Experiments tested the hypothesis that seeds of Mitella nuda are dispersed by the impact of rain on open seed-bearing capsules. Shielding plants from rainfall for a month did not significantly reduce seed dispersal. Reduction of surface wind and exclusion of large animal vectors did not decrease the number dispersed. Results from the field test were reinforced by discovering in the laboratory that simulated raindrops falling on open capsules from a height of 3.5 ft. could not cause dispersal. Several hypothetical mechanisms are offered as alternative to the idea that the seeds are dispersed by rainfall. The implications of the results in terms of evolutionary relationship between M. nuda and M. diphylla (a species considered to use a splash-cup mechanism) and an apparent adaptive shift within the genus are discussed.