Recent research has demonstrated that mortality patterns by marital status in Japan are different from corresponding patterns in other industrialized countries. Most notably, the magnitude of the excess mortality experienced by single Japanese has been staggering. For example, estimates of life expectancy for the mid-1900s indicate that single Japanese men and women had life expectancies between 15 and 20 years lower than their married counterparts. In addition, gender differences among single Japanese have been smaller than elsewhere, while those among divorced persons have been unanticipatedly large; and, the excess mortality of the Japanese single population has been decreasing over the past few decades in contrast to generally increasing differentials elsewhere. In this paper, we use a variety of data sources to explore several explanations for these unique mortality patterns in Japan. Undeniably, the traditional Japanese system of arranged marriages makes the process of selecting a spouse a significant factor. Evidence from anthropological studies and attitudinal surveys indicates that marriage is likely to have been and probably continues to be more selective with regard to underlying health characteristics in Japan than in other industrialized countries. However, causal explanations related to the importance of marriage and the family in Japanese society may also be responsible for the relatively high mortality experienced by singles and by divorced men.