Discussion

Abstract

Dr P. McL. D. D uff congratulated the author on his presentation of a most interesting paper and said that, in general, he welcomed all attempts to describe cyclic sedimentation in an objective manner. He would, however, appreciate clarification of certain basic points of the technique used. In at least one of the data matrices he noticed that shale never 'changed' into shale whereas limestone 'changed' into limestone. In more than one case sandstone 'changed' into sandstone. How did Dr Selley decide when one rock type followed itself? If there were differences in the sandstones they could not all be assigned to the same 'state' and presumably extra rows and columns should appear in the matrix corresponding to the different types of sandstones. If on the other hand the same sort of sandstone followed itself, what was the evidence for this and how did the criterion used compare in geological significance with a change from one rock type to another? The A uthor agreed with Dr Duff that the problem of recognizing multi-storey sequences in shales was difficult, if not sometimes impossible. In sandstone and limestone sections, however, bedding planes were generally present, and these could be used to distinguish both similar and dissimilar units. Geologists who were uncertain of how to recognize bedding planes in the field could construct a transition matrix using a fixed sampling interval. As discussed in the text of the paper, however, this procedure was suspect since the operator could alter the transitional probabilities at will by