The recent article by R. P. Saller on Roman dowry in the Principate makes some interesting and important suggestions about the function of dowry and its role in the devolution of property. I am in broad agreement with a good deal of what he says, and would not dispute his views that dowry was, as shown by the requirement of collatio dotis, regarded as in a sense part of a woman's patrimony, and that the rules for the recovery of dowry show that the purpose of giving dowry was not held to rest on one single principle, but included provision both for the expenses of the wife's maintenance during marriage and for a possible remarriage after divorce or widowhood. However, his remarks on both points need some qualification and amplification. Briefly, I hope to show (i) that the oddities and anomalies noticed by Saller in the rules governing the recovery of dowry at the end of a marriage are apparent rather than real, since these rules rest, not on conflicting views about the purpose of dowry, but on the fact that the husband had full legal ownership of the dowry during marriage, together with the right of the wife or her pater to an actio rei uxoriae for recovery of dowry; (ii) that the rules for collatio dotis applied only if the woman herself chose to claim a share in her father's estate on intestacy beyond the amount of her dowry; (iii) that the use of the dowry for the wife's support was an equitable, rather than a legal, requirement.