272 Universiti) of California Publications in Zoologij [\'oh. 15 



E. THE THERMOCLINE AND ACCOMPANYING SALINITY 

 DISTRIBUTION 



The temperatui'e iisually decreases continually as the depth in- 

 creases and our observations have never revealed a "reversal," i.e., 

 an increase in temperature with an increase in depth.- The depth- 

 interval in which the temperature changes most rapidly with the depth 

 is called the "thermocline." The water in this intei'val forms a 

 transition layer or boundary separating two masses of water of verj' 

 different temperatures. During the summer months, from May to 

 September inclusive, the thermocline is near the surface and the rate 

 of temperature change, 0?16 C per meter, is comparatively rapid; 

 while during the winter months, from October to April, it occurs at 

 a greater depth, and the rate of change is only 0;07 C per meter 

 (pi. 26, fig. 45). 



In August the thermocline is usually between 5 and 25 meters and 

 the rate of change, 0?30 C per meter, is greater than during any other 

 month, while in February it occurs between 25 and 100 meters, and 

 the rate of change, 0?05 C per meter, is smaller than during any other 

 month (pi. 36, fig. 98). 



The distribution of salinity with respect to depth is closely cor- 

 related with the temperature distribution. During August, when the 

 thermocline is most pronounced, the salinity decreases most rapidly 

 to a minimum at about 30 meters and increases from there to the 

 bottom, while during December, January, and February, when the 

 temperature in the upper 50 meters decreases very slowly with respect 

 to the depth, the salinity is nearly constant, though a minimum is 

 indicated at 30 meter.s (pi. 37. fig. 99). 



The existence of a minimum salinity between 20 and 60 meters. Ht 

 least in the region from Point Conception to the Coronado Islands, is 

 established beyond doubt, since it is clearly revealed by the monthly 

 average values (pis. 29-32). as well as by nearly all of the numerous 

 individual series of observations. A few such series are presented in 

 the following table. 



- A few such cases have Ijeeii inclicateil, but proveil to be due to errors.. 



