Abstract: Studies of parental investment in mammals have frequently used suckling behaviour to estimate energy transfer from mother to offspring, and consequently to measure maternal input. Such studies assume that the more an offspring sucks, the more milk it will receive. This assumption has been questioned, and a review of the literature found little support for it. To test if suckling behaviour provided an accurate index of milk or energy intake we used a radioactive isotope technique to label the milk of thoroughbred mares and to measure milk transfer to foals. We found no significant linear relationship between usual measures of suckling behaviour and milk or energy intake. No behaviours associated with suckling nor with characteristics of mares and foals improved the relationship; only the number of butts associated with each suck episode even approached significance. If we had used suckling behaviour to test theories on differential maternal investment our conclusions would have been in error. For example, female foals tended to suck for longer than males did but there was no difference in the amount of milk transferred. Consequently, we show that measures of suckling behaviour do not adequately predict milk intake in the domestic horse and we suggest that conclusions about differential maternal investment in mammals based on suckling behaviour are likely to be in error. Copyright 1999 The Association for the Study of Animal Behaviour.