I test a “clock-advance” model that implies triggered tremor is ambient tremor that occurs at a sped-up rate as a result of loading from passing seismic waves. This proposed model predicts that triggering probability is proportional to the product of the ambient tremor rate and a function describing the efficacy of the triggering wave to initiate a tremor event. Using data mostly from Cascadia, I have compared qualitatively a suite of teleseismic waves that did and did not trigger tremor with ambient tremor rates. Many of the observations are consistent with the model if the efficacy of the triggering wave depends on wave amplitude. One triggered tremor observation clearly violates the clock-advance model. The model prediction that larger triggering waves result in larger triggered tremor signals also appears inconsistent with the measurements. I conclude that the tremor source process is a more complex system than that described by the clock-advance model predictions tested. Results of this and previous studies also demonstrate that (1) conditions suitable for tremor generation exist in many tectonic environments, but, within each, only occur at particular spots whose locations change with time; (2) any fluid flow must be restricted to less than a meter; (3) the degree to which delayed failure and secondary triggering occurs is likely insignificant; and 4) both shear and dilatational deformations may trigger tremor. Triggered and ambient tremor rates correlate more strongly with stress than stressing rate, suggesting tremor sources result from time-dependent weakening processes rather than simple Coulomb failure.