Embryogenic cell suspensions of cotton (cv. Coker 312) were sieved to obtain a fine fraction. This fraction was then subjected to severe growth limiting high or low temperature stress for selection of resistant variants. Calls were placed in a controlled temperature incubator for various periods of time at either 45 C, 40 C, or 50 C to determine a desirable time/temperature exposure to obtain a severely limiting but not unrecoverable stress for the cell lines. Cells resistant to high temperature were then selected using three methods: continuous exposure at 45 C for 10 hours, three hour pulses at 45 C at 24 hour intervals until 100 hours of stress were accumulated, three hour pulses at 45 C at 24 hour intervals until 75 hours of stress were accumulated followed by exposure to a lethal temperature of 45 C for 24 hours. Cells were selected for low temperature resistance by prolonged exposure to 5 C. Cells were plated for recovery after 2 weeks and 4 weeks exposure. Cell lines were maintained in the absence of stress for 5 subcultures prior to retesting for sustained resistance to temperature stress. When tested, significant increases in cell viability over the control line were noted for all methods of high temperature selection. The growth curve of selected vs non-selected cell lines of low temperature selected lines were compared. While no growth was measurable for either line at 5 C, 10 C, or 15 C constant temperature, selected cell lines entered the log phase of growth more rapidly under cycling 30/15 C temperature than non-selected and produced a greater cell volume in a defined period of time. Additionally, when shoot tips of regenerated plants were rooted in vitro at 22 C (inhibitory to normal rooting) only 30% of non-selected shoot tips rooted whereas 100% of those from selected cell lines rooted. Of nonselected shoot tips that rooted the rooting pattern was very sparse, whereas the rooting pattern of selected tips was very normal.