x1 



that of the tube ; in this instance it is as 1 to 500, the diameter 

 of the tube being y% ; it is graduated to the five-hundredth of 

 an inch, and reads as 1000. From this arrangement it is evi- 

 dent that any deviation produced in the surface of the cis- 

 tern, from the rise or faUing of the mercury in the tube, 

 will be inappreciable, and does not amount to the errors of 

 observation. The rain-gauge is also similar to one which I 

 made for Dr. Apjohn at the same period. In superficial area 

 it measures 1000 inches; the rain is collected in a vessel 

 which is graduated into cubic inches, consequently, when 

 one inch by measure is indicated by the graduations, it de- 

 notes that Y(j\ju of an inch of rain has fallen on the surface 

 above, and as the receiver is graduated into cubic inches all 

 the way, it gives at once the decimal, until you come to 1000 

 cubic inches, which is equivalent to 1 inch of rain : this gauge 

 admits of the most simple verification. 



