I remember back in the late 1970s and early 1980s people were speculating that endorphins might be the source of runners high—and I remember how disappointed they were when experiments showed that exercise-induced levels of endorphins were much too low to produce the reported levels of euphoria.
They may simply have been looking for the wrong class of euphoria-inducing chemicals—it now looks like runners high may be produced not by endorphins but by endocannabinoids. Further, according to a paper in the Journal of Experimental Biology, there’s reason to believe that this euphoria is important enough in improving reproductive fitness that it is broadly selected for in cursorial mammals (that is, those with bodies suited for running). See Endocannabinoids motivated exercise evolution for an overview. (The paper itself is behind an annoying paywall for people without some sort of university affiliation.)
This could explain a lot of otherwise hard-to-explain behavior:
. . . a neurobiological reward for endurance exercise may explain why humans and other cursorial mammals habitually engage in aerobic exercise despite the higher associated energy costs and injury risks, and why non-cursorial mammals avoid such locomotor behaviors.
Sadly, the evolved system does not show much promise for turning couch potatoes into endurance athletes:
. . . couch potatoes are not about to leap suddenly out of their comfy chairs and experience the pleasurable effects of exercise, because they probably cannot produce enough endocannabinoids. . . . Inactive people may not be fit enough to hit the exercise intensity that leads to this sort of rewarding sensation.
This is why I’m so pleased at having come up with an exercise regimen that I can persist with over the winter—I enter spring already able to enjoy the euphoric joys of exercise without having to first get in shape.