70 REDFIELD, KETCHUM AND RICHARDS [CHAP. 2 



this zone will be referred to as the oxygen-minimum layer. In high latitudes 

 the oxygen-minimum layer is absent and the vertical distributions are more 

 uniform. 



There has been general agreement that the oxygen-minimum layer arises 

 from a dynamic equilibrium between the rate at which oxygen is removed 

 from the respective water layers and the rate at which it is renewed by the 

 circulation of water (see Richards, 1957). One group of investigators has 

 emphasized the latter factor and has concluded that the oxygen content is 

 reduced to the greatest extent where the motion of water is minimal. Seiwell 

 (1937) and Sverdrup (1938) have pointed out that this condition is contrary 

 to the facts in certain situations and that the observed distributions of oxygen 

 may be accounted for by assuming various other suitable relations between 

 the dynamic factors. 



When the concentration of any nonconservative element in a moving mass 

 of water is considered, its value will depend not only on the changes which 

 have resulted from the balance of dynamic factors operative along its course, 

 but also on the initial condition of the water. If the velocity of flow is great 

 the change due to biological activity may be small and the observed concentra- 

 tions may depend largely on a state established at the point of origin of the 

 water mass. For example, a core of water of low oxygen content present in the 

 Gulf Stream off New England may be tracked back toward its origin as far 

 south as the Caribbean Sea (Richards and Redfield, 1955). It has been shown 

 that the vertical distributions of oxygen and phosphorus in the Atlantic arise 

 in substantial part from conditions existing near the sea surface at the places 

 where the several water layers are formed, as well as from the decomposition of 

 organic matter in the course of their flow at depth (Redfield, 1942). 



The demonstration depends on separating the phosphorus present into its 

 several fractions, as discussed on page 33. The principal motion of the water, 

 by advection and horizontal eddy diffusion, may be assumed to be along 

 surfaces of equal density. It is consequently instructive to plot the data as a 

 function of a t rather than depth, in which case the principal motion is in the 

 horizontal and the continuity of motion becomes apparent. The convention is 

 not adapted to represent conditions in the deep water with a<>27.8 because 

 the variation in at is insignificant. The conditions below 2000 m are better shown 

 in the usual manner as a function of depth. The conditions at the depth of 

 200 m may be taken to represent the state of the water masses at the time of 

 their origin. At this depth seasonal variations in the surface layers are largely 

 eliminated. 



The distribution of inorganic phosphorus along a north-south section of the 

 Atlantic Ocean is shown in Fig. 17a, using this convention. The close relation 

 of phosphate concentration to density distribution is striking. Maximal con- 

 centrations occur in a zone between at 27.0 and at 27.5 which meets the sub- 

 surface in the subantarctic region between 42°S and 50°S, i.e. between the 

 approximate positions of the subtropical and antarctic convergences. This zone 

 will be referred to as the intermediate layer. Maximum concentrations occur at 



