Using the observations of line-profiles in flare spectra published by Ellison, Mustel, Severny, and Suemoto, we investigate the physical conditions in chromospheric flares. We find in agreement with Mustel and Severny that flares are optically thick in all few lines of the Balmer series, and that lines higher than Hα line is broadened by the radiation damping, and the lines higher than Hβ by the Stark effect. On the contrary, we differ from the Soviet writers in the conclusion that the state of excitation al least in the surface layers of a flare, is rather close to the state of thermal equilibrium.
Comparing the theoretical course of intensity at the Hα line with Ellison´s observation, we find systematic deviations, which may be explained by an influence of the Lα back-radiation of flares on the upper photospheric level.
The central absorption in the Hα line is due to a self reversal of the line due to a decrease of temperature towards the surface of the flare. In large chromospheric flares kinetic temperature is 9000°-10000° inside a flare, and decreases to about 6000° on its surface. Excitation temperature is about 8000° inside a flare and close to 6000° on its surface. Electron temperature is probably close to 10000°. The number of free electrons may be estimated at
6 . 10^12 cm^-3. The radial optical thickness in the center of the Hα line may be as high as 10^4 in the largest flares. In large flares normally observed the optical thickness is equal to 2 . 10^3 -
6 . 10^3. The lowest limit of the optical thickness is given by the thickness of bright flocculi, shich does not much exceed the value of 10; nevertheless in all the phenomena which are usually classified as flares the optical thickness is probably higher than 200.