4 THE PRINCIPAL SPECIES OF JVOOD. 



in that while the conclusions were correct as applied to the 

 specimens immediately studied, such specimens did not stand 

 for the species at large. The recognition of difficulties, the 

 selection of specimens, the scientific standardizing of methods 

 so that results could be generally utilized, as distinct from the 

 simple manipulation of specimens in testing-machines, have 

 not been exhaustively attempted until recently.* 



The experiments that have been made to determine the 

 strength of woods may be grouped into the four following 

 divisions : 



(i) E.xperiments conducted by the U. S. Division of Forestry 

 (Dr. B. E. Fernow, Chief), under the direction of Professor J. B. 

 Johnson. About forty thousand tests were made, distributed over 

 thirty-one American species, the results, so far as obtained, being 

 undoubtedly the most valuable in existence. The detail considered 

 and methods evolved have in a way reclassed the testing of woods 

 and must influence all future efforts, but results are disappointing in 

 that they have been obtained for so few species, and some of these 

 of commercially secondary importance. These experiments are 

 characterized as follows : 



Completeness and Reliability of Records. 



Large and Small Test Pieces. 



IMoisture Conditions Standardized at 1 2^ Dry Weight. 



Samples from Representative Portions of Tree. 



Selection of Representative Trees. 



Uniformity of Methods. 



Large Number of Individual and Total Tests. 



Small Number of Species Covered. 



Specific Gravity Determinations. 



Soil and Forest Conditions Indicated. 



Botanical Accuracy Assured. 



These experiments are originally described in Circular No. 15 

 and other publications of the U. S. Forestry Division, also in 

 " Materials of Construction," by Professor J. B. Johnson. 



(2) Experiments conducted for the Tenth U. S. Census by 

 Mr. J. P. Sharpless at the Watertown (Mass.) Arsenal upon speci- 

 mens botanically selected by Professor Sargent. These experiments 

 are less complete in detail, and averages are based upon infinitely 



* It should be noted that the selection and preparation of specimens require 

 the exercise of more judgment than the simple testing of specimens, if the con« 

 elusions are to be such that they can be generalized from. 



