144 DISCOVERY REPORTS 



It has, however, to be borne in mind that the surface of a non-homogeneous sea is, 

 as a rule, not horizontal. Suppose that in the upper layers the density varies horizontally 

 but is constant at great depths and that the current is practically zero at great depths. 

 The latter assumption involves that at great depths the isobaric surfaces are horizontal 

 or level surfaces, and it follows that the surface of the sea shows elevations where the 

 average density is small and depressions where the average density is great. Within a 

 convection current the light water is accumulated on the left-hand side (in the southern 

 hemisphere) and, therefore, the surface of the sea is higher on the left-hand side of the 

 current. The surface, therefore, is inclined even if no slope current in the sense of 

 Ekman is present, and in the presence of such a slope current the actual inclination is 

 the combined effect of the distribution of density and the piling up of the water. 



By means of very precise levelling the variations of the sea-level along coasts can be 

 determined. By comparing the results with the corresponding variations which can be 

 derived from the distribution of density, it should be possible to decide if this distribu- 

 tion is alone responsible for the variation, or if in addition a piling up effect has to be 

 considered. Some results of highly precise levelling in the United States and Central 

 America have been discussed by Avers (1927). Between Colon and Panama a difference 

 in sea-level of 17-8 cm. has been found, the level being highest on the Pacific side. By 

 means of the observations of the ' Carnegie ' in the Caribbean Sea and the Gulf of 

 Panama, a corresponding difference of about 35 cm. is found, supposing the pressure at 

 a depth of 3000 m. to be the same in both seas. If these figures can be brought in re- 

 lation to each other, they would indicate that the difference of level, which should be 

 expected because of the difference in density, has been reduced. Such a reduction 

 would be brought about if the water were piled up in the Caribbean Sea and drawn 

 away from the coast in the Gulf of Panama. 



Avers, furthermore, points out that the sea-level increases from south to north both 

 along the Atlantic and Pacific coasts. Along the Atlantic coast an increase of 3 1 cm. is 

 recorded from St Augustine, Florida, to Portland, Maine, while along the Pacific coast 

 an increase of 34 cm. is found between San Pedro, California, and Seattle, Washington. 

 The sea-level is from 49 to 59 cm. higher at the Pacific coast than at the Atlantic. It 

 seems possible that the increase along the Atlantic coast can be associated with varia- 

 tions in the density of the water, but observations which are suitable for a test are not 

 available. It may, however, be pointed out that to the north of Portland, Maine, the 

 sea-level must sink, if the height depends upon the density of the water, because in that 

 region the average density increases rapidly to the north. Along the Pacific coast the 

 increase in the height of the sea-level cannot be attributed to a variation in the density, 

 since this, according to the observations of the 'Carnegie,' should rather tend to cause 

 a decrease from south to north. It is, on the other hand, not very probable that the 

 prevailing winds will cause a considerable piling up of the water at the two coasts, be- 

 cause the general directions are such that the surface water is drawn away. The results 

 of these levellings, therefore, throw no more light upon the question of the relative im- 

 portance of the distribution of density and the piling up effect of the wind. Apart from 



