Temperature Measurement

III.

Temperature Measurement

III.

Submitted 1910 | SovietRxiv: ru-191001.07439 | Translated from Russian

Pre-reform orthography. The original is printed in the Russian spelling used before the 1918 reform (ѣ, і, ъ).

Full Text

Proceedings

of the Tomsk Technological Institute

of Emperor Nicholas II

Vol. 19. 1910. No. 3.

III.

V. L. Maltev.

TEMPERATURE MEASUREMENT.

81–227.

CONTENTS.

	Page
Preface	1

PART I.

Theory and construction of various instruments.

	Page
1. Temperature scales	3
Mercury scale	3
Hydrogen scale	4
Thermodynamic (absolute) scale	5

Gas thermometers.

	Page
2. General remarks. Various methods of measurement	7
3. Sources of errors	10
Variability of the coefficient of expansion of gases $\alpha$	10
Gas permeability of the vessel walls	11
Change in the volume of the vessel	11
Influence of the presence of water	12
Influence of harmful space	12
4. Calculation of corrections	13
5. Designs of gas instruments	21
Chappuis normal thermometer	21
Barus pyrometer	23
Joly air thermometer	24
Richard recording pyrometer	25
Siegert-Dürr pyrometer	25
Wiborg mercury pyrometer	27
Wiborg aneroid pyrometer	29
Jüling-Steinbart pyrometer	31
Pyrometer with a micromanometer	33

Instruments with liquids.

	Page
6. Subdivision	35
7. Glass thermometers	35
Alcohol thermometer	35
Mercury thermometer	35
Choice of glass. Thermal aftereffect	36
Ratio of the capacity of the vessel to the dimensions of the tube	37

Corrections of mercury-thermometer readings . . . . . . . . . . . . . . . . . . 38
    Influence of thermal aftereffect . . . . . . . . . . . . . . . . . . . 38
    Nonuniform expansion of mercury . . . . . . . . . . . . . . . . 39
    Influence of pressures . . . . . . . . . . . . . . . . . . . . . . . 39
    Correction for the projecting column . . . . . . . . . . . . . . . 40
Designs of mercury thermometers . . . . . . . . . . . . . . . . . . . . . 42
    Conditions for principal normal thermometers . . . . . . . . . . . 42
    Thermometers with a casing . . . . . . . . . . . . . . . . . . . . 43
    Beckmann thermometer . . . . . . . . . . . . . . . . . . . . . . 44
Use of thermometers . . . . . . . . . . . . . . . . . . . . . . . . . 44
    Measurement of temperature in tubes . . . . . . . . . . . . . . . 44
    Auxiliary body . . . . . . . . . . . . . . . . . . . . . . . . . 45
    Carpenter’s method for measuring high temperatures . . . . . . . 46
    Break in the mercury column and its reconnection . . . . . . . . 46
Maximum and minimum thermometers . . . . . . . . . . . . . . . . . 47
    Rutherford thermometer . . . . . . . . . . . . . . . . . . . . . 47
    Negretti and Zambra thermometer . . . . . . . . . . . . . . . . 48
    Six’s instrument . . . . . . . . . . . . . . . . . . . . . . . . 48
Liquid pyrometers . . . . . . . . . . . . . . . . . . . . . . . . . 49
    Thalpotasimeter . . . . . . . . . . . . . . . . . . . . . . . . 49
    Pabell’s pyrometer . . . . . . . . . . . . . . . . . . . . . . 50
    Steinle and Hartung pyrometer . . . . . . . . . . . . . . . . . 51
    Regener and Mach pyrometer . . . . . . . . . . . . . . . . . . 51
    Self-recording thalpotasimeter . . . . . . . . . . . . . . . . . 52

Expansion of solid bodies.

Metallic and graphite pyrometers . . . . . . . . . . . . . . . . . . . 52
Féry spiral pyrometer . . . . . . . . . . . . . . . . . . . . . 53
Graphite pyrometers . . . . . . . . . . . . . . . . . . . . . . 54

Instruments based on heat transfer.

Calorimetric methods and instruments . . . . . . . . . . . . . . . . 55
    Fischer calorimeter . . . . . . . . . . . . . . . . . . . . . . 55
    Fuchs instruments . . . . . . . . . . . . . . . . . . . . . . . 57
    Siemens pyrocalorimeter . . . . . . . . . . . . . . . . . . . . 59
    Sentignon method . . . . . . . . . . . . . . . . . . . . . . . 59
Determination of the water equivalent . . . . . . . . . . . . . . . . 60
Variability of heat capacities . . . . . . . . . . . . . . . . . . . . 60
Finding the heat capacity . . . . . . . . . . . . . . . . . . . 61
Pyrometers based on melting points . . . . . . . . . . . . . . . . . 63
    Metals and their alloys . . . . . . . . . . . . . . . . . . . . 63
    Seger cones . . . . . . . . . . . . . . . . . . . . . . . . . 64
    Le Chatelier’s experiments . . . . . . . . . . . . . . . . . . 65

Temperature Measurement

	Page
Brearley pyrometers “Sentinel” and the instrument for them	67
Ducomet instrument	68
18. Wiborgh thermophone	69

Instruments based on optical phenomena

	Page
19. Light pyrometers	70
Py’s table	71
Comparison of color scales	71
Mesuré and Nouel pyrometric tube	72
Féry spectral pyrometer	73
20. Photometric pyrometers	74
General remarks and Wien’s law	75
Le Chatelier and Borne instrument	75
Féry instrument	77
Holborn and Kurlbaum pyrometer	77
Wanner pyrometer	78
Wanner pyrometer for low temperatures	80

Electrical phenomena

Thermoelectric instruments

	Page
21. General remarks	81
Conditions for the choice of metals	82
Expressions for thermoelectric force	83
Choice of thermocouples	84
22. Methods of measurement	85
23. Corrections to the readings of thermoelements	87
Influence of electrical resistances	87
Influence of cold junctions	88
Influence of heating of the millivoltmeter	89
Observations of A. Baikov concerning the influence of the sheath	90
24. Handling thermoelements	91
Position of the instrument in the furnace	91
Protection of the iron tube from burning through	92
Connection of the ends—the measuring junction	92
Connection to the millivoltmeter	93
Insulation of thermocouple wires	93
25. Designs and details of thermoelements	93
Necessity of a sheath	94
Siemens and Halske instruments	94
Hartmann and Braun instruments	94
Heraeus instruments with insulation of quartz glass	95
Paul Braun instruments (nickel and carbon)	96

IV. V. L. Mal’kov.

	Page
Bristol’s compound thermoelement	97
Increasing the sensitivity of the thermoelement	98
Head with water cooling	98
Bristol’s compensator	98
26. Galvanometers for thermoelements	99
Hartmann and Braun instrument	99
Siemens and Halske instruments	99
Wall-mounted galvanometer of the Cambridge Scientific Instrument Co.	100
Siemens and Halske set for the compensation method	101
Siemens and Halske self-recording galvanometer	101
Hartmann and Braun self-recording galvanometer	102
27. Féry thermoelectric pyrometer	103
Idea, construction, and use of the instrument	103
Accuracy of the instrument and its merits	105

Measurement of electrical resistance.

	Page
28. General remarks	106
Siemens’ expression for resistance	106
Advantages of platinum	106
Expressions for the resistance of platinum	107
Platinum scale	108
29. Methods of measurement	109
Method 1—Siemens’ method with branching	109
Method 2—Callendar’s method with two current sources	110
Method 3—with Wheatstone’s bridge	111
Fisher-Crompton circuit with a differential galvanometer	112
Method 4—the null method	112
Comparison of the indicated methods	113
30. Sources of error and their elimination	114
Checking the voltage when working by method 3	114
Checking the voltage when working by method 2	114
Influence of harmful resistance and the Callendar loop	115
Influence of temperature fluctuations in the room	116
Influence of thermoelectric phenomena and insulation	117
31. Description of various designs	118
Instruments for measuring temperature by method 2	118
Instruments for measuring temperature by method 3	119
Instruments made of quartz glass	121
Instruments for measuring temperature by method 4	121
Use of high-voltage electric current	124
Self-recording instruments	125

Preparation of Reports at a Distance

		Page
32.	Instruments for measuring temperature at a distance	126
	Mercury thermometers with a switchboard	126
	Mercury thermometer with a large number of contacts	127
	Transmission by the Mönnich method	128

General Practical Instructions

		Page
33.	Accuracy of various instruments	128
	Conditions for accuracy of measurements	129
	Rapidity of following temperature fluctuations	130
	Comparison of the accuracy of various instruments	131
34.	Selection of the appropriate type of instrument	132
	Various examples of instrument selection	133

PART II

Verification of Instruments for Measuring Temperatures

		Page
35.	General instructions	136
	Various methods of verification	136
	General rules for verification	137
36.	Verification by means of comparison	137
	Apparatus with liquid air for temperatures from −190 to 0° C	138
	Apparatus with liquid carbon dioxide for temperatures from −67 to 0° C	141
	Apparatus with refrigerating mixtures	142
	Table of refrigerating mixtures	143
	Water bath of the German Physico-Technical Institute	145
	Oil bath for temperatures from +50° to +200° C	147
	Saltpeter bath for temperatures from +250° to +550° C	148
	Oil bath of the Physico-Technical Institute	149
	Apparatus of the Physico-Technical Institute from +100° to +550° C	150
	Thermostat operating with liquid vapors	152
	Thermostat for liquids with a high boiling point	154
37.	Verification by melting points	155
	Verification by the melting point of ice	155
	Verification by the freezing point of H₂O,HCl	156
38.	Verification by temperatures of chemical transformation	158
	Transformation temperature of NaBr,H₂O	158
	Transformation temperature of Na₂SO₄,H₂O	160
39.	Verification by boiling points	160
	Conditions that the corresponding bodies must satisfy	160
	Table of boiling points of various bodies	161

Table of boiling temperatures of water at various pressures . . . 162
Thermostat for water . . . 163
Thermostat for water of the Physico-Technical Institute . . . 164
Apparatus for boiling points above $200^\circ$ C. . . . 166
Apparatus for the boiling point of sulfur . . . 166

Particular features in testing glass mercury thermometers . . . 167
    Calibration . . . 168
    Determination of thermal aftereffect . . . 170
    Testing $0^\circ$ C. . . . 171
    Watson’s apparatus . . . 172
    Callendar’s method for increasing the accuracy of readings . . . 174
Particular features in testing thermoelements . . . 175
    Electric furnace for comparing thermoelements . . . 175
    Barus’s apparatus for boiling points . . . 177
    Apparatus for testing at the boiling points of Zn and Cd . . . 178
    Le Chatelier’s apparatus for the same purpose . . . 181
    Table of melting points of metals . . . 182
    Testing at melting points in a crucible . . . 183
    Testing at the melting points of noble metals . . . 184

Conclusion.

Final conclusions . . . 186

Appendix 1: list of firms and average prices . . . 188
Appendix 2: list of literary sources . . . 193
List of noted misprints . . . VII.

Noted Misprints

Page	Line	Printed	Should read
6	14 from top	by gas	by gases
6	9 from bottom	is denoted	is denoted
13—20	everywhere, where: $dT, dH, d\omega, dV, dG, d\theta,$	one must	$\Delta T, \Delta H, \Delta \omega, \Delta V, \Delta G, \Delta \theta$
22	9 from bottom	by means of	by means of
25	12 from bottom	by it	by it.
29	1 from top	volume	volume
29	9 from top	identical	identical
29	6 from bottom	empiri-	empiri-
29	4 from bottom	observations	observations
30	1 from top	see.	see.
30	14 from bottom	retur-	retur-
30	13 from bottom	$V,$	$V'$
30	7 and 6 from bottom	dial plate	dial plate
31	1 from top	dial plate	dial plate
32	2 from top	in velocity	in velocities
41	6 from bottom	insofar	insofar as
42	26 from bottom	German Imperial Institution	German Imperial Physico-Technical Institute
43	13 from bottom	closed	closed
43	12 from bottom	not of	not of
49	3 from top	the lowest	the lowest
49	12 from bottom	potassium	heating
64	3 from bottom	special	special
74	8 from bottom	$\left(\dfrac{1}{T_1}\quad \dfrac{1}{T}\right)$	$\left(\dfrac{1}{T_1}-\dfrac{1}{T}\right)$
82	10 from top	years	years
82	11 from top	to be made	to be made
82	3 from bottom	hardened	hardened
83	12 from bottom	for platinum	for a thermocouple of platinum and platinorhodium
84	11 from bottom	consecutive	consecutive
84	5 from bottom	Dewar	Dewar
84	1 from bottom	blockamil	blockamil
85	8 from top	parasitic	parasitic
85	13 from top	possessing	possessing
91	1 from top	the lowest	the lowest
91	1 from bottom	to dispose	to dispose
92	2 from top	insufficient	insufficient
93	15 from top	conductors	conductors
93	18 from top	soldering	soldering
93	19 from top	negligible	negligible
93	22 from top	sufficient	sufficient
95	4 from bottom	thickness	thickness
98	7 from top	permitting	permitting

VIII. V. L. Mal'ev.

Page	Line	Printed	Should be
98	15 from bottom	garmoelementu	termoelementa
98	4 from bottom	soedinennye	soedinennye
99	24 from bottom	zarkal'tsem	zerkal'tsem
99	20 from bottom	dorozhe	dorozhe,
100	23 from bottom	raspolozhennym	raspolozhennym
100	5 from bottom	gal'vonomentry	gal'vanometry
103	4 from bottom	zazhimami	zazhimami
104	1 from top	$i$	$d$
105	20 from top	millivot	millivol't
107	5 from bottom	nedostatkov	nedostatkov
110	11 from bottom	postoyannuyu	postoyannoyu
113	8 from top	stoyanykh	stoyannykh
113	1 from bottom	dabavleniya	dobavleniya
118	19 from bottom	metalla	metalla,
125	3 from top	otsutstviya	otsutstviya
127	10 from bottom	prodolzhitel	prodolzhitel'nom
128	7 from bottom	zachitel'noi	znachitel'noi
129	3 from top	prezhde	prezhde,
130	12 from top	proportsianal'no	proportsional'no
131	16 from top	$5,2$	$2,0$--$5,0$
132	11 from bottom	osobeno	osobenno
133	5 from bottom	temoelementov	termoelementov
138	9 from bottom	postepeno	postepenno
139	9 from bottom	vannoi	vannoi
144	12 from bottom	otni maet	otnimaet
146	6 from bottom	vodyannoi	vodyanoi
147	13 from top	tsilindriseskoi	tsilindricheskoi
148	8 from bottom	otverstie	otverstiya
150	5 from bottom	termomotrov	termometrov
153	11 from bottom	kodensiruetsya	kondensiruetsya
156	16 from bottom	vysovyvayushchii	vysovyvayushchiisya
156	3 from bottom	$2 \frac{0}{5}$	$2{,}5^\circ$
156	3 from bottom	$0{,}663,$	$0{,}663;$
157	3 from top	tritrovaniem	titrovaniem
158	11 from bottom	elementev	elementov
160	16 from top	bramistyi	bromistyi
160	18 from top	istinoi	istinnoi
160	14 from bottom	bramistago	bromistago
165	8 from top	termemetov	termometrov
165	15 from top	kondinsiruyutsya	kondensiruyutsya
167	14 from bottom	$d$	$e$
167	15 from bottom	$b\ a$	$d\ e$
171	2 from bottom	pol'zovat'sya	pol'zovat'sya
175	2 from top	primierov	primerno
176	6 from top	prikleivaemaya	prikleivaemaya
179	20 from top	stavit's	stavitsya
183	5 from top	okislyaetsya	okislyaetsya
186	7 from top	odostich	dostich,
189	11 from top	ukazaniem	ukazaniem
191	1 from bottom	tormoelektr.	termoelektr.
193	3 from bottom	Sewerbe	Gewerbe

... due to the fact that the weakening of the light intensity as a result of refraction, absorption, and reflection affects only the ray from the incandescent lamp, whereas the ray from the source under investigation passes through the instrument almost unweakened.

The rays from the source under investigation enter through the aperture $a$ and, after passing through the lens $d$, the upper half of the circular diaphragm $f$, the lens $b$, and the aperture $e$, fall into the observer’s eye. In the lower part of the T-shaped tube, which at the same time serves as the handle for the whole instrument, there is an osmium lamp $g$, powered by an accumulator battery; the light from it passes through the ground glass $h$, is polarized by the prism $k$, then passes through a second Nicol prism $l$, which can be rotated relative to $k$ and thereby, in the manner described above, weaken the polarized ray to the required degree; the ray then passes through the lens $m$, is reflected by the rectangular prism $c$, and then, through the lower half of the diaphragm $f$, the lens $b$, and the aperture $e$, enters the observer’s eye.

In observing, one proceeds as with the preceding instrument: the prism $k$ is rotated until the field of view illuminated by the lamp and the glowing body under investigation appear to have the same light intensity; from the angle of rotation, read on the scale $n$, the required temperature is determined, as before.

The selection of the instrument is carried out in exactly the same way as for the preceding one. The fall in voltage of the battery, which should give about 1.6 volts, is compensated by a gradual change of a special resistance.

Concluding this chapter on optical pyrometers, it may be said that in all cases in which they can be used—for example, in determining the temperature of furnaces, fireboxes, and incandescent metal, and in technical measurements—they, especially pyrometer-photometers, may confidently be given preference over all other instruments.

Electrical Phenomena

Thermoelectric Instruments

21. General remarks.—The thermoelectric method of measuring temperatures is based on the phenomenon discovered by Seebeck in 1821, consisting in the fact that if two rods of different metals are joined to one another at both ends in such a way that along the remaining...

if along their length they do not touch one another, then upon heating or cooling one of the junctions, an electromotive force appears in the closed circuit of the rods.

The magnitude of this force depends on the temperature difference of the two junctions and on the nature of the metals taken. By measuring this force and knowing the temperature of one of the junctions, it is easy, by one means or another, to determine the temperature of the other junction.

Many outstanding scientists soon began to concern themselves with methods of determining temperature on the basis of thermoelectric phenomena. However, only after the work of Le Chatelier, from the 1890s of the past century onward, did thermoelectric thermometers begin to be manufactured by factory methods, and at the present time they have received wide and quite deserved distribution.

In choosing metals for a thermocouple, two main circumstances must be borne in mind: on the one hand, the magnitude of the electromotive force and the law of its variation, and, on the other, the physical and chemical properties of the given materials; among other things, an important condition for a good thermoelement is the absence of secondary currents.

Table 12.

material	millivolt
iron	+2.10
silver	0.90
alloy of copper with 10% aluminum	0.70
gold	0.60
alloy of platinum with 10% rhodium	0.50
alloy of platinum with 10% iridium	0.50
alloy of steel with 5% nickel	0
alloy of steel with 13% manganese	−0.30
alloy of copper with 20% nickel	−0.60
nickel	−2.20
alloy of steel with 35% nickel	−2.70
alloy of steel with 75% nickel	−3.70

The magnitude of the electromotive force may be judged from the accompanying Table 12, in which are indicated, in millivolts, the electromotive forces of thermoelements composed of pure platinum on one side and the corresponding metal on the other, at junction temperatures of 0° and 100° C. The electromotive force $e$ of a thermoelement composed of any two materials of Table 12,

Measurement of Temperature

is obtained in the form of the difference of their electromotive forces with platinum. For example, for a silver—nickel element, $e=0.90-(-2.20)=3.10$ mV.

In view of the fact that the electromotive force increases not proportionally to the temperature difference, but somewhat faster, in order to choose the corresponding element it is necessary to know the dependence of the electromotive force on the temperature difference of the junctions.

Avenarius and Tait gave the following analytical expression for the electromotive force $e$:

\[ e=(t_1-t_2)[a+b(t_1+t_2)]; \tag{55} \]

where $t_1$ and $t_2$ are the temperatures of the junctions, and $a$ and $b$ are thermal coefficients depending on the nature of the bodies.

For the two most commonly used thermocouples this dependence may be expressed in the following manner, taking the temperature of one of the junctions as $t_2=0^\circ\text{C}$:

for a thermocouple of platinum and platinum-rhodium,

\[ e=-0.31+0.008048t+0.00000172t^2; \tag{56} \]

equation (56) is valid for temperatures up to $1600^\circ\text{C}$;

for a thermocouple of constantan and silver, we have approximately

\[ e=0.035t+0.000025t^2. \tag{57} \]

Still better results for platinum are given by the logarithmic expression found by C. Holmann,

\[ \lg e=1.2196\lg t+0.302. \tag{58} \]

However, the numerical coefficients in equations (55)—(58) are not entirely constant, since it is very difficult to have chemically pure metals, and still more difficult to obtain completely identical and homogeneous alloys. Therefore, for any reasonably accurate measurements, $f(e,t)$ must be found experimentally for each instrument, as will be discussed below, and for practical use the results obtained should be presented in the form of a diagram.

In Fig. 69, p. 84, examples of such diagrams are presented for several thermocouples; curve 1 is a thermoelement of copper and constantan;

curve 2 and $2'$—two pairs of silver and constantan; as can be seen, the curves generally coincide and diverge slightly only after $500^\circ$ C;

Chart 69.

curve 3—the couple of platinum and platinoiridium ($\mathrm{Pt}+20\%\ \mathrm{Ir}$) was obtained by Barus; curves 4 and $4'$—the couple of platinum and platinorhodium ($\mathrm{Pt}+10\%\ \mathrm{Rh}$); $4'$ was obtained by Le Chatelier, and 4 by Holborn and Wien at the Imperial Physico-Technical Institute; there, too, curves 1, 2, and $2'$ were obtained.

As regards the chemical and physical properties of the metals chosen, they must meet the following requirements: the metal must be as pure as possible, must, as far as possible, not oxidize, must not change its structure after successive heatings, must not be brittle, and, finally, must be ductile, so that wire of uniform cross-section can be made from it.

In view of all the foregoing, for measuring low temperatures from $-190^\circ$ to $+300^\circ$ C, one should use a thermoelement of copper and constantan; for the lowest temperatures, namely down to $-260^\circ$ C, J. Dewar advises using a thermocouple of gold and Neusilber, soldered with pure silver. For temperatures from $0^\circ$ to $+600^\circ$ C, a thermocouple of silver and constantan is used. A. Geyl advises replacing the silver with an alloy of 75% silver and 25% manganese; the alloy, obtained in a blowpipe flame, is very homogeneous and gives

flexible, strong wire; moreover, the electromotive force of such a thermocouple is considerably greater.

For temperatures up to $900^\circ$, platinum and an alloy of platinum with nickel were at one time used; the electromotive force of such a thermocouple is considerably greater than that of other combinations, and chemically the element is a very good one, but it has now been abandoned, since all alloys with nickel have the unpleasant property of producing very strong parasitic currents, which disturb the readings.

For temperatures from $+300^\circ$ to $+1600^\circ$, thermocouples made of platinum and platinorhodium or platinoiridium are the most suitable. As is seen from Fig. 69, for temperatures below $300^\circ$ these thermocouples are insufficiently sensitive. It is not possible to use, for high temperatures, couples made of constantan and silver, which have an electromotive force about 6 times greater, in view of the chemical properties of these bodies.

For the very highest temperatures, up to $2000^\circ$ C., the firm Heraeus manufactures thermoelements of iridium and an alloy of iridium with 10% ruthenium. Since iridium cannot be drawn because of its brittleness, it is rolled into the form of a square wire $0.8$ mm thick.

The electromotive force of this thermocouple is about $1/3$ that of a thermocouple of platinum and platinorhodium. Owing to the evaporation of ruthenium during melting, the thermoelement cannot be made as uniform as one of platinum and platinorhodium, and it must be calibrated for each thermocouple by comparison up to $1600^\circ$ with a standard thermocouple of platinum and platinorhodium, and thereafter by the melting point of platinum $+1780^\circ$; likewise, when working with this thermocouple, frequent checking of the latter point is useful.

The absence of a suitable material for the mounting of this thermoelement hinders its spread for technical purposes; for scientific purposes, however, it can be very useful, especially in view of the simplicity of handling, a general advantage of all thermoelements.

22. Methods of measurement.—The measurement of the electromotive force arising as a result of a temperature difference may be carried out in two ways: by direct measurement of the voltage of the current with the aid of a sensitive millivoltmeter, and by the so-called null method.

The null method, or the method of compensating connection, proposed by Kohlrausch and developed by Lindeck, is used for especially precise measurements, since it is more sensitive.

It consists in the fact that the unknown electromotive force of the thermoelement is balanced by another, known force.

Diagram 70.

Fig. 70.

The arrangement of this method is shown in Fig. 70: $t$ is the thermoelement, $a$ a small electric accumulator, $w_1$ a constant resistance, $w_2$ an adjustable resistance, $m$ a milliammeter, and $g$ a galvanometer. The current from the accumulator $a$, varied by means of the resistance $w_2$, produces in the constant resistance $w_1 = 0.1$ ohm a potential difference which must balance the unknown electromotive force of the thermoelement $t$; for this purpose the contact of the resistance $w_2$ is moved until the current disappears in the right-hand part of the circuit, as will be indicated by the return to 0 of the pointer of the galvanometer $g$; the reading on the milliammeter $m$ will make it possible to find the temperature $t$.

Diagram 71.

Fig. 71.

If in the circuit of Fig. 70 the resistance $w_2$ is made constant and, conversely, the resistance $w_1$ is varied, then the circuit of Fig. 71 is obtained. The current from the accumulator $a$, which gives a strictly definite potential difference at each element of the resistance $w_1$, passes through the main circuit $w_1 w_2 w_3 a$ and branches off into the section $t c g$; the greater the electromotive force of the thermocouple $t$, the farther the contact on the resistance $w_1$ must be shifted, i.e., it must be increased, so that the two electromotive forces balance each other, as indicated by the position of the pointer of the galvanometer $g$; the position of the contact $w_1$ gives directly the potential difference in millivolts, from which the temperature $t$ is found in the usual way.

The distinctive feature of this circuit is that the voltage drop in the main circuit across the section $w_1 w_2 w_3$ must always be exactly the same; since the electromotive force of the accumulator $a$ changes, the voltage drop has to be regulated, for which purpose part of the resistance $w_3$ is made variable; the regulation is carried out by comparison with the normal element $n$, while the branch containing it is closed by switching to it, by means of the switch $c$, the galvanometer $g$, whose pointer must again come to zero.

The method according to the scheme in Fig. 71 is more reliable than the method according to the scheme in Fig. 70, since in it the accuracy of the readings depends only on the constancy of the electromotive force of the normal cell $n$ and the resistance $w_1$, which in fact is the case; whereas in the method according to Fig. 70 the accuracy depends on the accuracy of the readings of such a delicate instrument as the milliammeter $m$.

With the null method it is possible to use thermocouples made of thinner wire, down to $0.2$ mm, which increases the sensitivity of the instrument and appreciably reduces the cost of platinum thermocouples.

The disadvantages of the null method are the complication of taking readings and the impossibility of measuring rapidly fluctuating temperatures.

23. Corrections to the readings of thermoelements.—When working with a thermoelement, the following two circumstances must be borne in mind: first, the certificate of the German Physico-Technical Institute, supplied with every good thermoelement, indicates the electromotive force $e$ of the element itself. A millivoltmeter connected to the thermoelement measures the voltage at the terminals, which is less than $e$ by the magnitude of the voltage drop in the element itself. This drop is equal to the product $ir$, where $i$ is the current strength and $r$ is the internal resistance of the element. Since $i$ depends on the external resistance, i.e. on the galvanometer, the voltage at the terminals of the element will also be different, depending on the galvanometer used. In addition, the galvanometer is usually placed at some distance from the thermoelement and is connected to it by copper conductors, whose resistance must likewise be taken into account; in the case of a considerable distance and a small resistance of the galvanometer, their influence may be quite significant.

Thus, in order to obtain accurate results, the readings of the instrument must be multiplied by a certain coefficient calculated on the basis of Ohm’s law. Namely, if $e'$ is the reading on the galvanometer, $r_1$ the internal resistance of the thermoelement, $r_2$ that of the copper conductors to the galvanometer, and $r_3$ that of the galvanometer itself, then the actual electromotive force $e$, by which the desired temperature must be determined with the aid of the table of the Physico-Technical Institute, is

\[ e = e' + i(r_1 + r_2); \]

substituting, according to Ohm’s law, $i$ by

\[ \frac{e'}{r_1 + r_2 + r_3} \]

and performing the addition, we obtain

\[ e = e' \frac{r_3 + 2(r_1 + r_2)}{r_1 + r_2 + r_3}. \tag{59} \]

The magnitude of the correction may be judged from the following example: a thermoelement made of silver—constantan; $d=0.5$ mm; length $l=1000$ mm; $r_1=2.49$ ohms; length of the copper leads $L=12$ m; their diameter $d_0=1.0$ mm.

The specific resistance of copper $c=0.017$; using the expression

\[ r=c\frac{L}{f}, \tag{60} \]

where $f$ is the cross-sectional area of the conductor in $\mathrm{mm}^2$, we obtain for both conductors

\[ r_2=0.017\cdot \frac{12}{0.7854}\cdot 2=0.52\ \text{ohm}; \]

let the resistance of the millivoltmeter be $r_3=539$ ohms; then

\[ e=e'\frac{539+2(2.49+0.52)}{539+2.49+0.52}=1.006\,e'. \]

For the convenience of practical use, on the basis of such a calculation, according to the table of the Physico-Technical Institute, a temperature curve is compiled, like those shown in Fig. 69, but already for the entire setup: the thermocouple, the conductors taken, and the millivoltmeter. Such a diagram makes it possible, from the millivoltmeter reading, to obtain at once the graphically corrected temperature.

Instead of calculating the correction for the resistances of the thermoelement, the conductors, and the galvanometer, one may construct a corrected temperature curve by points (the melting and boiling points of known bodies), found for the given combination by direct experiments.

The second important condition for the accuracy of the readings is that the free ends, the so-called cold junctions of the thermoelement, should have the same temperature as during calibration. At the German Physico-Technical Institute the thermoelements are calibrated by immersing these ends in melting ice. In measurements for technical purposes, one often confines oneself to cooling the ends with running water, whose temperature is measured in ${}^{\circ}\mathrm{C}$ and added to the readings of the thermoelement. When measuring temperatures by means of thermoelements made of platinum and platinum-rhodium, only half the temperature of the cold ends need be added to the readings of the thermoelement; this is explained by the very gentle course of the temperature curve up to $+200^{\circ}\mathrm{C}$ for these elements.

For especially precise measurements with platinum–platinum-rhodium elements, the true temperature $t$ is found, according to Holborn’s experiments, by adding to the temperature $t'$ obtained from the galvanometer the correction $k t_c$, i.e.

\[ t = t' + k t_c, \tag{61} \]

where $t_c$ is the temperature of the cold junctions, and $k$ is a certain coefficient depending on the temperature $t'$. Holborn gives for it the values indicated in Table 13.

Table 13.

$t'\ ^\circ\mathrm{C}$	0	100	200	300	400	500	600	700	800	900	1000
$k$	1.00	0.89	0.76	0.65	0.59	0.56	0.54	0.52	0.51	0.50	0.49

The same addition must, of course, also be made in the case when the ends are not cooled at all; their temperature is then taken to be the temperature of the surrounding air. However, this is permissible only for rougher readings, since, owing to radiation and heat transfer from the heated connected ends, the temperature of the cold junctions may be considerably higher than the ambient temperature, and it is difficult to measure it with sufficient accuracy.

In millivoltmeters having a temperature scale, the correction for the temperature of the free ends can be avoided by setting, at the beginning of the readings, the pointer not at $0^\circ$, but at the temperature of the free ends. This method of introducing the correction is the simplest and most accurate, since it also automatically takes into account the circumstance that the electromotive force does not increase directly in proportion to the temperature, but somewhat more rapidly.

Finally, it must be borne in mind that the resistance of the coil of the millivoltmeter, and consequently its readings, depend on the temperature. Therefore the millivoltmeter must be installed so that it has normal room temperature; if it is impossible to eliminate its heating in any way, then its readings will be lower than the true ones. Since the resistance of the coil winding amounts to about one third of the total resistance, while the added resistance is made of manganin wire or another material whose resistance is almost independent of temperature, the readings must be multiplied by $(1 + 0.0013\,\Delta t)$, where $\Delta t$ is the excess of the instrument temperature over average room temperature $(15^\circ\mathrm{C})$.

For more precise observations, the numerical coefficient for $\Delta t$ must be calculated; exact data on the magnitudes of the total resistance and of the coil winding separately should be requested from the instrument manufacturer.

Finally, in general it is necessary to make sure that all junctions where different metals come into contact have the same temperature, equal to the room temperature; otherwise thermoelectric currents will also arise at these points, affecting the readings of the instrument.

The accuracy of thermoelement readings, when all the indicated corrections are observed, in practical use reaches $\pm 1^\circ$ in thermocouples for measuring temperatures from $-200^\circ$ to $+600^\circ$ C. In instruments for high temperatures, up to $1600^\circ$ C., the accuracy is $\pm 5^\circ$ in absolute measurements; when using one and the same instrument, the difference of temperatures can be determined with an accuracy also of about $\pm 1^\circ$. Recently the firm Heraeus has obtained materials of such purity, platinum and rhodium, that the absolute accuracy of its thermoelements, of course with proper handling, is not less than $\pm 1^\circ$. With very sensitive galvanometers and thermocouples, and with proper handling, fluctuations of temperature of $0.1^\circ$, even $0.01^\circ$ C., can be measured. Such instruments are used in physicochemical, and partly in meteorological, observations.

Table 14.

No.	pyrometer winding	reading, °C
1	quartz . . . .	1022
2	platinum over quartz . . .	1211
3	quartz . . .	1022
4	porcelain . . .	973
5	platinum over porcelain . . .	1230
6	porcelain . . .	969
7	quartz . . . .	1026
8	platinum over quartz . . .	1217
9	quartz . . .	1024

In considering the question of possible errors in the use of thermoelements, one cannot pass over in silence the very interesting experiments of A. A. Baikov, who drew attention to the influence of the sheath in which a thermoelement of platinum and platinum-rhodium is enclosed upon its readings, in the case when the thermoelement is subjected to the direct action of a flame. He compared the readings of a thermoelement enclosed alternately in a sheath of quartz glass, of porcelain, and in the same sheaths but inserted also into a thin-walled platinum sheath, when heated in the very same flame. As an example, we give one series of his experiments, Table 14.

It turns out that the pyrometer indicates the lowest flame temperature when it is enclosed in a porcelain sheath; a quartz sheath gives a temperature $50^\circ$ higher, a platinum one over quartz $240^\circ$ higher, and over porcelain even $260^\circ$ higher.

A. Baikov explains this phenomenon partly by the radiation of quartz and porcelain, which lowers the temperature of the thermoelement below the flame temperature, but chiefly by the contact property of platinum, which assumes a temperature higher than that of the flame. As a result of his experiments, Baikov arrives at the following conclusions:

Solid bodies—platinum, quartz, porcelain—heated in one and the same flame assume temperatures different from the flame temperature; this difference may reach several hundred degrees;
The temperature to which a solid body is heated in a flame depends on the contact properties of the body and on the ratio of its surface to its volume;
Determination of the flame temperature by means of solid bodies heated in it is incorrect, and the true flame temperature is indeterminate;
The difference in the luminosity of solid bodies when heated in one and the same flame is due to the fact that solid bodies are heated to different temperatures.

These conclusions are very important. If they are correct, then many experimental data, such as the melting and evaporation temperatures of metals and the temperatures of transition into various allotropic states, require correction; moreover, it cannot even be stated immediately how such a correction should be undertaken, since we do not yet have the means to measure temperatures without resorting at all to solid bodies.

The indicated conclusions of A. Baikov were checked in the scientific laboratory of W. Heraeus, with special precautions taken to eliminate the influence of radiation and thermal conductivity. The phenomena observed by A. Baikov were not noted. Thus, without deciding to give preference to one or the other of the aforementioned observers, this question must for the present be regarded as open, and its prompt final resolution is to be desired.

24. Handling thermoelements.—When measuring very high temperatures, the cement of the thermoelement begins to soften. In order that the instrument not be damaged thereby, it is best to arrange...

that is, it must be suspended strictly plumb, since supports placed under the end and even the middle of the instrument may prove insufficient; the outer casing may bend.

In addition, in order to avoid damage to the porcelain tubes when measuring high temperatures, the instrument must be heated and cooled, as far as possible, slowly and gradually.

To protect the iron tube, into which the thermoelement is usually inserted in technical measurements, from rapid burning-out when measuring high temperatures—especially if the thermoelement is pushed into the furnace for continuous use—it is useful to coat the iron tube with a thin layer of liquid clay.

The wire for thermoelements is taken 100 to 200 cm long; the length itself is of no importance, provided only that it is sufficiently great so that the heat imparted to one end of the wires of the thermocouple does not heat the other end.

The joining of the ends subjected to the temperature being measured is achieved either by fusing them together or, at low temperatures, by soldering with pure tin; in doing so one must not use any acids, but only rosin, since acids, by corroding the wire, alter the resistance of the thermoelement. It is most convenient to join the ends of the wires by twisting them together over a length of about 1 cm and fusing the tips in a gas burner. Twisting alone is not sufficiently reliable, since the wires may become covered with oxide, which greatly increases the resistance of the thermoelement. If after some time the twisted end, under the action of high temperatures, changes its structure, which affects the indications of the instrument and is discovered upon recalibration, the twisted end must be cut off and a new junction made. Such a small change in the length of the thermoelement wires does not perceptibly alter its resistance and, consequently, its indications.

The junction of the ends of a platinum and platinum-rhodium thermocouple requires special attention. With prolonged heating, the platinum-rhodium wire, chiefly its end, becomes brittle and breaks off easily. The junction may be restored by twisting both wires. Despite the impossibility in this thermocouple of forming the oxide mentioned above, such a junction is unreliable, since on heating the turns of the spiral unwind, the contact becomes ever worse, and the resistance increases. It is better to restore the junction by fusion. For this purpose it is convenient to use a small arc lamp, or even simply a pair of carbons from an arc lamp:

the left hand into the outer part of the voltaic arc, first the end of the platinum-rhodium wire; then, when a droplet of molten metal begins to form on it, the end of the platinum wire is introduced into the flame with the right hand; when a droplet has likewise formed on it, the wires are brought close to one another until the two droplets coalesce, and immediately, by a smooth motion of both hands, the wires are withdrawn from the voltaic arc; the droplet quickly solidifies, and the junction is ready. During this operation one must look at the voltaic arc through densely colored blue glass.

In the absence of an electric current, it is possible, though more difficult, to fuse the wires in the flame of an oxy-gas burner. The wires, twisted together over a length of 5 mm, are introduced into the flame, at first admitting very little gas, then gradually increasing the inflow of gas, and only when a droplet of molten metal forms is the junction removed from the flame.

The free ends of the thermoelement are connected with insulated conductors leading to the instrument by means of which the electromotive force is measured. This connection is made either with clamps or, better still, by soldering with tin. In view of the negligible voltage of the current, the insulation of the conductors may be of the simplest kind, such as that used for electric bells.

The resistance of these conductors should not be large; it is desirable to make it not more than 1 ohm. For this purpose, with conductor lengths up to 20 m, it is sufficient to use copper wire 1 mm thick; with lengths up to 100 m—2 mm.

The wires of the thermocouple itself must be carefully insulated from one another along their entire length. For temperatures up to 600° this can be done by slipping a thin glass tube or glass beads onto one of them; it is also useful to insert the whole element into a glass tube sealed at one end. At higher temperatures, tubes of porcelain or quartz glass must be used, and at the highest temperatures—of Marquardt composition. A simple and good method of insulating the wires of a thermocouple consists in braiding them in a figure-eight pattern with a thin, about 2 mm, asbestos cord, Fig. 72. True, with frequent heating the winding begins to crumble, but it is easy to renew.

Fig. 72.

25. Construction and details of thermoelements.—The thermoelements themselves are most often made of wires 0.4 to 0.6 mm in diameter in the case of such expensive material as platinum and its alloys, and up to 1.0 mm in the case of less valuable metals.

One of the advantages of thermoelements is that, in contrast to most other pyrometers, they can be used to measure the temperature of even very small bodies or spaces.

Unfortunately, thermoelements cannot be subjected to the action of incandescent gases, since carbon monoxide, which may always be present in them, alters the platinum-rhodium alloy, and with it the electromotive force of the element. In view of this, at the expense of the sensitivity of the element, it is necessary to surround it with a porcelain or Marquardt tube closed at the end. To prevent mechanical damage, the element is often enclosed in an iron casing, up to $1000^\circ$ C, or a nickel casing, up to $1200^\circ$ C; and so that the porcelain tube does not shake, it is wrapped with asbestos cord or the void is filled with dry quartz sand. An instrument equipped in this way, manufactured by Siemens & Halske, is shown in Figs. 73 and 74: $a$ and $b$ are the wires of the thermocouple, $c$ is the inner porcelain tube, and $d$ the outer one; $f$ is an iron tube with a head, the upper part of which is made of ebonite $e$. Recently Siemens & Halske have made the head somewhat differently, as in Fig. 75. For factory practice the same firm makes very convenient wooden pear-shaped opening heads, Figs. 76 and 77.

Fig. 75.

Fig. 76 and 77.

Fig. 73–74.

The firm Hartmann & Braun manufactures thermoelements of a somewhat different construction. In Figs. 78 and 79 an instrument for up to $1000^\circ$ C is shown: the wires, insulated from each other by two porcelain tubes, are placed in an oval manganese-steel tube with a welded closed end. For temperatures up to $1200^\circ$ C, the steel tube is replaced by a nickel one; for

for the same temperature range, the apparatus shown in Fig. 80 is intended: the thermocouple wires $a$ and $b$ are passed through openings in cylindrical rods $c,c$ made of refractory material; these cylinders are set on a nickel rod $d$ and are pressed against one another by the spring $e$. Such a fitting is cheaper than whole long porcelain tubes, and much stronger than they are.

Instead of porcelain tubes, the firm W. C. Heraeus insulates thermocouple wires by means of tubes drawn from quartz glass. The advantage of such insulation is that quartz glass, being quite indifferent to abrupt changes of temperature, lasts much longer and, moreover, if it breaks, can be fused together again. At the same time, quartz-glass tubes, being considerably thinner than porcelain ones, follow temperature fluctuations much more quickly, so that with such a pyrometer one can make up to 15 readings in 1 hour, whereas with other mountings 3–4 hours are required.

The only necessary precaution is that the incandescent quartz tube should nowhere touch iron, since otherwise it melts and breaks at that point. In these types the quartz tube is separated from the surrounding iron tube by a chamotte interlayer. To attain special sensitivity, the tip of the thermoelement in the sealed quartz tube is made to project by 60–80 mm from the iron fitting.

Fig. 78–79.

Fig. 80.

For measuring the temperature of molten metals, the said firm also manufactures special thermoelements in a steel fitting, the lower part of which, over a length of 30 or 60 cm, is made of graphite. The graphite tip, whose outside diameter is about 40 mm, with a wall thickness of about 15 mm, can be immersed directly in the metal without fear that it will crack from a sudden change in temperature. The thermocouple, of platinum and platinum-rhodium, is for insulation also enclosed, of course, in the usual manner in a pair of porcelain tubes.

Recently, thermoelements have begun to be manufactured by very many firms, whence also the various names of these instruments; however, in essence they all consist of a thermocouple and a galvanometer. The differences among the thermoelements themselves are in themselves inessential; for the accuracy of the readings it is important only that the alloys from which the wires are drawn be as homogeneous as possible.

There are, however, thermoelements of a quite special construction. Thus, Siemens and Halske manufacture, for direct immersion in hot gases and liquids, a thermoelement consisting of a copper tube through which a constantan wire insulated with asbestos is passed; the wire is soldered to the lower, closed end of the tube. Owing to its direct contact with the medium under investigation, such a thermoelement follows all fluctuations of its temperature very rapidly. Of course, it is applicable only to liquids that do not act chemically on copper. Its range of application is up to $+500^\circ$ C.; the accuracy of its readings is $\pm 5^\circ$ C.

The firm Paul Braun in Berlin manufactures instruments for temperatures up to $1250^\circ$ C., in which the thermocouple is made of a wire of pure nickel $a$, Fig. 81, and a tube $c$ of artificially prepared carbon.

Fig. 81.

Such carbon can be heated, without causing it any changes, to white heat provided there is no access of air, i.e. of oxygen; the latter is achieved by the fact that the carbon is enclosed hermetically along its entire length in a steel tube $d$, in which, after the absorption of the initially negligible amount of oxygen, the carbon is in an atmosphere of completely indifferent gases. The whole element is wrapped with asbestos cord $e$ and inserted into a second iron tube $f$, which can easily be replaced. When measuring temperatures above $1000^\circ$ C., the iron tube wears out too quickly, and therefore its end is replaced by a tip made of porcelain or quartz glass.

Since at temperatures above $1000^\circ$ nickel suffers from contact with carbon, the lower junction (“soldered joint”) is made...

by means of an intermediate iron head $g$, into which, together with it, a porcelain tube $b$ is inserted, insulating the nickel wire $a$ along its entire length. The ends of the thermoelement are brought out to terminals and surrounded by a jacket of perforated sheet metal for better cooling.

The electromotive force of this element is 2.5–3 times greater than that of an element made of platinum and platinum-rhodium, but it varies with temperature so irregularly that it cannot be expressed by a single equation, and the calibration of the instrument has to be done purely empirically. In Fig. 82 is shown the scale of a galvanometer with a temperature scale for the thermoelement described.

Next let us also mention the thermoelement of V. Bristol, who attempted to create an inexpensive instrument for measuring temperatures up to $1600^\circ$. The idea of his instrument can be explained with the aid of Fig. 83: $a$ and $b$ are two short pieces of wire made of platinum and platinum-rhodium, $d$ and $e$ are two wires made of inexpensive alloys possessing the property that, when their junctions with $a$ and $b$ are heated no higher than $650^\circ$ C, the electromotive forces arising in these junctions are equal and mutually opposite, so that the needle of the galvanometer, connected to the ordinary conductors $f$ and $f'$, indicates only the electromotive force of junction $c$; for the correctness of the indications it is therefore important that junction $c$ have time to assume the desired temperature before junctions $a\!-\!d$ and $e\!-\!b$ heat up to $650^\circ$ C. This is achieved by the fact that the end of the element, junction $c$, has no insulation and is subjected directly to the action of the hot body or space.

Fig. 83.

Fig. 82.

To protect junction $c$ from mechanical damage, an iron tube is slipped over the ordinary mounting of the thermoelement; when the thermoelement is inserted ...

into the space under investigation, is moved back slightly, exposing junction $c$; since junction $c$ assumes the required temperature after the lapse of a few seconds, the above-mentioned destructive action of the incandescent gases on the thermocouple does not have time to make itself appreciably felt; moreover, when the wires $a$ and $b$ are short, replacing them is comparatively inexpensive. Besides its cheapness, the merit of this instrument is the rapidity of the readings, which makes it possible to follow more precisely the fluctuations of the measured temperature. Its indications are, of course, less accurate than those of an ordinary Le Chatelier thermoelement, but for practical purposes they are quite satisfactory.

In conclusion we shall also mention a way of increasing the sensitivity of a thermoelement, i.e., its electromotive force, for a given temperature difference. For this purpose a number of thermocouples are connected in series, according to the scheme in Fig. 84, where $a$ is, for example, a constantan wire, and $b$ is copper. The electromotive force of such a compound thermoelement is, of course, equal to the sum of the electromotive forces of all the separate thermocouples. With such thermoelements one can measure the very slightest temperature fluctuations, down to $0.01^\circ$, and even less.

Fig. 84

Fig. 84.

As was indicated in the preceding chapter, in order to maintain a constant temperature of the cold junctions of a thermoelement they are often cooled with water; or they are simply immersed in a vessel with water, the temperature of which is measured with a mercury thermometer, or else with running water, using a special vessel. Such a vessel, manufactured by the firm Siemens and Halske and attached to the frame of the thermoelement, is shown in Fig. 85.

Fig. 85

Fig. 85.

For proper cooling, the length of the wires washed by the water must not be very small. The cooling water is supplied and removed through rubber tubes.

In order completely to eliminate the need not only to maintain a constant temperature of the cold junctions, but also to introduce a correction into the readings of the millivoltmeter, Prof. Bristol proposed an ingenious “compensator,” consisting of an additional resistance that decreases with increasing temperature; it consists of a glass vessel filled with mercury, Fig. 86, p. 99, with a long, narrow neck; a loop of platinum wire is immersed in the mercury; the copper ends connected with it and passed through the glass are soldered—one to one of the free ends of the thermocouple, the other to one of the conductors leading to the millivoltmeter. With an increase of the surroun—

of the measured temperature, and the temperature of the cold junctions rises; consequently, the readings of the millivoltmeter decrease. But at the same time the temperature of the mercury in the compensator also increases; the mercury expands, its level rises, and the length of the free ends of the platinum loop decreases, and together with this its resistance also decreases, which in turn increases the deflection of the millivoltmeter needle. By selecting the length of the loop in dependence on the internal resistance of the millivoltmeter, one can achieve complete elimination of the influence of changes in the temperature of the cold junctions. Bristol’s compensator is very convenient, especially when using millivoltmeters with not particularly high internal resistance.

Fig. 86.

26. Galvanometers for thermocouples.—As the measuring instrument, a precision millivoltmeter with a pointer is usually used, most often of the Deprez-d’Arsonval system; sometimes millivoltmeters of the same system, but with a mirror, are used.

Of the pointer millivoltmeters, the instruments of the firm Keiser & Schmidt are fairly widespread; their better, though also more expensive, instruments are those of Hartmann & Braun. In the instruments of the first firm the axis of oscillation of the pointer ends in points resting on stone bearings; in those of the second, Fig. 87, 1:5 natural size, the pointer is suspended on a tightly stretched thread. The merits of instruments of the first kind are that they are more robust and give correct readings even if they are not set strictly level, or if the stand trembles slightly; instruments with a thread are more sensitive, but also more fragile; they require very careful installation and extremely cautious handling.

Fig. 87.

Fig. 88.

Very accurate and convenient instruments are manufactured by Siemens & Halske. In Fig. 88 is also shown, at 1:5 natural size, a galvanometer with

with an index pointer on knife-edges. In Fig. 89 an instrument of the same firm is shown, but with a raised pointer. The mirror situated next to the scale makes it possible to take very accurate readings: when taking a reading one must look at the pointer so that it covers its image in the mirror. Before taking a reading the instrument must be brought, by means of the leveling screws, into a strictly horizontal position, using for this purpose the circular level $a$ located on its cover; at the side there is the arrestment $b$, which must be released, by turning it to the left, before taking a reading, and then immediately closed again. Failure to observe this rule shortens the service life of such a delicate instrument and can also easily cause it to break down.

Fig. 89.

Some firms also manufacture wall-mounted galvanometers with the edge of the pointer arranged at the side. Such an instrument, made by the Cambridge Scientific Instrument Co., is shown in Fig. 90. Such instruments are especially convenient for factory and plant premises, where it is difficult to find a sufficiently suitable, secluded place for installing a galvanometer.

Fig. 90.

In order to reduce the influence of secondary conditions—for example, the length and cross-section of the connecting conductors—and also to make it possible to measure simultaneously, with one and the same galvanometer, the temperature of several thermoelements, using a switch of ordinary design, millivoltmeters are deliberately made with a large internal resistance, 300–600 ohms; for thermoelements up to $1600^\circ$ C, even 800 ohms.

The galvanometers described above usually have, in addition to a scale graduated in millivolts, a second scale graduated in degrees C, which makes it possible to take readings of temperatures directly, without consulting a table or a curve of the dependence between electromotive force and temperature. At the same time, however, one must not forget that the tempe-

the temperature scale is applied purely empirically for the given thermoelement; for another one it may be unsuitable.

In especially precise measurements a galvanometer with a mirror is sometimes used. However, taking readings with such an instrument is considerably more difficult, and therefore they are used very rarely, it being preferred in such cases to carry out measurements by the zero, compensation method, the arrangements for which were discussed above.

In Fig. 91 there is shown, at 1:5 natural size, a compact set of instruments for the arrangement of Fig. 70, manufactured by Siemens & Halske; the same letters denote the same parts as in Fig. 70.

Fig. 91.

In those cases where it is desirable to know the successive course of a change in temperature, self-recording galvanometers are useful. Such an instrument, manufactured by Siemens & Halske and shown with its protective cover removed in Fig. 92, p. 102, consists of a Deprez–d’Arsonval millivoltmeter with a pointer, connected to a universal self-recording attachment of the same firm. Its operation is as follows: a clockwork mechanism uniformly moves a transparent paper tape and, placed beneath it, a cloth tape impregnated with ink, and at the same time, at equal intervals of time, usually every 1 minute, sometimes every 30, 15, or even 12 seconds, causes an arc, extending across the entire width of the paper tape and pressing the voltmeter pointer against it for one instant, to descend. This pressure produces a point,

which, thanks to the coordinate grid printed on the paper, indicates the temperature of the thermoelement at the given moment.

Since the pointer, in its movements, does not touch the paper, the sensitivity of this instrument is the same as that of an ordinary millivoltmeter.

Fig. 92.

The speed of motion of the paper is 20 or, if desired, 60 mm per hour. The length of the roll of tape inserted is about 45 m, i.e., it is sufficient for 2250 or, respectively, 750 hours; moreover, when it is not necessary to make a record, the arresting device is lowered, stopping both the pointer and the clockwork. The clockwork must be wound every 8–10 days.

Up to 5 thermoelements may be connected to this instrument simultaneously; for this it is only necessary to include an automatic switch, likewise actuated by the clockwork. It is true that the temperature curves for each of the thermoelements will no longer be continuous: namely, from each point there will be 4 marks at intervals of 12 seconds, and then the next group of marks only after 10 minutes, etc. If there are no rapid temperature fluctuations, then in this case as well it is easy to obtain a continuous curve.

The firm Hartmann & Braun makes self-recording galvanometers according to Fig. 93, p. 103. The general arrangement is the same as—

...the preceding instrument; the chief difference is that the recording paper is fastened to a drum which normally makes one revolution in 24 hours; the reading—the pressure of a pointer—is made every 45 s. For measuring the resulting diagram there is a transparent glass scale. In addition, the instrument has a scale and a pointer indicating the temperature at every moment. The instrument is very convenient and less bulky than the preceding one; in Fig. 93 it is also shown at 1:5 of natural size.

Fig. 93.

Another way of recording the readings of a millivoltmeter is by means of photography. In this case it was natural to make use of the more accurate mirror galvanometers. However, the whole apparatus becomes so bulky, complicated, and expensive that it is used only for especially precise scientific work; in other cases, simpler self-recording pyrometers are preferred.

27. Féry’s thermoelectric pyrometer.—The instrument described below is a combination of optical and thermoelectric pyrometers. Its action is also based on Wien’s law, mentioned above.

The sensitive part of the instrument, which is subjected to the action of the “black-body” rays, consists of crossed and soldered wires, copper and constantan, forming a thermocouple and connected to plates $a$ and $b$, insulated from one another, Figs. 94 and 95, p. 104, with terminals $m$ and $n$ for connecting the wires leading to the galvanometer.

The junction must be located at the focus of the concave mirror $c$, which collects the rays of the source under investigation. Focusing is carried out...

is done with the aid of the rack-and-pinion with head $i$; near the junction there are also two small mirrors, which throw into the eyepiece $d$ the image of the body whose temperature is being measured, cut by a horizontal line into two halves; until the instrument is in focus, two halves of the image are visible in the eyepiece, shifted with respect to one another; with correct adjustment both halves give one whole image. In addition, care must be taken that the image of the incandescent body slightly overlaps the junction, which then appears as a black disk in the middle of the field of view.

Drawing 94 and 95.

Drawings 94 and 95.

The readings of this instrument do not depend on its distance from the body whose temperature is being measured, at least up to a certain limit. This is explained by the fact that in front of the thermocouple there is a small diaphragm, which admits only part of the rays reflected by mirror $c$. Only when the pyrometer is so far removed that the rays reflected from $c$ all fall on the thermocouple and do not entirely cover it, so to speak, do not fill the diaphragm, will the readings of the instrument become smaller than the true ones; this is recognized by the fact that the luminous image will then prove smaller than the image of the junction and will not cover it. Thus, when measuring the temperature of a large steel ingot, the same reading is obtained at distances of both 1 and 20 m.

At the front of the instrument there is a hinged cover $f$ with a diaphragm, thanks to which, with one and the same instrument, temperatures can be measured over very wide limits.

For this purpose the galvanometer has 3 scales: one, a temperature scale, for measuring from $500^\circ$ to $1500^\circ$, in which case the cover $f$ must be completely thrown back; a second, also a temperature scale, for $1500^\circ$ to $2300^\circ$, in which case the cover is closed, but the diaphragm is moved aside, leaving half of the whole area open; and a third—in millivolts, by which readings may be taken for temperatures from $2000^\circ$ to $2950^\circ$, closing the diaphragm halfway, i.e., leaving open only one quarter of the whole area; in this case the temperature is determined

by means of the auxiliary table. By covering the diaphragm still more, it is possible to measure still higher temperatures; thus, Prof. Féry himself, with the aid of this instrument, found the temperature of the sun to be equal to $7800^\circ$ C.

For convenience in measurements, the pyrometer is placed on a tripod stand. In Fig. 96 the external appearance of the instrument is shown with the cover closed, and in Fig. 97 with the cover thrown back.

Since the thermoelement, even when measuring the highest temperatures, is not heated above $80^\circ$—$100^\circ$ C., its properties and, consequently, its readings do not change even after prolonged use.

The galvanometer supplied with this pyrometer, of the Meylan–d’Arsonval system, has an internal resistance of 8.9 ohms; its scale is from 0 to 4 millivolts, with divisions of 0.1 millivolt; the other two scales are graduated in ${}^\circ$C; the accuracy of the readings of this instrument, when used properly, for temperatures from $700^\circ$ to $2300^\circ$ is $\pm 10^\circ$, and above this $\pm 50^\circ$. The graduation of the temperature scales may be performed either on the basis of Wien’s law or, more reliably, like Wanner’s pyrometer, by comparison, directing it at a thermoelement heated in an electric furnace.

Figs. 96 and 97.

Instead of the usual galvanometer, the instrument, if desired, may also be furnished with a recording galvanometer of the kind shown in Fig. 93.

The instrument described has, in general, the same advantages as photometric pyrometers. In comparison with ordinary thermoelements it is more convenient for measuring the highest temperatures, from which the sheaths of ordinary thermoelements suffer. However, it is not suitable for those cases when the temperature being measured fluctuates rapidly, since a certain, though small, time is required for the thermoelement to assume the corresponding temperature.

Measurement of Electrical Resistance

28. General remarks.—The instruments described below are based on the increase in the resistance to the passage of an electric current through metallic conductors as their temperature rises. Thus, knowing for a given conductor the relation between its resistance and temperature, one measures the resistance $r$ of the conductor subjected to the action of the temperature sought, and from $r$ calculates the corresponding temperature $t$.

C. W. Siemens found that the resistance $r_t$ at the temperature $T = 273^\circ + t^\circ$ C. can be expressed, for all the metals he investigated—platinum, silver, aluminium, copper, and iron—within the limits up to $350^\circ$ C., by the equation

\[ r_t = \alpha T^{\frac{1}{2}} + \beta T + \gamma . \tag{62} \]

Of all metals, platinum proved, according to the experiments of Siemens and subsequent investigators, to be the most suitable for the indicated purpose for the following reasons:

its electrical resistance undergoes a smaller change with increasing temperature than that of other metals;
the purity of the metal has a very great influence on the accuracy of measurement by this method, and platinum is easier than other metals to obtain chemically pure;
platinum does not oxidize and in general does not change chemically during operation.

For platinum the coefficients in equation (62), according to Siemens’s experiments, are equal to

\[ \alpha = 0.039369;\quad \beta = 0.00216407;\quad \gamma = -0.24127. \]

Subsequently, by comparing the Siemens instrument with other pyrometers, it was established that equation (62) is also valid for considerably higher temperatures, approximately up to $900^\circ$.

According to Callendar’s most recent experiments, somewhat different coefficients should be taken in equation (62), namely

\[ \alpha = 0.03771;\quad \beta = 0.00252;\quad \gamma = -0.245. \]

The difference is explained partly by the increased accuracy of the measurements, but mainly by the fact that at the present time it has become possible to obtain…

to regard the platinum as almost chemically pure. The extent to which the purity of platinum affects its resistance may be judged from the following figures: the temperature coefficient of resistance of ordinary platinum is 0.00300, while that of chemically pure platinum is 0.00346.

Moreover, Siemens’ expression cannot be considered successful for the very reason that at $T=0$ it gives a negative value for $t$, which is obviously impossible.

The dependence under consideration may better be expressed in the form

\[ r=r_0\left(1+\alpha t+\beta t^2+\gamma t^3\right), \tag{63} \]

where $r_0$ is the resistance at $0^\circ$ C, and $\alpha$, $\beta$, and $\gamma$ are constants calculated from observations.

Callendar’s very careful experiments with the resistance of platinum may, according to Eq. (63), be reduced to the expression

\[ r=r_0\left(1+0.0034508\,t+0.0000002\,t^2-0.00000000108\,t^3\right), \tag{64} \]

or, almost as closely, to an expression with only two constants,

\[ r=r_0\left(1+0.0034675\,t-0.0000000075\,t^3\right). \tag{65} \]

Unfortunately, Eqs. (64) and (65), while representing the law well within the limits of the experiments from $0^\circ$ to $600^\circ$ C, are entirely unsuitable for extrapolation, since they give a certain maximum for $r$ at $t$ somewhat above $1000^\circ$, which is obviously also impossible.

In view of this, at the present time use is made chiefly of an equation of parabolic form,

\[ r=r_0\left(1+\alpha t+\beta t^2\right), \tag{66} \]

which permits extrapolation.

Without dwelling on the values of $\alpha$ and $\beta$ given by various authors such as Benoit, Hempidge, and others, in view of deficiencies subsequently discovered either in their measurements or in the initial numbers, we shall confine ourselves to Callendar’s expression

\[ r=r_0\left(1+0.003505\,t-0.0000004\,t^2\right); \tag{67} \]

the only defect of Eq. (67) is that it does not fully agree with observations at low temperatures.

The best results, according to Callendar’s experiments, are given by the expression

\[ r = r_0 e^{\frac{\alpha t}{1+\beta t}}, \tag{68} \]

where for platinum $\alpha = 0.0034259,\ \beta = 0.001529$, but it is less convenient for practical use.

In conclusion, let us also indicate a quite peculiar method of calculating the temperature from the measured resistance, likewise proposed by Callendar. He introduces a new concept—the “platinum scale”—after the model of the mercury scale; namely, if the resistance of a certain length of platinum wire at $0^\circ$ is equal to $r_0$, and at $100^\circ$ is equal to $r_{100}$, then, assuming a linear increase of resistance, we obtain for $1^\circ$ C. an increase of resistance of

\[ \frac{r_{100}-r_0}{100}. \]

If at the temperature sought, $t$, the resistance proves equal to $r_t$, then under the indicated assumption

\[ t=\frac{r_t-r_0}{r_{100}-r_0}\cdot 100. \tag{69} \]

If one takes such a length of wire that $r_{100}-r_0=1$ ohm, then the whole measurement is reduced simply to determining the increase in resistance $r_t-r_0$. For example, $r_t-r_0=7.58$ ohms gives directly $758^\circ$ C.

As we have already seen, the relation between $t$ and $r$ is not linear; therefore the temperature found from equation (69) requires a correction. If the temperature on the “platinum” scale is denoted by $t_p$, then, according to Callendar’s experiments, the correction $\Delta t$ may be calculated from the equation

\[ \Delta t=t-t_p=\delta\left[(0.01t)^2-0.01t\right], \tag{70} \]

where $\delta$ is an empirical coefficient depending on the material.

Equation (70) is also the equation of a parabola and is in essence equivalent to equation (66). According to Callendar’s experiments, for $\delta$ one should take the value 1.57. The latest experiments at the National Physical Laboratory in England have confirmed the validity of equation (70) up to $1000^\circ$ C.

Taking for the platinum used by the Cambridge Scientific Instrument Company, the largest firm manufacturing thermometers based on electrical resistance, $\delta=1.5$, we finally obtain

\[ t=\frac{1025}{0.3}-\sqrt{\left(\frac{1025}{0.3}\right)^2-\frac{1000t_p}{0.15}}. \tag{71} \]

However, in practice, instead of computing the true temperature $t$ from eq. (71), having found $t_p$, the corresponding $t$ is taken directly from the table supplied by the firm with the instrument, or from a table calculated in advance, or from the corresponding diagram. Table 15 gives an idea of the magnitude of the deviations of $t_p$ from the true temperature. For practical use such a

Table 15.

$t_p$	$t$	$t_p$	$t$	$t_p$	$t$	$t_p$	$t$
°C	°C
0	0	300	309.7	600	654.4	900	1049.5
50	49.6	350	364.5	650	716.2	950	1122.0
100	100.0	400	420.2	700	779.4	1000	1197.0
150	151.2	450	477.0	750	844.3	1050	1274.5
200	203.1	500	534.9	800	910.8	1100	1355.1
250	256.0	550	594.0	850	979.1

table must, of course, be compiled for smaller temperature intervals, e.g., every $10^\circ$ or even $5^\circ$.

29. Methods of measurement.—All the numerous circuits for measuring changes in resistance, at first sight very diverse, can be reduced to the following 4 basic methods.

Method 1: the unknown, varying resistance $r_t$, Fig. 98, is determined from a known, constant resistance $r$; namely, the current from the storage cell $a$ branches at point 1, and, according to Kirchhoff’s law, through each of the branches there will flow a current whose strength is inversely proportional to the resistance of the given branch, i.e.

Fig. 98.

\[ i_1 : i_2 = \frac{1}{r} : \frac{1}{r_t}, \]

whence

\[ r_t = \frac{r i_1}{i_2}, \tag{72} \]

knowing $r$ and measuring the current strengths $i_1$ and $i_2$ in each of the branches by means of the corresponding instruments $i_1$ and $i_2$, Fig. 98, we shall find the required resistance $r_t$, and from it the corresponding temperature.

The indicated method was used in the first instruments of the type described by Siemens himself, who, in view of the absence at that time of sufficiently sensitive ammeters, took as instruments for measuring current two voltameters, in which the measure of current strength was the volume of gases formed during the electrolysis of water. This method required rather bulky and fragile instruments, a rather long time for each reading, and at the same time did not possess great accuracy, no more than $\pm 1$ to $2^\circ/0$. That is why it has now been completely abandoned.

If the voltameters $i_1$ and $i_2$ are replaced by two ammeters, the indicated drawbacks are eliminated, but the circuit, owing to the presence of two measuring instruments, nevertheless becomes more complicated, and therefore less reliable than other circuits. One may, of course, leave only one ammeter, but then the current strength of the cell $a$ must be known.

In such a case, in addition to equation (72), we have

\[ i = i_1 + i_2, \]

whence

\[ r_t = r \frac{i_1}{i - i_1} \quad \text{or} \quad r_t = r \frac{i - i_2}{i_2}. \tag{73} \]

Here, with known $r$ and $i$, it is sufficient to measure the current strength in one of the branches.

Such a circuit is quite convenient, but has the following disadvantages: $r_t$ is not obtained directly; it must still be calculated; for accuracy of the readings it is necessary to maintain the current strength of the cell $a$ constant; and, chiefly, the sensitivity of the instrument for small changes of $r_t$ is not great.

Method 2. Kapsel proposed the circuit, Fig. 99, which is obtained from the preceding one if one adds to it, in the upper section, a resistance $r_1$, and in the lower section one more accumulator cell $a_2$; if the electromotive forces of the cells are respectively $e_1$ and $e_2$, then through the branch $1\!-\!2$ there will flow a current whose strength, according to Kirchhoff’s law,

Fig. 99.

\[ i=\frac{e_1 r_t-e_2 r_1}{r_t(r_1+r_2)+r_1 r_2}; \tag{74} \]

if, for convenience of measurement, elements for which $e_1=e_2$ are taken, then

\[ i=\frac{e(r_t-r_1)}{r_t(r_1+r_2)+r_1 r_2}, \tag{75} \]

i.e., the current $i$, measured by the milliammeter $m$, is approximately proportional to the difference of the resistances; a small change in the resistance $r_t$ causes a considerable change in the current $i$, and together with it a large deflection of the pointer $m$. For example, a change of $r_t$ by $10\%$ may change $i$ by $100\%$.

From Eq. (76) we obtain the value of the resistance

\[ r_t=\frac{r_1(e+i r_2)}{e-i(r_1+r_2)}. \tag{76} \]

The circuit is very simple and, unlike method 1, very sensitive; therefore it is especially convenient for measurements of not very high temperatures, when the smallest fluctuations must be noted.

Method 3: this method is also obtained from the basic method 1 by connecting two branches by means of a Wheatstone bridge, Fig. 100.

If $r_0$ is the resistance of the bridge and of the milliammeter $m$, $r$ the resistance in the section $1\text{-}a\text{-}2$, $e$ the electromotive force of the element, $i$ its current, then, denoting the currents passing through the branches respectively by $i_1, i_2, i_3$ and $i_4$, and through the bridge by $i_0$, according to Kirchhoff’s law one may write the following 6 equations:

Fig. 100.

\[ \begin{aligned} i-i_1-i_3&=0; & i r+i_1 r_1+i_2 r_2&=e;\\ i-i_2-i_4&=0; & i_0 r_0-i_1 r_1+i_3 r_3&=0;\\ i_0+i_1-i_2&=0; & i_0 r_0+i_2 r_2-i_4 r_t&=0; \end{aligned} \]

eliminating from these equations the variable and unknown quantities of the current $i, i_1, i_2, i_3$ and $i_4$, we obtain for $r_t$ the expression

\[ r_t=\frac{e r_2 r_3+i_0 A}{e r_1-i_0 B}, \tag{77} \]

where

\[ A=r_1r_2r_3+r_1r_0r_3+r_0r_2r_3+rr_1r_0+rr_2r_0+rr_3r_0+rr_1r_2+rr_2r_3, \]

\[ B=rr_0+rr_1+rr_3+r_1r_0+r_2r_0+r_1r_2+r_2r_3+2r_1^2+3r_1r_3+2r_1^2r_3r^{-1}+2r^2r_3r^{-1}. \]

Despite their complicated form, the quantities $A$ and $B$ do not make the computations difficult, since for each given setup they do not change and can be computed in advance once and for all.

Since the voltage of the current $e$ must be known, then, by measuring $i_0$, we obtain all the data for determining $r_t$.

In practice the instrument is calibrated so that, for a definite $e$, the deflection of the milliammeter pointer indicates directly the desired temperature $t$. To this end, at some initial temperature, for example $0^\circ$ C., all 4 resistances are made equal; the pointer will stand at 0, deflecting to one side or the other when $r_t$ changes. The circuit is very convenient and widely used.

The Fisher-Crampton circuit shown in Fig. 101 differs in essence from the one just analyzed only in that $g$ is a differential galvanometer, whose pointer receives a deflection in one direction under the action of the current going through the branch $1-r_1-3$, and under the action of the current in the branch $4-r_2-2$—in the other direction; when the resistance $r_t$ changes, the current through $4-r_2-2$ will change, and the magnitude of the change can be judged from the change in the deflection of the pointer $g$. In comparison with the preceding circuit, this circuit has no advantages, but at the same time requires a more complicated differential measuring instrument.

Fig. 101.

Method 4 is obtained from method 3 if one of the constant resistances $r_1$, $r_2$, or $r_3$ is replaced by an adjustable one, Fig. 102, p. 113, and, when $r_t$ changes, it is adjusted so that the pointer of the galvanometer $g$ returns to 0, i.e. so that no current flows through the bridge; in Eq. (77) $i_0=0$, and

\[ r_t=\frac{r_2r_3}{r_1}. \tag{78} \]

If we set $r_1 = r_2$, then we obtain directly $r_t = r_3$.

This method is very convenient and accurate. The idea was first proposed by Spor shortly after the appearance of Siemens’ method. Instead of a galvanometer, Spor used a telephone, in which the electric current produced a noise that disappeared at the moment when, by gradually changing the constant resistances, i.e., those independent of temperature, the absence of current in the Wheatstone bridge was achieved.

$Diagram of a Wheatstone-bridge circuit with resistances $r_1$, $r_2$, $r_3$, $r_t$, a source $a$, and galvanometer $g$.$

Fig. 102.

Instead of changing $r_3$, Fig. 102, one may change $r_1$ or $r_2$, whereby neither the circuit nor equation (78) is altered in substance. Finally, an additional resistance may be inserted in the Wheatstone bridge, and by varying this additional resistance one may bring the galvanometer needle back to 0. But even such a circuit, proposed by some firms, is essentially no different from the one analyzed. Of course, for it the fundamental equations, and with them equation (78), take on a somewhat different form; the resistance $r_0$, which has now become variable, will also enter into it.

Comparing the individual methods with one another, one may say that method 3 is the most convenient and simple; method 2 is more sensitive and especially convenient for small changes of temperature; method 4—the null method—is somewhat more complicated and, moreover, is not applicable in measuring rapidly fluctuating temperatures, but its advantage is the greater accuracy of its readings. With it one can use a much more sensitive galvanometer, since its scale, for the same length, may have only a small number of very large divisions; at the same time, as will be seen below, it is more accurate, since it does not depend on changes in the voltage of the element $a$, and with it it is easy to eliminate the influence of added resistances. In addition, it at once gives the desired resistance $r_t$, whereas method 3 requires calculations or the use of an empirically graduated scale, which is always less accurate. Method 1 is at present to be set aside as insufficiently sensitive.

The circuits examined exhaust all those proposed up to now, since the differences among the various circuits encountered in the literature arise not so much from differences in their essence as from differences in their representation and from the addition to them of many secondary—

degree auxiliary devices, which only obscure the clarity of the diagram.

30. Sources of errors and their elimination.—In instruments based on the measurement of electrical conductivity, just as in all others, the accuracy of the measurement depends not only on the construction of the instrument, but also on its handling. As was to be expected, the chief attention when using the instruments now under consideration must be directed to the electrical side, with which we shall begin.

When using the most widespread method 3, as is evident from equation (77), not only the current strength but also the voltage $e$ in the main circuit enters into the expression for the resistance $r_t$. It is possible, of course, each time to measure this voltage with a special instrument and then calculate $r_t$ from equation (77), but in practice it is preferred to work with a constant $e$, which is all the easier since the current consumption during the measurement is negligible and, consequently, the drop in voltage of the storage cell is very slow. In order that, despite the change in the cell voltage, a constant voltage $e$ should remain in the circuit, the diagram of Fig. 100 must be supplemented by two additional resistances $r_4$ and $r_5$, Fig. 103. The control resistance $r_4$ is chosen of such a magnitude that, when it is switched in by means of the switch $c$ instead of the resistance thermometer $r_t$, with the proper $e$ the pointer of the milliammeter stands at a definite mark; if the voltage of the cell $a$ has decreased, the deflection will be smaller, and then part of the resistance $r_5$ is cut out until the pointer assumes the required position; if, with $r_5$ completely cut out, the pointer still does not reach the mark, the storage battery must be recharged.

Fig. 103

Fig. 103.

In Koepsel’s method 2, Fig. 99, the checking of $e$ is carried out by the same switching over to a control resistance, the setting of the pointer $m$ to the required mark being achieved by changing the resistance $r_2$, which serves as a shunt for the ammeter $m$. For use of equations (75) and (76) in this case it is also necessary to ensure that the voltages of both cells $a_1$ and $a_2$, or, more precisely, the currents sent by them into the branch $1$-$2$, are strictly identical. In order to regulate the small difference that is always possible, the connection at point 2 is made by means of a small half-coil resistance,

set so that, at a definite initial temperature, for the resistance $r_t$ the pointer stands at 0.

In method 4, the zero method, the voltage $e$ has no significance, which is one more advantage of it over method 3.

Independently of the method of measurement, in the instruments under discussion there exists a source of errors in measuring $r_t$, arising from the fact that the temperature changes not only of the measuring platinum spiral itself, but also of the conductors attached to it and leading to the measuring apparatus; regarding the temperature of these conductors one can say only that it varies from the desired $t$ to $t_0$ of the surrounding medium, exactly like the gas in the connecting tube of gas thermometers, with which these conductors present a complete analogy, and for this reason they may be called a “parasitic resistance.” As there the influence of the “parasitic” space, so here the influence of the “parasitic” resistance can be allowed for by adding part of it to the resistance $r_0$.

For the zero method this influence can be completely eliminated by means of Callendar’s compensating loop, Fig. 104: $c,c$ are the conductors whose “parasitic” resistance must be eliminated; they are made of wire of the same metal as the resistance $r_t$, but of large cross-section in order to lessen their influence; $e$ is a loop of the same metal, of the same cross-section and the same length as $c$, placed parallel to $c$ and connected into the section $4-r_3$. Thus, to the extent that the parasitic resistance of the conductors $c$ increases, the resistance $e$, i.e. that of the section $4-r_3$, increases by exactly the same amount, and, with $r_1=r_2$, equation (78) still gives exactly $r_t=r_3$. The only drawback of this loop is the necessity of having 4 conductors from the resistance thermometer.

Fig. 104.

For method 3 the use of a compensating loop cannot give complete elimination of the influence of the parasitic resistance, as is seen from equation (77), but it does give some improvement; the main thing is that the parasitic resistance must be made as small as possible, and then its change will not be so appreciable.

In order that temperature fluctuations should not affect the readings of the instruments, the remaining conductors, from the thermometer to the meas—

The connecting leads from the measuring device, the galvanometer, and the Wheatstone bridge must have as small a resistance as possible. For technical purposes it is usually recommended that these leads be copper, insulated, and of such a cross-section that their total resistance be no more than 1 ohm. In very precise experiments the leads are made short and so thick that their resistance is about 0.01 ohm. At the same time, if this condition is not observed, then one also cannot use, without considerable inaccuracy, the temperature scale applied when the instrument is manufactured; instead, the temperature must be calculated by means of the corresponding equations (76), or (77), or (78).

For a more considerable length of the leads, it is better to calculate their cross-section on the basis of the magnitude of the allowable error due to temperature fluctuations in the rooms through which the conductors pass. This is most simply explained by an example.

Example: the temperature is measured in a drying chamber, where it fluctuates from $+40^\circ$ to $+90^\circ\text{C}$; the resistance of the thermometer is equal, at $0^\circ\text{C}$, to 50 ohms, increasing for each $1^\circ$ by approximately 0.18 ohm. If the accuracy of the measurement is sufficient to $\pm 1\%$, then the resistance of the copper leads may vary by

\[ (90 - 40)\cdot 0.18 \cdot 0.01 = 0.09\ \text{ohm}; \]

suppose the temperature in the rooms fluctuates from $+5^\circ$ to $+25^\circ\text{C}$; assuming that the resistance of copper leads increases with temperature by $0.4\%$ for each $1^\circ$, we obtain, for a fluctuation of $25^\circ - 5^\circ = 20^\circ$, a change of $0.4 \times 20 = 8\%$.

Hence the total resistance $r$ of the leads must be no more than

\[ r \cdot 0.08 = 0.09\ \text{ohm}, \qquad r < 1.1\ \text{ohm}. \]

If the length of one lead is $L = 60\ \text{m}$, and the specific resistance of copper is $c = 0.017\ \text{ohm}$, then the cross-section $f$ is obtained from the equation (60) given above:

\[ f = \frac{c \cdot L}{r} = \frac{0.017 \cdot 2 \cdot 60}{1.1} \approx 2\ \text{mm}^2. \]

In order to be sure that the resistance of the conductors does not change with time, it is better to make all connections by soldering with tin, rather than by clamps and screws, whose contact surfaces easily become covered with oxide, greatly…

increasing the resistance. Where clamps are unavoidable, the surfaces must be inspected and cleaned.

For correctness of the readings, the constant resistances must not vary with temperature; in these types they are made of manganin, whose temperature coefficient is only 0.00001. Adjustable resistances, e.g. $r_3$, Fig. 102, for precise measurements are best made in the form of resistance boxes with plugs; the latter must have a sufficiently large, well-fitted surface.

Finally, care must be taken that there be no thermoelectric phenomena, for which purpose all connections should, as far as possible, be of identical metals and have the same temperature.

In especially precise measurements, the influence of thermoelectric currents can be eliminated by making two readings—one with the current from the accumulator in the forward direction, the other in the reverse direction. In this case it is convenient to use a mercury switch, either of a special construction or consisting of simple cups with mercury, into which the ends of the conductors from the accumulator are immersed alternately.

How much thermoelectric currents affect the readings may be shown by P. Rote’s observations when measuring the temperature of liquid air with the aid of an ordinary instrument and an improved one according to Fig. 124, p. 123. Reading I was made with one direction of current, reading II with the reverse; the difference, converted to $^{\circ}\mathrm{C}$, is $0.8^{\circ}$ for the English instrument and only $0.08^{\circ}\mathrm{C}$ for Rote’s instrument.

Table 16.

Reading No.	English instrument	Improved Rote instrument, millivolts
I	2.469	3.7816
II	2.472	3.7800
difference	$+0.003$	$-0.0016$

Regarding the strength of the current passed from the accumulator, it should be noted that it is generally not large, only about 0.1 ampere. For especially precise observations, e.g. by the zero method, the current strength is taken still smaller, about 0.01 ampere, since otherwise the current, in passing, heats the platinum spiral and thereby somewhat increases the readings of the instrument. However, with a current strength of 0.01 ampere this heating, according to Callendar’s observations, gives an error of less than $0.02^{\circ}\mathrm{C}$.

In conclusion, a few words must be said about the insulation of the platinum spiral: its turns are sometimes wound on a porcelain rod. However, at high temperatures the resistance of porcelain noticeably decreases; moreover, platinum becomes brittle and in places

adheres to the porcelain rod, which can be explained by admixtures in the porcelain. Very good insulators can be made of mica; unfortunately, at a temperature of about $+800^\circ$ C, mica gives off water and becomes very brittle. At the present time mica is being successfully replaced by quartz glass.

On the same question one may also point out the following: when measuring temperatures below $0^\circ$ C, the air enclosed in the tube with the platinum spiral releases moisture in the form of hoarfrost, which gives an additional, though weak, short-circuiting of the individual turns of the spiral; this is reflected in the readings, giving an exaggerated temperature. To avoid this phenomenon it is necessary to seal the tube and to admit air through a drying device. The construction of such an instrument is indicated below, p. 123.

Summarizing all the foregoing, one may say that all sources of error can in part be avoided, as, for example, in the zero method; in part they can be reduced to an insignificantly small value; and, in the extreme case, as, for example, the influence of temperature on the galvanometer and the conductors, they can be allowed for with sufficient accuracy according to the same rules as in thermoelectric measurements.

31. Description of various designs.—Although above we have been speaking all the time about measuring the resistance of platinum, as the metal most suitable for the indicated purpose, in practice instruments are also encountered with another metal: for measuring temperatures from $-50^\circ$ to $+100^\circ$ C the spiral is made of thin iron wire; for damp rooms or in the presence of acid vapors, instead of iron wire, gold wire is sometimes used; for temperatures from $+100^\circ$ to $+600^\circ$ C, nickel wire is often used.

Fig. 105.

Fig. 105.

Fig. 106.

Fig. 106.

The designs themselves are quite varied, depending on the purpose of the instrument and the method of producing the readings.

In Fig. 105 there is shown, approximately at $1:8$ of actual size, an instrument for measuring temperature in a room when using a circuit...

Kapsel, method 2; the resistance, made of thin iron wire, and in appropriate cases of gold wire, is surrounded by a casing of perforated sheet metal to protect it from mechanical damage. The instrument is intended for readings at a distance; a mercury thermometer attached to the casing serves for readings on the spot and for checking.

In Fig. 106 there is shown, approximately on the same scale, the corresponding galvanometer with a temperature scale directly on it; for measuring the temperature at 10 places with this same galvanometer, a switch is located beneath it. The extreme left-hand button, with the pointer at 0, serves to switch off the galvanometer; the next two buttons, $R_0$ and $R$, are used in regulating the current strength of both cells, as mentioned above, p. 114.

Fig. 107.

In Fig. 107 there is shown an apparatus of the same circuit, but for measuring, for example, the temperature of superheated steam; the instrument is screwed directly into the wall of the steam pipe; the resistance is made of nickel wire; the casing is of iron; the instrument is suitable up to $500^\circ$ C. The firm G. A. Schulze builds similar instruments for measuring temperatures in flues, but only with a longer rod, from 100 to 300 cm, and with an iron tube open at the bottom.

Now let us proceed to the description of the most widespread instruments operating according to method 3. In Fig. 108 there is shown a thermometer, manufactured by the firm P. Braun, for being screwed directly into a pipeline or into a flange in the brickwork of a boiler. The spiral is of nickel, the casing steel; length $l$ from 150 to 2000 mm; the instrument is suitable for temperatures up to $500^\circ$ C.

Fig. 108.

Fig. 109.

In Fig. 109 there is shown, at $1:2$ natural size, an instrument made by the same firm for measuring the temperature of a liquid directly—

by direct immersion of the instrument in it: a thin nickel wire a is wound on a porcelain tube open at both ends; the wire is enclosed in a hermetically sealed casing made of copper tubing b; the conductors to the galvanometer consist of a flexible two-core cable with a lead sheath c. With the indicated construction, by bringing the instrument into contact with the liquid both outside and inside, a very rapid following of changes in the temperature of the liquid under investigation is achieved.

Fig. 110.

In Fig. 110 there is shown, at 1:4 natural size, an instrument manufactured by the firm Hartmann & Braun: a thin platinum strip, wound spirally on a mica plate, is enclosed in a flat metal casing. The flange at the lower part of the head is fastened to a coupling for screwing into a pipeline; in the case of a flat wall of a vessel containing the liquid under investigation, the instrument can be screwed by its flange directly to this wall. This instrument is intended for measuring temperatures from $-100^\circ$ to $+500^\circ$ C. The galvanometer is supplied directly with a temperature scale. In Fig. 111 there is shown, at 1:4 natural size, a wall-mounted type of galvanometer for temperature limits from $-20^\circ$ to $+60^\circ$ C.

Fig. 111.

The most successful of the instruments of the type being examined must be considered the recently introduced design manufactured by the firm W. C. Heraeus, Figs. 112 and 113, p. 121: the platinum wire is wound in the form of a spiral on a quartz-glass rod heated to softening and is inserted into a thin-walled tube likewise of quartz glass; after this, the tube is heated to softening and the air is pumped out of it, as a result of which the tube fits tightly on the rod, so that the turns of platinum come close to the surface of the tube; owing to the absence of an air layer, such a thermometer follows almost

instantly to changes in the temperature being measured and is not afraid of the sharpest fluctuations of it, which is its great advantage over ordinary mercury thermometers; from the platinum resistance to the terminals in the head run two conductors; for temperatures up to $+400^\circ$ C the conductors are made of silver, up to $+900^\circ$ of gold wire; the connection with the platinum ends is made not by soldering but by fusion.

Figs. 112–113

Fig. 112–113.

Fig. 114

Fig. 114.

Fig. 115

Fig. 115.

Fig. 116

Fig. 116.

In Fig. 114 such a thermometer for scientific measurements is shown at $1:2$ natural size; the lower thin part has a thickness of only 4 mm, the length of the platinum spiral is 60 mm, and its resistance at $0^\circ$ C is 50 ohms; similar instruments are made with a thickness of only 3 mm, a spiral length of 20 mm, and a resistance of 25 ohms.

For technical purposes the instrument is provided with a casing made of an iron or steel tube, Figs. 115 and 116, both at $1:5$ natural size; the length of the spiral in them is 60 mm, and its resistance is 50 ohms. The temperature limits are from $-200^\circ$ to $+700^\circ$ C. For measuring temperature in a room, the instrument according to Fig. 117, $1:2$ natural size, p. 122, is intended.

In Fig. 118 is shown, at $1:5$ natural size, the external view of a Wheatstone bridge for the instruments according to Figs. 114–117.

Finally, the instrument for measurement by method 4, the zero method, manufactured by the Cambridge Scientific Instrument Company, is shown in Figs. 119–121, p. 122: $a$ is a platinum spiral wound on a cross of mica plates, $b$—four likewise platinum conductors, more precisely 2 conductors and one Callendar compensating loop; $c$ disks of

traces, supporting these conductors; $d$ is a porcelain tube, $e$ a steel tube-casing, which is advised to be removed when measuring temperatures above $700^\circ$ C.; $f$ is also a steel tube; the head $g$ is made of boxwood, and $i$ are clamps leading to the Wheatstone bridge. The instrument is intended for up to $+1200^\circ$ C.

Fig. 117

Fig. 117.

Fig. 118

Fig. 118.

In Fig. 122 the same kind of instrument is shown, but one more adapted for technical measurements; the same letters denote the same parts as in Fig. 119; $h$ is a porcelain insulator. The cast-iron head $g$ protects the clamps from mechanical damage, the action of moisture and acid vapors; for the same purpose the conductors to the bridge are made in the form of a cable; in cases where this need not be feared, the porcelain head may be installed according to Fig. 123.

Figs. 122–123

Figs. 122–123.

Figs. 119–121

Figs. 119–121.

For measuring temperatures below $0^\circ$ C. the thermometer mentioned above, p. 118, shown in Fig. 124, is used; for hermetic sealing the end $h$

the tube $d$ is filled with wax or paraffin and, in addition, is compressed by the nut $f$ with a conical thread; air can enter $d$ only after passing through the handle $k$, filled with phosphoric anhydride or calcium chloride; $i$ are clamps surrounded by a copper casing $g$, to prevent undesirable thermoelectric phenomena under the influence of radiation and air motion.

Fig. 124.

Fig. 125.

As a Wheatstone bridge and galvanometer one may use instruments of the ordinary type, but the Callendar construction, manufactured by the Cambridge Scientific Instrument Company, is also very convenient, where both the bridge, the galvanometer, and the cells are assembled in one small box, Fig. 125, about $1:7$ natural size: $P$ is the thermometer itself, $B$ the galvanometer, whose pointer is brought to zero by changing the resistance $r_3$, Fig. 104, for which the flywheel $H$ is turned; $F$ is a spring lever for closing the current; the movable scale $A$ gives the resistance $r_3$, or even directly the desired temperature $t_n$ in ${}^{\circ}\mathrm{C}$. $C$ is a pair of storage cells, and $R$ is the regulating resistance for setting the required voltage.

In order that the same Siemens instrument may be used for thermometers with different values of resistance at the initial temperature, i.e. that readings in $0^\circ$ C. may be obtained directly on scale $A$, the firm inserts in the branch $b\text{–}c$ of the bridge, Fig. 104, a special resistance of manganin, exactly equal to the resistance of the platinum spiral of the thermometer at $0^\circ$ C., the so-called ice coil. When changing the thermometer it is enough also to insert in the branch $b\text{–}c$ the corresponding ice coil supplied with each thermometer, in order to use the Siemens instrument without any changes of the scale.

As a source of electric current in modern designs, accumulator cells are most often encountered, one, sometimes two, with a voltage of about 2 volts and a capacity of 10–15 ampere-hours. Accumulators may be used in two ways: either in the ordinary manner, charging them beforehand and then, as they discharge during operation, compensating the fall of their voltage by changing the corresponding resistance $r_2$, Fig. 103, or, if near the place where the temperature is measured there is electric wiring with a constant current, for example for lighting, the accumulator may be used only as an equalizer, giving a current of small, constant voltage.

In such a case the current from the main wires $b,c$, Fig. 126, goes into the accumulator $a$ through the resistance $r$; to the thermometer, however, the current goes along the wires $d,e$. The resistance $r$, which is made, for example, of iron wires, takes up the larger fluctuations of voltage, while the accumulator evens out the small, more frequent fluctuations. With such an arrangement there is no need for the regulating resistance $r_2$, Fig. 103; the checking resistance $r_4$ remains only for checking the correctness of the whole installation. When using the scheme of Fig. 126, by changing the resistance $r$ and correspondingly applying a second temperature scale to the galvanometer, one can obtain an instrument either with two measurement ranges, i.e. with different limits, for example from $0^\circ$ to $+400^\circ$ and from $+400^\circ$ to $+800^\circ$, or, by somewhat modifying the bridges, with different divisions, coarser in some broader ranges and more fine in narrower ranges of greater interest for the given case, for example $0^\circ$ to $+800^\circ$ and $+500^\circ$ to $+700^\circ$ C.

Fig. 126.

Fig. 126.

The external appearance of such a compensator, consisting of a resistance and an accumulator, is shown at $1:5$ natural size in Fig. 127;

it completely equalizes fluctuations of the voltage in the network, amounting to as much as 10%.

In the absence of a constant electric current, one may use ordinary galvanic Leclanché cells or so-called dry cells.

In conclusion it should also be mentioned that thermometers with platinum resistance are very conveniently used as self-recording instruments, with temperature measurement by method 2 or 3. As the self-recording instrument an ordinary galvanometer is used, even the same kind as for thermoelements, Figs. 92 or 93; the only difference is in the graduation of the temperature scale.

Fig. 127.

Fig. 127.

The Cambridge Scientific Instrument Company manufactures self-recording instruments for measuring temperature by method 4, the zero method. In this instrument, Fig. 128, when the temperature of the thermometer changes, the galvanometer pointer is deflected to one side or the other and by this deflection closes the current flowing in one of two small electric motors, which moves in the horizontal direction a pointer at the end of which is a pen resting on a drum; one motor is set in motion when the temperature rises, the other when it falls. The drum, covered with graduated paper, moves under the action of a clockwork mechanism

Fig. 128.

Fig. 128.

one revolution in 25 hours; by a simple switching of the transmission mechanism this speed can be increased, and then one revolution will be made in 2 hours 5 min.

The firm indicated also makes such self-recording instruments with a paper tape instead of a drum; the speed of its motion is 10 mm per hour, and its length is calculated for a week-long record.

These instruments are very accurate, sufficiently sensitive, and may be recommended for laboratory scientific work. For technical purposes they are too delicate and expensive.

Production of Reports at a Distance

32. Instruments for measuring temperature at a distance.—Here we shall examine several instruments whose construction is also based on electrical phenomena, but which serve not for measuring temperature itself, but only for transmitting to a greater or lesser distance the reports obtained in one way or another.

In Fig. 129 there is shown schematically a group of mercury thermometers whose readings are transmitted over a distance by means of an annunciator similar to that used with electric bells. The annunciator is brought into action in the following manner: into the glass tube of the thermometer several platinum wires are sealed, one at the bottom, in constant contact with the mercury column, the others, from 3 to 8 in number, depending on the number of temperatures which the instrument is to indicate, higher up—opposite the corresponding divisions. These contacts are connected by wires to a small source of electric current $a$. When the mercury column reaches one of the contacts made opposite the divisions, it will close the current

Fig. 129.

from a, which in one way or another causes the first shield of the indicator to open, showing the temperature reached; as the temperature rises, the following indications open. To check whether the temperature has subsequently fallen, the shields of the indications that have opened must be closed; if the temperature has not fallen, they will all immediately open again; if it has fallen, only those will open whose contacts have remained closed by the mercury.

So that the current is not wasted needlessly, each shield, when opened, automatically interrupts it at once. The disadvantages of this device are the lack of continuity of the indications and, along with this, the large number of conductors and their considerable total length.

This last inconvenience is eliminated in the circuit in Fig. 130, in which, for further transmission, only two conductors are required, no matter how many contacts are sealed into the mercury thermometer. Here the temperature indicator is a galvanometer g, into which a current of greater or lesser strength enters depending on the magnitude of the resistance r_1: as the column of mercury rises, all new contacts are switched in; these, in turn, being connected by conductors e with the separate elements of the resistance r_1, switch in new elements of it; the total resistance thereby decreases, the current increases, and the deflection of the galvanometer needle increases; when the temperature falls, the reverse takes place.

Fig. 130.

Since the current from the element a, of very small voltage, is used, no oxidation occurs at the contacts even with prolonged connection of the instrument.

The switch c serves for checking the readings of the galvanometer: if, when c is turned to the right, the needle g does not come to a definite mark owing to a drop in voltage in the element a, then the corresponding supplementary resistance r_2 is changed.

In Fig. 131 a corresponding thermometer is shown, intended for controlling the temperature of steam in a superheater; e is a cable twisted from separate conductors leading to the parts of the resistance r_1, Fig. 130.

Fig. 131.

Menning’s apparatus is quite peculiar, Fig. 132; the actual temperature meter is a metallic thermometer, or talpotasimeter, which turns the indicating pointer $z$ together with the electrical coil $s$ connected to it; the latter is surrounded on the outside by another, larger coil $S$, through which an intermittent current is passed, from the Ruhmkorff coil $R$ and the element $a$, as a result of which an induced current is excited in the coil $s$, the strength of which depends on the angle of rotation of $s$ relative to $S$.

Fig. 132.

In the main circuit, together with $S$, there is included a coil $S_1$, identical with it in resistance and in number of turns, and in the auxiliary circuit, together with $s$, there is included an identical coil $s_1$, from which an induced current of the opposite direction is obtained, and the telephone receiver $T$. When $s$ and $s_1$ occupy the same positions relative to $S$ and $S_1$, induced currents of equal strength are excited in them and, therefore, mutually cancel one another; nothing is heard in the receiver $T$. When $z$, and together with it $s$, turns somewhat, the currents will be of different strength, and noise will be heard in the receiver $T$, which will disappear when $s_1$ is turned by hand into the same position as $s$; at this, the pointer $z_1$ will stand at the same division as $z$.

As is evident, Menning’s circuit can be used for transmitting over a distance not only the readings of thermometers, but also of other instruments, such as manometers, tachometers, water-level indicators, etc. Unfortunately, the accuracy of measurement according to Menning’s circuit depends to a considerable extent on the sharpness of hearing of the person making the reading.

General practical instructions.

33. Accuracy of various instruments.—When choosing, from among the very diverse instruments described above, the one most suitable for the measurements at hand, one must be guided by two circumstances: the convenience of using the instrument and the accuracy of its readings; for very many measurements the latter circumstance

is decisive. Above, in describing the individual instruments, we have already indicated the accuracy of each instrument; here we shall merely assemble these data into a more visually clear, orderly summary. But before doing so, let us first examine in general what is to be understood by the word accuracy of an instrument, and on what conditions it depends.

The accuracy of an instrument’s readings is characterized by the magnitude of the unavoidable difference in the readings when the same temperature is measured repeatedly, and moreover when the readings are taken by different persons. This accuracy may be called absolute.

In measurements of a changing temperature, especially one changing in a single direction, the accuracy of the quantities obtained in the form of differences between individual readings will be considerably greater than the absolute accuracy of the instrument taken as a whole, since many errors arising, for example, from an incorrect position of the $0^\circ$, from backlash in the mechanism, from the influence of external conditions, and so on, disappear when one reading is subtracted from another. If the successive readings are made by the same person, i.e. the subjective error remains constant, then the accuracy of the temperature changes found may be as much as 10 times greater than the absolute accuracy of the instrument.

When speaking of the accuracy of an instrument’s readings, one must understand, of course, not simply the possible deviation of the directly made reading from the true value of the temperature being measured, but the difference between the true temperature and that found with the aid of the instrument under certain specified conditions.

The principal among these conditions are the following:

the instrument itself, of the selected type, must be made as well and as carefully as possible;
the measurements must be carried out skillfully, i.e. by a person possessing the appropriate experience, with observance of all precautions indicated for the given type of instrument;
the reading taken must be corrected by introducing the corresponding unavoidable corrections indicated above;
the instrument must necessarily be checked from time to time in one way or another, as described in detail below; this checking must be carried out the more often, the higher the temperature being measured and the greater the desired accuracy.

In Table 16, p. 131, a summary is given of the absolute accuracy of readings for various types of instruments. For precision instruments the figures are taken chiefly from investigations carried out at the German Physico-Technical Institute; from there also are taken the figures of the last

of the column—the sensitivity of the instrument, i.e., the smallest difference of temperatures that can still be established with the given instrument. This quantity depends, of course, not so much on the observer as on the instrument, and moreover to a considerable extent on the care taken in its manufacture; in general it is almost proportional to the absolute accuracy of the instrument. The figures for the remaining instruments are taken partly from the most recent periodical literature, and partly determined by the author himself.

For less accurate instruments, used for technical measurements, the magnitude of the possible error, besides °C, is also indicated in % of the measured temperature. Such a definition of the error, strictly speaking, is even more correct, since the error usually increases with temperature, and not even proportionally to the temperature, but faster. In addition, such an indication in % may be useful for estimating the practical suitability of an instrument for the measurement at hand.

Table 16 is arranged in decreasing order of accuracy of the instruments, but only approximately, since a stricter sequence would force instruments similar in construction to be separated too much and, moreover, would hardly even be advisable, since the figures in the table, of course, cannot be regarded as unconditionally exact, but rather only as averages for the corresponding type of instrument, and with improvements in design, often almost imperceptible, they may change greatly.

In addition to the accuracy and sensitivity of the instrument indicated above, in many cases—especially when measuring a rapidly changing temperature—in order to obtain correct readings it is necessary that the instrument follow temperature fluctuations as quickly as possible, i.e., have as small a heat capacity as possible; otherwise its readings will be below the true temperature when it is rising and above it when it is falling.

For comparison of different instruments it is desirable to express this speed of response somehow numerically. For this one may make use of the following circumstance: if a thermometer is heated by some source of heat, for example the flame of a gas burner, to a certain temperature and then, after removing the burner, the time $\tau$ is measured with a stopwatch during which the thermometer falls by half the number of °C by which it had been heated, it turns out that this time, for the given instrument, depends almost not at all on the height of the temperature to which the instrument had been heated; this time $\tau$ depends only on the mass, the cooling surface, and the material of the instrument being tested.

In view of this, the value of $\tau$ found in this way, expressed in seconds, will sufficiently characterize the speed of response

Table 16.

Type of instrument	Measurement range, °C	Absolute accuracy, °C	Absolute accuracy, %	Sensitivity, °C
normal gas, with hydrogen	−250 to +100°	±0.01°	—	0.005°
precise gas, with hydrogen or helium	−260 to 0	0.1	—	0.01
gas, with nitrogen	+100 to +1600	5.0	—	1.0
change in the electrical resistance of platinum	−200 to +1200	0.1	—	0.02
thermoelement, gold and neisilber	−260 to 0	0.1	—	0.02
thermoelement, constantan and silver	0 to +600	0.8	—	0.3
thermoelement, platinum and platinum-rhodium	+300 to +1000	1.0	—	0.5
thermoelement, platinum and platinum-rhodium	+1000 to +1600	5.2	—	1.0
thermoelement, iridium and ruthenium	+1000 to +2000	10	—	—
normal glass, with mercury	−10 to +100	0.01	—	0.005
normal glass, with mercury	+100 to +300	0.1	—	0.05
glass mercury, with gas	+200 to +500–750	1.0	—	0.3
glass mercury, with pentane	−200 to 0	0.1	—	0.02
calorimeters	+300 to +1500	5–10	—	—
Wanner pyrom.-photom.	+625 to +1000	3–5	0.5–1.0	—
Wanner pyrom.-photom.	+900 to +1400	10–20	0.5–1.0	—
Wanner pyrom.-photom.	+1400 to +2000	30–60	2–3	—
Kurlbaum pyrom.-photom.	+600 to +1900	—	1–2	—
Féry photometric pyrometer	+700 to +2000	10	0.5–1	—
Féry thermoelectric pyrom.	+500 to +2000	10	0.5–1	—
Féry thermoelectric pyrom.	+2000 to +3000	50	1–2	—
Féry spiral pyrom.	+500 to +1100	5–10	1	—
Féry spiral pyrom.	+1000 to +1700	10–15	1–2	—
Wiborg gas pyrometer (with mercury)	0 to +1200	5–10	1	—
Holing–Steinbart gas pyrometer	+100 to +1000	5–10	1	—
Holing–Steinbart gas pyrometer	+1000 to +1650	15–25	1.5	—
Mesuré and Nouel optical pyrometer	+900 to +1500	50	3–5	—
Tempel spectral pyrom.	+700 to +1400	50	3–5	—
Tallyptazimeters	−20 to +600	10–20	3–5	—
graphite pyrometer	+350 to +800	20–50	5–10	—
Seger cones	+590 to +2030	20–40	—	—
Sentinelle cylinders	+500 to +1000	5–10	1–2	—
Wiborg thermophone	+300 to +1000	5–15	1–1.5	—
Wiborg thermophone	+1000 to +2400	15–50	1.5–2.0	—

of a given instrument behind the fluctuations of the temperature being measured, of course, only comparatively with other instruments working under similar conditions.

If one turns to existing instruments, it appears that the smallest lag is given by instruments based on the change in the resistance of platinum, and by thermoelements, for which $\tau$ is $10''$—$15''$, sometimes even only $5''$. It goes without saying that this is true for instruments without a casing, and inserted only into a thin-walled quartz or glass tube. The presence of a casing, especially a metallic one, increases $\tau$ to several minutes.

In good mercury thermometers with a small bulb, $\tau$ is from $30''$ to $60''$, while for instruments with a large bulb it is up to $100''$—$120''$ and even more.

For all the other instruments—gas instruments, thalpotasimeters, not to mention graphite pyrometers, for which $\tau$ may reach 10 minutes—$\tau$ is considerably greater, usually more than 1 minute.

34. Choice of the appropriate type of instrument.—In choosing the instrument most suitable for a given particular case, one must take into account a whole series of circumstances. First of all, the degree of necessary accuracy, or, more precisely, the permissible error, may be known; in this respect Table 16 can at once indicate which instruments must be excluded as insufficiently accurate for the given case. The same Table 16 will provide information on the temperature limits for which a given type of instrument may be used.

Next, the reliability of the instrument is of importance; in this respect one may safely say: in all cases where the temperature limits and other conditions permit, one should use a mercury-in-glass thermometer as the most reliable instrument, especially if it is sufficiently often checked and the required corrections are introduced. Next, almost equally reliable are instruments based on thermoelectric currents and changes in electrical resistance, and, for high temperatures, pyrometer-photometers.

Then, for technical measurements, the convenience of using the instrument and the possibility of placing it in comparatively inexperienced hands are often important. In terms of convenience, the possibility of measuring temperature at a distance is of great importance; for example, to monitor from an office the course of a chemical production process or the operation of a steam station. In addition to the instruments indicated in Chapter 32, for the same purpose

all thermoelements and instruments based on the measurement of electrical resistance are suitable; then, for distances up to 50–100 m, Thalpotasimeters and certain gas instruments of the type of the Uehling-Steinbart pyrometer and the pyrometer-micromanometer may be used.

In conclusion, we shall try to give indications as to which instruments should be used, on the basis of the considerations set out above, in the various most widespread cases:

Scientific and, in general, precise investigations—mercury-in-glass thermometers are the most suitable for temperatures from $-10^\circ$ to $+300^\circ$ C.; beyond these limits, thermoelements and instruments based on the measurement of the electrical resistance of platinum are suitable, and in some cases precise gas instruments; for the measurement of high temperatures, optical pyrometers-photometers may be useful, for example for determining the temperature of the filament of an incandescent lamp;
Gas production—for measuring temperatures in retorts it is most convenient to use optical pyrometers-photometers or Féry instruments; for measuring temperatures at various places in a furnace one may use either thermoelements of platinum and platinum-rhodium, or instruments with platinum resistance;
Power stations with steam engines—the determination of temperatures in the firebox of a steam boiler is carried out with the aid of thermoelements of platinum and platinum-rhodium, pyrocalorimeters, Wiborgh thermophones, or optical pyrometers; in flues—with the aid of thermoelements of silver and constantan, instruments with platinum resistance, and mercury-in-glass thermometers; these same last three types of instruments are used to determine the temperature of superheated steam and the temperatures in the condenser and cooling tower;
Power stations with internal-combustion engines—for observing temperatures in the gas generator it is most convenient to use a thermoelement of platinum and platinum-rhodium; the temperature of the combustion products is measured either by means of a thermoelement of silver and constantan, or by an instrument with platinum resistance, or, finally, by a mercury thermometer; with the last it is most convenient to measure the temperature of the water cooling the engine cylinders;
Blast furnaces—the temperature of the preheated air is measured by means of thermoelements of platinum and platinum-rhodium, or nickel and carbon, or instruments with platinum resistance, or, finally, by means of Wiborgh thermophones; the temperature of the bosh gases is most conveniently measured by means of thermoelements of silver and constantan, or by means of the Uehling-Steinbart gas pyrometer;

6, cast-iron and steel-casting work—the temperature of the flowing molten metal is most simply determined by means of one of the optical pyrometers; the temperature of the metal in crucible melting may be determined by means of thermoelements in a special graphite mounting;

7, rolling-mill work—the temperature of blanks is also determined by means of optical instruments;

8, various furnaces for the treatment of steel and cast iron—for hardening steel, for annealing it, for obtaining malleable cast iron—observations are made by means of thermoelements, or instruments with platinum resistance, or instruments of the “Sentinel” type with a definite melting point;

9, various ceramic industries—temperatures in furnaces are measured by means of thermoelements of nickel and carbon or of platinum and platinum-rhodium, or by means of optical instruments, or, finally, by means of Seger cones;

10, glass factories—the temperature in glass-melting furnaces, ranging from 1000°–1400° C., is measured by means of thermoelements of platinum and platinum-rhodium or of nickel and carbon; the temperature during casting, especially of mirrors, is measured by means of optical pyrometers; finally, the temperature in annealing furnaces, in which slow and uniform cooling must take place, should be measured by means of thermoelements of platinum and platinum-iridium or of nickel and carbon, or by means of instruments with platinum resistance, in which case it is useful to have a self-recording galvanometer for better monitoring of the lowering of the temperature;

11, various chemical industries—for proper control of temperatures, in most cases one may recommend instruments for producing readings at a distance—thermoelements made of one or another metal, depending on the magnitude of the temperature being measured, or instruments with platinum resistance; sometimes ordinary mercury thermometers are convenient; in view of the diversity of industries it is difficult to give general rules; thus, in breweries—for measuring the temperature in vats—one may use instruments with electrical resistance; when controlling the temperature of the wort it is useful to use the same instruments, but with self-recording galvanometers; temperatures in cellars, where in more modern plants a low temperature is maintained by means of refrigerating machines, are most conveniently controlled by means of instruments with platinum resistance according to the Kæpsel scheme, or, in view of

constancy of temperature, by means of mercury thermometers with contacts that actuate registering counters;

12, finally, observation of temperatures in large buildings, such as schools, hospitals, museums, etc., is conveniently carried out by means of instruments that produce reports at a distance: instruments with platinum resistance, or mercury ones with registering counters, or galvanometers with transmission according to Mönnich’s scheme.

In a word, at the present time one can hardly point to a case in which the measurement of temperature could not be carried out with the aid of existing instruments, and moreover with sufficient accuracy.

SECTION II.

Verification of Instruments for Measuring Temperature

35. General remarks.—By the word verification of a thermometer one usually means simply the determination of how correct the readings of the scale of the given instrument are. Besides this principal task, in verifying an instrument one may solve a whole series of others: to determine the accuracy of the instrument, absolute and relative, or its sensitivity; then the lag of the readings; the influence on the reading of the character of the temperature change, i.e., whether the reading is taken with rising temperature or with falling temperature, etc. All these and analogous tasks, while of undoubted interest for practical use, are at the same time, on the one hand, solved experimentally very simply, and, on the other, require the same instruments for obtaining the corresponding temperature as in verification in the above-mentioned narrow sense, with which we shall chiefly be concerned.

The methods of verifying thermometers may be divided into two groups: one is the verification of the given instrument by comparing it with an accurate, standard—gas, mercury, or thermoelectric—instrument; the other is direct verification, more precisely calibration, by known constant points—the melting and boiling temperatures of certain precisely studied bodies and the transformation temperatures of known chemical compounds. Such calibration, in essence, is of course nothing other than the same comparison, but not direct; rather, through a body taken with the precise instrument by which the corresponding points were determined in their time. The only instruments that can be verified independently of others, among thermometers, are gas thermometers, which serve as the fundamental, primary instruments in our temperature measurements.

As for calibration by constant points, it, in turn, may be divided into 3 groups: by melting points, by boiling points, and by transformation temperatures. Calibration by melting points is in general less convenient and less accurate than

calibration by boiling points, although the latter method requires the introduction of a correction for the vapor pressure; calibration by transformation temperatures is very convenient, but, unfortunately, at the present time these temperatures have been accurately studied for only a few compounds.

As general basic rules that must be observed in all methods of verification, the following may be indicated:

In making readings on the instrument being verified and on the standard instrument, these readings must be corrected by calculating, in the manner indicated above, all the corresponding corrections for both instruments; still better, if this is at all possible, is to conduct the test so that at least certain corrections—such as, for example, that for the projecting column in a mercury thermometer, or for the temperature of the cold junctions in a thermoelement—are equal to zero;
During verification it is necessary to take several readings at each temperature and to use their mean value; if the difference in the readings is due not to an actual fluctuation of temperature but to inaccuracy in the indications of the instrument—for example, when verifying by constant boiling or melting points—then from the readings obtained and from their mean value one must calculate the magnitude of the possible and probable error of the given instrument at the given temperature;
Having found the true indications of the given instrument for a series of temperatures, one must present them in the form of a diagram, plotting temperatures on the axis of abscissas and the corresponding readings on the axis of ordinates—for example, pressure in a gas pyrometer, the length of the column in a mercury thermometer, electromotive force in a thermoelement, resistance in an instrument with platinum resistance, etc.; by joining the individual points with a smooth curve, one obtains a diagram for intermediate temperatures and, moreover, when drawing the smooth curve it is easy to detect an error that may have crept into one or another reading;
Finally, when verifying instruments it is, of course, necessary to observe all the rules of handling and precautions given for the given instrument, such as not subjecting it to excessively rapid and abrupt temperature fluctuations; for thermoelements, connecting the galvanometer only at the moment of taking readings, etc.

We shall now proceed to the description of the individual methods of verification.

36. Verification by comparison.—For such verification it is necessary to have an appropriate standard instrument for measuring temperatures and a good thermostat, i.e., an apparatus in which one can maintain a strictly constant desired temperature. The most reliable method is comparison of the given instrument with a gas instrument—or, keeping them side by side in a water, steam, or other bath, or, at higher tempera-

inserting the given instrument inside the gas one, for example as shown in Fig. 6, p. 23. Comparison precisely with a standard gas instrument is the most accurate, because, as we saw at the beginning, gas instruments are the fundamental ones on which our entire temperature scale rests. Unfortunately, comparison with gas instruments requires rather expensive and complicated apparatus, and the comparison itself is rather difficult; therefore it is practiced chiefly only in institutions devoted to the verification of various instruments, such as the Main Chamber of Weights and Measures in St. Petersburg, the Imperial Physico-Technical Institute in Charlottenburg, the International Bureau of Weights and Measures in Sèvres, the Conservatoire of Arts and Crafts in Paris, the National Physical Laboratory at Teddington (London), the National Bureau of Standards in Washington (U.S.A.), and others.

A simpler verification is carried out by comparison with a standard mercury thermometer, or with a thermoelement, or in general with an instrument made with particular care, calibrated in one of the above-mentioned official institutions and furnished with a certificate, i.e. with a table of the corresponding corrections.

When comparing instruments, great attention must be paid to ensuring that the sensitive parts of the instruments—the mercury bulbs, the junctions of thermoelements, or the spirals of instruments with platinum resistance—are placed side by side and as close to one another as possible; it is especially important, when the instruments are in a vertical position, that their sensitive parts be strictly at the same height, in view of the possible formation, by height, of layers with different temperatures. Then one must be careful and take measures, by using the appropriate stirrers, so that the temperature of the medium in which the instruments being compared are immersed is as uniform and constant as possible; however, with particularly precise instruments, in order to eliminate the error from dead travel or from the equivalent change in the surface of the mercury column, it is useful to take readings while the temperature is rising gradually, but very slightly, by fractions of a degree over 10–15 minutes.

Depending on the field of application of the instrument being calibrated, one must use different standard instruments and different means for obtaining the required temperature.

Let us begin with the lower limit of temperatures.

To attain temperatures from $0^\circ$ to $-190^\circ$ C., one may recommend the devices developed after many years of experience at the Physico-Technical Institute in Charlottenburg.

The lowest temperatures, down to $-190^\circ$, are obtained with the aid of liquid air; for this purpose the apparatus shown at $1:5$ natural size in Fig. 133 is used: $G_1$ is a Dewar vessel with double walls, silvered inside and outside, and with an evacuated space between the walls; $G_1$ stands on a wooden base $F$, fitted tightly to it and lined with cloth, not shown in Fig. 133, and is closed by a wooden lid $D$ with a felt lining for better insulation; so that the lid, by its weight, does not press on the rather fragile vessel $G_1$, it is held by means of clamping screws on the wooden uprights $A,A$. The inner vessel $G_2$ is also made of glass and also has double silvered walls, but the space between them is filled with down or, in general, with a poor conductor of heat; attempts to make $G_2$ likewise a Dewar vessel with an evacuated space were not crowned with success: despite all precautions, the vessel burst under the action of the low temperature.

Into the vessel $G_2$ benzine is poured, and for the very lowest temperatures petroleum ether, which also serves as the bath for the thermometers being tested; through a special opening in the lid $D$ of $G_1$, liquid air is poured in by means of a cardboard funnel, after which it passes through the lateral openings of the stopper $K$ into the glass tubes $U,U$; the tubes $U,U$, brought out somewhat above the upper edge of $G_2$, are connected by a rubber tube with a glass three-way cock and can—

Figs. 133 and 134.

be separated from the outside atmosphere by means of screw clamps. When communication with the atmosphere is established, the liquid air in $U,U$ evaporates and produces the required lowering of temperature, which in the apparatus under consideration amounts to as much as $2^\circ$ per minute.

To obtain and maintain any low temperature, the vessel $G_2$ is provided with an electric heater, consisting of constantan wire 0.25 mm in diameter; the total resistance of this wire is about 55 ohms; for insulation this wire is wound with silk impregnated with shellac, and is wound on the metal ring $R$, Fig. 135, which at the same time also serves as a stirrer.

Fig. 135.

In view of the low temperature of the medium in which the heater is located, currents of comparatively considerable strength may be passed through its wire without danger of piercing the insulation; however, to maintain a constant temperature the current strength amounts, depending on the temperature level, to no more than several tenths of an ampere. The current is supplied through two thick copper wires, likewise wound with silk impregnated with shellac; these wires are passed through the glass tube $S\text{-}S$, which serves as the holder of the stirrer. The slots $P$ in the ring $R$ embrace the tubes $U,U$, which serve as guides for the stirrer.

Under the mutually opposite influence of the evaporation of air in the tubes $U,U$ and of the current heating the winding $R$, the liquid of the bath rather quickly assumes some constant temperature depending on the current strength; of course, regulation of the current strength must be very sensitive; under this condition the temperature can easily be maintained for 10–15 minutes constant with an accuracy of up to $0.1^\circ$–$0.2^\circ$ C.

In Figs. 133 and 134 one can see the method of actuating the stirrer: its rod $S$ is connected by means of the collar $C$, the connecting rod $P$, and a piece of wire—the clip $Z$—with a constantan ribbon of cross section $10 \times 0.1$ mm; the ends of the ribbon are joined by simply bending one of the ends passed through a slot in the other end. This ribbon embraces two rollers $B$ and $B_1$, of which the latter is set into rotation by means of a cord from a small electric motor. This method of actuation has the advantage over a simple crank mechanism that the stirrer has a uniform speed of raising and lowering, with abrupt changes of speed at the dead points, which ensures very good stirring.

The whole apparatus is extremely convenient; the consumption of liquid air is not great, about 4 liters for 10 hours of operation. Unfortunately, at temperatures below $-150^\circ$ C, petroleum ether, after several hours of operation, under the influence of absorption of atmospheric moisture, becomes turbid and thickens so much that the stirrer stops.

When obtaining temperatures not below $-79^\circ$, solid carbon dioxide may be used as the refrigerant; pieces of it are thrown directly into the liquid of the bath, gasoline or another light kerosene fraction, and thus the desired lowering of temperature is achieved with an accuracy of several degrees; the final setting and maintenance of the required temperature is achieved, as above, by an electric heater.

To obtain temperatures from 0 to $-79^\circ$, one may use the apparatus shown in Fig. 136: $a$, $b$, and $c$ are three concentric cylindrical copper vessels with diameters respectively of 180, 140, and 100 mm; the annular spaces between them are occupied by a coil of solid copper tubing of total length about 14 m, with an internal diameter of about 8 mm and a wall thickness of 1 mm; this tube withstands a pressure of several hundred kg/cm$^2$; the place where it passes at the bottom through the wall of the middle vessel $b$ is soldered; the three vessels $a$, $b$, and $c$ form two cavities: the inner one with the alcohol bath, into which the instruments being tested are directly immersed, and the annular cavity between $a$ and $b$, serving only as a jacket to reduce heating of the alcohol bath by the surrounding air; for the same purpose, the whole apparatus is also wrapped on the outside with several layers of felt. Through the coil flows carbon dioxide, entering through tube $d$ and through valve $f$ from a steel cylinder, in which it is in liquefied form, having a pressure of 50–70 atm., depending on the ambient temperature.

Fig. 136.

When passing through valve $f$, carbonic acid evaporates almost instantaneously and expands adiabatically to the external, atmospheric pressure, with a strong lowering of temperature; valve $f$ serves to regulate the amount of carbonic acid flowing out, and together with it the temperature of the bath. The stirrer $r$, consisting of a four-bladed screw and set in motion by pulley $s$ from a small electric motor, promotes better use of the cooling action of the coil, since it moves the alcohol counter to the flow of carbonic acid in the coil, as is shown by the arrows; at the same time it ensures the greatest possible uniformity of temperature throughout the bath. The cover $g$ of the apparatus is made of ebonite; in it there are 8 conical openings about 19 mm in diameter for the simultaneous testing of 6 instruments, assuming that one standard instrument is placed on each side; the thermometers are inserted with the aid of rubber stoppers $n,n$; for convenience of observation the cover can be turned, for which purpose it is provided with rollers $h,h$. To prevent a long thermometer from being pushed in too deeply and broken by the stirrer, an intermediate bottom $k$ of fine wire mesh is arranged above the latter.

Fig. 137.

It is true that this apparatus does not make it possible to obtain temperatures below $-67^\circ$ C, since with a further lowering of the bath temperature the carbonic acid begins to freeze and clog the narrow opening of valve $f$; on the other hand, of course, the apparatus cannot give as constant a temperature as the one described above, but for many cases, especially for testing alcohol thermometers, which in themselves are not very accurate, it is very convenient.

Finally, for temperatures from $0^\circ$ to $-21^\circ$ C one may use the apparatus shown in Fig. 137: $a$-$b$ is a cylindrical copper vessel about 190 mm in outside diameter and 265 mm high, with double walls for insulation from the surrounding air; to the ebonite cover $g$ is screwed a copper cylinder $c$, open at the bottom and having openings $e,e$ for the circulation of the refrigerating mixture under the action of a four-bladed stirrer ...

of the pulley $r$, driven by an electric motor of $1/20$ h.p.; on the same shaft there is also a four-bladed screw $p$. The cooling mixture is made of chipped ice with the addition, depending on the required temperature, of a known proportion of one or another salt, for example according to Table 17, compiled from the experiments of Rudorff and T. Moritz.

Table 17.

For 100 g of ice, salt is needed	g	temp. °C
potassium sulfate $K_2SO_4$	10	— 1.9
sodium carbonate $Na_2CO_3$ (cryst.)	20	— 2.0
potassium nitrate (saltpeter) $KNO_3$	13	— 2.85
potassium chloride $KCl$	30	— 10.9
ammonium chloride (sal ammoniac) $NH_4Cl$	25	— 15.4
ammonium nitrate $(NH_4)NO_3$	45	— 16.8
sodium nitrate (saltpeter) $NaNO_3$	50	— 17.8
sodium chloride (table salt) $NaCl$	33	— 21.3
calcium chloride $CaCl_2$	150	— 33.0
(alcohol $CH_3CH_2OH$	105	— 30.0)

In order that salt not yet dissolved should not settle, a scraper in the form of two brass strips $d,d$ is attached to the stirrer shaft near the very bottom of the vessel. The lid, in which there are 9 openings for the simultaneous comparison of up to 7 thermometers with 2 standard instruments, for ease of observation turns on 3 rollers $h,h$; to make it possible to check thick meteorological thermometers, the slightly conical openings into which the thermometers are inserted, passed through rubber stoppers $n$, are about 24 mm in diameter. For insulation the apparatus is wrapped on the outside in several layers of felt, which is made waterproof so that accidental splashes, unavoidable during prolonged operation, do not wet it and so that it would not lose its properties as a poor conductor of heat. To make the felt waterproof, it must be soaked in a solution of alum, then immersed in a solution of lead acetate, dried in warmth, and then beaten out.

Although the air layer between the walls $a$ and $b$ also provides good insulation, it is nevertheless useful to pump out this air, reducing the pressure to only 250 mm of mercury, in order to lessen the harmful circulation of the air.

The apparatus described is very convenient in operation; the temperature is kept very constant; with proper action of the stirrer its fluctuations may be less than $0.1^\circ$, even $0.05^\circ$ C.

In extreme cases, the testing of thermometers from $-20^\circ$ to $0^\circ$ C. may be carried out without any apparatus at all, simply by placing them in a mug with a cooling mixture of ice and the corresponding quantity of table salt NaCl or ammonium thiocyanate $(\mathrm{NH}_4)\mathrm{SCN}$.

Such a test requires special attention and even skill; it is especially important to make sure that the mercury bulbs are completely immersed in the liquid and are not in the air between pieces of the mixture.

Instead of taking different salts in order to attain different temperatures, one may use some one salt, for example sodium chloride or calcium chloride, adding different quantities of the salt to the ice. The approximate dependence between an increase in the salt content and a lowering of the temperature of the mixture is given in Table 18, according to Hammerl’s experiments on mixtures of crystalline calcium chloride and ice and according to Tollinger’s experiments on mixtures of ammonium nitrate; in the 3rd and, correspondingly, 6th columns is shown the number of heat units which 1 kg of the corresponding mixture can take from the surrounding medium or from a body immersed in this mixture.

Table 18.

$(\mathrm{CaCl}_2 + 6\mathrm{H}_2\mathrm{O})$ per 100 g of ice, g	Temperature, °C	1 kg of mixture removes, heat units	$(\mathrm{NH}_4)\mathrm{NO}_3$ per 100 g of ice, g	Temperature, °C	1 kg of mixture removes, heat units
20	−4.0	66.2	13	−4.0	76.0
37	−8.1	57.3	28	−8.0	71.0
41	−9.0	55.6	36	−10.0	68.3
53	−12.4	50.1	45	−12.0	65.6
61	−14.7	47.1	56	−14.0	62.2
72	−18.3	42.9	67	−16.0	58.4
81	−21.5	39.7	76	−17.0	57.1
84	−22.7	38.4
97	−28.0	35.2
103	−30.4	32.3

When comparing temperatures from $0^\circ$ to $+60^\circ$ C., one may use a water bath heated by a gas burner; the vessel for the bath may be a cylindrical tin one, well insulated from the sides, for example wrapped in felt and covered on the outside with asbestos cardboard. The diameter of the vessel is about 100–150 mm, its height 150–300 mm, depending on the length of the thermometers being compared and on the temperature up to which the verification is carried out; namely, the thermometers must be immersed in the water up to the top of the mercury column.

The comparison itself should be carried out as follows: heat the water in the vessel approximately to the desired temperature, move the burner aside and, while thoroughly stirring the water, immerse in it the instruments being compared; since it is impossible to avoid currents of different temperatures even with vigorous stirring, the mercury bulbs, or in general the sensitive parts of the instruments—both of the standard instrument and of the one being compared with it—must be kept as close as possible to one another and strictly at the same level. Readings should be taken when the thermometers fully assume the temperature of the bath, which requires 15–60 seconds, depending on the instrument. Then heat the bath to the next temperature, take the reading again, and so on. Instead of heating the water with a gas burner, one may simply gradually add a certain quantity of hot water; in this case it is only necessary to stir the water especially carefully and to insulate the vessel better also from below, for which purpose it may simply be placed on a piece of felt. The verification is carried out, depending on the accuracy of the instrument and the forthcoming measurements, every $5$–$10^\circ$. This method is applicable only up to $60^\circ$ C., since beyond this strong evaporation begins.

For the simultaneous comparison of a large number of thermometers, and, chiefly, in order to attain the greater constancy and uniformity of temperature that are necessary when verifying especially precise instruments—for example, for calorimeters—at the German Physico-Technical Institute an apparatus according to Fig. 138, p. 146 is used. The apparatus, shown approximately at $1:10$ of full size, consists of a cylindrical copper vessel $a$ with double walls in order to reduce heat loss by radiation; in addition, the vessel $a$ is, of course, wrapped in several layers of felt. This same annular space also serves for rapidly raising the temperature of the bath; for this purpose, not far from the bottom, there is an annular tube with openings into which steam is admitted from a special small boiler; the inflow of steam is regulated by the cock $c$; the condensate is discharged by means of the cock $d$.

The upper hexagonal glass part of vessel $a$, with posts $b,b$, makes it possible to take readings of thermometers with the mercury column completely immersed. Vessel $a$ is best filled with distilled water, which remains transparent for several weeks, whereas tap water quickly becomes turbid, which reduces the accuracy of the readings. To the upper part of the apparatus—a Y-shaped crosspiece $q$, resting on three columns—is attached a brass tube $e$, 55 mm in diameter and reaching almost to the bottom of vessel $a$; the lower part of this tube, for heating the water, is wound with a spiral $n$ of insulated constantan wire; the length of the wire is about 30 m, its thickness 0.8 mm, which corresponds to a resistance of about 16 ohms, owing to which, when a current of 110 volts is switched on directly, the temperature of the water is maintained at a constant level of about $85^\circ$ C.; for regulation and reduction of the temperature, rheostats $w$ serve—one for coarser and the other for precise regulation; to obtain temperatures above $85^\circ$—up to $95^\circ$—steam is admitted into the space between the double walls of vessel $a$, as when warming up the apparatus. Spiral $n$ is surrounded by a second brass tube; the space between the two tubes does not communicate with the water bath, but is filled with mineral oil for better insulation of the wire and for increasing heat transfer; the leads to $n$ are brought upward through special branch tubes and glass tubes.

Fig. 133.

To achieve a uniform temperature, a three-bladed screw stirrer $r$ is used, driven by pulley $k$ from an electro-

of a motor of about $1/20$ h.p.; the water is driven by the stirrer $r$ downward through the tube $e$, emerges from it through slits near the bottom of the vessel, then rises, and in doing so is mixed once more by the stationary inclined blades $s,s$. To prevent heating of the rising water, which acts directly on the thermometers, the spiral $n$ in its upper half is surrounded by a second, wider tube, open at the top and bottom, along which the water rises without touching the thermometers. For switching the current into the electric motor for the stirrer and for setting one of its three speeds, the switch $t$ is provided. The upper cover, into which the thermometers to be tested—up to 12 in number—are inserted by means of clamps, can, for convenience in taking readings, be turned by hand by means of the flywheel $m$ and a series of transmissions—cord conical $l$ and cylindrical $i$.

This apparatus satisfies the strictest requirements with respect to temperature regulation and its constancy. However, when testing mercury thermometers, as already mentioned, it is better to make the readings, for the sake of constancy of the mercury meniscus, at a slowly rising temperature; the electric heating device of this apparatus makes it possible to obtain a uniform rise of temperature of less than $0.1^\circ$ in 50–60 minutes, as required for testing one set of thermometers.

For comparing thermometers at higher temperatures, approximately from $+50^\circ$ to $+200^\circ$ C, the water bath has to be replaced by an oil bath. A simple, yet quite convenient and sufficiently accurate apparatus for most practical cases is shown semi-schematically, approximately at $1:10$ of natural size, in Fig. 139: in a cylindrical metal vessel $a$, for example an enameled saucepan, about 300 mm in diameter and 300–350 mm high, standing on a layer of sand about 50 mm thick, is placed a second similar vessel $b$, of somewhat smaller dimensions; into vessel $b$ mineral oil is poured, continuously stirred by the crank stirrer $r$, of the kind described above, set in motion by an electric motor by means of the cord pulley $k$, and surrounded by a brass tube $d$ with openings at the ends; to the same tube $d$ are attached the bearings of the stirrer shaft. In the cover of the apparatus—metallic,

Fig. 139.

covered, to reduce heat loss, with asbestos cardboard, has 6 openings with sleeves into which the instruments being tested are inserted by means of plugs rolled from asbestos cardboard moistened with water. To reduce heat loss the entire instrument is covered on top with a cap $c$ made of asbestos cardboard, further covered with thin sheet iron. The oil bath is heated by means of 1–3 gas burners, depending on the required temperature. Readings are taken every $30^\circ$–$50^\circ$; readings are made when, with the burner flame suitably regulated, the bath temperature has become established and has remained constant for 5–8 minutes.

If, instead of ordinary machine oil, one takes cylinder oil, or so-called viscosine, used for lubricating the cylinders of internal-combustion engines and machines working with superheated steam, then the bath temperature can be brought up to $250^\circ$–$300^\circ\text{C}$.

For still higher temperatures the oil bath must be replaced by a nitrate bath: sodium, or ordinary, nitrate $\mathrm{NaNO_3}$ melts at about $+313^\circ\text{C}$ and begins to evaporate noticeably at about $600^\circ\text{C}$; potassium nitrate $\mathrm{KNO_3}$ melts at $+329^\circ\text{C}$ and is also applicable up to $600^\circ\text{C}$. It is most convenient to take a mixture of these two substances in the ratio of their molecular weights, i.e., for 100 g of $\mathrm{KNO_3}$, about 84 g of $\mathrm{NaNO_3}$; such a mixture becomes liquid at a temperature of about $230^\circ$ and begins to evaporate noticeably only at $600^\circ\text{C}$.

Fig. 140

Fig. 140.

The vessel $a$ for such a bath, Fig. 140, must be very strong, in view of the great increase in the volume of the nitrate upon heating; it may be made either from a vessel in which mercury is sold, or from a piece of seamless steel pipe 125–150 mm in internal diameter and 350–400 mm long; the vessel must have a tightly fitted cover with 5 openings: 4 for the thermometers and a central one, about 60 mm in diameter, for the tube $d$ with a stirrer, as in the preceding apparatus.

Into the openings for the thermometers are screwed tubes open at the bottom, reaching almost to the bottom of the vessel. This achieves more uniform heating of the thermometers. Near the upper cover of the vessel, into the wall of the latter, a downward-bent drain tube $e$ is screwed in for removing excess nitrate. On top, to reduce heat loss, the apparatus is covered with a cap $c$ made of asbestos cardboard, covered with.sulake

thin white tin; an opening is made in the cover, from which a tin tube goes off for the removal of combustion products and saltpeter vapors. Heating is carried out by means of 2–3 powerful gas burners.

The thermometers must be verified in exactly the same way as with an oil bath.

At the German Physico-Technical Institute they formerly used an oil bath, shown approximately at 1:5 natural size in Fig. 141. Vessels $A$ and $B$ are made of copper for better heat transfer; heating was carried out by means of gas burners, through the medium of air circulating in the space between the walls of vessels $A$ and $B$; although such heating, despite the use of red-copper vessels, is uneconomical and slow, nevertheless

Fig. 141.

it gives a very constant temperature, which is also greatly aided by the use of two stirrers.

According to the same scheme, an apparatus was made for a saltpeter bath. At present, in the Imperial Physico-Technical Institute, for temperatures from $+50^\circ$, or rather, even from $+100^\circ$ to $+500^\circ$ C., use is made of the apparatus shown approximately at $1:10$ natural size in Fig. 142: $a$ is a cylindrical chemical beaker of Jena glass, with a capacity of about 4 liters, 28 cm high and with walls 1–2 mm thick. The beaker $a$ is wrapped with asbestos cord and inserted into a copper vessel $d$, with walls 1.5 mm thick. Vessel $d$ is fastened to the copper ring $f$, which, by means of 3 rollers $h$, rests on the copper ring $c$, closing the space formed by the double walls of the jacket $b$, which is also wrapped with asbestos cardboard. The jacket $b$ serves to reduce heat loss from the bath; at higher temperatures the jacket has to be heated from below with gas burners, letting the hot combustion products into the space between the walls $b$; the inlet and outlet openings with a pipe necessary for this are not shown in Fig. 142.

Fig. 142.

The beaker $a$ is closed with a light aluminum lid with a gasket of asbestos cardboard; in the lid there are 10 conical openings for the thermometers being tested; the lid rests on 3 iron posts $k$ and in turn supports a copper tube, about 40 mm in internal diameter, inside which 2 screw stirrers rotate in the usual manner; the length of the stirrer shaft is about 600 mm, so that even the longest thermometers do not reach the crossbar $q$.

The stirrers are rotated so that the liquid descends through the tube and rises along the glass walls; such motion is important not so much for the sake of uniformity of temperature as for avoiding waves on the surface, which interfere with readings or cause the mercury columns of the thermometers being checked to protrude.

Tube $g$ is connected either to an exhaust or to a cooler for condensing the vapors of the bath liquid at higher temperatures.

The liquid in the bath is heated by means of a spiral $e$ of constantan wire, wound on a clay tube with helical grooves, surrounding the tube with the stirrers. The thickness of the wire is 0.75 mm, its length about 24 m, which gives a resistance of about 22 ohms; in order to avoid the possibility of a short circuit, the ends of the spiral are led to the clamps $i,i$ through glass tubes. When this heater spiral is connected to a storage battery with a voltage of 110 volts, without any additional resistance, a current of about 5 amperes passes through the spiral, which gives a rise in the temperature of the liquid of about $5^\circ$ per minute; at higher temperatures, somewhat less. To obtain a definite temperature the current must, of course, be passed through a special adjustable resistance; the apparatus responds to a change in current strength almost instantaneously; by using the adjustable resistance, and after the required temperature has been reached by switching the spiral off and on from time to time, the temperature can be kept very constant: at $+150.0^\circ$ the fluctuations over 10 minutes, with skillful handling, do not exceed $0.01^\circ$. Since the wire of the spiral is not insulated, the bath must be filled with a non-conducting liquid: for temperatures up to $200^\circ$—palmin, an artificial edible fat, melting at $+30^\circ$C, producing little vapor and remaining transparent up to $+200^\circ$C; for temperatures up to $+300^\circ$C, olive oil may also be used, which is very convenient but only more expensive; from $300^\circ$ to $550^\circ$C the apparatus may be used as an air bath; the stirrers, whose bearings become hot, must in that case be lubricated with graphite, and even here they give a very uniform temperature.

Readings in this apparatus, as in other precise ones, are taken either through a sighting tube or at least through a magnifying glass.

Recently, in the literature, indications may be found concerning the use of the electric Heraeus furnace for checking thermometers. Indeed, such a furnace, consisting of a porcelain-

tube, wrapped with platinum foil through which a current passes, makes it possible very simply and quickly both to obtain a known temperature and to maintain it fairly constant; but the trouble is that the temperature is distributed very unevenly along the tube—in the middle it is higher, decreasing very sharply toward the ends; the difference may reach $100^\circ$ C. Therefore such a furnace is suitable only for instruments with a very small sensitive part, for example, for the calibration of thermoelements, where it will be discussed in more detail.

For calibration by comparison with other instruments, chiefly mercury quartz thermometers with divisions up to $+750^\circ$ C., and instruments with platinum resistance, there has hitherto been no suitable apparatus. In the author’s opinion, for this purpose one could use the above-mentioned furnace, but take it with as wide a tube as possible, not less than 80 or even 100 mm in diameter, hermetically closed at both ends, and insert a tube with a stirrer-fan, of the kind indicated in Fig. 142. With the furnace in a vertical position and with the fan properly operating, such an apparatus should give a very uniform temperature, easily and accurately regulated within the range approximately from $+200^\circ$ to $+1000^\circ$, or even $1200^\circ$ C.

Fig. 143.

In conclusion, one may also mention the thermostat that was formerly used at the German Physico-Technical Institute, Fig. 143. At the present time, with the introduction of the above-described thermostats with electric heating, it has been abandoned, although it possesses a whole series of valuable qualities: simplicity of maintenance, constancy of temperature, and a comparatively uncomplicated and

inexpensive construction. It is based on two phenomena: on the constancy of the temperature of saturated vapor at a definite pressure, and on the dependence of the temperature of vapors on pressure.

The apparatus, shown in Fig. 143 at approximately 1:10 natural size and made of thick sheet nickel, consists of a boiler $a$, in which a certain liquid boils, heated by a gas burner; of a cylindrical vapor jacket $b$ with double walls, the annular space between which is heated by the descending vapor and protects from cooling the inner cavity, into which the thermometers are lowered; of a condenser $c$, cooled by running water, which condenses the remaining vapors, which flow back into the boiler. To vary the vapor pressure and the temperature associated with it, the apparatus is connected by tube $g$ with the cylindrical vessel $f$, in which a rarefaction of the air can be obtained or, conversely, a certain excess pressure by means of the appropriate pumps, communicating with $f$ by means of tube $l$; tube $m$ leads to a mercury manometer; the lid of vessel $f$, made of corrugated iron, makes it possible, by disconnecting $f$ from the pump by means of cock $h$, to produce small changes of pressure by means of screw $c$, with handwheel $p$. In order that the pressure should not fluctuate as a result of changes of temperature in vessel $f$, the latter is surrounded by a jacket $G$, through which water of constant temperature is passed; in addition, an asbestos-cardboard screen is placed between $f$ and the thermostat. Particular attention in the apparatus is paid to the tightness of all connections; the lid of the vapor jacket $b$ is screwed down on a rubber gasket; the instruments being tested are inserted in the following manner: on the thermometer, or on the glass tube soldered below for the thermoelement, a ring of soft rubber is put, which is compressed between the lid and a special clamping flange; the sockets for these rings, in order to obtain pressure also on the thermometers, as may be seen in the drawing, are conical.

The pressure in the apparatus is maintained very constant; its fluctuations are not more than $0.5$—$1.0$ mm of mercury column, which, with the liquids used, corresponds to a change in temperature of $0.01^\circ$—$0.02^\circ$. The temperature over the entire height of the vapor jacket is completely uniform.

Table 19 shows the temperatures that are obtained in the apparatus described at the corresponding pressures in mm of mercury column, read on the manometer in this apparatus and reduced to $0^\circ$ C., when using the following liquids: methyl alcohol, distilled water, and amyl acetate. Since

the question of the dependence between pressure and temperature for water will be discussed further below; thus, in Table 19 for water only the temperature limits that can be obtained in the apparatus described are indicated. The data of Table 19 were obtained by V. Pomplun at the German Physico-Technical Institute.

Table 19.

working liquid	pressure, mm	°C	working liquid	pressure, mm	°C
methyl alcohol $\mathrm{CH_3OH}$	387.4	+48.7	distilled water from	390	82.8
methyl alcohol $\mathrm{CH_3OH}$	410.2	50.0	distilled water to	794	101.5
methyl alcohol $\mathrm{CH_3OH}$	446.8	52.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	398.1	118.5
methyl alcohol $\mathrm{CH_3OH}$	486.5	54.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	408.5	120.0
methyl alcohol $\mathrm{CH_3OH}$	529.4	56.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	486.0	125.5
methyl alcohol $\mathrm{CH_3OH}$	575.2	58.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	552.3	129.7
methyl alcohol $\mathrm{CH_3OH}$	626.6	60.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	640.6	134.0
methyl alcohol $\mathrm{CH_3OH}$	680.0	62.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	748.6	140.3
methyl alcohol $\mathrm{CH_3OH}$	732.0	64.0	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	768.4	141.2
methyl alcohol $\mathrm{CH_3OH}$	758.2	64.9	amyl-acetic ester $\mathrm{C_5H_{11}OC_2H_3O}$	797.2	142.6
methyl alcohol $\mathrm{CH_3OH}$	794.3	66.1

In Table 19 data are given for only three liquids; there is no difficulty in selecting liquids also for the intervals from 66°–83° and from 101° to 118°; in this way the apparatus can serve within the limits from +48° to +142° C.

In Fig. 144 is shown a thermostat, likewise constructed by the German Physico-Technical Institute, but intended chiefly for liquids boiling above +100° C. Its features are as follows: in view of the easier condensation of the vapors of such liquids, the lower part of the annular space is connected with the tube $f$, leading to the condenser $c$, by the branch pipe $d$ for vapors and by the downward-inclined branch pipe $e$ for the outflow of liquid; by this means continuous circulation of vapors and uniform boiling of the liquid, without jolts and explosions, are achieved; to attain the latter aim it is also useful, during the first heating of the apparatus, to set the flame of the burner in the proper way, to remove the first, more volatile particles or decomposition products by releasing them into a separate condenser attached to the tube $g$, and only after this to admit them

to the refrigerator with water and begin work. To make better use of the heat of combustion, the apparatus is provided with a casing $b$ of roofing iron, covered with asbestos board.

37. Checking by melting points.—At the present time the melting points of a whole series of bodies have been precisely determined, such as tin $\mathrm{Sn} +232^\circ$, lead $\mathrm{Pb} +327^\circ$, zinc $\mathrm{Zn} +418^\circ$, etc.; the melting point of water, i.e. the melting of ice, as is known, is adopted as the initial point of the centigrade scale and is considered $0^\circ$ C.

The indicated melting points, strictly speaking, depend on the pressure at which the transition of a body from the solid to the liquid state takes place; namely, the melting temperature increases with increasing pressure. However, within the limits of barometric, i.e. atmospheric, pressure fluctuations, the changes in melting temperature are very small, less than the degree of accuracy with which the melting points have so far been determined, so that their changes may be entirely neglected.

Drawing 144.

Drawing 144.

Of all melting points, the most important, of course, is the melting point of ice. The verification of this point, i.e. $0^\circ$ C., is carried out very simply in the apparatus according to Drawing 145: it consists of a copper or zinc cylindrical vessel $a$, 120–150 mm in diameter, 200–300 mm high, open at the top, with a conical bottom in which there is an opening and a tube $b$ for draining the water that forms, since otherwise, as water accumulates, its temperature will begin to rise owing to absorption of heat from the surrounding medium. However, only the excess water need be drained off, so that the sensitive part of the instrument does not come into contact with air. To slow the melting of the ice, which is especially important when checking a large number of instruments,

Drawing 145.

Drawing 145.

the vessel may be wrapped in felt, or at least in newspaper, in several layers; sometimes the vessel is even made with double walls, i.e., with air insulation.

For the accuracy of the calibration it is important that the ice be perfectly pure, i.e., obtained from distilled water; snow is not suitable for this purpose, since it usually contains impurities that change the melting temperature. For better contact of the ice particles with the instruments being calibrated and to avoid larger air spaces, the apparatus should be filled with pieces of ice as small as possible, or, still better, with ice shaved with a knife or with a device used for shredding cabbage.

The use of the melting points of other bodies, with a higher temperature, is very convenient for checking instruments with a small volume of the sensitive part, such as thermoelements and, in part, the newer instruments with platinum resistance, about which more will be said below; for the majority of instruments, however, such as mercury-in-glass thermometers, Tallqvist meters, calorimeters, etc., such a check is hindered by a whole series of circumstances: the difficulty, or rather the impossibility, of obtaining a uniform temperature in a vessel of the required, fairly considerable dimensions; the difficulty of catching the brief moment at which the rise in temperature stops in instruments whose indications do not follow the change in temperature instantaneously; in particular, for glass mercury instruments—the danger of crushing the instrument, especially the little vessel with mercury; the necessity of introducing a correction for the emergent column of mercury, etc.

We mentioned above that the presence of impurities affects the melting temperature of ice. The American chemists T. Richards and F. Jackson have recently proposed making use of this property for checking the most precise scientific thermometers with divisions down to $0.001^\circ$ in the range from $-4^\circ$ to $0^\circ$ C. Their method is based not on measuring the actual melting point of ice, but the equivalent freezing point of water when a certain amount of hydrochloric acid HCl is added to it. In the following table 20, p. 157, the dependence is indicated between the concentration of the solution $c$ and the lowering of the freezing point of water $\Delta t$; the concentration of the solution $c$ is given as normal, i.e., by the number of gram-molecules per liter, the volume being taken as the actual one, corrected for the compression of the glass at the temperature of the bath; for example, for a lowering by $2.5^\circ$, $c = 0.663$, and since the molecular weight of HCl is $36.46$, one must take a solution 1 liter of which at $-2.5^\circ$ C contains

\[ 36.46 \cdot 0.663 = 24.173 \text{ g HCl}. \]

In the last column of table 20 is indi—

...the quantity of silver chloride in grams that is precipitated from 10 cm³ of the cold solution is indicated; this column serves for preparing and checking, by titration, the required solution.

The actual checking of the thermometer is carried out as follows: the Dewar beaker mentioned above, with double silvered walls and of sufficiently large dimensions, is filled with a solution of hydrochloric acid in water according to Table 20; the beaker is placed in a refrigerating mixture at a temperature slightly below the freezing temperature of the solution taken—for example, of ice and a correspondingly stronger solution of the same hydrochloric acid.

Table 20.

depression of the freezing point $\Delta t$ °C	normality of the solution $c$	wt. of silver chloride per 10 cm³ of solution, g
0.500	0.139	0.199
1.000	0.277	0.397
1.500	0.409	0.587
2.000	0.538	0.772
2.500	0.663	0.951
3.000	0.773	1.109
3.500	0.886	1.271
4.000	0.997	1.430

The Dewar beaker is covered with a lid of asbestos cardboard with an opening through which the thermometer being checked is inserted; it is useful first to cool the latter to $0^\circ$C in a vessel with melting ice. At first, under the influence of the refrigerating mixture, the temperature of the solution in the beaker will fall, but when the freezing point of the given solution is reached, the temperature will stop and remain at one level until the whole solution has frozen. Of course, in order to obtain a uniform temperature, the solution in the beaker must be stirred all the time with a special stirrer-spinner. After some time, when the temperature of the instrument has fully settled, a reading is taken, which is compared with Table 20 or with a curve constructed from this table.

Readings must, of course, be made at a distance through a telescope, and in general all the precautions customary in precise thermometry must be taken.

Instead of a Dewar vessel one may also take an ordinary chemical glass beaker, but placed inside a second similar beaker of somewhat larger dimensions, in order to obtain air insulation.

The indicated method of checking by means of previously prepared solutions is especially convenient when many instruments must be checked frequently. In the case of an infrequent check, or especially,

if it is necessary to check some temperature lying between the numbers of Table 20, it is better to proceed in the reverse way: the Dewar vessel is placed in the appropriate cooling mixture, while a mixture of ice and water is put into the vessel itself and hydrochloric acid is gradually added until the temperature indicated by the thermometer falls to the desired degree and remains at it, which will show that equilibrium has been reached and the freezing process is taking place. Then, with a chilled pipette, a certain quantity of the cold solution is taken and its normality $c$ is determined by titration with silver nitrate. From the $c$ found and the normal curve constructed with the aid of Table 20, the corresponding temperature is found, against which the readings of the thermometer being checked are compared.

Of course, the method described is somewhat laborious and requires skill in performing the titration—and, moreover, a very accurate one—but it provides a reliable method for calibrating the most precise thermometers, and does so without the aid of complicated instruments or a normal thermometer.

38. Calibration by the temperature of a chemical transformation.—As is known, the temperature of transformation is the temperature at which a given chemical reaction begins to proceed; so long as the temperature of the bodies participating in the reaction is below the temperature of transformation, the reaction does not begin, no matter how long the bodies remain in contact. If the given reaction is endothermic, i.e., proceeds with absorption of heat, then, having begun upon attainment of the transformation temperature, it will proceed only so long as heat is supplied to it; in this case it is not difficult to supply heat in such a quantity that all of it goes only to maintaining the reaction and not to heating the bodies; in such a case the temperature of the bodies will remain constant the whole time, equal to the temperature of transformation. The latest investigations have shown that the temperature of transformation is a quite definite quantity, depending only on the chemical elements participating in the reaction. Thus, it has been possible to find with very great precision the transformation temperatures of many crystalline salts, which are especially easy to obtain chemically pure by repeated crystallization.

Of all the salts investigated, sodium bromide, $\mathrm{NaBr}$, is of special interest; its transformation temperature was determined with extraordinary precision by T. Richards and R. Wells and found to be equal to $50.674^\circ\mathrm{C}$ on the hydrogen scale, so that this temperature lies exactly midway between the fundamental points, $0^\circ$ and $100^\circ\mathrm{C}$, of mercury thermometers, which are the most widely used in scientific work.

In the present case the reaction consists in the conversion of the dihydrate $\mathrm{NaBr,2H_2O}$ into the anhydrous salt $\mathrm{NaBr}$ and its saturated solution in water; the reaction may be represented approximately, in units by weight, by the following equation

\[ 103\mathrm{NaBr},36\mathrm{H_2O}=61\mathrm{NaBr}+42\mathrm{NaBr},36\mathrm{H_2O}. \]

This equation is based on the observation that a saturated solution of sodium bromide contains, at the transformation temperature, about 116.8 parts of anhydrous salt per 100 parts of water; in other words, in two molecules of water (36 g) liberated from one molecule of hydrate, almost exactly 42 g of anhydrous salt are dissolved.

The thermometer may be tested in the apparatus shown in Fig. 146, approximately at $1:5$ of its natural size: a glass test tube $a$, for which another similar test tube $b$, attached to a cork ring, forms an air insulating layer to slow the reaction, is filled to somewhat less than half its height with a mixture of crystals of sodium bromide and water, cooled almost to $0^\circ$, carefully purified and boiled, with pieces of ice made from the same distilled water. Into this mixture is inserted the mercury bulb of the thermometer being tested, which, in order to avoid the correction $t$ for the projecting column of mercury, is usefully surrounded by a jacket formed by the glass tube $d$, through which water is passed, heated in the thermostat $g$ to such a temperature that in the jacket it has on average about $50.0^\circ$ C.; for this purpose, in the tube $d$, at the middle of its height, a small thermometer, not shown in Fig. 146, is suspended; the water enters through the glass tube $e$, goes upward, and flows down through the rubber tube $f$. To maintain a uniform course of the reaction and the same temperature throughout the reacting mixture,

Fig. 146.

the latter during the experiment must be stirred; for this purpose there is a stirrer $r$—a ring of platinum wire inserted into a glass handle $s$. The whole apparatus stands in a vessel $c$, filled with water heated by an electric current by means of a constantan-wire spiral not shown in the drawing. The temperature of the water must be somewhat higher than the transformation temperature, by $2^\circ$—$3^\circ$, depending on the amount of heat lost by the upper part of the apparatus. With the indicated small temperature difference and the presence of an air layer at the test tube $a$, the reaction proceeds very slowly; the temperature remains constant, equal to the transformation temperature, for more than an hour, and, with a well-chosen temperature of the bath, even up to two hours.

In order that the method described should in fact give accurate results, it is very important that the mixture be composed of substances as chemically pure as possible. It is especially difficult to obtain chemically pure sodium bromide. Even repeated recrystallization of the commercial salt gives a substance whose transformation temperature is up to $0.5^\circ$ below the true value. Pure sodium bromide can be obtained only from pure bromine and pure sodium carbonate $\mathrm{Na_2CO_3}$; a description of its preparation lies outside the scope of our task. The size of the crystals, or the degree of their pulverization in the mixture, apparently has no influence on the transformation temperature, provided only that the mixture is stirred properly at all times.

In conclusion it may also be mentioned that the same investigators determined with great accuracy the transformation temperature of sodium sulfate $\mathrm{Na_2SO_4}$; it is equal to $+32.383^\circ\mathrm{C}$. This substance can be used in the same apparatus as in Fig. 146. Obtaining pure sodium sulfate is easier than obtaining pure sodium bromide; moreover, because of the greater closeness of the transformation temperature to room temperature, it is easier to carry out the reaction and to maintain the temperature of the mercury column, but the point $+32^\circ$ itself is of less interest for calibration.

39. Calibration by boiling points.—Bodies whose boiling temperatures can be used for calibrating thermometers must satisfy the following conditions:

the boiling phenomenon of the given body must have been precisely studied, i.e., not only must its boiling point at normal barometric pressure be known, but the dependence between the boiling temperature and the vapor pressure must also have been determined;
the given body must be readily obtainable, as far as possible, in a chemically pure form, since the boiling temperature depends on the

of the purity of the liquid taken, and often even the slightest impurity greatly changes this temperature;

3, the given body must not change, e.g., decompose, under the action of the boiling temperature, since such a change affects the temperature of the vapors;

4, the vapors of the body taken must not have a destructive action on the ordinary materials from which the corresponding thermostats are made, and must not be poisonous, i.e., harmful to the health of the person making the observations.

On the basis of these requirements the following Table 21 has been compiled, in which the bodies most suitable for the purpose in question are indicated.

Table 21.

name	composition	$t$ °C	$\Delta t$ °C	$\varphi$ °C
methyl, or wood, alcohol	$\mathrm{CH_3OH}$	$+64_{,06}$	$0_{,034}$	$\pm 0_{,05}$
ethyl, or wine, alcohol	$\mathrm{C_2H_5OH}$	$78_{,4}$	$0_{,036}$	$0_{,1}$
water	$\mathrm{H_2O}$	$100_{,0}$	$0_{,037}$	$0$
amyl acetic ether	$\mathrm{C_5H_{11}OC_2H_3O}$	$140_{,53}$	$0_{,047}$	$0_{,1}$
aniline	$\mathrm{CH_5NH_2}$	$184_{,3}$	$0_{,054}$	$0_{,1}$
naphthalene	$\mathrm{C_{10}H_8}$	$218$	$0_{,060}$	$0_{,1}$
mercury	$\mathrm{Hg}$	$357_{,2}$	$0_{,075}$	$0_{,1}$
sulfur	$\mathrm{S}$	$445$	$0_{,082}$	$0_{,5}$
cadmium	$\mathrm{Cd}$	$765$	$0_{,12}$	$2$
zinc	$\mathrm{Zn}$	$930$	$0_{,15}$	$5$

For explanation of Table 21 it should be noted that it gives the boiling points $t^0$ at normal barometric pressure of 760 mm of mercury; the third column gives the approximate change of boiling temperature $\Delta t$ when the pressure changes by 1 mm of mercury; with the aid of the quantity $\Delta t$, the temperature $t'$ at any pressure $B$ is expressed in the form

\[ t'=t+\Delta t(B-760), \tag{62} \]

consequently, $t'>t$ at pressures $B$ above 760 mm, and $t'<t$ when $B<760$ mm; the last column gives the degree of accuracy $\varphi$ in °C with which the corresponding boiling point has at present been determined.

As is evident, in this respect there still remains quite a lot to be done.

In particular, with regard to individual bodies it should be noted as follows: for alcohol, the boiling temperatures are given on the assumption of anhydrous spirit, or absolute alcohol; the water content raises the boiling point, bringing it close to $100^\circ$; the boiling points of alcohol are used comparatively rarely, because of the great convenience, for these temperatures, of the method of comparison.

The use of the boiling points of aniline and naphthalene is very convenient, but the drawback of these substances is that they, like all organic substances in general, readily decompose; therefore, when using them, certain precautions must be taken, for example, the temperature should be raised only gradually and, most importantly, boiling should not be continued for too long.

When using mercury, it must be borne in mind that during boiling it becomes contaminated by the absorption of oxygen, and although its boiling point $t$ is not noticeably changed by this, the boiling itself becomes irregular and violent; therefore the mercury must from time to time be replaced by pure mercury.

As for the boiling temperature of water, which in the graduation of instruments serves for determining the principal, fundamental point of the scale, in view of the importance of this temperature and of its frequent use, its dependence on pressure is given separately in Table 22, according to the most recent experiments of H. Wiebe, carried out at the German Physico-Technical Institute.

Table 22.

Barometer height, mm	millimeters 0	millimeters 1	millimeters 2	millimeters 3	millimeters 4	millimeters 5	millimeters 6	millimeters 7	millimeters 8	millimeters 9
700	97,714°	754	793	832	872	911	950	989	*028	*067
10	98,106	145	184	223	262	301	340	378	417	485
20	98,494	532	571	609	648	686	724	763	801	839
30	98,877	915	953	991	*029	*067	*105	*142	*180	*218
40	99,256	293	331	368	406	443	481	518	555	593
50	99,630	667	704	741	778	815	852	889	926	963
60	100,000	037	074	110	147	184	220	257	293	330
70	100,366	402	439	475	511	548	584	620	656	692
80	100,728	764	800	836	872	908	943	979	*015	*051
90	101,086	122	157	193	228	264	299	335	335	405
800	101,441

Now let us proceed to the description of the thermostats themselves, i.e., the instruments in which the boiling of the indicated bodies is carried out. We shall begin with the water thermostat.

In Fig. 147 there is shown, at 1:10 natural size, the most widespread type of such an instrument, in general identical with the instrument in Fig. 143 and differing from it only in details of construction: $a$ is a copper boiler, onto the neck of which is fitted a cylinder $b$ of zinc or white tin with double walls; from above the instrument is closed by a lid $g$, also with a double wall to reduce cooling; stoppers are inserted into the tubes-openings, through which the instruments being tested are passed; the steam from the boiler rises upward to the lid itself, passes through the openings $e,e$, and descends through the annular space, protecting the inner cavity from cooling; the tube $d$ connects the inner cavity of the instrument with the outside atmosphere, and through it steam and condensate also leave; so that, during more prolonged operation, too much steam (moisture) is not released into the room, the tube $d$ may be connected to a Liebig condenser; $f$ is a glass water manometer, connected to the tube $c$, which communicates with the inner steam space; by this manometer the flame of the burner heating the boiler $a$ is regulated so that steam formation is not too vigorous and the steam pressure is equal to the barometric pressure; $h$ is a glass tube indicating the water level. To reduce fuel consumption it is useful to wrap the vessel $b$ with asbestos cardboard.

Fig. 147.

When using this instrument it is necessary to ensure that the sensitive part—the mercury bulb, the junction of the thermoelement, or the platinum coil—is by no means immersed in the water, since its temperature is not equal to the temperature of the steam, determined from the barometric pressure $B$ by means of Table 22; the temperature of the water is usually somewhat higher than the temperature of the steam, partly because of the salts contained in the water, and partly, after some time of boiling, because boiling is hindered by the absence of air bubbles.

On the other hand, although the outer casing and the annular steam space protect the inner cavity from cooling, nevertheless in the upper part of the apparatus the temperature of the steam may be somewhat lower; therefore the sensitive part should not be raised unnecessarily high above the water level. In the case of checking mercury thermometers, they should be inserted so that the end of the column is barely visible, or, better still, they should be immersed deeper and withdrawn only for reading; in this way the correction for the protruding column is avoided.

For control it is useful here also to keep in the apparatus an accurate normal thermometer. With proper use, the accuracy that can be attained with this instrument is about $0.04^\circ\mathrm{C}$.

In Fig. 148 there is shown, approximately at $1:10$ of natural size, a thermostat used at the German Physico-Technical Institute. Its chief feature consists in the fact that the water vapor, which is obtained by heating the small boiler $a$ with the gas ring burner $m$ and which may easily prove to be somewhat superheated, enters the upper cavity of the apparatus only after passing through the tubes $k$, through a layer of water, and, consequently, assumes the temperature of saturated steam; in order that the steam in the upper cavity $b$ should not become superheated under the influence of the products of combustion, the upper part $b$ of the apparatus is placed on an asbestos gasket on the flange of the upper boiler, which in turn also stands on an asbestos gasket; asbestos cardboard moistened with steam gives sufficient tightness under the action of the weight alone of the upper part of the apparatus. The boiling tubes, which pass through the water space of the lower boiler, and the casing directing the motion of the gases, ensure a possibly complete utilization of the burner. A series of water manometers $f, f$ and stopcocks $d, d$ makes it possible to vary somewhat the pressure in

Fig. 148.

upper plane, which makes it possible, when calibrating glass mercury thermometers, to take readings at a gradually increasing temperature; this, as was already mentioned above, p. 138, increases the accuracy of the observations.

Readings are taken only after the thermometers have been subjected to the action of the vapor for $1/2$ hour; the indication is calculated from at least three readings. When normal thermometers are calibrated with this apparatus, the accuracy reaches $0.01^\circ$.

The testing of thermometers by the boiling point of alcohol, amyl acetate, ether, and other liquids with a boiling temperature not higher than $+150^\circ$ to $+180^\circ$ C is carried out in the same apparatus as for water, Figs. 143, 147, and 148, but only with the release of the vapors into a refrigerator, as shown in Fig. 143 or 144.

When using substances with a boiling point approximately from $+180^\circ$ to $+500^\circ$ C, which condense strongly already on contact with air at room temperature, they may be heated on a gas burner simply in a large test tube with a diameter of 25–30 mm and a height of 200 mm; the substance taken should occupy 40–50 mm in height. In order to increase the height of the vapor space, it is useful to wrap the upper $2/3$ of the test tube in height with asbestos board.

The calibration of glass mercury thermometers, in order to avoid or at least reduce the correction for the emergent column of mercury, is better carried out in an even longer test tube, which one has to make oneself from thin-walled, hard-melting glass tubing, sealed at one end, likewise with a diameter of 25–30 mm and a length of 250–400 mm, depending on the length of the thermometer. Heating such a long test tube is more conveniently done not on a gas burner, but in the above-mentioned Heraeus electric furnace, of course placing the latter vertically. To avoid overheating the vapors, the end of the tube with the boiling liquid must be placed in the middle of the furnace in height, where the temperature is highest; in doing so the temperature of the furnace, i.e., the current strength, is regulated so that the boiling is only just able to proceed.

When using substances with high boiling points, such as mercury and sulfur, a high vapor space can be obtained with a gas burner only on condition that the loss of heat by the vapors through radiation is reduced as much as possible, and even that the walls of the tube are somewhat heated. In this case, simply wrapping the test tube with asbestos board is insufficient.

Fig. 149 shows a very simple device which nevertheless gives good results: on a glass test-tube with a diameter of 20–25 mm, an enlargement is blown out at a height of 70–80 mm from its upper, open end, by which it is held in the plaster cylinder $a$; this cylinder is made by pouring plaster into a tin vessel $b$ with the tube $c$ inserted in it. As soon as the plaster begins to harden, the tube is removed, turning it slightly; then one and the same plaster cylinder can serve for several tubes. The height $h$ of the cylinder is taken according to the tube, 150–250 mm. When checking a thermometer, the tube, filled to 40–50 mm with the working substance taken, is heated on an ordinary Bunsen burner, gently at first, and then, once boiling has begun, without special precautions; the flame of the burner is increased until the liquid produced by condensation of the vapors flows abundantly down the walls of the upper enlargement; the temperature of the vapors is then practically quite the same throughout the whole tube inside the plaster cylinder.

Fig. 149.

Fig. 149.

In view of the influence of the purity of the substance taken on the boiling points, for each substance—aniline, naphthalene, mercury, and sulfur—it is necessary to have its own separate tube, so as not to remove the working substance in order to obtain another temperature. Sulfur, however, must not be left in the tube, since on subsequent heating the solidified sulfur bursts the tube as a result of its great increase in volume on heating; instead of pouring out the molten sulfur after work, one may, of course, simply take out the tube and lay it on asbestos cardboard; owing to the enlargement, the tube will lie with its open end slightly upward, the sulfur from the lower part of the tube will spread in a thin layer almost along its entire length and, on heating, will gradually flow back down again without bursting the tube.

In particular, when checking thermometers at the boiling point of sulfur, instead of a test tube or a homemade tube one may use the flask from V. Meyer’s apparatus for determining the vapor density of bodies with a high boiling point. The flask $a$, Fig. 150, p. 167, has an enlargement at the bottom to hold a large quantity of boiling liquid and is made of Jena glass, which withstands strong and prolonged heating much better. At the top the flask is closed with a rubber stop—

with openings for 1 or 2 thermometers to be tested and for a curved tube $b$, which connects the interior cavity of the flask with the atmosphere; so that sulfur vapors cannot escape and ignite in the air, a small ball is blown onto tube $b$, which is filled with glass or asbestos wool. To increase the height of the vapor space, and at the same time to avoid overheating of the sulfur vapors, the cylindrical part of the flask must be wrapped with asbestos cardboard, $c$; nevertheless, in order to be quite certain that the bulb of mercury of the thermometer is actually heated to the boiling temperature, it should not be raised higher than 15–20 mm above the level of the boiling liquid.

A very convenient apparatus for using the boiling point of sulfur is shown in Fig. 151: this is also W. Meyer’s flask, which, instead of being raised on a stand, stands by itself on a tripod thanks to a cone $a$ of asbestos cardboard, fastened to the flask by means of a clamp with a tie $b$; cone $a$ serves at the same time as both a support and a casing reducing heat loss; so that the sulfur vapors do not overheat, the cylindrical part of the flask, as above, is also wrapped with asbestos cardboard. When testing an apparatus $t$ with platinum resistance, as shown, for example, in Fig. 151, a similar cone $c$ retains the sulfur vapors, which otherwise might damage its handle.

Fig. 150.

Fig. 151.

A description of apparatus for testing at the still higher boiling points of cadmium and zinc, which are used chiefly in the testing of thermoelements, will be given below in the corresponding chapter.

40. Special considerations in the testing of glass mercury thermometers. — The principal special considerations consist in the so-called calibration and in the change, over time, of the fixed points of the scale—

melting snow and boiling water, and together with them all the thermometer readings.

The calibration of a glass thermometer is the checking of the cylindricity of the capillary tube. In an ideal thermometer with an absolutely cylindrical tube, equal lengths of the tube would correspond to equal volumes of mercury; in actual instruments this cannot be achieved rigorously. Even in the case where, in graduating the thermometer, the divisions are not laid off at equal distances, but the irregularities of the capillary tube are taken into account, equal intervals in degrees still will not contain strictly identical volumes of mercury. Therefore, however carefully a thermometer may have been made, before undertaking precise measurements it is necessary to calibrate it.

The essence of calibration consists in separating from the mercury column a part of definite length and, moving this part along the entire tube, observing with the aid of a magnifying glass or a weak microscope how many divisions this small column successively occupies in the tube; from the observations obtained, the mean length in °C of the column taken is found, and the difference between the observed number of divisions and the mean value gives the correction of the caliber for the corresponding interval of the tube.

Table 23 shows an example of the calibration of a chemical thermometer up to $+300^\circ$C by means of a mercury column of $50^\circ$.

Table 23.

lower reading	upper reading	length of column	correction of interval	correction from $0^\circ$C
$-0.15^\circ$	$+50.13^\circ$	$50.28^\circ$	$-2.52$	$-2.52$
$+50.80$	$99.23$	$48.43$	$-0.67$	$-3.19$
$101.44$	$148.81$	$47.37$	$+0.39$	$-2.80$
$151.72$	$198.68$	$46.96$	$+0.80$	$-2.00$
$201.81$	$248.44$	$46.93$	$+1.13$	$-0.87$
$251.74$	$298.64$	$46.90$	$+0.86$	$-0.01$

Mean $47.76^\circ$

The first three columns of Table 23 are self-explanatory; the 4th column is obtained as the difference between the mean length of the column and the corresponding observed length; the 5th column is obtained by successive algebraic addition of the interval corrections.

In reality, calibration is carried out with the aid not of one mercury column but of several, of different lengths, since with a single column the errors of individual readings successively ...

are added. In column 2 of Table 24 are indicated the corrections of the corresponding scale readings, obtained for the same thermometer by means of a combination of observations with columns 50°, 100°, and 200° long.

When using the thermometer it is much more convenient if the calibration corrections are equal to 0 not at the beginning and end of the scale, but at fixed points 0° and $+100^\circ$C. This means that the mean value of the interval must be taken not for all 6 intervals, as was done in column 3 of Table 23, but only for the first two intervals. The corrections obtained in this way, using more accurate observations with the three indicated columns, are given in column 3 of Table 24. Then, determining the position of 0° and 100° and finding that the distance between these points on the scale of the given thermometer is equal to $100.67^\circ$, i.e., that the scale requires a correction of $-0.67^\circ$, or $-0.335^\circ$ for each 50°, we obtain the figures of column 4. The final calibration corrections for the corresponding points of the scale are obtained as the algebraic sum of the corresponding corrections of columns 3 and 4 and are given in column 5.

Table 24.

scale reading	exact correction	transformed correction	correction for the value of the division	final calibration correction
50°	$-2.52^\circ$	$-0.96^\circ$	$-0.34^\circ$	$-1.30^\circ$
100	$-3.18$	$\pm 0$	$-0.67$	$-0.67$
150	$-2.82$	$+2.01$	$-1.00$	$+1.01$
200	$-2.04$	$+4.46$	$-1.34$	$+3.12$
250	$-0.89$	$+7.29$	$-1.67$	$+5.62$
300	$0$	$+9.85$	$-2.01$	$+7.84$

As we see, for the thermometer taken the correction reaches a very considerable value. In good thermometers the corrections obtained by calibration are significantly smaller. As has already been indicated, the principal standard thermometers must not have calibration errors greater than $0.25^\circ$ for an individual reading; laboratory thermometers are still considered good if the calibration errors do not exceed 1% of the reading for a scale with divisions in whole degrees, and half that amount, i.e. 0.5%, for divisions of $0.1^\circ$.

The calibration itself, i.e. separating off a part of the column of the required length, moving it along the tube, taking an exact reading of the divisions, and calculating the corrections, is a rather complicated matter and requires considerable skill. Since, except for the most exact scientific observations, calibration almost never has to be resorted to, and moreover a detailed description of the methods of calibration and of the methods for calculating the corrections can be found in any more complete manual.

in experimental physics, we shall not dwell on this question in greater detail. Instead of calibration, even good working thermometers are simply compared with a standard one calibrated at the appropriate institution.

Regarding the influence of the thermal after-effect of glass on the readings of a thermometer, it was indicated above, p. 36, that it manifests itself in two ways: as a slow rise of the position of $0^\circ$, sometimes lasting for years, and as a lowering of $0^\circ$ as a consequence of each heating. Accordingly, in every precise glass thermometer one must distinguish two positions of the fixed points $0^\circ$ and $100^\circ$ C.—the highest and the lowered—which are determined in quite different ways.

To determine the first, highest position of $0^\circ$, the thermometer, after being left at rest for at least one year, is immersed in melting ice and kept in it for several days, until its reading, at first slowly rising, finally becomes constant.

The position of $0^\circ$ thus found has no special practical significance in itself; it gives only a characterization of the given instrument in the form of the difference between the highest and the lowered position of $0^\circ$: the smaller this difference, the smaller the thermal after-effect of the glass used, and the better and more reliable the thermometer itself.

The lowered position of $0^\circ$, which corresponds more closely to the working position of $0^\circ$, is found much more simply and quickly: the thermometer is heated to $100^\circ$, exposing it for $1/2$ hour to the action of the vapor of boiling water; it is removed from the water bath and, while still warm, after waiting only until the mercury has fallen to $25^\circ$–$30^\circ$, it is immersed in melting ice and the position at which the mercury again begins to rise is noted.

From the lowering $\Delta t$ found in this way and from the thermometer reading $t_{100}$ at $100^\circ$ found just before it, the value of one division, i.e. $1^\circ$ C., is found in the form

\[ 1^\circ = \frac{t_{100} - \Delta t}{100}. \tag{63} \]

By careful experiments, Pernet established that the value of $1^\circ$ found in this way remains practically constant with the passage of time.

Since the lowering of $0^\circ$ will be different if the thermometer is previously subjected not to $100^\circ$, but to some other temperature, then, in pro—

Measurement of Temperature

...in making especially accurate observations one must proceed as follows: knowing approximately the temperature $t$ to be measured, the thermometer is subjected to the action of this temperature in the corresponding thermostat, again for $1/2$ hour; then the corresponding lowering of $0^\circ$ is determined, after which the desired temperature is measured, and immediately afterward the lowering of $0^\circ$ is again determined. With sufficient accuracy it may be assumed that the mean of these lowerings gives the actual lowering $\Delta t'$ at the temperature $t$. Then, taking into account the value $1^\circ$ corrected according to equation (63), we obtain the true temperature $t$ from the reading $t'$

\[ t=(t'-\Delta t')\cdot \frac{100}{t_{100}-\Delta t}, \tag{64} \]

where for $\Delta t'$ and $\Delta t$ one must, of course, take not the absolute values, but those with the corresponding sign obtained in the verification.

As for the verification itself of glass thermometers, then besides the apparatus described above, p. 155, one may use, and perhaps even with greater convenience, the apparatus shown in Fig. 152, which is easily assembled in any laboratory: it is a glass funnel $a$ with a rubber tube, the stopcock $b$ on which makes it possible, as indicated above, to regulate the removal of water accordingly; a weak spring $f$ presses the thermometer under verification $t$ to the holders $e,e$; the opening $c$ and the adjusting screws on the stand make it possible to set the thermometer strictly vertical, which is important when readings are taken with an especially accurate thermometer.

Figs. 152 and 153.

For verification by means of perfectly pure ice one may use the arrangement shown in Fig. 153: distilled water is frozen in a small glass vessel $d$, after inserting into the vessel...

a glass rod, close in size to the vessel with the mercury of the thermometer; this rod is then slightly heated and removed, and in its place the thermometer $t$ is inserted, to which $d$ is attached by means of a piece of rubber tubing. The thermometer, thus surrounded by pure ice, is inserted into the apparatus described above, and the purity of the ice placed in it no longer has much significance.

In conclusion it may also be noted that thermometers must be immersed in the ice somewhat deeper than $0^\circ$, in order to avoid the influence of the projecting part of the thermometer, which has a higher temperature; this is especially important for precise thermometers with a scale enclosed in a second tube. Moreover, when checking $0^\circ$ in these latter thermometers, and in general in thin-walled thermometers, one must ensure that the ice does not press too hard on the walls of the small vessel with mercury, since otherwise for $0^\circ$ a reading may be obtained that is several hundredths of a degree higher than the true one.

Fig. 154

Fig. 154.

The checking of $100^\circ$ is carried out by means of one of the apparatuses described above and presents no special features; one need only not forget the correction for the projecting column of mercury or, still better, conduct the check so that the column does not project at all, and draw the thermometer out to the boiling temperature only at the moment of reading.

As has already been said more than once, when checking glass thermometers one must first check the upper fixed point, keeping the thermometer in the vapors of boiling water for not less than $1/2$ hour, and only then determine, by the method described above, the position of $0^\circ$C.

Here it is perhaps appropriate to describe also Watson’s apparatus, Fig. 154, constructed specially for checking precise glass mercury thermometers by the method of comparison. The main idea of this apparatus is to avoid completely the correction for the projecting column of mercury by heating the thermometers over their entire height to the same temperature. Both thermometers, the one being checked and the standard one, are bound together and placed in a little cup $a$ with mercury, in which the thermometers also float. The little cup $a$, together with the thermometers, is moved by means of a piston $b$ and a rod $c$ in the glass tube $d$, sealed at the top and open at the bottom, tightly enclosing $a$ and having a diameter of about 25 mm and a height of about 760 mm, i.e., suitable for the longest thermometers.

meters, provided only that they fit in it as regards their thickness. Tube $d$, in turn, is inserted, by means of a rubber stopper $f$, into the glass tube $e$ of about 47 mm diameter. Into the annular space between tubes $d$ and $e$ there is poured, through a funnel with stopcock $h$, Fig. 155, a small quantity of liquid, whose vapors heat both the mercury bath and the thermometers; the heating of the liquid is carried out in the usual way by means of an electric current passing through a spiral $n$ of constantan wire. Variation of the temperature of the vapor bath is achieved, as in the instruments described above, by changing the vapor pressure and by using various liquids.

In Fig. 155 there is shown, approximately at $1:15$ of actual size, a general view of the apparatus: the drawn-out end of tube $e$ is connected with a condenser $g$, which prevents the escape of vapors from $e$; next there is a branch to the three-way stopcock $i$, connecting the inner cavity of $e$ with the mercury vacuum gauge $m$ and with tube $l$, leading to the three-way stopcock $n$. In one position this stopcock connects the vapor space with a water-jet ejector, in the other—with the same pressure air pump. The glass vessel $o$, hermetically isolated from the external atmosphere, serves to equalize the operation of the pumps.

Fig. 155.

All glass parts are fastened to a wooden box—stand.

As working liquids for temperatures from $+20^\circ$ to $+46^\circ$ C, carbon disulfide $CS_2$ may be used; from $45^\circ$ to $80^\circ$, ethyl alcohol $C_2H_5OH$; and from $79^\circ$ to $120^\circ$ C, chlorobenzene $C_6H_5Cl$. Water is not suitable for this apparatus, since it forms dew on the glass, which interferes with readings on the thermometers. In order to protect the rubber tube from the corrosive action of the indicated

liquids, a little mercury is first poured onto the bottom of the annular space. In order that the stopper remain in place under fairly considerable fluctuations of pressure inside the apparatus, it must be glued on; the easiest way to do this is to smear the ends of the tubes with boiled oil; instead, the stopper may also be glued on with rubber cement made of natural rubber dissolved in benzene or chloroform; in this case the stopper must first be washed with gasoline. A stopper glued by one method or another must dry for 4–5 days.

When using liquids with a vapor temperature higher than $120^\circ$ C., the tubes $c$ and $d$ have to be soldered directly at the bottom.

The apparatus must be well insulated against heat loss and, under this condition, gives a very constant temperature, with fluctuations of less than $0.1^\circ$ over 2–3 hours of operation. When the pressure in the apparatus is changed, the thermometer readings are fully established after approximately half an hour.

It was already indicated above, p. 37, that owing to the increase, with rising temperature, of the coefficient of expansion of mercury, a mercury thermometer gives readings that are too high; as indicated there, for example, at $+500^\circ$ C. on the gas scale, a mercury thermometer made of Jena glass 59ʺ and having a uniform scale would read $+527.8^\circ$ C. Other observers, who have worked with thermometers made of other glass, give a different correction, usually somewhat smaller, but nevertheless quite significant for temperatures above $+200^\circ$ C. In view of the fact that this correction depends to a considerable extent on the composition of the glass, it is difficult to give a general formula for calculating it. For the same reason this correction is introduced into the scale at the time the thermometer itself is made: beginning from $+200^\circ$ C., the scale is made more and more stretched out, and the divisions larger and larger. This must be borne in mind when calibrating the thermometer.

In conclusion, it is useful to mention how to proceed in the case when the thermometer to be checked by comparison has finer divisions than the available standard thermometer; for example, one has to check a calorimetric thermometer with divisions of $0.01^\circ$, while the standard thermometer has divisions of $0.1^\circ$. In this case the readings on the standard thermometer need only be made with the aid of a cathetometer, which makes it possible to measure vertical distances with an accuracy of up to 0.01 or even 0.005 mm. Since the distances to be measured are very small, only a few mm, one should attach to the cathetometer, for moving the sighting tube accordingly, a micrometer screw with divisions on the flywheel, after the model of

...calibrate, presents no great difficulties and is inexpensive. The telescope with a magnification of about 40 times must, of course, have the usual system of intersecting threads; moreover, instead of one horizontal thread it is better to place two very close to one another and, when taking readings, not to cover the visible division of the thermometer with the thread, but to set the tube so that it falls exactly between these two threads. The tube is placed at a distance of 4–5 m from the thermometers.

The readings themselves are made in the following manner: when the temperature has become steady, one measures the distance $a$, in mm, of the end of the mercury column from the nearest lower division $t_1$, and the distance $b$ from the nearest upper division $t_2$; for checking, one also measures the distance $c$ between these two divisions. Then, obviously, $a+b=c$ and

\[ \frac{a}{c}+\frac{b}{c}=1. \]

If $m$ is the value of one division in $^{\circ}\mathrm{C}$, then the temperature sought is

\[ t=t_1+\frac{a}{b}\,m=t_2-\frac{b}{c}\,m. \tag{65} \]

This method, proposed by L. Calderon, is very simple and considerably increases the accuracy of measurement; thus, with a thermometer divided in $1^\circ$ it makes it possible to measure the temperature with an accuracy to $0.1^\circ$, or even $0.01^\circ$, and with divisions of $0.1^\circ$, with an accuracy to $0.01^\circ$, or even $0.005^\circ$.

The indicated procedure deserves attention for taking readings not only when checking thermometers, but also in general in precise measurements.

41. Special features in checking thermoelements.—The checking of thermoelements, within the limits from the very lowest temperatures to $+500^\circ\mathrm{C}$, can be carried out with the aid of the same devices as the checking of other instruments. It is necessary to resort to special devices only at higher temperatures. However, even for less high temperatures, when thermoelements are checked frequently, it is convenient to use separate devices necessitated by the construction of the thermoelements.

We shall begin with a description of the device used at the German Physico-Technical Institute when checking thermoelements by the comparison method. This is an electric furnace, drawing 156, p. 176, consisting of a tube $b$-$b$ of refractory porcelain material, about 600 mm long, with an internal diameter of about 25 mm, wound with...

in a helical groove with wire of pure nickel, 2 mm thick; to reduce oxidation of the nickel at high temperatures, tube $b$, over the wire, is coated with a mixture of clay and fireclay

Drawing 156.

in a layer 5 mm thick. Tubes $c$ and $d$, of ordinary porcelain, having a length of about 500 mm and internal diameters of, respectively, 50 and 100 mm, and also porcelain rings $e$ and $f$, glued to the tubes by means of a putty of finely ground fireclay and clay, form air insulation to reduce heat losses; for the same purpose tube $d$ is also wrapped on the outside with asbestos cardboard. Since at high temperatures the properties of porcelain as an insulator deteriorate, in order to avoid the passage of current from the nickel winding to the wires of the thermoelements, the latter, enclosed in thin-walled porcelain or quartz tubes, are placed not directly in tube $b$, but in tube $a$, made of the same refractory porcelain mass and having an internal diameter of about 10 mm. Tube $a$ is supported by porcelain sleeves $h,h$; the projecting ends of tubes $a$ and $b$ are almost not heated.

The thermoelements to be compared, up to 10 pairs at a time, are inserted into tube $a$ in a stretched-out form, i.e., so that one wire of each pair hangs down from one end and the other from the other; the junction is located near the middle of tube $a$. In order that the junctions of all the thermoelements being compared should assume strictly the same temperature, they are placed in the longitudinal slots of a star-shaped little holder $m$ of platinorhodium, drawings 157 and 158, and by means of a wire of the same alloy, laid in the circular groove $n$, drawing 158, are tightly drawn to the holder.

Drawings 157 and 158.

Regulation of the temperature in the furnace is carried out by means of two rheostats—coarse, through $1/4$ ohm, by means of an ordinary Siemens and Halske rheostat with buttons, and fine—by means of a special rheostat made of constantan ribbons with a sliding contact. The furnace described gives temperatures from $+200^\circ$ to $+1400^\circ$ C. The same

the furnace, but with a winding of iridium-platinum wire, makes it possible to bring the temperature up to $+1600^\circ$ C.

At the present time, instead of the furnace described, one may successfully use the Gerus electric furnace mentioned above.

Readings may be taken with a millivoltmeter, but for precise observations it is, of course, necessary to use the zero method described above.

When checking by the boiling points of mercury and sulfur, one may use one of the devices described above. The thermoelements must be introduced into the vaporization region in a thin tube of refractory or quartz glass sealed at the bottom, and all precautions must be taken so that the thermocouple wires cannot in any way come into contact with the vapors of mercury or sulfur.

For frequent checking of thermoelements, K. Barus advises using the apparatus shown approximately at 1:4 natural size in Fig. 159: $a$ is a thick-walled glass tube, sealed below with a concentric thin-walled tube $b$, open at both ends; from below, the thermoelement being checked is inserted into tube $b$ so that the junction is 10–15 mm above the surface of the boiling liquid, and from above it is closed, in order to eliminate the air current, with a plug of asbestos wool; the shield $n$ of asbestos cardboard protects the upper part of the apparatus from heating, which could easily cause superheating of the vapors; to raise the vaporization region, the upper part of the apparatus is wrapped, as always, with asbestos cardboard $p$, while the lower part, in case of breakage, is surrounded by a copper mesh $m$. Heating is carried out by means of an annular gas burner $r$, distinguished by great power and at the same time allowing especially the upper layer of the working substance $k$ to be heated, which gives a steadier boiling.

Fig. 159.

Fig. 159.

Instead of fusing tubes $a$ and $b$, they may be joined by filling with gypsum the annular space between them at the lower end of $a$ to a height of 20–25 mm.

The burner $r$ is shown separately in Fig. 160; it is an annular Bunsen burner with blowing through the nozzle $b$; $d$ is the gas pipe; along the inner annular surface up to 40 openings, 1 mm in diameter, are uniformly distributed. For blowing it is most convenient to use a small water-jet pump connected to a water tap.

Fig. 160.

According to Barus’s experiments, boiling in this apparatus proceeds very quietly, and the temperature of the vapors remains constant, with an accuracy to fractions of a degree, during uninterrupted boiling for several hours.

For mercury and sulfur it is, of course, necessary to have a separate apparatus for each, which, given its low cost, is not burdensome.

Inserting the thermoelement from below is very convenient with respect to the arrangement of the leads and protects the wires of the thermoelements well from the action of the vapors of the working liquid.

For testing thermoelements at the boiling points of such bodies as cadmium, zinc, and others, which boil at red heat and higher, it is necessary to construct special apparatuses which, in principle, are similar to those just described, but made of Jena glass, porcelain, or quartz, and refractory clay, and with an especially strong burner. The author obtained quite satisfactory results with the furnace shown at 1:4 of natural size in Figs. 161 and 162: $a$ and $b$ are two crucibles of refractory clay with holes drilled in the bottom; into the lower crucible, somewhat above half its height, is inserted from above a thin-walled porcelain or quartz tube $c$, closed at the top; for tightness the joint is sealed with putty $d$ made of feldspar.

Figs. 161–162.

The opening in the upper, inverted crucible is closed with a clay rod $e$, but not tightly, rather with a gap, so that, when boiling, the vapors of the metal can escape into the atmosphere; at the same time, the opening for the exit of the vapors must not be too small, otherwise it is easily clogged and the vapors, having no outlet, will blow up the crucible.

The proper operation of the instrument when working with cadmium and zinc can be judged by the small flame obtained at this opening, since the vapors of the named metals, on entering the air, ignite instantaneously.

Of course, for each metal one must have a special pair of crucibles, since the boiling temperature depends on the purity of the working substance.

The furnace into which crucible $a$ is inserted is made of refractory clay and likewise consists of two parts, fitting tightly one upon the other; in the lower half an opening is made for the burner, and in the upper one—for the release of the products of combustion; the furnace, both on top and on the sides, is covered with asbestos board and tightened on the sides with iron hoops $f, f$. A cap of asbestos board is also put on the upper crucible $b$.

The crucibles fit tightly into the furnace, but it is unnecessary to strive for hermetic sealing at the joints. The furnace is placed on a special stand, in the manner of a stool, so that from below there is convenient access to tube $c$ for the thermoelement; this stool must also be covered on top with asbestos board.

The essential part of the instrument is the burner. The latter is made in two designs: an ordinary one—for temperatures up to $1200^\circ$–$1400^\circ$ C., and one for burning gas in a jet of oxygen—for temperatures up to $1700^\circ$ C.

In the first case one may use either an ordinary blast burner, for example, one used for glassblowing work, or even a simple burner without blast, but a powerful one, for example, of the Fletcher system. A good burner can be made very simply from iron gas tubes and a T-piece, Fig. 163: $a$ is a tube, $3/8''$ or $1/2''$ in diameter, 80 to 160 mm long; the gas is supplied through tube $f$, $1/4''$–$3/8''$, and air is blown in through nozzle $e$; the nozzle is made of copper, but it can also be made from a $3/8''$ iron tube by inserting only a copper tip; the nozzle opening is made, depending on the strength of the blast, 1–4 mm. In the T-

Fig. 163.

... it is useful to make an opening $b$ in the nozzle $c$, through which an additional quantity of air from the atmosphere is drawn in by the stream from $e$. The supply of gas and the blast can be regulated by means of clamps with screws, placed on the rubber gas and air tubes.

The burner need not be inserted deeply into the furnace, since in this case the end of tube $a$ can easily melt; by thickening it, i.e., by putting a second iron tube over it, one can improve the heat transfer and prevent melting.

For temperatures up to $1200^\circ$ C. one may use a water-jet apparatus for the blast; to increase the strength of the blast, two or even three such air pumps may be connected in series. One may, of course, also use ordinary foot bellows; to obtain a more uniform jet it is better to take Fletcher bellows with a rubber bag.

Burners with an oxygen-jet blast are very convenient in operation. Such a burner is easily made according to Fig. 164: through nozzle $e$, with an opening not more than 1 mm, oxygen enters from a cylinder, where it is under high pressure, not less than 25 atm.; illuminating gas flows in through tube $f$, $1/4''$—$3/8''$. The principal advantage of such a burner is the possibility of quickly and easily obtaining both the highest temperatures, up to $1700^\circ$ and even higher, and moderate temperatures, regulating only the inflow of oxygen. By decreasing the inflow of gas at a fixed supply of oxygen, one can obtain an oxidizing flame; by increasing it, a reducing flame. To avoid melting the end $d$, this burner must by no means be inserted into the opening of the furnace.

Fig. 164.

Fig. 164.

To obtain $1700^\circ$ in the furnace described, 30 min. is sufficient, it is true, with the furnace preheated beforehand. Unfortunately, working with oxygen is somewhat expensive.

Barus’s careful experiments with a similar furnace showed that the temperature in the case where tube $c$ does not project into the crucible above the upper edge of the furnace is very uniform over the entire height of tube $c$, both in the region of the liquid metal and in that of the vapor, and between these regions it is even impossible to establish an exact boundary...

... These same experiments proved the complete constancy, to within $2^\circ$—$3^\circ$, of the boiling temperature during prolonged boiling, for example of zinc, over the course of several hours.

Instead of the apparatus in Figs. 161 and 162, according to Le Chatelier’s data, one may use an ordinary porcelain crucible $a$, Fig. 165, with a lid with an opening $b$, through which is inserted a thin-walled, slightly bent, glazed porcelain tube $c$, protecting the thermoelement from the action of the metal vapors. It is sufficient to take the crucible 70 mm high and fill it with metal to 50 mm; further, up to the top, in order to prevent oxidation of the metal, it is filled with coal dust. The crucible is heated in a small gas furnace, for example as in Figs. 166 and 167, with the aid of the same burner as indicated above. The temperature of the crucible is raised, as in the preceding apparatus, to the point at which metal vapors appear from the opening.

Fig. 165.

The furnace itself is made of refractory clay and is tightened with iron hoops; in order to increase the service life of the furnace, it is advisable to provide it with an inner lining $b$, which receives the direct action of the high temperature; this lining is made of coarse-grained, up to 1 mm, quartz sand, in the following manner: the sand is moistened with a syrupy solution of water glass; then, having spread on the bottom a layer of the required thickness, 10—15 mm, a bottle of appropriate diameter is placed in the furnace and the resulting annular space is packed with the same mass; so that it does not stick to the bottle, the latter

Fig. 166—167.

they are surrounded with oil paper. Such lining lasts a very long time and is easily replaced. The entire furnace is inexpensive and works well.

In conclusion, it is necessary to mention Barus’s experiments concerning the influence of pressure on the boiling temperature of sulfur and of certain metals. On the basis of his experiments Barus gives an analytical expression for the relation between pressure and boiling temperature. However, this equation as yet has no practical significance, since the change of temperature given by it within the limits of fluctuations of atmospheric pressure is considerably smaller than the degree of accuracy with which it has so far been possible to determine the corresponding boiling points indicated in Table 21, p. 161.

In general, however, the use of the high boiling points of metals for checking thermoelements has at the present time, with the availability of normal thermoelements supplied with certificates from the German Physico-Technical Institute, and of very convenient Heraeus furnaces, lost its significance to a considerable extent with the wide spread of electric energy.

Table 24.

Name	Symbol	$t$, °C	$\varphi$, °C
tin	Sn	$+232^{0}$	$\pm 0.5$
lead	Pb	327	1.0
zinc	Zn	419	1.0
antimony	Sb	630.5	1.0
silver in air	Ag	955	2.0
silver, pure	"	961.5	1.0
gold	Au	1064	1.0
copper in air	Cu	1065	1.0
copper, pure	"	1084	4.0
nickel	Ni	1482	6.0
palladium	Pd	1587	—
platinum	Pt	1780	5.0

The checking of thermoelements by the melting points of metals is carried out much more simply, not only in comparison with boiling points, but also unconditionally, and therefore may be especially recommended for technical purposes. In the table below, 24, are given the temperatures $t$ of several of the most convenient metals for this purpose; the value $\varphi$, in °C, in the last column gives the degree of accuracy with which the corresponding point is known.

has at present been determined. Table 24 has been compiled chiefly from the most recent experiments of Holborn and Day, carried out at the German Physico-Technical Institute.

The verification may be carried out in two ways, depending on whether or not the metal oxidizes in air.

When using an easily oxidized metal, it is melted in a crucible in a furnace of the kind shown in Figs. 161–162 or 166–167. A piece of the given metal, worked into the form of a solid cylinder, Fig. 168, 40–60 mm high, 20–30 mm in outside diameter, and with a bottom 15–20 mm thick, is placed in the crucible. The internal bore is made according to the diameter of the tube in which the thermoelement is enclosed. Owing to the position of the burner in the furnace, the heating of the crucible and of the metal in it takes place from below. When the metal begins to melt, the little cylinder will begin to settle, and the temperature indicated by the galvanometer will stop rising; the thermoelement is removed when the bottom of the little cylinder has melted.

Fig. 168.

Crucibles may be used, for temperatures up to $1000^\circ$–$1100^\circ$, of porcelain or fireclay, and for higher temperatures—of graphite or magnesite.

To avoid rapid oxidation of the metal, the burner flame must be reducing.

The absorption of gases by the liquid metal must influence the reading of the thermoelement, lowering it; however, at present this phenomenon has not yet been sufficiently studied, apart from the influence of atmospheric oxygen on the melting points of silver and copper, as indicated in Table 24. For metals with a lower melting point this influence must be still weaker than, for example, for silver.

When working with metals that are difficult to obtain in the form of thick cylindrical rods, one may, instead of preparing the above-mentioned little cylinder, melt a known quantity of the metal in the crucible, let it cool, and drill the required opening for the thermoelement. The same procedure may also be followed when working with metals from which the indicated little cylinder can easily be obtained, when the same piece of metal is used repeatedly. In that case the melting point is recognized solely by the cessation of the rise in temperature.

Le Chatelier, on the contrary, advises carrying out the verification not by the melting temperature, but by the temperature of solidification of the metal: melt the re-

the required quantity of metal in a crucible, insert the thermoelement into it, and stop the heating. The arrest of temperature during the slow solidification of the metal in the hot furnace is also observed quite distinctly for several, up to 10, minutes, depending on the mass of the metal.

Some authors find that the quantity of metal taken also affects the actual determination of the temperature, giving too low a temperature when only a small amount is used in melting. Thus, Holborn took up to 6 kg of copper when melting it. However, melting such large quantities presents considerable difficulties. In any case, in gas furnaces one should not take less than 0.7–1.0 kg.

Recently, electric furnaces have also begun to be used for melting metals. At the German Physico-Technical Institute, Holborn and Day built a furnace with a vertical fireclay cylinder wound with a spiral of pure nickel, following the model of the furnace in Fig. 156. For the indicated purpose one may use, with still greater convenience, so-called kryptol furnaces. Thus, a kryptol furnace of the Wuest system permits the temperature to be brought up to 1800°–2000° C. In view of the absence of gases, the very uniform temperature, and the convenience of regulating them, electric furnaces must certainly be preferred to gas furnaces. Moreover, in modern furnaces the cost of the electric current consumed is only comparatively slightly higher than the cost of working with gas.

It goes without saying that the aforementioned kryptol electric furnaces can be successfully used not only for melting points, but also for boiling points of metals.

When checking thermoelements made of platinum and platinum-rhodium, or platinum-iridium, by the melting point of non-oxidizing precious metals—gold and palladium—one may proceed as follows: the exposed junction of the thermoelement is wrapped with a thin leaf or wire of the indicated metal and pushed into a sufficiently preheated furnace or crucible; the open flame of a gas burner is completely unsuitable in this case. The readings of the millivoltmeter will begin to rise rapidly and uniformly; but this rise will pause from the moment the melting of the metal taken begins until it is completed, owing to absorption of the latent heat of fusion. When wrapping with wire, care must be taken that the end of the thermocouple is completely covered; with gold, 0.01 g is sufficient for this purpose.

Such a check, while very simple, is not very reliable and requires a certain amount of skill, since the pause in the readings of the milli-

Temperature Measurement

the millivoltmeter is very brief, especially when the amount of metal is small.

A considerably more convenient method is that of Le Chatelier, which consists in connecting the ends of the thermocouple—the platinum and platinum-rhodium or platinum-iridium wires—by means of a thin wire of a fusible metal, gold or palladium, 5–10 mm long. The junction is gradually inserted into a furnace heated in advance; the pointer of the millivoltmeter will begin to move until, at the moment when the connecting wire melts, the current is interrupted; this reading will correspond to the melting temperature of the metal taken. Since near the melting temperature the wire loses a considerable part of its strength, the interruption of the current may occur not from melting but from rupture of the wire, so that the millivoltmeter gives only the lower limit of the melting temperature. The experiment should be repeated three times in succession, and from the readings obtained one should take not the mean, but the greatest reading.

Such a check is quite convenient, but unfortunately it is not applicable to other metals, owing to their easy oxidation, which gives considerably lower temperatures; silver, however, gives by this method a temperature lower by only 1.5–2°.

The melting point of platinum may be determined by the same method, with only the difference that in this case one can no longer use a red-hot furnace, but must heat the thermocouple in an open flame, which must be sufficiently large for its temperature to be constant.

Instead of determining the melting point of platinum by the interruption of the current, in which it is rather difficult to catch the greatest deflection of the pointer, one may make use of the melting of platinum when forming the very junction of the thermoelement, proceeding as indicated above: twist the ends of the wires over a length of 5–10 mm, hold them in the flame of an oxy-gas burner, gradually increasing the gas jet until the platinum begins to melt. For such a check two persons are needed: one regulates the gas supply and watches for the formation of a droplet of molten platinum, the other observes the deflection of the millivoltmeter pointer.

The author obtained quite satisfactory results by this method when fusing the ends of a thermocouple made of silver and constantan. The deflection of the pointer gives, of course, the melting temperature of the more fusible metal, in the latter case silver.

With the described methods of checking by boiling points, and, with known precautions, also by the melting of metals in crucibles in

of a gas furnace or, still better, an electric furnace may, either directly or with insignificant modifications, also be used for checking instruments based on measuring the electrical resistance of platinum. In checking these instruments, especially those with a long platinum spiral, care must be taken that the entire spiral is subjected to the action of the given temperature; this is comparatively easy to achieve when metals are boiling, and much more difficult when they are melting. Instead of using the melting points of a metal in the present case, it is better to use its solidification point.

Conclusion.

42. Final conclusions.—Comparing everything set forth in the present work, we arrive at the following general conclusions:

1, the principal instruments for determining and measuring temperatures are instruments based on the expansion of gases;

2, for practical use, whenever the temperature limits and other local conditions allow, preference should always be given to glass mercury thermometers, as the simplest, most reliable, and most accurate;

3, thermoelements and instruments based on measuring the electrical resistance of platinum are just as reliable and accurate, although more complex; moreover, these two types of instruments are applicable over very broad ranges of temperatures;

4, for measuring high temperatures, optical pyrometers are in many cases very convenient;

5, the remaining instruments and methods are significantly inferior to those indicated above in respect of accuracy, and some of them, such as Seger cones, Wiborgh’s thermophone, etc., give only single readings;

6, existing instruments make it possible to measure temperatures within the broadest limits, from $-260^\circ$C to $+2000^\circ$C and even higher, and under the most varied conditions;

7, contrary to former opinions, it may be said that, with proper handling, observance of the necessary precautions, or introduction of the appropriate corrections, and sufficiently frequent checking, existing instruments possess entirely satisfactory accuracy, not only for technical but also for scientific purposes;

The verification of various instruments for measuring temperature is carried out comparatively simply and does not require overly complicated apparatus;
The verification of instruments in the appropriate special institutions has now been placed on the proper footing and is one of the reasons for the results achieved in recent times with respect to the accuracy of measuring various temperatures.

Appendix 1.

Desiring to facilitate the introduction into practice of various instruments for measuring temperature, to which at present in many industries, unfortunately, sufficient attention is not paid, the author has decided to give in Table 25 a list of the better-known European firms from which various instruments may be ordered.

Bearing in mind that in practice, when choosing the appropriate instrument, one often has to take its cost into account as well, the following Table 26, in which it is indicated from which firms the respective instrument can be obtained, is supplemented by an indication of prices. It goes without saying that the prices given are only average, approximate ones, for preliminary consideration. More detailed and exact information would increase the size of Table 26 too much and would not correspond to the purpose of the book; anyone can obtain such information by requesting a catalogue from the respective firm.

Table 25.

No. in order	Name of firm	Address	Representative in Russia, No.
1	Paul Altmann	Berlin NW. 6., Luisenstr. 47.	31
2	W. Apel	Göttingen.	—
3	C. W. Julius Blancke & Co., G. m. b. H.	Merseburg/Saale.	—
4	Paul Braun & Co.	Berlin N. 113, Seelowerstr. 6.	—
5	The Cambridge Scientific Instrument Co., Ltd.	Cambridge, England.	—
6	Chem. Laboratorium für Tonindustrie	Berlin NW. 21, Dreysestr. 4.	—
7	Dreyer, Rosenkranz u. Droop	Hannover.	—
8	I. C. Eckardt	Stuttgart-Cannstatt.	—
9	D-r Walter Gérard	Berlin W. 50, Kurfürstendamm 11.	—
10	Arno Haak	Jena Karl-Zeissstr. 12.	—
11	Hartmann & Braun A.—G.	Frankfurt-a.Main.	34

Table 25 (continued).

No. in order	Name of firm	Address	Representative in Russia No.
12	Cornelius Heinz	Aachen	—
13	W. C. Heraeus, G. m. b. H.	Hanau a. Main	—
14	Franz Hugershoff	Leipzig, Carolinenstr. 13.	28
15	Institut für Gährungsgewerbe	Berlin N. 65. Seestr.	—
16	Keiser & Schmidt	Berlin-Charlottenburg, Charlottenburger Ufer 53/54	—
17	Max Kohl, A.-G.	Chemnitz i. S., Adorferstr. 20	—
18	H. Maihak	Hamburg, Grevenweg 57.	—
19	F. Mollenkopf	Stuttgart, Torstr. 10.	—
20	Jules Richard	Paris, rue Mélingue 25.	—
21	D-r Siebert & Kühn	Cassel.	33
22	Schäffer & Budenberg, G. m. b. H.	Magdeburg-Buckau.	30
23	W-we Joh. Schumacher	Köln a. Rhein, Bayenstr. 57.	29
24	Steinle & Hartung	Quedlinburg a. H.	—
25	Vereinigte Fabriken für Laboratoriumsbedarf, G. m. b. H.	Berlin N., Scharnhornstr. 22.	32
26	Vereinigte Lausitzer Glaswerke, A.-G.	Berlin NW. 40, Heidestr. 55/59	—
27	Zabel & Co.	Quedlinburg a. H.	—
28	Franz Hugershoff	Moscow, Rozhdestvenskii Blvd., Mattern House	representative of foreign firm No. 14
29	G. H. Kohlstorn	St. Petersburg, Tserkovnaya 1–33.	23
30	Langenzippen & Co.	St. Petersburg, Kamennoostrovskii 11.	22
31	N. Lunz	St. Petersburg, Sadovaya 27/9.	1
32	Dr. O. G. Lunz	Moscow, Myasnitskaya, Langovoy House 40.	25
33	G. A. Pikersgill & Co., T./H.	Moscow, Varvarka, Srednie Ryady 178/9.	21
34	F. B. Raumb	St. Petersburg, Panteleimonovskaya 13.	11
35	O. Richter	St. Petersburg, Admiralteiskii Ave. 4.	—
36	R. Roevskii, technical office	St. Petersburg, B. Ruzheinaya 126.	—
37	Siemens & Halske	St. Petersburg, Admiralteiskaya Emb. 46.	—
38	E. S. Tryndina Sons	Moscow, Lubyanka, own house.	—
39	Thomas Firth & Sons	Riga.	—
40	F. Schwabe	Moscow, Kuznetskii Most, Golitsyna House.	—

Temperature Measurement

Table 26.

Name of instrument	Temperature limits, °C	Cost in rubles	Nos. of corresponding firms according to Table 25
Air thermometer (Joly and others)	0 to 1000°	50 to 100	10
Vessel for such an instrument	0 to 1000	4 to 40	19, 21–33, 25–32
Recording gas pyrometer of Richard	0 to 700	130 to 200	20, 35, 40
Air pyrometer of Siegert-Dürr	0 to 1500	250	35
Air pyrometer of Wiborgh	0 to 1500	90	1–31
Air pyrometer of Wiborgh, Uhling-Steinbart	0 to 1650	300	14–28, 17, 19, 25–32, 26, 38
Air pyrometer of Wiborgh, Krell	0 to 900	—	1–31, 10, 14–28, 19, 25–32, 26, 38
Alcohol glass thermometer	−100 to +30	4 to 10
Pentane glass thermometer	−200 to +30	9 to 23
Mercury glass normal thermometer with certificate	−5 to +105	21 to 33	1–31, 2, 10, 12, 14–28, 15, 17, 19, 21–33, 25–32, 26, 35, 38, 40
Mercury glass normal thermometer with certificate	+90 to +550	18 to 30
Mercury glass laboratory thermometer	−10 to 550	1 to 15	1–31, 2, 3, 5, 6, 7, 8, 9, 10, 12, 14–28, 15, 17, 18, 19, 21–33, 22–30, 23–29, 25–32, 26, 35, 38, 40
Mercury glass laboratory thermometer in a mounting	−10 to 550	6 to 20
Mercury thermometer of quartz glass	300 to 750	50	19, 21–33, 25–32, 26
Mercury calorimetric thermometer, with 0.01°	—	15–23	1–31, 2, 10, 14–28, 17, 19, 21–33, 25–32, 26, 35, 38, 40
Mercury Beckmann thermometer, 0.01°	—	11–25
Mercury Mallke thermometer	−10 to +550	6–18	1–31, 10, 14–28, 19, 21–33, 25–32, 26, 38
Mercury thermometer with contacts (numbered)	−20 to 500	1.50–20	14–28, 15, 19, 21–33, 25–32, 26, 35
Maximum and minimum thermometer (Rutherford)	−30 to 100	3–5	10, 14–28, 17, 19, 21–32, 25–32, 26, 35, 38, 40
Maximum and minimum thermometer of Six	−30 to 70	3–13	17, 22–30
Tallipotazimeters (with ether)	+35 to 180	21–25
Tallipotazimeters (with mercury)	360 to 750	21–45	6, 7, 8, 17, 18, 22–30, 24, 27, 35, 40
Tallipotazimeters, self-recording	−20 to +500	80–175
Graphite pyrometers	0 to 1000	17–45	1–31, 7, 8, 14–28, 17, 18, 19, 24, 24, 27, 35, 38, 40
Fischer pyrocalorimeter	300 to 1500	20–45	2, 14–28, 19, 25–32, 26, 35
Salleron pyrocalorimeter	300 to 1100	40	35, 40
Siemens pyrocalorimeter	300 to 1500	40–60	1–31, 14–28, 19, 25–32, 35, 37, 40
Sentinon instrument	0 to 2500	320	20
Metals and alloys	315 to 1775	1 gr. 0.30–2.80	14–28, 25–32
Seger pyramids	590 to 2030	100 pcs. 2.50–3.50	1–31, 6, 14–28, 19, 25–32, 35, 40
Sentina cylinders (Brerly)	300 to 1050	10 pcs. 1	6, 39
Wiborgh thermophone	300 to 2400	50 pcs. 7	14–28, 35, 40
Mezure and Nouel pyrometric tube	900 to 1500	65	25–32, 35, 40
Tempel spectral pyrometer	700 to 1400	125–140	1–31, 14–28, 25–32, 35
Le Chatelier-Korno pyrometer	600 to 1800	175	35, 40
Féry light pyrometer	600 to 1800	175	5
Spiral light pyrometer	500 to 1700	190	5
Thermoelectric light pyrometer	600 to 2000	200–270	5, 25–32, 35, 36

Table 26 (continued).

Name of instrument	Temperature limits, °C	Cost in rubles	Nos. of corresponding firms according to Table 25
Holborn and Kurlbaum pyrometer	600° to 1900°	300—450	37
Wanner pyrometer	625 to 1000	150—250	6, 14—28, 21—33, 25—32, 35
” ”	900 to 2000	180—250	6, 14—28, 21—33, 25—32, 35
thermoelement Cu + const.	−100 to 600	6—15	4, 5, 11—34, 12, 14—28, 16, 21—33, 22, 25—32, 26, 35, 37
” C + Ni (with galvan.)	0 to +1250	80	4, 21—33
” Pt + PtRh	+300 to 1600	27—120	1—31, 4, 5, 11—34, 12, 13, 14—28, 16, 21—33, 22, 23—29, 25—32, 26, 35, 37
” Pt + PtIr	300 to 1000	20—90	5, 21—33, 37
” Ir + IrRu	1000 to 2000	125	13
galvanometers for thermoelements	—	39—100	1—31, 4, 5, 11—34, 12, 13, 14—28, 16, 19, 21—33, 22, 25—32, 26, 35, 37
” self-recording	—	200—500	4, 5, 11—34, 12, 13, 16, 21—33, 25—32, 35, 37
set for the zero method	−200 to +1200	415	13, 37
instruments with measurement of electrical resistance	−260 to 900	20—150	4, 5, 11—34, 35
quartz instruments with platinum “spiral”	—	20—25	13, 37
set of galvanometers, resistance boxes, etc.	—	70—180	5, 11—34, 13, 37
self-recording set	—	280—600	5, 11—34, 13, 37
instrument for checking by the melting point of ice	0	5—10	5, 10
instrument for checking by the boiling point of water	100	5—15	1—31, 5, 10, 14—28
instrument for checking by the boiling point of sulfur	445	40	5
crucible with alloy Cu + Sn for checking by melting point	742	1	5
Heraeus electric furnace	0—1400	50—170	5, 13

Table 26 (continued).

Name of instrument	Temperature limits, °C	Cost, in rubles	Nos. of corresponding firms according to Table 25
Holborn and Kurlbaum pyrometer	600° to 1900°	300—450	37
Wanner pyrometer	625 to 1000	150—250	6, 14—28, 21—33, 25—32, 35
” ”	900 to 2000	180—250	6, 14—28, 21—33, 25—32, 35
Thermoelement Cu + konst.	−100 to 600	6—15	4, 5, 11—34, 12, 14—28, 16, 21—33, 22, 25—32, 26, 35, 37
” C + Ni (with galv.)	0 to +1250	80	4, 21—33
” Pt + PtRh	+300 to 1600	27—120	1—31, 4, 5, 11—34, 12, 13, 14—28, 15, 21—33, 22, 23—29, 25—32, 26, 35, 37
” Pt + PtIr	300 to 1000	20—90	5, 21—33, 37
” Ir + IrRu	1000 to 2000	125	13
Galvanometers for thermoelements	—	39—100	1—31, 4, 5, 11—34, 12, 13, 14—28, 16, 19, 21—33, 22, 25—32, 26, 35, 37
” recording	—	200—500	4, 5, 11—34, 12, 13, 16, 21—33, 25—32, 35, 37
Set for the zero method	−200 to +1200	415	13, 37
Instruments with measurement of electrical resistance	−260 to 900	20—150	4, 5, 11—34, 35
Quartz instruments with platinum “spiral”	—	20—25	13, 37
Set of galvanometers, resistance boxes, etc.	—	70—180	5, 11—34, 13, 37
Recording set	—	280—600	5, 11—34, 13, 37
Instrument for checking at the melting point of ice	0	5—10	5, 10
Instrument for checking at the boiling point of water	100	5—15	1—31, 5, 10, 14—28
Instrument for checking at the boiling point of sulfur	445	40	5
Crucible with alloy Cu + Sn for checking at the melting point	742	1	5
Heraeus electric furnace	0—1400	50—170	5, 13

Appendix 2.

List of Literary Sources.

In view of the fact that the journal literature on the question of measuring temperatures has by the present time reached very great proportions, so that a complete, so to speak exhaustive, list of it, from the moment scientific thermometry arose, i.e., at least from the beginning of the 19th century, would take up very many pages, and since, moreover, many earlier data have now become entirely obsolete, have retained only historical interest, and, being included in a general list, would only obscure it, the author has decided to confine himself to a list of the literature only for the last 30 years, beginning with 1880, while omitting the smaller articles, chiefly of a purely descriptive character and devoted to instruments that have by now lost their significance.

One of the motives that guided the author in choosing the indicated period of time was the circumstance that all the instruments which one now has occasion to use were invented, or at least brought to their present constructive perfection, precisely during this last period of time.

Persons who may wish to have a list of the literature for an earlier period, or an entirely exhaustive one, can easily supplement the list given below with the aid of Repertorium der technischen Journal-Litteratur, in which almost all articles for 1823–1908 are included, or with the aid of Fortschritte der Physik.

Sources with which the author, for one reason or another, was unable to become acquainted in the original are marked with an asterisk *.

In conclusion, the author permits himself to note that the reader will not find in the list of literature given below quite a number of data, descriptions of the newest instruments, and indications of methods of use and verification that are of known theoretical and critical elucidation. These were obtained by the author himself in the course of his laboratory work, and, in part, so far as concerns descriptions of the newest instruments, were kindly supplied by the respective firms, to whom the author considers it his duty to express his profound gratitude for this.

In order to reduce the number of pages in the list of literature, the titles of articles are given not verbatim, but with abbreviations; the following abbreviated designations have also been adopted for the journals most often encountered:

Am. Journ. — The American Journal of Science.
Ann. Gew. — (Glasèr’s) Annalen für Gewerbe und Bauwesen.
Ann. d. Phys. — Annalen der Physik und Chemie (Poggendorff’s, Widmann’s, Drude’s)

Abbreviation	Journal title
Beibl.	Beiblätter Ann. d. Phys.
Bull. d’enc.	Bulletin de la Société d’encouragement pour l’industrie nationale.
Centr. O. M.	Central-Zeitung für Optik und Mechanik.
C. r.	Comptes rendus hebdomadaires des séances de l’Académie des sciences.
Dingl.	Dingler’s Polytechnisches Journal.
Electr.	The Electrician; London.
ETZ	Elektrotechnische Zeitschrift.
El. Eng. L.	The Electrical Engineer; London.
Eng.	The Engineer.
Engng.	Engineering.
Erfind.	Neueste Erfindungen und Erfahrungen.
Frankl. J.	The Journal of the Franklin Institute.
Gas. Light	The American Gas-Light Journal.
Gén. civ.	Le Génie civil.
Iron A.	The Iron Age.
Iron & Steel I.	The Journal of the Iron & Steel Institute.
Ind.	Industries and Iron.
Inv.	Invention and industrial Record.
J. Gas. L.	Journal of gaslighting, water supply and sanitary.
J. Gasbel.	(Schilling’s) Journal für Gasbeleuchtung und Wasserversorgung.
Mechan.	Der Mechaniker.
Mech. Z.	Deutsche Mechaniker-Zeitung; Beiblatt zu Z. Instr.
Nat. The.	The Nature; London.
O. Z. Bergw.	Oesterreichische Zeitschrift für Berg und Hüttenwesen.
Phil. Trans.	Philosophical Transactions of the Royal Society of London.
Proc. Roy. Soc.	Proceedings of the Royal Society.
Sc. Am.	Scientific American.
Sitz. B. Preuss. Ak.	Sitzungsberichte der Königlichen Preussischen Akademie der Wissenschaften.
Stahl.	Stahl und Eisen.
Techn. Z.	Deutsche Techniker-Zeitung.
Uhl. T. R.	Uhlands Technische Rundschau.
Verh. V. Gew.	Verhandlungen des Vereines zur Beförderung des Gewerbefleisses.
Wien. Ak.	Sitzungsberichte der Kaiserlichen Akademie der Wissenschaften; Wien.
Z. V. d. I.	Zeitschrift des Vereines deutscher Ingenieure.
Z. Instr.	Zeitschrift für Instrumentenkunde.

Articles and books of a general nature, or devoted to the examination of many types of instruments.

* 1. Barnes, measurement of temperature for industrial purposes. Chemic. Ind. 27, 1908 p. 661.

* 2. Baty, methods of temperature indication. Chemic. Ind. 24, 1905 p. 307.

* 3. Bechstein, Instrumente zur Messung der Temperatur für technische Zwecke. Techn. Z. 22, 1905 S. 133.

* 4. ” die Entwicklung der Thermometrie und Pyrometrie. Prometeus, 16, 1905 S. 613; Braunk. 4, S. 425.

* 5. Beckert, Temperaturmessinstrumente. Centr. O. M. 8, 1887 S. 41.

Beguin, mesure des hautes températures. Gén. civ. 28, 1896 p. 388.
Behn u. Kiebitz, eine indirekte Methode zur Bestimmung der Temperatur von flüssiger Luft. Ann. d. Phys. 4, 12, 1903 S. 421.
Blass, Messung von Flammentemperaturen. Stahl. 12, 1892 S. 893.
Bliakher, Heat in the Factory. Riga 1905, pp. 74--89.
Bolz, die Pyrometer. Chem. Z. 12, 1888 S. 1395.
Bond, pyrometers and carbonization. J. Gas. L. 1905 p. 691.

* 12. Borchardt, Bestimmung hoher Temperaturen. Z. Beleucht. 6, 1900 S. 46.

Brand, Technische Untersuchungsmethoden. Berlin 1907, S. 145--161.
Bronn, Apparate und Methoden zur Messung hoher Temperaturen. Z. ang. Chem. 18, 1905 S. 462.

* 15. Browne, pyrometers. Sc. Am. Suppl. 1884 p. 7304; Canad. Mag. 13, 1885 p. 75.

* 16. ” thermometers. Sc. Am. Suppl. 1889 p. 11422.

* 17. Carnelly and Burton, new form of pyrometer. J. chem. soc. 1884 p. 237.

Carpenter, Experimental Engineering. New-York 1904 p. 341--359.
Chappuis, Traveaux du Bureau International des poids et mesures. 1888 A. VI.
Dreyer, on methods of measuring low temperatures. Proceedings of the St. Petersburg Polytechnic Institute III 3--4, 1905 p. 515.
Dewar, absolute zero of temperature. Physical Review 3, 1902 p. 125.

* 22. Evrard, Instrumente zur Messung der Temperatur von glühenden Körpern. Met. Arb. 10, 1889 S. 262.

* 23. Ferrini, measurement of high temperatures. Sc. Am. Suppl. 1880 p. 4031.

Frew’s pyrometer. Engng. 41, 1886 p. 8466; Iron A. 37 No. 5.

* 25. Fuc, Uebersicht der gebräuchlichen Fabrikthermometer. Z. Zuck. Böhm. 17, 1893 S. 500.

* 26. Fuchs, Schwellenthermometer. Chem. Z. Rep. 15, 1890 S. 149; Rep. Phys 27, S. 118.

Gheury, Notes on pyrometry. Engng. 77, 1904 p. 655.
Gramberg, Technische Messungen. Berlin 1900, S. 236--249.
Gray, high-temperature measurements. Chemic. Ind 23, 1904 p. 1192.

Groshans, absoluter Nullpunkt des hundertteiligen Thermometers.
Z. phys. Chem. 42, 1903, p. 626.

* 31. Guillaume, mesure des températures basses. Arch. sc. 20, 1888, p. 396.

Haber and Richardt, chemical determination of flame temperatures.
J. Gas. L. 89, 1905, p. 287.
Landolt u. Börnstein, Physikalisch-chemische Tabellen. Berlin, 1909.

* 34. Lauth, expériences pyrométriques. Compt. r. mineralog., 1887, p. 50.

Lechatelier & Boudouard. Mesure des températures élevées. Paris, 1900.

* 36. Legott’s heat indicator. Inv. 8, 1887, p. 2396.

* 37. Lomshakov. Testing of steam boilers and machines. St. Petersburg, 1897, pp. 97—133.

* 38. Lucas, mesures thermométriques et la température. Bull. Soc. él. 3, 1886, p. 420.

[Lucas], la thermométrie rationelle. Bull. Soc. él. 4, 1887, p. 15.

* 40. Mahlke, Pyrometer und deren Anwendung in der Technik. Polyt. CBl. 57, 1895, p. 13.

Moorwood, industrial pyrometers and modern furnaces. Am. Mach. 31, 2, 1908, p. 589.

* 42. Murrie’s, pyrometer. Mech. World. 20, 1886, p. 272.

* 43. Murrie, description of a series of new thermometers for high temperatures. Chem. ind. 4, 1886, p. 189; Bull. d’enc. p. 348.

Nimvitsky, methods and instruments for measuring temperature. Ural. Techn. 1909, VI, p. 3, VII, p. 13, VIII, p. 7.
Pirani, moderne Temperaturmessung. Mechan. 12, 1904, p. 217; Techn. Rundsch. p. 73.
Pöthe, Temperatur-Messung. Ulb. T. R. 1905, 3, p. 12; Z. Elektr. u. Masch. 8, p. 156.
Pullen. Experimental Engineering. Manchester, 1900, vol. I, pp. 52—69.
Quincke, acustisches Thermometer. Dingl. 307, 1898, p. 300; Z. kompr. G. 2, p. 63.
Richardson, application of the ionization from hot bodies to thermometric work. Phys. Rev. 26, 1903, p. 357; 27, p. 183.
Salomon, neue Thermometerscala. Z. ang. Chem. 1891, p. 409; Centr. O. M. 12, p. 261.

* 51. Schäffer & Budenberg, pyrometer. Ind. 2, 1887, p. 414.

* 52. Schmidt, Temperatur-Messapparate. El. Anz. 25, 1908, p. 1059.

Schütz, Fortschritte in der Messung hoher Temperaturen. Z. V. d. I. 48, 1904, p. 155.
Seger, Bestimmung hoher Temperaturen. Dingl. 284, 1892, p. 215.

* 55. Siemens & Halske, Pyrometer. Prometheus 16, 1905, p. 387.

* 56. Spohr, der heutige Stand der Temperaturmessung. Riy. Ind. Z. 9, 1885, p. 165.

* 57. Strube, Neuerungen an Pyrometern und Thermometern. Z. Rübenz. 4, 1880, p. 203.

Temperature Measurement

Taylor, a magnetic indicator of temperature for hardening steel. El. Rev. 59, 1906 p. 207; Electr. 57, p. 739; Pract. Eng. 34, p. 337; El. Eng. L. 38, p. 237; Mech. World 40, p. 88.
Taylor & Hobson, temperature gauge for steel-hardening. Engng. 85, 1908 p. 693.

* 60. Thumann, Temperaturmessungen in Tiefbohrlöchern. Glückauf. 37, 1901 S. 1105.

Waidner, methods of pyrometry. Iron & Steel Mag. 8, 1904 p. 539; Proc. Eng. Western Penns. 1904.

* 62. Waller, measurement of high temperatures. West. Electr. 24, 1899 p. 274.

* 63. Weber, Messung der Temperatur. Gaea. 26, 1891 S. 539.

* 64. “ Messung der Temperatur von 100° bis 1000°. Ind. Blätter. 28, 1890 S. 73.

* 65. Wiborgh, Pyrometer und ihre Fehler. O. Z. Bergw. 44, 1896 S. 404. Chem. Z. Rep. 20, S. 247.

Winkelmann, Handbuch der Physik. Leipzig 1906. III, 1, S. 1–41, 134–154.
Khvolson, Course of Physics. St. Petersburg 1899. III pp. 14–65.
Bestimmung hoher Temperaturen. Dingl. 244, 1882 S. 438; Chem. CBl. III, 13, 1882 S. 666.

* 69. Pyrometer. Ann. Gew. 13, 1883 S. 175.

* 70. Pyrometer. Mech. World. 17, 1884 p. 142; Nat. The. 30, p. 366; Iron A. 34, No. 10.

Appareils pour mesurer les hautes températures. Portef. écon. 29, 1884 p. 61.

* 72. Pyrometer. Ind. Blätter 22, 1885 S. 182.

* 73. A new pyrometer. J. Gas. L. 46, 1885 p. 966.

* 74. New inventions in thermometry. J. Gas. L. 46, 1885 p. 198.

* 75. The use of pyrometers. Mech. World. 19, 1885 p. 264.

* 76. Thermometers for high temperatures. J. Gas. L. 47, 1886 p. 958. Sc. Am. Suppl. 22, p. 8770.

* 77. Ein neues Pyrometer. Prakt. Phys. 1, 1883 S. 337.

* 78. Temperaturbestimmung bei metallurgischen Processen. Met. Arb. 14, 1888 S. 347.

* 79. Pyrometers in practice. Ind. 13, 1892 p. 241.

* 80. Pyrometers and their use. Ind. 19, 1895 p. 282.

* 81. Genaue Messung von Temperaturen. Centr. O. M. 21, 1900 S. 83; Z. Calciumcarbid 3, S. 334; Z. Glas. 9, S. 42.

* 82. Determination of temperature. Gorn. Zh. 1889, 3718.

* 83. Praktische Erfahrungen über Temperaturmessungen. Erf. 27, 1900 S. 304.

* 84. Messung hoher Temperaturen. Uhland’s W. T. 1897, 3, S. 13.

* 85. Pyromètre pour les hautes temperatures. Vie sc. 1901, 1, p. 206.

Self-recording pyrometers. Verh. V. Gew. Sitz. B. 1902, p. 196.

* 87. Modern pyrometers. Centr. O.M. 23, 1902, p. 158.

* 88. Pyrometers for metallurgical work. Iron & Coal. 68, 1904, p. 1428;
Iron & Steel I. 1904, 1, p. 98.

* 89. Thermometers and pyrometers for industrial purposes. Iron & Coal.
69, 1904, p. 1792.

*• 90. Pyrometer for hardening furnaces and in metallurgical operation. El. Rundsch. 22,
1905, p. 358; Erf. 32, p. 534.

* 91. Thermometers and pyrometers. Sc. Am. Suppl. 60, 1905, p. 24708.

* 92. Pyrometer and pyrometry. Braunkohle. 4, 1905, p. 425.

Pyrometers and their significance for the gasification of bituminous coals.
Z. Beleucht. 11, 1905, p. 264.

* 94. Application of the pyrometer. Met. Arb. 32, 1906, p. 19.

* 95. Pyrometers and temperature measurement. Iron & Coal. 77,
1908, p. 2518.

-* 96. The pyrometer and its applicability in foundry and metallurgical
operation. Giess. Z. 1908, p.

Gas thermometers and pyrometers.

Andrews, air thermometer. Chem. CBl. III. 22, 1881, p. 808.
” new air thermometer. Ber. Chem. G. 14, 1881, p. 2116.
Arndt, determinations of melting points at high temperatures. Verh. V. Gew. Abh. 1904, p. 265.
Baranov, pneumatic pyrometer. Tekhn. Sb. 1899, p. 131.
Barus, On the thermo-electric measurement of high temperatures.
Bull. U. S. Geol. Survey. 1889, No. 54, p. 657.
Bedford, Expansion of porcelain with rise of temperature. Phil. Mag. 49, 1900, p. 90.
Berthelot, on a new method for measuring temperatures. C. r. 120, 1895, p. 831.
” measurement of high temperatures by the interferential method. C. r. 126, 1898, p. 410.

* 105. Biltz, Sitz B. Preuss. Ak. 1895, Jan.

Bottomley, constant-volume air thermometer. Phil. Mag. 26, 1888,
p. 149; Z. Instr. 9, 1889, p. 106.
Bristol, recording air pyrometer. Engng. 71, 1901, p. 150; Foundry
21, 1902, p. 40.
Buckingham, thermodynamic corrections of the nitrogen scale.
Phil. Mag. 15, 1908, p. 526.
Cailletet, a new gas thermometer. Dingl. 269, 1888, p. 222;
Beibl. 12, p. 569; C. r. 106, p. 1055.
” measurements of low temperatures. Rev. ind. 20, 1889, p.
214; J. d. phys. 8, 1888, p. 286; Ann. ind. 20, 2, 1888
p. 375; Lum. él. 28, 1888, p. 528.

Cailletet & Colardeau, Messung niedriger Temperaturen. Naturw. R. 3, 1888 S. 378; Chem. techn. Z. 6, 1888 S 465; Chem. trade. 3, 1888 p. 145.
Callendar, practical measurement of temperature. Proc. Roy. Soc. 41, 1886 p. 231; Phil. Trans. 178 A, 1888 p. 161.
” compensated air thermometer. Proc. Roy. Soc. 50, 1882 p. 247.
” comparaison du thermomètre de platine et du thermomètre à azote. Éclair. él. 24, 1900 p. 80.
” thermodynamical correction of the gasthermometer. Phil. Mag. 6, 5, 1903 p. 48.
Callendar & Griffiths, determination of the boiling-point of sulphure, method of standardysing platinum resistance thermometers. Phil. Trans. 182 A, 1892 p. 119; Proc. Roy. S. 49, 1891 p. 56.
Carpenter, errors of the Jolly air thermometer. Eng. News. 29, 1893 p. 5.

* 118. Chappuis, le thermomètre à gaz. Arch. se. 20, 1888 p. 5.

” Vergleich des Quecksilberthermometers mit dem Gasthermometer. Z. phys. Chem. 2, 1888 S. 975; Beibl. 13, 1889 S. 13.
” notes on gas-thermometry. Phil. Mag. 50, 1900 p. 433.
” notes on gas-thermometry. Phil. Mag. 3, 1902 p. 243.
” nouvelles études sur les thermomètres à gaz. J. d. Phys. 43, 1904 p. 883.
Chappuis et Harker, comparaison entre le thermomètre de platine et le thermomètre à azote. Éclair. él. 24, 1900 p. 78; Phil. Trans. 194 A, 1900 p.; J. d. phys. 3, 10, 1901 p. 20.
Crafts, comparaison des thermomètres à mercure avec le thermomètre à hydrogène. C. r. 95, 1882 p. 836.
” thermomètres à gaz. C. r. 106, 1888 p. 1222.
Crafts et Meier, mesure des températures élevées C. r. 90, 1880 p. 606; Beibl. 4, 1880 S. 442.
Day and Clement, measurements with the gas thermometer. Am. Journ. 26, 1908 p. 405.
Dürr, Luftpyrometer. Z. V. d. I. 35, 1890 S. 791.
Ennis, measurement of draught and temperature of the flue-gases. Eng. Mag. 1907 p. 742.
Féry, la mesure des températures élevées. C. r. 134, 1902 p. 977; ind. él. 11 p. 252.
Gummach, Vergleichung von Quecksilberthermometern mit dem Luftthermometer. Chem. Cbl. III, 13, 1882 S. 226; Rep. an. Chem. 2, 1882 S. 136.
Gaab, Walther-Dürr’sche Pyrometer. Stahl 14, 1894 S. 432.
Harker, on the high-temperature standards of the National Physical Laboratory. Phil. Trans. 203, 1904 p. 343.

* 134. Harker and Chappuis, comparison of platinum and gas thermometers. Chem. News. 79, 1899 p. 301; Gas. Light. 71, p. 167.

Holborn and Day, air thermometers at high temperatures. Ann. d. Phys. 68, 1899, p. 815; Sitz. B. Preuss. Ak. p. 691.
” ” air thermometers at high temperatures. Ann. d. Phys. 4, 2, 1900, p. 505.
” ” Measurement of high temperatures. Z. Instr. 21, 1901, p. 84.
Holborn and Valentiner, temperature measurements up to 1600° with the nitrogen thermometer and with the spectral photometer. Sitz. B. Preuss. Ak. 1906, p. 811.
Holborn and Wien, measurement of high temperatures. Z. Instr. 12, 1892, p. 257; Ann. d. Phys. 47, p. 107.

* 140. ” ” Bull. d’enc. 5, 1896, p. 1012.

” ” Measurement of low temperatures. Sitz. B. Preuss. Ak. 1896, p. 673; Ann. d. Phys. 59, p. 213; Stahl, 16, p. 840.
Jacquerod and Richardt, boiling point of naphthalene, diphenyl, and benzophenone. Ber. chem. G. 37, 1904, p. 2531.
Job, new method for the measurement of high temperatures. C. r. 134, 1902, p. 39; Ind. él. 11, p. 42; Z. Instr. 22, p. 165.
Jüptner, Wiborgh’s air pyrometer. O. Z. Bergw. 36, 1888, pp. 603, 617.
Kammerling-Onnes, hydrogen thermometer for very low temperatures. Z. kompr. G. 2, 1898, p. 1.
Kapp, studies on the air thermometer. Ann. d. Phys. 4, 5, 1901, p. 905; Z. Instr. 22, 1902, p. 91.
Kersten, pneumatic pyrometer (Uehling–Steinhart). Rév. Min. 43, 1898, p. 44; Berg. Z. 57, 1898, p. 498.
Krell, hydrostatic measuring instruments. Berlin, 1897.
Ladenburg and Krügel, measurement of low temperatures. Z. kompr. G. 3, 1899, p. 61; Z. Kälteind. 6, p. 147.
Lebedev. Hydrogen thermometer of the Main Chamber of Measures and Weights. Vremenn. Palaty Mer i Ves. issue 4, p. 57; appendix to Zh. F.-Kh. O. 1899.
Le Chatelier, measurement of temperature from the time of flow of specified quantities of air. Dingl. 277, 1890, p. 46.
Lehrfeldt, a numerical evolution of the absolute scale of temperatures. Phil. Mag. 45, 1898, p. 363.
Leybold, measurement of high temperatures. Dingl. 265, 1887, p. 373.
Lomshakov, Uehling–Steinhart air pyrometer. Gorn. Zh., 1895, p. 349.
Lumsden, new form of pyrometer. J. Chem. Soc. 83, 1903, p. 349.

* 156. Meilink, comparison of platinum thermometers and thermoelements with the gas thermometer. Beibl. 65, 1904.

Mensching and Meyer, description of a pyrometer. Z. phys. Chem. 1, 1887, p. 145; Ber. Chem. G. 20, p. 582.
Meyer, small laboratory air thermometer. Ber. Chem. G. 26, 1893, p. 1047.

Michelson, air thermometer independent of the barometric pressure.
Am. Journ. 3, 24, 1882 p. 92; J. d. phys. 1, p. 183.
Miller, new construction of the air thermometer. Carl. Rep. 16.
1880 p. 251.
” measurement of high temperatures, J. Gas. L. 76, 1900 p. 634.
Morton, the pneumatic pyrometer (Uehling–Steinbart). Engng. 58,
1894 p. 251.
Olszewski, comparison of gas thermometers at low temperatures. Beibl. Ann. d. Phys. 10, 1886 p. 679.
Pellat, the absolute temperature deduced from the normal thermometer.
C. r. 136, 1903 p. 809.
Pinzger, the air pyrometer of Wiborgh. Z. V. d. I. 1894 p. 1547.
Randall, permeability of platinum. Bull. chim. 21, 1898 p. 682.
Rose-Innes, constant-volume gas thermometer. Phil. Mag. 50, 1900 p. 251.
” gas thermometer and the thermodynamic scale. Phil. Mag.
6, 2, 1901 p. 130.
” practical attainment of the thermodynamic scale. Phil. Mag. (6) 6, 1903 p. 353.
” practical attainment of the thermodynamic scale. Phil. Mag. 15, 1908 p. 301.

* 171. Schneebeli, air thermometer for high temperatures. Arch. sc.
9, 1883 p. 355.

Spring, on Wiborgh’s air pyrometer. O. Z. Bergw. 37, 1898 p. 20.
Stock u. Nielsen, a thermometer for low temperatures. Ber. chem. G. 39, 1906 p. 2066.
St. Claire Deville et Troost, determination of high temperatures.
C. r. 90, 1880 p. 727, 773; Beibl. 4, 1880 p. 443; Chem. CBl.
11, 1880 p. 305.

* 175. Travers, the obtaining and measurement of very low temperatures.
Rev. gén. d. sc. 14, 1903 p. 597.

Travers u. Jaquerod, temperature measurement. Z. phys. Chem. 45, 1903 p. 435.
Travers, Senter a. Jaquerod, measurement of temperature. Proc. Roy. Soc. 70, 1902 p. 484.
Uehling & Steinhart, pyrometer. Iron A. 53, 1894 p. 365; Sc. Am. Suppl. 37, p. 15170.
” ” pyrometer. Rev. ind. 29, 1898 p. 354; J. Gas. L. 71, p. 123.
” ” Pyrometer. Stahl. 19, 1899 p. 431; Tonind. 23, p. 833; Am. Mach. 22, p. 84; Iron & Coal 58, p. 553.
” ” Instrument for measuring high temperatures.
Gesundh. Ing. 24, 1901 p. 321.
Uehling, pneumatic pyrometer. Iron & Coal. 60, 1900 p. 160; 61, p. 673.

* 183. ” pyrometer. Iron A. 79, 1907 p. 136; 81, 1908 p. 1613.

Weber, Wärmemessung bei technischen Einrichtungen. J. Gasbel. 32, 1889 S. 1025.
Wiborgh, Luftpyrometer für den praktischen Gebrauch. J. Gasbel. 32, 1888 S. 7; Stahl. 8, S. 699; Berg. Z. 47, S. 468; Chem. Z. Rep. 12, S. 273.
” air pyrometer. Engng. 46, 1888 p. 214, 317; Sc. Am. Suppl. 26, p. 10752.
” air pyrometer. Iron & Steel I. 2, 1888 p. 110; Dingl. 271, 1889, S. 118, 163.
” Verbesserungen am Luftpyrometer. Stahl. 11, 1890 S. 913; J. Gasbel. 34, S. 703; Berg. Z. 50, S. 401.
” Verbesserungen am W. Luftpyrometer. Stahl. 12, 1891 S. 913.
” pyromètre à air. Bull. d’enc. 89, 1891 p. 173.
” air pyrometers. Ind. 12, 1892 p. 522; Eng. min. 54, p. 389.
” Pyrometer. Techniker, 15, 1893 S. 32.
Wiebe u. Böttcher, Vergleichung des Luftthermometers mit dem Quecksilberthermometer. Z. Instr. 10, 1890 S. 16, 233.
Winstanley, air thermometers. Engng. 30, 1880 p. 458; Phil. Mag. 5, 10, p. 380; Dingl. 237, S. 127.
Witz’s air thermometer. Sc. Am. Suppl. 1880 p. 3929; Engng. 30, p. 116; C. r. 91, p. 164.
Messung hoher Temperaturen. Dingl. 265, 1887 S. 373; Töpfer. Z. 18, S. 113.
Messung sehr tiefer Temperaturen. Z. kompr. G. 9, 1905 S. 1.

* 189. A new temperature recorder. Payes Weekly. 8, 1906 p. 191.

Mercury Glass Thermometers.

Adam, Korrektion des herausragenden Fadens mit Hülfe des Fadenthermometers. Z. Instr. 27, 1907 S. 101.
Allihn, Normalthermometer für Laboratorien. Chem. Z. 12, 1888; S. 1555.
” das Ansteigen des Eispunktes bei Quecksilberthermometern aus Jenaer Normalglas. Z. anal. Chem. 28, 1889 S. 435; Chem. Ztg. 33, 1909 S. 1301; Chem. ZBl. 81, 1910 S. 710.
” das Ansteigen des Eispunktes bei Quecksilberthermometern. Z. anal. Chem. 29, 1891 S. 381; Chem. Z. Rep. 14, S. 331.

* 203. Auzenat, vitesse de mise en équilibre des thermomètres. Monit. scient. 56, 1900 p. 753.

Baudin, Sur la diminution du coefficient de dilatation du verre. C. r. 116, 1893 p. 971.
” dépression du zéro des thermomètres. Inv nouv. 1896 p. 398.
Beckmann, Thermometer mit Hilfsteilung nach Kühn. Chem. Z. 26, 1906, S. 337.

Beckmann, Thermometer for the determination of molecular weights and small temperature differences. Z. phys. Chem. 51, 1905, p. 329.
Bergen, Krupp’s hot-blast pyrometers. Iron & Steel I. 1886, p. 207.
Betts, milligrade thermometer. Sc. Am. 82, 1900, p. 170.
Böttcher, depression of the ice point. Z. Instr. 8, 1888, p. 409.
” Testing and cooling of thermometers. Centr. O.M. 24, 1903, p. 207; Mech. Z. 1903, p. 154.
Brown, thermometry. Phil. Mag. 5, 14, 1882, p. 57.
Büchter, innovations in the manufacture of mercury thermometers. Polyt. CBl. 55, 1894, p. 109; Z. Dampfk. Ueb. 17, p. 107.
Calderon, determination of the value of the degree. Ber. chem. G. 21, 1888, p. 3303.
Claydon, determination of the volume of mercury in a thermometer. Phil. Mag. 5, 21, 1886, p. 248; Beibl. 11, 1887, p. 815.

* 216. Cleveland and Abbe. Rep. of. Sign. Office for 1888.

* 217. Chappuis. Archives d. sc. phys. 28, 1892, p. 204.

Chree, thermometry. Phil. Mag. 45, 1898, p. 293.
Crafts, cause of the variations of the fixed points in thermometers. C. r. 91, 1880, p. 370.
” elevation of the zero in mercury thermometers. C. r. 91, 1880, p. 291.
” on the accuracy of measurements made with the mercury thermometer. C. r. 95, 1882, p. 910.
” depression of the zero in mercury thermometers. C. r. 94, 1882, p. 1298.
” comparison of mercury thermometers with the hydrogen thermometer. C. r. 95, 1882, p. 836.

* 224. Crafts and Pernet. Thermometers becoming unusable. Wschr. Brauer. 1, 1884, p. 670; Z. Spiritusind. 7, p. 928.

Dufour, quartz thermometer for high temperatures. C. r. 130, 1900, p. 775; Rev. ind. 31, p. 130; Mech. Z. 1900, p. 89; Tonind. 24, p. 691; Engng. 69, p. 486.

* 226. Earnshaw, a blue-glass pyrometer. Gas. Light. 73, 1900, p. 803.

Gary, the Goerz thermograph. Mech. Z. 1908, p. 148.
Goetze, new Beckmann thermometer. Mechan. 12, 1904, p. 235; 13, 1905, p. 86.
Grieshammer, the manufacture of thermometer glasses. Mech. Z. 1904, p. 233.
Griffiths, comparison of a platinum thermometer with mercury thermometers. El. Rev. 27, 1891, p. 363; Electr. 25, p. 615.
Grützmacher, reduction of mercury thermometers to the air thermometer. Z. Instr. 15, 1895, p. 250.

Grützmacher, Thermometer with variable mercury filling. Z. Instr. 16, 1896, pp. 171, 200.
” thermometric corrections. Ann. d. Phys. 68, 1899, p. 769.
” Thermometer with paper scales. Mech. Z. 1902, p. 84.

* 235. Guillaume, Travaux et Mémoires du Bureau International des poids et mesures. IV, 1886, p. 18; V, 1, 27; VI, p.

* 236. ” le thermomètre à mercure. Arch. sc. 16, 1886, p. 517.

” les thermomètres à mercure. J. d. phys. 6, 1887, p. 228.
” thermometric investigations. Centr. O.M. 8, 1887, p. 235.
” le thermomètre à mercure. J. d. phys. 7, 1888, p. 419; Beibl. 13, 1889, p. 73.

* 240. ” Traité pratique de la thermometrie de précision, 1889.

” solution pratique du problème de la colonne émergente d’un thermomètre. C. r. 112, 1890, p. 87.

* 242. ” la colonne émergente et la mesure des températures. Bull. Soc. chim. 5, 1890, p. 547.

” practical solution of the problem of the emergent thread. Z. Instr. 12, 1892, p. 69.
” Determination of the correction for the emergent thread. Z. Instr. 13, 1893, p. 155.

* 245. Handl, Theory of the thermometer. Carl. R. 17, 1881, p. 300.

* 246. Hebe, most suitable form of the enlargements on thermometer capillary tubes. Z. Glas. 4, 1894, p. 35.

Hecker, Aging of thermometers. Mech. Z. 1901, p. 41.
” Investigation of the constancy of boiling thermometers 59^III. Z. Instr. 21, 1891, p. 133.
Hoffmann, on Schott’s compensation thermometer. Z. Instr. 17, 1897, p. 257.
” Schott’s compensation thermometer. Z. Glas. 7, 1898, p. 1.
Jäger and Gumlich, Production and investigation of mercury standard thermometers. Z. Instr. 15, 1895, p. 2.
Kuhn, practical hints for judging thermometers. Chem. Z. 26, 1902, p. 106; Sprechs. 35, p. 272.
” Use of high-grade factory thermometers. Chem. Z. 27, 1903, p. 54.
Lemke, Reduction of mercury thermometers 59^III to the air thermometer. Z. Instr. 19, 1899, p. 33.
Mahlke, Use of liquid carbon dioxide for the production of high-grade mercury thermometers. Z. Instr. 12, 1892, p. 402.
” Measurement of temperatures up to 550° by means of mercury thermometers. Ber. chem. G. 26, 1893, p. 1815.
” Auxiliary instrument for determining the correction for the emergent thread. Z. Instr. 13, 1993, p. 58.

Mahlke, comparison of mercury thermometers $59^{m}$ with air thermometers. Ann. d. Phys. 53, 1894, p. 965.
" determination of the scale of high-grade mercury thermometers. Z. Instr. 15, 1895, p. 171.
Marchis, thermomètre à zéro invariable. J. d. phys. 4, 1895, p. 217.
" le déplacement du zéro des thermomètres. C. r. 123, 1896, p. 799; 124, 1897, p. 493; 125, 1897, pp. 294, 434.
" déformations permanentes du verre. J. d. phys. 7, 1898, p. 573; Z. phys. Chem. 29, 1899, p. 1.
" contribution à l’étude expérimentale du verre. J. d. phys. 8, 1899, p. 193.
" permanent change of glass. Z. phys. Chem. 37, 1901, p. 533.
" precision thermometry. Z. phys. Chem. 37, 1901, p. 605.

* 266. Mare, Carl. R. 16, 1891.

McClellan, thermometer glass. Franc. J. 152, 1901, p. 63.
Mills, researches in thermometry. Trans. Edinb. 29, 1880, p. 567; Engng. 32, 1881, p. 216.
Minchin, expansion of quartz glass. Z. Instr. 27, 1907, p. 165; Phys. Rev. 24, 1897, p. 1.
Müller, thermometer with compensation of thermal aftereffect. Z. ang. Chem. 1898, p. 29.
" thermometer glass and thermometer cooling. Mech. Z. 1904, p. 202.
Nagle, rate of fall or rise of mercurial thermometers. Trans. Am. Min. Eng. 1892, p. 70.

* 273. Niehls, glass thermometers for high temperatures. Polyt. CBl. 60, 1899, p. 116.

" on the technique of high-grade thermometers. Mech. Z. 1903, p. 205.

* 275. Pernet, variations of the fixed points in thermometers. C. r. 91, 1880, p. 471.

" measurement of temperatures with the mercury thermometer. J. d. phys. 10, 1881, p. 520.

* 277. " Traveaux et Mémoires du Bureau Intern. d. p. et mes. IV, p. 31,

" on corrections caused by capillary forces. Z. Instr. 6, 1886, p. 377.

* 279. " standard thermometer. Chron. ind. 10, 1887, p. 102.

* 280. " Wissensch. Abh. d. Phys.-Techn. Reichsanstalt. 1894.

Philips, compensation of temperatures in thermometers. C. r. 90, 1880, p. 501.
Pickering, on sensitive thermometers for calorimetric purposes. Beibl. 10, 1886, p. 761.
" sensitive thermometers. Beibl. 11, 1887, p. 688; Phil. Mag. 23, 1887, pp. 401, 406.

1. Rabe, Measurement of higher temperatures in technology. Chem. Z. 28, 1904, p. 39; O. Z. Bergw. 52, p. 142; Erf. 31, p. 251.
Raoult, Precision cryoscopy. Z. Instr. 19, 1899, p. 219.
Rimbach, Correction of thermometer readings for the protruding thread. Z. Instr. 10, 1890, pp. 153, 292.
Scheel, Expansion of water with temperature. Ann. d. Phys. 47, 1892, p. 440.
” Tables for reducing readings with mercury thermometers to the hydrogen scale. Ann. d. Phys. 5, 1896, p. 168.
Schloesser, thermometric investigations. Z. Instr. 21, 1901, p. 281.
Schott, Studies of glasses. Z. Instr. 11, 1891, p. 330.
Swan-Dixon, researches in mercurial thermometry. Proc. Roy. Soc. 66, 1900, p. 86.
1. Warmbrunn & Quilitz, Standard thermometer for chemical use. Prakt. Phys. 1, 1888, p. 88.
1. Weber, Influence of the composition of the glass on the depression phenomena of thermometers. Sprechs. 19, 1886, p. 784.
” Depression phenomena of thermometers. Verh. V. Gew. Sitz. B. 1888, p. 135.
1. ” Influence of the composition of the glass on the depression phenomena of thermometers. Chem. Anz. 6, 1888, p. 217; Beibl. 12, p. 649; Sprechs. 21, p. 241.
” Influence of the composition of the glass on the reliability of thermometers. Reimann’s. Z. 20, 1889, p. 287.
1. ” on the manufacture of thermometers for higher temperatures. Glashütte. 19, 1889, p. 133; Sprechs. 22, p. 193.
1. ” Variability of the zero point in thermometers. Hann. Gew. Bl. 1889, No. 14, p. 229.
Weidmann, Connection between elastic and thermal after-effect of glass. Ann. d. Phys. 29, 1886, p. 214.
Wiebe, Influence of the composition of glass on the after-effect phenomena in thermometers. Mitt. Berl. Ak. 9, 1885, p. 629; Gew. Bl. Bayr. 6, 1886, p. 72.
” on thermometer glass. Z. Instr. 5, 1886, p. 167.
1. ” Change in the state of mercury thermometers after heating to higher temperatures. Prakt. Phys. 1, 1888, p. 373.
1. ” on boiling thermometers. Prakt. Phys. 1, 1888, p. 339; Z. Instr. 8, pp. 362, 377.
” Use of mercury thermometers at high temperatures. Z. Instr. 10, 1890, p. 207.
” Comparison of mercury thermometers. Z. Instr. 10, 1890, p. 435.

Wiebe and Böttcher, Comparison of the air thermometer with the mercury thermometer. Z. Instr. 10, 1890, p. 233.
Wiebe and Moeller, Linear expansion of scale glasses. Z. Instr. 28, 1908, p. 137.
Elevation of the zero point in mercury thermometers. Naturf. 13, 1880, p. 376.
Influence of the composition of the glass on after-effect phenomena in thermometers. Z. Instr. 5, 1885, p. 21.
On thermometers becoming incorrect. Wschr. Brauer. 3, 1886, p. 776.
On thermometer glass. Wschr. Brauer. 4, 1887, p. 854.

* 312. A novelty in thermometers. Chem. Rev. 16, 1887, p. 194.

Variability of mercury thermometers. Wschr. Brauer. 4, 1887, p. 180; Chem. Z. 11, p. 305.
Manufacture of thermometers for high temperatures. Verh. V. Gew. Sitz. B. 1888, p. 237.

* 315. Influence of the composition of the glass on the depression phenomena of thermometers. Glashütte. 18, 1888, p. 146.

On sensitive thermometers. Z. Instr. 8, 1888, p. 146.
On thermometers and temperature determinations. Z. anal. Chem. 27, 1888, p. 188.

* 318. Riga Ind. Z. 1894, p. 186.

On the sensitivity of thermometers. Z. Glas. 4, 1895, p. 81.
Thermomètre à mercure. Vie. sc. 1900, 1, p. 424; Ind. él. 9, p. 150.

* 321. Storage of fine thermometers. Z. chem. Appar. 1, 1906, p. 266.

Thermometer. Prakt. Masch. Konstr. 1907, No. 42, p. 30.
Experiments on heat transfer in thermometers. Z. Kält. Ind. 1907, pp. 187, 242.

In addition, Nos. 119, 131, 184, 193.

Glass thermometers with other liquids

Baudin, thermomètre à éther de pétrole. C. r. 133, 1901, p. 1207; Rev. ind. 33, 1902, p. 76.
Benninger, Thermometrograph System Six. Erfind. 14, 1887, p. 398.

* 326. Chappuis, Thermometer for low temperatures. Phys. Rev. 2, 1892, p. 672.

Ebermayer, Maximum and minimum thermometer. Z. Instr. 2, 1882, p. 134.
Hoffmann and Rothe, Expansion of technical pentane at low temperatures and the scale of pentane thermometers. Z. Instr. 27, 1907, p. 265.
Kohlrausch, Thermometer for very low temperatures. Ann. d. Phys. 60, 1897, p. 463.

Maximum-Minimum-Thermometer. Bierbr. 18, 1887 S. 219.
Mewes, Petroleum-ether thermometer. Dingl. 315, 1900 S. 785.
” the Kohlrausch petroleum-ether thermometer. Z. compr. G. 4, 1901 S. 140.
Mirus, Toluol thermometer. Chem. Z. 18, 1894 S. 331.
Rothe, Liquid thermometer for very low temperatures. Z. Instr. 22, 1902 S. 192.
” Pentane thermometer and platinum thermometer. Z. Instr. 24, 1904 S. 47.
Wiebe, Potassium-sodium thermometer. Z. Glas. 4, 1894 S. 1.

Other instruments with liquids.

Chramer, Thermometer for measuring ring-furnace gases (Steinle & Hartung). Tonind. 27, 1903 S. 904.
Debaecker, baroscopic thermometer. Gén. civ. 18, 1890 p. 308; Sc. Am. Suppl. 31, p. 12811.
Eschenhagen, Mercury remote thermometer. Z. Instr. 14, 1894 S. 378.
Henrich, the temperatures in the borehole (outflow thermometer). Z. Bergw. 52, 1904 S. 1.
Lafouque, microthermometer for measuring very small variations of temperature. C. r. 97, 1883 p. 1207; Rév. ind. 14, p. 504; Z. Instr. 4, 1884 S. 173.

* 342. Notice sur les instruments enregistreurs de Richards frères. Paris. 1889.

Schönlan, Mercury pointer thermometer. Prakt. Phys. 4, 1890 S. 155.
Steinle & Hartung, steel mercury thermometer. Ind. Ztg. 26, 1886 S. 254.
” ” steel mercury thermometer. Verh. V. Gew. Sitz. B. 1903. S. 52; Bierbr. 1903 S. 295.
Steinle’s steel mercury thermometer. Engng. 69, 1900 p. 415.
Pyrometer with superheated liquids. Maschinenb. 16, 1880 S. 77.
A recording thermometer. Nat. The. 68, 1903 p. 69.
Thermometers and thermoregulators operated by the pressure of saturated vapors. Sc. Am. Suppl. 60, 1905 p. p. 25048.

Expansion of solid bodies.

* 350. Ahlfelder, Metal thermometer. Prakt. Phys. 4, 1890 S. 154.

Beckert, the graphite pyrometer. Z. V. d. I. 25, 1881 S. 149, 279.
Féry, integrating thermometer. C. r. 140, 1905 p. 367.
” spiral pyrometer. Engng. 85, 1909 p. 663; Phys. Review 10, 1909 p. 169.
Helios-Upton, recording thermometer. Iron A. 70, 1902, 18/9, p. 11.

* 354. ” ” registering thermometer. Eng. Clev. 40, 1903 p. 320.

Hopkins, metallic thermometer. Gas. Light. 74, 1901 p. 10.

Lauth and Vogt, pyrometric measurements. Töpfer. Z. 18, 1887, pp. 49, 73, 98; Tonind. 11, pp. 71, 81; Chron. ind. 10, p. 268; Bull. d’enc. 86, p. 228; Mond. ind. 14, p. 14; Ann. Ind. 19, 1, p. 118.
Maurer, theory and practice of metal thermographs. Z. Instr. 3, 1883, p. 308.
Metal thermometer and metal pyrometer by Zabel. Masch. Constr. 14, 1881, p. 424.
Serpollet, pyrometer. Rév. d. Méc. 1897, p. 771.

Calorimetric instruments and methods.

Amagat, water-circulation pyrometer. Mondes. 3, 6, 1883, p. 463; C. r. 97, p. 1053; Z. Zuck. Böhm. 7, p. 105; Rév. ind. 14, p. 503.
Andrée, hydropyrometer for metallurgical purposes. Berg. Z. 43, 1884, p. 506; Stahl, 5, 1885, p. 144; Mech. World. 19, p. 271.
Braubach, water pyrometer. Stahl, 16, 1896, p. 207.
Fischer, calorimeter for high temperatures. Berg. Z. 50, 1890, p. 184.

* 364. Hoadley, the Siemens pyrometer. Iron A. 27, 1881, No. 21.

Hoadley, the calorimeter as a pyrometer. Am. Mach. 4, 1881, No. 23; Iron A. 28, No. 1.
Hoadley, observations with the platinum pyrometer, with heat carriers of platinum and of iron encased with platinum. Sc. Am. Suppl. 1882, p. 5711; Frankl. J. 114, 1882, pp. 169, 252.
Lauth, Boulier pyrometer. Bull. soc. chim. 90, 1883, p. 108; Rev. ind. 14, p. 399; Ann. ind. 15, 2, p. 314; Sc. Am. Suppl. 1883, p. 6541; D. Töpf.-Z. Ztg. 14, pp. 361, 387.
Lauth, Boulier pyrometer. Thonind. 8, 1884, p. 301; Bul. d’enc. 83, p. 38; Nat. 12, 1, p. 396; Sc. Am. Suppl. 17, p. 6956; 18, p. 7101.

* 369. Le Chatelier, XVI congress of the Société technique de l’industrie de gaz. VI, 1889.

Leclerc, determination of the temperature of furnaces. Rev. ind. 14, 1883, p. 203.
Meyer, on calorimetric temperature determination. Ber. chem. G. 13, 1880, p. 407.
Oberhoffer, the specific heat of iron. Metallurgie. 1908, p. 173; Phys. Rev. 9, 1908, p. 175.
Saintignon, water pyrometer. Chron. ind. 6, 1883, p. 315.
Saintignon, continuous-flow water pyrometer. Semaine. 9, 1884, p. 248; Ann. ind. 16, 2, p. 598.

* 375. Saintignon, pyromètre différentiel à eau. Mondes. 4, 1, 1885 p. 433; Iron A. 36, Nr. 20; Sc. Ann. Suppl. 20, p. 8154.

” das Wasserstrom-Pyrometer. Ann. f. Gew. 219, 1886 S. 54; D. Töpf.-Z. Ztg. 24, S. 251.
” pyromètre. Gaz. 30, 1887 p. 175.
” pyromètre à circulation d’eau. Gén. civ. 16, 1891 p. 328, Dingl. 276, S. 220; Dampf. 7, S. 410.
” pyrometer. Electrochem. Ind. 3, 1905 p. 359.
Calorimètre Salleron Gaz. 48, 1905 p. 194.
Siemens, pyromètre à eau. Gaz. 31, 1888 p. 200.
” Wasser-Pyrometer. Zuckerind. 12, 1887, S. 367.
” ’sche Wasserpyrometer; (separate company brochure).
Wagner, chemical technology. St. Petersburg, 1892, pp. 2—7.

* 385. Z. Dampfк. 1891 S. 130.

Juptner v. Jonstorff, Feneumgsaulagen. 1891, S. 253.

* 387. Techn. Coll. 1893, p. 493.

Mesures pyrométriques. Rev. ind. 18, 1887 p. 8.
Besides that, No. 196.

Pyrometers based on melting, and the thermophone.

Bischof, der Seger’sche Normalkegel Nr. 11. Sprechs. 20, 1887 S. 222; Töpfer. Z. 18, S. 142.
” Pyroskopenkegel. Thonind. 11, 1887 S. 83.
” Schmelzbestimmungen mit Normalkegeln. Töpfer. Z. 18, 1887 S. 12; Glashütte 17, S. 17; Thonind. 11, S. 41.
” Normalthon oder Kegel. Thonind. 22, 1908 S. 775.
Brearley and Moorwood, Sentinel pyrometers and their application to the treatment of tool steel. Iron & Steel J. 1, 1907 p. 261.
Claassen, Herzfeld, Martini. Temperaturmessungen in einem Kalkofen. Z. V. Rüb. Ind. 47, 1897 S. 218.
Cramer, Pyrometer. Thonind. 26, 1902 S. 301, 1090.
Gerbez, indicateur de température pour organes de machines. Rev. ind. 18, 1887 p. 285; Sc. Am. Suppl. 24, 1887 p. 9771.
Hecht, Schmelzkegel. Sprechs. 26, 1896 S. 911.
” Feuerfestigkeitsskala. Thonind. 22, 1898 S. 670; Chem Z. 22, S. 157.
Heintz, über Seger’s Pyroskope. J. f. Gasbel. 1886 S. 894; Wien. Gastechn. Nr. 5.
” Seger’s Pyroskope. Thonind. 11, 1887 S. 2.
Hofmann, Modification der Bischof’schen Feuerfestigkeitsbestimmungen. Thonind. 23, 1899 S. 196.
Lauth, mesures pyrométriques. Gén. civ. 10, 1886 p. 78.

Lauth et Vogt, mesures pyrométriques à hautes températures. Bull. soc. chim. 46, 1886 p. 786.
Ledieu, les pyroscopes. C. r. 94, 1882 p. 1274.
Loeser, die Schmelzvorgänge der Segerkegel. Keram. Rundsch. 1907, Nr. 36.
Rothe, Prüfung von Segerkegeln. Stein u. Mört. 10, 1906 S. 307; Thonind. 30, S. 1473.
Peyrusson, Verwendung von Pyroskopen. CBl. Glas. 12, 1898 S. 238.
Schüler, Temperaturmessungen an Sulfatöfen. Z. ang. Chem. 1897 S. 361.
Seger, zur Bestimmung hoher Temperaturen. Dingl. 256, 1885 S. 191.
" Pyrometer und Messung hoher Temperaturen. Thonind. 9, 1885 S. 121; D. Töpf.-Z. Ztg. 16, S. 146.
" Normal cones for determining temperatures in the kilns of the ceramic industry. Thonind. 10, 1886 S. 168, 245, 262, 280.
" Instruction for measuring the temperature in the kilns of the ceramic industry by means of test cones. Thonind. 23, 1886 S. 229.
" pyrometric measurements. Sprechs. 21, S. 1888 S. 52.
" Determination of the temperature in the kiln with the pyroscope cone. Töpfer. Z. 18, 1887 S. 87; Thonind. 11, 1887 S. 84.
" Normal cone. Sprechs. 20, 1887 S. 116.
" the measurement of high temperatures in the ceramic industry. Thonind. 15, 1890 S. 195; Eisen Z. 12, S. 748.
Substitute for Seger cones. Thonind. 22, 1898 S. 1045.

* 418. Comparison of the Seger cones with the Le Chatelier pyrometer. Berg. Z. 55, 1896 S. 179.

Sentinel pyrometer. Eisenz. 27, 1906 S. 580; Z. Werkzm. 10, S. 458.
Simonis, Seger cones 022 to 7. Tonind. 32, 1908 S. 1764; Sprechs. 41, S. 561.
" the arc for pyrometric determinations. Sprechs. 39, 1906 S. 1283.
Wiborgh, Thermophon. Berg. Z. 55, 1896 S. 257; Dingl. 301, S. 133.
" Thermophon. Z. V. d. I. 41, 1897 S. 297.
Jüptner, Wiborgh’s Thermophon. O. Z. Bergw. 45, 1897, S. 99.
Normal cones for determining the temperature in the kilns of the ceramic industry. Thonind. 1886 S. 135, 145.
On pyrometers, especially on the manufacture of normal cones. Mitt. Zigel. 9, 1886 S. 6.
Pyroscope cones. Töpfer. Z. 18, 1887 S. 35.
Pyroscopes and pyrometers. Berg. Z. 46, 1887 S. 150.
Pyrometric measurements. Sprechs. 21, 1888 S. 105; Thonind. 12, S. 61.

* 430. Engineer, 1894 No. 10.
431 Schmelzpunkte von Metallen. Dingl. 320, 1905 S. 489.
432. Schmelzpunktbestimmung feuerfester keramischer Produkte. Z. Heiz. 10, 1906 S. 231.
433. Pyrometers for metallurgical purposes. Engng. 82, 1906 p. 92.
* 434. Pyrometers for measuring kiln temperature. Clay-Worker. 49, 1908 p. 58; Brik. 28, p. 19.
In addition, Nos. 184, 384, 388.

Optical pyrometers.

Ballois, la mesure des hautes températures. Éclair. él. 46, 1906 p. 484.
Berthelot, méthode optique pour la mesure des températures. Ann. d. Chim. 26, 1902 p. 58.
Khvolson, black temperature. Phys. Rev. 7, 1906 p. 235.
Crova, mesure spectrométrique des hautes températures. Mondes. 51, 1880 p. 433; C. r. 91, p. 252; Ann. d. Chim. 5, 19, p. 167; Z. anal. Chem. 19, 7330.
„ mesure optique des hautes températures au Creuzot. Ann. ind. 13, 1, 1881 p. 433; Rev. ind. 12, p. 169.
„ expériences faites dans les usines du Creuzot pour la mesure optique des hautes températures. C. r. 92, 1881 p. 36,707.
„ sur la mesure optique des hautes températures. C. r. 114, 1892 p. 941.
Darling, Pyrometry in modern workship practice. Eng. 103, 1907 p. 594.
De Grahl, optische Pyrometer. Z. Dampfk. 28, 1905 S. 53,
* 444. Deprez et d’Arsonval, Soc. d. phys. 5. II 1886.
Doelter, Bestimmung der Schmelzpunkte mittels der optischen Methode. Z. Elektrochem. 12, 1906 S. 617.
Ductretet, pyrometrisches Sehrohr. Gew. Bl. Bayr. 21, 1889 S. 293; O. Z. Bergw. 37, S. 326; Rev. ind. 20 p. 74; Man. Build. 21, p. 84.
Ducretet-Lejeune, pyrometrical lunette Iron. 41, 1893 p. 422.
Feld, Temperaturbestimmungen mittels der LeChatelier und Wanner-Pyrometer. Chem Ind. 26, 1903 S. 256.
Féry, nouveau pyromètre. Bull. soc. chim. 31, 1904 p. 701; ETZ, 25. S. 864; Rev. ind. 35, p. 461; J. d. phys. 4, 3, p. 701.
„ optisches Pyrometer. Dingl. 318, 1903 S. 717; Gén. civ. 43, p. 72.
„ la température des flammes. C. r. 137, 1903 p. 909.
Féry, optical pyrometry. Electrochem. Ind. 3, 1905 p. 478; Am. Mach. 28, 2, p. 668; Eisenz. 26, S. 927; J. Gas. L. 90, p. 360; Iron. & Coal. 71, p. 1452.

” das optische Strahlungspyrometer. J. Gasbel. 49, 1906, p. 500; Bayr. Gew. Bl. 1906, p. 394; Uhl. T. R. 1905, 3, p. 5.
” Nouvelles méthodes pyrometriques. Rev. Scientif. 8, 1907, Phys. Rev. 10, 1909, p. 169.
Foiret, the Féry pyrometer. J. Gas. L. 87, 1904, p. 31.
Grauger, nouveaux pyromètres optiques. Rev. chim. 6, 1903, p. 141.
Hase, Pyrometer Wanner. Z. ang. Chem. 15, 1902, p. 715.
Hart, technical thermometry. Electrochem. Ind. 6, 1908, p. 453.
Hempel, Spectralapparat. Z. ang. Chem. 14, 1901, p. 237.
Henderson, pyromètres à radiation de Féry et Holborn & Kurlbaum, Electricien 32, 1906, p. 401.
Holborn u. Kurlbaum, Pyrometer. Z. Instr. 22, 1902, p. 55.
” ” über ein optisches Pyrometer. Ann. d. Phys. 4, 10, 1903, p. 225.
” ” über ein optisches Pyrometer. Sitz. B. Kgl. Preuss. Ak. 1901, p. 712.
Holborn u. Valentiner, Vergleichung der optischen Temperaturskala mit dem Stickstoffthermometer bis 1600°. Ann. d. Phys. 22, 1907, p. 1.
Iklé, optische Methoden zur Messung hoher Temperaturen. Phys. Z. 6, 1905, p. 154.

* 466. Imshenetskii, measurement of high temperatures. Proceedings of the Imperial Technical Society. 1893, p. 72.

Kroupa, thermoelektrisches Pyrometer von Féry. O. Z. Bergu. 53, 1905, p. 298.
Kurlbaum, Methode die Temperaturen leuchtender Flammen zu bestimmen. Phys. Z. 3, 1902, p. 187.
Ladenburg, die Temperatur der glühenden Kohlenstoffteilchen leuchtender Flammen. Phys. Z. 7, 1906, p. 697.
Lampen, electrical resistance furnace for the measurement of higher temperatures with the optical pyrometer. J. Am. Chem. Soc. 28, 1906, p. 846.

* 471. Langley, Am. Journ. 31, 1881, pp. 1, 32, 90.

* 472. Le Chatelier, mesure optique des températures élevées. C. r. 114, 1892, pp. 214, 391; Dingl. 286, p. 43; Rev. ind. 23, p. 94; J. d. phys. 1, p. 185; Ind. électr. avril 1892.

Leduc, pyromètre Féry. Rev. Techn. 25, 1904, p. 190.
Leisse, das Wanner.sche Pyrometer. J. Gasbel. 47, 1904, p. 862.
Lummer u. Kurlbaum, bolometrische Untersuchungen. Ann. d. Phys. 46, 1892, p. 204.
Lummer u. Pringsheim, Temperaturbestimmung mit Hilfe des strahlungsgesetzes. Phys. Z. 3, 1901, pp. 36, 97; Chem. Z. 25, 1901, p. 888.

Lummer and Pringsheim, Temperaturbestimmung von Flammen. Phys. Z. 3, 1902, p. 235.
Lummer and Pringsheim, Strahlungstheoretische Temperaturscala. B.d. D. Phys. G. 1903, p. 13.
Mesuré, mesure des températures des corps incandescents. Gén. civ. 13, 1888, p. 43.
Mesuré, lunette pyrométrique. Nat. 17, 1888, p. 139.
Mesuré & Nouel, pyrometrisches Sehrohr. Bergz. 49, 1891, p. 457; Thonind. 14, p. 575; Z. V. d. I. 34, 1890, p. 1335; Z. Instr. 11, 1890, p. 379.
Morse, thermo-gage. Am. Mach. 26, 1903, p. 1515; Iron. A. 72, 22/10, p. 1; Gén. civ. 44, p. 77; El. Rev. N. Y. 43, p. 792.
Nernst, determination of high temperatures. Phys. Z. 4, 1903, p. 733; Z. Electrochem. 1903, pp. 622, 889.
Nernst, the brightness of incandescent black bodies and a simple pyrometer. Phys. Z. 7, 1906, p. 380.
Nichols, optical method for the measurement of high temperatures. Am. Journ. s. 19, 1880, p. 42; J. d. phys. 10, 1881, p. 92.
Parsy, pyrometers. J. Gas. L. 1905, p. 114; Gaz. 49, p. 68.
Paschen, bolometric investigations. Ann. d. Phys. 48, 1893, p. 286.

* 488. Planck. Verh. d. D. Phys. G. 2, 1900, p. 202.

Schäfer, results of recent radiation measurements. Z. V. d. I. 1902, p. 17.
Siemens & Halske, pyrometric measurements. Centr. O. M. 23, 1902, p. 2.
Thuring, radiation pyrometer. J. Frankl. 165, 1908, p. 363.

* 492. Töllner, temperature-indicating colors. Erf. 28, 1901, p. 422.

Tompson, Mesuré and Nouel’s pyrometer. Iron A. 55, 1895, p. 374.
Violle, la lunette pyrométrique de Féry. Bull. d’enc. 106, 1904, p. 933.

* 495. Waidner and Burgess. Phys. Rev. 19, 1904, p. [[unclear: missing page number]].

* 496. Waidner and Burgess. Bullet. No. 2; Bureau of Standards Washington, 1905.

Wanner, photometric measurement of the radiation of black bodies. Ann. d. Phys. 2, 1900, p. 141.
Wanner, photometric pyrometer. Phys. Z. 3, 1901, p. 112; Chem. Z. 25, p. 1029.
Wanner, the measurement of high temperatures. O. Z. Bergv. 50, 1902, p. 99.
Wanner, pyrometer for high temperatures. Stahl 22, 1902, p. 207; Thonind. 26, p. 779; Z. ang. Chem. 15, p. 44; Portef. éc. 47, p. 73; Riga. Ind. Z. 28, p. 96; Z. V. d. I. 46, p. 616; Mining Journal 1902, p. 280.
Wanner, pyrometer, measurement of high temperatures with it. Stahl 23, 1903, p. 77.

Wanner, pyrometer. Eng. min. 76, 1903 p. 971.
” pyrometer. Iron A. 73, 1904, 18/2 p. 24; O. Z. Bergw. 52, S. 419.
” Pyrometer. Bayr. Gew. Bl. 1906, S. 147.
” ” J. Gasbel. 50, 1907, S. 1005; Z. V. d. I. 52, S. 156.
” ” Centr. O. M. 29, 1908 S. 2.
” das optische Pyrometer fur 625–1000° C. Electroch. Z. 15, 1908 S. 106; Bayr. Gew. Bl. 1908 S. 26.
Wanner, a new pyrometer. Mining Journal, 1902, 4, p. 280.
Wedding, selbstverzeichnende Pyrometer. Stahl, 22, 1902 S. 1309.
White and Taylor, colour temperatures. J. Gas. L. 74, 1899 p. 1549.
Wien, eine neue Beziehung der Strahlung schwarzer Korper zum zweiten Hauptsatz der Warmetheorie. Sitz. B. Preuss. Ak. 1893, S. 55.

* 512. ” Ann. d. Phys. 52, 1894 S. 132.

Wien u. Lummer, Methode zur Prufung des Strahlungsgesetzes absolut schwarzer Korper. Ann. d. Phys. 56, 1895 S. 451.
Wilson a. Gray, effective temperature of the sun. Phil. Trans. 185 A, 1894 p. 361.
Wolfmann, das Wannersche Pyrometer. Sitz. B. V. Bef. Gew. 1902 S. 143.

* 516. Optische Messung hoher Temperaturen. Naturf. 14, 1881 S. 232.

The optical pyrometer. Canad. Mag. 17, 1889 p. 298.

* 518. Notes of the Imperial Russian Technical Society, 1889, III.

* 519. Technical Collection. 1891, p. 482.

Optische Pyrometer. Z. Heiz. 7, 1902 S. 103.
Nouveau pyrometre a radiations thermiques. Ind. el. 13, 1904 p. 253.
Pyrometer. Thonind. 29, 1905 S. 993.

* 523. Optische Pyrometer. Z. Beleucht. 12, 1906 S. 164.

Optical pyrometry. Engng. 83, 1907 p. 323; 84, p. 345, 539.
Optical pyromers for temperatures above 600° C. Iron & Coal. 76, 1908 p. 1696.

In addition, Nos. 130, 353.

Thermoelectric Instruments.

d’Arsonval, mesure des temperatures par l’electricite. Lum. el. 5, 1881 p. 40; Maschinenb. 15, S. 384; Z. Instr. 1, S. 90.
Aten, Anwendung des Heraeus Thermometer. Z. phys. Chem. 1907 S. 66, 643.
Baikov, on contact phenomena in a flame. Bulletin of the St. Petersburg Polytechnic Institute. II, 1–2, 1904, p. 137.
Baikoff, Pyrometer. Sprechs. 37, 1904 S. 1912; Thonind. 28, S. 1671; Chem. Z. 28, S. 1107; Z. anorg. Chem. 38, S. 161.

Ballois, thermoelectric pyrometers. Éclair. él. 48, 1906, p. 372.
Barus, the physical treatment and measurement of high temperatures. Leipzig, 1892.
Barus, measurement of high temperatures. Phys. Rev. 2, 1892, p. 295; Phil. Mag. (5) 34, 1892, pp. 1, 376.
Becker, the Le Chatelier pyrometer in its quartz-glass mounting. J. Gasb. 1907, p. 895.
Bequerell’s, electric thermometer. Thonind. 5, 1881, p. 283.
Bequerell’s, electric thermometer. Electr. 2, 22, 1901, p. 389.
Berkenbusch, measurement of flame temperatures by thermoelements. Ann. d. Phys. 67, 1899, p. 649.
Braun & Co., improvements to electric pyrometers. Thonind. 32, 1908, p. 102.
Bristol, thermometer-thermostat. Eng. min. 77, 1904, p. 839; Street. R 23, p. 871; Eng. News. 51, p. 514.
Bristol, low-resistance thermoelectric pyrometer and compensator. El. Rev. N. Y. 48, 1906, p. 732; J. Nav. Eng. 18, p. 636; Sc. Am. 94, p. 415; Iron A. 77, p. 1610; Eng. News. 55, p. 159; Foundry 28, p. 252; Iron & Coal. 72, p. 2039.
Bristol, electric pyrometer. Iron A. 83, 1909, p. 625.
Bunte, Le Chatelier pyrometer. J. Gasb. Wass. 39, 1886, p. 382.
Charpy, study of the hardening of steel. Bull. d’enc 4, 10, 1895, p. 666.
Chauvier et Arnoux, industrial thermoelectric pyrometers. Electricien, 35, 1908, p. 51; Mechan. 16, p. 158.
Crompton & Co., electric pyrometer. Electr. 56, 1906, p. 808.

* 545. Damour, Bull. d. l’Ass. de l’Ecole de mines. 1889, III.

Day & Allen, temperature measurements to 1600°. Phys. Rev. 19, 1904, p. 177.
De Grahl, temperature measurements by electrical means. Z. Dampfk. 27, 1904, p. 59.
De Grahl, combustion losses and endothermic reactions. Z. Dampfk. 30, 1907, p. 237.
Dewar and Fleming, thermoelectric power of metals and alloys between −200° and +100° C. Phil. Mag. 40, 1895, p. 95.
Dewar, low-temperature investigations. Chem. News. 91, 1905, p. 216.
Dewar, thermoelectric determination of the lowest temperatures. Proc. Roy. Soc. 76, 1905, p. 316; Chem. News. 92, p. 169.
Guillaume, measurement of temperatures by electrical methods. Lum. él. 28, 1888, pp. 201, 312, 499.

* 553. Harris, electric pyrometers for manufacturing processes. El. Mag. 9, 1908, p. 275.

* 554. Hartmann u. Braun, temperature remote-measuring apparatus. El. Anz. 16, 1899, p. 2491.

Hartmann u. Braun, elektrische Temperaturmessapparate. Elt. 31, 1904 S. 305.
Hausrath, Messung kleiner Temperaturdifferenzen mit Thermoelementen. Ann. d. Phys. 17, 1905 S. 735.
Hecht, Temperaturbestimmung mit dem Thermoelement. CBl. Glas. 11, 1896, S. 311; Prom. 8, S. 74.
Heraeus, Thermoelemente. Z. ang. Chem. 8, 1895 S. 430.
", Bestimmung der Schmelztemperaturen von Gussmetallen. Giess. Z. 1, 1904 S. 45; Z. Dampfk. 27, S. 210; El. Rev. N. Y. 45, p. 936.
", zur Messung höherer Temperaturen. Chem. Z. 29, 1905 S. 48.
Heraeus und Keiser & Schmidt, Pyrometer. Z. Instr. 15, 1895 S. 373.
Hirschson, neue elektrische Pyrometer. Eisenz. 26, 1905 S. 583; Chem. Z. 29, S. 185.
", registrierende Galvanometer für pyrometrische Zwecke. Chem. Z. 30, 1906 S. 1093; Z. Dampfk. 29, S. 249.
Holborn, Messungen mit dem Le Chatelier Thermoelement. Z. V. d. I. 41, 1897 S. 226.
Holborn u. Wien, Messung hoher Temperaturen. Ann. d. Phys. 56, 1895 S. 360.

* 566. Howe, Eng. min. 50, 1890 p. 426.

Keiser, ein neues Pyrometer. Chem. CBl. 3, 18, 1887 S. 1335; Am. Chem. J. 9, p. 296.
Küch u. Retschinsky, Temperaturmessungen im Quecksilberlichtbogen der Quarzlampe. Ann. d. Phys. 22, 1907 S. 595.
Küpper’s Pyrometer. Giess. Z. 2, 1905 S. 158.
Kurnakov, a new form of recording pyrometer. Izv. SPb. P. I. I, 1—3, 1904 p. 138; Z. anorg. Chem. 42, S. 184.
Lake, electric furnaces used in treating steel. Am. Mach 31, 2, 1908 p. 234.
Lambert, the pyrometric installation in the gun section Woolwich. Iron & Steel I. 76, 1908 p. 109.
Ladenburg u. Krügel, Messung tiefer Temperaturen. Ber. Chem. G. 32, 1899 S. 1818.
Le Chatelier, pyromètre thermoélectrique. C. r. 102, 1886 p. 819.
", Messung hoher Temperaturen durch Thermoelemente. Beibl. 11, 1887, S. 351.
", les couples thermoélectriques. Gén. civ. 10, 1887, p. 291; C. r. 104, p. 1443; J. d. phys. 6, p. 23; Lum. él. 24, p. 71; Engng. 43, p. 239.
", thermoelektrisches Pyrometer. Berg. Z. 51, 1892 S. 277.
", mesure des températures développées dans les foyers industriels. Bull. d’enc. 91, 1892 p. 276; Chem. Z. Rep. 16, S. 111.

Le Chatelier, Pyrometer. Berg. Z. 55, 1896 p. 344.
” Erfahrungen mit dem —Pyrometer. Dingl. 307, 1898 p. 19.
” Pyrometer. Uhl. W. T. 1900, 3, p. 30.

* 582. ” pyrometer. West. Electr. 31, 1903 p. 94.

” Gebrauch des —Pyrometer. Z. Heiz 7, 1902 p. 127.
” nouveaux pyromètres thermo-électriques industriels. Ind. él. 15, 1906 p. 228; Brick. 24. p. 13.
Lewis, new form of recording apparatus. Chem. Ind. 21, 1902 p. 844.
Lowth Bell, application of the Le Chatelier pyrometer for blast furnaces. Iron, 40, 1892 p. 445.
Marguerre, Versuche an Dampfturbinen. Z. V. d. I. 52, 1908 p. 1347.
Lommuïr and Swinden, the Bristol recording pyrometer. Iron & Steel I. 1, 1909 p. 444.

* 589. Meissner, über eine Fehlerquelle bei thermoelektrischen Messungen. Wien. Ak. Ber. 115, II a, 1906 p. 847.

Meslin, une forme de thermomètre électrique. Electr. 23, 1902 p. 196; Rev. ind. 33, p. 109; C. r. 134, p. 412.
Miller, heat variations in hot processes and means for its determination. Am. Mach. 31, 2, 1908 p. 395.
Moldenke, melting point of cast iron. Iron A. 62, 1898 No. 16 p. 7; Engng. 67, 1899 p. 330.

* 593. Munnoch, Le Chatelier pyrometer. Am Electr. 11, 1899 p. 112.

* 594. Northrup, measurement of temperature by electrical means. Proc. El. Eng. 25, 1906 p. 219; El. World. 47, p. 1191.

* 595. ” electrical methods of measuring temperatures. Eng. Rec. 54, 1906 p. 394.

” direct-reading electric thermometer. Electroch. Ind. 4, 1906 p. 286.
Offerhaus and Fischer, cold-junction temperature corrections. Electroch. Ind. 6, 1908 p. 362.
Palmer, thermoelectric determination of temperatures 0°—200° C. Phys. Rev. 21, 1905 p. 65.
Pecheux, détermination à l’aide de pyromètres thermo-électriques des points de fusion des alliages de l’aluminium. C. r. 143, 1906 p. 397.
” mesure des températures élevées à l’aide des couples thermo-électriques. Lum. él. 2, 1, 1908 p. 29.

* 601. Pellat, méthode à évaluer les très basses températures. Ind. él. 11, 1902 p. 14.

Pfeiffer, Handhabung des Le Chatelier Pyrometers. Z. ang. Chem. 14, 1901 p. 390; Mitth. Dampfk. 24, p. 301.
Pillier, nouveaux pyromètres thermoélectriques industriels. Bull. Soc. él. 6, 1906 p. 183.
„Queen“, electrical pyrometer. Eng. min. 73, 1902 p. 80.

Quincke, Temperaturbestimmungen mit Thermoelementen. Z. V. d. I. 40, 1906, p. 101.
Ramakers, measurement of high temperatures. El. Rev. N. Y. 46, 1905, p. 780; Sc. Am. 92, p. 427.
Roberts-Austen, use of the Le Chatelier pyrometer. Iron & Steel I. 1891, p. 90; Proc. Inst. M. Eng., p. 543.
—— —— measurement of high temperatures. Proc. Civ. Eng. 110, 1892, p. 152; Iron. 40, p. 268.

* 609. —— —— Chem. Ind. 16, 1896, p. 1.

—— —— recording pyrometer for industrial use. Chem. Ind. 16, 1897, p. 5; Proc. Inst. M. Eng. 1897, p.

* 611. Rosset, measurement of high temperatures in electrical laboratories. CBl. Akkum. 6, 1905, p. 23.

Russner, Le Chatelier Pyrometer. Z. Oest. I. V. 49, 1897, p. 9.
Sauveur and Whiting, detection of the finishing temperatures of steel rails. Railr. G. 1903, p. 587.
Schuyler B. Serviss, the internal temperature gradient of metals. Am. Journ. 4, 24, 1907, p. 451.
Siemens & Halske, thermoelectric apparatus for temperatures up to 600° C. ETZ. 1881, p. 246; Electr. 7, 1881, p. 221.
—— —— thermoelectric measuring apparatus up to 600° C. Maschinenb. 19, 1884, p. 232.
—— —— Pyrometer. Centr. O. M. 20, 1899, p. 76; Uhl. W. T. 1899, 3, p. 6.
—— —— a recording pyrometer. Z. Instr. 24, 1904, p. 350; Iron & Coal. 68, p. 41.
Siemens and Halske, pyrometer. Obshched. Tekh. 1899, p. 92.
Stansfield, improvements in the thermoelectric pyrometers. Phil. Mag. 46, 1898, p. 59.
Thuring, electric pyrometers. Iron A. 1907, p. 304.
Waggener, measurement of flame temperatures by thermoelements. Ann. d. Phys. 58, 1896, p. 579.

* 623. Waldo, electrical measurement of furnace temperatures. El. Rev. N. Y. 38, 1901, p. 82.

Wedding, application of the Le Chatelier-Heraeus pyrometer. Stahl. 16, 1896, p. 660.
—— Determination of the temperature of metal baths with the Le Chatelier pyrometer. Glückauf. 34, 1898, p. 394.
—— Demonstration of the Le Chatelier-Heraeus pyrometer in metal baths. Verh. V. Gew. Sitz. B. 77, 1898, p. 106.
White. Constancy of thermoelements. Z. Instr. 27, 1907, p. 259; Phys. Rev. 23, 1906, p. 449.

Wien, Thermoelement Platinrhodium. Chem. Ind. 18, 1895 S. 214.
Wolff, Feussner’s Kompensationsapparat. Z. Instr. 21, 1901 S. 227.
Zehnders Pyrograph. Masch. Constr. 253, 1882 S. 188.
Zeleny, galvanometer scale for the direct reading of temperatures.
Phys. Rev. 27, 1908 p. 141.
Temperaturmessungen in elektrischen Rohröfen. J. Gasbel. 39,
1896 S. 833.
Messung hoher Temperaturen. Dingl. 303, 1897 S. 39.
Thermoelement für sehr tiefe Temperaturen. Dingl. 304, 1897 S. 57.
Elektrische Temperatur-Messapparate. Glückauf. 36, 1900 S. 118.
* 636. Measurement of temperature by electrical means. Chem. News.
86, 1902 p. 273.
* 637. Elektrische Temperatur-Messapparate. Centr.-O.M. 25, 1904 S. 74.
* 638. Thermoelektrische Pyrometer. Turb. 1905 S. 282.
Thermoelektrische Pyrometer. Thomind 29, 1905, S. 993.
* 640. Electrical pyrometers. Iron & Coal. 72, 1906, p. 1861.
* 641. Recent electric pyrometer applications. El. Rev. 59, 1906 p. 198.
* 642. An electric recording pyrometer. El. World. 48, 1906 p. 929.
Thermoelektrisches Pyrometer mit kompensiertem Element. ETZ,
1907 S. 155.
Thätigkeit der Reichsanstalt für 1906. Z. Instr. 27, 1907, S. 109.
Thätigkeit der Reichsanstalt für 1907. Z. Instr. 28, 1908, S. 174.
* 646. Eng. Mag. 1909 p. 174.
In addition, Nos. 99, 101, 130, 133, 135, 136, 139, 141, 149, 156,
284, 369, 442, 448, 458, 466, 486, 490, 509, 517.

Instruments with a change in electrical resistance.

Barnes and McIntosh, platinum thermometer. Phil. Mag. 6, 1903
p. 550.
Braun, das elektrische Pyrometer. CBl. Electr. 10, 1888 S. 898; ETZ.
9, S. 421; Z. phys. Chem. 2, S. 976; Lum. él. 30, p. 65.
„ über ein elektrisches Pyrometer. Z. Instr. 9, 1889 S. 150;
Dampf. 6, S. 808.
„ Widerstand-Pyrometer. Töpfer. Z. 33, 1892 S. 302; Rig.
Ind. Z. 18, S. 7.
Brown, pyrometer. Iron A. 63, 1899, No. 6—4, p. 10.
Bruger, registrierendes elektrisches Widerstandsthermometer. Phys. Z.
7, 1906 S. 775; ETZ. 27, S. 531.
Callendar, mesure des températures par les variations de résistance.
Lum. él. 23, 1887 p. 78.
„ construction of platinum thermometers. Phil. Mag. 32,
1891 p. 104.

Temperature Measurement

Callendar, melting point of gold and silver. Phil. Mag. 33, 1892 p. 220; Iron, 39, p. 489; Engng. 53, p. 762; Z. Instr. 12, p. 213.
” pyromètre électrique. Lum. él. 44, 1892 p. 73.
” practical thermometric standard. Phil. Mag. 48, 1899 p. 519, El. Eng. L. 24, p. 465; Z. Instr. 19, p. 184.
” recorder and platinum thermometer. Engng. 67, 1899 p. 675.

* 659. ” Proc. Roy. Inst. 16, 1901.

” indicator for platinum thermometer. Engng. 72, 1901 p. 644.
” portable circuit for platinum thermometers. Z. Instr. 22, 1902 p. 164.
” electrical methods of measuring temperature. Engng. 77, 1904 p. 402.
Callendar & Griffiths, determination of the boiling point of sulphur. Proc. Roy. Soc. 49, 1891 p. 56.
” ” platinum thermometers. Bull. Soc. él. 16, 1899 p. 410.

* 665. ” ” pyrometer. Iron & Coal. 69, 1904 p. 1958.

Campbell, direct reading resistance thermometers. Phil. Mag. 9, 1905 p. 713.

* 667. Chappuis et Harker. Trav. Mém. Bureau. int. d. P. et M. 12, 1900 p. 18.

Chree, investigations on platinum thermometry. Proc. Roy. Soc. 67, 1901 p. 3.

* 669. Crompton’s direct reading pyrometer. Mar. Eng. 27, 1905 p. 289.

Darwin, alarm device for high or low temperatures. El. Anz. 22, 1905 p. 1285; Engng. 79 p. 329.
Dewar, electric resistance thermometry at the temperature of boiling hydrogen. Proc. Roy. Soc. 73, 1904 p. 244.
Edward Brown & Son, a new electric pyrometer. Clay Worker. 50, 1908 p. 272; Brick. 29, p. 443; Iron A. 82, p. 846.
Fory, comparison of platinum thermometers. Phil. Mag. 50, 1900 p. 421; Éclair. él. 24, p. 80.

* 674. Garden’s electrical thermometer for determining temperatures at a distance. Nat. The. 23, 1881 p. 464.

Haagen, a new quartz-glass resistance thermometer. Z. ang. Chem. 20, 1907 p. 565.
Hartmann & Braun, a new pyrometer. Thonind. 15, 1888 p. 29.
” ” electric pyrometer. Stahl. 12, 1892 p. 656.
” ” pyrometer. Neuzeit. 2, 1893 p. 445.

* 679. ” ” electric resistance thermometer. Schw. Elektrot.-Z. 5, 1908 p. 133.

Heraeus, temperature measurements. Dampfk. Masch. 30, 1907 p. 308.

1. Heusock and Neville, J. Chem. Soc. 1890 p.
Hoadley, the platinum-wire pyrometer. Chem. News. 47, 1883 p. 171.
Holborn u. Henning, Platinthermometer. Ann. d. Phys. 26, 1908 S. 833.
Jäger, Empfindlichkeit der Widerstandsthermometer. Z. Instr. 26, 1906 S. 278.
1. Jäger u. Steinwehr, D. phys. Ges. 1903 S. 353.
„ „ Erhöhung der kalorimetrischen Messgenauigkeit durch Anwendung von Platinthermometern. Z. Instr. 24, 1904 S. 28; Beibl. Chem. Z. 1905, S. 744.
„ „ Anwendung des Platinthermometers bei kalorimetrischen Messungen. Z. Instr. 26, 1906 S. 237.
1. Kammerlingh-Onnes, Vergleich des Platin-Thermometers mit dem H-Thermometer und dem Goldwiderstandsthermometer. Beibl. 1907 S. 772.
Krellsen, Empfindlichkeit der Thermometer. Z. Veriz. 11, 1906 S. 33.
Maercks, Feuerungskontrolle und Dampfkesseluntersuchung. Z. V. d. I. 53, 1909 S. 121.
Meilink, Vergleich des Widerstandes von Gold- und Platindraht. Z. kompr. G. 9, 1906 S. 163.
Mendenhall, differential resistance thermometer. Phil. Mag. 5, 2, 1885 p. 334; Am. Journ. 3, 30, 1885 p. 114; Beibl. 1, 1886 S. 11; Chem. News, 1886 p. 293.
Meyer, Bestimmung tiefer Temperaturen. Elektroch. Z. 5, 1898 S. 6.
Montpellier, thermomètres et pyromètres de Hartmann & Braun. Electr. 20, 1900 p. 97.
1. Priestley, Inst. of Publ. Health Congress. 1901.
Puluj, Telethermometer. Wien. Ak. Ber. 98, 1891 S. 1502.
1. Rosenhain, Electro-Chemist and Metallurgist, v. 1, 1901.
Shaw, appareil mesurant la température par la variation de la résistance électrique. Lum. él. 30, 1888 p. 79; Electr. 21, p. 667.
Siemens, pyromètre électrique. Ann. ind. 14, 1, 1882 p. 372.
Spohr, Abänderung des Siemens’schen Pyrometers. Dingl. 257, 1885 S. 315.
Thiesen, Platinthermometer. Z. Instr. 23, 1903 S. 363.
Travers u. Groyer, Vergleich des Platinthermometers mit dem Normalthermometer zwischen −190° und +440°. Z. phys. Chem. 52, 1905 S. 437; Proc. Roy. Soc. 74, p. 528; Z. kompr. G. 9, S. 43.
1. Wade, Survey Dept. Cairo. 1905.
1. Waidner and Mallory, Phys. Rev. 8, 1899 p.
Whipple, temperature indicator. Sc. Am. 87, 1902 p. 69.
1. „ Proc. Clev. Inst. Eng. Dec. 1904.
„ resistance and radiation pyrometers. Electroch. Ind. 4, 1906, p. 438.
Adjustable electrical thermometer. Iron, 23, 1884 p. 246.

Resistance pyrometer. Erf. 17, 1891, p. 62.
* 710. Electrical pyrometers and telethermometers. Sprechs. 1892, p. 89.
Electrical pyrometers. Electr. 52, 1903, p. 100.
Electrical methods of measuring temperature. Engng. 77, 1904, p. 370.

In addition, Nos. 99, 110, 112, 116, 133, 139, 140, 141, 156, 230, 335, 442, 458, 466, 547, 555, 565, 594, 635, 640, 644, 645.

Instruments Indicating at a Distance

Advance alarm pyrometer. Iron A. 82, 1908, p. 1375.
Chenu, téléthermomètre du Théâtre de la Monnaie. Electr. 8, 1884, p. 535; Ingén. 7, p. 23.
Berthelot, sur les points de fusion de l’argent et de l’or. C. r. 126, 1898, p. 473.
Blanche’s apparatus for the automatic reporting of specified temperatures. Chem. Ztg. 8, 1884, p. 692; Ann. f. Gew. 14, p. 168.
Ferrini, distance indicator for temperatures. Z. Instr. 2, 1882, p. 41.
Grummach, an electrical contact thermometer. Z. Instr. 9, 1889, p. 296.
Hartl, new remote thermometers. Z. V. d. I. 35, 1891, p. 1399.
Marmor, les thermomètres avertisseurs d’Eon. Cosm. 42, 1900, p. 421.
Marting, remote and signal thermometers. El. Anz. 23, 1906, p. 1.
Meyer, electrical telethermometers. El. Anz. 14, 1897, p. 1358.
Moennich, the remote-measurement inductor. Z. Instr. 9, 1889, p. 122; ETZ. 10, p. 278; Hopfen. Z. 29, p. 2280; CBl. Bauw. 11, 1890, p. 21.
Recknagel, remote-measuring and remote-control devices for heating installations. Gesundh. Ing. 31, 1908, p. 83.
Richard, thermomètres avertisseurs métalliques. Nat. 32, 2, 1904, p. 363.
Schultze, remote-measurement thermometer. Uhl. W. T. 1896, 2, p. 68.
Ubrig’s remote thermometer. Verh. V. Gew. Sitz. B. 1889, p. 234.
* 728. Standard thermometer Co. Telemeter system. Sc. Am. 1889, p. 63; El. World. 14, p. 199.
Ward’s distance thermometer. Iron A. 48, 1890, p. 159.
The electrical remote thermometer for central heating. Z. Blechind. 16, 1887, p. 441; Gesundh. Ing. 10, p. 378.
Indicateur de température à distance. Gén. civ. 37, 1900, p. 222.
Electrical remote thermometers. Schw. Elektrot. Z. 5, 1908, p. 616.

In addition, Nos. 555, 630, 696, 710.

Verification, Calibration, and Graduation.

Angot, graduation des thermomètres à alcohol. J. d. phys. 10, 1890 p. 399.

* 734. Barus, Am. Journ. 48, 1894 p. 332.

Berthelot, graduation des couples thermoélectriques. C. r. 134, 1902 p. 983.
Brown, calibration of thermometers. Nostr. M. 29, 1883 p. 1.
Callendar and Quain, platinum thermometry. Phil. Mag. 47, 1899 p. 191.
Crafts, use of thermometers with particular reference to the determination of melting and boiling points. Chem. News. 48, 1883 p. 286; 49, 1884, p. 6, 15; Chem. J. 5, 1883 p. 107. Bull. Soc. chim. 39, 1883 p. 196, 277.
Mendeleev, obtaining high temperatures in laboratories. Phys. Rev. 7, 1906 p. 252.
Galitzine, über das Dalton’sche Gesetz. Ann. d. Phys. 41, 1890 S. 588.
Griffiths, determination of some boiling and freezing points by means of the platinum thermometer. Proc. Roy. Soc. 48, 1891 p. 220; Phil. Trans. A. 182; 1891 p. 43.

* 742. ” Nature, 1895.

* 743. Grützmacher, Wiss. Abh. d. Phys.-T. Reichsanstalt. 3, 1900 S. 248.

” neuere Thermostaten. Mech. Z. 1902 S. 184.
Guillaume, la précision atteinte dans la mesure des températures. Lum. él. 35, 1891 p. 88.
Guttmann, determination of melting points at low temperatures. J. Chem. Soc. 87, 1905 p. 1037.
Heycock and Neville, platinum resistance pyrometers, J. Chem. Soc. 1895 p. 160, 1024.
Holborn, Platinwiderstände und Petrolätherthermometer. Ann. d. Phys. 4, 6, 1901 S. 242, 255.
Holman, calibrating thermometers. Am. Journ. 3, 23, 1882 p. 278; Phil. Mag. 5, 14, 1882 p. 294.

* 750. Jäger u. Gumlich, Wiss. Abh. d. Phys.-T. Reichsanstalt 1, 1894 S. 83.

* 751. Jäger u. Diesselhorst, ” ” ” ” 3, 1900 S. 305.

Johnson, new apparatus to determine the melting points of slags. Electroch. Ind. 4, 1906 p. 262.
Kammerlingh-Onnes, Kryostat für Temperaturen unter −210° C. Z. Instr. 27, 1907 S. 254; Comm. Phys. Lab. Leid 1905.
Ladenburg u. Krügl, Messung tiefer Temperaturen. Ber. Chem. G. 33, 1900 S. 637.
Lindeck u. Rothe, die Prüfung von Thermoelementen für hohe Temperaturen. Z. Instr. 20, 1900 S. 285.
Mahlke, Thermometervergleichungsapparat für Temperaturen zwischen 250—600° C. Z. Instr. 14, 1894 S. 73.

Marie et Marquis, thermostat électrique (Berlemont). Cosm. 1903, 1, p. 387; Dingl. 318, p. 494.
Meyerhoffer and Sanders, fixed point for thermometers, Glauber’s salt and chromic acid. Z. phys. Chem. 27, 1898, p. 367.
Mills, melting and boiling point, as related to chemical composition. Phil. Mag. 5, 17, 1884, p. 173.
Pickering, calibration and standardizing of mercurial thermometers. Phil. Mag. 5, 21, 1886, p. 180.
Pomplun, comparison of thermometers at temperatures above 40° C. Z. Instr. 11, 1891, p. 1.
Richards, transition temperature of sodium sulfate. Z. phys. Chem. 26, 1898, p. 690; Chem. News, 78, p. 229; Am. Journ. 6, p. 201.
Richards and Churchill, use of the transition temperatures of complex systems as fixed points. Chem. News. 79, 1899, p. 149; Z. phys. Chem. 28, p. 313.
Richards and Jackson, new method of calibrating thermometers below 0°. Z. phys. Chem. 56, 1906, p. 362.
Richards and Wells, transition temperature of sodium sulfate. Z. phys. Chem. 43, 1903, p. 465.
Richards and Wells, transition temperature of sodium bromide. Z. phys. Chem. 56, 1906, p. 348.
Rose-Innes, Kelvin’s absolute method of graduating a thermometer. Phil. Mag. 45, 1898, p. 227.
Rothe, thermostat with electrical heating device up to 500° C. Z. Instr. 19, 1899, p. 143.
Rothe, thermostat for low temperatures. Z. Instr. 22, 1902, p. 14.

* 770. Rücker, calibration of mercurial thermometers. Chem. News. 46, 1882, p. 142.

* 771. Russel, Neumann, calibrating thermometers. Am. Journ. 3, 21, 1881, p. 373.

* 772. Schreiber, on testing thermometers below the ice point. Z. Instr. 8, 1888, p. 206; Beibl. 13, 1889, p. 70.

* 773. Thiesen, Scheel and Sell. Wiss. Abh. d. P.-T. Reichsanstalt. 2, 1895, p. 12.

Walter, vessel for comparing thermometers at arbitrary temperatures. Z. Instr. 12, 1892, p. 342.
Watson, instrument for the comparison of thermometers. Phil. Mag. 44, 1897, p. 116.
Watt, boiling temperature of metals. Phys. Review 9, 1908, p. 174.
Whipple, verification of thermometers. Phil. Mag. 5, 21, 1886, p. 27.

* 778. Whipple, verification of thermometers at the freezing point of mercury. Beibl. 11, 1887, p. 813.

Wiebe, official testing of thermometers. Z. Instr. 1, 1886, p. 22.

* 780. — official testing of thermometers. Chem. Z. 15, 1890, p. 66.

— official testing of thermometers. Z. anal. Chem. 30, 1891, p. 1.
— on the vapor pressure of water vapor at temperatures 82°–100°. Z. Instr. 13, 1893, p. 329.

* 783. Wild. Mélanges phys et chim. VIII, St. Petersburg, 1893, p. 263.

Witt, chemical apparatus. Ber. d. Chem. G. 1893, p. 1696.

* 785. Verification of thermometers. Man. Build. 12, 1880, p. 183.

* 786. Dangers from an incorrect thermometer; testing. Brenn. Z. 9, 1880, p. 16.

* 787. Adjustment of mercury thermometers. Hopfen-Z. 23, 1883, p. 86.

* 788. Checking of thermometers. Hopfen-Z. 23, 1883, p. 875.

* 789. Checking of the thermometer. Z. Spiritusind. 6, 1883, p. 852.

* 790. Testing of thermometers. Brenn. Z. 13, 1884, pp. 68, 102; Hopfen-Z. 24, p. 922.

* 791. Testing of thermometers in the Imperial Normal Weights and Measures Commission, Hopfen-Z. 25, 1885, p. 1468.

Heat regulator for calibration offices. Dingl. 262, 1886, p. 218.
Regulation for the testing and certification of thermometers. Z. Instr. 8, 1888, p. 25.

* 794. Testing of thermometers. Z. Rübenz. 22, 1888, p. 47.

* 795. On the testing of glass thermometers. Chem. Z. 12, 1888, p. 1521; Ind. Z. 30, p. 27.

Official testing of thermometers. Z. Instr. 8, 1888, p. 27.

* 797. Official testing and certification of thermometers. Hopfen-Z. 28, 1888, p. 2151.

* 798. Testing and certification of thermometers by the Physico-Technical Imperial Institute. Hann. Gew. Bl. 1888, p. 414.

* 799. Regulations of the Physico-Technical Imperial Institute for the testing and certification of thermometers. Rep. Phys. 25, 1889, p. 322; Gesundh. 14, p. 24.

* 800. Regulations for the testing of thermometers at the Saxon Testing Institute. Rep. Phys. 26, 1891, p. 171.

Testing regulations for thermometers. Z. Instr. 18, 1898, p. 76; ZBl. f. Deutsch. Reich, 26, p. 76.
Testing of thermometers. Centr. O. M. 23, 1902, p. 255.
Activity report of the Physico-Technical Imperial Institute for 1900. Z. Instr. 1901, p. 145.

* 804. Practical advice on the testing of thermometers. Khim. 1902, p. 529.

Government standardisation of electrical pyrometers. El. World. 44, 1904, p. 20.
Die französischen Prüfungsbestimmungen für Thermometer. Mech. Z. 1905, p. 170.
Die englischen Prüfungsbestimmungen für Thermometer. Mech. Z. 1905, p. 187.

In addition, Nos. 101, 116, 135, 136, 139, 140, 141, 142, 193, 211, 257, 272, 289, 305, 532, 549, 565, 589.

Submission history

[v1] 1910

lower reading	upper reading	length of column	correction of interval	correction from \(0^\circ\)C
\(-0.15^\circ\)	\(+50.13^\circ\)	\(50.28^\circ\)	\(-2.52\)	\(-2.52\)
\(+50.80\)	\(99.23\)	\(48.43\)	\(-0.67\)	\(-3.19\)
\(101.44\)	\(148.81\)	\(47.37\)	\(+0.39\)	\(-2.80\)
\(151.72\)	\(198.68\)	\(46.96\)	\(+0.80\)	\(-2.00\)
\(201.81\)	\(248.44\)	\(46.93\)	\(+1.13\)	\(-0.87\)
\(251.74\)	\(298.64\)	\(46.90\)	\(+0.86\)	\(-0.01\)

scale reading	exact correction	transformed correction	correction for the value of the division	final calibration correction
50°	\(-2.52^\circ\)	\(-0.96^\circ\)	\(-0.34^\circ\)	\(-1.30^\circ\)
100	\(-3.18\)	\(\pm 0\)	\(-0.67\)	\(-0.67\)
150	\(-2.82\)	\(+2.01\)	\(-1.00\)	\(+1.01\)
200	\(-2.04\)	\(+4.46\)	\(-1.34\)	\(+3.12\)
250	\(-0.89\)	\(+7.29\)	\(-1.67\)	\(+5.62\)
300	\(0\)	\(+9.85\)	\(-2.01\)	\(+7.84\)

For 100 g of ice, salt is needed	g	temp. °C
potassium sulfate \(K_2SO_4\)	10	— 1.9
sodium carbonate \(Na_2CO_3\) (cryst.)	20	— 2.0
potassium nitrate (saltpeter) \(KNO_3\)	13	— 2.85
potassium chloride \(KCl\)	30	— 10.9
ammonium chloride (sal ammoniac) \(NH_4Cl\)	25	— 15.4
ammonium nitrate \((NH_4)NO_3\)	45	— 16.8
sodium nitrate (saltpeter) \(NaNO_3\)	50	— 17.8
sodium chloride (table salt) \(NaCl\)	33	— 21.3
calcium chloride \(CaCl_2\)	150	— 33.0
(alcohol \(CH_3CH_2OH\)	105	— 30.0)

Name	Symbol	\(t\), °C	\(\varphi\), °C
tin	Sn	\(+232^{0}\)	\(\pm 0.5\)
lead	Pb	327	1.0
zinc	Zn	419	1.0
antimony	Sb	630.5	1.0
silver in air	Ag	955	2.0
silver, pure	"	961.5	1.0
gold	Au	1064	1.0
copper in air	Cu	1065	1.0
copper, pure	"	1084	4.0
nickel	Ni	1482	6.0
palladium	Pd	1587	—
platinum	Pt	1780	5.0

Temperature Measurement

Full Text

Proceedings

of the Tomsk Technological Institute

of Emperor Nicholas II

TEMPERATURE MEASUREMENT.

CONTENTS.

PART I.

Theory and construction of various instruments.

Gas thermometers.

Instruments with liquids.

Temperature Measurement

Instruments based on optical phenomena

Electrical phenomena

Thermoelectric instruments

Preparation of Reports at a Distance

General Practical Instructions

PART II

Verification of Instruments for Measuring Temperatures

Noted Misprints

Electrical Phenomena

Thermoelectric Instruments

Measurement of Temperature

Measurement of Electrical Resistance

Production of Reports at a Distance

General practical instructions.

SECTION II.

Verification of Instruments for Measuring Temperature

Measurement of Temperature

Temperature Measurement

Conclusion.

Appendix 1.

Temperature Measurement

Table 26.

List of Literary Sources.

Articles and books of a general nature, or devoted to the examination of many types of instruments.

Temperature Measurement

Mercury Glass Thermometers.

Glass thermometers with other liquids

Other instruments with liquids.

Expansion of solid bodies.

Calorimetric instruments and methods.

Pyrometers based on melting, and the thermophone.

Optical pyrometers.

Thermoelectric Instruments.

Instruments with a change in electrical resistance.

Temperature Measurement

Instruments Indicating at a Distance

Verification, Calibration, and Graduation.

Submission history

Access Paper

Citation

Share

Related Papers

Feedback