22 ROYAL SOCIETY OF CANADA 



raiglit resuit iii fracture at a point lower even than "would have been the 

 case had the timber been allowed to remain in its normal state. 



There are many cases in practice where the timber is alternatel}' Avet 

 and dry. These are cases in which the moisture is greater than in the 

 normal condition, and tests at these stages, as well as in the normal and 

 completely submerged conditions, become therefore of great practical 

 value. 



The load on a kiln-dried specimen, so long as it does not exceed, say 

 about 60 per cent of the ultimate strength, produces a compression directly 

 proportional to the load, and the element of time seems to be no longer an 

 appreciable factor. The set, too, when the specimen is relieved of load, 

 is very small. The coefficient of elasticity can therefore be readily 

 obtained. As the load increases beyond the 60 per cent, the rate of increase 

 of the compression is much less in proportion to the rate of increase of 

 the load, and the resistance becomes similar to that oftered by a hard 

 material like stone. When the maximum load is reached, the specimen 

 fails instantaneously, by breaking np into longitudinal splinters, in some 

 cases as small as match wood. 



If the specimen is in the normal state, the amount of the compression 

 is found to vary directly with the amount of moisture present in the 

 atmosphere, and the average coefficient of elasticity can be easily found. 

 When the moisture in the specimen exceeds that corresponding to the 

 normal state, it is ditficult if not impossible, to make even an approxi- 

 mate determination of the coefficient of elasticity. Time now becomes 

 an extremely important factor. The amount of the compression (and 

 the set) increases with, but much more rapidly than, the load, and hence 

 it becomes of the utmost consequence, as the ultimate strength of the 

 si^ecimen is approached, to carefully note the point at which failure 

 actually occurs, as it is quite po.ssible through inexperience, inattention, 

 or by design, to obtain a result which is inaccurate to the extent of 8 or 

 10 per cent, or even more. 



The character of the failure of a wet sjiecimen depends upon its length, 

 its diameter, and upon the proportion of hard and soft fibre. Should 

 there exi.st any great difference in the density of the annual rings, failure 

 will first be ap])arent on that side on which the density is least. In long 

 specimens crippling is accompanied by a bending over towards the side 

 of least resistance. In short specimens, failure usually takes place by the 

 folding or crippling of the layers at or near both ends. 



Transverse Tests. 



Transverse tests upon kiln-dried timbers are of importance as stand- 

 ards of reference, but as this condition cannot be maintained, and the 

 timber gradually tends to pass into the normal state, transverse tests 

 taken between the two conditions are of little, if any, practical value. 



