between Metallic Masses having different Temperatures. 449 



We have already dismissed the supposition that it has any con- 

 nection with the passage of air through that groove, and referred 

 the effect solely to the actually observed increase of velocity in 

 the oscillations ; it still remains to explain this result. After a 

 very careful consideration of the phenomena, I am disposed to 

 account for it entirely upon the diminution of the surfaces in 

 contact. It may at first sight be thought that the adhesion of 

 two metallic surfaces must be too small to influence sensibly the 

 time of an oscillation ; when the enormous velocity of these 

 oscillations is however considered, there can be no room for 

 astonishment. We have shewn that there are frequently more 

 than five hundred contacts and separations in a single second. 

 The most minute adhesion must therefore clog the energy of the 

 impulse in a way nearly insensible to our ordinary modes of im- 

 pression : yet cases are not awanting where such adhesion is 

 abundantly sensible, and especially when a metal so soft as lead 

 is one of those employed. It is not difficult to perceive how the 

 position of the groove or separation of pieces (for we have seen 

 that the effect is absolutely independent of the form of the 

 groove, provided the contact of the bar and block for a certain 

 space be avoided) is the most favourable for producing the vi- 

 brations. The separation of surfaces may either be in the block 

 as Fig. 7, or in the bar, as indicated by the dotted lines at a in 

 the same figure: the surface of contact will thus be reduced, as 

 there shewn, to about one half. If, instead of this, the space be- 

 tween the solid angles b and c had been reduced to one-half, the 

 stability of the bar would have been materially changed, and the 

 requisite distance between the points d'appui for producing an 

 active vibration would have been deranged. By cutting out the 

 interior space of contact, the other conditions remain unimpaired, 

 and the adhesion is diminished to almost any required extent ; 

 in fact, the note has been most clear and steady, when the two 

 points of contact of the block had almost the whole intermediate 



