Results of daily observations of the binary system Cyg X-3 using the R
ATAN-600 radio telescope from May 21 to July 20, 1997 are presented. T
hese observations were part of a patrol program to study the flare var
iability of the radio emission of X-ray sources. Two short, moderate (
similar to 1 Jy), optically thick flares were detected on May 22 and M
ay 26 at six frequencies (0.96, 2.3, 3.9, 7.7, 11.2, and 21.7 GHz). Th
e flux decay of the second flare was exponential (S-v= S(0)e(-t/tau))
with tau = 1.03 days. After this flare, there was a sharp decrease in
the radio intensity on May 29-31, to 15-20 mJy at 4-11 GHz, 10-12 days
before a powerful flare on June 12 (similar to 4 Jy). In this hare, t
he flux density of Cyg X-3 rose from the quiescent level by a factor o
f 20-30 over 1-2 days. This was the only flare in our observing period
which was optically thin, with spectral index alpha = -0.34 to -0.50
at high frequencies during the first four days. Four days after the on
set of the June 12 flare, its flux density began to decay exponentiall
y at all frequencies, and the characteristic decay time tau decreased
with frequency according to the power law tau = 11.5(days)v(-0.29+/-0.
03), where v is in GHz. Four days later, the decay changed to a power
law at all frequencies except for 21.7 GHz, where a new outburst began
. The average behavior for all the frequencies for the final stage of
the flare was a decay according to the power law S-v = S(1)t(-4.0+/-0.
2) Later, the spectral index alpha = +0.64 was observed for two, short
, optically thick Bares on June 22 and 29. The June 29 flare displayed
a power-law decay S-v = S(0)t(-2.0+/-0.1) to the same very low flux l
evel from which it began, similar to 35 mJy. During the deep pre-flare
minimum of May 29-31, the spectral index was alpha = +1.23+/-0.05, in
dicating that the radio source was strongly absorbed in this period, P
ossibly due to synchrotron or thermal self-absorption of the central s
ource of Cyg X-3. At frequencies higher than 2.3 GHz, the delay in the
flare flux-density maxima was substantially smaller than the interval
between our observations (one day); this delay was similar to 1.0-1.5
days at 0.96 GHz for local maxima over the entire observation period.