64 REPORTS ON THE STATE OF SCIENCE, — 1915, 



(6) The Milne-Shaw begins to fall about 15 minutes after the thermo- 

 graph, and falls pretty rapidly ; the fall had not stopped when the 

 observations were closured, and it is clear that the coefficient deduced 

 from the fall would be greater than that found from the rise. 



(c) The disturbance of the Milne-Burgess machine is much more 

 complex. There are waves on the rise, and the main maximum which 

 follows that of temperature by nearly an hour is followed by another 

 1^ hour later, and again another at 4-|- hours after the temperature 

 maximum. There may be others later still. A reasonable explanation 

 is that the warmth reaches different parts of the instrument at different 

 times, the separate parts producing separate maxima. This would fully 

 account for the curious discrepancy between the former results, and the 

 large range of values for the interval between M.-B. and M.-S. We have 

 only to suppose that sometimes one part of the instrument is pre- 

 dominantly affected and sometimes another. 



The main conclusion is that internal temperature can only be re- 

 sponsible for a part, and probably only a small part, of the diurnal de- 

 flections of the instruments. The main cause seems to be external, 

 and is probably the daily opening and shutting of the valley to which 

 Milne drew attention many years ago. 



Before leaving these deflections, a word or two may be said as to the 

 tidal effects, so conspicuous on the Milne-Shaw machine at Bidston, but 

 almost hidden by the temperature effects at Shide. Some trouble was 

 taken to disentangle them, and it was found that the lunar tide could 

 be identified by means of the progressive phase. For the Milne-Burgess 

 instrument the coefficient was about 1 mm. ; but it is mixed up with 

 temperature effects which may vary widely in character. The tidal 

 coefficient for the Milne-Shaw instrument at Shide is not very much 

 larger than 1 mm. — perhaps 2 mm. at most. But the material available 

 at the time of the discussion was not large and further investigation is 

 desirable. 



Finally, some particulars may be given as to the deflections at other 

 stations, where the simple Milne instruments are in use. The magnifica- 

 tion being small, the disturbance of the trace by either temperature or 

 tide is not noticeable to a casual glance ; but if careful measures are 

 made of the travel at every 2 hours (or 4 hours in some cases) through- 

 out the day, the general nature of the movements can be detected. 

 Such information may be of value in arranging for the setting up of 

 instruments of higher magnification. 



The diurnal changes were deduced from the means for all the days 

 measured ; as also the semi-diurnal changes. 



As regards the lunar tide, the simplest way of detecting it is to form 

 the differences between readings for 2 days separated by any con- 

 venient interval from 4 days to 11 days ; for in TJ days the lunar semi- 

 diurnal tide reverses itself, maxima falling on the former minima. We 

 thus get twice the effect by the subtraction. If the interval of 7 or 8 

 days is not available, the factor will not be so large as 2, but is easily 

 calculated from the relation 



Cos {6— a) -cos {0 + a) = 2 sin a sin 0. 

 For 7* days, a = 90° ; for 4 days, a = 4 X 9077-| = 48°. 



