THE FACT that salty water is denser than fresh water provides a basis for a number of experiments, some of them quite surprising. For example, one might suppose that if a Layer of salty water was put above a layer of fresh water, the salty water would simply sink. If the layering is done in a certain way, however, the result is not so simple. What one finds is that the system starts to oscillate and may continue to do so for a remarkably long time.

Pour cold tap water into a tall clear glass until the glass is nearly full. Put a hole in the bottom of a paper cup or a plastic cup with the point of a safety pin or something of similar size. Make a salty solution (about equal to half the volume of the cup) from cold tap water and a teaspoon or so of salt, taking care that the salt dissolves fully. Add a bit of food coloring to the solution so that later, when the experiment is under way, it will be easy to distinguish the salty water from the fresh.

Now start lowering the cup into the glass, simultaneously pouring the salty solution into the cup. You want the cup to reach a depth where the fresh water outside it is about halfway up the sides of the cup, and you want the salty solution in the cup to be at about the same level as the fresh water outside. At that point, since the effect is sensitive to jostling, you should fasten the cup. I fastened it with a beaker clamp attached to a ring stand of the kind used in chemistry laboratories, but you might try taping the cup to two flat kitchen knives laid across the top of the glass.

One would suppose that in this situation the dyed salty water would stream through the hole into the fresh water until the level of the salty water was reduced enough so that the pressure at the level of the hole was the same in both bodies of water, whereupon the stream would grow thinner and finally stop. Indeed, you will see these events over a period of several minutes, but do not turn away. In a short time the stream will begin flowing again! Thereafter, with a period of from 15 to 20 seconds, the stream will begin abruptly, wane disappear and then start up abruptly again.

What is happening is that during the time when the dyed stream is not visible, fresh water is streaming upward into the cup through the hole. The system is oscillating. This effect was discovered in 1970 by Seelye Martin of the University of Washington [see "The Amateur Scientist," June, 1971].

The period of oscillation depends mainly on the size of the hole and to a lesser extent on the concentration of salt in the cup. My oscillator continued overnight with approximately the same period in spite of the continuous reduction in the concentration of salt due to mixing. Martin's oscillator, which was in a tin can instead of a paper cup, continued for four days. He also ran a test with a hypodermic syringe replacing the cup; with this system he obtained a period of about four seconds and a lifetime of 20 cycles.

It is easy to understand the first part of the flow. The salty water was initially as high in the cup as the fresh water outside the cup. Since the salty water was denser than the fresh water, the pressure at the hole was higher from the salty water than from the fresh water. The salty water therefore poured through the hole until the pressures equalized. The puzzle is why the salty water did not merely stabilize then as the downward flow slowed to a stop.

In spite of the equalization of pressures at the hole a layer of denser fluid overlying a layer of lighter fluid is unstable and subject to any small, random disturbances to the system. Any such disturbance creates a small wave on the interface between the two types of fluid. The wave grows in amplitude exponentially with time, at least initially, because of the density difference. As a result some of the lighter fluid protrudes upward across the old interface and some of the denser fluid protrudes downward. This instability to small disturbances and the resulting intrusion of each fluid on the other's domain are responsible for the oscillation in the salt oscillator.

The fresh water protruding upward accelerates through the hole because it finds itself lighter than the salty water at the same level on the other side of the hole. Apparently the fresh-water stream pinches off the salt-water stream soon thereafter, and there is then a single up-going stream of fresh water moving through the hole. This addition of water to the cup gradually increases the height of the fluid in the cup and thus the water pressure at the level of the hole. The loss of water from the glass, however, barely lowers the water level there because the glass is wider. Eventually the pressure from the salt-water side becomes great enough at the level of the hole to overwhelm and halt the up-going stream of fresh water. We are then essentially back at the start of the cycle. There is too much water in the cup, and a stream of salt water emerges downward for a while. Gradually the flow of the stream decreases as the pressures once again become equal at the level of the hole. Then some random disturbance sets up a wave on the interface that once again sends up a stream of fresh water. With a rate depending on the diameter of the hole and of the cup and on the viscosity of the water, the streams alternate up and down. You have a salt oscillator.

The instability at the interface between a denser fluid overlying a lighter fluid, when the interface is otherwise in hydrostatic equilibrium, is called a Rayleigh instability (or sometimes a Rayleigh-Taylor instability). The salt oscillator is an example of a system that oscillates after exciting itself, in this case through the Rayleigh instability and the consequent rapid growth of a disturbance on the interface between the two fluids.

Apparently Martin's paper is the only one that has been published on salt oscillators, and you might be interested in doing more work on them. For example, how does the period of oscillation depend on the diameter of the hole? Martin's use of a hypodermic syringe and needle is advantageous because needles of different radii could be easily interchanged, the radius of each was well known and the needle was certainly more circular in cross section than my pinholes. Nevertheless, you can establish the general dependence of the period on the diameter of the hole with a range of pinholes.

The dependence of the period of oscillation on the concentration of salt can be checked by adding salt to an initially weak concentration of salt in the cup. Take care, however, that the added salt does not merely clog the hole.

You might also try other fluids. They do not have to mix. All that is required is that they differ in density and that the denser fluid be in the cup. I tried water and Karo light corn syrup (a clear syrup used in cooking or on pancakes), dyeing the syrup a light red before I poured it into the cup. With a hole of suitable size the oscillations of the viscous red fluid and the clear fresh water were almost majestic. The upward plume was much easier to see in the corn-syrup oscillator than in the salt-water oscillator. I lightly dyed the fresh water blue and could then easily see the upward plume of fresh water periodically break through the hole, push its way up through the syrup and then lie on top of it.

When the downward stream begins during the cycle of a salt-water oscillator, the dyed salt water appears to burst downward, taking the form of an inverted umbrella. During the part of the cycle when the flow is steadier, and in steady flow without oscillations, the stream may fall several centimeters and then separate into ring vortexes. The rings behaved much as my grandfather's rings of cigar smoke did on those quiet days in Aledo, Texas, when his ring-blowing was our only amusement. The rings would race after each other, wrap themselves about each other and then race away. This play is easier to see in the rings of dyed water if you use a round glass container, such as a large tea glass, because the curved surface magnifies the image.

Another example of Rayleigh instability can be observed in a common parlor trick. Partly fill a glass with water, place a piece of paper over the mouth and, while holding the paper in place, invert the glass of water. Now remove your hand from the paper. The water stays in the glass thanks to two effects. First, surface tension between the water and the paper and between the water and the rim of the glass helps to hold the paper in place. Second, in larger glasses the water column falls somewhat and thus reduces the pressure in the air still trapped in the glass. The difference in air pressure between the top and bottom of the column of water also helps to hold the water in place.

Suppose the paper were to disappear suddenly without disturbing the water surface. What would happen? Common sense tells you that the water would fall out. If the water surface could remain perfectly undisturbed, however, there is no reason why the water should fall. The pressure difference between the top and the bottom of the column should still maintain the water in the glass.

The water does fall, though, because it cannot remain undisturbed. Soon after the paper is removed a small wave will develop on the bottom surface because of some small random disturbance to the water. The wave will grow in amplitude, at first exponentially with time, exactly as we have seen at the interface in the salt oscillator, because once again we have a Rayleigh instability. A bubble rises up on one side of the column of water as part of the water begins to descend on the other side, and then we rather quickly have a wet floor.

Salt fingers involve a layering of saltier water over fresher water, similar to the layers in the salt oscillator. Although the fingers do not have the fascination of an oscillation, they are important in the microstructure of the ocean. Consider a layer of hot salty water over a layer of cooler fresh water. The salt makes the top layer denser than the bottom layer, but the difference in temperature more than compensates to give a net situation where the top layer is less dense than the bottom one.

One would suppose that if a small disturbance sends part of the top layer downward and part of the bottom layer upward, the difference in overall densities should restore the original boundary. For example, the hot salty water intruding downward would find itself less dense than its surroundings in the cool fresh water and would be buoyed back upward.

During the intrusion from the initial disturbance on the boundary, however, heat is exchanged relatively quickly between the hot water and the cool water. The hot water releases heat as it moves downward. The cool water absorbs heat as it moves upward. As a result the intruding salt water suddenly finds itself more dense than its surroundings rather than less, and instead of being buoyed back it is accelerated downward. Similarly, the intruding fresh water quickly becomes warm, finds itself less dense than the surrounding hot salty water and is accelerated upward. The initial disturbance to the interface between the two layers is therefore enhanced to produce fingers of salty water stretching downward and fresh water stretching upward. The fingers eventually stretch themselves to the point where they turn over to create a layer of mixed water with a salinity and temperature intermediate to those in the top and bottom layers.

The salt fingers are not themselves directly observable in the ocean, but the resulting temperature and salinity pro files have been detected [see "The Microstructure of the Ocean," by Michael C. Gregg; SCIENTIFIC AMERICAN, February, 1973]. For example, the warm salt water of the Mediterranean sets up the conditions for salt fingering as it flow through the Straits of Gibraltar an over the fresher, cooler water of the Atlantic. Probes measuring the salinity and temperature as functions of the depth show steplike profiles where the salinity and temperature decrease wit depth. Such a step has one layer of higher salinity and temperature than the layer below, with a layer between the two having intermediate values.

You can produce salt fingers quit easily in your kitchen by carefully pouring hot salty water (dyed with food coloring) over cooler fresh water. In pouring you want to minimize the initial turbulent mixing of the two types of water. If the container is a standard drinking glass, I tilt it so that the poured water falls only a centimeter or less. With a bigger container I can lower the container of salty water to nearly the level of the fresh water to accomplish the same thing. You can instead float a thin piece of wood on the surface of the fresh water and then pour onto the wood.

At first the interface between the dyed hot salty water and the clear fresh water merely indicates some mixing. After a few minutes the salt fingers, from one to five centimeters long and about a millimeter wide, develop at the interface, lasting from several minutes to several hours. I find that the fingers are easier to see if I shine a flashlight through the water somewhat toward my eyes.

The key to the salt fingering is the difference in the rate of diffusion of heat and salt. The heat that made the top layer lighter in spite of its salt concentration moved to the cooler fresh water about 100 times faster than the salt. Th rather quick transfer of heat is what made the downward-intruding salt water suddenly find itself heavier than it surroundings and what made the up ward-intruding fresh water suddenly find itself lighter than its surroundings

You can accomplish the same sort of convective fingering with solutions of sugar and salt. Make the sugar solution less dense than the salt solution by put ting somewhat less sugar than salt in given amount of water. Dye the sugar solution so that you can follow the motion, and gently put the sugar solution over the salt solution in a container (Since both solutions are at the temperature of cold tap water, this salt fingering is easier to do in the kitchen than the preceding experiment, where rapid cooling of the hot water can be a problem.)

Once again, as with the arrangement of hot salty water and cold fresh water, the layers of sugar solution and salt solution should be stable because the top layer is the less dense one. As before, however, any initial disturbance is enhanced because of a difference in the rates of diffusion: Salt diffuses faster than sugar. Thus when the initial disturbance sends a small amount of sugar solution downward the salt from the surroundings moves into that intruding bulge faster than the sugar can move into the surroundings. The bulge, with salt added, suddenly finds itself denser than its surroundings and is accelerated downward to produce a finger. Similarly, a small bulge of salt water intruding upward loses its salt faster than it gain sugar, finds itself lighter than its surroundings and is accelerated upward to produce a finger.

The interaction of salt and heat on the density of seawater and the difference in their rate of diffusion also lead to a curious "perpetual" fountain you can simulate in the kitchen. In a tropical ocean the water near the surface can be relatively warm and salty whereas the water near the bottom is relatively cold and fresh. Imagine a pipe lowered vertically almost to the bottom and a pump initially used to bring bottom water to the surface. Theoretically the pump could be removed and the flow would continue by itself. Since a single pipe would hardly alter the characteristics of the ocean, this flow would continue "forever."

To understand how the flow continues to pump itself consider a small parcel of water beginning its journey up the pipe. As that bit of cold water rose, it would gradually gain heat from the warmer water outside the pipe at the same depth. The parcel of water would then be lighter than the saltier water at the same depth and temperature just outside the pipe. As a result the parcel of water would accelerate upward. Assuming the parcel is always warmed in this manner throughout the trip upward, it would always find itself lighter than the outside water and would continue to accelerate. In principle the water would spurt as much as a couple of meters above the surface of the ocean, forming a perpetual fountain.

Henry M. Stommel of the Massachusetts Institute of Technology tried some years ago to set up such a fountain in the deep ocean near Martinique. With about 1,000 meters of flexible hose 5/8 inch in inside diameter he and his colleagues obtained a fountain some 24 inches high, but they are doubtful that it resulted from differences in density. The top of the hose was attached to a float, which rose and fell with the waves, stretching the upper part of the hose and creating a sort of pumping action. Stommel thinks the pumping action may have created the fountain.

Now pour a thin layer of hot salty water over the top of the warm water and thus also over the top of the cup, again being careful to avoid excessive mixing. Place one or two drops of dye just over the hole in the inverted cup. The dye indicates that a slow fountain of water spreads from the hole through the top water, a simulation of the oceanic salt fountain. The water rising through the cup is warmed to about the same temperature as the water at the same depth outside the cup, but it lacks the salt that the outside water has and is therefore lighter than the outside water. Hence the water in the cup is forced upward through the hole. This kitchen fountain will continue until the heat and salt are more evenly distributed.

As a final note on effects due to the difference in densities of salty and fresh water, let me describe the push a ship receives when the gates of the last lock in the Panama Canal open. As a ship approaches the Pacific end of the canal it is progressively lowered through locks to sea level. The water in these locks is mostly fresh water fed by rain-filled lakes in Panama. In each lock the ship must wait until the water level is lowered to that of the next lock. Then the water pressure on the two sides of the gate separating the two locks is equal, and the gate is easily pulled open.

In the last lock the water on the other side is the salty ocean water. Since saltier water is denser than fresher water, equilibrium pressure on the gate is reached when the fresh water is still a bit higher than the ocean level. As the gate swings open, this extra height of fresh water flows seaward, carrying the ship along with it on a brief but free ride.

Bibliography

AN OCEANOGRAPHICAL CURIOSITY: THE PERPETUAL SALT FOUNTAIN. Henry Stommel, Arnold B. Arons and Duncan Blanchard in Deep Sea Research, Vol. 3, No. 2, pages 152-153; February, 1956.

THE "SALT-FOUNTAIN" AND THERMOHALINE CONVECTION. Melvin E. Stern in Tellus: A Quarterly Journal of Geophysics, Vol. 12, No. 2, pages 172-175; May, 1960.

OBSERVATIONS OF THE CELL STRUCTURE OF SALT FINGERS. T. G. L. Shirtcliffe and J. S. Turner in Journal of Fluid Mechanics, Vol. 41, Part 4, pages 707-719; May 15, 1970.

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.