DAVIS STRAIT AND LABRADOR SEA 191 



the bottom- water constituent. As has been mentioned, the northern 

 limit of the area in which vertical convection to bottom probably 

 occurs is only slightly north of Godthddb station 24 and is probably 

 closer to it than to Manon station 984 (fig. 135). A southern limit at 

 about 55° N. latitude might be postulated from a consideration of the 

 horizontal motion of the Atlantic Current border. However, be- 

 cause of the tempering effect of the more southerly latitude on the 

 severity of the winter, the southern limit of the area of vertical 

 convection to bottom lies more to the north and is estimated to be 

 between General Greene station 2035 and Godthaab station 10 (fig. 

 135). Such limits would seem to be borne out by a consideration of 

 the midlongitudinal salinity section (fig. 143) and the longitudinal 

 oxygen section (fig. 148). The area in question is shown shaded on 

 the small inset on figure 149. This shaded area is considered by the 

 authors to represent the region in which the bottom water of the 

 Labrador Sea is most probably formed in the wintertime. It is an 

 area whose size will vary from winter to winter and in some years 

 will certainly be smaller and in other years may possibly be some- 

 what larger than the area shown on figure 149. 



An increase in the density through an increase in salinity resulting 

 from ice formation is a factor which assists wintertime convectional 

 sinking as has been pointed out by Helland-Hansen and Nansen 

 (1909) and Mosby (1934). The areas of ice formation in the north- 

 western North Atlantic, hoM^ever, are largely non-coincident with the 

 area in which we have assumed the bottom water of the Labrador 

 Sea to originate, and therefore this phenomenon of salt concentration 

 is considered inconsequential there. 



Adjacent to our area of bottom-water production, particularly to 

 the north and east, are areas in which vertical convection probably 

 penetrates to considerable depths. Figure 152 shows the temperature 

 distribution found by the Meteor in March 1935 along a section ex- 

 tending southward from Cape Farewell, the data for which wer3 

 kindly supplied by the director of the Institut fiir Meereskunde an 

 der Universtiit von Berlin. An inspection of the section indicates 

 that stations 121 and 122 are in the comparatively quiet water north 

 of the Atlantic Current and south of the Irminger Current past Cape 

 Farewell. These stations then should be expected to be most favor- 

 able for the establishment of vertical convection in the wintertime. 

 Furthermore, as the date of the observations was probably onlj^ 

 slightly past the coldest part of the winter, one might expect to find 

 evidence of vertical convection at stations 121 and 122 if it occurs 

 in this region. Such evidence seems to be present, for at station 



121 between about 725 meters and about 1,650 meters and at station 



122 between about 1,100 meters and about 1,850 meters, the tempera- 

 tures actually observed were slightly lower than the maximum tem- 

 perature necessary to produce vertical convection to those depths on 

 the assumption of no horizontal motion and on the basis of average 

 observed salinities and densities. The observed densities at stations 

 121 and 122 showed a very weak stability, the change in o-t from 

 surface to 2,000 meters being but 0.03 and 0.04, respectively. A 

 slight apparent instability was found at about 1,500 meters at sta- 

 tion 121. These densities combined with the foregoing indicate that 

 vertical convection extended to depths of about 2,000 meters shortly 

 prior to the Meteor's observations. 



