156 Intelligence and Miscellaneous Articles. 



as easy to resolve. This spectrum having the appearance of a co- 

 lonnade, I may add that, starting from the small column near D in 

 the yellow, three magnificent columns are found on the side towards 

 the red, and five on the side towards the violet, all of them resolvable 

 into finer lines — in all at least nine columns. It is one of the most 

 curious spectra which the observation of the heavens presents. I may 

 add that the variable stars (except Algol) belong to this type — 

 Comptes Rendus, December 9, 1867. 



ON AN IMPROVED METHOD OF DIVIDING ALCOHOL AND OTHER 

 THERMOMETERS. BY WILLIAM ACKLAND, ESQ. 



In the manufacture of thermometers for low temperatures, con- 

 siderable difficulties are experienced owing to equal increments of 

 temperature not causing an equal increase of volume of the fluid 

 employed ; and this unequal expansion gives rise to considerable 

 errors even within limited ranges of temperature. 



If, in constructing an alcohol thermometer, we select a tube 

 with uniform bore, we find, on giving it certain fixed points, cor- 

 responding to an equal number of degrees between each point, that 

 the spaces so obtained are each diminishing in length as we de- 

 scend ; hence the ordinary mode of equal divisions between the 

 ascertained points is found to give inaccurate results. An im- 

 portant question therefore arises, — Can we so subdivide our tube 

 that each subdivision shall be in its proper place ? 



During the autumn of last year, Mr. Grlaisher mentioned this 

 subject to me, and at the same time offered some valuable sugges- 

 tions to remove this difficulty, and caused me to direct my atten- 

 tion to this subject. Many subsequent interviews with that 

 gentleman, extending over some months, enabled me to mature 

 rny first idea, and aided by his valuable assistance and counsel, the 

 success aimed at has been attained ; and I am now in a position 

 to state that I have completed a dividing- engine on an entirely 

 new principle, which may strictly be termed "Interpolating;" 

 for, whilst between any consecutive terms taken at regular 

 and fixed intervals we can by numerical interpolation find 

 intermediate terms conformable to the law of the series ; so, be- 

 tween fixed and determined points of our thermometer, we can, 

 by using the " interpolating-engine," so divide the intervening 

 spaces that they shall conform to the varying nature of the series, 

 or, in other words, that the divisions shall increase in magnitude 

 as we ascend the scale in such proportions as are demanded by 

 the expansion of the alcohol and the bore of the tube employed. 



To illustrate this new method I have invented, let us suppose 

 we have ascertained the fixed points, P, Q, R, S, of a thermometer 

 corresponding to equal differences of temperature. And let 



J? Q K S 



It is desired to divide the intervening spaces each into ten divi- 

 sions, so that each division shall occupy its true position. 



