Shrews have very high metabolic rates and are often unintentionally starved in rodent live-traps during capture-mark-recapture (CMR) studies. Here, we suggest a shrew exit as a modification to rodent traps. To test whether this modification is (1) saving shrews and (2) not jeopardizing results of rodent captures, we compared captures in Ugglan traps with and without shrew exits, studying bank voles (Myodes glareolus) in a spruce forest in central Finland. Numbers of captured bank voles and body size of smallest juvenile bank voles were not affected by the shrew exit, while the number of captured common shrews (Sorex araneus) was reduced from 31 to 0 individuals per 100 trap nights. However, rare larger shrew species (> 8 g body weight) could not escape through the exit. A shrew exit can, therefore, save smaller shrew species in standard live-trapping of vole-sized rodents without affecting CMR data of the rodent.