The influence of well construction on bacterial contamination of private water wells in Pennsylvania.

Introduction

Approximately one million private water systems in Pennsylvania provide water for approximately 3.5 million rural residents (U.S. Census Bureau, 1990). Drilled wells account for over 95 percent of these rural water supplies, and an additional 20,000 private wells are drilled each year in the state.

All aspects of private-water-system management in Pennsylvania, including construction, testing, and treatment, are the voluntary responsibility of the homeowner. Not surprisingly, surveys have shown that the majority of private water systems in Pennsylvania fail to meet at least one drinking-water standard (Bickford, Lindsey, & Beaver, 1996; Francis et al., 1982; Sharpe, Mooney, & Adams, 1985; Swistock, Sharpe, & Robillard, 1993; Zimmerman, Zimmerman, & Lindsey, 2001). Coliform bacteria are a frequent problem encountered in private water supplies in the state (Bickford et al., 1996; Durlin & Schaffstall, 2001; Jones 1989; Sharpe et al., 1985).

The voluntary nature of private-water-system management makes these systems vulnerable to bacterial contamination and waterborne-disease outbreaks. A recent U.S. Geological Survey (USGS) study identified the occurrence of culturable viruses and H. pylori in noncommunity water supply wells in Pennsylvania (Lindsey, Rasberry, & Zimmerman, 2002). Craun (1986) also concluded that private water systems or small, semipublic supplies were most vulnerable to waterborne disease. More recent data from Lee, Levy, Craun, Beach, and Calderon (2002) found that 64 percent of the identified waterborne-disease outbreaks in the United States during 1999-2000 occurred in unregulated private water wells.

Bacterial contamination of groundwater wells can occur from both above and below the surface. Large-scale pollution of entire groundwater aquifers may occur from failing septic systems or animal waste in runoff from farms or urban areas. Contamination of individual wells may be caused by surface water flowing along the well casing or from a loose fitting or absent well cap that allows insects, animals, or surface water to directly enter the well. Jones (1973) provided early evidence of the importance of well construction practices in preventing bacteria entry into private wells.

The type of bedrock around the well is also an important factor in determining its susceptibility to bacterial contamination. Carbonate bedrock (limestone and dolomite), which is common in much of central and eastern Pennsylvania, is typically more fractured and connected to the surface, allowing for rapid movement of groundwater and contaminants over large distances. Noncarbonate bedrock, which in Pennsylvania is predominantly shale and sandstone, generally has slower movement of groundwater and less susceptibility to surface contaminants. Notable exceptions are shallow shale aquifers, where septic system contamination may be quite severe.

Surface contamination of wells may be prevented by installation of a grout seal around the casing and a vermin-proof or "sanitary" well cap on the top of the well casing. Unlike most states, Pennsylvania does not have statewide construction standards requiring a grout seal or sanitary well cap on private wells. In the absence of regulations, cost competitiveness dictates that most well drillers construct a "nonsanitary" well without a grout seal and with a loose-fitting well cap. Although the benefits of well construction regulations are theoretically sound, there is a lack of data on the effect of well construction on bacterial contamination. A recent USGS study compared bacterial contamination in 39 sanitary wells that had a grout seal with contamination in 39 nonsanitary wells that did not have a grout seal (Zimmerman et al., 2001). Coliform-bacteria contamination was similar for the two groups of wells, but E. coli contamination was more likely in nonsanitary wells. Most of the sanitary wells in the USGS study had loose-fitting rather than sanitary well caps. As a result, insects were frequently observed inside both groups of wells, which could explain the similar coliform-bacteria results (Zimmerman et al., 2001).

The prevalence of bacterial contamination in private water systems and its role as a potential rural health issue have been demonstrated. It is still unknown, however, how much of a role individual well construction practices play in providing a path for bacteria to enter the drinking-water supply. The objective of this study was to determine if disinfection of existing contaminated wells and retrofitting with a simple sanitary well cap could significantly reduce bacterial contamination. A second objective was to determine if the use of a grout seal and sanitary well cap on new water wells could potentially reduce or eliminate bacterial contamination.

Methods

Well-Capping Study

A recent USGS study in Pennsylvania by Zimmerman and co-authors (2001) used an exhaustive search to identify wells in southeast Pennsylvania with varying bedrock geology and well construction. Thirty-eight wells that tested positive for coliform bacteria during the USGS study were selected for the study reported here, including 17 with a grout seal and 21 without a grout seal. The wells differed in geology and construction, but they were selected to minimize other variables. Data collected during the USGS study indicated that well depths ranged from 63 to 260 feet, but 33 of the 38 wells were 120 to 225 feet deep. The pH of groundwater in these wells ranged from 5.24 to 9.01, but 30 of 38 wells had similar pH levels, of 6.7 to 7.8. There were no statistical relationships between either well depth or water pH and the occurrence or prevalence of coliform bacteria in these wells.

Field personnel visited each of the 38 homes during February 2002 to collect water samples to confirm the presence of coliform bacteria prior to disinfection and installation of a sanitary well cap. Collection of water samples from each well followed standard methods recommended by USGS (Myers & Wilde, 1997). The wells were each purged for 5 to 10 minutes before sampling. During the well purging, specific conductance and water temperature were measured every 1 to 2 minutes. The authors also noted the pumping rate by measuring the time required to fill a 5-gallon bucket. The total volume of water pumped, along with stability in the conductance and temperature readings, was used to decide that samples had been collected from the aquifer or borehole rather than from water stored in the pressure tank or home plumbing system.

Water samples from most wells were collected from the same spigot that had been previously sampled by USGS in 2000-2001. USGS personnel had visually inspected each of the home plumbing systems to confirm that sample locations were not influenced by water treatment equipment. Homeowners were interviewed at the beginning of the study to ensure that treatment equipment had not been installed since the prior USGS study. Most of the wells were sampled at outside spigots or pressure tanks if water treatment equipment was present. Because of scheduling difficulties, a few homes had to be sampled at locations different from the USGS location.

Once the well was purged, a sample was collected with rubber gloves into a sterile, Whirl-Pak[R] plastic bag. Water samples were placed in a cooler and processed within six hours of collection. Analysis for total coliform bacteria involved filtration of 100 mL of sample through a membrane filter with a hand-filtration apparatus to capture bacteria. The membrane filters were placed on petri dishes with m-Endo growth media and incubated at 35[degrees]C for 22-26 hours. Samples testing positive for coliform bacteria were then cultured for the presence of E. coli bacteria by immediate transfer of the membrane filter to a separate petri dish with NA-MUG culture media. These samples were incubated for an additional 4 hours at 35[degrees]C. Results of both total coliform and E. coli bacteria were reported in the standard units (number of bacteria colonies per 100 mL of water).

To ensure data quality, two samples from each well were filtered, a 100-mL sample and a 200-mL sample. Bacteria numbers in each well were computed as the average of the results from the 100 mL and 200 mL samples. Averages of less than one colony per 100 mL were rounded up and reported as 1 colony per 100 mL. In two cases, E. coli could be measured only on a 200-mL sample because of equipment availability problems.

During the initial sampling in February 2002, water samples that tested positive for E. coli were submitted to the Gastroenteric Disease Center (GDC) at Penn State University. Each E. coli colony was tested for five common virulence factors to determine if the E. coli strain was potentially pathogenic to humans. This testing was done to provide some insights into the occurrence of pathogenic bacteria in home water wells.

Wells that tested negative for coliform bacteria during February 2002 were resampled in May 2002 according to the procedures outlined above. Wells that did not contain coliform bacteria in February or May were removed from the study. Wells that tested positive entered the Well-Capping Study. Two professional well contractors were hired to disinfect each well with 5.25 percent chlorine solution or chlorine tablets (shock chlorination). The well contractors used the standard method for disinfecting contaminated or newly constructed wells by adding chlorine solution or tablets on the basis of well depth and diameter measurements to produce at least 100 parts per million (100 ppm) of chlorine in the well water (Makuch & Sharpe, 1988). After the shock-chlorination procedure, each well was fitted with a sanitary (vermin-proof) well cap by the professional well contractors during June-October 2002. Each well was then resampled for bacteria during early December 2002 (about 30-60 days after disinfection and capping). The wells were sampled again in October 2003, approximately one year after disinfection and capping, to determine the long-term impact of disinfection and installation of a sanitary well cap.

The objective of the shock chlorination procedure was to kill bacteria existing in the well before installation of the sanitary well cap so as to test the effectiveness of the cap in preventing entry of bacteria into the well. No attempt was made, however, to determine if the shock chlorination had immediately disinfected each well. Theoretically, a well drilled into an aquifer with widespread bacterial contamination can never be adequately disinfected for any period of time with shock chlorination. Instead, the study focused on the efficacy of the standard method of disinfection used by homeowners and well contractors. The study also made no attempt to distinguish between the individual effects of shock chlorination and installation of the sanitary well cap since all of the study wells received both treatments.

Data quality was monitored continuously during the study with frequent use of sterile blank samples (sterile buffered water from Hach Co.). A total of 31 sterile blanks were filtered and measured for coliform bacteria during the study. Eighteen of these samples involved filtration of 100 mL of sterile water, and 13 involved filtration of 200 mL of sterile water. Blanks were analyzed during each of the sampling periods; 13 during February 2002, six during May 2002, eight during November-December 2002, and four during October 2003. None of the blank samples tested positive for total coliform bacteria.

Quality assurance was further measured through the duplicate analysis of each sample (100-mL and 200-mL filtration). This provided a total of 120 duplicate samples. Of these, 71 produced negative results for both 100-mL and 200-mL volumes. A paired t-test on the 49 samples with positive bacteria results produced no statistical difference between results from filtering 100 mL and 200 mL. A similar paired t-test on duplicate samples for E. coli also produced no statistical difference, but the sample size (nine wells) was much smaller.

Although there were no statistical differences between results from 100-mL and 200-mL samples, there were some indications that 100 mL of sample was not always sufficient to detect small numbers of coliform bacteria. There were 10 cases in which coliform bacteria were detected in the 200-mL sample but not in the 100-mL sample. In each of these cases, the bacteria numbers were small (<5 colonies) in the 200-mL sample.

New-Well Study

New wells constructed with a grout seal and sanitary well cap are not required in Pennsylvania except under a few local ordinances. The authors were able to locate 24 newly constructed residential wells with grout seals and sanitary well caps through a search of well-drilling records provided by the Pennsylvania Geological Survey. The 24 wells were located in five counties in southeast Pennsylvania. These wells had generally been drilled within the last two years, although a few had been drilled up to 10 years ago. Eighteen of them were located in noncarbonate bedrock, and six were located in carbonate bedrock. Well depths ranged from 140 to 850 feet, although 20 of the 24 wells were 140 to 340 feet deep.

Homeowners with new wells were contacted by phone during October 2002 so that the authors could obtain permission to sample their wells and determine proper sample collection locations that would avoid water treatment equipment. Water samples were collected at outside taps where water treatment equipment was absent or could be bypassed. Samples were collected at the pressure tank from the few homes where water treatment equipment existed and could not be bypassed. Attempts were made during collection of samples to visually confirm homeowner reports about water treatment equipment. Samples were collected from each home during November 9-December 4, 2002, and were analyzed for total coliform and E. coli according to the protocols described above.

Results

Well-Capping Study--Initial Testing

The sampling of 38 wells in February-May 2002 was intended to confirm the presence of coliform bacteria in these wells before they were disinfected and fitted with sanitary well caps. All of these wells had previously tested positive for coliform bacteria during 2000-2001 (Zimmerman et al., 2001). Surprisingly, only 17 (45 percent) tested positive for total coliform bacteria. All but two wells contained very low numbers of bacteria, less than 9 colonies per 100 mL. By comparison, during the 2001 USGS study, nearly half of the wells had more than 10 colonies per 100 mL. The disappearance of coliform bacteria from 21 of the 38 wells between the 2000-2001 USGS sampling and the sampling in early 2002 was unexpected and provided a smaller set of wells for evaluation of the effectiveness of disinfection and sanitary-well-cap installation.

Coliform bacteria were slightly more common in wells with a grout seal (53 percent) than in wells without a grout seal (38 percent), and they were more common in wells located in carbonate bedrock (55 percent) than in noncarbonate bedrock (33 percent) (Table 1). Four of the 38 wells (11 percent) contained E. coli bacteria during the February-May sampling. This contamination rate is similar to rates reported by other Pennsylvania studies (Bickford et al., 1996; Durlin & Schaffstall, 2001; Zimmerman et al., 2001). Since E. coli bacteria can only originate from human or animal wastes, these wells represent more serious contamination problems. All four were found in carbonate bedrock, and three of the four did not have a grout seal. Analysis of six E. coli colonies from the four wells by the Penn State Gastroenteric Disease Center identified one colony in one well that was potentially pathogenic. The serotyping analysis tested for five common virulence factors related to E. coli. These tests targeted the most common pathogenic strains but did not completely eliminate the possibility of other virulent E. coli.

Effects of Disinfection and Sanitary-Well-Cap Installation

Sixteen of the 17 wells that tested positive for coliform bacteria were disinfected and fitted with a sanitary well cap during June-October 2002 (one homeowner chose to be removed from the study). All 16 wells were resampled in November-December 2002 to determine the short-term impact of disinfection and sanitary-well-cap installation. Coliform bacteria were still present in nine of the 16 wells (56 percent) within this 30-60 day period. Wells in carbonate bedrock were slightly more likely to still contain bacteria than wells in noncarbonate bedrock (Table 2). This finding may have been because wells in carbonate bedrock generally had higher bacteria concentrations before capping and were more likely to contain E. coli bacteria. The capping procedure was also more effective on wells that had small numbers of bacteria. Of the 10 wells that had [less than or equal to]2 coliform bacteria per 100 mL before capping, only three still contained coliform bacteria after capping. All six of the wells that had [greater than or equal to]3 coliform bacteria per 100 mL before capping still had bacteria after disinfection and capping. The wells in which capping was successful had higher coliform-bacteria concentrations (up to 13 colonies per 100 mL) during the prior USGS study of 2000-2001, suggesting that capping can also be successful on wells in which bacteria levels are intermittently higher. None of the wells tested positive for E. coli in Fall 2002 after capping (Table 2).

Of the 16 capped wells, seven did not have a pre-existing grout seal. Five of these seven wells (71 percent) were re-contaminated quickly after capping (Table 2). The two wells without a grout seal that were successfully treated by capping had small numbers of bacteria before capping. Only four of the nine wells with a pre-existing grout seal (44 percent) still contained coliform bacteria within 30-60 days of capping, and three of these four were located in carbonate bedrock.

Results One Year After Capping

Thirteen of the 16 wells were sampled again approximately one year after they were disinfected and fitted with a sanitary cap. Of the three wells that could not be resampled, two were decommissioned by the homeowners because they had tested positive for bacteria after disinfection and capping. The third homeowner simply chose not to participate further in the study.

While disinfection and installation of a sanitary well cap had some initial success in eliminating bacteria, the results after one year showed less long-term benefit. Of the 13 wells, 11 were contaminated with coliform bacteria, and E. coli had returned to three wells (Table 2). The differences in bedrock and grout presence observed earlier were not as apparent after one year. The two wells in which bacteria were still absent represented both types of bedrock and a well with and without a grout seal. However, E. coli bacteria were again only found in wells in carbonate bedrock.

Bacteria numbers, in general, were higher during the final sampling of these wells (Figure 1). The overall temporal variability in bacteria numbers during the study may be explained by weather conditions. Weather during the 2000-2001 sampling done by USGS included near-normal precipitation. The lower-than-expected detection of bacteria during the February 2002 sampling period occurred in conjunction with a serious drought in the region. The counties where the study wells were located were all under a state Drought Emergency during February 2002, primarily because of extremely low groundwater levels. Since coliform bacteria are generally surface organisms, they are more likely to occur in groundwater during periods of rapid groundwater recharge by surface water. During periods of drought, the lack of groundwater recharge could reduce bacterial contamination of wells by removing the source of the bacteria (i.e., direct surface-water contamination) or by reducing the movement of underground contamination sources (i.e., septic-system effluent, etc.). The lack of precipitation during the fall and winter may have prevented groundwater recharge to these wells for several months. In 2003, precipitation in the study area was much above average. Bacteria levels in most of the wells rose dramatically in October 2003 in response to these conditions (Figure 1). Although it has long been assumed that bacteria levels vary seasonally with weather conditions, these results provide evidence of the relationship between weather conditions and the occurrence of coliform bacteria in groundwater wells. A few studies have documented similar results on small numbers of private water wells (Jones, 1973; Oliphant, Ryan, Chu, & Lambert, 2002).

Bacterial Contamination of New Wells with Sanitary Construction

Sampling of 24 new wells constructed with a grout seal and sanitary well cap was completed during early December 2002. Coliform bacteria were found in 29 percent of new wells, and 17 percent contained E. coli (Table 1). Contamination rates for both total coliform and E. coli bacteria were higher for new wells constructed in carbonate bedrock, although the sample size was small (n = 6). These results, along with the results from the capping study, suggest that well construction practices may be less successful in preventing bacterial contamination in areas with carbonate bedrock. New wells with a grout seal and sanitary well cap that were located in noncarbonate bedrock had a lower incidence of bacterial contamination (Table 1). Of the four new wells in noncarbonate bedrock that were contaminated with bacteria, two were neighboring wells within 200 yards that were grossly contaminated with both coliform and E. coli bacteria. These two wells were apparently affected by aquifer contamination from a local agricultural or septic-system source. Well depth may also have played a role in bacterial contamination. Coliform bacteria were found in wells up to 340 feet deep, but the five wells greater than 400 feet deep did not contain coliform bacteria.

[FIGURE 1 OMITTED]

Discussion and Conclusions

Disinfection and installation of a sanitary well cap on private water wells containing coliform bacteria was effective in removing coliform bacteria from some wells for a few months, especially on wells that originally contained low numbers of coliform bacteria and no E. coli. After one year, however, only a few of the wells were still free of total coliform bacteria. In wells in carbonate bedrock, bacteria were detected after less elapsed time, but within one year, contamination rates showed no differences related to bedrock type or the presence of a preexisting grout seal.

Bacterial concentrations in the wells varied considerably during the study, perhaps in response to varying climatic conditions. Coliform-bacteria levels were very low during an extreme drought in 2002 but much higher after above-normal precipitation in 2003. The effects of disinfection and installation of a sanitary well cap reported here could represent a conservative assessment since all of the study wells contained bacteria during an extreme drought. In this respect, the authors' pool of study wells may have been skewed toward wells that were affected by more serious aquifer contamination. In any case, the apparent climate-related differences in bacterial contamination indicated that sanitary surveys of wells might best be conducted subsequent to active periods of groundwater recharge.

Wells in carbonate bedrock areas were more likely to contain coliform bacteria and E. coli bacteria than wells in noncarbonate bedrock regardless of well construction, a result suggesting that aquifer contamination was more prevalent in carbonate aquifers. New and existing wells constructed in noncarbonate bedrock had lower bacterial contamination rates.

Overall, aquifer contamination from areas other than the wellhead appeared to be an important source of bacterial contamination in the study wells. This source was evidenced by the fact that both new wells with sanitary construction and retrofitted wells with sanitary construction (pre-existing grout seal and new sanitary well cap) still had bacterial contamination rates of 29 to 71 percent. Contamination in wells influenced by aquifer contamination cannot be fully addressed with improved well construction practices alone, and greater attention to curtailment of polluting activities appears warranted.

Although the wells included in this study were carefully selected to represent both existing and new wells in the two major types of bedrock found in Pennsylvania, the small sample size justifies caution in interpreting the results. Intensive research studies on private water wells typically choose to work with small numbers of carefully selected wells because of the logistical problems inherent in working with homeowners and the travel costs associated with spatially representative samples. The pool of potential wells for the study reported here was further restricted because new wells with sanitary construction are rare in Pennsylvania. Further research using emerging bacteria-source-tracking methodologies may shed more light on the sources of bacterial contamination in groundwater wells.

Acknowledgements: The authors wish to thank Chad Voorhees, Angela Schreffler, and Bill Wert for their help in the field collection of data. They also thank Tammy Zimmerman, Michele Zimmerman, and Bruce Lindsey of the USGS Water Resources Division, New Cumberland, Pennsylvania, for their assistance in all aspects of the project. Eichelbergers, Inc., and Raab Well Drilling were helpful in locating study wells and installation of well caps. The project was made possible, in part, by a grant from the Center for Rural Pennsylvania, a legislative agency of the Pennsylvania General Assembly.

Corresponding Author: Bryan R. Swistock, Extension Associate, School of Forest Resources, Pennsylvania State University, 133 Land and Water Building, University Park, PA 16802. E-mail: [e-mail].

REFERENCES

Bickford, T.M., Lindsey, B.D., & Beaver, M.R. (1996). Bacteriological quality of ground water used for household supply, Lower Susquehanna River Basin, Pennsylvania and Maryland (USGS Water-Resources Investigations Report No. 96-4212). Lemoyne, PA: U.S. Geological Survey.

Craun, G.F. (1986). Waterborne diseases in the U.S. Boca Raton, FL: CRC Press.

Durlin, R.R., & Schaffstall, W.P. (2001). Water resources data for Pennsylvania, water year 2000, Delaware River Basin (Vol. 1). (USGS Water-Data Report No. PA-00-1). New Cumberland, PA: (U.S. Geological Survey).

Francis, J.D., Brower, B.L., Graham, W.F., Larson, O.W., III, McCaull, J.L., & Vigorita, H.M. (1982). National statistical assessment of rural water conditions (EPA 570/9-84-004). Washington, DC: The Office of Drinking Water, U.S. Environmental Protection Agency.

Jones, E.E., Jr. (1973). Well construction helps determine water quality. Journal of Environmental Health, 35(5), 443-450.

Jones, M.C. (1989). Occurrence of nitrates and pollution indicator organisms in private water supply wells of rural Berks County, Pennsylvania. Master's thesis, Pennsylvania State University, University Park.

Lee, S.H., Levy, D.A., Craun, G.F., Beach, M.J., & Calderon, R.L. (2002). Surveillance for waterborne disease outbreaks--United States, 1999-2000. Morbidity and Mortality Weekly Report, 51(SS-8), 52.

Lindsey, B.D., Rasberry, J.S., & Zimmerman, T.M. (2002). Microbiological quality of water from noncommunity supply wells in carbonate and crystalline aquifers of Pennsylvania (USGS Water Resources Investigations Report No. 01-4268). New Cumberland, PA: U.S. Geological Survey.

Makuch, J., & Sharpe, W.E. (1988). Safeguarding wells and springs from bacterial contamination (Penn State Cooperative Extension, Extension Circular 345). University Park, PA: Penn State College of Agricultural Sciences Cooperative Extension.

Myers, D.N., & Wilde, F.D., eds. (1997). Biological indicators (USGS Techniques of Water-Resources Investigations, Book 9, Handbooks for Water Resources Investigations, Chapter A7, Report No. 09-A7.1). Denver, CO: U.S. Geological Survey.

Oliphant, J.A., Ryan, M.C., Chu, A., & Lambert, T.W. (2002). Efficacy of annual bacteria monitoring and shock chlorination in wells finished in a floodplain aquifer. Ground Water Monitoring and Remediation, 22(4), 66-72.

Sharpe, W.E., Mooney, D.W., & Adams, R.S. (1985). An analysis of groundwater quality data obtained from private individual water systems in Pennsylvania. Northeastern Environmental Science, 4(3-4), 155-159.

Swistock, B.R., Sharpe, W.E., & Robillard, P.D. (1993). A survey of lead, nitrate and radon contamination of private individual water systems in Pennsylvania. Journal of Environmental Health, 55(5), 6-12.

U.S. Census Bureau. (1990). 1990 Census of Population and Housing, Pennsylvania Housing Characteristics. http://factfinder.census.gov/servlet/QTTable?_bm=n&_lang=en&qr_name=DEC_1990_STF3_DP5&ds_name=DEC_1990_STF3_&geo_id=04000US42. (4 Apr. 2004).

Zimmerman, T.M., Zimmerman, M.L., & Lindsey, B.D. (2001). Relation between well-construction practices and occurrence of bacteria in private household-supply wells, south-central and southeastern Pennsylvania (USGS Water-Resources Investigations Report No. 01-4206). New Cumberland, PA: U.S. Geological Survey.

Bryan R. Swistock, M.S.

William E. Sharpe, M.S., Ph.D.

TABLE 1 Results of Bacteria Testing for 38 Private Wells in the
Well-Capping Study and 24 New Wells Constructed with a Sanitary Well Cap
and Grout Seal*

                                                          Number Testing
                                                          Positive for
                                                Number    Coliform
                    Bedrock/Well Construction   of Wells  Bacteria

Well-capping study
                    Carbonate/no grout seal     13         6
                    Carbonate/grout seal         7         5
                    Noncarbonate/no grout seal   8         2
                    Noncarbonate/grout seal     10         4
                    Totals                      38        17
New-well study
                    Carbonate with grout seal    6         3
                      and sanitary well cap
                    Noncarbonate with grout     18         4
                      seal and sanitary well
                      cap
                    Totals                      24         7

                                                Number Testing
                    Bedrock/Well Construction   Positive for E. coli

Well-capping study
                    Carbonate/no grout seal     3
                    Carbonate/grout seal        1
                    Noncarbonate/no grout seal  0
                    Noncarbonate/grout seal     0
                    Totals                      4
New-well study
                    Carbonate with grout seal   2
                      and sanitary well cap
                    Noncarbonate with grout     2
                      seal and sanitary well
                      cap
                    Totals                      4

*Results are from initial bacteria testing conducted February-May 2002
on the private wells and from bacteria testing conducted December 2002
on the new wells.

TABLE 2 Occurrence of Coliform Bacteria and E. coli Bacteria in Wells
Approximately 30-60 days (Fall 2002) and One Year (Fall 2003) After
Disinfection and Installation of a Sanitary Well Cap

Bedrock/
Well Construction             Fall 2002 Sampling Period
                            Total  Positive Total  Postive
                            Wells  Coliform        E. coli

Carbonate/no grout seal      5        3              0
Carbonate/grout seal         5        3              0
Noncarbonate/no grout seal   2        2              0
Noncarbonate/grout seal      4        1              0
Totals                      16        9              0

Bedrock/
Well Construction               Fall 2003 Sampling Period
                                         Positive Total  Positive
                            Total Wells  Coliform        E. coli

Carbonate/no grout seal        4           4                2
Carbonate/grout seal           3           2                1
Noncarbonate/no grout seal     2           2                0
Noncarbonate/grout seal        4           3                0
Totals                        13*         11                3

*Two homeowners drilled new wells, and one chose not to participate in
the study after the Fall 2002 sample collection.