On the base of photoelectric observations (VAssILYEVA, 1967) we have drawn the conclusion, that an excess of brightness of some inhomogeneities (with r.m.s, of the O/ fluctuations ~ 2/0) is observed in a limited spectral range, and that the pictures of the inhomogeneities in two different colors are different.
However, the structure of the solar granulation has been shown by EVANS (1968) not to be greatly different in widely separated colors.
In my opinion, there are at least three sources for this discrepancy: (1) the influence of differential atmospheric refraction which we did not take into account and which was discussed by Evans; (2) the difference of the instruments used; (3) differences in the exposure times.
In actual fact, the influence of the atmospheric refraction is included in our results. But although its value is known the question remains to what extent the real fluctuations in brightness are distorted by this phenomenon. We could not directly eliminate it from our recordings, but we tried to estimate the influence of the earth's atmosphere in an indirect way, and to show the regularities of solar origin, using the specific advantages of our method.
The effect of atmospheric refraction on our results seems to be overestimated by Evans. First, there is a clear dependence of the radius of correlation on the color, obtained from more than 60 recordings in seven spectral regions. As the average sizes of the inhomogeneities are determined from the autocorrelation functions for each recording separately, and as they are free from the influence of atmospheric refraction, such a dependence means that the absolute identity of two recordings in two colors is impossible in principle for our observations. So, the low cross-correlation between the brightness fluctuations in two colors may be due to other reasons than atmospheric refi'action only.
Second, numerous recordings often show the inhomogeneities at different wavelengths to be displaced relative to each other in individual cases, the values of this displacement being larger in the West-East direction than those of the projection of the atmospheric refraction in the direction of scanning.
Moreover, if the displacements were due to atmospheric refraction the maximum of cross-correlation might be expected to occur at a lag, comparable with the value of differential atmospheric refraction. But we have obtained the maximum of crosscorrelation at a lag equal to zero.