A scheme to systematize the classification of the hydrography of the oceans developed by Montgomery in the late 1930s. He developed this to overcome limitations he saw in the earlier core layer method of Wust. In this method variable properties (e.g. salinity, temperature, etc.) are examined on surfaces of constant potential density along which it is assumed that maximum mixing and flow occur. The variations in depth of such surfaces can also be used as a diagnostic tool for locating geostrophic currents since rapid changes in depth are indicative of their presence.

The isopycnal method was originally applied such that all densities were calculated relative to the ocean surface, i.e. as either sigma-t or sigma- . This was later modified when it was found that, due to the nonlinearity of the equation of state for seawater, maximum values could exist well above the ocean bottom - ostensibly signifying a hydrostatic instability. The problem was rectified by the use of potential densities calculated relative to different pressures as was required by the situation, i.e. sigma-1 for densities relative to 1000 decibars (db), sigma-2 for 2000 db, etc., a procedure that is still followed today.