This week's paper is from Biology Letters in which Pfaller and Gil examine resource use by tiny (<10 mm) crabs that live in the shells of loggerhead turtles. These "columbus crabs", Planes minutus, live in the tail area of the shell, which is likely why they were long assumed to feed on turtle poop. In the 1990's, however, John Davenport discovered that they really serve as minuscule roombas, eating other animals that attempt to colonize the turtle's shell such as barnacles and amphipods. (N.b. Although this is referred to as a symbiotic relationship, there is no data quantifying mutual benefits. It is possible that harboring a crab roomba is actually more costly than the barnacles would have been.) Pfaller and Gil were interested in whether crab population size on turtles or ocean flotsam (floating debris) was determined by the total surface area of the habitat (turtle or flotsam) or the available refuge area (space under the turtle's tail or area of flotsam covered by sheltering barnacles). They show that the refuge area is more important than the total surface area in predicting crab population size. The smaller refuge available on turtles resulted in predominantly one male-female pair of crabs living on a turtle, versus a large piece of flotsam might have up to 233 crabs. The behavior of turtle-cleaning might therefore also generate social monogamy in crabs, with all kinds of consequences for behavior and evolution. In all fairness this had been proposed previously by Thomas Dellinger, but this new paper supports Dellinger, and quantifies the refuse space.