Since the introduction of the measles, mumps and rubella (MMR) vaccine in Sweden in 1982, a yearly evaluation of the immunity pattern and seroconversion rate of 12-year-old children has been carried out. To be able to realize large-scale, follow-up studies, techniques have to be used which are not too labour-intensive and time-consuming but are sensitive enough to detect past immunity and post-vaccination titres. We report tests with haemolysis-in-gel (HIG) technique and an enzyme-linked, immunosorbent assay (ELISA) compared with neutralization (NT) for the detection of mumps antibodies. This study comprises 226 paired serum samples obtained between 1985 and 1989. One hundred and forty-one of the paired samples had been selected because they had given negative or borderline readings, using HIG technique. The remaining samples were consecutive pre- and post-vaccination sera obtained in 1989 from 85 vaccinees from one area in Sweden. HIG technique gave both false positive and false negative readings, compared with NT as also the false positive sera were detected. Non-inactivated sera could not be used in the NT test against mumps virus, owing to unspecific NT reactions. No differences between non-inactivated and inactivated sera by NT were seen, as against other paramyxoviruses. Cross-reactivity between mumps and parainfluenzae in NT tests was not demonstrated. The ELISA test proved more reliable and specific than HIG and was more sensitive than NT. Some post-vaccination sera from vaccinees who failed to seroconvert by NT contained high levels of mumps antibody detectable by the ELISA.(ABSTRACT TRUNCATED AT 250 WORDS)