57° 
entire couple were of platinum, platinum-alloy. For 
accurate work it cannot be assumed that variations 
in the temperature of the vicinity of the galvanometer 
are of no importance, Corrections must be applied 
to reduce the readings to the correct values at 0° C. 
In practice it will be found that the simplest way of 
maintaining the cold junction at a constant tempera- 
ture is to use a Dewar vessel or thermos-flask filled 
with oil (see Fig. 1) into which the cold junction of 
the thermo-couple, or of the composite thermo-couple, 
is placed. 
A simple form of potentiometer in conjunction with 
a direct moving-coil galvanometer, has largely in- 
creased the usefulness of the direct-reading instru- 
ments by opening out the scale to any desired extent. 
For the autographic recording of temperature the 
photographic arrangement originally due to Le Chate- 
lier is still the only way of recording very small and 
NATURE 
[JANUARY 15, I9I4 
identical at o° and 100° C., and experiment has shown 
that the formula :— ° 
1—pr=df cea 
100 100} 
when 6 is a factor depending on the purity of the 
wire for making the thermometer, expresses the 
relationship between them in other parts of the scale. 
The t—pt curve being a parabola, it is only neces- 
sary to determine the resistance at three different 
temperatures in order to ascertain the value of 6, and 
thus to standardise the thermometer completely. The 
three temperatures usually employed are 0°, 100°, and 
444-70° C. (the boiling point of sulphur). 
The resistance of the thermometer is usually 
| measured by the ordinary Wheatstone 
Bridge methods, and several instruments 
have been designed for this purpose. 
The methods of measuring temperature to 
a very high degree of precision are outside 
Co 
Fic. 1.—Lhermos-flask cold-junction control. 
rapid changes in temperature. In the majority 
of recorders now in general use the galvanometer 
pointer is depressed intermittently by clockwork, or 
some electrical mechanism on to either an inked 
thread or typewriter ribbon which is pressed on to a 
chart mounted on a rotating drum (clock-driven), the 
resulting record being a series of ink marks. The 
thread recorder (Fig. 2), designed by Mr. Horace 
Darwin, may be taken as a typical example 
of one of these recorders. The action of 
the clockwork in depressing the galvano- 
meter boom on to the paper is so rapid 
that the boom is only under control by 
the chopper-bar for less than two seconds out of the 
minute. The figure illustrates a double recorder in 
which two galvanometers connected to two separate 
thermo-couples are recording on the same drum. 
Resistance Thermometers. 
Sir William Siemens was the first to suggest, in 
1871, that the change in the electrical resistance of a 
wire with temperature might be employed as a means 
of measuring temperatures. In the hands of Cal- 
lendar and Griffiths this has become the most accurate 
method of measuring temperatures up to 1200° C. 
Callendar pointed out that if R, denoted the resist- 
ance of the thermometer coil at o° C. and R, its 
resistance at 100° C., a temperature scale could be 
established for that particular wire which might be 
called the scale of platinum temperatures, such that, 
if R were the resistance of the coil at any tempera- 
ture ¢ on the gas scale, the temperature on the 
platinum scale would be R—R,/R,—R,x100. For 
this quantity he employs the symbol #t, its value 
depending on the sample of platinum chosen. 
In order to reduce the temperatures on the platinum 
scale pt to the gas scale, it is necessary to know the 
law connecting t and pt. They are, of course, 
NO. 2307, VOL. 92] 
INKED 
THREAD MOVED FORWARD 
O'S MM PER MINUTE 
the scope of this paper, but reference to them 
will be found in Dr. Burgess’s book,’ which also 
contains a full bibliography of papers on the subject. 
Both thermo-electric and resistance thermometers 
have a distinct upper limit of temperature beyond 
which they should not be employed. The resistance 
thermometer cannot be used beyond 1200° C., owing 
to the disintegration of the mica frame, and even the 
thermo-couple can rarely be used above 1400° C., 
because of the impossibility of finding a gas-tight 
protecting envelope that will last above this tempera- 
ture. 
The porcelain tubes made by the Royal Berlin 
Porcelain Manufacturing Co. are on the whole the 
most satisfactory. This firm have comparatively re- 
cently introduced a tube made of a new material 
called ‘‘ Marquardt,’ which will resist temperatures 
4) REVOLUTION EVERY 
2 MINUTES 
} REVOLUTION IN 
25 HOURS 
Fic. 2.—Diagram of double-thread recorder, = 
up to 1700° C. (approximately the melting point of 
platinum). Unfortunately tubes made of this mate- 
rial are very brittle, and great care must be taken in 
handling them, especially in allowing them to cool 
slowly. Tubes made of fused quartz are also em- 
ployed, but it will be found that these tubes soon 
devitrify and become brittle if used continuously at a 
temperature of g00° to ro00° C. 
9 “The Measurement of High Temperatures,” pp. 212-218, 470-471. By 
G. K. Burgess and H. Le Chatelier. (London: Chapman and Hall, Ltd.) 
In this connection it may be mentioned that Principal E. H. Griffiths suc- 
cessfully made d fferential temperature measurements between the freezing 
points of two liquids to one-millionth of a degree Centigrade by means; of 
resistance thermometers and a sensitive bridge. 
