A mark-recapture field study was conducted to determine fish passage at 5 concretebox culverts and 5 low-water crossings (concrete slabs vented by culverts) as well as 10 controlsites (below a natural riffle) in Flint Hills streams of northeastern Kansas. Additionally, we testedthe upstream passage of four fish species native to the Great Plains streams (Topeka shiner(Notropis Topeka), green sunfish (Lepomis cyanellus), red shiner (Cyprinella lutrensis), andsouthern redbelly dace (Phoxinus erythrogaster)) through three simulated crossing designs (boxculverts, round corrugated culverts, and natural rock riffles) at water velocities of 0.1 to 1.1 m/s inan experimental stream. The field study indicated that cyprinids were twice as likely to moveupstream of box culverts than low-water crossings and 1.4 times as likely to move upstream ofcontrol reaches than any crossing type. The best models indicated that the proportion ofcyprinids that moved upstream increased with decreased culvert slope and length, perching, andincreased culvert width. Our controlled experiment indicated that fish can move throughvelocities up to 1.1 m/s in a 1.86-m simulated stream and that the proportion of fish that movedupstream did not differ among crossing designs for southern redbelly dace, green sunfish, orTopeka shiner; however, natural rock riffles had lower proportional movements (mean=0.19) thanthe box (0.38) or corrugated culvert designs (0.43) for red shiners. Water velocity did not affectthe proportional upstream movement of any species except that of Topeka shiners, whichincreased with water velocity. Crossing design alone may not determine fish passage, and water velocities up to 1.1 m/s may not affect the passage of many Great Plains fishes. Barriers to fishmovement may be the result of other factors (e.g., perching, slope, and crossing length). The useof properly designed and installed crossings has promise in conserving Great Plains streamfishes.