2 As is true of much of the work reported here, the preceeding paragraph owes much to many discussions with Michael Humphreys whose contribution is gratefully acknowledged.
 Readers familar with the the work of Atkinson and Birch (1970) will no doubt recognize that this emphasis upon rates of change is adapted from the their theory of the dynamics of action. (See also Revelle, 1986).
 Venables (1984) elaborates on the difficulties of taking a within-subjects construct and making inferences between subjects.
For those who find their favorite personality trait not listed in the three-seven dimensions of the taxonomists, it is necessary to remember that the dimensional space defined by these primary dimensions can include any number of combinations of the primaries. (In fact, a major difference between the American and European taxonomists is the emphasis in Europe of searching for causal bases of the dimensions as a way of resolving the indeterminacy of rotation problem that besets the American taxonomies. Consider, for example, the debate between Eysenck (1981, 1987) and Gray (1981, 1987) as to the biological signifigance of introversion-extraversion and neuroticism-stability versus impulsivity and anxiety.) While perhaps not appealing to all, such a dimensional framework does provide more coherence to theory development than does the alphabetical arrangement of traits favored by some (e.g., London and Exner, 1978) or the historical arrangement favored by most undergraduate texts.
 The interested reader should see the very thorough review by Stelmack, 1990, of basal and phasic arousal differences as they relate to introversion.
 In a personal communication, Robert Crowder has pointed out that this hypothesis may be understood in terms of the stimulus sampling theory of Estes (1959) as it is applied to the problem of spacing repetition effects. See Crowder (1976, pp. 277-300).