It is self-evident that we fall asleep more readily when lying down than when standing up. Nevertheless, the influence of this and more subtle changes in posture, activity and the situation in which sleep propensity is measured have been largely ignored. The term somnificity is introduced here, defined as the general characteristic of a posture, activity and situation that reflects its capacity to facilitate sleep-onset in a majority of subjects. The relative somnificities of different activities and situations in the Epworth Sleepiness Scale (ESS) were investigated in 23 groups, involving 2802 subjects from seven different countries. The means of the different ESS item-scores were ranked from highest to lowest in each group. There was a high concordance (Kendall's C=0.84, P < 0.0001) among these ranks for all groups, whether of normal subjects or patients with sleep disorders, regardless of age, sex, or average sleepiness in daily life assessed by total Epworth scores. The ESS item-ranks formed an ordinal scale of somnificities with five different levels. Analysis of raw ESS item-scores for all 987 individual Australian subjects showed the same pattern of somnificities with six different levels, but with a lower concordance (C=0.39, P < 0.0001). This was probably because of subject × situation-specific interactions that were averaged within groups. A conceptual model of sleepiness is outlined that includes interactions between separate sleep and wake drives as a possible way of including behavioural and situational influences on sleep propensity.