1. The North American Great Plains contains thousands of lakes that vary in salinity from freshwater to hypersaline. Paleolimnological studies show that salinity levels in these lakes are tightly linked with climate, and current projections point to a more arid future in the region due to natural and anthropogenic climate change, potentially influencing lake salinity.2. Many zooplankton species are sensitive to changes in salinity, and their position near the base of the aquatic food web makes it important to understand how they might respond to increasing salinity levels. Zooplankton communities in lakes with rising salinity levels may exhibit changes in structure, including a shift toward more salinity-tolerant species and a reduction in abundance, species richness, and diversity. However, it is possible that dispersal of zooplankton among lakes could mitigate such community changes when migrant populations replace sensitive zooplankton with those that are locally adapted to higher salinities.3. To test if dispersal could reduce salinity-induced changes in zooplankton communities, we ran a field enclosure experiment at a freshwater lake in southern Saskatchewan where we manipulated salinity levels and zooplankton dispersal.We evaluated how salinity and dispersal influenced species identities and relative abundances (community structure) using multivariate statistics and comparing taxonomic and functional compositions among the different treatments (richness, diversity, and evenness). 4. We found that increasing salinity levels in our enclosures above that in our study lake resulted in lower zooplankton abundances and species richness levels, primarily due to the loss of cladoceran species. However, patterns in our multivariate analyses suggested that cladocerans were maintained in enclosures with salinity levels of 2.5 and 5.0 g/L when those enclosures received immigration from nearby lakes.5. In contrast, our univariate analyses failed to find evidence that immigration affected community structure (richness, diversity, evenness). The lack of significant statistical differences could suggest that dispersal does not have an effect, or it may have been a problem with statistical power, as a power analysis suggested that fairly large effect sizes would have been required to achieve statistical significance.