People with panic disorder-agoraphobia and their relatives often react anxiously to CO(2)-enriched gas mixtures. Available data are not suited to disentangle genetic from common environmental causes of familial aggregation of CO(2) reactivity, nor provide quantitative estimations of the sources of trait variation. Three-hundred-forty-six twin pairs belonging to the general population-based Norwegian NIPH Mental Health Study underwent self-assessments of anxiety and of DSM-IV panic symptoms after inhalation of a 35%CO(2)-65%O(2) mixture. Two thresholds were employed - at sample's 75th and 90th percentiles of responses - to define provoked panic attacks and to calculate polychoric correlations. Variance components were estimated by structural equation modelling (SEM). For definitions of responses based on the sum of all 13 panic symptoms, SEM could not discriminate between shared environmental versus genetic causes of familial resemblance for provoked attacks. For definitions of responses based on global anxiety, or on the sums of those symptoms (dyspnea, dizziness, palpitations) with highest variance post-CO(2), the best-fitting models indicated additive genetic factors as the sole causes for within-family resemblance. Best-fit heritability estimates ranged from 0.42 to 0.57. Genetic and idiosyncratic environmental factors explain most of individual differences in reactivity to hypercapnia. Within-family similarities for this trait are largely explained by genetic determinants.