4
Bruce A. Warren
(1962) told the whole story, with references and quotations], and so it had no influence
on the development of oceanography.
2. WATER PROPERTIES
Direct measurement of subsurface currents did not become feasible in any significant way until the mid-twentieth century, so oceanographers studied what they could
actually measure, namely, water properties—at first only temperature, but then salinity and oxygen concentration as well. Before the Second World War, high-quality
temperature measurements were mostly made with mercury thermometers. Until the
introduction in 1874 of reversing thermometers for determining both temperature in
situ and depth of observation (formerly estimated from amount of line payed out),
subsurface temperatures were measured with maximum–minimum thermometers
at depth, or from water samples brought to the surface in insulated water bottles.
Maximum-minimum thermometers could never detect temperature inversions, and
so, for example, the Challenger Expedition (1873–1876), using only these during the
first part of the voyage, missed the temperature maximum associated with the North
Atlantic Deep Water in the western South Atlantic. On the other hand, the warm deep
water in the Southern Ocean had already been found in the late eighteenth and early
nineteenth century with insulated water bottles (G. Deacon, 1984, p. 23).
But Nansen, engaged with his Fram Expedition to the Arctic (1893–1896),
realized that, to get accurate in situ temperatures from insulated water bottles, the
temperature read on deck had to be increased by the amount of adiabatic cooling the
sample had undergone in being raised from depth. Ekman (1905) therefore prepared
tables and graphs for the addition. Helland-Hansen (1912) then reversed the calculation, subtracting Ekman’s increments from in situ temperatures to obtain a quantity
preserved during adiabatic displacements, which, adopting current meteorological
terminology, he called the potential temperature.
The salinity determinations made on the Challenger were actually shipboard
measurements of specific gravity at a standard temperature, and were not very reliable.
Since the relative proportions of the major ions dissolved in seawater had been found
during the course of the nineteenth century to be virtually constant around the ocean,
oceanographers in the 1900s began titrating for chlorinity instead. At very best (not
always achieved) the accuracy of titration salinities was ± 0.02. This was adequate
for studying the Atlantic, but it is close to the level of actual salinity variation in
the deep Pacific, which was thus obscured from view until conductivity methods of
salinity measurement, an order of magnitude more accurate, became practical in the
1950s.
Hydrographic stations, comprising discrete, serial observations of these properties in the vertical, either to middepth or to the bottom, were occupied along lines
across currents, basins, or entire oceans. The sections of property fields so obtained
showed the shapes and sizes of the vertical variations, and the quasi-horizontal
Bruce A. Warren
(1962) told the whole story, with references and quotations], and so it had no influence
on the development of oceanography.
2. WATER PROPERTIES
Direct measurement of subsurface currents did not become feasible in any significant way until the mid-twentieth century, so oceanographers studied what they could
actually measure, namely, water properties—at first only temperature, but then salinity and oxygen concentration as well. Before the Second World War, high-quality
temperature measurements were mostly made with mercury thermometers. Until the
introduction in 1874 of reversing thermometers for determining both temperature in
situ and depth of observation (formerly estimated from amount of line payed out),
subsurface temperatures were measured with maximum–minimum thermometers
at depth, or from water samples brought to the surface in insulated water bottles.
Maximum-minimum thermometers could never detect temperature inversions, and
so, for example, the Challenger Expedition (1873–1876), using only these during the
first part of the voyage, missed the temperature maximum associated with the North
Atlantic Deep Water in the western South Atlantic. On the other hand, the warm deep
water in the Southern Ocean had already been found in the late eighteenth and early
nineteenth century with insulated water bottles (G. Deacon, 1984, p. 23).
But Nansen, engaged with his Fram Expedition to the Arctic (1893–1896),
realized that, to get accurate in situ temperatures from insulated water bottles, the
temperature read on deck had to be increased by the amount of adiabatic cooling the
sample had undergone in being raised from depth. Ekman (1905) therefore prepared
tables and graphs for the addition. Helland-Hansen (1912) then reversed the calculation, subtracting Ekman’s increments from in situ temperatures to obtain a quantity
preserved during adiabatic displacements, which, adopting current meteorological
terminology, he called the potential temperature.
The salinity determinations made on the Challenger were actually shipboard
measurements of specific gravity at a standard temperature, and were not very reliable.
Since the relative proportions of the major ions dissolved in seawater had been found
during the course of the nineteenth century to be virtually constant around the ocean,
oceanographers in the 1900s began titrating for chlorinity instead. At very best (not
always achieved) the accuracy of titration salinities was ± 0.02. This was adequate
for studying the Atlantic, but it is close to the level of actual salinity variation in
the deep Pacific, which was thus obscured from view until conductivity methods of
salinity measurement, an order of magnitude more accurate, became practical in the
1950s.
Hydrographic stations, comprising discrete, serial observations of these properties in the vertical, either to middepth or to the bottom, were occupied along lines
across currents, basins, or entire oceans. The sections of property fields so obtained
showed the shapes and sizes of the vertical variations, and the quasi-horizontal