order to overcome the friction of the mercury against the glass, it is lubricated with alcohol. The upper part of the tube, which is not blackened, also contains air. A dextrous shaking of the tube will transfer bubbles qf air from one end to the other, thereby holding the column of mercury to any desired height, or to any desired distance from the circuit points.
The stand should be mounted in a locality that is not shaded. The tube should be in the plane of the true meridian, and face the south. The angle of inclination should be roughly about 45 degrees. In summer it should be a little nearer to the vertical; in winter a little more to the horizontal. The angle depends partly on the amount of air in the lower bulb and partly on the position that gives it the maximum of insolation. In general, the results are best when the top of the mercury is from half an inch to an inch above the circuit points during the warmest part of the day.
The recording device is attached to that of the anemometer, using the same sheet but a different recording pen. When the sun is shining, a contact of the clock completes the circuit. The movement of the armature operates a jigger, which moves the pen once every minute. The jigger and the progressive movement of the drum cause the pen to make steps in series of five, back and forth. Each series represents five minutes of time. These continue while the sun is shining. When the sun is not shining, the mercury in the thermometer drops away from the platinum circuit points. The pen then draws a straight line.
Even when carefully adjusted, the recorder will not begin to register for some time after sunrise; it ceases to register a short time before sunset. These periods must be measured from time to time by the observer, taking the time of sunrise and sunset from a reputable almanac for the approximate latitude of the station. These intervals are the twilight corrections.
The morning twilight correction is usually somewhat greater than that of evening. Both vary slightly between winter and summer. In localities where city smoke and floating dust do not contribute to instrumental sluggishness, the morning correction should not exceed one hour; the evening correction should not be more than half as much. The presence of smoke,
