42 CHAMBERLIN— A POSSIBLE REVERSAL [April i8 



ture density of the Antarctic regions from which the Pacific and 

 Indian oceans are not separated by appreciable barriers. 



An interesting ilhistration of the close balance between salinity- 

 density and temperature-density is presented by the saline waters 

 that issue from the Mediterranean in which evaporation is in excess 

 of combined precipitation and inflow from adjacent lands. As a 

 result, the concentrated waters that form the deeper body of the 

 Mediterranean creep out through the bottom section of the Straits of 

 Gibraltar, while the upper section is occupied by a compensating 

 inflow from the Atlantic. Although the straits are shallow, the out- 

 creeping current does not appear in the upper horizons of the adja- 

 cent Atlantic waters, according to Buchan's charts, but descends to 

 depths of three thousand to five thousand feet before it finds a 

 horizon of density-equilibrium. It then spreads westerly in a great 

 spatulate wedge across the north Atlantic and occupies the larger 

 part of its area between the depths of four thousand and five thou- 

 sand feet. (See the maps of Buchan.) It is warmer and more 

 saline than the normal oceanic waters at its horizon, and lies on 

 colder but less saline waters below. 



These and similar phenomena point to a notable closeness of the 

 balance between the density effects of salinity and of temperature 

 respectively. More saline but warmer waters both overlie and 

 underlie less saline but colder waters. On the whole, however, at 

 present, the temperature effects are dominant and cold waters occupy 

 the abysmal depths of all the great oceans. 



A comparative computation of salinity-effects and of temperature- 

 effects on density, from such data as are now available, leads to a 

 similar conclusion relative to the closeness of balance between the 

 opposing agencies, but this cannot be entered upon here. 



Now, as previously remarked, the geological record gives good 

 evidence that in the majority of known periods the temperatures in 

 the polar regions were subtropical or warm temperate. Freezing 

 must apparently have been a trivial factor, if not quite absent, and 

 low temperature was robbed of its chief densifying effects. Evapo- 

 ration in the zones of descending air currents in low latitudes must 

 apparently have been operative, in some degree at least, to furnish 

 the geological agencies which the record implies. Deposits of salt 



