The writer has established reference intervals in the subsurface on the basis of apparent systematic interlayer water loss of swelling clay minerals. The intervals are used in much the same manner as the familiar indicators for metamorphism, but are present at sufficiently shallow depths to be evident within oil-bearing strata of the Gulf Coast. The resulting conclusion is that clay-mineral diagenesis indicators may prove to be important petroleum-evaluation markers as well as fundamental properties of sedimentary basins.

Sedimentary basins are viewed as combinations of gases, liquids, and semisolids distributed through a solid matrix. During geologic development the interstitial components segregate by migration and produce various commercially exploitable concentrations. Water, the principal fluid component of the sedimentary section, is thought to migrate in three separate stages. Initially, pore water and excessive (more than two) clay-water interlayers are removed by the action of overburden pressure. This initial water flow (which is essentially completed after the first few thousand feet of burial) reduces the water content of the sediment to about 30 percent, most of which is in the semi-solid interlayer form. A second stage of dehydration is thought to occur when the heat absorbed by the burie sediment becomes sufficiently great to mobilize the next-to-last water interlayer in an M(H₂O)_x + ^DgrH_r = 1 + XH₂O fashion. The final stage of sediment dehydration which extracts the last remaining water monolayer from clay lattices is apparently very slow, even by geologic standards, requiring tens or possibly hundreds of millions of years depending upon the geothermal and burial history of the sediment.

Petroleum hydrocarbons which are distributed throughout the matrices of potential source beds in normal frequencies of 300-3,000 ppm are thought to be too sparse to initiate continuous fluid flow. In normal marine sediments, however, the water associated with clay minerals is present to a considerable depth in the order of 200,000 ppm, and therefore it is reasoned that this phase forms the connection between petroleum source and reservoir beds.

The first and last dehydration stages are probably unimportant in Gulf Coast oil migration, inasmuch as they occur, respectively, at levels too shallow and too deep to intersect the interval of maximum liquid petroleum availability. The amount of water in movement during the second stage, at a level which does intersect this interval, is 10-15 percent of the compacted bulk volume and represents a significant fluid displacement capable of redistributing any mobile subsurface component. A measure of the degree to which the second-stage interlayer water has been discharged into the system can be noted on X-ray diffractograms. The movement appears to occur in a relatively restricted, depth-dependent temperature zone in which the average dehydration temperature of the points measured is 22 °F. With the use of an empirically derived P/T curve and a geothermal-gradient map, a set of regional subsurface dehydration contours can be constructed. A plot of 5,368 liquid petroleum production depths referenced to this dehydration "surface" shows an almost perfect Gaussian distribution. It seems significant that, although the dehydration depths range from 4,000 to 16,000 ft, hydrocarbon production depths are distributed in a statistically consistent relation to the calculated clay-dehydration contour surface.