To test a model which describes the factors expected to predict parent and child behaviors in separations and reunions in daycare, 88 dual-career mothers (n= 54) and fathers (n= 34) were observed during interactions with their infants and toddlers in this context. Simultaneous regression analyses and path analyses provided mixed support for the model. In general, separation behaviors were predicted better by the model than reunion behaviors. At separation, parent sensitivity was predicted by parent gender and increased levels of child distress was predicted by lower levels of parental involvement in child-care and increased parental separation anxiety. Parents who were less involved in their children's daily care had children who were more distressed at separation as did parents who were anxious about the effect of separation. At reunion, parent sensitivity was predicted by separation anxiety. Parents who were more anxious about employment-related separations were more sensitive in their interactions with their children. Child happiness at reunion was not predicted by any variables included in the model. There were mean differences between mothers and fathers on a number of variables including parent sensitivity (mothers were more sensitive), involvement in child-care (mothers were more involved), and anxiety about employment-related separations (mothers were more anxious). There were no significant differences in regression coefficients between mothers and fathers suggesting that the model held equally well for mothers and fathers. The results are discussed in terms of their support for the model and attachment theory.