Excerpts from Origin of the Carolina Bays

The following text is from Douglas Johnson's work as presented in his book "The Origin of the Carolina Bays", Volume IV of Columbia University's Columba Geomorphic Studies series, pp 116 - 120.

Johnson is using these arguments to dismiss the meteorite impact theory. We see his observations as supporting our ejecta blanket hypothesis, where the bay depressions are imperfections in the surface of the blanket created during emplacement. The entire bulk of the sand stratum in which the bays are contained is considered to be ejecta by us.

Dr. Johnson’s observations regarding the passage of fluvial channels into and across Carolina bays seems to us to be predicated on a short lifespan of these landforms. Considering our hypothesis that they are indeed quite old (800,000 years), there is adequate time for normal headward erosion to breach into a bay and eviscerate it’s interior.


Were the Carolina craters the product of meteoritic impact, the Coastal Plain beds exposed in the sides of the craters should show the effects of violent impact of a great meteoritic mass, or of violent explosion associated with such impact. It is scarcely conceivable that so catastrophic an event would leave sedimentary beds in the crater walls quite undisturbed. We should expect to find them badly shattered with fragments irregularly disposed, or turned upward as in Meteor Crater, Arizona, and the craters on the island of Osel. It appears to be well established, however, that the Coastal Plain beds are normally undisturbed about the Carolina bays. True, in a majority of bays the surrounding beds are not exposed at all; but exposures occur in so many that if disturbance were common the fact would be readily apparent. So far as the writer is aware not a single case of disturbance has thus far been reported. On the other hand, many instances of horizontal stratification or of horizontal plain surface have been observed beneath sand rims in the infacing walls of steep-sided bays, in road cuts trenching such walls, and in drainage ditches cut through bordering rims and several feet into the underlying Coastal Plain formations.” Melton noted that “beds of Coastal Plain sediments in a number of cases seem to lie flat beneath the bay rims." He recognized that this fact imposed a burden on the meteoritic hypothesis, for he added: “Should this condition prove to be common it would be necessary, in order to retain the meteoritic hypothesis, to postulate Widespread reduction of the original surface through marine planation.” We have just seen that the condition is common; and we have earlier shown that the conception of widespread planation beneath the sea, with later elevation of the bays and their rims through the zone of destructive shore processes, is untenable. We can only conclude that the common occurrence of undisturbed Coastal Plain beds bordering the bays is a fact difficult to explain on the basis of the meteoritic hypothesis.


If the Carolina bays are meteorite craters, they might well contain lakes, or might formerly have contained them. But such lakes should not commonly have surface outlets or form part of any integrated surface drainage system. The reasoning which underlies this deduction from the meteoritic hypothesis is simple. Prior to impact, the groundwater level and tl1e surface drainage of the Coastal Plain must have been established. Wherever a meteoritic mass excavated a crater the bottom of which extended below groundwater level, water would stand in the depression to form a lake. But we should not expect the water to rise above groundwater level and overflow the crater rim. Normally tl1e craters would have neither incoming nor outflowing surface streams. Even if a meteorite struck in a preexisting stream course, we should expect formation of the crater to deflect the stream. Crater lakes are relatively common phenomena in SOIHC volcanic regions, but surface streams draining such lakes are rare. We are accustomed to finding crater lakes, sinkhole lakes, kettlehole lakes of glacial regions, and similar water bodies without numerous integrated drainage connections. A remarkable feature of the Carolina bays, of which no account has previously been taken, is the frequency of channels affording ingress or egress to streams which now or formerly drained into or out of the bays. Such channels are beautifully shown in Figures 5, 15, 17, 32, 40, and 42, and less clearly in the case of many other bays figured in this volume. They are sometimes represented, although not always accurately, on soil maps.(13) Topographic maps show outlet streams in many cases and occasionally reveal a well-integrated drainage system draining into and out of a succession of bays.(14) Although there is little surface drainage into White Lake and Suggs Mill Pond (see Bladen County soil map, North Carolina), both these bays have outlet channels carrying considerable volumes of water. In both cases the natives report large springs “boiling” or “fountaining" on the lake floor due to strong upwelling of water from below. These two examples suggest that similar conditions formerly existed in many bays which do not now contain lakes, and that outflow of water supplied by bottom springs carved the channels so frequently observed leading outward from the craterlike depressions. The large bay north of Pinewood (Fig. 17) contains no large lake, but its outlet channel was long ago dammed to impound artificially the water entering the basin. Later the dam was destroyed, and at present a stream of sufficient volume to serve as the home of fish six or eight inches long flows freely from the great bay.” The channel traversing the rim of the large bay near Wilmington, North Carolina (15), is particularly broad and sharply carved in places. In many cases outlet channels, while clearly visible, are partly or completely obstructed by accumulations of rim sand. Doubtless many former channels have thus been wholly obliterated. Many bays have two or more outlets, draining either into adjacent streams or into neighboring bays (Fig. 4).

Whatever the cause of the Carolina craters, it evidently produced a vast number of basins into which or out of which water flowed freely through channels still visible and often still functioning. Because such drainage channels are not commonly found in known meteorite craters or, for that matter, in craters produced by volcanic eruptions, artillery fire, or other violent explosions or impacts, it seems doubtful whether the channeled craters of the Atlantic Coastal Plain can reasonably be attributed to meteoritic impact, whether of the gouging or explosive type. To account for the observed facts by the meteoritic hypothesis we would have to assume that in every case where a crater now possesses, or formerly possessed, one or more outlet channels, the meteorite excavated a depression deep enough to tap artesian water under sufficient pressure to cause surface outflow. Where inlet channels are found, we would have to assume further that impact produced no rims which could impede the flow of surface water into the craters.


If we admit, for the moment, that meteoritic impact could somehow produce the observed craters and their rims, and that peculiar geological conditions could cause waters to rise in these craters and overflow their borders, it is obvious that such outflow must take place where rims or crater borders were lowest. The water could not overflow where rim accumulation was at a maximum so long as other parts of the rim, or portions of the crater wall devoid of any rim, offered a lower outlet. Studies of aerial photographs (Figs. 4, 5, 15, and 17) suggest, and field examinations confirm, that outlet channels repeatedly traverse rim barriers. The writer was unable to run lines of level about the bays, so cannot offer definitive proof that the rim barriers where traversed by outlets are higher than other parts of bay borders. But the relations of bay borders and outlets to the marshy surfaces of the bays leave little doubt that, were the outlet channels filled to the surface level of the rim on either side, the position of the outlet would in a number of cases be materially higher than other parts of the bay borders. This indicates that the outlets are, in fact, antecedent streams, having begun to function before the rim came into existence, and having continued to flow while the rim on either side was built up to its present altitude. If this be true, the craters as well as their outlets are older than the rims, and the hypothesis which makes both craters and rims the simultaneous results of meteoritic impact appears incompetent to explain the relations actually observed.

(12) Examples of bays revealing exposures of undisturbed Coastal Plain beds are: the bay northeast of McBride Church, Hilltonia quadrangle, Ga.-S.C.; bay one mile northwest of Eureka Springs, Oliver quadrangle, Ga.; bay southeast of Reynold, and large bay just east of Blackville, both on Williston quadrangle, S.C. (in last case apparent stratification of beds at one point dips faintly downhill toward bay, as if slight slumping might have occurred); compound bay with medial sand rim east of Govan, Olar quadrangle, S.C.; Dial Bay, Mayesville quadrangle, S.C.; Devil`s Woodyard Bay near Springerville, north of Darlington, S.C.

(13) See, for example, the Bladen and Cumberland County soil maps, N_C.; Barnwell, Williamsburg, and Florence County soil maps, S.C.

(14) See, for example: Alligator Bay and Doussoss Bay, Olar quadrangle, S.C.; Saint George Church bay, Rowesville bay, Black Bay, and others, Orangeburg quadrangle, S.C.; the bay east of West Middle School, Wadboo Swamp bay, Polk Swamp bay, and the bays drained by Sandy Creek, Bowman quadrangle, S.C.; Big Junkyard Bay, Guys Branch Bay, Islanded Bay, and others, Manning quadrangle, S.C.

(15) Girard Wheeler, manuscript report on observations made at the writer's request.