26 On the EXPANSIVE FORCE 



dining in the leaft from that direction to the horizon. The 

 plug was loft. 



Jan. 2. 1785. — Being colder than 31ft December, in order 

 to haften the efFed, I put a mixture of common fait and fal 

 ammoniac to the water, and tied a long pack-thread, with a 

 piece of red rag at its end, to the fuze, in order to find where 

 it fell in the fnow. This plug made its efcape, like the reft ; 

 for at half paft fix it was flown, and a cylinder of 8^- inches of 

 ice ftanding over the fuze-hole. The plug was loft ; for the 

 red rag appeared no where on the furface of the fnow. 



4., — Tried a plug made with fprings, in the manner of a 

 fearcher, only very fliort and ftrong. Added the freezing mix- 

 ture. The fiiell gave a fudden crack at a quarter after nine^ 

 and inftantly Ihot from its furface two thin plates of ice, re- 

 fembling fins, about 2 inches in the higheft parts. On exa- 

 mining the fliell I found it burft, and the plug forced up about 

 half an inch ; and, on breaking the fliell, the fprings were con- 

 fiderably bent, fo as not to have recovered their firftfituation. 



g. — Repeated the laft experiment, with a fimilar plug and 

 the freezing mixture. It was thrown out, as before, and the 

 projeding icy cylinder was ^^ inches high. 



Similar experiments were afterwards made with all the lefler 

 ftiells ; yet, though one or more of each fort were adlvially 

 burft, more plugs were projeifled than produced that effect. 

 As foon as the fnow began to difappear from the furface, I 

 fearched carefully for the plugs, and found fix of them ; which^ 

 being all marked with notches after the firft experiment, I eafily 

 formed from them the following table. 



Plug, 



