On Mean Sea Level and its Fluctuations. 19 



features to which we should naturally look for a meteorological 

 explanation. The chief nieteorolo<i;ical causes which remain to be 

 considei'ed are barometric pressure and wind. That the former has 

 a marked influence on the mean level of the ocean was shown lon<^ 

 ago by Sir James C. Ross from observations made on his Antarctic 

 Expedition in 1848. Here Sir James Ross showed that for maiiy 

 consecutive days the Mean Sea Level vaiied inversely as the heijiht of 

 the barometer, and that the difference of level corresponded to the 

 difference of barometric height approximately in the ratio of the specific 

 gravities of sea- water and of mercury. In short, as Sir James Ross 

 said, his observations seemed to indicate " that the ocean is a water- 

 barometer on a vast scale of magnificence." * Similar observations 

 had already been made by ]\l. Daussy and by Sir John Lubbock (Sen.) 

 in his woik on the Tides. 



Professor Geelmuyden [of. cit. pp. 19, 53) has considered tJiis 

 question in connection with the tides at Christiania. His method 

 was simply to reduce the observed tide-levels to a constant baro- 

 metric pressure. " A set of barometric readings, registered during 

 the same period, was handed over by the Norwegian Meteorological 

 Institute. On the approximate supposition that the action of air 

 pressure on the water level is instantaneous, they were analysed in 

 exactly the same manner as the tides, and the values of A, B, etc. 

 [the harmonic components | multiplied by 13'25 (which was taken as 

 the specific gravity of mercury as compared with sea-water) were 

 added with their sign to the corresponding tidal constants. The 

 effect on the terms of astronomical origin is mostly insignificant, even 

 on terms whose amplitude is so small that they have been retained 

 only for the purpose of this comparison. But all five tides of long 

 period came out with dim.inished co-efficients, and three of them with 

 strongly altered phase ; only for the annual tide was the phase sensibly 

 the same, which seems to prove that otiier meteorological causes, 

 principally, perhaps, prevailing winds, act on the whole in the same 

 direction as the direct air pressure, but stronger still." 



If we take, by interpolation, from Dr. Buchan's Meteorological 

 Atlas, the mean monthly barometric heights at Dundee, we find 

 that they follow a mean annual fluctuation which shows a maximum 

 in May, and a minimum in November or December, the total range 

 being from about 757 "4 to 750-4 mm.f The cm've is not unlike in 

 form to our curve of Mean Sea Level, and is nearly opposite to it in 

 phase. But, on the other hand, when we compare the two curves 

 we see that an amplitude of 3-0 mm. of mercury in the one corresponds 

 to about 244 mm. of water in the other, or one mm. of mercmy to over 

 80 mm. of water. This is just about six times the ratio which Sir 

 James Ross found in the Antarctic, — six times, that is to say, the ratio 

 of the two specific gravities. Moreover, though the phase of the annual 



* Phil. Trans., vol. 144, 1854, p. 295. 



t The actual monthly means at Dundee for the period 1897-1912, kindly 

 given me (since the above was written) by Dr. A. Watt of the Royal Scottish 

 Meteorological Society, are somewhat different from the above, viz., Jan., 759".3; 

 Feb., 756-4; Mar., 756-6; Apr., 758-4; May, 760-5; June, 7609; July, 7606; 

 Aug., 7582; Sept., 7616 ; Oct., 7584 : Nov', 7578; Dec, 754-6. Here we have 

 high values in September and in January ; but there happen to have been some 

 exceptionally high readings in both months in the course of the period of si.xteen 

 years. 



