47 



deeper reference surfaces have not been used for this region chiefly 

 because of the errors which would have been introduced through neces- 

 sary extrapolations in the many cases where observations did not ex- 

 tend to deeper levels. Since, in tliis survey, all stations were occupied 

 to as near bottom as was practicable the question arose as to whether 

 the deepest common level would be the best. In order to find the 

 most nearly motionless level in the area as a whole the least variability 

 in vertical differences in dynamic height for any 500-meter layer was 

 selected as a criterion. This investigation showed that the desired 

 level was at about 2,000 meters. This method was checked by in- 

 spection of horizontal gradients of anomaly of specific volume. Figures 

 45 to 48 show the horizontal distribution of anomaly of specific volume 

 for the 1,500, 2,000, 3,000, and 4,000 decibar surfaces and demonstrate 

 that the horizontal gradients are least in the 1,500 and 2,000 decibar 

 surfaces. It will be seen that with the exception of the region im- 

 mediately south of Cape Farewell, the 2,000 decibar level is indicated 

 as most nearly motionless. 



This level was then used as a reference surface in the preparation of 

 the dynamic topographic maps of the 3,000, 3,500, and 4,000 decibar 

 surfaces shown in figures 49 to 51. Admittedly, because of the small 

 density gradients involved and the probably undulatory character of 

 the motionless surface if one exists, the accuracy of these maps is 

 highly questionable as to velocity. It is believed, however, that the 

 indicated directions of flow are reliable in the main. They show the 

 southward movement along the American side of the basin consistent 

 with other criteria of the movement of the deep and bottom waters of 

 the Labrador Sea. They also show that this water is joined by other 

 water from the southern side of the end of Reykjanes Ridge. As to 

 the origin of the latter stream, the absence of observations to the 

 eastward of the area surveyed does not permit of more than a conjec- 

 ture. The higher temperature points to some source other than the 

 northwestern arm of the Labrador Basin. Possibly, the water in 

 question comes from the northeastern arm of the Labrador Basin to 

 the latitude shown in figures 49-51, by a route eastward of the fortieth 

 meridian. There is even the possibility that it crosses the Mid- 

 Atlantic Ridge from the eastward between latitudes 50° and 55° N. 

 The need for a further survey to determine this question is plainly 

 indicated. 



Bottom temperatures. — As previously mentioned, the observations 

 were carried to as near bottom as was practicable at all stations. 

 Thus we have data on which to base good extrapolations to bottom, 

 supported by 20 cases in which the lowest bottle bore evidence of 

 actually having been on bottom. (Such as mud or sand in the lip of 

 the bottle, in the water sample or in the glass case of the unprotected 

 thermometer.) From this material it has been possible to construct 



