tapped for this purpose.

     The study was divided into eight runs starting in August,

1986 and ending in February, 1987. During the first three runs,

no effluent from the filter was recycled to the septic tank.

Nitrification was sought in the bottom ash filter and

denitrification in the rock filter. Nitrification worked

reasonably well in the bottom ash filter as could be seen by the

reduced ammonia-N levels of 3.4 to 14.9 mg/l (Average 9.6).

Ammonia-N levels in the septic tank on the other hand were 33.3

to 51.5 mg/l (Average 43.6) However, denitrification in the rock

filter was incomplete allowing effluent nitrate-N concentrations

from 9.4 to 14.6 mg/l (Average 11.9).

      In runs 4 to 8, a portion of the BRBAF effluent was

 recirculated to the septic tank. Total nitrogen removal of the

 system was much better during these runs. For runs 4 - 8 the

 average TN values for the system effluent was 7.2 to 9.6 mg/l

 (Average 8.4) as compared to runs 1 - 3 in which the values

 ranged from 15.8 to 25.7 mg/l (Average 22.1). Apparently the

 anoxic conditions and organic carbon in the septic tank provided

 suitable conditions for denitrification to occur. A modification

 of this design has been utilized for this research.