55 



and very accurate readings can yet be made. It is a bad practice 

 to clean the mercury in barometers as soon as it becomes sli|t^htly dull 

 and tarnished. Leaks are apt to be started in the joints of the 

 cistern, and slight changes in the position of the ivory point give 

 rise to new and unknown corrections for instrumental error. The 

 mercury itself is apt to become contaminated with impurities and 

 afterwards will remain bright only a very short time. 



118. Comparrative harometer readings. — Each regular Weather Bu- 

 reau station is supplied with two good barometers to lessen the 

 chances of a break in the record and to guard against erroneous 

 reports from the use of imperfect instruments. Monthly, therefore, 

 and on other special occasions, as further specified in paragraph 122, 

 five comparative readings of all barometers on station should be 

 made at uniform intervals of hours, half hours, or quarter hours, as 

 may be most convenient to the observer. 



119. As the object of the comparative readings is to ascertain 

 accurately the amount of discordance between the barometers and 

 enable the main office to replace defective instruments, it is impor- 

 tant that the observer use more than ordinary care to read the barome- 

 ters exactly as they are. He should not feel biased or disposed, in 

 the slightest degree, to make the readings come out one way more 

 than another. His whole endeavor should be to make the settings 

 and readings as accurately as possible, without any regard as to 

 how the readings may differ in the end. When the pressure is 

 found to vary rapidly, make the readings of the two or more instru- 

 ments as quickly as possible, and throughout the series endeavor to 

 keep the temperature stationary. 



120. Before each reading the cistern of the barometer should be 

 unscrewed so as to lower the mercury one-sixteenth to one-eighth of 

 an inch below the ivory point and the setting then carefully made. 



121. An interval of two or three hours should intervene after 

 barometers are unpacked, cleaned, or moved to a new office and hung 

 in position before comparative readings are commenced. 



122. Comparative readings should always be made with new 

 barometers whenever received at a station, and also both before 

 and after instruments are removed from one location to another or 

 cleaned. 



123. Suggestions and instructions for cleaning harometet^s. — In a 

 few cases observers are authorized to clean cisterns of barometers 

 that are very dirty and can not easily be replaced. (See par. 187.) 

 The following instructions will then guide in the proper performance 

 of the work: 



124. Take a series of five comparative readings before the work is 

 begun. 



125. Provide one or more very clean, dr^ porcelain or glass cups 

 or saucers. Avoid the use of damp, unclean, or metal vessels. 

 Cleanse the vessels by thorough washing in soap and water and 

 wipe dry with a clean towel, finally polishing the vessel with tissue 

 or similar soft paper. Provide, also, some pieces of clean cloth and 

 sheets of tissue paper for cleansing the glass parts of the cistern, also 

 a few small sheets of clean white paper about 4 by 6 inches for 

 use in filtering the mercury. Calendered letter paper is not so 

 good as the ordinary so-called book paper. A most convenient 



