414 Mr. A. J. Ellis. On Influence of Temperature [Feb. 3, 



1° F. (ibid., p. 523). By adding or subtracting the mean of the 

 observed beats from this calculated pitch of the fork, the pitches of 

 the harmonium reeds at those temperatures were determined, and the 

 mean of all the determinations for each reed was taken. In all cases 

 I calculated to two places of decimals, but the second place cannot be 

 depended on when counting the beats ; and as the result is consider- 

 ably affected by the second place, the process is not so satisfactory as 

 could be wished, and must be regarded as only preliminary. The loss 

 of pitch in proceeding from the lower to the higher temperatures thus 

 determined was divided by the number of vibrations at the lower tem- 

 perature, and also by the number of degrees F. of difference of tempe- 

 rature. The result or coefficient of temperature, being the alteration 

 for 1 vibration and 1° F., would serve to reduce one pitch to the other, 

 on the supposition, which cannot be more than approximately correct, 

 that the alteration for temperature is uniform, and is the same for 

 reeds of very different pitches and makes. Such a coefficient is, how- 

 ever, clearly better than none at all, and is especially useful in deter- 

 mining pitch by Appunn's instrument. 



In the following table the number of the reed is that marked on 

 Appunn's treble tonometer (ibid., p. 527). The "pitch" means the 

 number of double vibrations in a second made by the reed, on the 

 three occasions of observation already mentioned, and as the tempe- 

 rature was variable during the first observations, made on different 

 days, though constant on the same day, this is annexed in a separate 

 column. As harmonium reeds are subject to rather sudden small 

 alterations from causes not yet investigated, it is not possible to be 

 perfectly sure that all the reeds would have shown precisely the same 

 pitch at the same temperature for observations made at intervals of 

 more than nine months. I am inclined to think that reeds 22 and 23 

 must have so altered. In other observations on reed 22, made 14th 

 July, 1880, at 71°"5F., I obtained practically the same results as here, 

 differing from those of all the other reeds. The observations on that 

 day were not sufficiently numerous nor exact to be here recorded, but 

 they agree very well with those now given. The flattening of the 

 reeds for each increase of temperature is quite unmistakable, even in 

 the passage from pitch II to III, with a difference of only 6° F. The 

 pitch III of reed 26 is certainly a bad observation, as the results of 

 the determinations by the two forks differed much more than usual, 

 and it should, therefore, be thrown out of consideration. Altogether 

 for such a small difference as 6° F. the observations could not be made 

 with sufficient accuracy to secure trustworthy results. Even for the 

 greatest difference of temperature, 26° F., the difference of pitch never 

 amounts to so much as *9 vibration in a second for the reeds observed. 



The three last columns give the coefficients of temperature arising 

 from comparing these three pitches two and two, namely, I and II, 



