BACKGROUND There is considerable uncertainty in the current literature about the relationship between personality dimensions and affective temperaments. METHODS We compared-in a non-ill 14-26-year-old Italian student population of 1010-the affective temperaments of classic psychiatry conceived as subaffective traits [and measured through the Temperament Assessment of Memphis, Pisa, Paris and San Diego-Interview Version (TEMPS-I) in the Akiskal and Mallya Operationalization] with Cloninger's revised Tridimensional Personality Questionnaire (TPQ) deriving from the experimental psychology tradition. RESULTS The Depressive Temperament (DT) and Harm Avoidance (HA), loaded positively on the same canonical variate, whereas the hyperthymic (HT) strongly, and Novelty Seeking (NS) moderately, loaded negatively. In contrast, the Cyclothymic Temperament (CT) loaded highly positively on a second variate, on which both Novelty Seeking strongly and Harm Avoidance moderately loaded positively. Reward Dependence (RD), Persistence (P), and Irritable Temperament (IT) did not significantly relate to any temperamental and personality constructs. At a subdimensional level of TPQ 'shyness with strangers', 'stoic rigidity', 'detachment', 'fear of uncertainty', 'reflection', and 'anticipatory worry' correlated best with the DT. 'Gregariousness', 'exploratory excitability', 'uninhibited optimism', 'attachment', 'confidence', 'extravagance', 'independence', 'vigor', and 'impulsiveness' correlated best with HT. Lastly, 'anticipatory worry', 'disorderliness', 'sentimentality', and 'fatigability' correlated best with CT. CONCLUSIONS The data provide concurrent validity to TEMPS-I and, as earlier suggested by Cloninger, indicate that (as expected) high HA and DT are related. High NS is both related to the HT and CT, and (somewhat unexpectedly), the CT is related to high HA. In a more theoretical vein, hyperthymic-novelty seeker can be predicted to be overrepresented among those with high achievement; on the other hand, a moody, restless disposition (a cyclothymic-harm avoidant type) may engage in outrageous behavior and be liable to negative affective arousal. We submit that these considerations could shed some light on the origin of socially adaptive behavior ('sunny' or sanguine types) on the one hand, and borderline conditions, anxious-hostile bipolarity ('dark' types) on the other.