"My stream table," writes Schwartz, "is some three feet wide, 10 feet long and six inches deep. It is hinged so that either end can be raised and it holds 500 pounds of dry sand. The weight of the completed assembly when wet, together with accessories, approaches half a ton. I first made two boxes, each open at one end, of 20-gauge galvanized iron and after soldering the corners I etched the metal surface with dilute muriatic acid and applied two coats of epoxy resin paint. The sheet-metal boxes were then nested for support in close-fitting wooden boxes. Half-inch plywood was used for the bottoms. The ends and sides, which measure two by six inches in cross section, were made of Douglas fir. One box is about four feet long and the other about six feet long. The open end of the larger box was beveled at an angle of 20 degrees to allow the hinged sections to be folded. Three equally spaced cross cleats were spiked to the bottom of each box to give added strength and to provide air space so that water could not collect between the bottom and the supporting surface. One could reduce the cost somewhat by constructing the boxes of rough planks and lining the wood with plastic sheeting. The sand is shifted frequently, however, and thin plastic is not particularly durable. The end sections were joined by a heavy piano hinge and covered by a wide strip of heavy sheet rubber that extends across the bottom and over the sides. The strip was first cemented over the hinge and then secured at the edges by metal strips screwed to the wood. Ordinary door hinges, although not as strong, could be substituted and covered with a strip of opened automobile inner tube. In this case the crack in the bottom and the V-shaped openings at the sides should be covered with loose pieces of plywood to prevent the weight of the sand from pushing the rubber into the openings.

"The larger section of the stream table is equipped with pipe racks for supporting shower heads about two feet above the sand and with a clip on the upper edge of the end for holding a garden hose that leads to a water tap. Rain is simulated by the sprinkler system and stream sources by the open hose. An outlet at the bottom of the smaller section is fitted with a short length of garden hose that can be clipped to the end at various heights for adjusting the depth of accumulated water [see Figure 1].

"Accessories include plastic sheeting, a cement trowel, a few pounds of dry sand and a wave generator. My wave generator was improvised from a small aquarium pump of the type that has a leather diaphragm for a piston. The cylinder, piston and connecting rod were disconnected and a metal rod about an eighth of an inch in diameter and five inches long, threaded at each end, inserted in the end of the connecting rod to replace the screw that held the leather diaphragm. During reassembly the metal rod was passed through a hole drilled in the bottom of the cylinder and a small wooden block was attached to the free end of the rod to serve as a paddle. The completed mechanism was then mounted on a plywood base that is clamped to the side of the stream table where the paddle makes contact with the water. The speed of the motor, and hence the size of the waves, can be controlled by a rheostat.

"The apparatus rests on a sturdy platform that is about waist-high. The larger end is raised and lowered by means of an automobile jack and is supported in elevated positions on blocks. In most experiments the smaller end of the table remains level and the larger end is tilted to represent elevated terrain. Coast lines are formed by pushing sand away from the lower end of the table. Sand of any grade may be used, but I prefer the fine white variety used in hourglasses known as Ottawa sand. This material is clean, flows readily and holds its shape nicely when wet. Ottawa sand can be procured from suppliers of building materials. It is currently priced at $5 per 100 pounds in the New York City area. Five hundred pounds fills the stream table to a depth of about four inches.

"The operation of the table makes some demand on the experimenter's manipulative skills because the production of realistic effects involves art as much as science. Most experiments are set up by first modeling wet sand to represent some combination of geological features such as mountains, valleys plains or coast lines. The features are then subjected to erosion by streams, waves or mechanical forces. The investigation of land erosion in rainy regions makes a nice introduction to the technique.

"The introductory series of experiments assumes a region that has recently been elevated above sea level by natural forces: a uniform, slightly rolling terrain that slopes gradually to the sea. First elevate the larger end of the table about 18 inches and block it in place. Place about half of the sand in the elevated section so that it extends slightly past the hinge into the lower section and cover it with plastic sheeting. The edges of the sheeting should be turned up about an inch on all sides. Place the remainder of the sand over the sheeting so that it extends to within about three feet of the lower end and make a few small piles of sand here and there in the empty area. Wet the sand thoroughly and smooth it from side to side to form a shallow curve down the middle with the edges about an inch above the center. Mold the lower end of the sand to form a steeply sloping beach and give the inland topography a bit of moderate relief here and there. Adjust the outlet tubing so: that the water that accumulates in the smaller section will reach about half the depth of the sand and thus establish an initial shore line. Connect the inlet hose to the shower heads and open the tap to simulate a brisk rain.

"Initially the falling drops will merely smooth the surface and the water will soak into the sand. After a short interval, however, small gullies will form near the lower boundary of the rain belt and gradually become longer and deeper until they unite in a single, fast-running stream. As the rainfall continues, erosion will extend the gullies slowly upstream.

"In the meantime the main stream deepens rapidly and within a few minutes will have cut a relatively straight channel to the 'sea,' a channel with steep banks and few sand bars. The runoff accumulating at the lower end of the stream table gradually submerges the marine features; bays, peninsulas and islands appear, depending on the prearranged pattern of sand.

"After the basin fills note in particular the mechanism of delta formation at the mouth of the river. The current quickly loses its velocity when it encounters the still water, and it deposits a thin, fanshaped pattern of sand on the bottom that extends a considerable distance beyond the mouth of the stream. Geologists refer to this initial deposit as the 'bottom set.' A second deposit, known as the 'fore set,' is characterized by a steep, straight edge that advances over the bottom set. As the top of the fore set approaches 'grade' (the average level of the stream), a thin, gently sloping deposit is laid down on top of the fore set. Thereafter erosion cuts a channel in the top set, and the velocity of the stream, now confined to the channel, increases and initiates a new cycle with the formation of bottom set at the mouth of the new channel. When this cycle has progressed through the stages of fore set and top set, still another channel forms. In this way the exposed surface of a mature delta becomes a fan-shaped patchwork of small islands interlaced by streams.

"When the delta has matured, lower the larger section of the stream table so that the terrain slopes about three inches from end to end. This simulates a later stage in the history of the stream's development, when erosion has leveled the uplands. Initially the gullies that appeared beneath the shower heads had joined to form a relatively narrow, fast-running stream that tended to deepen faster than it widened. After the straight channel reached a certain depth, however, its steep banks were undercut here and there. Material breaking free at these points was swept downstream. Other portions of the bank, where the sand was packed more densely, resisted the current for a time, and erosion proceeded fastest at the opposite bank. Slight bends therefore appeared along these stretches of the river. Such departures from straightness had the effect of directing the water still more forcefully against the points of weakness and accelerating erosion. With the simulated erosion of the uplands, the narrow, fast-running stream becomes a slow, meandering river. After some hours it will observed that the meanders move slowly downstream and bends form near the mouth of the river. The result is a broad valley dotted here and there by former islands, sand bars and strange crescent shaped lakes that mark channels originally occupied by the river.

"Other major features of rainy regions are induced by ground water that collects between the surface of the soil and a layer of relatively impervious material underneath. This layer is simulated by the plastic sheeting of the stream table. To set up a demonstration ground water, first shut off the shower heads and drain the accumulated water. Then lower the larger section so that the sand is about six inches higher than it is at the outlet end. Mold a terrain that slopes gradually and, beginning at the high end, form in succession a lake bed, a lowland, a hill with gently sloping sides, and a small basin behind a ridge near the coast line [see illustration]. Detach the garden hose from the shower heads, clip it to the elevated end of the stream table and open the tap. Regulate the flow so that seepage into the sand just balances the inflow and causes the level of the lake to remain constant. Within a matter of minutes the underground water table will form, seep into the lowland and create a swamp. Shortly thereafter a pond will form in the depression near the coast line and one more springs will appear downgrade from the pond, if the edge of the plastic sheeting has been brought close to the surface. Runoff from the spring will gradually fill the basin at the outlet and form a beach. A well dug in the face of the hill above the pond will fill to the height of the water table.

"Of the several geological agencies responsible for large-scale alterations of the landscape none is more spectacular in its effect than the action of glaciers. So far I have discovered no way to simulate the slow deformation and flow of ice in the stream table. But it is easy to duplicate the effects of a glacier in the region of its most extreme advance-where the flow of ice has stopped. The glacier simply melts and deposits its burden of transported material. The associated runoff erodes the downward-sloping terrain.

"To perform this experiment drain the stream table, remove about 75 pounds of sand, lower the larger section so that it is level with the smaller section and embed the plastic sheeting about an inch below the surface of the remaining sand which should slope gently from one end to the other. Next build a ridge about six inches high across the mouth of a U-shaped valley, near the hinge, using mixture of sand and pebbles. Then fill the upland space to a depth of eight inches with a mixture of four parts of crushed ice (or snow, if it happens to be available outside) to one part of sand. Bury several fist-sized lumps of ice in the sand downgrade from the ridge and let the table stand until the ice melts. The resulting surface will be marked by many features typical of glacial land forms. Several small lakes may form in the 'till plain' above the ridge of sand and pebbles. The ridge, which will have become somewhat eroded, represents a terminal moraine: the rubble deposited at the edge of a glacier. An 'outwash plain' will have formed downgrade from the terminal moraine, marked here and there by random pebbles and perhaps by a stream bed and depressions, known kettles, that mark points where lumps of ice were buried-all typical remnants ice erosion that can be observed in northeastern U.S. as results of the continental glaciation that ended about 10,000 years ago.

"Some features of the crust's vertical excursions, such as the great uplift currently under way in the southwestern U.S., are quite easy to simulate. These uplifts invariably open cracks, or faults. In nature displacement at the interface can amount to as little as a fraction of an inch or to many feet. In the case of Sierra Nevada a block of granite hundreds of miles long has split from the rest of the crust and has tilted in such a way that successive displacements have raised one side of the block thousands of feet and have similarly lowered the other side.

"To investigate faulting of three principal types remove the plastic sheeting from the stream table, place the automobile jack under the larger section and pile the wet sand level with the edges of the table as far as it will extend above and below the hinge. As you jack up the larger section, observe the sand. A crack will appear in the Vicinity of the hinge, and as the larger section continues to rise the block of sand in the larger section will slide below that in the smaller section. This is an example of compression faulting. Note the angle of the fault with respect to the horizontal. Without disturbing the position of the table, smooth the sand at the fault and tamp it down lightly. Then reverse the jack and lower the larger section. A tension fault will now appear. Observe that the angle of this fault makes a large angle in relation to that of the compression fault. Jack up the larger section again, make a deep V-shaped trench across the table above the hinge and fill it with dry sand. Reverse the jack and lower the larger section. Two faults will appear, one on each side of the trench. The surface of the dry sand will drop below that of the wet material and form a 'graben,' a flat-bottomed trench with sloping walls. The upper valley of the Rhine River is a typical example, a depressed area some 200 miles long and 20 miles wide between faults bounding the Black Forest on the east and the Vosges Mountains on the west.

"A particularly fascinating series of experiments can be based on the erosive action of water waves on coast lines. Set up the stream table to form a deep basin some three feet square at the outlet end of the smaller section and mold wet sand into a sheer cliff about six inches high. Fill the basin to a depth of some four inches and set up the wave generator to produce waves about two or three inches from crest to crest. After five minutes lower the water to a depth of two inches. Observe the shape of the beach that emerges and particularly how the cliff has been undercut here and there. Such undercutting frequently carves sea caves from the rock, such as the Grotta Azzurra (Blue Grotto) on the island of Capri. Beach formations that stand high above present water levels are common features, such as those surrounding Great Salt Lake in Utah. Wave action is also responsible for such features as the long, thin offshore sand bars that characterize the eastern coast of the U.S. To observe the formation of sand bars, shut off the wave generator and build a smooth bank of sand the width of the tank and extend it into the water about a foot so that the surface of the sand is submerged an inch. Start the wave generator. Wave action will quickly transport sand from the leading edge of the bank to a region behind it, where some of it will form a long offshore sand bar. After a few minutes the growing sand bar will begin to interfere with the transport mechanism and the bottom will stabilize as a profile of equilibrium. Thereafter, if the amplitude of the waves does not increase, the sand bar and shallow basin behind it will remain as permanent features.

"Wave transport frequently links a small offshore island to the mainland by a sand bar, known as a tombolo. The transported sand may come from the mainland or from the island, depending on the wave action and the nature of the terrain. To simulate the formation of a tombolo, shape an offshore island on a shallow submerged shelf and generate waves that approach the island obliquely. In time the tombolo will form on the leeward side of the island in line with the advancing wave front. The agency of turbidity currents is responsible for the large-scale transport of submarine sediments. These can be simulated by pouring dilute sirup into the water at the edge of a steeply sloping beach. The currents are easy to follow by eye if a few drops of food coloring are added to the sirup.

"Numerous other experiments have been made. It is possible, for example, to simulate the formation of pillow lava: the irregularly rounded deposit that forms when molten lava flows into a lake or the sea. The glassy exterior of such pillows is caused by the rapid crystallization that occurs when the lava makes contact with the water. Cavities in the pillows are usually rich in minerals that are derived from gaseous elements trapped in pockets of the molten material. Realistic pillows are formed when molten solder is poured down a steep beach of the stream table.

"The apparatus invites much experimental work that remains untried. It should be possible to simulate the flow of glaciers by means of silicone putty or other plastic media, and the features that derive from glacial melting could be developed in greater detail, perhaps by supplementing the melting ice with a water source. Then too, dunes and other effects of wind erosion can be investigated by the experimenter who has access to a source of compressed air. I have reproduced some formations characteristic of arid regions, such as the piles of debris known as talus that accumulate at the bottom of steep cliffs, simply by pushing dry sand over the top of a rise. We are continuing to develop new stream-table demonstrations here at Columbia University that are intended to serve primarily as teaching aids. The apparatus will also be used in a series of research projects."

Bibliography

CLASSROOM STREAM TABLE. Walter C. Brown in Journal of Geological Education, Vol. 8, No. 2, pages 63-64; Fall, 1960.

ELEMENTS OF GEOGRAPHY. Vernor C. Finch, Glenn T. Trewartha, Arthur H. Robinson and Edwin H. Hammond. McGraw-Hill Book Co., Inc., 1957.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.