AbstractThere recently has arisen substantial concern for the reproducibility of scientific findings, but the discussion has not significantly impacted Earth science. We consider repetition, replication, and reproducibility in Earth science, using an example from geomorphology. Repetition repeats the program of observations in the same exercise to establish precision of results. Replication is duplication of observations using similar resources but in an independent program. Reproduction is confirmation of scientific principles using different resources in an independent program. We conclude that results will mainly be limited to reproduction—confirmation of principles—and that this is the essential goal for advancing the science. We illustrate these concepts by review of our experiments on the infiltration of fine grains in flowing water into a bed of coarser grains, conducted using glass beads in a laboratory flume.