608 ROSSiTER [chap. 16 



spirit levelling, has recently developed from a need to confirm the findings of 

 the Unified European Levelling Network. This network, which for scientific 

 purposes has crossed political and geographical boundaries in an attempt to 

 integrate the levelling nets of individual countries, has produced provisional 

 results indicating that sea-level is something like 0.5 m higher in the northern 

 Baltic than in the Mediterranean. 



Here one is concerned with the deviations of observed sea-level from a 

 geopotential surface, from place to place, but before these can be ascertained 

 it is necessary to utilize a long series of records at many gauges, and to eliminate 

 from them as many of the time variations as possible. Many of the results 

 given in section 4 of this Chapter originate from an investigation of this 

 type, and it will be seen that if equation (1) is formed for each of a pair of 

 stations, one can obtain the mean difference in level between their respective 

 bench marks for a tmiform atmospheric pressure over the region concerned. 

 Doodson (1960) has given tenative values for certain stations, and has also 

 commented upon the feasibility of attaining, observationally, a fundamental 

 reference plane in the sea (the geoid). Whether a similar treatment can be 

 applied so as to allow for the semi-permanent surface gradients caused by 

 differences in the density of sea-water rests largely upon whether sufficient 

 observational data of temperature and salinity are available. A theoretical 

 statement of the problem has been given by Vantroys (1958). 



The amount of material assistance the oceanographer can offer the geodesist 

 must be limited by the standard of accuracy the geodesist demands. There is a 

 particular levelling problem, however, where the geodesist must rely much 

 more upon the oceanographer. Levelling connexions across even quite narrow 

 bodies of water provide formidable obstacles to spirit levelling. Such a body of 

 water is the Straits of Dover; Cartwright (1960) hopes to demonstrate that, 

 with the aid of current observations from the Straits and sea-level observations 

 from opposite shores, it is possible to carry a level across the 21 miles of water 

 to an accuracy of the order of a centimetre. 



6. Conclusion 



In this chapter an attempt has been made to provide a broad outline of the 

 subject, but it should be read in conjunction with that on seasonal variations in 

 level (Pattullo in Vol. 2). Although a certain emphasis has been placed upon 

 the influence of winds and air pressure, it is not suggested that these factors 

 necessarily predominate in all seas and oceans to the exclusion of, for example, 

 density variations, though in shallow seas this is highly probable. In deep 

 untrammelled waters the density factor is possibly more important. 



Mention has not been made of any periodic variations other than the astro- 

 nomical and pole tides. Oceanographical literature, like meteorological litera- 

 ture (Shaw, 1928, 2, p. 320), contains many examples of the search for periods 

 such as the sunspot cycle and even the fourth harmonic of Gabriel's cycle of 

 744 years ; as with the nodal and pole tides, however, the oscillations are too 

 small to be verified conclusively with existing data. 



