413 



much more than half as much, their moisture contents being 

 now 78 per cent. The canariensis blocks carried only 55 per 

 cent. The blocks were now allowed to dry gradually over a 

 period of about eight weeks and were weighed and measured 

 at intervals. Finally, when they had dried down to less than 

 their original weights when freshly cut out of the beams, they 

 were put in a drying oven and kept at a temperature of a 

 little over 100° C. for seven hours. They were then taken 

 out one by one and rapidly weighed and measured. From this 

 series of measurements the curves shown 'in fig. 3 have been 

 plotted, showing the relation between the moisture contents, 

 as expressed in percentage of the dry weights, and the 

 diameters expressed as percentages of the diameter of the dry 

 block. 



As soon as the blocks were taken out of the water they 

 at once started to dry out and decrease in weight, but, curi- 

 ously enough, continued to still further expand for a day or 

 two, although they were losing moisture. After that the 

 insignis blocks, which had absorbed the greatest quantity of 

 water, remained practically of the same diameter until the 

 moisture contents were reduced to about 50 per cent., as 

 measured on the dry wood, when contraction began to take 

 place. Contraction then took place at an accelerating rate as 

 the wood further dried, and in all cases the greatest amount of 

 contraction for 1 per cent, loss of moisture took place as the 

 wood finally dried down to the 10 or 12 per cent, of moisture 

 that is permanently contained in seasoned timber. This ex- 

 plains why the doors of our houses sometimes stick in the 

 winter. The alteration of the moisture contents of seasoned 

 wood with the humidity of the air only ranges over 2 or 3 

 per cent., but it occurs just at the point where the rate of 

 contraction or expansion is greatest. 



The somewhat remarkable behaviour of the wood under 

 the conditions of the tests seems to be capable of explanation 

 when the fact is taken into consideration that the water in 

 the wood exists partly as free water within the cells and 

 partly as absorbed water in the cell walls. The contraction 

 or expansion of the wood is due to a change in the moisture 

 contents of the cell walls. An alteration of the amount of free 

 water within the cells will of itself produce no effect on the 

 dimensions of the block. The complete saturation of the cell 

 walls evidently takes time and when the insignis blocks were 

 first removed from the water, although the cells were full 

 the walls had not yet absorbed quite as much as they were 

 capable of absorbing. The process of saturation of the walls 

 would then still go on, as long as there was free water within 

 the cells, and the blocks in consequence still expanded. After 



