>v. 



i f* 



Sect. IX. 1 



METEOROLOGY. 



281 



tliermometer he employs should be eonipafed with it, so 

 us not to leave a doubt as to the amount of their constant 

 diffi^rence exceedino- a tenth of a degree. To n::ke such 

 comparisons, long rest, in contact, in a box staffed with 

 cotton, allowing only tbp portions of the scales where 

 read off to be from time to time uncovered for that 

 purpose, is the best mode of insuring their perfect identity 

 of temperature. If, in any instance, the zeros differ in 

 different parts of the scales, a table of reduction to the 

 standard will require to be constructed. The com 

 parisons should be repeated at not very distant interval 

 of time, especially in the case of self-registering " 

 mometers, whose index-error is constantly changing, and 

 requires great watchfulness. In registering thermo- 



C3 



ther- 



meters record but do not appl^ their zeros. 

 Ill placing the M 



Thermometer, an exposure 



should\e chosen perfectly shaded both from direct sun- 

 shine, and that reflected from the sea, or radiated from 

 any hot object. It should be especially guarded from 

 rain and from spray, so that the bulb should never be 

 wetted, also from warm currents of air and from local 

 radiation; completely detached from contact with the 

 ship's side, and fully exposed to the external air. In 

 readins? it the observer should avoid touching, breathing 



V 



on, or in any way warming it by the near approach of his 

 person ; and in night-observations particular care should 

 be taken not to heat it by approach of the lignt. 

 quicker tlie reading is done the better. 



The 



At nieiit it 



fm 



should be completely screened from the sky, so as to 

 annihilate all loss of heat by upward radiation ; a light 

 frame-case uf double wire-gauze will perhaps be found 



\ . .-h 



