SOUTH AUSTRALIAN WATER SUPPLY. 631 



The third state is that either the water passes too slowly or the 

 hed becomes so foul as to be dangerous. It then becomes necessary 

 to prepare a new bed either by thoroughly cleansing the old mate- 

 rial or providing fresh. The great objection to filter beds appears 

 to be that there is always a chance of danger from the concentration 

 of germs in them, and from the possibility of the films of organic 

 matter affording a suitable material for the growth of pathogenic 

 organisms in large quantities, some of which may be passed into 

 the service by some irregularity of action. Domestic filters, unless 

 frequently cleansed and intelligently used, are even more liable to 

 the same danger. This danger arises from the fact that a few 

 germs of a disease introduced into the system are generally easily 

 overcome without inconvenience, but if introduced in greater 

 numbers they produce their specific illness. 



4. Anderson's Process. — This process consists in passing the 

 water through a revolving cylinder containing a quantity of small 

 pieces of iron ; air is in some cases pumped through the water at 

 the same dme. After leaving the cylinder, in which it remains ou 

 an average three and a half minutes, it is again aerated if necessary 

 to peroxidise the iron which has been dissolved. The oxidised 

 ii'on precipitates and subsides quickly, carrying down with it much 

 of the organic matter present, and the water may be readily filtered 

 or allowed to clarify by subsidence for a day or two. 



From the analyses in Table 5 it will be seen that the Adelaide 

 service water treated by this process is changed from an impure 

 water to one of a high degree of organic purity, leaving nothing to 

 be desired either chemically or biologically. Tests by this process 

 have been made with much more impure waters, such as those of 

 Kapunda and Nelshaby, and it was found that a clear and colorless 

 effluent could be obtained even with these waters if aeration 

 during the agitation with iron were employed. As an example of 

 the especial adaptability of this process for this colony I will point 

 out how it mitjht be advantageously used for the large reservoir 

 at Happy Valley. This reservoir will have a capacity of about 

 3,000 million gallons, and during the four or five winter months 

 about thirty million gallons a clay can be taken from the Onka- 

 paringa River. To treat this quantity the water would be passed 

 through a battery of revolving cylinders having a total capacity of 

 80,000 gallons, in which it would remain on an average three and 

 a half minutes ; after leaving the cylinder the dissolved iron would 

 be peroxidised by forcing air through the water, which would then 

 be led to a settling reservoir having a capacity of thirty million 

 gallons. Two such settling reservoirs would be required, one being 

 filled while the clear water was running from the other into the 

 large reservoir. The reservoir would thus be filled in about three 

 months. In Table 2 it will be seen that the Onkaparinga water 

 during these months is about half as saline and less than half as 

 hard as the average Hope Valley reservoir water, but is much more 

 impure. I have calculated from the literature on the Anderson 



