In live cell imaging of fluorescent proteins, phototoxicity of the excitation light can be problematical. Cell death is obvious, but reduced cell viability can make the interpretation of observations error prone. We characterized the phototoxic consequences of 488 and 546 nm light on untransformed human cells and tested methods that have or could be used to alleviate photodamage. Unlabeled RPE1 cells were given single 0.5-2.5 min irradiations in early G1 from a mercury arc lamp on a fluorescence microscope. Four hundred eighty-eight nanometer light produced a dose-dependent decrease in the percentage of cells that progressed to mitosis, slowing of the cell cycle for some of those entering mitosis, and a approximately 12% incidence of cell death for the highest dose. For 546 nm light we found a 10-15% reduction in the percentage of cells entering mitosis, no strong dose dependency, and a approximately 2% incidence of cell death for the longest irradiations. For cells expressing GFP-centrin1 or mCherry-centrin1, fewer entered mitosis for each dose than unlabeled cells. For constant total dose 488 nm light irradiations of unlabeled cells, reducing the intensity 10-fold or spreading the exposures out as a series of 10 sec pulses at 1 min intervals produced a minor and not consistent improvement in the percentage of cells entering mitosis. Reducing oxidative processes, by culturing at approximately 3% oxygen or adding the reducing agent Trolox noticeably increased the fraction of cells entering mitosis. Thus, for long-term imaging there can be value to using RFP constructs and for GFP-tagged proteins reducing oxidative processes.