Berti's tube extended up the side of a multistory building and therefore was too big to serve as a useful barometer. A few years later his compatriot Evangelista Torricelli proposed that the water be replaced with mercury, which is 13.6 times denser, so that the instrument's height could be reduced. Torricelli's associate Vincenzo Viviani constructed such an apparatus. He sealed one end of a glass tube about 30 inches long and filled it with mercury. Pressing a finger over the open end, he inverted the tube and lowered its end into an open container of mercury. When he released his finger, mercury flowed from the tube until the pressure from the weight of the mercury left in the tube matched the atmospheric pressure. This instrument was the first true barometer. By noting the height of the mercury column one could measure the atmospheric pressure. High pressure pushed mercury into the tube, increasing the column's height. Low pressure allowed mercury to leave the tube, decreasing the column's height.

Because mercury is so dense, the variations in column height are usually small and sometimes must be recorded in measurements of less than a millimeter. Why not stick with water as the working fluid? Although the barometer would be quite tall, the variations in atmospheric pressure should yield changes in column height 13.6 times greater than those in a mercury barometer. Unfortunately the water barometer falls quite a bit short of the barometric ideal.

The explanation for this failure lies in the fact that the space above the water column is not a full vacuum as Berti conjectured. Some of the water vaporizes, filling the space with vapor. Whenever the temperature changes, the amount of vapor varies. Hence the pressure the vapor exerts on the water column also varies. For example, in warm weather the increased vapor pressure reduces the height of the column. The device is therefore not accurate as a barometer.

Sam Epstein of Los Angeles has found a way around this difficulty. His barometer contains common antifreeze (ethylene glycol) diluted with water. The antifreeze reduces vaporization at the top of the barometric column. The barometer's residual dependence on temperature is accounted for by a table that Epstein formulates when he calibrates the instrument.

Epstein has made two such barometers. One of them is described as a barometer of the cistern type and the other is a barometer of the siphon type. The cistern barometer is similar in design to Berti's apparatus. To accommodate the low density of the antifreeze mixture (1.07 times the density of water), the device stands about 35 feet high. Since monitoring the fluid level can be a nuisance if one must climb several flights of stairs, Epstein designed a remote sensor to read the column height from the base.

If you want to build a similar apparatus, your first move should be to find a suitable place for it to stand. It should be fastened to a wall by brackets. It should not be in direct sunlight because of the heating effect. Although the freezing point of the antifreeze mixture is low (about-34 degrees Fahrenheit), you must be careful not to mount the device where the mixture might freeze in winter.

The barometer is made out of six sections of polyvinyl chloride (PVC) tubing, each section six feet long and one inch in internal diameter. If you decide to do without a remote sensor, make the top section of transparent plastic. Join the sections with union pieces sealed with primer and cement. Be certain the seals are tight; after the cement sets, the sections can be separated only by sawing them apart. For the working fluid mix four gallons of good-quality antifreeze with an equal amount of distilled water.

The sensing elements are 24 platinum electrodes spaced along the top section. Each electrode is numbered according to its height. An additional electrode, mounted on the opposite side of the section, serves as a common electrode. When a voltage is applied between the common electrode and a numbered one, a small current is conducted through the antifreeze solution. The numbered electrodes are connected to a rotary switch at the base of the barometer. When you turn the switch, you put a different electrode in the circuit. By rotating the switch to find the highest electrode that conducts electricity you can ascertain the approximate height of the fluid.

In order to make the electrodes, buy from a laboratory-supply house a nine-inch length of platinum wire (No. 24 gauge). Also get 25 glass tubes, each tube six inches long and five millimeters in internal diameter. Cut the wire into nine-millimeter sections and bend each section into the shape of a Z. Its two outer arms should each be four millimeters long. Heat the end of a glass tube in a gas-and-compressed-air flame from a Meker burner. Hold the end just above the tip of the blue part of the flame, rotating it slowly to provide uniform heating. The end soon softens and closes. When the opening is only slightly wider than the arms of the wire, insert a wire section into the tube and continue heating the glass so that it seals around the middle section of the Z.

When the tube closes around the wire, gradually diminish the air supply to the burner while rotating the tube in the flame. After another minute remove the tube from the flame and let it cool. Be careful not to overheat the wire or it will melt, ruining the device.

After the tube has cooled cut the end containing the wire to a length of one inch. Then fire-polish the ragged end left by the cut.

Each electrode is connected to the rest of the sensor circuit by hookup wire, which has strands and is insulated. Have a six-inch length of this hookup wire handy with one end stripped of its insulation. Put approximately half an inch of 50/50 rosin-core solder the electrode. Hold the section in a gas flame until the solder melts. Then quickly remove the section from the heat and push the stripped end of the hookup wire into the molten solder. After the device has cooled make sure the wire is firmly embedded; if it is not, repeat the procedure. When the wire is properly set, seal the platinum end of the tube with PVC cement.

Prepare the electrodes and drill holes for them in the top section of PVC tubing. Make each hole just large enough to accept an electrode. Space the centers of the holes for the numbered electrodes at 2.4-inch intervals, starting about a foot below the top of the tubing. Then seal each electrode with PVC cement. When the barometer is later assembled, put the highest electrode 33.4 feet from the level of the fluid in the cistern.

From the stub of hookup wire at each electrode run enough additional wire to reach the base of the barometer. Each wire is attached to a resistor connected to one of 24 positions on a rotary switch. Number the electrodes, beginning with the lowest one. The resistors connected to the lowest six electrodes are rated at 90 kilohms. The higher electrodes, which are farther from the common electrode, require resistors with less resistance. Group the 24 electrodes into four sets of six. Decrease the resistance by 20 kilohms for each successively higher set. Each of the resistors has a power rating of .5 watt.

The rotary contact of the switch is connected to a double-pole, single-throw toggle switch. A wire runs from the toggle switch to the common electrode. Along the route install a small neon lamp with no internal resistor; it signals when current is conducted through the antifreeze solution. The toggle switch is connected to an isolation transformer that is plugged into a wall outlet. When you want to take a measurement, turn on the toggle switch. (It should be off at other times.) Select a position for the rotary switch. If the neon lamp lights, turn the switch counterclockwise until the lamp goes out. Otherwise turn the switch clockwise until the lamp goes on. The last position of the switch indicates the height of the fluid in the barometer to an accuracy of approximately two inches.

The rest of the construction does not depend on the remote sensor. Mount the PVC tubing on a wall and plug its lower end with a rubber stopper taped on or with a threaded cap. The cistern is a wide-mouthed two-gallon jar with a screw-on lid. Drill in the lid a hole slightly larger than the tubing. Half fill the jar with the antifreeze mixture. Slip the lid over the lower end of the tubing and have someone hold it well above that end.

Raise the support of the jar until the lower end of the tubing is in the antifreeze mixture. Put a funnel into the top of the tubing and slowly pour in the antifreeze mixture, trying to avoid air bubbles. After the tube is full allow 30 minutes for any air bubbles to dissipate. Then cap and seal the top of the pipe. Unplug the lower end and screw the lid of the jar into place. Allow one day for the vapors of ethylene glycol and water to saturate the region above the fluid column.

If you do not build the remote sensor, add a scale alongside the upper end of the fluid column by which you can measure the atmospheric pressure. The height of the fluid column is measured from the fluid level in the cistern. The scale should extend in height from 28.8 to 33.4 feet, the same range covered by the placement of electrodes for the remote sensor. If the cistern is wide enough, the fluid level in it hardly changes as atmospheric pressure varies, and so you can attach the scale to the barometer.

The scale can be marked in pressure units called inches of mercury. The full length of the scale, 4.6 feet, represents a pressure change of 4.6 inches of mercury. Mark the scale in .1-foot sections. Each section then represents a pressure change of .1 inch of mercury. For example, if the atmospheric pressure increases so that the column of antifreeze mixture climbs by .1 foot, the column in a mercury barometer would climb by .1 inch. (You could mark the scale in any other units of pressure.)

If the barometer is exposed to a considerable range of variations in temperature, you will need to record the height of the fluid throughout the entire range of temperatures. Compare the reading with the atmospheric pressure measured by means of some other device, such as a mercury barometer, that is not appreciably influenced by temperature. At very low temperatures the vapor pressure above the fluid hardly alters the height of the column. At higher temperatures the vapor pressure lowers the height.

After you have constructed a table of corrections for the temperature, you no longer have need for the second barometer. Each time you make a reading on the antifreeze barometer note the temperature on a thermometer placed next to the base. Use the table to determine the correction that must be added to your reading to get the true atmospheric pressure.

Epstein's second device, the siphon barometer, is read at the base without a remote sensor and does not require a cistern. It is in the shape of a U but with one side shorter than the other. The short side is made of transparent plastic pipe that is six feet long with an internal diameter of one inch. The long side consists of seven sections of PVC tubing. The sides are connected by PVC elbows and a short length of PVC tubing. When you have cemented the tubing, pipe and elbows together, mount the device on a wall, insert a funnel in the short side and pour in enough antifreeze mixture to fill both sides to a height of five feet.

Now a calculation must be made to ascertain how much additional fluid will be needed to balance the normal atmospheric pressure at your altitude. At sea level the normal atmospheric pressure will balance the weight of 29.92 inches of mercury. To find the equivalent height of the antifreeze mixture, multiply the height of the mercury by the ratio of the density of mercury to the density of the antifreeze. Dividing the result by 12 to convert it into units of feet, you will find that the equivalent height of the antifreeze mixture is 31.7 feet.

Less fluid is needed above sea level, more below it. Determine your elevation with respect to sea level. For every 100 feet above sea level subtract 1.3 inches from the benchmark height of 31.7 feet. For every 100 feet below sea level add 1.3 inches to the benchmark height. When you have determined the appropriate column height for your elevation, calculate the volume of fluid you will need to fill the PVC tubing to that height.

To add fluid to the barometer begin at the short side. Bring the fluid level up to the top of the pipe and close that side with a rubber stopper. Insert a glass stopcock through a hole in the stopper, the bore of the stopcock should be two or three millimeters in diameter. Close the stopcock, push the stopper into place, seal it with electrical tape and add the rest of the fluid to the long side of the barometer.

Allow 30 minutes for air bubbles to escape from the fluid and then attach a bicycle air pump to the protruding end of the stopcock. As you open the stopcock, pump air into the barometer, forcing the fluid level in the long side to the top. Immediately close the stopcock. Cement a cap to the top of the long side. When the cement has set, slowly open the stopcock. Wait one day for the vapors to saturate the space above the fluid in the long side. Replace the stopcock and the rubber stopper with another stopper through which you have inserted a two-inch glass tube with an internal diameter of one millimeter. Fasten the stopper with tape and cover it with a cap that is loose enough to pass air.

The scale for measuring the atmospheric pressure is mounted along the short side. It should run in the opposite direction from the scale on the cistern barometer. The reason is that an increase in atmospheric pressure drives the fluid down in the short side. Since the pressure variations are split evenly between the two sides of the barometer, the distance between marks on the scale is half the distance on the scale for the cistern barometer. Hence a change in height of .05 foot now represents a pressure change of .1 inch of mercury.

To calibrate his barometers Epstein built a mercury barometer. Since he is a chemist, he is well acquainted with the precautions that should be taken in work with mercury. If you decide to build a mercury barometer, you should get advice from someone similarly knowledgeable about mercury. Do the work in a well-ventilated area. Do not touch the mercury or get any of it in your mouth.

Epstein's barometer holds about two pounds of mercury, which he bought from a dental-supply house. The backboard support for the barometer is half-inch wood mounted on a wall with two picture hooks, each one strong enough to support 50 pounds. The barometer must be exactly vertical. The cistern, a jar with a metal screw-on lid, sits in a well that is supported by a wood extension from the backboard. Drill the extension so that a round steel stock one inch in diameter can be press-fitted through it. Drill and thread the stock to accept a 3/8-by-16-inch steel bolt, which should be greased before it is put in place. The end of the bolt fits securely in a threaded hole at the bottom of the well.

Two holes are drilled in the lid of the cistern. One hole, exactly centered, is slightly larger than the mercury tube that fits through it. The other hole is slightly larger than a plastic knitting needle that extends into the cistern. A rubber liner cemented inside the lid has similar holes but fits tightly around the mercury tube and the needle. The holes are punched with a laboratory cork borer. Add a barely visible V notch to the hole through the liner for the mercury tube. The notch passes air so that the air pressure inside the cistern matches the external air pressure. The notch must be small to keep the escape of mercury vapor to a negligible amount.

A support for the needle extends from the backboard. The tip of the needle indicates the height of a reference level scratched on the backboard. Whenever the atmospheric pressure is read, the height of the cistern is adjusted until the level of the mercury in the cistern is at the tip of the needle.

The tube holding mercury rests inside a guard tube made from thin-walled electrical conduit with an internal diameter of three-quarters of an inch. Make the guard tube 34.5 inches long. It has two viewing slots near the top through which the mercury level can be monitored. A scale marked in inches or millimeters is positioned along the front side of the guard tube. The wall behind the slots should be white to provide contrast with the mercury. A cap for the guard tube has a No. 10 eyebolt mounted in its center and is fastened to the tube by three sheet-metal screws.

Two supports for the barometer extend from the back wall. One of them is made of wood; its outlet end is cut to fit around the lower end of the guard tube. A clamp made out of sheet metal holds the tube snug against this wood support. The other support is a 3/8 inch bolt that goes through the backboard, extends over the barometer and passes through the eyebolt. The eyebolt is held in position by nuts on the support bolt.

The mercury tube, made of flint glass, is 35 inches long and five millimeters in internal diameter. For safety Epstein decided that the wall thickness should be at least one millimeter. He sealed one end of the tube with a Meker burner. Three half-inch wood dowels were slipped over the tube to serve as spacers in the guard tube. They have a centered hole slightly larger than the mercury tube. Two of them are glued in place with epoxy at the middle and bottom of the tube. The third dowel is glued 4.5 inches below the open end of the tube.

In filling the mercury tube, Epstein held it by clamps so that its sealed end was at the bottom. To catch any mercury that might spill he placed a wide glass container holding about half an inch of water under the tube. He also took the precaution of filling the tube in a carpetless room; if mercury spilled, he could amalgamate it with dust or filings of copper or zinc moistened with vinegar.

Epstein filled the tube by slowly pouring the mercury into a funnel held in place by a piece of rubber tubing in the open end of the tube. To facilitate the flow of mercury he inserted a four-foot length of piano wire into the tube through the funnel. Tapping the tube with a pencil, he worked the wire up and down, being careful not to scratch the glass. He continued the process until the tube was filled to within an inch of the top and the air bubbles in the mercury were eliminated.

With the help of an assistant he inverted the guard tube, placed its lower end on a soft surface, slipped the tube of mercury into it and pressed a wood dowel into the upper end of the guard tube. After sliding the inverted cistern lid over the protruding open end of the mercury tube, Epstein completed filling the tube with mercury. Wearing a surgical glove, he closed the open end of the mercury tube with a finger as he and his assistant inverted the assembly and lowered the tube into the mercury in the cistern. He removed the finger when the mercury tube was below the fluid level in the cistern. Being careful not to let the lower end of the mercury tube move above the fluid level in the cistern or touch the bottom of the cistern, he clamped the guard tube to the support extending from the back wall. Then he placed the cap on the guard tube and adjusted the height of the tubes until he could run the support bolt through the eyebolt.

After screwing the cistern lid to the cistern, Epstein positioned the knitting needle. To make a reading he rotated the bolt below the cistern until the fluid level in the cistern was at the tip of the needle. He then read the height of the mercury column to an accuracy of about one millimeter.

Bibliography

THE HISTORY OF THE BAROMETER AND ITS USE IN METEOROLOGY. W. E. Knowles Middleton. The Johns Hopkins Press, 1964.

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.