observed with boiling condensate flowing on both probes. 
tivand|i2"should be! the isame during °solution runs’ and 
distaliledswatervruns. J1f they do notWcorrespond, ‘the 
boilingzises topped, Vandtthe ebulliliometer is! checked for 
possibile malfunctions. . The difference; T1-T2, indicates 
Cheeconrection) to be applied: to the Pl-1T2%obsexnved ‘when the 
solution is flowing on the probe. 
hemperatunes ares cecorded tor at least half -an’ ‘hour. 
ithensthe valvelonvthe condensate line to *the “solution 
temperature probe is iclosed, ‘the pump heater is switched on, 
and the power “adjusted!© (Ihe condensate ‘supply reservoir ‘to 
SOMUttOn! pork tousmstile Lulelow? the (heater init is “kept on 
for some time after the flow into the reservoir is stopped 
so that condensate does not spill over and dilute the 
solutiononthe probe. The system requires about 15 minutes 
to reach equilibrium after the pump is started. The sampler 
iineas opened when a steady tlow2is obtained oll, T2<and 
MriZeanre recorded when circulation 1s maintained. 
Sampling procedure - First, the sampler is filled 
with fresh solution and put in place on the ebulliometer. 
When the pump maintains a steady flow, the valves are 
opened and care is taken to remove all air bubbles in the 
sample line. With the continuous flow, the temperatures 
are observed. When T1-T2 remains steady for a satisfactory 
length of time (about half an hour), the valves are closed. 
The sampler is taken out and cooled to room temperature. 
The flange and coupling ends are taken out and the exterior 
of the sampler is washed and dried thoroughly. This is 
done to prevent any contamination of the sample. The dry 
ends are put back on the valves and the valves opened 
slowly so that the sample reaches atmospheric pressure. 
Then the contents are emptied into a clean dry flask and 
labelled. The sampler is put back on the ebulliometer 
with fresh solution. 
Once the set of readings for one pressure and samples 
have been taken, the pressure is changed to the next 
desired value and the procedure continued. 
Analysis - In the initial stages, the samples 
were analyzed by the Knudsen method (12), i.e. titration 
with silver nitrate solution in standard equipment. This 
hasan yaccunacy of +0502)0/700 injchlorinity,or salinity: 
In the later stages a Bissett-Berman, Model 6230, Inductive 
Salinometer was used. The accuracy is claimed to be 
+0.005 parts per thousand (0/00) in salinity. 
47 
