dhe effect of the experimental uncertainty in salinity 
was found by differentiating Equation 24 with respect to 
x 
URE err a pes 
dx 
(29) 
At 80° C the following errors were found: 
Dia Laney, 5.5 0: 00009 
Salami ey ie. 020001 
These errors did not change appreciably with temperature. 
ihe estimated error im the viscosity was +0001). )-This 
was somewhat larger than most of the standard deviations 
obtained for the fitted data. 
Conclusions 
The apparatus was disappointing, as the precision of 
the results was +0.001 centistokes. A precision of +0.0001 
centistokes had been hoped for in order to work with 
electrolyte solutions made up in the lab. Improvements that 
can be made on the existing apparatus are: 
1. Use of the silver/palladium probe to obtain a 
constant high pressure for all the readings. Using the 
probe necessitates the use of hydrogen in the system which 
would reduce the vapor column correction. 
Zs Improvement of the temperature control in the bath. 
One possibility is to put fiberglass insulation around the 
oil bath and use the hot air outside that. 
oe USE VOL a photocell device to. time ihe vVascometcrs. 
Comparison with the work of Korosi and Fabuss (20) and 
Grunberg (15) showed agreement in the viscosity at the same 
1OnLE “Strewoth Locwichan 25.) in preparing their feest 
solutions, Grunberg left out the Ca** completely whereas 
Korosi and Fabuss replaced it with equivalent Mg**. In 
increasing viscosity values the order was (1) Grunberg (as 
would be expected), (2) Korosi and Fabuss, and (3) the 
present work. The agreement between the last two was better 
than with the first, but one would have expected a reverse 
order. <A possible explanation is the organic material in 
natural sea water. 
94 
