Early Work in the United States on Wood Properties
Prior to the early part of the 20th century, no design values were published, nor did building codes govern construction. Yet many buildings were built that still stand today. Although tests to determine mechanical properties of timber were conducted as far back as the 1700s, as discussed above, it was not until the late 19th century that research scientists conducted tests specifically designed to provide more reliable data on properties of wood for use in buildings. Generally unknown today, the research of Roth and Fernow (1895), Hatt (1904, 1905), Talbot (1909), Cline and Heim (1912) and others provided the means of designing with timber based on an understanding of material performance.
In 1895, Filibert Roth and B.E. Fernow published Forest Service Bulletin No. 10, which reported on the influence of weight and moisture content on the strength of clear wood specimens and summarized much of the available knowledge on wood behavior to that point. That same year, the committee of the American International Association of Railway Superintendents of Bridges and Buildings presented a report (finally published as Berg et. al., 1907) which also included a summary of available data on wood properties and behavior. These data pointed to 15 recommendations made by the railway committee, some of which foreshadowed modern design and grading rules.
Beginning in the late 1800s and continuing into the early 1900s, a concerted effort was made by U.S. Division of Forestry (under B.E. Fernow) and several universities (including Washington University, Yale, Purdue, and the University of California at Berkeley) to systematically extend this knowledge through well- designed testing programs. However, there was some disagreement about whether the best approach was to test small, clear specimens or larger wood members. Proponents of the first approach thought it important to identify clearly how wood behavior differs among various species, while proponents of the latter approach thought that the behavior of large members with defects such as knots, wane, and slope of grain, would not be adequately predicted from the characteristics of small clear specimens. An excellent summary of this historic controversy is provided by Green and Evans (2001).
In the early 1900s, W. K. Hatt (in charge of the timber testing program at Purdue University), who considered both approaches valid and significant, designed a testing program to address both methods. Forest Service Circular 38 (Hatt, 1905) was an attempt to summarize existing data and standardize programs among the various testing laboratories. At the University of Illinois Engineering Experiment Station, Arthur Talbot was also working on testing of timber beams. Talbot conducted horizontal shear tests on large timbers and small samples cut from the larger timbers to determine correlative values (Talbot, 1909). Talbot also examined the strength-limiting properties of knots, shake, and cross-grain, as well as moisture content, seasoning, and the strength difference between creosote treated and untreated timbers.
After data on both small clear specimens and large timber beams began to accumulate, other investigators looked at more subtle properties of wood. For instance, McGarvey Cline and Harry Tiemann conducted research on the effect of load rates on the strength and stiffness of wood (ASTM Proceeding 1908; Green and Evans, 2001), while Rolf Thelen investigated the testing of green and partially seasoned timbers (ASTM Proceedings 1908). While much testing had been conducted prior to 1910, little had been done to consolidate the findings into any sort of grading criteria based on the research results.
In 1912, Cline and Heim produced Forest Service Bulletin 108, which was a complete summary of the full-sized testing program, and provided an important reference for subsequent grading rules because it identified the strength of beams tested in bending and compression, both parallel and perpendicular to grain, as a function of their characteristics (moisture content, splits, knots, etc.).
There were two distinct mindsets developing in conjunction with these early investigations into the properties of wood. Many researchers felt that timber testing programs and any subsequent grading standards should develop data necessary for engineers and architects to reduce both the waste involved in overbuilt structures and to ensure minimum standards for structural integrity and safety. Timber tests, the proponents of this perspective argued, should therefore be conducted on full size members available to consumers (Green and Evans, 2001).
Many foresters and others, however, felt that timber tests should focus on the qualities of trees rather than the potential design uses for lumber. Timber tests, according to their perspective, should serve to provide data on the average strength properties of different species and demonstrate how growth conditions and other factors can affect the averages. Proponents of this perspective believed that multiple small sample tests should be conducted to provide strength averages without incurring the expense and waste that a full size test program would generate (Green and Evans, 2001).
These two perspectives led to considerable conflict as organizations began to develop grading standards for the lumber industry. Additionally, the discrepancy between test results for small clear samples and full size members was problematic. In order to correlate the data from both types of tests, researchers began to focus on so-called “defects” such as knots, splits, checks, wane, shake, and cross grain that seemed to limit the stress properties of full size members. This focus, along with the development of strength ratios based on results of both tests, provided the historical data that led to modern grading standards based on visual characteristics and mechanical tests.