1 Institute of Nuclear Physics of Moscow State University,119899, Vorobievy Gory, Moscow, Russia

2 Space Research Institute of Russian Academy of Scienses

In this part of the study we have investigated the chromospheric response to the injection of accelerated particles in the flare. We used maps of the active region's vector magnetic field obtained with the MSFC magnetograph to select 14 areas corresponding, mainly, to zones of enhanced magnetic field strength of both polarities. For each of these 14 areas we derived light curves of brightness in the center of the Ha line with a temporal resolution of 0.27s from images obtained by registering 7000 TV frames taken with the telescope co-aligned with the magnetograph. A comparison of the individual Ha light curves with the radio and SXR light curves indicated which of these 14 areas showed chromospheric responses in the initial phase of the flare and thus allowed us to localize the magnetic structures in which the energy release occurred. Comparison of the Ha light curves with the gamma-ray data showed that the statistically significant pulses of the gamma-ray count rates in the impulsive phase coincided with pronounced energy deposits into an Ha area situated near a root of a large filament located along the magnetic neutral line. The spatial-temporal structure of the Ha emission was studied in detail. The temporal behavior of the emission exhibits three characteristic scales: short (less than a second), middle (several seconds), and long (gradual variation spanning the entire event). We considered the short-scale variation as an indicator of the accelerated charged particle injections into small areas of the chromosphere at the footpoints of magnetic structures. We have found that correlations of the short and middle-scale structures in different areas vary significantly during the flare and we attribute these variations to reconfigurations of the active region magnetic field leading to changes in magnetic connections between the areas. These magnetic field reconfigurations are followed by changes in characteristics of all types of emission. In particular, a short dip in the value of the mean correlation coefficient separates the impulsive and the delayed phases of the flare and the most pronounced dip in about 100 s after it is followed by a significant hardening of the high energy gamma-ray spectrum. We suggest that our method of analysis outlined here can be fruitfully applied to other solar flares with registered high-energy emission as we continue our search for common features of these flares.