van de Rijt-Plooij & Plooij (1992) have found periods of 'regressive behaviour' (Regression Periods), which accompany developmental transitions in infancy. In fullterm normal infants these periods occur at specific ages. The present study had two central aims. Firstly, to see if the Plooijs' finding of Regression Periods at 12, 17 and 26 weeks postpartum could be replicated. Secondly, to investigate the relationship between the length of Regression Periods for a control group of participants and a group of participants at heightened risk of developing insecure mother-infant attachment: mothers presenting with symptoms of post-natal depression. Forty-five mother-infant dyads participated in this prospective, longitudinal study. After seeing mothers at home, they were interviewed weekly, for approximately 15 weeks, about specific infant behaviours and their reactions to their infant. Following two types of manipulation of the data, Regression Periods for control group participants were detected at weeks 12, 16,20 and 24, whereas for participants in the post-natal depression group, Regression Periods were detected at weeks 14, 17 and 25, supporting the Plooijs' findings. Regression Periods were longer in the post-natal depression group. Depressed mothers were also less flexible in their mothering style, measured on the Facilitators & Regulators questionnaire. Based on the findings of this study, the development of insecure attachment is discussed. A clinical implication of this work is that information about Regression Periods could be made available to new-mothers, using Regression Period knowledge to focus on the prevention of insecure attachment.