We present the unfiltered ROTSE-III light curve of the optical transient associated with GRB 050319 beginning 4 s after the cessation of gamma-ray activity. We fit a power-law function to the data using the revised trigger time given by Chincarini and coworkers, and a smoothly broken power-law to the data using the original trigger disseminated through the GCN notices. Including the RAPTOR data from Wozniak and coworkers, the best-fit power-law indices are alpha = -0.854 +/- 0.014 for the single power-law and alpha(1) = -0.364(-0.019)(+0.020), alpha(2) = -0.881(-0.031)(+0.030), with a break at t(b) = 418(-30)(+31) s for the smoothly broken fit. We discuss the fit results, with emphasis placed on the importance of knowing the true start time of the optical transient for this multipeaked burst. As Swift continues to provide prompt GRB locations, it becomes more important to answer the question, "when does the afterglow begin?'' in order to correctly interpret the light curves.