Handbook of Meteorology/Thermometers
PART II
THE MEASUREMENT OF TEMPERATURE:
THERMOMETERS
Quantitative measurements in temperature, based upon the calorie, have a definite place in physics; and those based on the British thermal unit have a broad application in various economies. Human sensitiveness to heat does not pertain to quantity but to intensity. A large block of ice at 30° may contain more heat, quantitatively considered, than a red-hot horseshoe. If carried into a room whose air was far below freezing, the ice might warm the room to a greater degree than would the horse-shoe. Humanity, and indeed, all living things require the intensity of heat that enables living organisms to function naturally and properly. The vital questions therefore are—“How warm is it?”—or, “How cold is it?”—or, “Is physical comfort satisfied?” These conditions depend upon intensity of heat—that is, upon temperature.
Temperature.—The term temperature has a broad application. It is an expression of the varying warmth of earth, air and water, with relation to life.
The unit of temperature is a degree, a measured part of the expansion which a column of mercury within a tube undergoes when heated from the melting point of ice to the boiling point of water at sea level.[1]
Three scales of degree measurements are more or less in use. In the Reaumur scale, now rarely used, the expansion is divided into 80 parts; in the centigrade, into 100 parts; in the Fahrenheit, into 180 parts. In the Reaumur and the centigrade scales the zero, 0°, of temperature is at the melting point of ice; in the Fahrenheit, 32 degrees below it. The centigrade scale is used chiefly in Continental Europe. The degree values are inconveniently large and, in many cases, fractional units of the degree must be expressed. Winter temperatures require the use both of positive and negative quantities, and this adds to the labor of computation and to the likelihood of error.
Absolute Temperature.—The fact that neither the centigrade nor the Fahrenheit scale per se expresses the relation of the volume of a gas to its temperature has led to the establishment of a theoretical absolute zero of temperature. The following demonstration and the accompanying cut explain the method of its determination. A glass tube about 50 inches long and closed at one end contains a free-moving piston of mercury, resting about 20 inches from the open end of the tube. The tube is first placed in a container filled with melting ice. When
The empiric determination of the absolute zero of temperature.
the piston has reached the low point its position is marked. It is then transferred to a container of boiling water, and the position of the mercury piston is again marked. The amount of expansion is divided into 180 equal parts or units. If, now, the distance between the first mark and the lower end of the tube be measured, it will be found to contain almost precisely 459 of these units. Each unit corresponds to 1 degree Fahrenheit. Hence from this experiment absolute or natural zero would be—459.4°. On the centigrade scale it is—273.13°. Absolute temperature is commonly expressed as A°; or 459.4°.
The natural zero deduced from an investigation of the pressure of a gas at constant volume has the same scale value as the absolute zero. It is inferred, in consequence, that the absolute or natural zero is the temperature at which molecular motion ceases.
The Thermometer.—The ordinary thermometer consists of a bulb or reservoir fused to the end of a glass tube or stem. The tubing from which the stem is drawn is a wedge-shaped prism with a strip of white enamel fused to the convex surface. When the prismatic tube is drawn out into thermometer stems the wedge angle becomes a lens which magnifies the fine thread of mercury; the enamel becomes an opaque, white background against which the thread of mercury is plainly visible. The bulbs of ordinary thermometers are commonly blown at the end of the drawn tube. Those of the best thermometers are made of a specially constructed glass and are fused to the end of the drawn tube.
Most solids, in cooling from fusion or from intense heating, suffer what is known as “hysteresis”—that is, molecular changes continue for a considerable time. These changes alter the size of the bore of the tube. In order to overcome them, the tubes of thermometers of the highest grade are laid away to “season” or “temper” for a period of two years. The shrinkage of an unseasoned tube is likely to cause the readings to register as much as 6 degrees too high.
The bore of the thermometer is microscopic in diameter; in thermometers graduated to fractions of a degree, it may be less than o.ooi inch; ordinarily it is from 0.002 inch to 0.005 inch. In the construction of precision thermometers the bore is measured under the microscope, and a bulb of hard glass of the required size is fused to the end of the tube. Cylindrical bulbs are preferable to spherical bulbs; they present a greater surface and therefore are more sensitive. The expansion and contraction of the glass with changes of temperature is somewhat greater than with spherical bulbs, but thermometer scales are compensated for this correction.
The bulb of the thermometer is usually filled with mercury at the time it is fused to the stem. While hot, the open end of the stem is inserted in a vessel containing pure mercury. As the air in the bulb cools, its contraction causes a small quantity of mercury to be forced through the bore of the stem into the bulb. The mercury in the bulb is then heated to its boiling point and the open end of the stem again dipped into the mercury. This process is repeated until both bulb and steam are completely filled.
The mercury and the tube are apt to contain a minute quantity of moisture. If this remains it is pretty certain to cause a broken column of mercury in the tube, thereby rendering the thermometer imperfect. To prevent this, the filled tube is kept for some hours at a temperature above the boiling point of water. The “roasting” requires care and experience.
When the roasting process is completed the bulb is again heated and as soon as the mercury is expanded to the top of the stem the latter is sealed, leaving an angle or “hook” at the top. The hook holds the tube fast to the scale.
The tube is now ready to be graduated. For this purpose it is placed successively in brine at 2°, melting ice at 32°, and water baths at 62° and 92°. The position of the top of the column is marked for each temperature and usually for each tenth degree of the scale. An engine ruling machine divides each division into 30 parts. The division marks are also scaled below 2° and above 92°. The metal scale is subdivided according to the marks on the stem.
Several manufacturers produce three grades of instruments. On those of the first class minute spots may be found at the various testing temperatures. If the marks do not appear, the thermometer is not trustworthy for precision purposes. Weather Bureau thermometers are specially tested, and a certificate showing the error for each 10 degrees accompanies each instrument. A thermometer which does not comply with these requirements is not a standard instrument.
Thermometers of the second grade are not engraved on the stem; the divisions are on the metal scale to which the tube is attached. The tubes are seasoned; but trifling inaccuracies unfit them for use where precision is required. They are sufficiently accurate for ordinary uses.
Thermometers of the third grade are “rejects” and rarely bear the maker’s name. The inaccuracy is always more than 1 degree, and it is likely to vary in different parts of the scale. In many instances these thermometers are sold to jobbers and retailers who stamp fictitious names on them. Experience and practise will usually enable one to discover the maker. The scope of thermometer scales varies greatly. Practically every industry employs temperature measurements which require specially constructed thermometers.
Weather Bureau Thermometers.—Several kinds of thermometers are necessary for the requirements of a weather service; and these are practically the same in all parts of the world where a weather service is maintained. In weather stations the daily maximum and the daily minimum are required; a continuous graphic record is desirable, but the instrument for this purpose is supplied to stations of first-class equipment only.
A standard thermometer of the ordinary type, that is, one. which shows existing temperature at any time, is desirable. A minimum thermometer can be used for this purpose, but a standard instrument is preferable. The scale, divided to single degrees, should be engraved on the tube and on the metal
Maximum and minimum registering thermometers.
Weather Bureau patterns.
strip as well. Readings are made to the nearest degree mark. If the fraction is exactly half a degree the preceding figure, if odd, will be increased by 1 degree; if even, it will remain unchanged.[2]
The maximum thermometer is so called from the fact that the mercury in the tube is not drawn back into the bulb when the temperature lowers. The expansion of the mercury in the bulb forces the flow into the bore, as with ordinary thermometers. A slight constriction of the bore at the top of the bulb prevents a backflow, thereby leaving the mercury in the bore at the maximum temperature since its last previous setting. Usually the maximum temperature of the day occurs between 2:30 p.m. and 4:00 p.m. and the thermometer should be set late in the evening or early next morning so as to record the maximum for the next day.
To insure accuracy of readings the thermometer should not be higher than the eye of the observer; and to avoid error of refraction by the lens front, the observer should stand squarely in front of the end of the mercury column. An error of refraction may amount to half a degree.
The “setting” of the maximum thermometer is accomplished by whirling it around a stud and bearing at the end opposite the bulb; centrifugal motion forces the mercury past the stricture, and back into the bulb.
The maximum thermometer of the Weather Bureau type should rest in a nearly horizontal position, the bulb slightly higher than the farther end of the tube. Ordinarily this will prevent any back flow. Occasionally, however, a maximum thermometer fails to leave the mercury in the bore at the point of maximum expansion; for reasons not fully explained the column of mercury is drawn back toward the bulb as the temperature falls. A maximum thermometer of this sort is known in Weather Bureau cant as a “retreater.” The retreating of the column of mercury may be overcome by a slight increase in the elevation of the bulb. When, however, it is discovered that the retreating is habitual, the thermometer should be returned to the maker.
Some maximum thermometers set easily; others require to be whirled vigorously. Observers differ in their likes and dislikes. If the column of mercury moves very easily there is always danger of error. If the tube is held in a horizontal position there is the possibility of a slight retreat of the column. And if the bulb is elevated there is always the possibility that it may slide toward the far end of the tube. The maximum thermometer readings are not trustworthy for any except maximum temperatures. When set, it should agree with the reset minimum thermometer.
It is advisable to bring the thermometer very carefully to a vertical position when the reading is made. This corrects at once any sliding of the column that may have occurred. But there is always danger that a part of the mercury in the tube may flow into the bulb when the thermometer is brought to a vertical position; this is most likely to occur in hot weather when the column is long.
Taking everything into consideration, a maximum thermometer that requires a moderately vigorous whirling is preferable. The instrument with which the observer can work to best advantage is the one with which he should be provided.
The minimum thermometer is for the purpose of registering the lowest temperature between settings. It is exposed in a nearly horizontal position with the bulb slightly higher than the opposite end of the tube. Inasmuch as winter minima are sometimes lower than the freezing temperature of mercury, and because the instrument here described requires a transparent column, minimum thermometers usually contain alcohol instead of mercury. The alcohol of American Weather Bureau thermometers usually is colored; in most foreign made instruments it is uncolored. In the matter of visibility the gain of a colored liquid is material, but coloring matter is not essential. In spite of care, a precipitation of the coloring matter occasionally may occur, and this is likely to cause a slight constant error. The space in the tube above the alcohol contains air, more or less saturated with alcohol vapor.
The essential feature of the minimum thermometer is a small black index within the bore and also within the liquid. As a lowering temperature contracts the liquid column, the cohesion of its surface drags the index toward the bulb. When the liquid expands, however, it flows around the index without moving it. The index therefore shows the lowest temperature between settings. The minimum temperature is read at the end of the index farthest from the bulb. The temperature may be read from the minimum thermometer at any time, reading from the end of the column of alcohol, and not from the index.
The minimum thermometer is usually attached to a strip of brass, bent so that the instrument is held about an inch from the board support. It is fastened so that the end containing the bulb may be swung to an inverted position. The maximum thermometer is fastened to the same support. The stud on which it whirls is about 2 inches long. The free end rests on a pin which is removed when the thermometer is set.
Instead of the fasteners described above, clamps, sometimes called the Townsend supports, may be used to hold the thermometers. The clamps are fastened to the board support and permit the setting of the thermometers. The clamps are issued as a part of station equipment.
Care and Adjustment of Thermometers.—The thermometers, being exposed to the weather, accumulate dust; the metal parts may become tarnished or even rusty. It is advisable to use a soft camel's hair brush for removing the dust, and this should be done two or three times a week. When occasion requires, a polishing brush may be used on the metal parts. It is more desirable to prevent than to remove rust and tarnish.
Unless the maximum thermometer becomes a retreater, it is not likely to get out of order. Even if drifting snow blown into the shelter incrusts it, no damage is likely to result. It is better to allow the snow to melt off than to attempt to remove it by force.
If a maximum thermometer has not been set for a long time, a break in the column which refuses to unite may result. The same may occur if the moisture has not been wholly expelled during the roasting process. In such a case it is usually possible to drive the space to the small chamber at the end of the tube. It may be driven into the bulb; if this is done the break is likely to work back into the column again. If the instrument is held in a vertical position, bulb down, at a distance of 1 or 2 inches from a table, and is allowed to fall with vertical blows so as to hit a thickness of blotting paper placed on the table, the broken space gradually displaces the mercury until it reaches the top of the column.
A break in the column of mercury in the tube is not necessarily a defect; it is only when the break will not close—that is, when it leaves an open space—that error in the reading results. In such a case the thermometer should be discarded, or else returned to the maker for repair.
The minimum thermometer is usually out of order when it is received at its destination. The index may be fast at the top of the tube, or in the bulb; most likely the alcohol column is broken, a half dozen or more bubbles occurring; possibly some of the alcohol is lodged in the chamber at the farther end of the tube.
To loosen the index, tap the edge of the metal scale with a small piece of wood—say, a clothes pin—until it becomes free.
With bulb end down, let the thermometer fall vertically an inch or more so that it strikes endways on the table or the shelter floor. Little by little the alcohol will flow along the tube; the broken parts of the column become shorter; and the bub- bles disappear.
If this fails, hold the thermometer at its upper end and bring it down forcibly as though striking hard blows with a hammer—being careful, however, not actually to strike anything. It may require vigorous exercise, but the centrifugal force will finally bring the broken parts to the rest of the column. It usually requires from a quarter to half an hour to put a minimum thermometer with a broken column in order. Great care must be used that no part of the alcohol remains in the chamber at the farther end of the tube.
Thermometer Shelter.—Maximum and minimum thermometers should be sheltered from the sun and from direct contact with precipitation of every sort. They also must be placed so that they are in contact with free air. They must be sheltered from heat radiated from buildings, metal roofs, and pavements. The board support should not be attached directly to the wall of a building; if on a porch it should be attached to an outrigger that leaves a space of a few inches from the building. A wide, covered porch with a northerly or an easterly exposure is the best position about a building.
The daily maxima on the south side of a house, within 3 feet of the wall will be from 2 degrees to 6 degrees higher in clear weather than those on the north side, close to the house. The minima will vary but little. If only a window exposure is available, a north-facing window should be selected, and the thermometers should be screened from the window if the room is heated. There should be several inches of space between the shelter front and the window.
In cities, the flat roof of a building frequently offers the best position for thermometers. A graveled roof reflects less heat than a metal roof, and should be preferred when possible. In any case the shelter should be placed where the thermometers are not affected by heat reflected from nearby walls. The best position on the roof must be determined by judgment and experience.
In open country and sparsely built localities, the shelter built after the plans recommended by the Weather Bureau 
A Thermometer Shelter.
It is placed in the shade of tall trees, and receives direct sunlight a few hours in the morning only. should be used. This consists of a miniature house, 3 by 2 by 2 feet, with louvered sides and a removable top. The front may be let down when readings are made. The front should face the north. The shelter rests on braced legs which should be anchored firmly to the ground. The top of the shelter is likely to become hot enough to radiate heat to the thermometers. This may be prevented in part by a double roof with an air space, or by covering the roof loosely with asbestos cloth or with lino- leum. There can be no objection to placing the shelter in the shade of a tree that shields it from the afternoon sun, provided it is not less than 8 feet from trunk and branches.
In locating a place for a shelter it is a good plan to use a second thermometer in various positions, checking and comparing maxima and minima. Reflection and absorption sometimes bring about unexpected results. Observers with experience are alert to these possibilities; the inexperienced observer must learn them. In general, if the shelter is distant twice the height of an object there will be no errors caused by reflection or by absorption from that object.
Anomalies of Temperature.—As a rule, minimum temperatures—and they usually occur just before sunrise—are less apt to be affected by unusual conditions than are maximum temperatures. The minimum temperatures on the south side of a building are usually the same as those on the north side. In prolonged hot spells, however, this does not always hold true. The walls of the southerly exposure may absorb so much warmth during the day that not all of it is radiated at night. As a result, a minimum registered under such conditions will be too high.
The prevalence of a stiff wind, especially the northwest wind of cold waves, equalizes temperatures to a remarkable extent; the minima of stations covering considerable areas rarely vary more than 1 or 2 degrees. On the other hand, on very still nights the minima of stations only a few miles apart may vary several degrees.
On very cold, still nights cold air tends to settle by gravity into low spots. This condition is so marked that the minima of localities only a few rods apart may vary as much as 2 or 3 degrees. This difference is very noticeable in mountain valleys where the cold air is apt to flow down the valleys at night. Frosts occur much more frequently along valley floors than in the foothills and the benches higher up.
City and suburban temperatures usually have about the same daily means, and their monthly averages should not vary more than a degree. The daily maxima and minima not infrequently vary several degrees. This is due chiefly to the fact that the less amount of smoke and dust in suburban localities favors absorption of heat in the day time and permits radiation at night.
In early fall and also in late spring, frost may be observed in sheltered places on the ground when the thermometer registers several degrees above freezing. An observer may therefore conclude that his minimum thermometer is not registering correctly. It is not likely than an error has occurred; ground surface temperature, especially on northerly exposures, may register several degrees—on occasions, as many as 10 degrees—lower than the thermometers 6 feet or more above the surface.[3]
Thermometers on the business streets of cities, especially those in which the blocks are solidly built, register from 3 degrees to 6 degrees too high as a rule, owing to radiation and reflection from nearby buildings. They indicate the temperature of the street, but not of the free air.
The Six Maximum and Minimum Thermometer.—A maximum and minimum thermometer of the Six pattern consists of a glass tube bent in two or three sections as shown in the accompanying figure. The tube in the center is a cylindrical bulb about 0.1 inch internal diameter; the bulb at the top of the right-hand column is large enough to have a volume of about 1 cubic centimeter. The bore in the U part of the tube is about 0.02 inch in diameter; it is filled with mercury. The central bulb is completely filled with a solution of creosote, or with alcohol. The expansion of the liquid in the central bulb pushes the mercury down on the left side of the U and up on the right side; it also pushes liquid into the air bulb on the right side, slightly compressing the air and vapor in the bulb. Lowering temperature causes a contraction of the liquid in the central bulb, thereby drawing back the mercury in the U. This is made more positive by the compressed air and vapor in the right-hand bulb. The scale reads downward on the left and upward en the right side. These are marked respectively “cold” and “heat,” or “night” and “day.”
Maxima and minima are recorded by separate indices within the bore of the tube. The indices are pieces of steel wire coated with glass—in some thermometers they are plain wire—each armed with two appendages. On one end the appendage points upward; on the other, downward. Their object is to hold the index lightly to the place in the bore to which the mercury pushes it. Pushing the indices is the only work the mercury in the U tube performs. Rising temperature pushes the index in the right-hand tube upward; falling temperature pushes the index in the left-hand tube upward. The indices are set by the use of a small magnet which accompanies the thermometer. The poles of the magnet are hollow-ground, so as to fit closely to the tube.
In many respects the Six thermometer is preferable for ordinary 
Maximum and
Minimum Thermometer.
Six’s pattern uses. It is not so likely to be broken as the regular Weather Bureau thermometers; it is very readily set; and it is more nearly “fool-proof” than the delicate Weather Bureau instruments. It is not so sensitive as the Weather Bureau thermometers; it is slow in registering; and the indices are occasionally caught in the mercury from which they are separated with difficulty. A violent jar may break the hair-like appendages that cause the indices to register. If this happens to the index in the right-hand tube its repair by a thermometer maker is possible; if in the left-hand tube the case is hopeless.
In selecting a thermometer of this type one should note first whether or not the readings of the two tubes are the same. When a thermometer has lain edgewise, or on its side, for a number of days—and this may occur when it is in transit on a railway—a flow of the liquid past the mercury, from one tube to the other, may take place. As a result the readings on the two sides do not coincide. An expert in the mechanics of thermometry can make the necessary adjustment, but it should be done by an expert and not an experimenter. In selecting a Six thermometer, a comparison of the readings of the two sides should be the first care.
The Six thermometer may not be quite up to the standard of accuracy. If the error is small the thermometer needs not be condemned, however, for an allowance can be made therefor.
The Thermograph.—The thermograph is both a registering and a recording thermometer. The essential part of the mechanism consists of two thin strips of metal having different coefficients of expansion. The metal strips are brazed or soldered surface to surface, bent to a quadrant or curled into a coil, and annealed. The type used by the United States Weather Bureau consists of a curved flat tube filled with mercury or with alcohol. In either type of thermograph expansion causes a warping of the metal which is communicated to a lever, whose long arm is a recording pen.
The recording part of the thermograph is a drum containing a clock. The clock is geared so as to cause one nearly complete
Thermograph—nigh drum.
revolution of the drum in a week. The slight shortage of a complete revolution is an allowance for the margin of the fastening of the paper on which the record is made.
The paper strips upon which records are made are about 12 inches long. They vary in width according to requirements. Horizontal lines lithographed from engine-ruled plates divide the width of the strip into degree spaces. Arcs of circles, whose radii are the length of the pen, divide the length of the strip into day spaces, each of which is subdivided into two-hour intervals. High drum record sheets are ruled for temperatures varying from −50° to 120°; low drum strips are usually ruled from 0° to 100°.
High drum thermographs are used very generally in meteorological work. Low drum strips have all the temperature range necessary for greenhouses, refrigerating establishments and freight cars containing perishable goods. They are used in many Weather Bureau stations where the yearly range does not materially exceed 100 degrees. For most stations, and for general military use, a high drum thermograph is advisable.
Thermographs are not so accurate as standard thermometers. In very damp weather the expansion and swelling of the paper on which the record is made affects the accuracy. Inasmuch as the paper rests on the lower collar of the drum, the upward expansion of the paper may render the record of the maximum 2 or 3 degrees too low. In any case, the maxima and minima should be compared with those of the registering thermometers, and the corrections, plus or minus, noted on the record sheet.
If the thermograph record does not coincide with the thermometer readings an adjustment screw will bring the pen to the proper position. It is a good plan to adjust the pen so that the minimum coincides with that of the minimum thermometer. The time of the minimum may always be determined from the thermograph sheet, and this is one of its important uses. As a rule, the minimum temperature occurs a short interval before sunrise. During the progress of a cold wave there may be a steady fall of temperature covering a period of two days. Frequently the fall of temperature continues from 12:01 a.m. to 11:59 p.m.[4] The “lowest this morning” is therefore not the minimum of the day; and though this fact may escape the notice of the observer, it will not escape the record of the thermograph.
The recording pen of the thermograph may lag anywhere from five minutes to forty-five minutes behind the actual temperature. In very damp weather the lag is usually the greatest, and, in fixing the time at which a given temperature occurred, this fact must be taken into the calculation. An observer who studies the vagaries of his instruments—and they are many—will learn how to master them.
High-air Thermographs.—High-air temperature observations are usually obtained by thermographs secured to kites or balloons. In manned balloons a very sensitive thermograph is contained in a tube through which a current of air is forced. This instrument, the Assmann aspirator, is far more convenient than an ordinary thermograph. Experience has shown that unless the air is in rapid motion, registration is too slow to be trustworthy. A mechanical fan moves the air through the tube at the rate of about 12 feet per second.
In another form of instrument decreasing pressure moves a plate in one direction while the stylus of a bi-metallic thermometer records with a motion at right angles thereto. A clock is not required in this type of instrument; it is therefore lighter and more convenient.
The Black-bulb Thermometer.—This instrument, now little used, consists of a maximum standard thermometer, the bulb of which is covered with a coat of lampblack and encased in a vacuum tube. Originally it was designed for the measurement of solar radiation. A thermometer of this sort, exposed to direct sunshine, registers a temperature many degrees higher than does an ordinary thermometer, but the degree varies according to the thickness and the quality of the lampblack. It is therefore a very imperfect instrument for the purpose designed.
The black-bulb thermometer roughly measures the temperature which popular tradition terms “sensible” heat; exposed to direct sunshine, the temperature registered is from a few degrees to 60 or more degrees higher than the temperature registered by the ordinary thermometer. With a high humidity, or in smoky, dusty or foggy air, the black-bulb thermometer registers much lower than in clean air.
- ↑ The standard conditions are somewhat complex. The real test in thermometry is a comparison of a thermometer in the process of manufacture with an accurately made standard. A standard thermometer is a part of the equipment of the manufacturer.
- ↑ This rule applies in all Weather Bureau computations.
- ↑ In the latitude of middle England Sir Napier Shaw notes that thermometers on the grass register lower by 20 degrees than those in the shelter, a difference of 8 degrees being very common.
- ↑ These figures are generally employed in weather bureau and in meteorological time to avoid the confusion that results from the use of the term “midnight,”