The computer aided design approach used in current applications of semicrystalline polyoxymethylene (POM) requires high strain-rate mechanical data. The primary aim of this work has been to measure the room temperature modulus and tensile strength of injection molded samples of POM of different molecular weights at cross-head speeds of between 10(-5) ms(-1) and 10 ms(-1). We observe no major transition in bulk mechanical behavior in this range of test speeds, the Young’s modulus E, in particular, showing little strain rate dependence. This is rationalized on the basis of tensile tests over a range of temperatures, these indicating room temperature to correspond to the plateau in the E(T) curves (T-g for these materials is taken to be - 70 degrees C, and the DSC melting onset occurs at similar to 170 degrees C). The tensile strength increases as similar to log(d epsilon/dt) and the behavior is found to be highly nonlinear, strains to fail of the order of 1 being observed even at the highest strain rates, depending on the molecular weight;. It is believed that the yield stress of the crystalline regions determines the tensile strength above T-g, the higher degree of crystallinity associated with lower molecular weights resulting in a slightly higher tensile strength. Nevertheless, failure is qualitatively brittle, with no necking and relatively little permanent deformation. This behavior is discussed in terms of morphological investigations of the fractured samples by optical and scanning electron microscopy (SEM). In attempting to relate ultimate failure to the molecular/crystalline structure of the samples, measurements of the critical stress intensity for crack initiation in mode I opening, K-IC, as a function of crystallization temperature T-c have been carried out using compact tension specimens machined from injection molded and compression molded plaques. K-IC increases with molecular weight and decreases with T-c at low test speeds (in spite of an increase in crystallinity with T-c). This is accounted for in terms of a crack shielding model for crack initiation and of molecular rearrangements occurring during crystallization which lead to a decrease in the effective entanglement density with T-c. The implications of this model are then compared with K-IC results over a range of cross-head speeds and temperatures.