SECT. 5] 



SEASONAL CHANGES IN SEA-LEVEL 



487 



Both the distribution of gauges and the large variations in level with time and 

 with space make it questionable whether contours of the sort given here are 

 worthwhile. Our reasons for presenting the results in this way are (1) to make 

 the large-scale, general features easy to comprehend, and (2) to indicate clearly 

 the areas where deviations are particularly large or irregular. 



The principal features shown on Fig. 1, the chart for March, are that sea-level 

 is lower than its annual mean elevation almost everywhere in the Northern 

 Hemisphere, but higher than average almost everywhere south of the equator. 

 Appreciable exceptions in the Northern Hemisphere occur in the Arabian Sea, 

 in the Gulf of Siam and between 40° and 60°N in both the Pacific and the 

 Atlantic Oceans. In the Southern Hemisphere the only negative values of any 



Fig. 1. Deviations from mean annual sea-level, averaged for the months February, March 

 and April. Heavy lines, zero deviation; light solid lines, positive deviations (sea-level 

 higher than the annual mean) ; dashed lines, negative deviations. Contour interval 

 3 cm. 



magnitude occur along the south coast of Australia. The largest deviations 

 observed anywhere in the world occur in the Bay of Bengal ; deviations from 

 the annual mean at some gauges there are —40 cm. Gauges along the coast of 

 Mexico and Central America also show large negative values, — 16 cm, as does 

 the one on the northeast tip of Siberia, — 19 cm. The largest positive values 

 observed anywhere occur around Australia, particularly on the northeast 

 coast, +16 cm. An interesting detail small in area but having large deviations 

 is found in the Indian Ocean. The variations east of Sumatra are of different 

 sign from those on the west. Deviations in the Bay of Bengal are negative while 

 those in the Gulf of Siam are positive. This sharp reversal in this short distance 

 can be found in all four seasons. 



If we now compare this chart with Fig. 2, the chart for September, we find 

 almost identical general features, except that, where deviations were positive 

 in March, they are negative in September, and vice versa. Also, the locations 

 of the maximum deviations are virtually the same — the Bay of Bengal has 



