From Madeira to Tristan d'Acunha. 79 



bottom-current traced by H. M.S. " Challenger " from the Falk- 

 land Islands alona the east coast of South America as far as 

 Cape S. Roque, and thence across to St. Paul Rocks, and along 

 the southern slope of the equatorial plateau as far as Station 

 346, with the bottom-temperatures slowly rising from — o°.4 C. to 

 -f o°.4 C, in this long course of several thousand miles, proves that 

 there is such an inflow of cold water from the south towards the 

 north. Such a current, after crossing the equator, would have a 

 tendency to flow eastward and press up against the coast of 

 Africa, owing to the same cause which compels the Arctic current 

 to press up against and to flow along the coast of North 

 America. Unfortunately, we do not possess on the African 

 coast such a complete series of temperature-soundings as that 

 furnished by the officers of the U.S. Coast Survey, and future 

 observations can alone show how far the above theoretical con- 

 clusions are supported by facts. 



The portion of the Atlantic section from Stations 84 to 95 is 

 a continuation southwards of the eastern basin represented in 

 the section from Teneriffe to Sombrero, Station 1 to Station 

 10. It is probable that the combined warm and cold currents 

 flowing: southwards from the Azores encounter, when off the 

 Cape de Verde Islands, the cold equatorial stratum, and turn 

 westwards, along the northern limit of the equatorial belt. 



Before proceeding further, it may be necessary to allude to 

 an appearance in the Atlantic section (Plate 9) which is also 

 strongly marked in the Pacific section (Plate 19). It is the 

 increasing undulations of the isotherms, as we enter the lower 

 strata, at first sight suggesting the completely erroneous idea 

 that the fluctuations of temperature increase with the depth, 

 the very opposite being the case. This appearance is due partly 

 to the excess of the vertical over the horizontal scale, the former 

 being to the latter as 1 200 to i in the Atlantic section, and as 

 1500 to I in the Pacific section — a disproportion inevitable on 



