\section{Introduction}

\subsection{Observation of GRBs}

The MAGIC telescope's support structure and mirrors have been designed exceptionally light in order to
to react quickly to GRB alerts from satellites. \cite{design} and~\cite{PETRY} set 
the objective to turn the telescope to the burst position within 10-30\,sec.
in order to have a fair chance to detect a burst when the emission is still ongoing.
During the comissioning phase it could be proven that that goal was reached.
The telescope is able to turn 180 degrees in azimuth within 20\,sec. and 80 degrees in zenith within 10\,sec.\\


Very high energy (VHE) GRB observations have the potential to constrain the current GRB models
on both the prompt and extendend phases of GRB emission~\cite{HARTMANN,MANNHEIM,SALOMON}. 
Models based on both internal and external shocks predicts VHE fluence comperable to, 
or in certain situations stronger than, the keV-MeV radiation, 
with duration ranging from shorter than the keV-MeV burst to extended TeV afterglows~\cite{DERMER, PILLA, ZHANG1}.

\par

In many publications, the possibility that more energetic $\gamma$-rays come along with the 
(low-energy) GRB, have been explored. Proton-synchrotron emission~\cite{TOTANI} have been suggested 
as well as photon-pion production~\cite{WAXMAN,BAHCALL,BOETTCHER} and inverse-Compton scattering 
in the burst environment~\cite{MESZAROS93,CHIANG,PILLA,ZHANG2}.
Long-term HE $\gamma$ emission from accelerated protons in the forward-shock has been predicted in~\cite{LI}. 
This model predicts GeV inverse compton emission even one day after the burst.
Even considering pure electron-synchrotron radiation predicts measurable GeV emission for a significant fraction of GRBs~\cite{ZHANG2}.\\

GeV emission in GRBs is particulary sensitive to the Lorentz factor and the photon density of the emitting material - 
and thus to the distance of the radiating shock from the source - due to the $\gamma~\gamma \rightarrow$
\textit{e$^+$~e$^-$} absorption in the emission region. Direct comparison of the prompt GRB flux at $\sim$ 10\,GeV and $\sim$ 100\,keV 
may allow to determine the magnetic field strength~\cite{ASAF}.

\par

Several attempts have been made in the past to observe GRBs in the GeV range,
each indicating some excess over background but without stringent evidence. 
The only significant detection was performed by EGRET which detected seven GRBs emitting high energy (HE) 
photons in the 100\,MeV to 18\,dGeV range~\cite{EGRET}. The data shows no evidence of a HE rollover 
in the GRB spectrum~\cite{DINGUS}. Recent results indicate that the spectrum of some GRBs contains a very hard, 
luminous, long-duration component~\cite{GONZALES}.
There have been results suggesting gamma rays beyond the GeV range from the TIBET air shower array
in coincidence with BATSE bursts~\cite{AMENOMORI}, rapid follow-up observations by the 
Whipple Air Cerenkov Telescope~\cite{CONNAUGHTON1}, and coincident and monitoring studies by HEGRA-AIROBICC~\cite{PADILLA}, 
Whipple~\cite{CONNAUGHTON2} and the Milagro prototype Milagrito~\cite{MILAGRO}.
The GRAND array has reported some excess of observed muons during seven BATSE bursts~\cite{GRAND}. 
In this context, especially the publication from the TASC detector on \eg is important~\cite{GONZALES}, 
finding a HE spectral component presumably due to ultra-relativistic acceleration
of hadrons and producing a spectral index of $-1$ with no cut-off up to the detector energy limit (200\,MeV).\\

Concerning estimates about the MAGIC observability of GRBs, a very detailed study of GRB spectra obtained from the 
third and fourth \ba catalogue has been made in~\cite{ICRC,NICOLA}. The spectra were extrapolated to GeV energies 
with a simple continuation of the observed high-energy power law behaviour and the calculated fluxes compared with \ma sensitivities. 
Setting conservative cuts on observation times and significances,
and assuming an energy threshold of 15~GeV, a GRB detection rate of $0.5-2$ per year
was obtained for an assumed observation delay between 15 and 60 sec. and a BATSE trigger rate ($\sim$\,360/year).

Taking into account the local rate of GRBs estimated in~\cite{GUETTA}, late afterglow emission from few tens of GRBs per year 
should be observable above our energy threshold. The model of~\cite{ASAF2} predict delayed GeV emission that 
should be significantly detectable by MAGIC in 100\,seconds.

\subsection{Observation of XRFs}

While the major energy from the prompt GRBs is emitted in $\gamma$-rays ($E_p \sim$ 200~keV), XRFs are characterized
by peak energies below 50~keV and a dominant X-ray fluence. Because of similar properties a connection between XRFs and GRBs is 
suggested. The most popular theories say that XRFs are produced from GRBs observed ''off-axis''.
Alternativly, an increase of the baryon load within the fireball itself or low efficiency shocks can produce XRFs. 
If there is a connection between the XRFs and GRBs, they should originate at rather low redshifts (z $<$ 0.6).\\

Gamma-ray satellites react in the same way to XRFs and GRBs. In case of a detection the coordinates are distributed 
to other observatories (see section 2.1). Only from later analysis the difference can be established.

\par

In this case we include also the observation of XRFs by MAGIC in our proposal.