42 



the pen carrier not abolutely essential to constraining the pen to 

 the desired vertical line. Finally the contact pressure of the pen on 

 the record sheet is no greater than essential, and results from a small 

 residual gravitational tendency of the carrier to rotate the pen point 

 against the record sheet with a very gentle pressure. The siphon 

 tubes are about 1^4 inches in diameter and the float is only slightly 

 smaller. This gives a moving force capable of overcoming the una- 

 voidable friction in a highly satisfactory manner, and the absence of 

 any complicated mechanisms renders false and interrupted records 

 almost an impossibility. 



83. Time checks on record sheet. — ^As thus far described the baro- 

 graph is complete, and with the aid of the driving clock and drum, 

 which require no further description, gives highly accurate and con- 

 tinuous records. The detailed analysis of barometric records gener- 

 ally requires hourly readings. When record sheets are employed with 

 ruled scale for pressure and time there is always a difficulty in setting 

 the record so that the ruled hour lines on the sheet indicate the true 

 time. A similar difficulty arises in setting the pen to the correct 

 point on the pressure scale. This, however, is of slight consequence 

 if sheets are properly printed and cut with uniform margins and 

 carefully placed on the cylinder. 



To secure easily an equally satisfactory result with the time record, 

 the driving clock is provided with a dial and hands in the usual fash- 

 ion. These moving continuously day after day enable the clock to be 

 regulated to keep correct time, a result very hard to secure when the 

 rating is done on record sheets that are frequently changed. More 

 especially, however, the barograph is equipped with a special time- 

 marking device which automatically operates once each hour at the 

 instant the minute hand of the clock reaches XII, or the zero point 

 of the hour. Nearly all the time lines are omitted from the printed 

 rulings of the record sheet, and the marke-r operates so as to lift the 

 float a few hundreds of an inch and immediately release it. This 

 causes the recording pen to oscillate a few times up and down and 

 to inscribe a short transverse line across the pressure record. These 

 transverse strokes are, in fact, the hour lines for the entire record 

 and are inscribed with all the accuracy required. 



84. There is a further advantage from the action of the time 

 marker. The float is raised from the mercury momentarily and 

 subsequently oscillates slightly. This causes a general renewal of 

 for the forces of buoyancy and capillarity which determine the exact 

 positions of the float, and any failure of the pen at the end of oscil- 

 lations to return exactly to its original position is an index of the 

 magnitude of errors that arise partly from friction and partly from 

 variation in the capillary and buoyant forces. 



Discontinuities of several thousandths of an inch, due to these 

 causes, are sometimes found in the records, rendering apparent the 

 existence of small errors which would otherwise be only conjectured. 



85. The time marker is shown in Figure 27, and consists of an 

 electromagnet, the circuit of which is momentarily closed by a spring 

 contact operated by the minute hand of the clock as it passes the 

 XII point of the dial. 



The armature of the magnet is L-shaped, as seen in the picture. 

 A long, light, horizontal rod. is pivoted at the depressed end of the 



