In this chapter we described the constructs of temperament and attachment and have discussed similarities and differences between the two. We addressed the issue of whether temperament contributes to overall attachment security or to the specific type of attachment that children display. We conclude that although temperament may influence the type of secure and insecure attachment relationship children form with their parent, temperament alone will not determine if a child is classified as securely or insecurely attached. We presented evidence suggesting that certain dimensions of temperament, specifically negative emotionality, may be associated with infants' behavior during the Strange Situation, such as proneness-to-distress during separations. However, we noted that these temperament dimensions do not predict overall security of attachment. It is likely that although no single temperament characteristic, such as proneness-to-distress, in and of itself determines overall attachment security, it is possible that a constellation of temperament characteristics may be more strongly related to attachment security. The examination of constellations of temperament characteristics may be particularly useful for furthering our understanding of individual differences within attachment classifications. Such an approach may elucidate the reasons why infants are classified into one subgroup of secure, insecure-avoidant, or insecure-resistant attachment versus another subgroup. Furthermore, we suggest that the collection of findings regarding temperament and attachment not only underscores the importance of a transactional approach to early social-emotional development, but emphasizes that temperament and attachment can make unique and interactive contributions to children's social-emotional functioning. That is, the goodness-of-fit between infant and parent characteristics may best predict security of attachment. Although child characteristics clearly contribute to the development of the parent-child relationship, we believe that the effects of infant temperament on infant-caregiver attachment may well be indirect, and may be moderated by such variables as maternal personality and social support. Thus, taken together, a growing literature clearly indicates that although temperament and attachment security are interrelated, they are by no means interchangeable constructs. To return to our guiding question, "Temperament and attachment: One construct or two?" We reply, "two."