Isolates of Phytophthora sojae were collected during 1995 to 1998 from soil samples collected in 23 Arkansas soybean fields in 14 counties, and characterized by race. A total of seven races (races 2, 10, 14, 15, 24, 26, and 38) were found. Races 10, 24, and 15 were the most common and comprised 47, 22, and 9% of the 32 isolates, respectively. A single isolate each of races 2, 14, 26, and 38 also was found. Three of the isolates collected could not be characterized to race due to inconsistent results. In 1997 and 1998, a portion of a single soybean field at the University of Arkansas Southeast Research and Experiment Center near Rohwer, AR was surveyed intensively for P. sojae. The area was planted each year to the P. sojae-susceptible cv. Williams and both plants and soil were collected to assay for P. sojae from 16 and 28 plots (4.9 by 7.6 m) in 1997 and 1998, respectively. A total of 83 isolates were collected (11 from plants and 72 from soil), and found to represent 13 pathotypes, including 6 with virulence formulae that have not been described previously. Nine commercial soybean cultivars representing a range of reported resistance and tolerance to Phytophthora root and stem rot were screened for resistance to races 10, 15, and 26 of P. sojae using both hypocotyl injection and inoculum layer techniques. Cvs. Manokin, Hartz Variety 5545, and Riverside 499 were consistently resistant to all of the races using both inoculation methods. These results indicate that, although considerable pathogenic variability in P. sojae exists in soybean fields in Arkansas, cultivars with effective resistance are available to help growers manage Phytophthora root and stem rot.