31 



surface of the mercury and the lever /, poised in equilibrium, in which 

 case the platinum tip h will stand about midway between the plat- 

 inum-pointed screws above and below it, respectively. The pen car- 

 rier R' must also then occupy such a position on the screw that the 

 tracing pens will indicate the true barometric pressure upon the rul- 

 ings on the record sheet. Appropriate connections with an electric 

 battery being made, the action of the mechanisms will be as follows : 

 Any minute change in the level of the mercury will alter the posi- 

 tion of the iron float, in consequence of which the platinum tip of the 

 lever I will move into contact with one or the other of the platinum- 

 tipped screws, causing a current of electricity to be directed through 

 the electromagnet connected therewith. The action of either electro- 

 magnet releases the clockwork P. In doing this, however, the one 

 magnet shifts the lever L laterally toward the back of the instru- 

 ment, while this lateral movement will be toward the front if brought 

 about by the action of the other magnet. The movement of the clock- 

 work causes the lever L to advance or recede so that a pawl upon the 

 end, engaging a tooth of a ratchet wheel upon the end of the screw 

 S' revolves the latter a fraction of a turn. This fractional turn will 

 be in one direction if the level L is drawn backward or in the oppo- 

 site direction if L is pushed forward, according as the level I has 

 made contact with the upper or under screw. This movement of the 

 screw S' shifts the pen carrier R' and the pens upon the record 

 sheet, and, being communicated to the screw S causes a proportion- 

 ate change to take place in the float-carrier E. The clockwork auto- 

 matically stops after one such cycle of actions. If after these move- 

 ments the lever I is again poised in equilibrium no further action 

 ensues until the contact of h with one or the other of the screws is 

 again made, whereupon the cycle of actions will again be set up and, 

 if necessary, repeated in quick succession until the equilibrium of 

 the poised lever I is restored. The movement of the pen carrier E.', 

 corresponding to a change in the position of the float, is four times 

 as great as the change in the height of the mercurial column. A 

 change of 1 inch in pressure, therefore, is represented as a change of 



4 inches on the sheet. Each closure of the circuit producing one 

 twenty-fifth revolution of the screw S', or one-fiftieth revolution of 



5 represents a change in the height of the mercurial column of 0.001 

 inch, which is the nominal sensitiveness of the instrument. Owing, 

 however, to unavoidable imperfections in screw threads and electric 

 contacts and to the capillary action of the mercury in the barometer 

 tube, the probable error of the instrument is much greater than this, 

 no doubt amounting to at least 0.01 inch. 



By selecting proper proportions for the long and short legs of a 

 siphon barometer, the effects of temperature can be almost perfectly 

 eliminated. This, however, appears not to have been considered 

 when Foreman's barograph was designed, and the records are subject 

 to small periodic errors due to temperature changes. For the further 

 elucidation of the automatic compensation of siphon barographs for 

 temperature see paragraph 80. 



Owing to the presence of the float on the surface of the mercury 

 in the short leg, and to other causes, it is practically impossible to 

 make a direct measurement of the actual height of the column of 



311681° — 41 5 



