This paper reports the results of the second collaborative study of radioimmunological parathyroid hormone (PTH) determination in twenty-two serum samples, performed by twenty-four laboratories in the European PTH Study Group (EPSG). It demonstrates that radioimmunoassay systems can yield results that are clinically relevant in the differential diagnosis of hypercalcaemia. However, it also emphasizes, that numerical values, generated by a spectrum of different immunoassay systems cannot be compared directly. Numerical values can be compared, if the same reagents and method, provided as a kit, are used. Using their 'in house' reagents, fourteen out of twenty laboratories discriminated between low and normal iPTH serum, eleven of them between normal and marginal hyperparathyroid sera, seventeen of them between normal and grossly hyperparathyroid sera. Four laboratories agreed to a ranking order of serum samples, which was expected to cover all ranges on an average standard calibration curve, six were mistaking once, and ten two or more times. Using the kit, all participants but one discriminated between sera with low, normal, marginal and grossly elevated PTH and all agreed to the above ranking order of these sera. As far as these criteria are concerned, there was no relation between the quality of results, obtained with 'in house' reagents, and differences in antisera (but one, which was an anti human 1-34 PTH antiserum), methods of tracer preparation, and peptides used for tracer preparation. But only few laboratories, using bovine PTH standards were able to satisfy the above described conditions, whereas all of the laboratories, using human PTH standard preparations, did so with minor exceptions.