We discuss ROSAT and ASCA observations of the young active star Gl 355 .
During the ROSAT observation a strong flare was detected with a peak flux more than an order of magnitude larger than the quiescent level .
Spectral analysis of the data allows us to study the temperature and emission measure distribution , and the coronal metal abundance , for the quiescent phase and , in the case of ROSAT , also during the evolution of the flare .
The global coronal metallicity Z / Z _ { \odot } \sim 0.1 derived from both ROSAT and ASCA data is much lower than solar and presumably also much lower than the photospheric abundance expected for this very young star .
The temperature structure of the quiescent corona was about the same during the various observations , with a cooler component at T _ { 1 } \sim 7 MK and a hotter component ( to which only ASCA was sensitive ) at T _ { 2 } \sim 20 MK .
During the flare , the low temperature component remained approximately constant and equal to the quiescent value , while the high-temperature component was the only one that varied .
We have modeled the flare with the hydrodynamic-decay sustained-heating approach of Reale at al .
( 1997 ) and we have derived a loop semi–length of the order of \sim 1.5 stellar radii , i.e .
much larger than the dimensions of flares on the Sun , but comparable with the typical dimensions inferred for other stellar flares .
We have compared the derived loop size with that estimated with a simpler ( but physically inconsistent ) approach , finding that for this , as well for several other stellar flares , the two methods give comparable loop sizes .
Possible causes and consequences of this result are discussed .