In addition to the above microbiological species, samples were collected and analyzed for chemical constituents, such as nutrients, total organic carbon, and tritium, that may serve as indicators of possible surface contamination of the aquifer. All chemical analyses were done by USGS laboratories in Arvada, Colorado, or Ocala, Florida. Onsite analysis of specific conductance, pH, temperature, dissolved oxygen, and alkalinity were done at each PWS well according to procedures described by Wilde and Radtke (1997).

Results from the first round of enteric virus, coliphage, and fecal indicator bacteria sample collection are shown in figure 4. Of the 109 PWS wells sampled, 86 showed no presence of microbiological contamination. However, human enteric viruses were present in 9 of the 109 PWS wells at densities near the lower level of analytical detection. Contrary to what might be expected, most of the enteric virus contamination was observed outside the areas characterized as primary and secondary karst. Coliphage were observed in samples collected from 9 of the 109 wells--1 of which was a control site located in the Mississippi Alluvial Plain. Coliphage and enteric viruses were present in two wells located near the northern boundary of the Ozark Plateau aquifer system along the Missouri River. Fecal indicator bacteria

Virus sampling equipment and public-water-supply well.

Ozark and St. Francois aquifers. In addition, four PWS wells were selected as control sites. Two of the control sites are located in western Missouri where the Ozark aquifer is confined by the Ozark confining unit, the Springfield Plateau aquifer, and the Western Interior Plains confining system, and two are located in the Mississippi Alluvial Plain of southeastern Missouri (fig. 2).

Only community wells-municipal, PWS district, subdivision, and mobile home park-constructed in the last 15 years were

Forested land use in the Ozark Plateaus.
Photo courtesy of S.R. Femmer,
U.S. Geological Survey.

Agricultural land use in the Ozark Plateaus.
Photo courtesy of B.J. Smith,
U.S. Geological Survey.

considered. This study was designed to eliminate type of well construction as a variable for microbiological contamination so that interpretation could be limited to hydrogeology and land use. Well-selection criteria were based on availability of well construction and completion records (well and casing depth, size, and type), well operation and maintenance history, current usage, and geology. Finally, an attempt was made to obtain fairly even areal and vertical coverage in the aquifer and to represent the different land uses in the Ozark Plateaus.

Sample Collection and Analysis

The first round of sampling was completed in July 1997. Samples from each PWS well were analyzed for the following microbiological species-total human enteric viruses, male-specific and somatic coliphage, and fecal indicator bacteria [fecal coliform, Escherichia coli (E. coli), and fecal streptococci]. Coliphage are viruses that attack E. coli and are being studied as possible indicators of the presence of human enteric viruses. The enteric virus and coliphage samples were collected by the USGS at the selected wells. Samples represent untreated water from the aquifer. Enteric virus and coliphage samples were analyzed by the University of New England in Biddeford, Maine, according to procedures described in the