We investigate the possibility of detecting HI emission from gravitationally lensed HI clouds ( akin to damped Lyman- \alpha clouds ) at high redshift by carrying out deep radio observations in the fields of known cluster lenses .
Such observations will be possible with present radio telescopes only if the lens substantially magnifies the flux of the HI emission .
While at present this holds the only possibility of detecting the HI emission from such clouds , it has the disadvantage of being restricted to clouds that lie very close to the caustics of the lens .
We find that observations at a detection threshold of 50 \mu { Jy } at 320 MHz ( possible with the GMRT ) have a greater than 20 \% probability of detecting an HI cloud in the field of a cluster , provided the clouds have HI masses in the range 5 \times 10 ^ { 8 } { M } _ { \odot } \leq M _ { HI } \leq 2.5 \times 10 ^ { 10 } { M% } _ { \odot } .
The probability of detecting a cloud increases if they have larger HI masses , except in the cases where the number of HI clouds in the cluster field becomes very small .
The probability of a detection at 610 { MHz } and 233 { MHz } is comparable to that at 320 { MHz } , though a definitive statement is difficult owing to uncertainties in the HI content at the redshifts corresponding to these frequencies .
Observations at a detection threshold of 2 \mu { Jy } ( possible in the future with the SKA ) are expected to detect a few HI clouds in the field of every cluster provided the clouds have HI masses in the range 2 \times 10 ^ { 7 } { M } _ { \odot } \leq M _ { HI } \leq 10 ^ { 9 } { M } _ { \odot } .
Even if such observations do not result in the detection of HI clouds , they will be able to put useful constraints on the HI content of the clouds .