We analyze 117 echelle spectra of the T Tauri star DF Tau, concentrating on variations in the optical continuum veiling and the strong emission lines. Although this star was the inspiration for the original suggestion of magnetospheric accretion in T Tauri stars (TTSs), this hypothesis is only partially supported in our data. We find that variations in the Ca II infrared triplet lines correlate with the veiling variations; there is some evidence that the broad component of the He I line does, too. The narrow component of He I is shown to arise at the stellar surface, but it correlates with the broad component. There is a surprising lack of periodicity in the lines, and it does not occur where expected when seen. The correlation between continuum veiling and the line components expected to be most related to the veiling is poor. There is a great deal of variability in all the lines and line components; a snapshot spectrum is a poor way to characterize the star as a whole. The total Balmer line fluxes are poorly correlated with the veiling, unlike previous results on a large sample of TTSs. Redshifted absorption components are found in the weaker lines but are not common. The strength of the blueshifted absorption feature in H alpha is correlated with the veiling, but changes in it perhaps occur before veiling changes by about one day. This time delay supports the idea that the wind originates at some distance for the stellar surface and is related to accretion. Spherically symmetric wind models are unable to reproduce well the relative absorption levels on the blue side of the H alpha and H beta lines simultaneously. H alpha does not display the asymmetries expected of magnetospheric accretion, but it is sometimes suggestive of azimuthally asymmetric corotating structures. The lines wings indicate that the formation region of the H alpha lines is dominated by high turbulence. H beta does show more of the asymmetry expected of magnetospheric accretion.
