A lot of variables come into play with feeding behavior. Even food preference can have a genetic link. Then there's the availability of specific prey items, which will vary from location to location, specific local adaptations, etc. To say garters do not constrict is too general. I'm sure that such behavior would develop in a population that relies heavily on rodent prey. Even though concinnus' I've kept do not rely heavily on rodents where I find them, they will eat them given the opportunity, and I have seen them pin dead rat pinkies against the glass while they eat them. Presumably to assist with swallowing though. I've never personally seen them coil around prey, but maybe a different garter, from a different place, that relies heavily on rodents would develop constriction behavior over generations. Then when you use vagrans as an example, well, one of the first vagrans I found in WA this year vomited up a full grown mouse when I got it home. If that mouse was alive when the snake attacked it, there had to have been a struggle and the snake had to subdue the mouse somehow. I get the overall impresson that vagrans are more apt to be rodent eaters in the first place.

But lets make no mistake. Any prey item that leaves wads of undigested material is more likely to cause an impaction than say, worms and fish. Let me rephrase. Not "cause" an impaction, but rather it provides the necessary solid undigested material for a blockage to consist of.