284 Mr Adam'*s Description and Explanation 



tance between the divisions marking ^j^xh of a cubic inch is 

 equal to about Jth of an inch, and by inspection he can easily 

 observe to ^ih of this distance, or the depth of rain falling around 

 the rain gage to the ten-thousandth part of an inch. But a 

 smaller measure, with larger divisions, is occasionally used for 

 minute quantities of rain. The depth of rain is therefore en- 

 tered in the register in inches, and four decimals of an inch. 

 The first two decimals being obtained from the cubic inches, 

 and the last two from the tenths and decimals of the tenth of a 

 cubic inch, on the graduated measure. Thus a quantity of rain 

 measuring 17 cubic inches ^''oths and yg^dthsof a cubic inch 

 in the graduated measure, is marked .1779 of an inch deep 

 in the register: and a quantity of rain measuring only yl^dths 

 of a cubic inch in the graduated measure, is marked .0002, or 

 2 ten-thousandth parts of an inch in the register. This latter 

 portion is an extremely minute quantity of rain. But there 

 are many such minute entries in the register, varying from 

 1 to 12 ten-thousandth parts of an inch. And it is remarkable, 

 that many of them are noted as miniature showers, or depositions 

 of dew, from the air enclosed in the bottle ; these depositions 

 having been observed to vary in quantity with the variations of 

 sky and temperature, and to take place frequently at periods 

 when the sky was clear, and always, when thus entered, at times 

 when no rain fell from the external air. These entries may 

 therefore, it is thought, be considered as measurements of mi- 

 niature showers, or dew, deposited by the air contained in the 

 bottle, under the varying temperatures and states of weather 

 noted in the Register. 



In order to graduate the cylindrical glass measure K L with 

 the utmost accuracy, Mr Adam constructed two cubic tin-plate 

 measures A and B, the one upon a cubic inch of hard wood, and 

 the other upon a similar cube made exactly equal to the ^-'gth 



part of a cubic inch, each of its sides being = sj ^q- = v^lOO 



= 4-6415 tenths of an inch, so that, when both measures were 

 carefully levelled upon a plane table ab, supported by three 

 screws c, d, ^, ten fills of the latter filled the former so exactly, 

 that a single drop added to, or subtracted from, the amount, 

 caused the surface of the water in the cubic inch measure A to 



