125
18. At Green Island the wind, clouds, sea-surface and weather are observed at 4 a. 10 a. 4 p. and 10 p. As the island is within 4 miles of the Observatory the station is not supplied with instruments, but the observer appears to be doing his best. At Stone Cutters' Island, which is within 2 miles of the Observatory but at a much lower level, the rain is measured at 10 a. as published.
19. All the observations made at these four stations are revised, corrected and reduced at the Observatory, and the instruments &c. are occasionally inspected.
20. The barograph was erected in March and worked without interruption since the 1st April. The slab is placed on a teak-wood table, which is firmly screwed to the floor.
21. The image of the flame of a Kerosine lamp, enlarged by a condenser, is thrown upon the void space, narrowed by a slit, above, the mercury of a barometer of about three quarter inches internal diameter. By means of a photographic lens an image of this illuminated slit is thrown upon a cylinder covered with a sheet of sensitised paper, 23 inches long and 5 inches wide, which is revolved by a clock work, so that the portion covered by the image moves 0.364 inch an hour, the clock work moves also a shutter, that cuts off the light of the lamp from two minutes before, till two minutes after every even hour. The upper edge of the inverted image of the slit rises or falls as the mercury falls or rises in the barometer, but the lower edge is not permanently fixed, but rises or falls as the temperature of two zinc rods, fixed beside the barometer, falls or rises, and by aid of an adjustable glass lever the amount of displacement of the edge is made exactly equal to the temperature correction, that otherwise would have to be applied to the hourly readings.
22. The paper may be kept on the cylinder for two days, after which it has to be changed,--this being invariably effected between 10 a. and 11 a.,-developed, fixed, washed and dried. The photograph is ready to be measured three days after being removed from the cylinder. It is then placed between two glass plates in the tabulator, and the distances between the upper and lower edges of the blackened portion of the paper, which is interrupted by the two-hourly white lines, are read off at every hour or oftener, if required, by aid of a vernier capable of being read to 0.001 inch, two fine wires, fixed in empty sight-tubes being made to cover the respective edges.
23. The standards of reference are obtained from 10 a. 1 p. 4 p. and 10 p. readings of the standard barometer, corrected and reduced to 32° Fahrenheit. From the 1st April till the 1st September the 1 a. readings were also made use of, but experience proved this to be superfluous. The nominal inches on the tabulator should be greater than true inches in the same proportion as the magnified image of the slit is greater than the true image, which is about 14. Experience shows, that this has not been strictly attained. The nominal inches are 1.594, whereas they should be 1.534 inches long. But as the pressure here nearly always changes very slowly and regularly within 24 hours, it is not necessary to know this proportion with great accuracy, and it is for the same reason difficult to determine it. The above number was derived from observations made during the Typhoon in September and agrees with other observations.
24. The room in which the barograph and the standard barometer are placed is carefully shut up, so that the daily range of temperature is reduced below half a degree. Three large Kerosine lamps, always burning in the room, raise its temperature in winter a couple of degrees above the temperature of the air outside, while in summer the room is colder than the air. The temperature is observed by reading a carefully verified thermometer immersed in mercury in a tést-glass of the same diameter as the barograph barometer. The constancy of the temperature favours the accurate co-operation of the different parts of the apparatus, which are at a uniform temperature, just as a clock goes better in a room, where the temperature does not change much, because the different parts of the pendulum have the same temperature.
25. The barogram readings are entered in a journal kept in the computing room.
The figures are corrected for the scale-error of the tabulator, and when reduced to standard by comparison with the readings of the standard barometer (corrected and reduced to 32° Fahrenheit), they are entered in the tables printed in the monthly reports.
26. The thermograph was erected in March and worked without interruption since the 1st April. The slab is placed on massive teak-wood blocks, firmly screwed to a slab of granite, which rests on solid masonry.
The bulbs of the recording thermometers (dry and damp bulb) are placed in a zinc screen outside the northern window of the instrument room, which is substantially boarded, and in which are also placed two thermometers with bulbs,-dry and damp,-of similar dimensions. These have been care- fully verified at different temperatures by comparison with our standard thermometers. The tubes of the recording thermometers are bent and enter the instrument room through two slots (5.6 inches long, 1.2 inches broad and 9.2 inches asunder) bored in the boards. They then rise vertically and are held by pieces of brass, which may be raised or lowered to some extent. The slots are filled with non- conductive material, so that no air can pass out from the room. An airspeck is introduced into the mercury of each thermometer. These airspecks are photographed on the cylinder. A lamp is placed on each side of the thermograph, whose lights are condensed by lenses and reflected towards the cylinder from mirrors, placed on the slab behind the thermometer tubes. The light penetrating through