Light to moderate grazing in grasslands can create vegetation mosaics of short grazed vegetation and tall ungrazed vegetation. These mosaics have been proposed to maximize plant and animal species richness, yet experimental evidence, especially regarding arthropods is scarce. This study compares abundance, richness and species composition of arthropods in grazed mosaics to those of homogeneous short and tall vegetation. We sampled arthropods on three German coastal salt marshes where grazing with three densities (high, moderate and none) was installed in 1989 on previously intensively grazed plots. Stable vegetation mosaics had developed under moderate stocking densities. We collected spiders, beetles, bugs and moth larvae by suction sampling in a stratified random sampling design. Treatments had caused large differences in plant composition after 20 years, which were reflected in the arthropod community. Most species showed a clear preference for either short or tall vegetation, but some species were most abundant in grazed mosaics. Arthropod richness and composition were similar in patches of short vegetation in moderately and highly stocked plots, while patches of tall vegetation were similar to ungrazed plots. Surprisingly, however, grazed mosaics were not richer in species than homogeneous tall vegetation, despite the co-occurrence of species from short, tall and mosaic vegetation. We conclude that, although arthropod richness of salt marshes is greatly enhanced when stocking density is decreased, this cannot substitute ungrazed marshes for conservation of arthropod diversity. However, long term cessation leads to the disappearance of several species, and therefore the possibilities of rotational grazing should be explored.