The sexually size-dimorphic grasshopper Tetrix japonica exhibits variation in body-color markings on the pronotum even within a single local population. Such markings have been suggested to reduce the visual detectability of grasshoppers. However, some grasshoppers have no markings. In the present study, we examined the effect of the sex-related difference in body size and the spotted markings on the degree of camouflage. We hypothesized that: (i) large individuals (females) are potentially more readily detectable than small individuals; (ii) large individuals (females) with spotted markings would realize a moderate degree of the camouflage effect, whereas large individuals (females) without spotted markings would be quite poorly camouflaged; (iii) small individuals (males) would be sufficiently less detectable, with or without markings; and (iv) large individuals (females) would tend to have spotted markings in the wild. To test this hypothesis, we conducted a morph-frequency field survey and laboratory experiments on the body-size dependency of the spotted markings. The field survey confirmed that all females exhibited spotted markings and that the majority of males were non-spotted morphs. Next, to determine whether body size and the spotted markings affected crypsis, we conducted detection task experiments using humans as dummy predators by manipulating the body size, presence/absence of spotted markings, or both, of printed grasshoppers. The absence of spotted markings increased the detection risk in large and small grasshoppers, particularly in large-sized females. These results suggest that female-biased selective predation could have eliminated non-spotted female morphs because they were too conspicuous.