[bovey] presidential ADDRESS 19 



In order to eliminate these errors, I substituted for these rods, wooden 

 rods with light steel plates at the ends. The wooden rods were made 

 impei'vious to moisture by boiling them in parafine wax until they were 

 so completely saturated that they sank to the bottom. They Avere then 

 allowed to cool and afterwards properly prepared for the extensometer 

 work. It was found that with these rods the error due to a variation of 

 1° of temperature did not exceed l/300.000th of an inch, whereas with 

 the ordinary steel rods the error was tifteen to eighteen times as great. 

 The same extensometer was used both in the tension and in the compres- 

 sion experiments, the scale in the latter case being of course inverted. 



For the purpose of obtaining an accurate comparison, the greatest 

 care was taken that the specimens in the original timber were in the 

 same continuous line, and that, consequently, they had the same propor- 

 tion of hard and soft fibre and were alike in clearness and straightness of 

 grain. 



The experiments may be divided into four classes, namely, (1) direct 

 tensile ; (2) direct shear ; (3) direct compressive ; (4) transverse. The 

 conditions of the timber ranged from a frozen and from a saturated 

 condition to a kiln-dried condition, in which the whole of the moisture 

 had been eliminated. The observations as to the effect of freezing and 

 excessive moisture upon the strength of the timber were made on hemlock 

 and spruce specimens only. I shall designate as the normal state that 

 condition in which the specimen contains the greatest amount of moisture 

 consistent with the hygrométrie state of the surrounding atmosphere. 



The wet specimens used in the present experiments were cut from logs 

 taken from the river, in which they had been submerged for several months, 

 and contained their natural moistxire in addition to the water absorbed 

 whilst in the river. When these spruce and hemlock logs were cut through , 

 they were found to be completely saturated and also frozen. Some of the 

 comparative results were obtained from specimens cut from the logs after 

 they had been transversely tested. While still saturated, they were 

 worked into suitable forms with as little delay as j)ossible and as each 

 specimen was made it was buried in snow until required for the test. 

 Many of these specimens were tested when in a frozen condition, others 

 when saturated, but not frozen, and again some were kiln-dried at 212° 

 Fahren., until every trace of moisture had been removed. This point was 

 considered to be reached when repeated weighings showed no appre- 

 ciable difference. They were then taken out of the kiln and allowed to cool, 

 a necessary precaution in order to eliminate any ]iossible effects of heat 

 on the specimen and extensometer. No sensible change in weight could 

 be detected during the cooling process and the specimen remained 

 apparently devoid of moisture at least long enough to allow of one 

 complete test. As most of these tests were cari-ied on during the winter, 

 at a time when the atmosphere contains a minimum amount of moisture, 



