The primary direction of this thesis was towards exploring qualitative and quantitative characteristics necessary for refining and understanding the flat wood double cantilever beam (DCB) as a valid means for testing Mode I fracture energy in wood adhesive bonds. Southern yellow pine (SYP) adherends were used with epoxy and phenol formaldehyde (PF) impregnated films, providing two systems with different characteristics for investigation.

An adhesive penetration analysis was performed for both the epoxy and PF bonds. The PF penetration into the SYP was shown to be relatively shallow. The epoxy penetration was shown to be deeper. Epoxy-SYP DCBs were quasi-statically tested with varying widths (10 mm, 15 mm, and 20 mm), showing decreases in scatter of critical and arrest strain energy release rates, GIc and GIa, with increases in specimen width. Quasi-static fracture testing was also performed on PF SYP-DCBs, showing much higher critical and arrest fracture energy values than the epoxy-SYP DCBs, indicating that deep adhesive penetration is not necessarily a requisite for higher Mode I fracture energy values.

Grain distribution influences were computationally investigated because of the stiffness difference between latewood and earlywood growth and the grain angle along the length of the beams. The grain angle and the stiffness difference between latewood and earlywood growth caused the effective stiffness, (ExxI)eff, to vary along the length of the beam. The effective stiffness variation caused variations in the beam's ability to receive and store strain energy, complicating and confounding determination of experimental results.

Cyclic loading tests were performed on PF-SYP DCB's. The cycle frequency was 3Hz, with a valley to peak load ratio of R = 0.5. Specimen softening was observed with cycling, with re-stiffening occurring with crack growth. Contrary to expectations, specimen compliance occasionally decreased with small crack extensions. A toughening mechanism was frequently observed, whereby subsequent crack lengths required more cycles to failure than the previous crack length. Monotonically extending the crack length far from the fatigued region created a fresh crack that did not show the toughened behavior. But toughening did resume with subsequent crack lengths.