Dec. 5, 1889] 



NATURE 



lOI 



It is natural for a '• Traits pratique " to refer mainly to 

 the mercurial thermometer : for the great majority of 

 practical thermometric measurements lie within its scope. 

 Having a range from - 40' to at least 360' C, and a 

 possible sensitiveness of about o^'ooi, it rarely has to be 

 exchanged for more delicate or larger-scaled appliances. 

 Even the air thermometer— a sort of general appeal court 

 in measurements of heat — is always accompanied by a 

 number of ancillary mercurial thermometers. 



To begin at the beginning (which, by the way, the 

 author has not done), a thermometer has to be made ; 

 and the method of making it has a serious influence on 

 the result. One maker will overheat his glass, and thus 

 make the bulb harder than the stem ; another will leave 

 irregularities in the bulb which will cause the zero to rise 

 irregularly ; a third can never perfectly " deprive," as it 

 is termed, the stem of air ; the breath of a fourth is con- 

 stantly leaving fatty matter in the capillary tube. In 

 short, there are endless variations in technique, to which, 

 for delicate instruments, attention should be drawn. 



The division of the thermometer is, as might have been 

 expected, well described ; and minute details of calibra- 

 tion (chiefly by the method of broken threads) are duly 

 set forth. Then follows a notice of a less familiar 

 correction — that, namely, which depends on internal 

 pressure when the thermometer is in a vertical position, 

 and that which is produced by the (external) pressure of 

 the air. Two methods of ascertaining the thickness of the 

 bulb are given, but they are both inferior to Stokes's, 

 which turns upon measuring angularly the distance be- 

 tween a spot on the outside of the glass and its reflec- 

 tion from the inner surface. Then ensues a description 

 of the usual apparatus for determining the zero (which 

 M. Guillaume seems to read somewhat too soon after 

 immersing the bulb in the bath) ; and the method of 

 ascertaining the boiling-point of water accompanies this. 

 In the comparison of thermometers, which is next treated 

 the present writer prefers an air current to the metal 

 plunger figured on p. 125. 



If we observe the zero of a thermometer soon after 

 manufacture, and subsequently at frequent intervals, we 

 shall find that it is continually rising. The late Dr. Joule 

 observed this ascent in one of his thermometers for more 

 than seven-and-twenty years. There can be no doubt 

 that it is due to a kind of setting of harder silicates in 

 presence of softer or more viscous silicates in the mixture 

 of which the bulb is composed. The softer glasses show 

 it more than the harder ones ; but in all exact work, it has 

 to be determined and allowed for. This variation takes 

 place at the ordinary temperature. If now we heat the 

 thermometer moderately (say to loo') and cool it, we 

 shall notice a temporary depression due to a tem- 

 porary set. If, again, as is often the case in factory 

 work, we heat the thermometer for a long time to a high 

 temperature (say ISC'") the glass of the bulb (especially if 

 soft) will become sensibly more plastic ; and will some- 

 times yield sufficiently to external pressure to cause an 

 ascent of 6°. At higher temperatures the ascent is still 

 greater. Measurements of zero are therefore exceedingly 

 important, even for moderately accurate work, and the 

 author does not fail to draw minute attention to them. 

 We should have been glad if at this point he had said 

 something about the form of thermometer bulbs. Bulbs, 



for instance, which have their sides concave vary readily 

 in capacity with barometric changes. 



The exposure correction has exercised the minds of 

 physicists for a great many years. When the bulb but 

 not the stem of a thermometer is in a bath, the stem 

 may clearly have a different temperature from the bulb, 

 and the reading as a whole will be too low. In most 

 chemical and physical laboratories, it is usual to follow 

 Regnault, and to add, to the otherwise corrected reading 

 T., the quantity 



a(T - /)N. 



(N is the length in degrees of the exposed column, / is 

 its mean temperature, and a is the difference between 

 the expansion coefficients of glass and mercury.) There 

 can be no doubt that this correction gives too low a result 

 at high temperatures. It has been shown that if instead 

 of a we simply write (a + PN)— calculating a and p from 

 the results- the demands of experiment are fulfilled with 

 all desirable accuracy. The author, however, is disposed 

 to leave the reader pretty much to his own devices for 

 this correction. 



The remainder of M. Guillaume's work is chiefly de- 

 voted to the comparison of the mercurial with the gas 

 thermometer, and the measurement of dilatation of solid 

 bodies : there are some valuable tables at the end. 



A perusal of this " Traitd pratique " will perhaps cause 

 some regret that in most of our measurements of 

 temperature we should be obliged to employ a material 

 that is constantly undergoing physical change, and that 

 necessitates in instruments constructed of it so many 

 corrections. It is, on the other hand, a fortunate circum- 

 stance that we have in the mercurial thermometer an 

 admirable means of establishing and measuring the 

 corrections necessary to be imposed wherever glass is 

 accurately worked with. For it cannot be too em- 

 phatically pointed out that every lens, cylinder, flask, or 

 other glass instrument we employ is more or less 

 amenable to these corrections. M. Guillaume's work, 

 therefore, should command, as it deserves to command, 

 a very wide interest. Edmund J. Mills. 



THE FA UNA OF BRITISH INDIA. 

 The Fauna of British India, including Ceylon and 

 Burma. Edited by W. T. Blanford. ' Vol. I. Fishes. 

 By Francis Day. Pp. 548 ; 164 Figs. (London : 

 Taylor and Francis, 1889.) 



THE first volume of this, the last work of the well- 

 known Indian ichthyologist, Francis Day, was 

 issued under particularly painful circumstances, viz. 

 almost on the very day of the author's death. The 

 state of Mr. Day's health during the last few months 

 had prevented him from attending to the correction of 

 the proofs beyond the middle of this volume, which deals 

 with the Chondropterygians, the Physostome9-,-and the 

 Acanthopterygian family Percidce; and the task of car- 

 rying the remainder through the press has fallen on 

 the editor. This work is but a condensation of the 

 author's quarto " Fishes of India," completed in 1878, so 

 valuable for the copious and beautifully-executed litho- 

 graphic plates which accompany it. And, fortunately, a 

 number of these excellent illustrations (one for every 



