A method for screening the tolerance of Rosa rootstocks and species to Pratylenchus penetrans was developed after observing the effects of nutrition and initial nematode population density on damage to Rosa corymbifera cv. Laxa. The correlation between initial P. penetrans densities and the weight of host plants fitted (R-2 > 0.90) the Seinhorst model, Y = y(m) for Pi less than or equal to T, and Y = y(m) . m + y(m)(1 - m)z((Pi-T)). All inoculated plants of R. virginiana and R. eglanteria survived while there were differences in the mortality after inoculation of 12 other Rosa rootstocks or species. Pratylenchus penetrans caused significant reductions in the numbers of the leaves and roots, and in fresh weights of shoots, roots and whole plants of all 14 accessions. Among these species and cultivars, reductions in root number were least for R. virginiana whereas fresh weights of roots, shoots and whole plants were least affected for R. eglanteria, and the greatest reductions were of R. corymbifera cv. Laxa and R. glauca. Plant mortality and the reduction of shoot and plant weights of surviving plants were positively correlated with the reduction in numbers of secondary roots. The numbers of secondary roots at inoculation were negatively correlated with the degree by which plant weight was reduced and with plant mortality observed 50 days later. Differences in tolerance among species as observed by differences in reduction of plant growth parameters were confirmed by differences of the parameters derived from applying the Seinhorst model to the species.