Ip5 Mr. Thompson's Experiments 
This apparent deficiency remains now to be accounted for \ 
and, firft, it cannot be fuppofed, that it arofe from any imper- 
fection in Mr. robins’s method of determining the velocities of 
bullets ; for that method is founded upon fuch principles as leave 
no room to doubt of its accuracy ; and the pra&ical errors that 
occur in making the experiments, and which cannot be intirely 
prevented, or exactly compen fated, are in general fo fmall, that 
the difference of the velocities in queftion cannot be attributed 
to them. It is true, the effedt of thofe errors is more likely 
to appear in experiments made under fuch circumftances a9 
thofe under which the experiments we are now fpeaking of 
were made, than in any other cafe ; for the bullets being very 
light, the arc of the afcent of the pendulum was but fmall, 
and a fmall miftake in meafuring the chord upon the ribbon 
would have produced a very confiderable error in computing the 
velocity of the bullet : thus, a difference of one tenth of an 
Inch, more or lefs, upon the ribbon in the 85th experiment, 
would have made a difference in the velocity of more than 1 20 
feet in a fecond. But independent of the pains that were 
taken to prevent miftakes, the ftriking agreement of the ve- 
locities determined by the two methods in the experiments 
which immediately follow, as alfo in all other cafes where they 
could be compared, affords abundant reafon to conclude, that 
the errors arififtg from thofe caufes were in no inftance very 
confiderable. 
But if both methods of afceftaming the velocities of bullets 
are to be relied on, then the difference of the velocities, as de- 
termined by them in thefe experiments, can only be accounted 
for by fuppofing that it arofe from their having been diminiflied 
by the refiftance of the air in the paffage of the bullets from 
the mouth of the piece to the pendulum ; and this fufpicion 
wall 
