In the deep water of the eastern North Atlantic below 2000 m. the variations with depth of salinity, temperature, density, oxygen, phosphorus compounds and nitrate are quite small. By contrast the silicate content is doubled in a descent from 2000 to 4000 m.

The distinctive behaviour of silicate is revealed by diagrams (Fig. I) relating it to salinity, temperature, density and total phosphorus at station 2659 worked by R.R.S. Discovery II on 12 May 1950 (Armstrong, 1951; Cooper 1952, Table IV). The temperature-salinity diagram (Cooper, 1952, fig. 15, to 1500 m. only) suggests that between 1200 and 2000 m. we have to deal with simple mixing of the mean Gulf of Gibraltar and North Atlantic Deep waters. If silicate concentration were subject only to mixing processes the curves in Fig. I between these depths would be straight lines. They are not—consequently it would seem that solution of either particulate silica or of aluminosilicates may be occurring. As yet, clear interpretation is not possible. At least five hypotheses may be erected to explain, in whole or in. part, the observed distribution: (i) solution of bottom deposits; (ii) solution of ‘clay’ and of silica in suspension; (iii) concentration by vertical partition; (iv) tundra drainage; (v) sinking of surface water. These are examined in turn.