We describe an immunoblotting method for examining the electrophoretic properties of urinary Tamm-Horsfall protein. Using this method we tested frozen urine samples from a group of 37 thoroughly investigated recurrent idiopathic renal calcium stone formers and compared this with fresh urine from 19 non-stone forming laboratory staff. We found that there was a statistically significant different pattern of Tamm-Horsfall protein bands in the two sets of urines, with stone formers tending to have two bands and non-stone formers tending to have three bands. This could have been due to storage artefact and therefore a further group of 13 fresh urines from unselected renal stone formers was tested. A smaller proportion of these cases showed the two-band pattern, possibly because not all of this group were idiopathic calcium stone formers. This suggests that but does not prove that there is no significant storage artefact and that there may be an in vivo effect causing stone formation.