We have measured the ratio of Ha to H/~ central intensities in the peak kernels of 14 flares, using simultaneous filtergrams. The ratio is typically one with some scatter. By contrast, in stellar flares the ratio is about 0.8.In recent years we have seen Zirin (1978) that the intensity ratio of hydrogen lines is not always what we expect it to be. Important physical information can be obtained from these ratios. Most attention lately has been given to the Ha/La ratio, but one should be able to learn from Balmer ratios as well. In stellar flares, for example, it has been found (Kunkel, 1970; Mochnacki and Zirin, 1980) that H3, is the most intense Balmer line; this is attributed to the high temperature of the stellar flares. Thus in these events we have a Balmer increment for n < 6.There is little published data on the Ha/Hfl ratio in solar flares. The most complete data are given by Smith (1963) on the great flare of September 2, 1960. Smith found central intensity (relative to local continuum) ratios of 1.13, 0.848, and 1.27 in start, maximum and postmaximum flare stages, respectively. These ratios must be corrected for the 6563/4861 continuum ratio and also corrected to the frequency scale by dividing by A2; an overall factor 1.18 is obtained, giving 1.33, 1.0, and 1.50, for the Ha/H/3 ratios of F~ at start, max and decline. Smith gives line widths to zero intensity for the Balmer lines; the widths for Ha and Hfl are equal, which means that the Hfl width in the frequency scale is 1.8 times greater than that for Ha, and the integrated intensity correspondingly greater. Jefferies et al. (1959) give similar widths, but give no intensities. Smith found Ha/Hfl (central intensity) = 1.13, 0.848, and 1.27 at the start, peak, and decay of the flare. Reduced to F, these ratios are in good agreement, with H/3 about 1.4 stronger than Ha. We have used Smith's central intensities rather than the intensity above background, because the flare emission is obviously optically deep. The stellar flare data has subtracted the continuum, because we do not know the relative areas of line and continuum sources; the possible error introduced this way is only significant for the post flare measurements. Note that although the half-widths given by Smith decrease in the higher Balmer lines, the corrected widths dA/A increase. Svestka (1960) measured Ha and H/3 simultaneously in a flare and found Ha 3 times the local continuum and Hfl 4 times the local continuum, giving a ratio of F~ of 1.5. Svestka (1976) gives line width data from himself and Suemoto and Hiei indicating a half-width ratio of 1.3 corresponding to constant dA/A. The consensus of these data appears to be that dA/A for Hfl is equal to or greater than that for Ha, a remarkable and hitherto unnoticed result.
