HARPSICHORD. 187 
years, fubject to no variations but fuch as may be occafi- 
oned by alterations in the ftate of the air, to which all 
known fubftances are more or lefs liable.. 
————_ —_——_ 
« 
ra. _ IN the beginning of laft winter, I had the honour to 
- lay before the fociety an improved method of quilling 
a Harpsicuorb. . Withing to bring my difcovery to the 
teft of full experiment and to the judgment of abler critics, 
I forwarded a defcription and a model of my improvement. 
toa friend in London, requefting that it might be fubmitted 
to the examination of proper judges, and direCting, in 
cafe it fhould be approved of, that an inftrument made by 
one of the firft artifts and quilled according to my pro- 
pofed method, fhould be fent to me. I have accordingly 
received an excellent double harpfichord, made by Mefirs 
Shudi and Broadwood of London, and quilled according 
to my method; with this difference, I had rounded off the 
top of the tongue, and bending the quill over it, kept itina 
horizontal pofition by means of a fmall wire ftaple; as 
will be more fully underftood by referring to my former 
defcription. But Mr. Broadwood has left the tongue of 
its full length and ufual form: But made the hole, in 
which the quill is commonly fixed tight, fo large, that the 
quill has free room to play therein; and then fixing the 
quill below, has bent it round and brought it through 
this hole; which renders a ftaple unneceflary; the top of 
the tongue an{fwering the fame purpofe. ‘The principle 
on which the improvement depends is the fame in both; 
but his is the beft method of executing it. 
He informs, however, that’one inconvenience occurs 
viz. the quills being fo forcibly bent in the curved part, 
are liable, in fome inftances, to fpring back, and fo be- 
come not only too fhort to reach the ftring it fhould ftrike, 
but the projection of the curve will be apt to touch the 
firing behind it, when the ftop is pufhed: back. 
To 
