Filter 
The precipitation of calcium sulfate as the temperature 
increased was the main difficulty in working with natural 
sea water. If the precipitation was allowed to occur in the 
capillary, it might become plugged completely or else the 
readings would have been very erratic. From this point 
of view therefore, it was impossible to measure the viscosity 
of natural sea water by this method. The next best thing 
seemed to be to remove some of the calcium sulfate beforehand 
by heating the solutions to 150°C and filtering them under 
pressure. As 150°C was the highest experimental temperature, 
there was no subsequent precipitation during test runs. 
Basically, the filter was a pressure vessel divided into 
two chambers by a fiberglass filter paper which was 
supported by wire screens (Figure 25). All metal parts were 
of 316 stainless steel. Each chamber consisted of a 2" long 
industrial glass pipe spacer, 2" i.d. with 3/8" wall. These 
were closed at the outer ends by Steel end-preces, with 
washers of ethylene-propylene rubber preventing direct con- 
tact between the ends of the pipe spacers and the metal end 
pieces. The seals between the pipe spacers and the end 
pieces were made by o-rings around the pipe compressed on to 
the end pieces by circular clamps. 
The center piece contained the filter paper held between 
two wire gauze screens and was shared by both chambers. The 
method of sealing was the same. The entire assembly was 
held together by three rods which ran the length of the 
filter and were bolted across the end preces. Ihe only 
difference between chambers "A" and "B" was that there was 
a capillary tube welded through the center of the endpiece 
in "B" and this was connected to a small throttle valve on 
the outside. 
To operate, the filter was assembled with chamber "A" 
almost full of sea water. Then it was held in an oven at 
TS0°G for several hours. To filter, at was inverted so 
that "A'' was then uppermost and the pressure released 
slightly in "B'. A slow release of pressure was enough to 
drive the liquid through the filter. 
Results 
Distilled water runs and calibration constants: The 
calibration constants for the two viscometers used are given 
below. The viscometer with the swagelok fittings was used 
76 
