20 INSTRUCTIONS TO MARINE METEOROLOGICAL OBSERVERS 



parts of the instruments. This device, therefore, when perfect, not 

 only corrects the errors which would otherwise arise when ex- 

 pansion of the parts occurs during rising temperature, but also 

 those due to the contraction which takes place when the temperature 

 falls. When the device is imperfect, there is only partial compensa- 

 tion for temperature, and the readings of the instrument suffer in 

 accuracy on account of it. The aneroid may continue to be service- 

 able, however, but its value as a meteorological recorder has been dis- 

 tinctly lessened. When there is no compensating device— and there 

 are many of these instruments which lack it, or have it so inade- 

 quately developed that it is useless — readings from such aneroids can 

 never be depended upon for accurate pressure indications. 



Measures for remedying defects. — Aneroid barometers that are de- 

 cidedly erratic, because of one or more of the defects mentioned, are 

 subject to one of two remedies, viz, I'epair or replacement by a better- 

 performing instrument. This is in justice to the Weather Bureau, in 

 the interest of accurate observations, and of benefit to the ship, in 

 the possession of a satisfactory navigating instrument. 



Some aneroids, after becoming disordered, may be satisfactorily 

 repaired. If the trouble is due to gumminess in the w^orks, which 

 condition may occur in the aneroid on a tanker, cleaning will effect 

 the desired result. Tightness or looseness in the springs and linkage 

 uiay be relieved. When, however, there are defects in manufacture, 

 the results of attempted repairs are doubtful. Much more fre- 

 <[uently than otherwise, as shown by actual experience, they are 

 fruitless. Such delicate work as the repair of an aneroid barometer 

 should be entrusted to the makers of the instrument or their repre- 

 sentatives, or to a repairman who is fully equipped for such w^ork. 

 This rule applies also to the resetting of an aneroid, since many a 

 barometer has been impaired by resetting and has had to be subse- 

 quently recalibrated before its readings could be trusted. 



Testing of aneroids. — Some aneroid tests may be made simply and 

 ■\Aithout resort to apparatus. For instance, the freedom of move- 

 ment of the hand of an aneroid can be quickly tested by tapping 

 the instrument on the side or bottom with the fingers or knuckles, 

 or perhaps better by lifting the instrument about one-fourth of an 

 inch from a table or cane-seated chair and placing it back again 

 somewhat sharply. Under this treatment, if the joints and levers 

 are perfectly free, the hand will jump away and then return quickly 

 with a vibratory movement to its original ]30sition. If the instru- 

 ment is defective, the hand in some cases will not respond to the 

 slight knocks, or will do so without exhibiting any vibratory move- 

 ment, or upon being disturbed it ma}^ move a little, but will not 

 return to its original position. 



When aneroids fail, or appear to fail, to register correctly con- 

 siderable changes in pressure, as in passing through a deep cyclonic 

 storm, they should be tested for the real or apparent defect. The 

 Weather Bureau maintains apparatus in its offices at New York, 

 New Orleans, and San Francisco for this purpose. The instru- 

 ment to be tested is placed under a bell jar connected to an air 

 pump, and the air is gradually removed from the jar. If the an- 

 eroid is a good one, it should register the same falling pressure 



