O F S I L K. 269 



fliake off its fupeifliious molflure, which 

 will be further promoted by the preffure 

 which it fuffers in pafTing over each pully -, 

 this prefTure will at the lame time help to 

 make it compa6l ; and the groove of the 

 pullies being the angle of an equilateral 

 triangle will throw it into a figure which 

 by the time it arrives at the reel will be 

 fufficiently round : yet this roundnefs is 

 not only to the grooves moulding the 

 thread into a compa<5l figure, but to a 

 caufe lefs apparent which is this : the plane 

 of the pully fliould not be exadlly parallel 

 to that of the thread by which means the 

 thread will always roll a little on the fide 

 of the groove, and fo will receive a fmall 

 twift as it approaches to, and as it leaves 

 each pully. 



That this theory is alfo true in fa6l, you 

 may eafily prove by itretching an untwif- 

 ted filk-thread on two pinsj if you then 

 rbll one of the above-mentioned pullies 

 along it, holding its axle between your 

 finger and thumb, you will perceive that 

 any fmall particle of down, which you 

 may fallen to the thread, will begin to 

 turn round as the pully draws near it, 

 provided you give the pully a very fmall 



in- 



