etation, slowly advances seaward by the consolidation of peaty growth and by trapping finegrained inorganic sediment (Vaughan 1909). Shells are added by the accumulation of small species of Mollusca on the roots (Davis 1940). It may be assumed that the maximum of accumulation of marl, clay, and silt takes place always somewhat forward, that is, gulfward, in the slowly advancing swamp. Storm waters and tidal oscillations combine to permit the up-building of the accumulating swamp materials by these inorganic sediments which, in turn, may promote, at and above high tide levels, a denser undergrowth of the less aquatic plants.

As the zone of maximum arrest and accumulation of inorganic sediment advances Gulfward, lack of accretion or a decrease in rate of accretion in the zone nearer the original mainland permits normal compaction of peat and marl to show in an invasion of groundwater and brackish marine waters. The swamp is abundantly penetrated all along the western peninsula coast by transverse tidal scour channels, permitting Gulf waters to enter the rear zone.

If the foregoing processes and results depict the true history of the formation of the mangrove barrier ridge and lagoon on the western coast of peninsular Florida during the stillstand for the 3,000 to 5,000 years of Fisk's (1944) determinations, then the considerable width of 5 to 10 nautical miles of the mangrove belt is a product of the extended period of time during which approximate stillstand has persisted. If minor oscillations have occurred during this period, then some of the alternations of peat and of peat marl and in the types of peat reported by Davis may have been related to changes of sea level. An end condition of seaward advance may be found where the bottom slopes too steeply, or the growth has finally reached a zone where the processes outlined no longer produce bottom offshore that is sufficiently shoal to support mangroves. Under this hypothesis, we may understand why the mangrove growth on the southern part of the Gulf shore of Florida is exceptionally wide, as the combination of conditions required for the full formation of a mangrove ridge and lagoon are exceptional. Under this hypothesis the rate of Gulfward advance is the ratio between the width of the ridge, 30,000 to 50,000 feet, and the duration of stillstand, 3,000 to 5,000 years. Using the

figure 5,000 from Fisk's estimate, we find that the net outward advance of the mangrove forest has been between 6 and 10 feet per year. This is a measurable quantity.

It has been said that Davis (1940) finds the present swamp forest to be resting on and rooted through a surficial zone of marl a few feet thick without appreciable peat deposition in it. Hence, his interpretation that the accumulation of the average of 7.5 feet of buried peat and marl took place mainly during a rise of sea level seems not to conflict with the present writer's hypothesis for forward growth during stillstand. Accumulating datings by the deterioration of radiocarbon may permit a rate of upward growth, less compaction, to be made, Davis having found no means of doing so.

The mangrove barrier ridge and coastal lagoon are similar, in accomplishing appreciable shoreline prograding, to the other barriers known, the barrier island of sand, the barrier coraline reef and the rare barrier oyster reef, noted in the Gulf of Mexico where a large reef of Crassostrea virginica forms a bay barrier 25 miles long across the mouth of Atchafalaya Bay.

EMERGENT AND SUBMERGENT SHORELINES OF THE GULF

USE OF TERMS

Johnson (1919) laid much stress, in his shoreline studies, on the determination of whether the features of a shoreline were dominantly those of the relative submergence of a land surface or the relative emergence of a sea bottom. His interest centered in the immediate history of sea level and its effect on shorelines. Others have found it impracticable to discriminate entirely on many coasts between the exact form of the present shoreline and the topography of the coastal zone which has determined major features of both coast and shoreline (Shepard 1937a, Price 1939). This distinction involves the difficulty in determining for each sector whether the several submergences or emergences of the coasts during the Pleistocene have produced its dominant features.

A major shortcoming of the Johnson classification or the way in which it came to be applied was its use of the common and widely developed barrier island as either a major criterion or a

positive indicator (Price 1939, Johnson 1938) of emergence. Later work, here discussed, seems to invalidate this criterion entirely as an indicator of sea-level movement except where it may be found wholly emerged or submerged.

While not finding the concepts of submergence and emergence as valuable for shoreline classification as some others, we may well inquire what features plainly indicate such items of shoreline history.

SUBMERGENT SHORELINE FEATURES

14

Pleistocene entrenchment of stream valleys.-It is well established that the accumulation of large amounts of ice in the arctic and circumarctic regions several times during the Pleistocene, or Great Ice Age, caused strong lowerings of sea level. The latest well-established major lowering occurred in the late Wisconsin or Würm glaciation and amounted to about 450 feet in the Gulf of Mexico (Fisk 1944, 1952) and the Gulf of Paria, south of Trinidad, Venezuela. In some regions the figure is set at between 240 and 350 feet (Flint 1947). Fisk (1944, 1948, 1952) has shown by borings cited in various reports of the Corps of Engineers that the northwestern Gulf coast had a large number of entrenched Pleistocene stream valleys that have now been filled with sediments. Configurations of branch estuaries show that entrenchment was general in the Gulf. Only on the hard shelf off peninsular Florida are such valleys found submerged and fairly well outlined by depressions. These valleys are marked by depths of as much as 10 to 15 feet on the coast charts off northern peninsular Florida.

Embayed drowned valleys.-Incompletely filled drowned valleys in the Gulf take at least two forms, those in which the branching, dendritic pattern of drowned tributaries is still prominent (Baffin Bay in Texas) and those in which waves and currents have broadened the valley at shallow depth, producing oval, rounded or other equidimensional shapes. The writer (Price 1947) has shown that on the northwestern Gulf coast elongated drowned valleys tend to become segmented by spits and other obstructions, separately embayed by segments and the bay bottoms made flat under a dynamic equilibrium between erosion and deposition. This equilibrium of basin shape

14 Personal communication, T. H. Van Andel.

is actually the result of the formation of equilibrium bottom-profiles along most bay radii.

Embayment of drowned streams is most prominent in the Gulf on the compound Pleistocene-to-Recent deltaic coast of Texas and southwestern Louisiana. There, the rivers are large and the gradient of the Beaumont is not steep (1 to 3 feet per mile). On the steeper plain of Sector 1.2 (fig. 14), only one large, transverse valley bay (Mobile Bay) occurs. Where the plain is composed dominantly of active deltas or hard rocks or has only relatively minor streams, Sectors 1.11, 2.0 and 3.0, long broadly embayed stream valleys are absent. The writer has further considered local meteorological influences in the shaping of the bays of the northwestern coast (Price 1952).

The harbor of Matanzas, Cuba, is thought to be a drowned valley cut in a structural depression. or in a structurally weak zone.

Submerged base of mangrove peat.-Davis' (1940) conclusion that the mangrove swamps and peat of Florida formed during a gradual, more or less uninterrupted rise of sea level from about -8 feet relative to mean sea level has been mentioned.

Drowned lacustrine plain of Florida Bay.— Anaylsis of this unusual type of marine area needs somewhat extended exposition. The entire water area (Trask 1939, pp. 292, 293) is a honeycomb of shallow, rimmed basins individually upwards of 10 miles wide and 11 feet deep, the bottoms bare with a cover of soft marl or shell sand. The narrow rims are of marl and mangrove peat (Davis 1940). The writer's interpretation is that a rising sea moving up and across a very gently sloping shoal surface carried with it a transgressive shoreline zone of mangrove swamp. This coastal swamp belt, the mangrove ridge, moved slowly north and northeastward and is now present along the north shore. This ridge is of irregular shape and the marsh and swamp back of it to the north are now and were probably at all times honeycombed with lakes. Such lakes tend to become enlarged by wind scour if the banks are not encroached on too strongly by marsh and swamp growth. The result is that some large lakes occur among the innumerable small ones. It is further postulated that, as the Gulf waters invaded the swamp, more and more deeply, vegetation was slowly killed, and the lakes gradually widened by drowning and wave erosion.

The outlines of the lakes and rimmed basins of Florida Bay today show the characteristic coalescing of small basins with each other and with large ones, the intervening rims being removed. The lacustrine plain, the so-called bay, with its network of marl ridges prevents the development of appreciable tidal flow and scour in and between basins except along the border of the bay at the south. Here tidal channels have been scoured through breaks in the line of the Florida Keys, locally deepening the rimmed basins.

Statistical study of the relation between width and depth in the rimmed basins shows a rough approximation to the progressive deepening with increasing size characteristic of the bays of the northwestern Gulf (Price 1947). In the Bay of Florida this relation is modified on the southeast by the limiting depth of the hard Miami oölite and at the extreme west by an excess of sandy or marly deposition in the relatively large basins that there border the Gulf. Some of the western basins are completely filled with sediment.

Partly submerged eolian sand plain of Rio Grande delta region.-Stretching inland across the Pleistocene plains of this delta in Tamaulipas and Texas to their inner erosional scarp is a plain of eolian sand, or erg, with scattered dune fields. All, except small blowout fans of bare sand (about 1 by 3 miles in size) and their fields of bare dunes, is stabilized by grassy vegetation, thorny brush and live oaks. The coastal lagoons now form traps for eolian sand blowing inland from the beaches of the barrier islands. Only in droughts is some of this sand able to cross to the mainland over narrow flats that locally close the coastal lagoon. This immense sand plain must have come on shore before the barrier island was formed. The simplest explanation follows that of Daly (1934, pp. 197-201) that large amounts of sand probably blew on some shores when the sea level was low during one or more of the glacial periods. Other possible explanations are that the sand has come from the reworking of successive barrier island sands and other beach deposits or from sandy sediments in the walls and on the floors of entrenched valleys.

EMERGENT SHORELINE FEATURES

Pocket harbors (emergent rimmed basins) of Northwestern Cuba.-Several writers on Cuba (as Hayes, Vaughan, and Spencer 1901) have referred

to the purse-shaped or pocket harbors of Cuba. Those of the sector from Havana to Bahia Honda (fig. 12) on the northwest coast are of an unusual, petal-shaped type. They lie in a plain from 3 to 5 feet above sea being upwards of 6 miles long. A small stream usually enters one or more of the several marginal indentations of the small rounded-to-oval basin, not always in the axial position. Other similar marginal indentations have either no appreciable inflowing drainage or receive very slight drainage. Yet well-formed submerged channels converge from all these indentations toward a central channel of tidal type. This channel may be as deep as 8 fathoms. Such harbors do not seem to the writer to be explicable as normal embayments of drowned stream courses or of stream confluences, as some have suggested.

If the coast of Cuba west of Bahia Honda (3.1, fig. 14) is examined on the navigation charts, basins similar to the pocket harbors and the rimmed basins of the Bay of Florida, lying in swampy terrain, will be seen here and there behind the Colorados Barrier Reef, mostly clustering toward the mainland shore. These basins have axial or radial tidal channels draining to the Gulf below sea through passes or breaches in the reef. These small, rounded and rimmed basins of the Colorados lagoon seem to be features of a present mangrove-lined shoreline like those along the north shore of Florida Bay. The writer interprets the pocket harbors of the Havana type, surrounded by a slightly elevated plain (Palmer 1945), as similar mangrove lakes scoured out at sea level in the midst of a saline swamp and then slightly elevated on the unstable young orogenic coast (Sector 3) of Cuba.

Barrier Island not an indicator of long-period sea-level change.-Johnson (1919, 1925) thought that his offshore bar, called barrier island by the writer (Price 1951a, Shepard 1952), was a feature predominantly of an emergent shoreline. He believed that the structure was formed by a semipermanent sea level change, a slight worldwide lowering of sea level or an upwarping of the crust, along an offshore bar formed originally as a submarine feature. Fenneman (1938, p. 4), following some early writers, believed, however, that a barrier island was formed merely as an equilibrium structure produced on a shallow shelving coast by the balance between wave attack and bottom resistance regardless of any history of sea level

change. The writer's study of bottom profiles in the Gulf (fig. 15) indicates that barrier islands are (1) associated with well-developed equilibrium profiles, (2) on a shallow coast where the bottom is now at least 15 to 45 feet deep within one to two miles of shore and (3) thereafter slopes outward between about 2.0 and 5.0 feet per mile, (4) where sand, gravel or cobble are abundant along shore, and (5) where onshore wave attack is strong. These observations tend to confirm Fenneman's interpretation. Other observations, briefly stated, indicate that the barrier island does not require a worldwide or other semipermanent fall of sea level to bring it above sea, but that the only change in level needed is a local, short-period change between storm levels and normal sea levels taking place during periods of a few hours or days.

A series of aerial photographs taken at intervals of several years over the period 1934 to 1949 (Bates 1953) shows that a bar formed just below the intertidal zone off a new mouth of Brazos River remained submerged until a hurricane had occurred, after which it became a typical emergent barrier island of cuspate outline. A second bar then formed off a breach in this barrier, after which other hurricanes occurred before the second bar was, in turn, raised above sea to form a second line of emergent barriers. The inference is strong that, in each case, a pre-existing submarine bar was built higher during a hurricane, so that during the storm it bore the same height relation to the elevated storm sea level as it had formerly borne to the normal level of the Gulf. The bars emerged as barrier islands after the subsidence of the temporarily high sea levels.

On October 3, 1948, a hurricane passed about 100 miles off the coast of southwestern Texas, causing a high sea level or storm tide of some 3 or 4 feet for two days or more along the barrier islands. A week later, the writer found that the summit of the beach, the beach ridge, in front of the shore dunes had been built up and remade by the storm and was slightly farther inland than its former alignment. The shift in position was evidenced by erosion of dune faces. The convexly rounded beach ridge then rested where the front part of the dunes had been.

The raising of the beach ridge, previously described, to an elevation above its position during normal times was shown by the rapid mass-wasting that had affected it in a single

week. The beach ridge on this island formerly had, in places, a fairly well developed pavement of shell, but now the pavement had just begun. to be formed on the newly made ridge. The pavement was formed from disseminated shell by the washing and blowing away of sand, according to a well-established process. It was evident that this ridge had lost some 6 inches of its height and would lose another foot or a foot-and-a-half before a pavement would be formed to protect it. The former paved beach ridges had evidently lost similar heights.

Reports and illustrations of hurricane damage to New England beaches (Brown 1939, Howard 1939) show that the beach ridges were remade at higher levels, moved inland from their former positions and their axes rotated slightly by the hurricane waters.

These observations indicate clearly that the summit ridge of a barrier island functions briefly during storm tides as an underwater offshore bar and thereafter emerges as a barrier island.

Evans (1942) found that waves operating at a steady sea level tend to modify the slopes and positions of underwater bars, but not to build. them up above water.

The great development of active barrier islands on the Gulf coast, dominating the shorelines of the alluvial sectors (1, fig. 14), does not then, in the writer's opinion, tell a story of permanent or semipermanent sea level change, or mark either a submergent or an emergent shoreline condition.

The question of the source of the supply of material for the barrier, long thought to be a critical factor, is found to be secondary. Thus, barriers occur in the Gulf where longshore sediment drift is prominent (Sectors 1.12, 1.2, fig. 14) the sand derived largely from rivers (Bullard 1942), also where a longshore drift from a land connection (Chandeleur Islands, La.) is absent and where no land-derived sediment is present but onshore waves are strong and the barrier is built of broken shells from the adjacent bottoms, as on the north shore of Yucatán (2.2, fig. 14).

Emergent shoreline terraces and notches.-The lowest well-established elevated shoreline is that of the Pamlico of the Atlantic coast and Florida, standing at about 25 feet above mean sea level. This shoreline is marked in many regions by a well-cut and well-preserved terrace or by a broad elevated lagoon flat with a barrier island. Less

well-developed and somewhat controversial shorelines are reported from many places in the interval from about +3 to +10 feet. However, carefully selected, stable, protected, inner shoreline sectors on North Pacific Islands (Stearns 1941, 1945) and in Australia (Fairbridge 1948) exhibit shoreline cliffs in even-grained limestones with solution notches at about +3, +5 and +8 feet. These seem to represent worldwide stillstands of the sea (eustatic shorelines). Shorelines reported at +16 and +20 feet (Daly 1934 and others), are not as yet substantiated by data of unquestioned accuracy.

In places around the shores of the Gulf, there are definite indications of shoreline terraces that seem to indicate stillstands at about +5 and +8 feet. An elevated barrier island and coastal lagoon caught by the 10-foot contour has been mapped in Florida by MacNeil (1950) as the "Silver Bluff shoreline" (Parker and Cooke, 1944, pl. 4, fig. B). He did not follow it across southern Florida or on the west side of the peninsula.

Low shoreline flats appear in many places around the Gulf, but have not been critically studied in the field. Such a low bench shows in air photographs along the base of the high bluffs of the Champoton-Campeche limestone fault-block salient (fig. 12; Sector 2.2, fig. 14). It seems to have a gray, sandy soil. A flat along the front of the elevated Ingleside shoreline between the Rio Grande and Brazos-Colorado deltas (fig. 12) lying at from about 1 to about 5 feet above mean sea level has low, subdued spits and bars on its surface 15 and seems to be an emergent marine plain. It is about 0.3 mile wide. This flat may be a nondeltaic part of the original Pleistocene surface in front of this barrier. Deltaic deposits appear along the Gulf side of this barrier east of Galveston Bay.

Marsh borders the Pleistocene delta of Brazos River in Texas to an elevation of 2 to 3 feet above mean sea level. Just behind the marsh is a bench 1.0 to 1.5 mile wide at 3.0 to 4.5 feet with a low nip or wall between 4.0 and 6.5 feet above sea. This bench may be a low Silver Bluff representative.

At Buhler, a few miles northwest of Lake Charles, Louisiana, the Ingleside barrier and lagoon clays are well preserved. The top-of-clays,

15 Obscured by mima (pimple) mounds higher and wider than the spits (Price 1949).

representing the approximate shoreline position, lies between 22 and 25 feet above sea. This shoreline and the associated features are well defined running at the same elevation from near Lake Charles west to Beaumont and thence southwest through Fannette, Jefferson County, Texas. Where the shoreline comes within about 10 miles of Anahauc, Chambers County, it is sloping down to the southwest at about 1.5 feet per mile and reaches sea level at Smith Point on the shore of Galveston Bay. Before the formation of the bay, it was formerly tied there to the Brazos delta. The Ingleside shoreline seems to correlate with the Pamlico through the emergent barrier of Gulfport and Biloxi, Mississippi.

The deltaic plain lying south of the Ingleside in southwestern Louisiana and in Jefferson and Chambers counties, Texas, is of the same age as that immediately to the north of it, Prairie or Beaumont (Hayes and Kennedy 1903, pp. 27–38; Deussen 1924, p. 110). Along the shore of Jefferson and Chambers counties, it is a partly submerged deltaic plain.

The Ingleside appears again south of the Brazos-Colorado delta along the coastal lagoon that opens from Matagorda Bay at its southeast extremity and runs from there to the north flank of Rio Grande delta, the shoreline (top of clay) being at approximately 5 to 10 feet above sea.

The disagreements in the shoreline data for the northern Gulf coast would be removed if the coast from Florida to the Mississippi delta had been stable since Pamlico time, but a slight amount of gulfward downwarping had occurred between the vicinity of Galveston Bay and the coast of Mexico at some point north of Tampico.

The post-Pleistocene gulfward downwarp of the Beaumont Pleistocene plain (Doering 1935) increases in amount from about 1 foot per mile in southeastern Texas to 2 feet per mile southwest of Matagorda Bay. This downwarp seems to mark the influence of the young orogenic coast of Mexico, which it is approaching.16 This interpretation suggests that the emergent shoreline flat on the Gulfward flank of the Ingleside barrier may be either of Ingleside age, downwarped some 15 feet, or a younger post-warping shoreline, possibly of Silver Bluff age. Against a Recent age for the low bench is the seeming absence of marine fossils

16 Corpus Christi lies 175 miles east of folded Cretaceous rocks at the surface in Mexico and 125 miles northwest of submerged mountaius in the Gulf.