British South Asians (with ancestry from the Indian subcontinent) provided a puzzling exception to the British class gradient in mortality during the 1970s. On the assumption that class gradients in health are produced mainly by gradients in standard of living, this might be due to a break in the relation of class to standard of living (change in class structure), or by a break in the relation of standard of living to patterns of health behaviour and health risk (change in class lifestyles). Data on these characteristics are available from the West of Scotland Twenty-07 Study, where 159 South Asians aged 30-40 (mean age 35) were sampled alongside 319 of the general population in Glasgow. As regards changes in class structure, results indicate that the underclass thesis, which suggests that ethnic minorities are forced into less eligible jobs or into a separate labour market or into unemployment, resulting in a standard of living below that of the general population, still holds good for British South Asians in categories from social class III non-manual downwards. It does not hold good for owners of small businesses, where Sikhs and Hindus in particular have a standard of living equivalent to general population counterparts. However, prosperity is not predictable from levels of education in the subcontinent and from this and other signs it appears that a wholesale redistribution of class chances is occurring among British South Asians, disrupting inter-and intra-generational continuities in the relation between class and standard of living. There is little sign of change in class lifestyles, i.e. in the relation between standard of living and health behaviour or health risk. As yet, though, the new distribution of standard of living is affecting patterns of health behaviour and health risk more strongly than symptom experience or chronic illness, suggesting that a class gradient in health will re-emerge.