How do protist species differ in dispersal ability? We propose to conduct a field assay to compare the dispersal abilities of at least five protist species exposed to their natural dispersal agents. The small body size and rapid generation time of protists means that it is not possible to physically tag individuals of each species and positively identify the populations they give rise to in nearby empty pitchers. Instead, we will set up an artificial “population” of potted pitcher plants that enables us to fully control the source of dispersing protists and availability of dispersal agents they can use to move to new protist-free habitat patches. These plants will remain in their pots for the duration of the experiment. We will control for any potential gene flow by pollen or seed by clipping all flowers prior to opening. All pitcher plants will be sterilized and removed from the field site at the conclusion of the experiment. Our array consists of a group of “source” pitchers at the center surrounded by eight radial arms along which we will place 32 protist-free “recipient” pitcher plants. To initialize the array, we will add 25mL of each protist culture to a single pitcher on each of three plants located at the center of the array following a randomized block design. Each recipient pitcher in the array will be filled with the same bacterially-inoculated, but protist-free, media to promote rapid growth of arriving protists. The field dispersal experiment will use laboratory cultures of bacteria and protists originally isolated from pitcher plants in northern Michigan. These microbes would be introduced to a new site with zero barriers preventing escape into the environment. We expect that these microbes have adapted to lab conditions, but this has most likely reduced their ability to establish and compete in the wild where they interact with other species and experience strong fluctuations in the abiotic environment. Furthermore, the protist species are all pitcher specialists not known to occur anywhere else, including aquatic habitats immediately adjacent to their host plants (Sleith & Katz 2022). Because we must conduct this experiment away from natural pitcher plant populations, we do not expect there to be any naturally occurring specialist mosquitos (a hypothesized protist dispersal agent) at the array site. Instead, we will collect late instar larvae and pupae from the nearest natural Sarracenia population and rear them to adulthood in laboratory flight cages before releasing them in the center of the experimental array. Previous research suggests that this mosquito species has poor dispersal capacity and remains within 45 m (Krawchuk 2009). We expect the adults will die after ~38-42 days (Bradshaw, Holzapfel, & Davidson 1998). Because they are specialists, they will not be able to breed at the array location once we remove pitchers at the end of the experiment. We will allow wind and ovipositing mosquitoes to disperse protists around the array for up to 7 days. We will visit the array twice daily, sampling from three flagged pitchers per plant. During each visit, we will draw a 1mL sample, replace the lost volume with fresh media, and return to the lab to screen samples for protists. Samples that indicate successful protist invasion has occurred will trigger a sterilization/resetting process for the pitcher they were taken from. Using the data collected from this experiment, we will estimate each species' dispersal kernel. What mechanisms do protists rely on for dispersal between habitat patches, and does the importance of each mechanism differ among species? We will exploit differences in abiotic versus biotic dispersal agents to determine how species move among patches and test whether these explain overall differences in dispersal ability. Alongside each recipient pitcher plant in the dispersal array (Fig. 1), we will place a 50mL centrifuge tube containing the same bacterially-inoculated media. Centrifuge tubes lack the characteristic color, form, nectar production, and specialized surface texture mosquitos likely use to find suitable egg-laying habitat (Heard 1994). Therefore, we expect that mosquitoes will not oviposit in them and all protist arrivals will be due to wind. To ensure that this assumption holds, we will conduct a brief pilot prior to deploying centrifuge tubes in the field. We will enclose a protist-bearing pitcher plant and protist-free centrifuge tubes in a flight cage with a divider to limit airflow while allowing mosquito movement between the plant and tubes. We will introduce 20 ovipositing mosquitoes for 24h, then monitor for the presence of protists and larvae in the centrifuge tubes. Provided that our validation of this method works (i.e., neither protists nor mosquito larvae are detected in tubes from the cage experiment), we will separate the analysis of pitcher and centrifuge tube data from our field array and separate out species reliance on wind versus insect vectors. Is protist dispersal ability correlated with competitive ability? We will test for a diversity promoting trade-off in dispersal and competitive ability by combining field dispersal estimates with lab measurements of species interactions. We will use linear regressions to test for a relationship between species’ competitive ability and their (i) maximum dispersal distance and (ii) dispersal frequency. This section of the proposal requires no manipulations or collections at the proposed research site. How do realistic species arrival orders constrain end state diversity and species composition? Our lab-based protist growth and competition assays provide all of the parameters needed to map the “invasion graph” for our focal protist species, showing the full suite of possible community assembly pathways and end states. We hypothesize that differences in dispersal ability between protist species (estimated in our field experiment, Q1) will bias which assembly pathways are more likely and change the end state diversity and species composition. To test this hypothesis, we will compare simulations of the community assembly process for our focal protist species under empirical differences in dispersal ability versus equal dispersal abilities. We will initialize each new community assembly process as an empty community containing no protists, akin to a newly produced pitcher. At each step, we will draw a species to invade the existing community. In the empirical scenario, we will weight the probabilities that each species is drawn by their dispersal frequency (i.e., the integral of the dispersal kernel). This will allow the better dispersers to arrive earlier and more often, on average, compared to poorer dispersers. In the equal scenario, we will assign identical weights to each species, meaning that each will be equally likely to arrive at the first and each subsequent stage. We will repeat the addition of a new species until the community reaches a stable end-state, and we will conduct 10,000 community assembly simulations under each scenario. This section of the proposal requires no manipulations or collections at the proposed research site.