Larry Alexander's basic worry about adding thresholds to deontology is that to do so is hopelessly ad hoc and arbitrary – these are alien additions made by a theory desperate to avoid otherwise devastating counterexamples, but additions having no resonance with the theory that they are saving. Whether this is so depends on how one conceives of a more basic issue for deontology, namely, how its obligations win out over consequentialist considerations in more everyday situations, situations that are below whatever threshold one posits to exist. The guiding thesis is that if one rightly conceives of how deontology wins out over consequentialism below the threshold, one will have less difficulty in smoothly conceptualizing how deontology loses out to consequentialism in situations above the threshold. In each of these two scenarios, the most stringent obligation prevails, where "stringency" needs to be cashed out (both for deontological and for consequentialist obligations) in non-question-begging terms.