Fig. 1. We employ two kinds of schematics for representing phase-space dynamics. In the first (Fig. 1, Fig. 2 and Fig. 3) cells or D-conditions are represented by flat regions and flow by arrows. In the second (Fig. 4 and Fig. 5) cells or D-conditions are represented by darkened ovals and the phase flow by expanding conical shaded regions. The shaded region to the right of a condition represents the expansion of phase fluid from the condition; the region to the left, the contraction of phase fluid to the condition, or equivalently, the expansion from the condition under the anti-flow. In either representation a dashed path represents the development of an individual system.

Fig. 2. An individual phase point can behave t-asymmetrically by traveling through progressively larger regions. All or most of the fluid in a smaller region can flow into a larger region, but only the tiniest fraction of a larger region can flow into a smaller one. There results an asymmetry of explanation: a low entropy condition can account for a high entropy one but not vice versa. Moreover, for the case of entropy increasing toward the future, this means past conditions explain future ones but not vice versa.

Fig. 3. C is all ways the rock can fly at the window. E, the coarse covering of , contains the possible effects of C plus all states indistinguishable from these effects. A sub-region of E that lies in , such as M has to be exceedingly narrow.

Fig. 4. Only with a prior condition such as P can we narrow down the superabundant possible causes within .

Fig. 5. Because of the t-symmetry of the phase flow, the bundle of trajectories passing through D-condition contracts to enter A then expands after leaving it.