POLYZOA OF WATERWORKS 433 



In this system the water is introduced into shallow filtering- 

 beds, which are of large sujDerficial area and have a bottom of 

 sand through which the water passes befoi'e it is admitted to the 

 pipes. The efficiency of the filter does not depend on a mechanical 

 straining action of the sand, but is a biological process, princi- 

 pally confined to the surface-layer of the sand. In this layer 

 grows an abundant flora, consisting of Algfe and Bacteria ; and 

 the beneficial action is the result mainly of the presence of 

 Diatoms. By means of the slimy or gelatinous investment 

 formed by these organisms, the intervals between the sand- 

 grains ai-e filled up, thus increasing the eflectiveness of the 

 filter as a strainer ; while the surface-film of living organisms has 

 a very marked eflect in reducing the number of Bactei'ia, and 

 pi'obably aftects the character of the water in other ways. As 

 the filter continues in action, the number of organisms in its 

 floor increases and the surface-film grows in thickness in a corre- 

 sponding proportion. The filtering surface also tends to become 

 choked by the accumulation of detritus which has been strained 

 ofl", until from these causes it ceases to transmit the water so fast 

 as is necessary. The filter then has to be put out of action and 

 the surface-film must be removed. The film has to be allowed to 

 re-form for two or three days before the filter is in a condition 

 to be used once more. 



The organisms which are of importance in the filtering process 

 undergo seasonal variations. Thus at one time of year Kemna 

 finds that one species of Diatom is dominant, while at another 

 time the eflectiveness of the layer depends on some other species 

 or group of species. The filter thus needs constant observation, 

 since it may readily get out of working order. An interesting 

 case is recorded by Kemna (99, p. 48), who found that Dipterous 

 larvee {Chironomus) may have an important practical eft'ect. In 

 the observation alluded to, the surface-film was inhabited by 

 numerous larva3, which were living in tubes constructed out 

 of sand-grains. So long as the larvpe inhabited their tubes the 

 filter worked well. But when they left the surface-film in order 

 to undergo their metamorphosis, their empty tubes formed a 

 number of perforations which were naturally not closed by 

 Diatoms, etc., and the filter then ceased to work properly. 

 Information that the insects were undergoing their meta- 

 morphosis was given by the appearance of Swallows circling 

 round the filters in order to feed on the emerging Diptera. On 

 noticing this indication the filter was put out of use and given 

 three days' rest, during which the holes left by the emergence 

 of the Chii-onomus became closed by a new growth of the 

 surface-film. 



The practical importance of sand -filtration is well brought out 

 by the facts which are cited by Kemna in the two memoiis 

 already referred to. One of the most instructive cases is the 

 history of the Hamburg Waterworks during the years which 

 preceded Kraepelin's investigation of the pipe-fauna. In 1842 



