8 



is that material even of the same species was found to give strength 

 factors of such a wide range of values that scientific investigation was 

 discouraged. The elements affecting the strength were not known, 

 and there was no assurance that the piece to be used would develop a 

 strength within narrow limits comparable with that shown by the test 

 piece. To-day the elements affecting the strength of wood are much 

 better known. While we still obtain the same wide range of values we 

 are able to assign causes therefor. 



It is safe to say that the time is coming when, by occular inspection, 

 an expert will be able to predict the strength of a piece of wood more 

 accurately than can be done by the same method with iron or steel. 

 Rules of inspection may then be formulated which, if carefully followed, 

 will yield timber of comparatively uniform strength. But even with 

 the knowledge we now have it is possible to improve greatly upon the 

 present practice. 



With this object in view a queiy sheet was sent to about thirty of 

 the principal railroad companies of the United States for the purpose 

 of determining what the prevailing practice now is. Answers were 

 received from twenty of these companies, and serviceable information 

 from fifteen, ten of which represented 500 miles of trestle. The other 

 five did not report mileage. This information has been compiled in 

 Tables I and II. 



DATA FOR TRESTLE CONSTRUCTION. 

 PRESENT PRACTICE. 



Table I gives the different species now employed in the various parts 

 of these structures and a mean estimate of the length of life of each. 

 These separate estimates, however, were very erratic, in many cases 

 being little better than a guess, so that the mean given in the table is 

 by no means reliable. 



This lack of information is scarcely less remarkable than it is unfor- 

 tunate. Although for more than fifty years railroad companies have 

 been using timber, no accurate, classified knowledge exists as to its 

 length of life; yet this could be easily obtained if each member of a 

 trestle were given a number, as is done in iron structures. The length 

 of life of timber is, of course, not an exact quantity, being a function 

 not only of the various conditions of use, but also those of growth and 

 treatment previous to use. For a given locality and treatment the 

 length of life of originally sound timber of a given species should not 

 exhibit such a remarkable difference in durability as indicated on the 

 separate returns from the railroad companies. 



From an examination of the mean value given at the bottom of Table 

 I we see that, in general, the piles and posts outlast the stringers and 

 caps by from one to two years. This is reasonable and undoubtedly 

 correct in kind; probably an underestimate in degree. 



