Decembek 9, 1898.] 



SCIENCE. 



813 



earth, as pictured in the minds of most 

 people. 



Tracing the actual tidal wave as best we 

 can by direct observation, we find that 

 by starting at a point off the west coast of 

 South Africa the wave is separated into 

 two waves, one of which goes westward 

 and the other eastward. If we follow that 

 portion of the wave which has the least ob- 

 structed path, passing across the open 

 Pacifip to the southward of Australia and 

 to the Cape of Good Hope, we iind that 

 after thirty-six hours it has not yet circum- 

 navigated the globe, but instead his just 

 reached the neighborhood of Cape Horn 

 and is there apparently lost in a collision 

 with an eastward-hound tidal wave, which 

 started in the Pacific, off the west coast of 

 South America, twenty-four hours later than 

 the wave which we have followed. Look- 

 ing to other portions of the oceans we find 

 that the tidal wave moves northward rather 

 than westward over the whole Atlantic 

 and a part of the Pacific — in fact, on about 

 one-third of the total ocean surface ; and 

 that the progress has a decided eastward 

 tendency in at least two large areas, in the 

 Arctic ocean north of Europe, and in the 

 Pacific off the west coast of South America. 



If we examine the relative amplitudes of 

 the harmonic components of the tide at 

 different stations we shall find further 

 strong evidence of the radical modifications 

 made in the astronomical tide by the in- 

 fluence of shores and bottom. As a typical 

 case we may note that in the Atlantic the 

 semi-diurnal components are large as com- 

 pared with the diurnal components, so that 

 there are always two tides of nearly the 

 same height per day, whereas in the Pacific 

 the amplitude of the diurnal components 

 approaches that of the semi-diurnal com- 

 ponents and in some cases exceeds it, and 

 as a result the two Pacific tides of the same 

 day are in general of decidedly different 

 Iieights and in certain extreme cases but 



one tide occurs per day. In one of the ex- 

 treme cases, at Batavia, Java, the com- 

 ponent which has one high water per sidereal 

 day is more than six times as large as the 

 component having two high waters per 

 lunar day, although in the astronomical 

 tide the latter predominates over all others. 



Out of the confiict with the shores and 

 bottom the astronomical tide preserves only 

 its periods, and hence in making tidal pre- 

 dictions at a given station theory can 

 furnish us nothing but said periods. In 

 the confiict no period is lost, though the 

 amplitudes corresponding to certain periods 

 may be greatly decreased or increased, and 

 no new periods are known to be produced 

 save certain multiple periods, or overtones, 

 so to speak, produced by friction, and an 

 annual period fixed probably in the main 

 by meteorological causes. 



When the prediction is made it applies 

 to one point only, the point at which the ob- 

 servations were made, and with our pres- 

 ent lack of ability to predict the effect of 

 boundaries upon the range, the shape, 

 and the rate of progress of the tidal wave 

 that prediction can be extended even by 

 careful study of charts but a very few miles 

 from the stations of observation before ac- 

 quiring large errors. 



The effects of the boundaries— bottom 

 and sides— can best be studied in bays and 

 rivers, in bodies of water in which observa- 

 tions are available at many points, and in 

 which the direct effect of the Moon and Sun 

 is small as compared with that of the wave 

 transmitted from the ocean. The tide 

 tables and charts issued by the Coast and 

 Geodetic Survey furnish many fine oppor- 

 tunities to study this problem. For ex- 

 ample, the charts issued by that Survey 

 give complete and detailed information as 

 to soundings in the Chesapeake Bay and its 

 tidal tributaries, and the tide tables give 

 complete harmonic data for Old Point Com- 

 fort, Baltimore and Washington, and data 



