Sedimentary sequences show statistical evidence of dependence between successive lithologies; no truly random sequence has been reported. Explicit, although idealized, models were examined to determine the sensitivity of chi-squared tests to both random and dependent components of a depositional system. A series of simple models with the following characteristics were tested.

P is the probability of transiting from one state to the next in an ideal system, q is the probability of returning (feedback) to an initiating state, and r is the probability of a random transition to any state (including that just exited). P + q + r = 1.0. While varying P, q, and r, I simulated the performance of the system and tested the resulting frequency matrix against random models. P, that part attributed to a "cyclic" process, can be very small (about 0.1 or 0.2) and still result in significant differences from the random model. Even small contributions from a dependent process, superimposed on a largely random process, can cause the appearance of great order.

Permian Salado Formation beds exhibit values for P in the general range of 0.1-0.3 and, both visually and statistically, attain considerable order.

End_of_Article - Last_Page 519------------