This analysis uses prefecture-level data on deaths by cause and marria
ge type in Japan to test hypotheses which relate the 'arranged marriag
e' system to cause-specific mortality patterns among single Japanese m
en and women. The results from this analysis, combined with earlier fi
ndings, confirm the importance of the mate selection process in produc
ing atypically high and rapidly declining mortality rates among Japane
se singles, and suggest that the presence of tuberculosis, along with
several other diseases, was an important component of the screening pr
ocess for potential spouses. The findings also highlight the difficult
ies of identifying marriage selection mechanisms in industrialized soc
ieties from cause of death data.