Description of an improved Rain Gage. 331 
ual result, and that which is given by the tables: thus if the 
area of the gage be fifty square inches, as this is the half of one 
hundred, we must take half the sum of the tabular heights 
for the true altitude. 
It is not necessary to be very particular in the choice of 
a balance ; a pair of good common scales will answer, with 
true weights, either troy or avoirdupois. e gage may 
be made of tin, or sheet iron painted or japanned, but cop- 
per is more durable. The area of the funnel, and that of 
the top of the body part, are the only parts that need atten- 
tion in the construction. ese ought to be made tolerably 
exact. A strong hoop should be fixed around them on the 
outside to preserve their figure true. 
In every operation of weighing, the weight of the gage, 
moistened in the inside, must be deducted from the gross 
weight; the remainder is the corrected weightof the water 
with which the tables must be entered. © ¥ Lea? 
In the case of hail, snow, sleet, or frozen water, being in 
the gage, it isnot necessary to melt its contents into water, 
as. the changes effected by temperature, and _ pressure 
make no difference in the weight, ’ 
The use of scales and weights may be dispensed with, by 
substituting a steelyard, so constructed that the moveable 
weight on its arm might indicate by its positon, 3 the 
weight, but the inches and decimal parts of its correspond- 
ing altitude, without reference to the tables, and without 
calculation. : aE 
The advantages of this method of finding the quantity of 
Tain, in linear inches of altitude, will be appreciated by ad- 
Verting to the circumstance of our having a zangible quantity, 
9s an unerring guide to that which is nearly imperceptible. 
Twenty five grains and a half,a sensible quantity in a : 
balance, pointing out the difficultly visible division of the 
tesa part of an inch. Su the problem rev ersed; 
that the cubical contents of the water, or its weight, were 
required, from the observed attitude. The chances of ere 
tor would all be against the accuracy of such aga gt 
tion. The difficulties of the task, independently of th 
aforenientioned causes of variation, would evidently be in- 
surmountable. iit 
i had gage constructed on this principle, twelve or 
fourteen years ago, for my friend Dr. Akerly, who informe 
