DAVIS STRAIT AND LABRADOR SEA 193 



senting the annual cycles of their production. As the rates of south- 

 ward progress of these waters are most probably different, it is not 

 to be expected that an exhaustive survey made in any one year will 

 show a correspondence, in a horizontal projection, between the loca- 

 tions of successive temperature minima in the deepest bottom water 

 and the location of successive salinity maxima in the deep water. 

 Furthermore, the gradual mixing with surrounding water masses 

 tends to erase the identity of these maxima and minima as they pro- 

 ceed further from their sources. 



The arm of the Labrador Basin between southeast Greenland and 

 Eeykjanes Ridge is a possible source of formation of saltier deep 

 water such as is formed in the colder parts of the year in the north- 

 west arm of the Labrador Sea, that is, that portion southward of 

 Davis Strait Ridge. The saltier deep water, if any, so formed in 

 this northeastern arm of the Labrador Basin, is contributed, at least 

 in part, directly to the southern part of the Labrador Basin. Some 

 of the deep water so formed at this source may occasionally round 

 the southern end of Greenland and enter into the deep-water circu- 

 lation of the central part of the Labrador Sea somewhat northward 

 of Cape Farewell. A part of the bottom w^ater, some of which is 

 possibly formed in the wintertime in the northeast arm of the Labra- 

 dor Basin according to our idea, probably escapes into the Atlantic 

 Basin eastward of longitude 38° W. through possible deep channels 

 which may cross the southwestern end of Reykjanes Ridge. A minor 

 part of the bottom water of the northeastern arm of the Labrador 

 Basin may possibly enter the central part of the basin around the 

 southern end of Greenland. 



The salinity maxima representing annual cycles of production of 

 deep water from the northeastern source are apparently, because of 

 the location of their sources nearer to unmodified Irminger Current 

 water, usually higher in salinity than the maxima of the deep water 

 produced in the northwestern arm of the Labrador Basin. 



An assumption of no horizontal motion, which was made when 

 considering the area of wintertime vertical convection to bottom is 

 of course inaccurate and justifiable only because of the complete 

 absence of midwinter observations from the area in question. The 

 reality of such horizontal components is undoubted and, in fact, re- 

 quired for the reestablishment of summertime equilibrium. Their ef- 

 fect is to retard vertical convection and to restrict the area in which 

 it is produced. It is emphasized that their equalizing effect is de- 

 pendent upon the removal of their driving source with the cessation 

 of vertical convection, that driving source being the abstraction of 

 heat. 



In conclusion the authors wish to call attention to the hypothetical 

 nature of many other parts of this chapter. Some of the main fea- 

 tures, however, such as the formation of bottom water during winter- 

 time in the Labrador Sea and its eventual run-off into the deeper 

 Newfoundland Basin, are certainly indicated by the observational 

 data already collected. Final confirmation awaits future surveys 

 when subsurface observations must be made during the coldest time 

 of winter in the Labrador Sea. 



