New Method for obtaining a Constant Temperature. 89 



consists. Are we, then, any nearer an explanation of the 

 phenomena than in the older theory, seeing it does not appear 

 that there is anything in this form of experiment which 

 supports the free ion theory in preference to the older one? 



Prof. Ostwald here states that zinc [and hy implication any 

 other metal acted on by an electrolyte] becomes negative to 

 the electrolyte, while, according to his experiments with 

 mercury-dropping electrodes*, the metals were considered 

 positive. In a criticism of these mercury-dropping experi- 

 ments already published f , I pointed out that the latter con- 

 clusion, in the case of mercury when thus acted on, was 

 probably incorrect, and that, if so, the hypothesis on which 

 the results were explained was erroneous. It would be 

 interesting to know if the two theories can be reconciled. 



In conclusion it will, I hope, be apprehended that the above 

 criticisms are offered, not to prove or disprove any particular 

 theory, but rather to point out that in support of certain 

 theories there is being advanced a quantity of evidence that 

 appears entirely indecisive. So long as no experimentum 

 crucis can be pointed out, and all the theories explain known 

 experimental results, we advance no nearer a real knowledge 

 of the truth, and it seems a pity (supposing the dissociation 

 theory contain any truth) to overweight it at present with 

 evidence the inapplicability of which must be continually dis- 

 couraging to the student, 



IX. On a New Method for obtaining a Constant Temperature. 

 By Mr. Henky Crew, Instructor in Physics in Haverford 

 College %. 



ri^HE following work was suggested by an attempt on the 

 \ part of the writer to determine the coefficient of ex- 

 pansion of water by the areometric method of Matthiessen, 

 using, instead of a solid piece of glass, the hollow glass bulb 

 of a weight thermometer. 



It was very soon found, however, that the errors introduced 

 by temperature-variation of the water in which the bulb was 

 weighed far surpassed all other errors involved. 



The problem which must be solved, before this or any 

 other method yet devised is available for accurate work, is to 

 produce, throughout a certain limited space, a constant tem- 

 perature, and to maintain this temperature for a time sufficient 



* Zeitschrift fur physikalischc Chemie, i. p. 583, 1887. 

 t Phil. Mag." [5] xxvii. p. 390, 1889. 

 % Communicated by the Author. 



