In order to make certain corrections, it is necessary to know the absorption coefficient of neutralized sea-water for Since the half-normal HC1 added to the sea-water has about the same absorption coefficient for atmospheric gases as neutralized sea-water, we assume that it does not seriously affect the absorption coefficient of the sea water when 1 c.c. is added to 10 c.c. of the latter. Bohr's data were plotted for this purpose on the assumption that the absorption coeffi cient of neutralized sea-water for is the same as that of NaCl solution of the same Cl content. The isotherms form characteristi cally curved lines, but in the small portion included in figure 22 the curvature was hardly perceptible and was later obliterated by the use of a straight-edge in inking-in the pencil drawing.
It is necessary to know capillary depression of the mercury in B, which is easily determined by opening all stopcocks and measuring the vertical distance between the mercury meniscus in B and the plane of the mercury surface in the leveling bulb. A mercurial barometer with glass scale is preferable, and the capillary depression must be sub tracted from the reading.
In the following table for reducing the final gas-volume to 0°, 780 mm., and dryness, the temperature correction for the glass scale barometer is included. The moist gas-volume at designated temperature and pressure is multiplied by the correction in the table. Degrees and fractions of a degree may be estimated, because makes about as much difference in the correction as 4 or 5 mm. pressure.
The mode of calculation is perhaps beet shown by the following example: The analysis was made on 10 c.c. of sea-water of C1=19.375 per kilo, and the room was 20°, so the absorption coefficient was 0.765. The length of the graduated portion of the burette B was 9 cm., so the meniscus correction (Hg-H20 and was 0.0125 c.c. and the capillary depression 3 mm. The barometer minus the capillary depression was 725 mm. The water above the mercury meniscus was 0.0565+0.0125=0.069 c.c. The total gas (0.54) minus the gas after absorption of (0.179) equals the (0.361 c.c.) as measured directly; but the 0.069 c.c. of water was in equilibrium with a gas, a large fraction (0.669) of which was hence the in that water was 0.069X 0.669 X 0.765 = 0.353. In other words, the in the water = its volume X the fraction+ the absorption coefficient.
Practically all of this was pumped out and then reabsorbed; hence the amount originally pumped out was 0.361+0.0353 =0.3963 c.c., and this amount expanded to 39 c.c. was in equilibrium with 11 c.c. of water at the end of the first shaking. Hence this 11 c.c. finally trapped off in T contained 11 X0.765= 0.0855 c.c. of 39 This added to the 0.3963 c.c. brings the total up to 0.4818 c.c., and this X0.8644=0.4165 c.c. per 10 c.c. or 41.65 c.c. per liter at 0° and 760 mm. This correction of 0.8644 goes outside the table, but the table is probably extensive enough for most determinations, especially at a lower altitude, and the logarithm of the corrections for a more extended range may be found in the Chemiker Kalender or Landolt-Bomstein.
The apparatus was tested by making an artificial sea-water of neutral salts, distilled water that had just been boiled 15 minutes in a quartz vessel and cooled with air passing through it, and some of the used in standardizing the HCI, and weighed with standardized weights. The total as calculated and found in this
solution differed by about 0.5 per cent.
Dole used a simpler (titration) method for determining the in sea-water, but we have not yet had time to compare it with our results. We could not use it in these experiments on account of the large sample required.
The oxygen in the sea-water may be determined on the same sample after the by absorbing it with an alkaline solution of pyro gallol, introduced in the same manner as the NaOH solution, and the nitrogen + other inert gases may be determined by the residue. The whole calculation must be repeated for these gases, using their respec tive absorption coefficients, as given in the following table, calculated from Fox's data, for sea-water of C1=20. Fox does not state whether this is Cl per liter or per kilogram.
One of the most important requirements in such investigations as are described in this paper is accurate temperature control. All of the determinations in Minneapolis were made in a laboratory that was converted into a constant-temperature room by cutting off the steam heat and regulating the temperature by means of a nichrome wire stretched through the air in front of an electric fan and carrying 1,600 watts. The relay controlling the heating current was burned by arch ing because it was too small and we did not obtain a large one in time. But we found it possible to control the temperature to 0.1° by personal attention to a sensitive thermometer (graduated in tenths of a degree) placed in front of a one-sixth horse-power electric fan, and the heating current was controlled by means of a push-button.
The only time the electric heating failed to be sufficient was on two successive days when we were required to finish the experiments at 30°, although the outdoor temperature was about — 30° C. It was then necessary to fill the space between the windows and dark curtains with cotton batting and to turn on the steam heat. The radiator was inclosed in a paper covering and a hole left in this of such size that the temperature did not quite reach 30°, but could be quickly raised to 30° by means of the electric regulation. It was necessary to keep a one sixth horse-power electric fan and one or two other fans blowing on the apparatus all the time in order to keep all pieces of apparatus at the same temperature. The experiments at 10° were postponed until the outside temperature fell to about —30° C. The experiments at 20° caused little worry in regard to temperature regulation, except on some days when the outside temperature rose to about 10°. Owing.to conduction through hollow-tile walls from rooms at about 22°, it was sometimes necessary to leave the window open about an inch. Great care had to be taken then to equalize the temperature in different parts of the room. Several thermometers were compared with a standard and distributed around the room, one or two being placed on every piece of apparatus in order to detect any inequalities in temperature.
Great care was taken to prevent temperature changes in the appa ratus. When necessary to handle it, this was done very gingerly, with the tips of the fingers on the extreme ends. Clamps were devised for holding the apparatus in the various necessary positions. During the night the temperature was controlled within 2° of that in the day by means of a thermostat connected with the steam-heat radiator.