This work examines the usefulness of the single-fiber fragmentation test in studying the durability of fiber/matrix interfaces/interphases. This test measures the critical length/diameter ratio (L/D) of the fiber fragments formed in the test and relates this length to the interface's strength, or ability to transfer load. In the work reported here, we immersed samples of epoxy containing a single-glass fiber - that was previously sized with an epoxy-compatible coating - in either 65 or 75 °C water and tested after different times of exposure. In general, this ratio increased as a function of time of exposure to water. During exposure at 75 °C, the fibers' L/D in the samples did not increase significantly until after the sample reached its “apparent” equilibrium content of water ∼ (3.0 wt%). Because there was no significant measurable change in the tensile modulus between wet and dry samples, we cannot attribute these differences in L/D to changes in the resin's properties due to plasticizing by water. A small percentage of samples exposed at 65 °C did not show a significant increase in L/D, and in these cases the moisture produced a marked roughening of the fiber surface along the fiber/matrix interface. One possible explanation is that the attack by moisture degrades the interface, thus reducing its strength with a corresponding increase in the L/D. To varying degrees, however, the attack by moisture also degrades the E-glass fiber. This attack by moisture roughened the surfaces of the fibers and increased the distribution and/or size of the critical flaws, thus reducing both the strength of the fiber and the L/D. Based on our preliminary results, it appears that the singlefiber test has the potential to be useful for studying the durability of the resin/matrix interface providing that the influence of the environmental agent on all of the components of the model composite: resin, fiber, and interface/phase, is considered.