29 
Estimated "best fit" curves drawn through the arithmetic and 
geometric averages of the dye patch data collected at Orange County, 
plotted as the log of the most probably number of coliforms per ml 
against time are shown in Figure 3. The points fall surprisingly 
close to a straight line using either means of plotting, and the 
slopes of the lines are essentially the same, 
The other two curves on the graph are plots of individual 
dye patch experiments at Hyperion and Whites Point. The rate of 
decrease of coliforms was less than at Orange County in both cases, 
although the slope of the Hyperion curve is not greatly different. 
The Whites Point experiment differed from the others in two impor- 
tant ways, both of which may have contributed to the less rapid 
disappearance observed. First, the zero time count was lower by 
a magnitude or more than in the other dye patch experiments. Since 
this count was also lower than that of other samples taken in the 
Whites Point boil, it might not be representative of the dye patch 
and an initial sharp decrease may have been masked. Second, the 
dye patch was started at the edge of the inner boil and the patch 
of travel was such that it intersected "younger" sewage being 
released at the outer boil. This could mean that the dye patch 
was being mixed with water of higher coliform count which would 
explain the slower decline in coliform numbers. 
Of the three "atypical" sets of results, one was from an 
Orange County experiment in which the zero-time count was low and 
the chlorinity data showed little sewage present in the water. In 
this instance it can be concluded that the dye was introduced into 
an unrepresentative part of the boil and the results of this experi- 
ment should be eliminated. The other two atypical results were in 
