If GRB970228 was the first GRB event for which a X-ray and
optical counterpart was found, the event that triggered the
BeppoSAX/GRBM on 8 May 1997, had to carry
even more striking results.
The GRBM detected the trigger for GRB970508 at 21:41:50 UT,
just few minutes before than the satellite was passing over
the ground station in Malindi ([Piro et al. 1998b]). The
event was simultaneously detected in one WFC, thus providing
a prompt localization of the event at the coordinates
R.A.=06h53m28s and
Decl.=79
17'.4 (equinox 2000.0),
with a 99% error radius of 3'.
In figure 5.7 the light curve
of the event is shown in the energy
ranges of GRBM and WFC, showing a single-peak structure in the
GRBM and a double-peak structure in the WFC. This event was
much weaker than the February event. The
peak flux was 
in the 40-700 keV energy range and
in the 2-26 keV band.
Given the favorable detection condition (i.e. just before the
data down-load in Malindi) and the experience gained from
the BeppoSAX team, the NFI were pointed to the GRB location
just 5.7 hours after the event was triggered, and stayed on source
for an exposure time of 28.000 s. This was again
a world record at that time. As in the GRB970228 case
(and in the GRB970402 case, see below) a previously
unknown X-ray source was detected at the position
R.A.=06h53m46.7s and
Decl.=+7916'02" (equinox 2000.0),
consistent with
the GRB error box, 1SAXJ0653.8+7916.
The BeppoSAX NFI pointed the source again
three more times, after 2.7 days (24.000 s exposure),
4.1 days (12.000 s exposure) and 5.7 days (73.000 s exposure).
The source was detected in each of the four NFI pointings.
Also, an analysis of the WFC data have revealed that the X-ray
counterpart of the GRB was detected in the WFC up to 4.000 s
after the GRB.
As in the case of GRB970228 we have plotted the mean source flux in a log-log plot, and the result is shown in the upper panel of figure 5.8. In this case the time history of the GRB in the 2-10 keV band is plotted. Unlike the GRB970228 the source power law decay is far from being smooth and uniform, but shows the presence of new activity after a first decay. The ability of the BeppoSAX team to point the source very quickly and the peculiar decay law have also allowed for the first time to study the spectral evolution of the afterglow of a GRB over the four NFI pointings ([Piro et al. 1997]). A clear evolution can be seen in X-ray afterglow data ([Amati, L. et al. 1998]), with the hard-to-soft trend typical of the GRBs, that therefore seems to apply at their afterglows as well.
Also the optical follow-up of this event benefited of the
experience and better organization of the BeppoSAX team.
The first observations of this event started as early as 4 hours
after the event ([Castro-Tirado et al. 1997]).
The optical counterpart was identified
independently by several
observers ([Bond 1997,Djorgovski et al. 1997,Sokolov et al. 1997,Pedersen et al. 1997,Castro-Tirado et al. 1997])
at a position R.A.= 06h53m49.43s and
Decl.= +7916'19".6 (equinox 2000.0).
This world-wide effort allowed to follow the temporal
behavior of the optical transient from the very beginning to
the current days. In figure 5.9
we show the collection of the
observations from many telescopes. The time history of the source
appears very complicated, and qualitatively different from
what detected in the case of the February event, or better, it looks
like in that case we missed the first brightening of
the optical transient.
The most important thing, however, is that the optical counterpart was catched up so quickly that it was still bright enough to allow for an optical spectrum to be taken. This was actually done at the Keck telescope (Hawaii) from Metzger et al. ([Metzger et al. 1997]) and revealed the presence of some absorption line features. These were identified to be FeII and MgII absorption lines cosmologically redshifted, because of a distance corresponding to z=0.835. This is the first direct measurement of the distance of a GRB, and it definitely shows that GRB970508 came from outside our own Galaxy.
This GRB was also seen with the Hubble Space Telescope in order to find evidence for a host galaxy, like the one supposed to be detected in the GRB970228 case. The results ([Pian et al. 1997]) of the observation show a perfectly point-like source, with definitely no evidence for any host galaxy.
Moreover, evidence for a possible redshifted iron line (with z consistent with the value measured with optical spectra) has been recently discovered by means of a more time resolved analysis of the X-ray afterglow spectra ([Piro et al. 1999]).
The GRB970508 has entered the astronomy books not only because of its X and optical counterpart, but also because it is the first GRB with a radio afterglow. In fact, the radio observations of the GRB970508 field started at the VLA just 3.7 hours after the GRB trigger ([Frail et al. 1997b]). A new radio source, VLA J065349.4+791619, was seen in a position consistent with the optical transient associated to the GRB. The source flux was monitored at the frequencies of 1.43, 4.86 and 8.46 GHz. Soon after the GRB the source was not detectable, but after about one week it started brightening, reaching a mean flux of about 0.6 mJy at 4.86 GHz. The source exhibited short-term variations with amplitude of a factor of 2 or more. These variations can be interpreted as due to diffractive scintillation because of scattering of the radio wave with the interstellar matter. If this interpretation is correct, and if the source distance is taken to be z=0.835 (this would correspond to a distance of 1028 cm), then the short-term variations allow to estimate the source size: after few weeks the linear dimension of the source should have been of the order of 1017 cm.
Furthermore, a monitoring of the same source with the Very Large Baseline Interferometer (VLBI) ([Taylor et al. 1997]) has allowed to derive a positioning as accurate as 0.2 milliarcseconds. The measurement of a null proper motion at the level of 50 milliarcseconds per year is fully consistent with a cosmological distance to the object.