A survey of lead, nitrate and radon contamination of private individual water systems in Pennsylvania.

Private individual water systems throughout Pennsylvania were sampled for dissolved lead, nitrate-N and radon to determine the prevalence of these primary pollutants. Approximately 1,600 sources were tested for lead and nitrate and 989 were tested for radon. Twenty-eight percent of sampled homes had lead concentrations above 10 |mu~g/L and 19 percent were above 15 |mu~g/L. These percentages increased to 60 and 34 percent respectively when calculated total (digested) lead data were used, suggesting that total lead analysis may be appropriate even when dealing with relatively clear, low turbidity samples. Nitrate contamination was less prevalent and more regional than lead. Nine percent sampled homes contained nitrate-N above 10 mg/L with nearly all (96 percent) of these homes located in the agricultural southcentral and southeastern regions of the state. Nearly 80 percent of the groundwater wells tested contained radon concentrations above the proposed MCL of 300 pCi/L. Excessive radon concentrations existed in all regions of the state but were most prevalent in the eastern regions near the Reading Prong geologic formation.In Pennsylvania there are approximately one million private individual water supplies which provide water for about two million of the state's rural residents. Unlike municipal water supplies which must meet federal Safe Drinking Water Standards, management of private water supplies is voluntary. Past studies have shown that numerous health related (primary) and aesthetic (secondary) pollutants can occur in these unregulated water supplies (1, 2). Of the pollutants encountered in private water supplies, lead, nitrate and radon may present the greatest risk based on their estimated prevalence and health effects.Although lead is often thought of as an urban problem, some lead can probably be found in the paint, soil, dust, housewares or drinking water of many rural homes. Drinking water contributions to blood lead burdens are typically less than 20 percent, but have been reported as high as 40 percent depending on the amount contributed by other sources (3, 4).Recent guidance from the Centers for Disease Control (CDC) suggests that blood lead concentrations over 10 |mu~g/dL may be indicative of lead poisoning (5). Research has shown that blood lead concentrations increase by about 1 |mu~g/dL for every 5 |mu~g/L of lead in drinking water (6). At a conversion rate of 5:1, 50 |mu~g/L of lead in water would produce an unacceptable 10 |mu~g/dL of lead in blood. In order to allow for lead contributions from other sources (e.g., food, dust, etc.) the maximum allowable concentration of lead in drinking water should be no more than 10 |mu~g/L.The overwhelming source of lead in both private and municipal water supplies is corrosion of plumbing materials. Households typically contain various mixtures of metal pipes, faucets and solders which each can contain lead or lead impurities. Research has shown that lead based solders, which are present in 70 percent of American homes, and brass fixtures are the most significant sources of lead in tap water (7, 8, 9). Brass fixtures were found to be especially important contributors in first flush water samples from homes using plastic pipes (8). Corrosion of lead from metal plumbing increases as the pH, hardness and alkalinity of the source water decrease (10). Sharpe et al. (1) determined that corrosive drinking water was the most common problem in private water supplies of Pennsylvania, occurring in 60 percent of the homes tested.The prevalence of leaded plumbing components and corrosive groundwater in Pennsylvania argue that lead contamination is common in private water systems; however, few studies have actually documented its occurrence. Francis et al. (2) reported 16 percent of rural water supplies in the U.S. contained lead concentrations above 50 |mu~g/L and approximately 50 percent were above 10 |mu~g/L. A lab contamination problem, however, made these results suspect. Based on the limited available data, the U.S. Public Health Service estimates that 20 percent of the U.S. population has tap water lead concentration in excess of 20 |mu~g/L (11).Amendments were recently added to the National Primary Drinking Water Regulations to control lead in municipal water supplies through a complex scheme of regulations rather than the simpler, more direct maximum contamination level (MCL) approach (12). The U.S. Environmental Protection Agency (EPA) opted for an "action level" of 15 |mu~g/L which will require municipal water utilities to add corrosion control chemicals at the water treatment plant and take other actions if more than 10 percent of the tested homes are above 15 |mu~g/L. This action level replaces an MCL for lead of 50 |mu~g/L. It's estimated that these regulations will reduce the lead exposure of 130 million Americans (13), but they provide little protection for private water systems. Recent legislation in Pennsylvania and other states has extended lead regulations to private water systems by banning the sale of leaded solder, leaded pipe and limiting lead impurities in metal fixtures (14). Additional federal legislation has been introduced that has proposed a lead MCL of 10 |mu~g/L to provide a more appropriate level of safety from lead exposure (15).Unlike lead, nitrates usually occur in the source water as a result of pollution from animal waste, human sewage, fertilizers and natural organic matter. Increases in the use of agricultural fertilizers have been implicated in steady increases in groundwater nitrates over the past 50 years (16). In Pennsylvania, nitrogen fertilizer applications increased by 150 percent from 1960 to 1982 (17). Because nitrogen recovery in crops is typically only 50 percent, much of the applied nitrogen ends up in groundwater as nitrate (18).Health concerns about nitrate in water are usually related to young infants and methemoglobinemia (19). Two thousand cases of methemoglobinemia were reported internationally from 1945 to 1970, with an average fatality rate of eight percent (19). Other less proven effects from nitrates include hypertension, cardiovascular disease, impaired growth and central nervous system damage (20). The drinking water standard for nitrate of 10 mg/L as nitrate-N (45 mg/L as nitrate) was established primarily to protect young children from methemoglobinemia.The predominance of agriculture makes some areas of Pennsylvania susceptible to nitrate contamination, however, surveys of nitrates in private water supplies have reported diverse results. Sharpe et al. (1) found six to 14 percent of private water sources in Pennsylvania contained high nitrates. Francis et al. (2) reported only 0.3 percent of eastern U.S. private water supplies had nitrate-N above 10 mg/L. Some researchers in midwestern agricultural states have reported nitrate-N contamination rates (|is greater than~ 10 mg/L) as high as 70 percent (21). A U.S. Geological Survey study of Pennsylvania groundwater found 30 percent had nitrate-N levels indicative of human activity (|is greater than~ 3 mg/L) (22).Radon is perhaps the least understood and potentially most dangerous contaminant in private individual water supplies. The calculated occurrence of fatal cancers due to radon exposure from water may be larger than the sum of all other carcinogens known to exist in water supplies (23). It's estimated that 137 people die every year in the U.S. from exposure to radon in water (24).Radon becomes dissolved in groundwater and, to a much lesser extent, in surface water through radioactive decay of radium in rocks and soil (25). Once radon is dissolved in water, exposure to radon can occur through ingestion or, more importantly, through inhalation from showering and other activities which agitate the water. Research has shown that 10,000 pCi/L of radon in water will produce about 1 pCi/L of radon in indoor air (26). Drinking water, however, generally only represents about one percent of the total exposure from all sources of environmental radon (23). Based on the proposed radon air standard of 1 pCi/ L and to allow for other more important sources of radon, the EPA has proposed a radon in water standard of 300 pCi/L (27). It's estimated that even consumption of water with 300 pCi/L of radon involves a cancer risk of 1 in 10,000.Although there is an extensive data base of radon in municipal water supplies, little data on private water systems is available. Hess et al. (28) reported a mean radon concentration of 910 pCi/L in 16 private groundwater wells and 380 pCi/L in 105 public groundwater sources in Pennsylvania. In a nationwide study of community groundwater wells, Pennsylvania was among the top eight states for radon occurrence with a mean concentration of 1,570 pCi/L (29). Dixon and Lee (29) also reported higher radon concentrations in eastern Pennsylvania wells near the Reading Prong geologic formation and lower radon values in the sandstones, shales and carbonates of the Appalachian Plateau in western Pennsylvania. A review of the limited available data suggests that the radon in water risk is much higher in low-yield private groundwater systems (30).The objectives of this study were to provide an extensive and representative survey of the prevalence of lead, nitrate and radon contamination in private water supplies in Pennsylvania and to relate contamination from lead or nitrate with various water supply and land use characteristics.Lead and nitrate-N samplingWater samples were collected from 1,595 private water systems in 35 of the 67 counties of Pennsylvania from March 1989 to April 1992. Water sampling was conducted as part of county-sponsored Safe Drinking Water Clinics which were advertised to the general public. Clinic attendees interested in lead and nitrate-N testing were given clean, 125 ml plastic sample bottles, sample collection instructions and a survey form. Participants were instructed to collect first-draw water samples from the following morning and return the samples and completed surveys to a pre-designated drop-off point. A survey question was used to verify that samples were collected properly. Samples not collected from first flush were excluded from lead analyses. First flush rather than running water samples were collected because lead concentrations from first flush samples are more closely correlated with blood lead concentrations (31, 32, 33, 34).Samples were kept cool and returned to the Water Laboratory at the Environmental Resources Research Institute at Penn State University for lead and nitrate analysis within 72 hours of collection. Samples were filtered using 0.45 |mu~m filters and acidified prior to analysis. Nitrate analysis was performed according to the EPA Cadmium Reduction methodology 353.2 (35) on a Technicon Autoanalyzer II. Nitrate results are reported in mg/L of nitrate-N with a detection limit of 0.005 mg/L. Total dissolved lead was analyzed on a Perkin Elmer Atomic Adsorption Model 703 using the EPA AA Furnace method 239.2 (35). Lead results are reported in |mu~g/L of total dissolved lead with a detection limit of 1 |mu~g/L.Samples were analyzed for total dissolved lead rather than total lead to control costs. Because virtually all of the samples contained low turbidities, it was anticipated that little or no suspended lead would be present. To confirm this, a subset of 126 samples were aliquoted and analyzed for total and dissolved lead. Total lead samples were digested using Ultrex II Ultrapure Nitric Acid and filtered through 0.45 |mu~m filters.Data quality assurance and quality control were measuring using blind split samples and deionized water blanks. Twenty split samples and 42 blanks were randomly arranged among the actual samples throughout the duration of the study. Blanks had a median lead concentration of 1 |mu~g/L (range |is less than~1 to 15 |mu~g/L) and a median nitrate-N concentration of 0.015 mg/L (range |is less than~0.005 to 0.200 mg/L). The median difference between split samples was 2 |mu~g/L for dissolved lead (range 0 to 22 |mu~g/L) and 0.05 mg/L for nitrate-N (range 0.002 to 1.25 mg/L). Four deionized water blanks were also digested and analyzed for total lead. The median result was 2 |mu~g/L (range |is less than~1 to 4 |mu~g/L) implying that no significant lead contamination resulted from the digestion process.Survey forms were used to provide basic information on the water samples such as water supply type, depth of well, age of system and any previous water testing. Additional questions concerning symptoms of lead contamination, proximity to agriculture, and family health were also used for correlation with lead and nitrate-N results. Eighty-nine percent of the distributed surveys were at least partially completed and returned. Total response rates for given questions ranged from 65 to 88 percent of the samples tested.Data were analyzed on a statewide and regional basis. Pennsylvania was divided into six regions along county lines corresponding to the major physiographic provinces of Pennsylvania. Regional sample sizes roughly correspond to the land area occupied by each region. The northcentral region included 81 samples collected from Cattaraugus and Allegany counties in extreme southern New York state bordering Pennsylvania. Samples sizes from the 37 counties varied from one to 180 samples with 26 counties having more than 20 samples.Radon samplingGroundwater radon data were provided by the Pennsylvania Department of Environmental Resources (PADER), Bureau of Radiation Protection. Radon samples were collected from 989 private individual groundwater wells by PADER personnel from 1985 to 1989. Samples were collected from 48 of the 67 Pennsylvania counties although only 12 counties had more than 20 samples. The most sampled counties were located primarily in the southcentral and southeastern regions of the state where radon in air is most prevalent. Samples were analyzed using the liquid scintillation counting method (36) at the PADER Bureau of Laboratories, Harrisburg, Pennsylvania. All samples were analyzed twice with a median difference of 43 pCi/L (range 0 to 5,542 pCi/L).Data analysesStatistical analyses were conducted using the SAS statistical package. Data normality was tested using the Kolomogorov normality test (37). The Tukey multiple means comparison method (38) was used to statistically compare lead, nitrate-N and radon data among geographic regions and to correlate survey responses to lead and nitrate-N results. Pearson correlation coefficients (r) were used for correlation of nitrate-N results to land use data. All statistical tests were conducted at the 95 percent confidence level.ResultsThe majority of samples from the lead and nitrate-N survey originated from drilled wells (86%) with smaller percentages from springs (9%), hand-dug wells (4%) and cisterns (0.3%). Thus, 99.7 percent of the homes tested utilized groundwater for their water source. All of the PADER radon samples originated from groundwater wells. These results agree with 1980 census figures for rural Pennsylvania water supplies (39). Virtually all (88%) of the sources were constructed between 1950 and 1990 with similar numbers from each decade. Wells were generally shallow (median = 120 ft.) but ranged from |is less than~ 10 feet to 800 feet deep. Approximately half of the homes (41%) had never had any water quality testing done prior to this survey.LeadSurvey respondents indicated that 104 samples were not first flush samples; consequently, only 1,351 samples were included in the statistical analyses presented for lead. The distribution of lead concentrations for these samples is shown in Figure 2. Eighty percent of the samples had detectable amounts of lead (|is greater than~ 1 |mu~g/L). Approximately 28 percent of the samples were above the proposed MCL (10 |mu~g/L), 19 percent were above the present action level (15 |mu~g/L), and seven percent were above the previous MCL (50 |mu~g/L). The highest concentration sampled was 2,800 |mu~g/L with 26 samples above 100 |mu~g/L and nine samples above 200 mug/L.Lead concentrations were similar among the six regions of Pennsylvania. Median concentrations in each region ranged from 4 |mu~g/L to 10 |mu~g/L. Lead concentrations were significantly higher in southcentral and southeastern counties despite previous studies which have reported a lower incidence of corrosive groundwater in these regions (1). The percent of samples above 10 |mu~g/L in each region ranged from 16 to 49 percent while four to 38 percent were above 15 |mu~g/L. These percentages concur with regional median comparisons indicating similar percentages among regions with slightly higher values in the southcentral and southeastern regions.Significant differences in lead concentrations were also observed between the various types of water supplies. Springs and hand-dug wells had the highest median lead levels, likely the result of the more acidic, corrosive nature of these generally shallow water supplies (1). No correlation was observed between drilled well depth or drilled well age and dissolved lead concentrations.Corrosive water supplies may produce bluish stains and a metallic taste as a result of dissolution of copper pipes in the plumbing system. Since the copper pipes are often joined together with lead solder, these symptoms may be good indicators of lead contamination. In this study, homeowners noticing bluish stains had significantly higher lead concentrations (12.5 |mu~g/L) compared to those without this symptom (4 |mu~g/L). Homes with a metallic tasting water also had significantly higher lead concentrations (8 |mu~g/L vs. 5 |mu~g/L). Attempts were also made to correlate high lead concentrations to the occurrence of high blood pressure but no significant correlation was observed.Dissolved vs. total leadThe relationship of digested (total) lead concentrations to dissolved lead values for 126 samples is shown in Figure 5. The resulting regression equation predicted a difference between dissolved and total lead of six to 18 |mu~g/L within the range of measurements (0 to 100 |mu~g/L). Although small, these differences are significant when compared to the proposed lead MCL of 10 |mu~g/L. Use of dissolved rather than total lead could cause a significant underestimation of homes at risk from chronic exposure to low concentrations of lead. For example, 28 percent of the tested homes had dissolved lead concentrations above 10 |mu~g/L and 19 percent were above 15 |mu~g/L; however, these values increase to 60 and 34 percent, respectively, when calculated total lead concentrations are used. These data suggest that measurement of total lead is preferable to dissolved lead even for generally clear, low turbidity groundwater samples.NitratesOf the 1,583 water samples analyzed for nitrate-N, 9.4 percent were above the MCL of 10 mg/L and 30 percent were above 3 mg/L. The 30 percent above 3 mg/L agrees remarkably well with the earlier USGS study of Pennsylvania groundwater wells (22). Although less prevalent than lead, nitrates were much more regionally influenced. Median nitrate-N concentrations were |is less than~ 1 mg/L in every region of the state except the southcentral and southeast where they were significantly higher at 2.18 and 11 mg/L, respectively. Regional comparisons of contamination percentages illustrate that water supplies above the nitrate-N MCL were nearly nonexistent in all regions except the southcentral and southeastern. Ninety-six percent of the homes found to have excessive nitrates were in the southcentral (25%) and southeast (71%) regions. Not surprisingly, these regions also have the highest percentage of intensive agricultural land use in Pennsylvania.Nitrate concentrations were not strongly influenced by water supply type although hand-dug wells did contain significantly higher concentrations than other types. Wells with nitrate-N above 10 mg/L were significantly shallower (100 vs. 125 ft.) and older (23 vs. 19 years) than those with nitrate-N below 10 mg/L.Observed nitrate-N concentrations were correlated with several agricultural land use variables. Homeowners were asked to group their water supply into one of four distance categories to the nearest cornfield. The distance to the nearest cornfield was used as an indicator of agricultural impacts on groundwater nitrates since corn crops typically receive large amounts of nitrogen fertilizer. Median nitrate-N levels significantly increased as the distance to the nearest cornfield decreased. Median nitrate-N concentrations for each county were also positively correlated with percent agricultural land use (r=0.57, p|is less than~0.0014) and pounds per acre of nitrogen fertilizer used (r=0.82, p|is less than~0.0001).RadonThe distribution of radon concentrations from the PADER survey is shown in Figure 8. Approximately 78 percent of the samples were above the proposed MCL (300 pCi/L), 53 percent were above 1,000 pCi/L, and six percent were above 10,000 pCi/L. The highest concentration sampled was 141,270 pCi/L with 13 samples above 50,000 pCi/L.Although regional comparisons are limited because of the skewed geographic distribution, the radon data are useful nonetheless. The median radon concentration for all of the samples was 1,100 pCi/L with regional medians ranging from 187 to 2,555 pCi/L. Significantly higher radon concentrations generally occurred in the eastern regions of the state which agrees with the earlier study of Dixon and Lee (29). The percent of radon samples above the proposed MCL of 300 pCi/L in each region ranged from 52 to 88 percent for the four regions with sufficient data. The highest percentages occurred in the southeastern and southcentral regions which include the Reading Prong and other radon producing geologic formations.It should be noted that the skewed nature of the radon data set likely overestimates the prevalence and magnitude of the problem on a statewide basis. Recent, more representative data collected by the PADER suggest that the groundwater radon problem is present in all regions of Pennsylvania but is most severe in the southeastern portion of the state.Summary and conclusionsBased on their prevalence and health effects, lead and radon appear to be serious health threats to private individual water system users in Pennsylvania. Twenty-eight percent of the tested private water supplies contained dissolved lead concentrations above acceptable levels and approximately 78 percent contained radon concentrations above the proposed MCL. This translates into approximately 600,000 rural Pennsylvanians exposed to excessive lead and over one million exposed to dangerous radon concentrations in their drinking water. Lead concentrations were similar among the six regions of Pennsylvania with the highest values occurring in the southeastern region of the state. Radon concentrations tended to be higher in the eastern portion of the state near the Reading Prong geologic formation. The occurrence of lead contamination from corrosion of plumbing components should decline in the future as new lead-free plumbing replaces leaded components and as water systems with leaded components age. Lead in drinking water, however, still represents a very clear health threat to rural Pennsylvanians which needs to be addressed in rural health programs. Because radon is a naturally occurring problem, it will always be a problem in groundwater supplies from certain geologic formations.Digestion and analysis of total lead on 126 samples increased dissolved lead values by six to 18 |mu~g/L These differences are large in comparison to the proposed MCL of 10 |mu~g/L, strongly suggesting that total lead should be analyzed even for relatively clear, low turbidity groundwater sources.Nitrate-N was a less common problem than lead or radon with only nine percent of samples above the existing MCL. However, thousands of rural Pennsylvania children under six months of age may be exposed to dangerous levels of nitrate in their drinking water. Virtually all of the samples with high nitrate-N originated from the agricultural southcentral and southeastern regions of the state. It is anticipated that the incidence of nitrate contamination of groundwater in Pennsylvania will decline in the future in response to reduced use of nitrogen fertilizers and better animal waste management on farms.References1. Sharpe, W.E., D.W. Mooney and R.S. Adams (1985), An analysis of groundwater quality data obtained from private individual water systems in Pennsylvania, Northeast. Environ. Sci. 4(3/4):155-159.2. Francis, J.D., B.L. Brower, W.F. Graham, O.W. Larson III, J.L. McCaull and H.M. Vigorita (1983), National statistical assessment of rural water conditions, Dept. of Rural Sociology, Cornell University, Ithaca, NY, 1900 pp.3. U.S. Environmental Protection Agency (1986), Air Quality Criteria for Lead, NTIS PB87-142410.4. Ryu, J.E., E.E. Ziegler, S.E. Nelson and S.J. Fomon (1983), Dietary intake of lead and blood lead concentration in early infancy, Amer. J. Diseases of Children 137(9):886-891.5. Centers for Disease Control (1991), U.S. Department of Health and Human Services, Public Health Service. Preventing Lead Poisoning in Young Children, 108 pp.6. U.S. Environmental Protection Agency (1988), Relationship between infant blood lead concentration and lead in water or liquid diet (draft report). Prepared by A. Holtzamn and A. Marcus. Battelle - Columbus Division. EPA Contract No. 68-02-4246.7. U.S. Environmental Protection Agency (1986), Safe Drinking Water Act Amendments. EPA-570-986-002.8. Lee, R.G., W.C. Becker and D.W. Collins (1989), Lead at the tap: Sources and control. J. AWWA 81(7):52-62.9. Schock, M.R. (1985), Treatment or water quality adjustment to attain MCLs in metallic potable water plumbing systems. Plumbing Materials and Drinking Water Quality: Proceedings of a Seminar. EPA-6-600/9-85/007.10. Moore, M.R. (1973), Plumbosolvency of waters. Nature 243(5404):222-223.11. U.S. Public Health Service (1988), Toxicological Profile for Lead. Agency for Toxic Substances and Disease Registry, Atlanta, GA.12. U.S. Environmental Protection Agency (1991), Final Lead Rule, National Primary Drinking Water Regulation. Task Force on Lead in Drinking Water, Office of Drinking Water, Washington, D.C.13. U.S. Environmental Protection Agency (1986), Reducing lead in drinking water: A benefit analysis. EPA-230-09-86-019.14. Pennsylvania Plumbing System Lead Ban and Notification Act (1989), Act No. 1989-33. 35 P.S. |sec~ 723.1.15. Environmental Reporter (1991), The Bureau of National Affairs, Inc., Washington, D.C. November 1, pp. 1651-1652.16. Hallberg, G.R. (1986), From hoes to herbicides: agriculture and groundwater quality. J. Soil and Water Cons. 41 (6):357-364.17. Makuch, J. and J.R. Ward (1986), Groundwater and Agriculture in Pennsylvania. Circular No. 341. Pennsylvania State University. College of Agriculture, Cooperative Extension. 21 pp.18. Hallberg, G.R., R.D. Libra and B.E. Hoyer (1985), Nonpoint source contamination of groundwater in Karst-carbonate aquifers in Iowa. In: Perspectives on Non-point Source Pollution. EPA 440/5 85-001. Washington, D.C. pp. 109-114.19. Fraser, P. and C. Chilvers (1981), Health aspects of nitrate in drinking water. Sci. Total. Environ. 28:103-116.20. Terblanche, A.P.S. (1991), Health hazards of nitrate in drinking water. Water SA 17(1):77-82.21. Exner, M.E. and R.F. Spaulding (1974), Groundwater quality of the Central Platte region. Resource Atlas No. 2. Cons. and Surv. Div., University of Nebraska at Lincoln. 48 pp.22. U.S. Geological Survey (1984), National water summary 1984: Hydraulic events, selected water-quality trends, and ground-water trends, and ground-water resources. United States Government Printing Office, Washington, D.C. pp. 95-97.23. Crawford-Brown, D.J. (1992), Cancer risk from radon, J. AWWA 84(3):77-81.24. U.S. Environmental Protection Agency (1991), Proposed Revisions in USEPA Estimates of Radon Risks and Associated Uncertainties. Unpublished report.25. U.S. Environmental Protection Agency (1985), Nationwide occurrence of radon and other natural radioactivity in public water supplies.26. Gesell, T.F. and H.M. Prichard (1978), The contribution of radon in tap water to indoor radon concentrations. Proc. Sym. Natural Radiation Envir. III, U.S. Department of Energy Conference.27. U.S. Environmental Protection Agency (1991), National Primary Drinking Water Regulations: Radionuclides; Proposed Rule. Fed. Reg. 56:138.28. Hess, C.T., J. Michel, T.R. Horton, H.M. Prichard and W.A. Coniglio (1985), The occurrence of radioactivity in public water supplies in the United States. Health Physics 48(5):553-586.29. Dixon, K.L. and R.G. Lee (1988), Occurrence of radon in well supplies, J. AWWA 80(7):65-70.30. Lowry, J.D. and S.B. Lowry (1988), Radionuclides in drinking water, J. AWWA 80(7):50-64.31. Bacon, A.P.C., K. Froome, A.E. Gent. T.K. Cooke and P. Sowerby (1967), Lead poisoning from drinking soft water, Lancet i:264-266.32: Moore, M.R. (1977), Lead in drinking water in soft water areas -- Health Hazards, Sci. Total Environ. 7:109-115.33. Worth, D., A. Matranga, M. Lieberman, E. De Vos, P. Karalekas, C. Ryan. G. Craun (1981), Lead in drinking water: The contribution of household tap water to blood lead levels. Environmental Lead, Lynam, D.R., L.G. Piantanida and J.F. Cole, eds. Academic Press, New York. NY.34. Benson, J.A. and H. Klein (1983), Lead in drinking water: Investigation of a corrosive water supply. J. Env. Health 45(4):179-181.35. U.S. Environmental Protection Agency (1983), Methods for chemical analysis of water and wastes. U.S. EPA, Environ. Monitor. Support Lab., Cincinnati, OH. EPA 600/4-79-020.36. Lowry, J.D. (1991), Measuring low radon levels in drinking water supplies, J. AWWA 83(4):149-153.37. SAS Institute Inc. (1985), SAS User's Guide: Statistics. Version 5 Edition. Cary, NC. 957 pp.38. Neter, J., W. Wasserman and M. H. Kutner (1985), Applied Linear Statistical Models, Second Edition. Richard D. Irwin. Inc., Homewood. IL. 1,127 pp.39. U.S. Census Bureau (1980), Pennsylvania Housing Census, Table 94: Equipment and Plumbing Facilities for Counties.Bryan R. Swistock, The Pennsylvania State University, 132 Land and Water Research Building, University Park, PA 16802.