60 FR1DTJOF NANSEN. M.-N. Kl. 



that was very different from the water at 20 metres, where a temperature 

 of i.i i C. and a salinity of 34.11 / 00 were observed, giving a density of 27.34. 

 It seems probable that in the upper homogeneous layer, about 10 

 metres thick, the ice and the uppermost water-strata, with a thickness of 

 some few metres, were moving in one direction, while the water in the 

 lower part of the layer, about 10 metres below the surface, moved ap- 

 proximately in the opposite direction, in order to compensate the surface- 

 movement. 



While the central vector-diagrams for 20, 50, and 100 metres (Figs. 

 39 41) represent only the relative movements of the water at these depths, 

 the diagram for the greatest depths, between 300 and 470 metres (Fig. 43) 

 represents the actual movement of the ice-floe, provided that the move- 

 ments of the water at these depths were negligible. If this be correct, 

 it is simple, by a combination of this diagram with the other diagrams, to 

 construct the vector-diagrams for the actual movements at the other depths. 

 The simplest method is to take the diagram, e. g. for 20 metres, and place 

 it on the diagram giving the drift of the ice. If the mark for any certain 

 hour of the one diagram be placed exactly over the corresponding mark 

 for the same hour of the other diagram, and both diagrams are placed 

 in their correct position with their north-south co-ordinates parallel to each 

 other, then the line between the centre of the diagram for 20 metres, and 

 the centre of the other diagram, give the velocity and direction of the 

 actual movement at 20 metres at that moment. By doing this for all hours 

 during the period of observation, we obtain a vector-diagram showing the 

 velocity and direction of the actual movement at 20 metres during that 

 time. The directions are those from which the water was moving. If, 

 however, it is desired to have the direction towards which the water 

 was moving, all that is necessary is to turn the vector-diagram upside down, 

 so that North takes the place of South. In this manner our vector-diagrams, 

 Figs. 46 50, showing the actual movement at 20, 50, and 100 metres, 

 have been constructed. The black discs in these figures indicate the results 

 obtained by the observations at those depths. 



By taking the values of velocity and direction from these vector- 

 diagrams, the curves of velocity and direction, Figs. 51 & 52, have been 

 constructed, and also the progressive vector-diagrams, Fig. 53, showing the 

 actual movements at the depths of 20, 50, and 100 metres. 



As might be expected the total movements at these depths, given by 

 the latter diagrams, agree fairly well with the mean velocity and direction 

 of the actual movements at the different depths, obtained by the combina- 



