisely the same type of graph. Owing to limitations 

 of space and material, only single specimens could be used at each 

 humidity, but the greatest care was taken in the selection and cutting 



7 



