Theses and Dissertations

Issuing Body

Mississippi State University


Barnes, H. Michael

Committee Member

Seale, R. Dan

Committee Member

Shmulsky, Rubin

Committee Member

França, Frederico J.

Committee Member

Hopper, George M.

Date of Degree


Document Type

Dissertation - Open Access


Forest Resources

Degree Name

Doctor of Philosophy


College of Forest Resources


Department of Sustainable Bioproducts


This research approach focused on 2x8 and 2x10 lumber sizes, in lengths ranging from 12- to 16t. Lumber sample data collection included: growth rings per inch (RPI), specific gravity (SG) and presence of pith. The lumber was tested in static bending per ASTM D 198. Small clear samples were then cut and used for compression testing using ASTM D 143. Lumber for this project was collected from retail lumber stores across the Southeastern United States. The larger 'host' study focused on generating current material property data from over 2000 southern pine boards to help establish more accurate and up to date lumber standards. The sampling, dubbed the 'production-weighted sampling' procedure, utilized in this study mimics the requirements of ASTM 1990. Moreover, the findings from this study are beneficial for industry and research communities alike as several smaller projects, including this one, highlight advances ranging from new lumber design values to non-destructive testing technique innovations evaluating wood. All data were analyzed using IBM SPSS version 25 software. Bivariate correlations were run in conjunction with linear regressions. Compression parallel to grain (C?) and compression perpendicular to grain (C?) were statistically tested with RPI, SG and the presence/absence of pith as independent variables. Initially, the independent variables were tested across all data for each lumber size. Then, a series of data segmentation techniques were employed among each of the independent variables: RPI, SG and pith presence. For RPI, the segments were categorized: less than 3 RPI, 3 to 5 RPI, and more than 5 RPI. Secondly, SG segments were developed in groups: less than 0.4, 0.4 to 0.5, more than 0.5. Lastly, the pith data were segmented: no pith, pith one end and pith both ends. Compression parallel and perpendicular to grain were not predictable by RPI, although the SG values by RPI segments did exhibit a stronger statistical relationship. The SG segments themselves did not reveal statistical significance; however, the pith segments yielded higher r2 values. For compression parallel to grain, r2 values were almost double than perpendicular to grain r2 values, across all pith segments and both lumber sizes.



Pine lumber