FEBRUARY 8, 1901.] 
Fahrenheit hydrometer, or a hydrometer 
of the variable immersion type with the 
diameter of the neck reduced to the least 
practicable size. 
delicacy to tax the manipulative skill of 
the observer, and the investigation is on 
this account worth the doing simply as 
practise work, while the results of a care- 
ful study of the subject would be welcomed 
by any of our standard journals of physics. 
By a similar method, studies may be made 
of the density and coefficient of expansion 
of liquids having low freezing points, such 
as alcohol, ether and carbon disulphide. 
Something has been done in this line, but 
the subject is far from being exhausted. 
By the use of liquid carbon dioxide and 
ether one can readily reach the tempera- 
ture —80° C., and itis probable that the 
time is near when liquid air will be avail- 
able for the extension of such work to much 
lower temperatures. Data upon subjects 
such as these are useful even though no 
startlingly new phenomena be brought to 
light, and the observer has before him the 
possibility of discovering new and important 
relations which may have a bearing upon 
our theories of the nature of matter. The 
verification through this extended range of 
temperatures, viz., from —80°, and ulti- 
mately from —200°, upwards, the law al- 
ready theoretically established by Van der 
Waals * for the relation of the expansion of 
liquids to their critical temperatures; or, 
failing in that, the experimental demon- 
stration of the necessity of a modification 
of the theory, would be in itself ample in- 
centive for the investigation. 
The subject of specific heats at low tem- 
peratures is still awaiting the attention of 
our experimental physicists. H. F. Weber t 
a quarter of a century ago studied boron 
* Van der Waals, ‘Continuitat des gasformingen 
und fliissigen Zustandes.’ Leipzig, 1881. 
{+ H. F. Weber, Poggendorff’s Annalen, 154, pp. 367 
(1875). 
SCIENCE. 
The work is of sufficient - 
205 
and silicon, in this respect, down to —40° 
C. and carbon to —50° C., with most inter- 
esting results, after which comparatively 
little was done until 1898, when Behn * ob- 
tained values for several metals down to 
—200°. The specific heat of a great variety 
of solids and liquids still remains to be de- 
termined through the greater range of 
temperatures now within our reach, and the 
calorimetric observations are no more diffi- 
cult nor elaborate at low than at high tem- 
peratures. Any one who can set up and 
ealibrate a Bunsen ice calorimeter is in 
position to make the measurements. A 
cylinder of carbon dioxide and a can of 
ether will give refrigeration to —80° C. and 
perhaps by the time the work for this range 
of temperatures is completed it may be pos- 
sible to order a gallon of liquid air by tele- 
phone at a reasonable cost and thus readily 
extend the research to —200° C. Tempera- 
ture measurements are best made in such 
work by finding the change of resistance in 
a coil of fine copper wire. 
What I have attempted to point out, in 
a fragmentary way in the case of two or 
three particular problems is true of the 
whole domain of physics. No research is 
ever complete. However exhaustive it 
may at first sight appear, it will, when crit- 
ically considered, be found to afford merely 
a starting point from which to push further 
out into the infinite region of the unknown 
which lies beyond the boundaries of our 
present knowledge. Every theoretical dis- 
cussion is based upon assumptions which 
must be tested experimentally, and such 
tests usually lead to new and more accurate 
knowledge of the properties of matter and 
ultimately to modifications of the theory. 
Thus Poisson, long since, pointed out that 
the numerical value of the ratio for the rela- 
tive contraction of the diameter of astretched 
rod to the elongation would be one-fourth 
for all substances for which the assumption 
* Behn, Wiedemann’s Annalen, 66, pp. 237 (1898). 
