364 



FORESTRY INVESTIGATIONS IT. S. DEPARTMENT OF AGRICULTURE. 



The, great variation in strength which is noticeable in timber of the same species makes it necessary to accept 

 with caution the result of a limited number of tests as representing the average for the species, for it may have 

 happened that only all superior or all inferior material has been used in the tests. Heuce we would not be entitled 

 to conclude, tor instance, that pignut hickory is 14 per cent stronger than shagbark, as it would appear in the table, 

 for the 30 test pieces of the former may easily have been superior material. Only a detailed examination of the test 

 pieces or a fuller series of tests would enlighten us as to the comparative value of the results. 



The following datii, therefore, are not to be considered as in any sense final values for the species, except where 

 the number of trees and tests is very large : 



Results of teits in compression endwise. 

 [Pounds per square inch.] 



a Actual tests on "dry" material not reduced for moisture. 



The variation in strength in wood of the virgin forest, as will be seen from the tables, is in some species so 

 great that by proper inspection and selection values differing by 25 to 50 per cent may be obtained from different 

 parts of the same tree, and values differing 100 to 200 per cent within the same species. These differences have all 

 their definite recognizable causes, to find and formulate which is the final aim of these investigations. 



The tests are intentionally not made on selected material (except to discard absolutely defective pieces), but on 

 material as it comes from the trees, so as to arrive at an average statement for the species, when a sufficient number 

 of trees has been tested. How urgent is the need for data of inspection as above indicated will appear from the 

 wide range of results recorded. 



To enable any engineer to use the data here given with due caution and judgment, not only the ranges of values 

 and the average of all values obtained, but also the proportion of tests which came near the average values, have 

 been stilted, as well as the average results of the highest and lowest values of 10 per cent of the tests. With this 

 information and a statement of the actual number of tests involved, the comparative merit of the stated values can 

 be judged. With a large number of tests, to bo sure, it is more likely that an average value of the species has been 

 found. The actual test results have been rounded off to even hundreds in the tables. 



FACTORS OK SAFETY. 



With such lowest standard values, also lowest factors of safety could be employed. As to factors of safety, it 

 may be proper to state that the final aims of the present investigations may be summed up in one proposition, 

 namely, to establish rational factors of safety. It will be admitted by all engineers that the factors of safety as used 

 at present can hardly be claimed to be more than guesswork. There is not an engineer who could give account as 

 In the luisis upon which numerically the factors of safety for wood have been established as "8 for steady stress; 

 10 for varying stress; 15 for shocks" (see Morrimau's Testbook on the Mechanics of Materials); or as I to "> fur 

 "dead" load and 5 to 10 for "live" load (see Rankiue's Handbook of Civil Engineering). 



