of products of elemental graphically determined areas of cross section 

 enclosed between two isotherms and two lines of equal velocity multi- 

 plied by the average velocity in this elemental area and by the average 

 temperature in the elemental area. The first two factors give an 

 elemental volume transport and their summation gives a figure for 

 the volume transport across the section. To reduce the effect of 

 planimeter errors arising from the measurement of small areas a 

 figure for the mean temperature is obtained by dividing the summation 

 of area times velocity times temperature by the summation of area 

 times velocity. The figure thus obtained for mean temperature 

 is then used to multiply the best value of volume transport to obtain 

 the heat transport. It has been considered that this best value of 

 volume transport is obtained from computations of the difference 

 between the second depth integrals of specific volume at the stations 

 near the current boundaries, adjusted by graphicall}^ determined 

 transports between these station verticals and the appropriate zero 

 velocity lines. 



Thus both the mean temperature and the heat transport figures 

 dealt with here do not take into account gain or loss of heat at the 

 water-air interface, the heat of fusion from melting or formation of ice 

 or the heat of compression involved in any vertical component of water 

 motions. It is pointed out, however, that the vertical component of 

 velocity is small compared with the horizontal component, that during 

 the part of the year covered by these observations the seasonal warm- 

 ing of the surface produces a thin layer of large and increasing stability 

 which layer protects the bulk of the Labrador Current from heat 

 exchange with the atmosphere and that the heat of fusion associated 

 with the melting of ice principally affects this surface layer even in 

 the case of large bergs where the disintegration proceeds most rapidly 

 at the surf line and on the air-exposed surface. For the Labrador 

 Current during the summertime, then, the mean temperature and 

 heat transport, derived as explained above, are considered usefully 

 to approximate conservative properties. 



At the Bonavista triangle the volume transport entering the triangle 

 usually differs from the volume transport leaving the triangle. Similar 

 discrepancies occur with heat transport. Some of these discrepancies 

 may be the result of time changes, inasmuch as about 3 days are 

 required to complete the observations. There is also the possibility 

 that there may be transfers across the reference surface of 1,000 deci- 

 bars. In figure 25 the volume transport shown as "mean triangle" 

 is the mean of the volume transports entering and leaving the triangle. 

 Similarly, the mean temperature shown for the mean triangle is the 

 mean of the heat transports entering and leaving the triangle divided 

 by the mean volmne transport. 



Bearing in mind that true normal seasonal variation relationships 

 must be curvilinear, and that the tentative normals presented here as 



48 



