Long-term creep and recovery models (master curves) were developed from short-term data using the time temperature superposition principle (TTSP) for kiln-dried southern pine loaded in compression parallel-to-grain and exposed to constant environmental conditions (~70°F, ~9%EMC). Short-term accelerated creep (17 hour) and recovery (35 hour) data were collected for each specimen at a range of temperature (70°F-150°F) and constant moisture condition of 9%. The compressive strain was measured using bonded electrical resistance strain gages. For each specimen, the compliance curves for all the temperature levels were plotted against log-time on the same plot. The curve segments at successively higher temperature levels were shifted along the log-time axis with respect to the curve section at 70°F to construct a master curve for each specimen. The extrapolation of the developed master curves ranged from 0.23 to 6.4 years. The requirement that the shift factors below glass transition temperature follow Arrhenius formulation was satisfied by the empirical shift factors. The activation energy for creep and recovery of kiln-dried southern pine derived from the slope of the plot of horizontal shift factor and the inverse of the absolute temperature was 28 KCal/mole. Creep and recovery master curves were represented by power functions and the nonlinear regression analysis was used to estimate the model parameters. Linear regression models were developed to predict one parameter in creep and recovery models from Young's modulus. The other model parameter showed weak correlations with material properties; therefore, an average value was recommended. The validity of the master curves for predicting creep of wood exposed to normal interior environmental conditions in buildings was tested by conducting long-term (10 month) creep tests in a heated/cooled laboratory environment. The fluctuating test environmental conditions caused geometry changes in the surface of the wood specimens in addition to mechanosorptive creep leading to fluctuating long-term data. Therefore, a good agreement between the master curves and long-term data was not found. Creep behavior of shallow southern pine arches was studied to demonstrate the application of the finite element method, incorporating the long-term curves based on TISP, to predict creep in wood structures. Creep tests were conducted at various load levels applied at ambient environmental conditions for two months. One arch failed (i.e., snapped-through) nine days after the tests began indicating that creep can indeed cause instability failure in shallow structures. It was found that the supports in the arch test fixture deflected elastically; therefore, the arches were modeled as three pin structures with base pin joints supported by zero-length linear elastic springs. However, the elastic analysis results revealed the presence of other factors affecting the experimental response which complicated the modeling procedure. The creep analysis was performed using a finite element model incorporating the developed creep master curves; however, due to the complexity in the creep experimental apparatus, the numerical predictions were not validated experimentally. / Ph. D.