62 Mr. F. D. Brown's Notes on Thermometry. 



the same in five tins, A, B, C, D, E. The thermometer placed 

 in these tins marked l°-30, l°-34, l°-26, l a 30, and l°-27 re- 

 spectively (these numbers are not comparable with the former, 

 as the experiments were made a month or so later, when the 

 zero of the thermometer had altered its position). These ex- 

 periments showed that distilled-water ice gave the same results 

 as ordinary ice, and that the melting-point of different speci- 

 mens of ice, when mixed with distilled water, was the same 

 within 0°'005 C. The exceptionally low reading obtained with 

 the tin E in the first series of experiments was probably due 

 to the fact that the ice, having been made by means of a 

 freezing-mixture, was not at its maximum temperature. 



In subsequent determinations of the zero of thermometers 

 I have always used ordinary block ice, washed and mixed with 

 sufficient distilled water just to fill up the spaces between the 

 pieces, and have not allowed the water to drain away. These 

 results are in accord with those obtained by M. Pernet. 



Zero-movement .% and Substitution of the Determination of the 



Steam-point for that of the Zero-j>iint. 



In considering the well-worn question of the zero-move- 

 ments of thermometers, it is important to distinguish between 

 its practical and theoretical aspects. To make a study of zero- 

 movements from an abstract point of view, to find out equa- 

 tions expressing these movements under different circum- 

 stances and with different thermometers, to learn that when a 

 certain thermometer has been subjected to a certain series of 

 temperatures at certain intervals of time its indications on 

 next changing its temperature will be affected with a certain 

 index-error, may possibly be of some utility, but it does not aid 

 us much in the endeavour to free the readings of thermometers 

 from the errors with which the}' are surrounded. When once 

 we have acquired the information that a thermometer sub- 

 jected only to those changes of temperature Avhich are due to 

 the weather exhibits a gradual rise of zero, that the rise thus 

 taking place in a given time diminishes as the age of the ther- 

 mometer increases, but differs for different thermometers, when 

 Ave also know that a thermometer subjected to a high tempe- 

 rature after a considerable period of rest exhibits a decrease in 

 its zero-reading, dependent on the thermometer itself and also 

 on its previous history,— we know all, or nearly all, that we 

 can put to practical use. 



Thus, for example, the thermometer attached to my standard 

 barometer was verified at Kew Observatory when it was first 

 supplied to me, some four or five years ago. Since then I 

 have from time to time observed its reading in melting ice, 



. 



