494 



NATURE 



[August 21. 19 19 



lent in the level of the sea. But, whatever may have 

 been the exact circumstances and conditions of that 

 experiment, its result is very far from holding good 

 within our narrow seas. According to Prof. Witting, 

 both in the North Sea (on the average) and in the open 

 basin of the Baltic the slope of the sea-surface is 

 equivalent to a column of water about 35 times that 

 of the barometric gradient measured in mercury; 

 whereas, hydrostatically, the ratio should be only 

 about 13-5 times, or in the simple ratio of the specific 

 gravities of mercury anH water; in still narrower 

 areas, as in the gateways of the Baltic, the ratio 

 may be much greater, amounting to as much as 

 100 times. In the oi>en basins the direction of the 

 water-slope deviates from that of the barometric 

 gradient by about 55° ; and this value is again exceeded 

 in the narrower channels. In other words, then, we 

 find that in the opener basins the disturbance of the sea- 

 level is nearly three times as great as the direct hydro- 

 static effect due to atmospheric pressure, which is as 



region the height of the dam will depend not only 

 on the atmospheric pressure, but also on the quantity 

 of \yater which passes over into the North Sea, the 

 level of which varies according to its local conditions 

 and to more distant phenomena in the ocean. In a 

 general way, the peculiarities of the narrow entrances 

 to the Baltic are not difficult to comprehend. 



But let us return to the internal forces, or, prac- 

 tically speaking, to the distribution of densities in the 

 sea. We have here a direct cause of variation of 

 surface-level, giving us, as it were, a theoretic or ideal 

 sea-level at any particular place, largely affected in 

 actual fact by extrinsic forces which move and heap 

 up the waters. Now, we began by saying that mean 

 sea-level, properly so-called, is a constant as yet un- 

 determined. But, nevertheless, we know a very great 

 deal more about it than we did even a few years ago, 

 and can give a very fair approximation to it in a con- 

 siderable number of localities. Even at Dundee or 

 Aberdeen mean sea-level is a very different thing from 



Fig. 1. — Sea-level and pressure-distribution in March, 1906; represented bV 

 isanomalies from the means for the whole period 1898-1912. Thick 

 lines, sea-level in half-centimetres ; broken lines, barometric pressure in 

 tenth-millimetres. 



much as to say that the dynamical effect of the wind 

 must account for well-nigh two-thirds of the total 

 effect produced. Moreover, in these waters the current 

 produced by the wind deviates by about 20° to the 

 right of the direction of the wind itself; and the 

 current thus moves sideways .up the slope, about 45° 

 to the right of its uphill direction, which is an oceano- 

 graphical paradox. 



In the gateways of the Baltic these external forces 

 by no means fully explain the phenomena which occur. 

 We may calculate the amount of water which escapes 

 from the Baltic, on an average, month by month ; 

 and, proportional to these amounts, we then discover 

 residual deviations of slope in the water-levels. Within 

 its gates the Baltic may be regarded as a basin the 

 levels of which (except for perturbations at the mouths 

 of rivers) are mainly determined by the barometric 

 field. But the gateways themselves constitute a 

 sluice, damming up the waters within; and in this 



NO. 2599, VOL. 103] 



Fig. 2.— Mean sea-level, calculated values ; referred (in centimetre-units) to 

 a geoidic surface touching the North Sea in the middle of its northern 

 part. 



that mark at Liverpool from which all our Ordnance 

 levels are reckoned, and which (though by a happy 

 accident it is not far from the truth) we all know to be a 

 purely empirical or conventional datum. Now, as 

 Prof. Witting argues, we do already know enough 

 about this subject to be able to define, with a fair 

 amount of accuracy, a "zero-pressure level" in the 

 sea at a point where no movements are caused by the 

 distribution of densities ; to be more specific, we may 

 choose a point somewhere towards the middle of the 

 northern part of the North Sea, describe a geoidic 

 surface touching it, and call that (to the best of our 

 present knowledge) our datum level. Proceeding out- 

 wards from this zero point, we may calculate the 

 hydrodynamical part of the slope due to the distribu- 

 tioti of densities; and again we may calculate, and 

 superadd to this, the slope due to barometric pressures, 

 either for some particular epoch or in average values 

 corresponding with long-period barometric means. In 

 Fig. 2 Prof. Witting shows, after this manner, the 



