Seeds of many plant species are dispersed by seed-caching rodents that place groups of seeds in superficially buried scatterhoard caches. A case in point is Indian ricegrass (Achnatherum hymenoides), an important forage plant on arid western rangelands for which seedling recruitment comes largely from scatterhoards made by desert heteromyid rodents. A “diversionary seeding” strategy has been attempted for enhancing Indian ricegrass seedling recruitment by deploying commercially available seeds on the soil surface to divert rodents from recovering scatterhoards of Indian ricegrass seeds. The probability of such a passive restoration approach succeeding is likely affected by the relative desirability to rodents of Indian ricegrass seeds versus diversionary seeds. We conducted laboratory experiments to test preferences of Merriam's kangaroo rat (Dipodomys merriami), a primary dispersal agent of Indian ricegrass seeds, for Indian ricegrass seeds versus seeds of 5 diversionary seed candidates. Indian ricegrass seeds were consistently preferred over only 1 of the 5 alternate seeds in pairwise trials. In multiseed trials that presented all seed types (i.e., Indian ricegrass and the 5 alternates) simultaneously, ranking patterns of individual kangaroo rats varied significantly and Indian ricegrass was not preferred by any of the animals tested. Because individual kangaroo rats differed in seed preferences and all animals consumed certain seed types in greater amounts than they consumed Indian ricegrass, we suggest that using a mixture of different seed types in diversionary seedings is superior to deploying a single type of diversionary seeds. Understanding how population-level niche breadth is affected by dietary variation at the individual level can thus have important management implications.