4 current approaches to understanding temperament are discussed in the roundtable. In an introductory overview, Goldsmith outlines some of the major convergences and divergences in the understanding of this concept. Theorists representing 4 positions--Goldsmith, Buss and Plomin, Rothbart, and Thomas and Chess--outline their views by responding to each of 6 questions: How do you define temperament and explain the boundaries of the concept? What are the elements of temperatment? How does the construct of temperament permit you to approach issues or organize data in ways that are possible only if this construct is invoked? How does temperament develop? To what extent do you consider temperament to be a personological versus a relational or an interactional construct? and How does your approach deal with issues of temperamental "difficulty"? In 2 commentaries on the theorists' answers, Hinde highlights differences among their positions and indicates issues that current theories of temperament must take into consideration, and McCall draws on common aspects to propose a synthesizing definition that draws on all 4 approaches.

Using samples of twins and singletons totaling 715 individuals, the authors document heritable influences on various temperamental dimensions during the toddler and preschooler age ranges, which have been somewhat understudied relative to infants and older adolescents. In contrast to instruments on which prior literature is based, the Toddler Behavior Assessment Questionnaire and the Children's Behavior Questionnaire offer assessment of positive affectivity (separately from negative affectivity) and of emotional regulation. Positive affect reveals substantial shared environmental influence, and emotion regulation reveals additive genetic influence. Evidence for genetic variance in temperament is strengthened because intraclass correlations from many of these questionnaire scales show no evidence of "too-low" dizygotic correlations that imply contrast effects. Suggestive evidence is offered that psychometric characteristics of the questionnaires can affect biometric inferences.

No abstract

4 current approaches to understanding temperament are discussed in the roundtable. In an introductory overview, Goldsmith outlines some of the major convergences and divergences in the understanding of this concept. Theorists representing 4 positions--Goldsmith, Buss and Plomin, Rothbart, and Thomas and Chess--outline their views by responding to each of 6 questions: How do you define temperament and explain the boundaries of the concept? What are the elements of temperatment? How does the construct of temperament permit you to approach issues or organize data in ways that are possible only if this construct is invoked? How does temperament develop? To what extent do you consider temperament to be a personological versus a relational or an interactional construct? and How does your approach deal with issues of temperamental "difficulty"? In 2 commentaries on the theorists' answers, Hinde highlights differences among their positions and indicates issues that current theories of temperament must take into consideration, and McCall draws on common aspects to propose a synthesizing definition that draws on all 4 approaches.

Because of the possible implications of infant-mother attachment for later adjustment, we examined the extent to which it could be predicted by mother interactional variables and infant proneness to distress. The meta-analysis demonstrated that sensitive, responsive maternal interaction predicted the security of attachment in Ainsworth and Wittig's (1969) "strange situation." However, the strength of the relation was less than many narrative reviews have suggested. Proneness to distress, which is conceptualized as a temperamental variable, predicted resistance, which is a behavioral pattern in the strange situation that is thought to indicate one variety of insecure attachment. The strength of this association was low but was roughly comparable to that in the maternal domain. In both the maternal and infant domains, we attempted to predict the effect sizes by study characteristics such as method of assessment, sample composition, and age of subjects at the time of study. The pattern of results highlights several continuing methodological problems in the field and suggests that additional explanantory concepts are needed.Variation in maternal caregiving and infant temperament are two potential predictors of the quality of the mother-infant relationship in early life. Our review of these issues is motivated, in part, by the widespread belief that mother-infant attachment influences later adaptation, including the development of behavioral disorders (Sroufe, 1983). Although the cumulative evidence for the effect of variation in attachment quality on psychopathology is still inconclusive, serious anomalies of attachment that may result from maltreatment foreshadow poor outcomes (e.g., Cicchetti & Rizley, 1981; Oaensbauer & Harmon, 1982;Schneider-Rosen & Cicchetti, 1984). Study of the typical range of variation in attachment has yielded enough suggestive evidence to justify a volume called the Clinical Implications of Attachment (Belsky & Nezworski, in press). Although we must avoid uncritical attribution of later outcomes to early developmental influences (Rutter, 1981), the organizational construct of attachment is a potentially crucial determinant of later personality, both normal and disordered. Another potential determinant is the multidimensional construct of infant temperament, which also may influence later behavioral disorders (Bates, 1987;Chess & Thomas, 1984).Because these relationships are potentially relevant to later functioning, investigators have attempted to study the origins of We appreciate the efforts of many investigators who facilitated this report by sending us additional data from their studies. We also thank Loretta Rieser-Danner, who helped initiate the project. H. H.

Although several studies have examined anterior asymmetric brain electrical activity and cortisol in infants, children, and adults, the direct association between asymmetry and cortisol has not systematically been reported. In nonhuman primates, greater relative right anterior activation has been associated with higher cortisol levels. The current study examines the relation between frontal electroencephalographic (EEG) asymmetry and cortisol (basal and reactive) and withdrawal-related behaviors (fear and sadness) in 6-month-old infants. As predicted, the authors found that higher basal and reactive cortisol levels were associated with extreme right EEG asymmetry. EEG during the withdrawal-negative affect task was associated with fear and sadness behaviors. Results are interpreted in the context of the previous primate work, and some putative mechanisms are discussed.

The putative association between fear-related behaviors and peripheral sympathetic and neuroendocrine reactivity has not been replicated consistently. This inconsistency was addressed in a reexamination of the characterization of children with extreme fearful reactions by focusing on the match between distress behaviors and the eliciting context. Eighty 24-month-old children were observed in 4 mildly threatening contexts, and the relations among different measures of fear-related behaviors, reactive and basal cortisol levels, and baseline cardiac measures of heart rate, respiratory sinus arrhythmia, and preejection period (PEP) were examined. The hypothesis that only behaviors under the less threatening context would be associated with higher cortisol and sympathetic cardiac activity was confirmed; only task-specific freezing behavior predicted higher reactive and basal cortisol levels and resting PEP measured 1 week later. Implications for the conceptualization of dysregulated fear behaviors in the classification of extremely fearful children are discussed.

After screening 368 toddlers and selecting 77 into extremely inhibited, extremely uninhibited, and intermediate groups, 63 children (82%) were followed up at 4 and 7 years. Minority subgroups of both the inhibited and uninhibited children showed continuity on outcomes consisting of questionnaire measures of shyness, inhibitory control, and impulsivity, as well as multiepisode observational measures of behavioral inhibition and exuberance. Change from both inhibited and uninhibited status from the toddler age was more common than remaining extremely inhibited or uninhibited, but that change was largely constrained to the middle of the distribution.