The investigation of these questions requires the statistical analysis of drop sizes as a rainstorm develops. It also requires an intimate knowledge of the mechanisms responsible for the growth of snowflakes. Recently Norman J. Wilder of Corvallis, Ore., and Richard W. Wahl of Golden, Colo., respectively chose the problems of gauging raindrops and growing snowflakes as projects for their Junior Research Fellowship in the Atmospheric Physics Program at the University of Nevada. They worked under the supervision of Vincent J. Schaefer.

The program provided only limited time for experimentation. For this reason the procedures devised by Wilder and Wahl require additional work before they can be evaluated. Even so, the experiments demonstrate that some of the more fascinating aspects of weather can be studied indoors and with modest facilities.

Several techniques have been used for measuring the diameter of drops. In one a flat plate coated with oil or a similar substance that resists water is exposed to the drops as an airplane flies through a cloud. Some of the drops stick to the plate and are promptly photographed. The resulting images are later measured. The procedure is not altogether satisfactory. Large drops tend to break up on impact with the plate. Moreover, the airstream separates at obstructions such as flat plates. The smallest drops are therefore diverted, and many do not come in contact with the oil. Hence the collection efficiency of the plate declines in proportion to the diameter of the drops. Finally, the smallest drops evaporate quickly and may disappear before they can be photographed. For these reasons the observations are biased in favor of the larger drops.

An indirect technique has been used to measure drop sizes in supercooled clouds when icing conditions exist. It is based on the fact that the thickness of the ice that forms on a cylinder exposed for a certain time to freezing drops of a certain size varies with the diameter of the cylinder. When cylinders of assorted diameters are exposed to freezing drops, the ice collected by each cylinder varies with the distribution of drop sizes in the cloud. The distribution is calculated by melting the ice and weighing the water that is recovered from each cylinder. Still other methods, such as direct photography, have been tried, but none is entirely satisfactory.

In the method devised by Wilder drop diameter is indicated by the size and character of splash patterns made when drops strike a sheet of smoked glass. His experiments were done indoors; the method has not yet been tested in a natural rain cloud. The equipment consists of an apparatus stand with a scale for determining the height from which a drop of water of predetermined size falls on a sheet of smoked glass below [see Figure 1]. The drops are released by a pipette attached to the vertical rod of the apparatus stand. The resulting splash patterns are examined with a microscope and measured by a calibrated reticle in the eyepiece.

Pipettes of polyethylene are available in a range of nozzle diameters for releasing drops of a desired size. Droppers can also be made easily by heating a narrow section of glass tubing approximately eight millimeters in diameter until the glass softens; one then pulls the ends apart quickly and breaks the tube at its thinnest point to form two pipettes of identical nozzle diameter. When the glass is pulled to about the diameter of a human hair, drops less than a millimeter in diameter are released if the pipette is filled with distilled water. Drops of maximum uniformity are released if the nozzle is coated with a film of silicone grease. Tubes for releasing proportionately larger drops can be made by breaking off the capillary at points of larger diameter. To make a pipette for releasing drops of maximum diameter hold the end of a length of six-millimeter tubing in a gas flame until it softens and almost closes. Attach to the opposite end of the tubing a rubber squeeze bulb for releasing the drops.

Wilder found that the most distinctive splash patterns are formed on glass that has been treated with a nonwetting agent such as ferric stearate prior to the application of soot. The ferric radical clings to the surface and causes the stearate to clump as a water-resistant film. Wilder places three or four small grains of the compound on a microscope slide and heats the glass until the grains melt. The resulting film is spread evenly with a piece of paper toweling folded into a small square. When the slide has cooled, the glass is buffed with a soft cloth until the waxy streaks made by the stearate disappear. The prepared face of the slide is then smoked by passing the glass back and forth through a yellow gas flame.

During the operation the glass can be held by a pair of tweezers or by a fixture improvised from wire. Reasonably uniform coats of soot can be applied by passing the slide back and forth completely through the flame at the rate of about two strokes a second. This motion, which should be continuous, will also cause the slide to heat uniformly and so prevent the glass from cracking.

A thickness gauge is automatically built into the carbon layer. Any thinly coated slide, when held almost parallel to the rays of a light, reflects one or more bands of iridescent color, depending on the uniformity of the film. The appearance of many bands indicates either that the slide is not being moved through the flame quickly enough or that the flame is too luminous. To correct the latter condition open the air adjustment of the burner. The colors will begin to fade as the thickness of the film increases. The best thickness for making splash patterns is reached when the color just disappears.

Both the height from which the drop is released and the diameter of the drop influence the size of splash patterns. "One day," Wilder writes, "I noticed that the clear inner ring of a pattern made by a drop that fell from a height of 1.5 meters was larger than one made by a drop that fell only 25 centimeters, indicating that the size of the center ring varies with the terminal velocity of the drop. Twenty runs were then made from various heights but the results were meaningless. The outer diameter of the patterns also appeared to increase uniformly with height. A series of runs was then made to check the relation. This dimension, when plotted against the height from which the drops fell, resulted in a straight graph for falls between 10 and 25 centimeters [see Figure 3]. Below five centimeters the diameter decreased disproportionately. There was not enough time to investigate the effect at heights greater than 25 centimeters.

"Similarly, both the diameter of splash patterns and their overall configuration vary with the diameter of drops released from a predetermined height. In other words, the size and configuration of splash patterns bear a direct relation to drop diameter. A test made by releasing from a height of one meter drops that ranged in diameter from two to four millimeters resulted in a linear graph that included the diameter of substantially every drop. Drops of a given size always made similar patterns. I have had no opportunity to check the procedure with cloud droplets."

While Wilder was investigating splash patterns, Wahl developed a miniature wind tunnel for the production of snowflakes. Essentially his device consists of an open-end cylinder lined with cellulose tubes containing water. The assembly is suspended vertically in a chamber maintained at freezing temperature. The temperature difference between the water and the air generates an updraft through the cylinder. Evaporation from the permeable cellulose tubes causes a supercooled cloud to form in the cylinder.

This cloud can be "seeded" with particles of dry ice, just as natural clouds are seeded to trigger the precipitation of rain. The resulting crystals of ice are carried upward by the convection current and grow into snowflakes. Heavy flakes fall to the bottom of the cold chamber, where they can be collected in a shallow dish that contains a film of Formvar solution diluted with chloroform. When the film solidifies, impressions of the snowflakes remain. They can be examined at room temperature.

The cylinder consists of a tin can at least 10 centimeters in diameter and as tall as it is wide. The inner wall of the can is lined with closely spaced casings of the kind used by meat packers for making sausages [see lower illustration at left]. If sausage casings are not available, similar tubes can be made of sheet cellulose by joining the side edges of a sheet to form a tube; the edges are held together with plastic adhesive tape that resists water.

Wahl designed the assembly so that the cellulose tubes could be easily dismantled for replacing water lost by evaporation. He tied the bottom of each tube shut with a rubber band to which was attached a wire hook. When the tube is filled, the top is folded over and clamped shut with a bobby pin. After the filled casing has been placed in the assembly the rubber band is stretched around the outside of the can and attached by the wire hook to the loop of cellulose made by the top fold. The completed tunnel is suspended in the cold chamber by a sling of three wires attached to the upper rim of the can.

A cold chamber can be made of a pair of boxes-an outer one that is well insulated and an inner one, preferably of metal, that transmits heat readily. The chamber can be refrigerated by filling the space between the boxes with dry ice. The temperature need not be lower than 10 degrees Fahrenheit. The outer box can be inexpensively insulated with rock wool or with foam plastic of the kind used as packing material.

At subzero temperature on the Fahrenheit scale the apparatus operates for only short periods. At 10 degrees below zero ice forms on the casings in 15 minutes, and the water freezes solid within 25 minutes. The freezing could be prevented by inserting a short length of nickel-alloy resistance wire in each casing. and applying enough electric current to maintain the water above 32 degrees F. This would allow longer cycles of operation and the production of larger snowflakes.

None of the snowflakes grown by Wahl were larger than one millimeter in diameter. "In repeated cold-chamber experiments," he writes, "many different shapes and sizes of crystals were observed. Most were typically hexagonal, some were nearly triangular and on several occasions nearly all were rectangular. Although the temperature of the chamber could not be regulated, exceptionally large crystals appeared when the temperature of the air was just below the freezing point.

"I observed that the convection current tended to blow the crystals up out of the cloud and over the sides of the chamber. Some crystals were swept back into the updraft at the bottom of the tunnel and made a second trip through the cloud, growing larger en route. This recirculation was obviously desirable in such an experiment. Several schemes were tried in an effort to establish continuous recirculation, but without much success. First a small polyethylene bag with a little hole in the top for seeding was tied over the top of the apparatus. Vapor condensed on the plastic, heat accumulated and the air became stagnant. A second tunnel was then fitted with a cylindrical extension at its upper end. I had hoped that the convection current would lose enough velocity in the extension to prevent the escape of crystals. This arrangement did not work either. Next a funnel extension was tried with substantially the same result.

"Because of limited time I did not have an opportunity to try another idea-the use of an electric field to control the movement of the flakes. Most crystals of ice carry an electric charge that might possibly be used to control their movement. A boxlike arrangement could be improvised with electrodes in the form of wire screening at the top and bottom. This apparatus would be placed in the cloud. A potential difference applied across the screens could be adjusted experimentally to balance 6 the force of gravity as the ice crystals gain weight. With this arrangement it should be possible to grow snowflakes of any desired size."

Roger Hayward, whose illustrations regularly appear in this department, has devised a unique model that simulates the scattering of subatomic particles by a nucleus. "About nine months ago," he writes, "I bought a number of magnetic blocks that appear to be made of finely divided iron particles embedded in plastic. The larger units measure 1 1/16 inches long and 1/2 inch wide; the smaller ones are 1/4 inch square. All are about 3/32 inch thick. I understand that the units are commonly used for clamping papers to bulletin boards made of sheet iron. My objective was to construct a dynamic model for demonstrating some of the properties of a single atom of gas by setting up a rotating multipole magnetic field that would act on a set of steel balls 3/16 inch in diameter. The balls are supported just above the magnets by a level sheet of glass.

"My first model consisted of 14 pairs of 1/4-inch magnets, cemented in an annular groove machined in a disk of clear plastic [see Figure 5]. The stacks were arranged so that the magnetic poles faced alternately up and down. A hole was drilled in the center of the disk to fit a vertical spindle on which the assembly could be rotated. The speed of rotation could be varied continuously from about 10 to 100 revolutions per minute. The steel balls rolled on a thin sheet of plastic, and the plastic was in turn held in place by the ring support of an apparatus stand.

"The arrangement works quite well when the magnets are rotating in the range from 20 to 40 r.p.m. Balls launched squarely across the magnetic field are either captured or deflected in random directions. Balls can be launched in a desired direction at constant and reproducible speeds by means of a small inclined track.

"I next constructed a similar model equipped with the larger magnets, which were rotated by means of a nonmagnetic phonograph turntable [see Figure 6]. The disk that supported the magnets was cut from 1/2-inch plywood and was slotted radially to take 24 magnets. The poles of these magnets appear on the large faces of the rubber strips. Eight radial grooves, which terminate near the center of the disk, each hold an array of magnets assembled so that the north pole of each magnet faces the south pole of its neighbor.

"A second set of radial grooves was cut midway between the first set and spaced about half an inch farther from the center than the first set. Each of these grooves holds two magnets with opposite poles in contact. The combined field of these magnets points in the same direction as that of the inner set of magnets.

"The center of the disk was drilled to make a snug fit with the spindle of the phonograph. I think that the spindle is steel and that there may be a friction clutch of steel under the aluminum turntable. If so, these parts do not appear to distort the magnetic field significantly. I blocked a 12-inch square of window glass over the turntable so that it cleared the top of the magnets by about 1/32 inch. Incidentally, most window glass is slightly concave on one side and slightly convex on the other. The manufacturer's label is usually fixed to the convex side. The glass should be placed over the magnets with the concave side upward. Both the turntable and the glass were leveled to within about .005 inch by means of jacks improvised by threading a set of small blocks to accommodate cap screws. When the balls are placed on the leveled glass, they do not roll unless started, and once they are started they seem to roll equally well in all directions.

"In one sense this model is more limited than the first one because the speed of rotation is restricted to 33 1/3, 45 and 78 r.p.m. On the other hand, the larger size presents a more attractive display. The accompanying illustration [Figure 7 ] depicts the typical deflection path of a 'particle' that penetrates the 'nucleus,' the capture and subsequent emission of a particle and the trajectory of a sharply deflected particle. When the field is rotated at 78 r.p.m., a large number of balls are captured. The balls seem to become sufficiently magnetized to attract each other. At 45 r.p.m. the mass becomes unstable and a few balls escape. When the speed is reduced to 33 1/3 revolutions, the mass becomes very unstable, and within seconds balls are scattering in all directions.

"Balls occasionally enter the capture zone and are immediately ejected. One would be hard put to show a statistical relation between the trajectory of the balls and the nature and size of the magnetic array. I suppose it would be possible to demonstrate some effective diameter of the nucleus by a statistical analysis of the trajectories. The stack of three magnets that appears in the illustration at the half-radius point merely creates activity when there are a number of balls on the glass. These magnets might represent another near atom that would be expected to detect free particles."

Bibliography

PHYSICAL METEOROLOGY John C. Johnson. John Wiley & Sons, Inc., 1954.

THE PREPARATION OF SNOW CRYSTAL REPLICAS. Vincent J. Schaefer in Weatherwise, Vol. 9, No. 4, pages 132-135; June, 1956.

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.