460 



KKPORT — 1880. 



The observation of the angle betw"eeu the two reflexions is perfectly easy with 

 a full-sized sextant, if the trough of mercury be so propped up that the reflexion 

 from the pool can be viewed underneath the trough. For this purpose I use a 

 tripod stand with a bar of rough wood, say 18 inches long, 3 wide, and 2 thick, 

 secured horizontally across its top. I lay the mercurial horizon on one of its pro- 

 jecting ends and Ijetween a few studs that have been driven in to prevent its 

 accidentally slipping oft'. The edge of the bar is bevelled, and its thickness is 

 reduced at the place where the mercury trough is set. Then the observation is 

 taken, just as any other sextant observation would be. The reflexion from the 

 mercury falls upon the index-glass, and that from the pool is viewed directly 

 through the object-glass below the trough and its supporting bar. 



Unless the sextant be a full-sized one, this operation cannot be effected, be- 

 cause the index-glass will not stand high enough above the line of siglit to catch 

 the reflexion from the mercury. It will simply reflect the side of the trough. 



If there be no tripod stand, and it becomes necessary to lay the trough on 

 the ground, an observation can still be made, but in an inconvenient fashion. 

 The sextant will have to be held topsy-turvy, that the brighter reflexion of the 

 cloud from the mercury, and not the feebler one from the pool, should fall 

 on its index-glass. The angle read will be negative ; it will be what is com- 

 monly called an 'ofi'' angle. A small sextant may be used in this method, 

 because the rim of the trough is narrow that intervenes between the further 

 edge of the mercury and the objects seen beyond and over it. 



The most convenient method of measuring the rate of movement of clouds, after 

 the height of the cloud plane has been once determined, is to watch the movements 

 of a patch nearly overhead, and passing away from the zenith, as seen reflected 

 in the mercury, and measuring its angle of depression ( =its altitude) with some 

 simple and suitable instrument, such as the pocket alt-azimuth already mentioned. 

 Two measurements, n^ and a„ are taken, as well as the intervening time, t seconds, 

 whence we obtain rate of movement = height of cloud x (cotan Oj — cotan a^ in 

 t seconds. 



When the water is almost wholly calm, I find that 2' of error is the utmost 

 that need be feared. If wholly calm V would be ample to make allowance for 

 in a set of three or four observations. Now suppose we wish that our determina- 

 tion shall never be more than, say, 10 per cent, in error, we can easily find from 

 the tables what the minimum height of the station must be in any given case, 

 to secure this result. In the first instance we should require a parallax of 10' and 

 in the second of 20'. This is obtained b}' an elevation of 10 or 20 feet as the case 

 may be, when the height of the clouds in feet corresponds to the tabular numbers ; 

 that is, when it is between 2000 and 3000 feet. At 100 or 200 feet elevation, 

 clouds of ten times that height could be observed with equal accuracy. Numerous 



