A prospective longitudinal study assessed the development of the mother-child relationship within the context of other important aspects of transition to parenthood. In the prenatal phase, 238 women who were 8 months pregnant with their first child completed a set of questionnaires. At 1 month postpartum, 165 of these women completed a second set of questionnaires, and a subset of 86 were observed for 1 hour at home with their infants. When their child was 2 years old, 62 of the 86 mothers completed questionnaires and were observed in interaction with their toddler in the laboratory. The several hundred individual variables were reduced to internally consistent composite variables, 5 in the prenatal phase, 7 postpartum, and 14 in the toddler phase. Preliminary analyses reported in this paper assessed the structure of relationships among the composite variables within each phase (Pearson correlations) and whether there was significant prediction from one dimension to another across the phases, beyond stability in the predicted dimension (multiple regressions). Causal modeling analyses are still in progress.The results indicate that the development of maternal attachment is a gradual process and that feelings of attachment are related to a number of other measures of women's psychological well-being. Parenting confidence was found to play a central role in adaptation to motherhood within each phase, as well as from the prenatal to postpartum and postpartum to toddler phases. R!%UM& Une ktude longitudinale a kvalut le dkveloppement de la relation mbre-enfant dans le contexte d'autres aspects importants de la transition ii la condition de parents. Dans la phase prknatale, 238 femmes enceintes de huit mois de leur premier enfant ont rempli une serie de questionnaires. Un mois apres la naissance, 165 de ces femmes ont rempli une deuxitme skrie de questionnaires et un groupe de 86 mbres a ttk observe pendant une heure chez elles avec leur enfant. Quand leur enfant avait deux ans, 62 des 86 meres ont rempli des questionnaires et ont ttk observkes, en laboratoire, dans leur interaction avec leur enfant qui commence a marcher. Les centaines de variables individuelles ont t t t reduites A quelques

Mood fluctuations in women and men were studied both prospectively and retrospectively to determine whether cyclic changes occur over phases of the menstrual cycle, lunar cycle, and/or days of the week. The participants (1 5 women using oral contraceptives, 12 normally cycling women, and 15 men), who did not know the purpose of the study, recorded the pleasantness, arousal, and stability of their moods daily for 70 days (concurrent data). Later they recalled (retrospective data) their average mood for each day of the week and phase of the menstrual cycle (women only). The only evidence of mood fluctuation over the menstrual cycle in the concurrent reports was that normally cycling women reported more pleasant moods in the follicular and menstrual phase than did men and women on oral contraceptives. Women's moods fluctuated less over the menstrual cycle than over days of the week. Recollections of menstrual mood changes differed from actual changes: Women recalled more pleasant moods in the follicular phase and more unpleasant moods in the premenstrual and menstrual phases than they had reported concurrently. Bias also was evident in recollections of weekday mood fluctuations: Weekend highs were exaggerated and Monday blues were reported even though they were not reported concurrently. There was no evidence of mood fluctuations over the lunar cycle and the groups did not differ in mood stability. The retrospective reporting bias for both the This research is based on a thesis done by the first author under the co-direction of the second and third authors and submitted to the University of British Columbia in partial fulfillment of the requirements for the M.A. degree. Portions of this work were presented at the meeting of the Canadian Psychological Association, Halifax, June 1985.We gratefully acknowledge the helpful suggestions of Boris Gorzalka, Phillip Smith, Lawrence Walker, Janet Metcalfe, and James Russell.Requests for reprints should be sent to Jessica McFarlane, Psychology Department, 2136 West Mall, University of British Columbia, Vancouver, B.C., Canada V6T 1Y7.

The past decade has seen an accumulation of evidence that TV viewing influences beliefs about social reality, particularly in areas related to violence. Detailed content analyses of the portrayal of aggression on North American television were conducted in an attempt to document some of the symbolic messages in television content and the extent to which those messages vary according to program category. The TV sample analyzed consisted of 109 programs (81 hours) chosen on the basis of audience viewing figures; 76% were produced in the U.S.A. and 22% in Canada. An average of 9 acts of physical aggression and 7.8 acts of verbal aggression per program hour were observed, but both type and rate of aggression varied by program category. Other methods of conflict resolution occurred rarely, and few witnesses sought alternatives to aggression. Less than 2% of the aggression observed was accidental, and most (69%) was incidental to the plot. Aggression, especially verbal abuse, was often portrayed as humorous, and there was little evidence of consequences. Some differences between Canadian and U.S. programming were noted. The possibility of using schema theory to understand the role of television in the development of beliefs about social reality is explored.

The development of schematic concept formation was studied in adults and in first, third, and fifth graders. The task, based on the generation of two classes of complex patterns, required identifying the odd form in a set of three 8-sided polygons. Problems designed to be least, more, and most difficult were indeed so. Accuracy improved slightly between the third-and fifth-grade levels; but the results generally indicated remarkable consistency across age in types of error, concept learning over time, and strategies used. In particular, subjects at all age levels used two types of pattern cue: the general similarity between pattern and prototype and some specific physical features unrelated to class membership. Neither type of cue alone was sufficient to account for performance.

In a longitudinal prospective study, mood fluctuations were assessed for evidence of premenstrual syndrome (PMS) as well as other menstrual, day of week, and lunar cyclicity. Volunteer participants from the cornmunity (60 women and 10 men) provided daily data for 12 to 18 weeks. Significant mood fluctuation was determined by a new nonparametric method using each individual's own standard deviation as a measure of "marked" change. Cyclicity was the norm; two thirds of both the women and men had one or more menstrual or lunar phases or days of the week that were markedly positive and/or negative, relative to their own range, but few experienced stereotypical cyclicity (PMS, Monday blues, full moon). About half the women whose prospective data met conservative criteria for PMS, met liberal criteria, or met neither criteria said they had PMS, and half in each group said they did not.The supposedly "classic" mood pattern over the menstrual cycle is characterized by pleasant affect in the ovulatory phase (approximately midcycle), followed by negative affect in the premenstrual phase and men-