Deter rii in iufj the Specific Heat of a Liquid. 105 



The quantity entering into the result is 0^ - 0, and being between 

 30° and 40°C, an error of •02''C in the reading of the tenth ther- 

 mometer would only aftect the result to the extent of -06 per 

 cent.; on the other hand, the quantity 6 - 6^^ is the difterence of 

 two readings of the one-hundredth thermometer in the calori- 

 meter, and its value is approximately 2°C. Reading this ther- 

 mometer with a lens and taking the greatest possible care to 

 avoid any errors from parallax, it is perhaps possible to estimate 

 its readings to 'OOFC, but an error of '001° is very easily made 

 in the reading, and it is very probable that the instrument itself 

 is not really reliable to more than -OOo^C, and this really gives a 

 limit to the accuracy which the apparatus can be relied upon to 

 give, of about "2 per cent. Consequently the arrangements for 

 determining 6^ were considered to be amply accurate. From the 

 table of specimen results given at the end of this paper it will be 

 seen that sets of results were actually obtained that showed a 

 greater degree of accuracy than that claimed above, but these 

 results are obtained under exceptionably favourable circumstances, 

 and serve rather to show the possibilities of this method when all 

 such causes as tend to accentuate instrumental difficulties, viz.. 

 variations of temperature range or masses used, are eliminated, 

 than what can be regularly obtained when such exceptional con- 

 ditions are not available. As an instance of what is here 

 referred to, one example of such a special circumstance may be 

 mentioned. The large hundredth thermometer used in all these 

 experiments has a range of only 6°C, and is fitted with an 

 arrangement for changing the value of its zero by adding mercury 

 to the column in the stem or by removing some. Now one of 

 the greatest difficulties in the use of such a finely-divided instru- 

 ment is the " sticking " of the mercury and the consequent 

 motion of the thermometer by "jei-ks." To avoid errors due to 

 this cause, the water-motor driving the stirrer was so arranged 

 as to keep the thermometer in a slight tremor ; this is not a good 

 condition for taking an accurate reading, and the motor had 

 always to be stopped at the instant of taking the reading. This 

 left room for a small amount of sticking, and it was found that 

 at some times this would be very much more marked than at 

 others, depending apparently upon what particular portion of the 

 mercury was in the stem. This alone accounts for the fact that 



