Eighteenth-Century Development of Temperature Scales
Eighteenth-Century Development of Temperature Scales
Development during the early eighteenth century of practical thermometers with stable temperature scales by Daniel Fahrenheit (1686-1736), Anders Celsius (1701-1744), and others made possible reproducible, intercomparable temperature measurements. This had immediate practical applications in many branches of science as well as catalyzing work on the nature of heat and temperature and leading to the formulation of the concepts of specific and latent heat. Further efforts to refine these scales eventually led to the development of an absolute temperature scale.
Devices that merely indicate changes of temperature—known as thermoscopes—have existed since antiquity. Such devices are not thermometers since they provide no means of identifying specific temperatures or quantifying temperature changes. This requires a temperature scale. Galileo Galilei (1564-1642) is generally credited with constructing the first thermometer (1592), although Cornelius Drebbel (1572-1633) may have preceded him. Regardless, Santorio Santorre (1531-1636) was the first to publish a description of a thermometer (1612), and by the middle of the seventeenth century it was a well-known and widely used device. However, early temperature scales were completely arbitrary. They were often attached to thermometer stems capriciously or calibrated against a single temperature, such as the hottest day of the summer, with the degree being some arbitrarily selected distance. Consequently, even measurements made with instruments produced by the same maker were not guaranteed to be comparable.
Two strategies were pursued to overcome this state of affairs: calibrating instruments against some standard reference instrument and constructing fixed-point-scales from first principles. The former approach was undertaken by the Academia del Cimento around 1654 and the Royal Society of London about 1665. Distribution of their instruments made possible the first intelligible meteorological temperature records, but intercomparison with instruments not calibrated to their standards was difficult if not impossible. The latter approach was pursued with scales having either one or two fixed points. Robert Hooke (1635-1702) constructed a scale using one fixed-point and having degrees correspond to equal increments of the thermometricsubstance volume. Scales with two fixed points had their degree markings defined as some fraction of the distance between the fixed points. Carlo Rinaldini (1615-1698) first suggested the melting point of ice and boiling point of water as appropriate fixed points. He divided his scale into twelve degrees (1694).
Though much progress was made, confusion was still the order of the day. Because volume is that property of substances which is most readily perceived to change with temperature, it was the preferred means of measuring temperature in early thermometers. However, instrument makers employed different thermometric substances. Failure to understand the difference between heat and temperature made it difficult to select appropriate fixed points and standardize calibration procedures. Furthermore, lack of appreciation for how scales depend on the properties of thermometric fluids made intercomparison between different types of thermometers, even when well calibrated, difficult. Thus, by the beginning of the eighteenth century we find air, alcohol, mercury, and even linseed oil employed as thermometric substances with scales of anywhere from 50 to 300 degrees and calibrated with respect to anything from the freezing point of water to the melting point of butter.
Daniel Fahrenheit was the first to manufacture uniform thermometers with stable, reproducible scales. His scale evolved from Ole Röer's (1644-1710), who took the freezing and boiling points of water as fixed. Fahrenheit mistakenly took Röer's upper fixed point to be normal body temperature. Having done so he fixed the freezing point at 32° and upper point at 96°. The latter point was calibrated by placing a thermometer in a healthy person's mouth. Water's boiling point on this scale was approximately 212°. Though Röer clearly never took water's boiling point as fixed, the Royal Society of London officially established it as such in 1777, fixing the value at 212°. (Normal body temperature on this scale is 98.6°.) The revised scale was quickly adopted in England and Holland and became the standard throughout the English-speaking world.
Anders Celsius also made reliable thermometers with two fixed points. His lower fixed point was determined by immersing the instrument in ice, the upper point by placing it in boiling water. He set the upper point at 0° and the lower at 100°, thus producing the first centigrade thermometer (1741). Carolus Linnaeus (1707-1778) inverted the scale shortly after Celsius' death and is sometimes given credit for the present centigrade scale. But a centigrade thermometer with the freezing point at 0° had been built sometime before 1743 by Jean Pierre Christian (1683-1755).
The existence of stable, reproducible scales such as those of Fahrenheit and Celsius made it possible to compare readings from thermometers with different scales. George Martine (1702-1741) took advantage of this to prepare conversion tables, which he included in his Essays Medical and Philosophical (1740). Subsequently, by mid-century thermometers were widely used in chemistry, medicine, meteorology, oceanography, and many other disciplines as well as finding industrial applications in London breweries and elsewhere. The development of reliable temperature scales also allowed heat to be appropriately distinguished from temperature and led to the formulation of the concepts of specific and latent heat.
Eighteenth-century opinions about heat were largely derived from the medical writings of Galen (c. 130-200), who conceived of bodies as possessing differing amounts of heat, which was thought to be a material substance much like a fluid. Thermometers were thought to measure the quantity of heat in bodies, with temperature being proportional to the amount of fluid present. The amount of fluid in turn was believed to be proportional to volume. Joseph Black (1728-1799) challenged these views armed with the experimental results of Fahrenheit and others.
Fahrenheit had shown that a mixture of three parts water and two parts mercury reached an equilibrium temperature midway between the two initial temperatures—a result only possible on the fluid view if initial volumes were equal. Additionally, George Martine had experimentally demonstrated that when equal volumes of water and mercury are heated the mercury temperature rises faster (1739)—a result not possible if volume alone determines the amount of heat an object can absorb. Black concluded that the quantity of heat in an object was not proportional to its volume or mass (1760). He argued that different substances have different affinities or capacities for heat and that thermometers actually measure the density of heat in objects. Thus, mercury's temperature rises faster than water's because its affinity for heat is greater, meaning mercury accumulates heat faster than water. Black established that the amount of heat in any body is proportional to its temperature, mass, and heat capacity, defining "heat capacity" as amount of heat required to raise the temperature of a unit mass of a given substance one degree. Johann Carl Wilcke (1732-1796) independently arrived at the same conclusion in 1781, though referring to a substance's heat affinity as "specific heat."
The ability to clearly conceive of heat as a measurable physical quantity distinct from, though related to, temperature also made it possible to formulate the concept of latent heat, which ultimately helped undermine the fluid theory of heat. Employing Fahrenheit's mercury thermometers, Black showed that when substances go through phase changes (e.g. from solid to liquid, as when ice turns into water) they absorb heat without changing temperature. According to the fluid theory this heat in some way combines with the substance so as to remain concealed from the thermometer. Because this hidden heat could still potentially affect a thermometer Black referred to it as "latent heat."
If heat were a material substance as maintained by fluid theorists, then it would be expected to have weight. It follows that the weight of substances acquiring latent heat should increase, e.g., a fixed volume of water should weigh more as ice than liquid. George Fordyce (1736-1802) claimed to have measured such increases. Benjamin Thompson (1753-1814), better known as Count Rumford, repeated Fordyce's experiments with greater care and detected no such difference (1787). He concluded that heat must be a mode of motion, not a substance. Rumford provided a more convincing proof when he observed that as long as the mechanical action required to bore a cannon continued, heat was generated (1798). This suggested that heat was a form of motion. It also undermined the fluid theory, which assumed conservation of heat.
As thermometers saw increasing application and were required to measure temperatures over a wider range, it became evident that the values assigned to temperatures by the scales of Fahrenheit and others had no absolute physical significance. In 1739 René-Antoine de Réaumur (1683-1757) realized that the scales of alcohol and mercury thermometers do not coincide because each substance has a different thermal expansion. It was also recognized that the working range for any thermometric substance was restricted by thermal resistance. For example, mercury-in-glass thermometers have a lower range defined by the freezing point of mercury and upper range circumscribed by the glass resistance. To cover the range of temperatures that scientists wished to measure, instruments with different working substances were required. This, however, introduced insuperable practical problems. The ideal solution would be a scale valid over the entire temperature range and independent of any particular thermometric substance.
William Thomson (1824-1907), also called Lord Kelvin, provided such a scale with his thermodynamic definition of temperature based on the efficiency of a reversible Carnot cycle (1848). Known as the Kelvin scale, any thermal state can be assigned a thermodynamic temperature using it. It thus provides a standard and meanings by which to unify thermometry. The problem is that the Kelvin scale is impossible to realize in practice. Fortunately, the Kelvin scale is equivalent to the ideal-gas scale, which, though impossible to realize in practice as well, is nicely approximated by hydrogen thermometers except at very low temperatures.
STEPHEN D. NORTON
Middleton, W. E. Knowles. The History of the Thermometer and Its Use in Meteorology. Baltimore: The Johns Hopkins Press, 1966.
Quinn, T. J. Temperature. London: Academic Press, 1990.
Roller, Duane. "The Early Development of the Concepts of Temperature and Heat." Case 3 in James B. Conant, ed., Harvard Case Histories in Experimental Science. Cambridge: Harvard University Press, 1950: 9-106.
Van der Star, Pieter. "Life and Works." In P. van der Star, ed., Fahrenheit's Letters to Leibniz and Boerhaave. Amsterdam: CIP-Gegevens, 1983.
Landsberg, H. E. "A Note on the History of Thermometer Scales." Weather 19 (1964): 2-7.
Middleton, W. E. Knowles. "More On 'Celsius or Linnaeus.'" Bulletin of the American Meteorological Society 59 (1978): 190.
Taylor, F. Sherwood. "The Origin of the Thermometer." Annals of Science 5 (1942): 129-56.