ATLANT. DEEP-SEA EXPED. 1910. VOL. i) PHYSICAL OCEANOGRAPHY AND METEOROLOGY 



71 



and assuming the temperature at 9788 metres to be as 

 observed (2-60^), or if ttie temperature at ttie latter depth 

 were 0-323'' lower than it is, with ,a= 1-61°. In the 

 first case {/a = 1-928=, , t, - 2-60°) we obtain 

 S' = ft — 1-321°, and in the second (r„ = 1-61°, 



a — ^ O i > O 



r'b ^ 2-277 °): 0^0' = 1-016", the difference be- 



a — > o b — >• o 



tween tiiese values being 0-305° or the same as is found 

 between the two actual values of (•). 



m — ^ o 



The conditions demonstrated by the observations from 

 the Bougainville Deep are analogous. Here the salinity 

 is, also according to Schott, uniformly equal to 34-69 %o 

 between 5000 and 8400 metres. The minimum value of 

 is found at 5000 metres, and the state of "over-adia- 

 bacy" in the deeper layers is nearly the same as in the 

 Philippine Deep. The following temperatures would give 

 adiabatic equilibrium between 5000 and 8400 metres: t„ 

 — 2-107° instead of 1-81° if n, — 2.66° as observed, or 

 Tb =■ 2-358° instead of 2-66° if r^ = 1-81° as observed. 

 In the first case we have = 0== 1-631° and 



a — > b — >• 



in the second = 1-345°, the difference being 0-286°. 



These examples show that e.xcept in cases of adia- 

 batic equilibrium the differences found by a comparison 

 of potential temperatures of different water-masses vary 

 according to the pressure to which the potential tempera- 

 tures are referred. It is convenient to take one atmos- 

 phere as the pressure of reference, and to calculate the 

 values of 0. When nolliing is stated to the con- 



m — -> o 



tiary, potential temperature meatis the temperature which 

 a water-particle attains when it is raised adiabatically 

 to the surface of the sea. A difference between two such 

 potential temperatures does not show the exact amount 

 by which either of the two temperatures observed has 

 to be reduced in order to show adiabatic equilibrium. 

 This point may be of some interest when deep-sea tempe- 

 ratures are determined within a few thousandths of a 

 degree (as can now be done), but otherwise a comparison 

 of potential temperatures referred to the surface is suffi- 

 cient for our discussions of the temperature conditions 

 in the deep waters of the oceans. 



There are two main questions which arise in this 

 connection: — 1. What is the temperature of the deep 

 water before descending, if we assume that it is cooled 

 somewhere at the surface and that its temperature alters 

 on the way downwards merely as a result of adiabatic 

 processes? F^ven if these assumptions do not actually 

 hold good, it may be useful to compare the potential 

 temperatures from different levels and stations so far as 

 the deep water is concerned (cf. section 35j. — 2. Is the 

 deep water in a state of stability or not, if the tempera- 



ture increases downwards and the salinity is uniform? 

 On the latter assumption we have a state of stability until 

 the vertical gradient of temperature reaches the adiabatic 

 gradient. With a larger gradient of temperature in situ 

 the conditions are instable and a vertical convection takes 

 place. 



Provided that the above-mentioned observations from 

 the two deeps in the Pacific are nearly correct, we have 

 here an example of instability which must be due to 

 constant heating from the bottom. By means of these 

 observations the virtual coefficients of friction and of tem- 

 perature conductivity may be computed (both coefficients 

 probably being in this case identical) as pointed out by 

 W. Schmidt [1917] and more recently by Hi;sselbero. It 

 has been assumed that the salinity is uniform in these 

 deep strata, but it is open to question whether this is 

 true, even if the amount of ciilorine be constant. In this 

 connection it may be mentioned that probably the quan- 

 tity of lime is comparatively great, and increasingly so 

 downwards, in the deep parts of the ocean, making the 

 total salinity and density somewhat higher tiian a titration 

 of chlorine shows. Not impossibly, therefore, the density 

 at the potential temperature and atmospheric pressure — 

 which density may be termed the "potential density" — 

 approaches uniformity at the depths in question. ') 



3S. The Deep Water of the North Atlantic. 



Defant [1928] has proposed to use for the different 

 groups of water in the ocean a terminology analogous 

 to that used for the atmosphere: the troposphere including 

 the water-layers from the surface downwards as far as 

 their temperatures exhibit vertical variations in any es- 

 sential degree, and the stratosphere with but small 

 variations of temperature. The uppermost part of the 

 troposphere, on an average from the surface down to 

 about 200 metres, is called the zone of variability 

 ("Storungszone"). Such a grouping is very convenient, at 

 any rate formally. It may sometimes be better to use the 



') In a paper just published, 0. WuST [1929] has discussed the 

 observations from the Pliihppine Deep and the Bougainville Deep, 

 and has corrected the values given by Schott, quoted in the table 

 above. According to WOst the temperature in the Philippine Deep 

 rises from 1-55° C. at 5000 metres to 2-iS'' C. at 9788 metres, while 

 the salinity decreases from 34-()iS to 34-66 "/oo. From the Bougain- 

 ville Deep WOst finds at 5000 metres -m =- 1-97° C. and 5 = 3-1.71 

 °/oo, and at 8-100 metres -m ^ 2-60° and S = 34-67 ",00. These 

 values mean a smaller deviation from adiabatic equilibrium than sup- 

 posed in the discussion above, but the instability is nnich the same on 

 account of the decrease downwards of salinity. My intention here, 

 however, has only been to use the data as a formal example to 

 elucidate some general principles, and for this purpose it is unne- 

 cessary to discuss the reliability of the absolute values of -.,„ and S. 



