ean LP) Sie iB 
NATURE 
[Wov. 24, 1881 
; 
THE PRESSURE ERRORS OF THE 
“ CHALLENGER" THERMOMETERS} 
i 
I. The Pressure-Corrections supplied to the *‘ Challenger” along 
with the Thermometers 
HEN I was first asked to examine the thermometers I 
judged, from the appearance and nature of the protection 
over the bulbs, that very slight corrections only would be re- 
quired, even for the greatest pressures to which they had been 
exposed. But Sir Wyville Thomson told me that a correction 
of at least half a degree Fahr. had been assigned for them for 
every mile under the sea. This correction had been given him 
by Capt. Davis of the Admiralty, who had in his experiments 2 
the assistance and advice of such exceedingly able experimenters 
as the late Prof. W. Allen Miller and others. 
Hence, although it appeared to me at first sight incredible that 
any such correction should be required for thermometers with 
protected bulbs, I considered it absolutely necessary to try Capt. 
Davis’ experiments over again, under the same conditions as 
those which he had adopted in conjunction with Prof. Miller. 
My object was, of course, to find out whether I could again 
obtain these results, and, if I could obtain them, to discover 
what were the causes which led to their being so exceedingly 
different from what I should have expected. I felt assured that 
the results were much too large ;—and I had therefore, if ] could 
reproduce them, to trace the various possible causes of divergence 
between the re-ults of experiments conducted in a hydrostatic 
press and of other similar experiments mace at the same pressures 
in the deep sea. 
Half-a-degree Fahrenheit per mile of depth may seem to be a 
matter of very little consequence ; but when we recollect that 
some of the Challenger soundings were made at depths nearly 
appreaching :ix miles, we find that we have sometimes to deal 
with a correction of 3° F., enough to modify seriously our 
- theories of ocean circulation, For it can never be too strongly 
impressed on the student of science that there is no such thing as 
greatness or smallness in itself ; what is very small relatively to 
one class of quantities may be very great relatively to another 
and different one. All the temperature differences, except rear 
the surface of the sea, though important in their consequences, 
are very small relatively to differences of temperature in the 
atmosphere ; but, ju-t because they are so small, small errors in 
the determination of their values are important :—thus it was 
imperative to decide whether the corrections assigned by Capt. 
Davis are necessary. 
At first sight one might think that by far the best way of con- | 
ducting an inquiry of the kind would be to carry it out under 
circumstances nearly the same as those of the Challenger obser- 
vations. No doubt, if we had at hand a coal-pit or mine-shaft 
full of water, and of six miles or so in depth, we might make 
the experiments without the aid of presses, and under circum- 
stances far more favourable than those in which I was obliged 
to operate. The reasons for this statement will appear presently. 
There are great objections to making test-observations at sea, 
The Challenger observations themselves had, of course, to be 
made at sea, but to make under similar circumstances experi- 
ments for the purpose of determining corrections would be a 
perfectly hopeless attempt. The circumstances under which 
thermometers are let down and drawn up again at sea are ex- 
tremely unfavourable to accuracy of observation, I had, there- 
fore, to content myself with such conditions as could be procured 
by means of hydrostatic presses. 
II. Construction of the Thermometers.—I will now say a word 
or two about the construction of the thermometers themselves ; 
and I shall thus have an opportunity of pointing out some of the 
peculiarities of construction to which I have traced the greater 
part of the very large effects obtained by Capt. Davis, and given 
by him as corrections which required to be made. 
The Challenger thermometers are all of the Six pattern: there 
is a highly expansible liquid in the large bulb, which projects to 
a certain extent into the narrow U-tube. Then there isa column 
of mercury occupying the bend of the U and part of each stem. 
Above that, on the maximum side, there is some more of the 
* By Prof. Tait. Abridged by the Author from a forthcoming volume of 
the Reports of the Voyage of H.M.S. Challenger, by permission of the 
Lords Commissioners of H.M. Treasury. 
# “On Deep-Sea Thermometers,” by Capt. J. E. Davis, R.N. (Proceed- 
ings of the Meteorological Society, April 1871). 
| filled with alcohol. 
sensitive liquid ; and at the ends of the mercury column are the 4 
maximum and minimum indices, each containing a piece of steel, — 
so tbat they can be set by means of an external magnet. The © 
large bulb on which the temperature effects are mainly procuced 
is protected by an exterior shell of glass strong enough to resist © 
a pressure of at least 5000 fathoms of sea-water; that is to say, 
approximately, somewhere about six tons weight per square inch, — 
This external shell is nearly filled with alcohol. The main dif- 
ference between this and the first invented form of protected 
thermometer, which (so far as I know) was introduced by Sir 
William Thomson,? is simply that the bulb only is protected, the 
stem being exposed, and therefore the effects produced directly 
by compression are due solely to the stem of the instrument: 
unless, indeed, there be a strain produced on the protected bulb 
(altering its volume) by the wry-neckedness of the protecting 
shell, 
Now, as arule, till quite recently, practical workers in glass 
supposed that no effect at all would be produced by pressure 
upon an ordinary thermometer stem, simply because the external 
diameter is so much greater than the internal; and, in fact, so 
little was the nature of the effects of hydrostatic pressure known 
to practical glass-blowers that one of Mr. Casella’s workmen 
undertook in 1869 to furnish Capt. Davis with thermometers 
whose bulbs should he so thick as to ‘‘ defy compression”! It 
will be seen presently that :uch an idea is entirely absurd :— 
that, however thick is an unprotected thermometer, it will still 
have its indications altered by compression, and very nearly as 
much as a thinner one, unless that be extremely thin. So far as 
the Challenger instruments are concerned, the only effect that 
can be expected to be preduced directly by pressure is the 
diminution of the bore and length of the narrow tube, and the 
consequent forcing of the liquid which occupies it to fill a greater 
length in it. I made at starting a rough calculation of the 
amount of effect of this kind which was to be expected ; taking 
average data as to the compressibility and rigidity of glass. I found 
it to be a small fraction only of a degree for each ton-weight of 
pressure, except on those thermometers which had very short 
degrees. It was clear to me, therefore, that (unless the wry- 
neckedness already mentioned was the cause) the larger part of 
Capt. Davis’ result was not due to pressure directly. 
Ill. Wholly Protected Instruments. Their Defect,—For the 
purpose of comparison with the Chad/enger instruments, so far 
as regards the effect on the unprotected stem, Sir Wyville 
Thomson sent me two mercury thermometers constructed after 
Sir William Thomson’s device. In these instruments the whole, 
bulb and stem alike, is inclosed in a strong glass tube, nearly 
The effects of pres: ure on these instruments 
were very much smaller than on the thermometers of the 
Challenger. This re ult was so unexpected that I at first thought 
it due to defects in the new instruments. But, as will be seen 
later, it is quite consistent with the final result of my investiga- 
tions, It is, however, very difficult to obtain good results from 
these instruments under the circumstances in which I was work- 
ing. Their recording adjustment is constructed on a new plan, 
in which a little portion of mercury is detached from the rest ; 
and separated from it by a small quantity of air, which does not 
move it until compressed to a definite amount. To set the index 
before an observation, the instrument has to be swung round 
somewhat sharply at arm’s length. It was scarcely ever possible 
under these circumstances to adjust it to the temperature of the 
water in the press. The indices in the Challenger thermometers, 
on the other band, consist each of a piece of enamel with a 
couple of hairs attached to it so as to fix itself in the tube and 
retain a record of the observation. They have also a little piece 
of needle inside, and can thus be moved from the exterior by 
means of a horse-shoe magnet, so that the adjustment can be 
made at pleasure, and without any alteration of the temperature. 
The thermometers are plunged for some hours in the water in 
the press, and the indices are set in an instant while the instru- 
ment is partially lifted out for the purpose. With the other 
instruments one might spend days before he could get them 
1 “ The Effect cf Pressure in Lowering the Freezing-point cf Water expe- 
rimentally demonstrated,” by Prof. W. Thomson (Proc. R.S.£., February 
1850). See also the paper by Parrot (1833) quoted below. In this a pro- 
tected thermometer was undoubtedly employed; but the protecting sheath 
was part of the wall of the compression apparatus and was not ettached to 
the thermometer itself. From a reference in this paper I was led to consult 
Lenz’ observations on deep-sea temperatures. He appears to have measured 
these temperatures by bringing to the surface, with great care, a considerable 
quantity cf water from each depth. There was a thermometer in the col- 
lecting apparatus, with a bulb of extra thickness; but no recording index 
was employed, so as to show what was its indication under pressure. 
