disinfection

Detecting fecal waste contamination of groundwater is usually accomplished by testing for

fecal “index” or “indicator” bacteria. The authors examined 128 samples from 26 public
groundwater sources in New Jersey and compared the ability of nine potential indicators to

detect fecal contamination. Each sample was tested for heterotrophic plate count; total
BY THOMAS ATHERHOLT,
coliform (TC); fecal coliform (FC); Escherichia coli, enterococci, and Clostridium perfringens
ERIC FEERST,
bacteria (100 mL); somatic and F+ coliphage (100 mL); and coprostanol (1 and 4 L). Data from
BRUCE HOVENDON,
a source with known contamination showed that the TC test was the most reliable indicator,
JAE KWAK, AND
followed by the FC test. Fecal pollution contains high concentrations of TC bacteria, but
JOSEPH D. ROSEN
because some TC bacteria in the environment may not be of fecal origin, a positive TC test

result should be followed by an FC, E. coli, or enterococci test to confirm contamination.

The sanitary significance of the occasional presence of TC bacteria or coliphage without other

fecal indicators was uncertain. Methods capable of analyzing 1-L volumes should be
investigated. A groundwater should be analyzed multiple times (i.e., 10 or more) to confidently

determine its sanitary status.

Evaluation of
indicators of
fecal contamination
IN GROUNDWATER
ecal waste from humans and animals may contain disease-causing

F microorganisms, and there are several ways fecal waste can contami-
nate groundwater used for drinking (AWWA, 1999; Moe, 1997). Illness
can occur if contaminated water is consumed without prior or ade-
quate treatment to remove or inactivate the pathogens. Therefore, it is
important to detect fecal waste contamination in groundwater, especially for sys-
tems that do not treat water prior to delivery. Fecal contamination is usually deter-
mined by testing water for the presence of certain fecal-derived “index” or
“indicator” bacteria, such as total coliforms (TC), fecal coliforms (FC), or
Escherichia coli (Leclerc et al, 2001; Toranzos & McFeters, 1997). For exam-
ple, all public water systems in the United States that provide piped (delivered)

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 119
water for human consumption are required
to routinely test for TC bacteria to ensure
that the water is free of pathogens (USEPA,
1989b). A public water system is defined as one that has This study compared the ability of nine potential indi-
15 or more service connections or that regularly serves cator tests to detect fecal contamination in groundwater
25 or more people 60 or more days per year. sources used by public water systems in New Jersey. In
US drinking water regulations assume that all sur- addition to the “traditional” bacterial indicator tests such
face water and groundwater under the direct influence as TC, FC, E. coli, and enterococci, the authors included
of surface water (GWUDI), as defined by the US Envi- tests to detect heterotrophic plate count (HPC) bacteria,
ronmental Protection Agency (USEPA, 1989a), contain Clostridium perfringens (CP) spores, somatic and F+ col-
fecal contamination, and therefore testing for indicator iphages, and coprostanol, which is a fecal-derived chemi-
bacteria is not required. All surface and GWUDI waters cal. The “fecal specificity” of each test is shown in Table 1.
must be filtered and/or disinfected to remove or inacti- The authors included testing for HPC bacteria after the
vate pathogens (USEPA, 2002; USEPA 1998; USEPA, study had begun. HPC bacteria are any bacteria able to
1989a). There are no similar US regulations for ground- grow on an enriched growth medium. They have no direct
water sources not under the direct influence of surface sanitary significance, but it was hypothesized that ground-
water, but in 2000 the USEPA proposed the Ground water sources with fecal contamination might have higher
Water Rule (GWR) to protect these sources from fecal HPC concentrations than noncontaminated wells.
contamination (USEPA, 2000). In the proposed GWR, CP is an anaerobic bacterium derived from human
groundwater is assumed to be free of fecal contamina- and animal fecal wastes. Vegetative cells of CP do not
tion. Testing for specified microbes would be required survive in the aerobic environment, but CP forms hardy
only if the source (e.g., well, cistern) is located in a “sen- endospores that survive for many months or years in soils
sitive” hydrogeological location, such as karst, gravel, and sediments (Toranzos & McFeters, 1997; Bisson &
or fractured bedrock, or if there is a total coliform–pos- Cabelli, 1979).
itive sample in the delivered water and no microbe There is some evidence that pathogenic human enteric
removal/inactivation process is in place. Each state would viruses (HEV) may be present in groundwater in the
specify the use of either E. coli, enterococci, or col- absence of indicator bacteria because they are smaller
iphage testing. (and thus can travel farther) and remain infectious longer
E. coli testing (along with fecal coliform testing) is than the bacteria (Borchardt et al, 2003; USEPA, 2000;
specified as one of two alternative tests to be conducted Moe, 1997). However, in groundwater studies comparing
to confirm fecal pollution in a total coliform–positive the detection of HEV and indicator bacteria, HEV are
sample in delivered water (USEPA, 1989b). However, the analyzed from multiple-litre samples whereas the bacte-
specification of enterococci and coliphage, as in the pro- ria are analyzed from 100-mL samples. Detecting HEV is
posed GWR, is novel within a drinking water regulatory difficult, slow, expensive, or in some cases impossible.
context. Since 1986, in order to comply with the Clean Coliphage may be a suitable indicator of HEV contami-
Water Act, the USEPA has recommended that surface nation (Havelaar, 1993). Coliphage are viruses that infect
waters be tested using enterococci (or E. coli in the case E. coli. They are similar to HEV in size and environ-
of fresh waters) (USEPA, 1986). mental survival characteristics (Armon & Kott, 1996;

2003 © American Water Works Association

120 OCTOBER 2003 | JOURNAL AWWA • 95:10 | PEER-REVIEWED | ATHERHOLT ET AL

 The study collected 128 samples from 23 community water systems and 3 noncom-
munity water systems located throughout New Jersey. Sampling points included
cisterns (far left), wells located in pump houses (left), wells located in pits (above),
and wells that were not enclosed in any way. All of the sources of the water were located in unconfined aquifers—meaning that there were
no geological barriers preventing the movement of pathogens from a contaminant source to a potable water source.

Snowdon & Cliver, 1989). Coliphage can be either somatic collected between June 1999 and February 2002. In addi-
or F+ coliphage. Somatic coliphage infect their bacterial tion, 12 sterile, distilled water samples, purchased from
hosts through the cell membrane, whereas F+ coliphage retail food stores in 4-L sealed plastic containers, were
infect certain host cells that possess F pili. F pili are spe- poured into sample containers identical to the source
cialized appendages used by F pili-containing E. coli to sample containers at eight well locations. These field con-
transfer genetic material to recipient E. coli cells lacking trol samples and the four duplicate samples were labeled
F pili. Leclerc et al (2000) have identified a number of with fictitious source names and were analyzed in a
actual or potential limitations of coliphage as reliable “blind” fashion. Eight samples were collected from a
indicators of HEV contamination. Therefore, coliphage wastewater treatment plant. The wastewater samples
should be investigated further as indicators of fecal waste were analyzed to document the concentrations of
(not just HEV) contamination. coprostanol and the indicator microbes in human fecal
Human and animal fecal waste also contain sterol waste before, during, and after treatment.
compounds (Murtaugh & Bunch, 1967). Coprostanol The number of times each source was sampled is shown
(5ß-cholestan-3ß-ol) is the predominant member of a in Table 2. All of the sources are located in unconfined
family of “fecal sterols,” produced from cholesterol in aquifers, which means that there are no geological barriers
the intestines by certain microbes (Leeming et al, 1996; preventing the movement of pathogens from a contami-
Walker, 1982). Coprostanol is a nonpolar compound nant source to the potable water source. Twelve of the
and, like microbes, is associated with the particulate phase sources are classified as GWUDI (Table 2) by the New Jer-
of water samples. It is biodegraded over time under aer- sey Department of Environmental Protection (NJDEP)
obic conditions but is quite stable in anaerobic environ- in accordance with USEPA guidance (USEPA, 1992;
ments (Walker, 1982). Coprostanol was one of the most USEPA 1991). A population of GWUDI sources was
frequently detected compounds in a large survey of streams selected because it was anticipated that such sources
in the United States considered susceptible to contami- would have a higher likelihood of having fecal contami-
nation from human, industrial, and agricultural waste- nation than would non-GWUDI sources.
water (Kolpin et al, 2002). Sampling was conducted in the morning, on a monthly
or less frequent schedule, except for source 6-1, which was
SAMPLING sampled more frequently during the brief periods when
One hundred twenty-eight untreated groundwater the well was in operation. Untreated water sampling taps
samples, including four duplicate samples, were collected were wiped with a 70% ethanol-saturated pad or swab
from 23 community water system and 3 noncommunity and then run for several minutes before the sample was
water system sources as shown in Figure 1 and Table 2. drawn. The water temperature was measured at this time.
A public community water system serves a year-round Samples for microbial analysis were collected in two ster-
(resident) population. A public noncommunity system ile, reusable, 1-L polypropylene containers. Samples for
serves water to a nonresident population (i.e., school, sterol analyses were collected in 1-L (June 22, 1999–Mar.
church, factory, or roadside rest stop). The samples were 22, 2000) and/or 4-L (Mar. 22, 2000–Feb. 5, 2002) one-

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 121
use-only, amber glass, certified chemical-free bottles.1 Sam- after 57 samples had been processed, and CP monitor-
ples were stored in iced coolers between 0 and 10oC and ing was not performed on the three noncommunity sam-
were delivered to the analytical laboratories within 6 h. ples or on the last 12 samples collected (December
2001–February 2002). Bacteria data are reported as
ANALYSIS cfu/100 mL, except for HPC bacteria that are reported
All microbe tests were initiated immediately following as cfu/1 mL. Coliphage data are reported as pfu/100
sample delivery. Samples for sterol analysis were stored at mL. Data for samples in which no microbes were found
2–5oC (36–41oF) until analysis, which was several days are reported as 0/100 mL rather than <1/100 mL per
or weeks following delivery. Atherholt and Korn (1999).
Microbe tests. All of the microbe tests used in this Coprostanol. Samples were spiked with a sterol surro-
study are quick (48 h or less) and inexpensive ($50 or gate (cholesterol-2,2,4,4,6-d5), and the sterols were sep-
less), and none require special skills or costly equip- arated from the aqueous phase by solid-phase extraction
ment (Table 1). All tests involved membrane filtration of using a precleaned, 50 mm (2 in.) polymer disk3 as per
100-mL aliquots, except for a few coliphage samples in Kwak and Rosen (2002). The disks were extracted with
which 500- or 1,000-mL volumes were filtered (Tables acetone followed by dichloromethane (DCM), concen-
3 and 4). Two modifications of the coliphage assay were trated, filtered, and reconcentrated. The sterol hydroxyl
used. The E. coli host cultures were grown for 21–24 h group was reacted with bis-(trimethylsilyl)trifluoroac-
rather than 4 h, and one rather than two 5-mL beef etamide + 1% trimethylchlorosilane and an internal stan-
extract/surfactant2 eluting aliquots was used. Coliphage dard (chrysene-d12) was added before analysis.
monitoring was not initiated until after 16 samples had Extracted sterols were separated and identified using
been collected and analyzed for the other indicators. gas chromatography–mass spectrometry (GC–MS) using
HPC testing, not part of the original study design, began a 30 m (98.4 ft) × 0.25 mm (0.01 in.), 0.25 µm film thick-

TABLE 1 Indicator tests

Does a Positive Test Indicate

Indicator Method Fecal Contamination? Notes

Bacteria
HPC* SM9215D; m-HPC agar† No† No direct sanitary significance
Total coliforms SM9222B† Suggestive‡ High concentrations in fecal
(Leclerc et al, 2001; Toranzos waste but nonfecal sources also
and McFeters, 1997)
Fecal coliforms SM9222D† Likely‡ Minor nonfecal component in
(Leclerc et al, 2001; Toranzos many but not all waters
and McFeters, 1997)
E. coli USEPA§ 1103.1 (Dufour et al, 1981) Yes Considered fecal-specific but some
(Leclerc et al, 2001; Toranzos evidence of multiplication in
and McFeters, 1997) warm waters and sediment
Enterococci USEPA 1600 (USEPA, 1997a) Likely Minor nonfecal component in
(Leclerc et al, 1996) many but not all waters
Clostridium perfringens Bisson and Cabelli (1979) Yes Spores able to survive in
(Leclerc et al, 2001; Toranzos environment longer than most
and McFeters, 1997) enteric pathogens
Bacterial virus
Somatic coliphage Hsu et al (1998); on E. coli CN-13, Uncertain (suggestive or likely);‡ Potential human enteric virus
Payment (1991) (Leclerc et al, 2000; Stewart, indicator?
1998; Armon and Kott, 1996;
Snowdon and Cliver, 1989)
F+ coliphage Hsu et al (1998); on E. coli Famp, Yes Potential human enteric virus
Debartolomeis and Cabelli (1991) (Leclerc et al, 2000; Stewart, indicator?
1998; Armon and Kott, 1996;
Snowdon and Cliver, 1989)
Chemical
Coprostanol See text Yes Nonpolar, particle-bound, very
(Walker, 1982; Murtaugh and stable in anaerobic environment
Bunch, 1967)

*HPC—heterotrophic plate count

†SM—Standard Methods, 1998
‡Suggestive—confirmatory testing necessary; likely more definitive than total coliform test but can contain nonfecal sources
§USEPA—US Environmental Protection Agency

2003 © American Water Works Association

122 OCTOBER 2003 | JOURNAL AWWA • 95:10 | PEER-REVIEWED | ATHERHOLT ET AL

TABLE 2 Source information

Source

System/ Well Depth to Pump Number

Well Start Depth TS or BC Rate Wellhead of Times
Number Date ft (m) ft (m)* gpm (L/min) Enclosure† GWUDI‡ Sampled Notes

Community
wells
1-1 1951 57 (17) 47 (14) 1,000 (3,785) Pump house No 5 Periodically out of service
1-2 1971 167 (51) 132 (40) 700 (2,650) Pump house No 5 Periodically out of service
1-3 1963 198 (60) 166 (51) 700 (2,650) Pump house No 1
1-4 1965 176 (54) 145 (44) 1,000 (3,785) Pump house No 1
1-5 1968 133 (40) 100 (30) 1,000 (3,785) Pump house No 1
2-1 1983 35 (11) 26 (8) 500 (1,893) None Yes 8 20 ft (6.1 m) diameter infiltration
cistern with a roof, internal catwalk
to sampling pipe, and open to
atmosphere via screened vents
2-2 1952 198 (60) 173 (53) 700 (2,650) Pump house No 4 Taken out of service during project
2-3 1977 187 (57) 132 (40) 800 (3,028) Pump house No 10§
3-1 1965 90 (28) 70 (21) 325 (1,230) Pit** Yes 4 Chemical contamination; periodically
active
3-2 1958 76 (23) 65 (20) 350 (1,325) Pit Yes 6 Chemical contamination; periodically
active
3-3 1980 291 (89) 213 (65) 1,250 (4,731) None No 6 Out of service in winter
3-4 1982 141 (43) 102 (31) 400 (1,514) None No 6
3-5 1973 136 (42) 78 (24) 500 (1,893) None No 1
4-1 1965 80 (24) 50 (15) 500 (1,893) None Yes 3
4-2 1983 85 (26) 60 (18) 700 (2,650) None Yes 3
4-3 1983 87 (27) 67 (20) 700 (2,650) Pump house Yes 7
4-4 1988 66 (20) 66 (20) 2,800 (10,598) Pump house Yes 7 Collector well: vertical shaft with two
pumps; connected to five horizontal,
radiating, screened pipes beneath
an artificial recharge basin
5-1†† 1971 355 (108) 52 (16) 250 (946) Pit No 6
5-2 1944 502 (153) 462 (141) 185 (700) Pump house No 10§ Chemical contamination
5-3 1913 125 (38) 44 (13) 70 (265) None No 10§ Chemical contamination
5-4 1965 160 (49) 96 (29) 600 (2,271) Pump house No 3 Taken out of service during project
6-1†† 1990 400 (122) 152 (46) 243 (920) Pump house Yes 7 Infrequently active; known
contamination
7-1 pre-1925 NA‡‡ NA 20 (76) Pit Yes 7§
Noncom-
munity wells
NC-1†† 1988 75 (23) NA 10 (38) None Yes 1 Known contamination
NC-2 1996 450 (137) 60 (18) 25 (95) None Yes 1
NC-3 1997 75 (23) 30 (9) 5 (19) None Yes 1 Seasonal use only
Quality 12
control—
distilled
water
Sewage 3
treatment
plant

*Depth to top of the screened interval (TS) or to the bottom of the casing for wells in consolidated formations (BC)
†All wells grouted, have a pitless adapter and/or a sanitary seal or equivalent; wellheads extend aboveground except those in pits.
‡Designated by New Jersey Department of Environmental Protection as ground water under the direct influence of surface water (GWUDI) (USEPA, 1992;
USEPA,1991).
§Duplicate sample taken once; not included in total
**Pit—wellhead terminates belowground in a cement vault with a sump pump and with an entry door and vault top aboveground.
††Multiple fecal indicators found in one or more samples
‡‡NA—unknown

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 123
ness capillary GC column4 and with the mass spectrom- Sources could be grouped into three categories. The
eter operated in the electron ionization mode.5 Copro- three categories into which the sources could be grouped
stanol was detected by monitoring the 370 and 215 ion were: sources in which multiple indicators were detected
peaks. The recovery efficiency for the sterol surrogate (Table 3), sources with TC bacteria and/or coliphage
and for coprostanol in distilled water was 79% (coefficient but no other indicators (Table 4), and sources free of all
of variation [CV] = 19%) and 93% (CV = 22%), respec- indicators.
tively. The coprostanol detection limit in distilled water Indicator performance could be compared by examining
was 2 ng/L. Sample matrixes significantly influenced sterol the data from the sources in which multiple indicators were
recovery efficiencies, probably because of the solid-phase detected. Because all E. coli and the majority of FC and
extraction method used. Surrogate sterol recovery effi- enterococci are of fecal origin (Table 1), the presence of
ciencies for the field samples ranged from 7 to 157% any of these microbes in a source is strong evidence of fecal
(median 64%). contamination. Finding these indicators on multiple occa-
sions adds weight to the evidence. The data show that
RESULTS AND DISCUSSION source 6-1 is subject to frequent if not continuous fecal
HPC bacteria, but no other indicators, were detected in contamination (Table 3). This source was sampled seven
some of the field control samples. Surprisingly, HPC bacteria times during its brief periods of operation. TC bacteria
were detected in 5 of the 12 sterile, distilled water control were observed in all seven samples, FC in six samples,
samples at concentrations of 1, 2, 2, 2, and 3 cfu/mL. E. coli and enterococci in four samples, CP and coliphage
HPC bacteria, but no other indicators, were also observed in two samples, and coprostanol in one sample. Well 6-
in the four sets of duplicate source samples. The concen- 1 is classified as GWUDI with a known history of fecal
trations were within normal intersample variation: source contamination.
5-3 (sampled Feb. 16, 2000), 18 and 28 cfu/mL; source CP, coliphage, and coprostanol were observed only
7-1 (June 27, 2000), 14 and 26 cfu/mL; source 2-3 (Dec. when the other indicators were at comparatively high
17, 2001), 53 and 49 cfu/mL; and source 5-2 (Feb. 5, levels. The absence of coliphage and CP in some sam-
2002), 0 and 1 cfu/mL. ples may be because coliphage and CP are present in fecal

TABLE 3 Sources with multiple indicators detected*

Water E CP Som F+
Sample Temperature HPC TC FC E. coli cfu/ cfu/ Phage Phage Coprostanol
Date oF (oC) cfu/mL cfu/100 mL cfu/100 mL cfu/100 mL 100 mL 100 mL pfu/100 mL pfu/100 mL ng/L

Source 6-1†
07/13/99 53 (11.5) ‡ 8 2 2 2 0 0 0 None
07/20/99 52 (11) ‡ 4 3 0 0 0 0 0 None
07/26/99 52 (11) ‡ 1 1 0 0 0 0 0 None
08/04/99 52 (11) ‡ 1 0 0 0 0 0 0 None
07/12/00 52 (11) 3,800 3,620 38 23 5 0 10 137 13§
08/23/00 53 (11.5) 168 19 1 4 2 3 0 0 None §
09/13/00 52 (11) 2,600 5,300 262 31 405 2 55 7 None §
Source 5-1
06/28/99 56 (13.5) ‡ 0 0 0 0 0 ‡ ‡ None
08/04/99 55 (13) ‡ 0 0 0 0 0 0 0 None
10/06/99 55 (12.5) 66 0 0 0 0 0 0 0 None
03/15/00 55 (13) 65 0 0 0 0 0 0** 0** None
05/16/00 54 (12) 782 20 2 0 5 0 9 0 None §
10/12/00 55 (12.5) 156 0 0 0 0 0 0 0 None§
Source NC-1
02/21/01 53 (11.5) 578 14 5 2 0 ‡ 31** 0†† 3.5§

*HPC—heterotrophic plate count; TC—total coliforms; FC—fecal coliforms; Escherichia coli; E—enterococci; CP—Clostridium prefringens; Som phage—Somatic
coliphage on E. coli CN-13; F+ Phage—F+ coliphage on E. coli HS(pFamp)R; none—none detected
†Reserve well (not usually in operation); operational 7/5/99–8/29/99 because of drought and demand; operational 7/10/00–7/17/00 for maintenance; operational
8/13/00–11/15/00 because of sudden loss of primary (surface) source
‡Not done
§From a 4-L sample
**Per 500 mL
††Per 1,000 mL

2003 © American Water Works Association

124 OCTOBER 2003 | JOURNAL AWWA • 95:10 | PEER-REVIEWED | ATHERHOLT ET AL
pollution at lower densities than the other indicator bac-
teria as shown by Chauret et al (1999) and this study New Jersey public water system source locations and
FIGURE 1
(Table 5). Kwak and Rosen (2002) showed that TC and identification numbers. Sources labeled with a “G” are
FC bacteria could be consistently detected in 1-mL or groundwaters under the direct influence of surface
greater aliquots and coprostanol in 50-mL or greater water
aliquots of river water, making the latter a less-sensitive
indicator of surface water contamination. Thus, if the
amount of fecal concentration in a source is low, it is
NC–3 (G)
possible that TC, FC, E. coli, and enterococci could be NC–1 (G)
NC–2 (G)
detected in the absence of other indicators. The absence 6-1 (G)
of CP in groundwater has been observed by several inves-
7-1 (G)
tigators (Francy et al, 2000). CP and coliphage may be
detected more frequently by analyzing larger volumes, 5–3
but filter clogging can become a problem, as described later 5–2 3–3
in this article. Also, if a different coliphage assay had 5–1 3–4
3–5
been used, such as method 1601 (USEPA, 2001a) or 2–1 (G)
method 1602 (USEPA, 2001b), different results might 2–2
2–3
have been obtained. However, these methods were not 5–4 3–1 (G)
3–2 (G)
available when this study began. 4–1 (G)
Each state now has a source water assessment program 1–1 4–2 (G)
4–3 (G)
(SWAP) to protect public drinking water sources (USEPA, 1–3 1–2 4–4 (G)
1–4
1997b). In New Jersey, SWAP information, in electronic
format, includes a susceptibility ranking for each well
based on three delineated wellhead protection zones (cal-
culated using hydrogeological information) (NJGS, 2001)
and existing contaminant source inventories (NJDEP, Community wells
Noncommuntiy wells
1999). New Jersey used a conservative two-year time-of- Counties
travel to calculate wellhead protection zones for pathogens 1–5
(tier 1 zones). Water systems should use SWAP informa- Wastewater
tion to help identify sources of fecal pollution if present. treatment
plant
However, when the SWAP information was examined,
the origin of the contamination in source 6-1 was not
apparent. The well is located within limestone geology that 20 0 20 40 Miles
is subject to chemical solution and the formation of karst
landscape, but the well is also located in a sewered resi-
dential area (i.e., no nearby septic tanks). Portions of
wetlands, streams, and a small lake lie within the tier 1
protection zone. Thus, this well could be affected by one the well’s tier 1 zone, and stormwater from a road drains
or more of the surface water sources, by stormwater infil- into the stream within the tier 1 zone above the source.
tration or leaking stormwater drainpipes, or by leaks in As with source 6-1, surface water, stormwater, or leaking
the sewage collection system. In areas with trees, identi- sewage-collection pipes could be the pollution source.
fication of homeowners that have had waste drain pipes The mean HPC bacteria concentration (267 cfu/mL) was
clogged by tree roots may be a cost- and time-effective way significantly greater than the mean concentration of the
of identifying some sewage collection system leaks, but this other non-GWUDI sources (44 cfu/mL; duplicate data
was not attempted. averaged) (Wilcoxon Rank Sum, 2-sided: p = 0.003), sug-
Contamination can be sporadic and/or low. TC, FC, ente- gesting that this source might be misclassified as to
rococci, and somatic coliphage were observed in source GWUDI status.
5-1, but in just one of six samples (Table 3). The presence Analysis of operational water quality and meteoro-
of both FC and enterococci is strong evidence of fecal logical data failed to identify stormwater as the pollu-
contamination. A source water assessment was conducted, tion source in source 5-1. The well operators monitor
but, as with source 6-1, the origin of the contamination the untreated water twice per week for TC but not for
is not apparent. The well is deep (although some water- other fecal contamination indicators. TC was detected
bearing zones may not be deep) and is not designated as in 12 of 97 samples analyzed during the study period
GWUDI (Table 2). It is located in argillite (shale) in a (June 28, 1999–Oct. 12, 2000; Figures 2 and 3). These TC
sewered residential area, next to a small stream that has detections support the validity of the May 16, 2000, pos-
no flow during dry periods. Portions of wetlands lie within itive test results (Table 3). Daily precipitation data from

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 125
 treated sources in most instances. It
TABLE 4 Sources with TC and/or coliphage but no other indicators* is interesting that four of these sources
(80%) are classified as GWUDI (Table
Date TC† 2). To confirm contamination, it
and/or Water Total Somatic F+ would have been desirable to collect
Samples Coliphage Temperature Coliforms Coliphage Coliphage
Source Collected Observed oF (oC) cfu/100 mL pfu/100 mL pfu/100 mL additional samples following the TC-
positive samples and test them for TC
09/09/99 55 (13) 0 0 1
and either FC, E. coli, or enterococci
2-3 10 03/22/00 54 (12) 0 1 0
in a manner similar to the Total Col-
06/23/99 ‡ 3 ‡ ‡
iform Rule (USEPA, 1989b), but this
3-1 4 08/19/99 75 (24) 0 2 1
was not part of the study design.
3-2 6 08/19/99 59 (15) 2 0 0
Somatic and/or F+ coliphage were
3-4 6 08/19/99 56 (13.5) 0 1 0
found in six sources (eight samples)
4-4 7 03/22/00 52 (11) 0 1§ 0
in the absence of other indicators
09/01/99 55 (13) 0 0 1
except TC bacteria (Table 4), despite
5-2 10 11/17/99 54 (12.5) 0 1 0
the fact that the coliphage method
09/01/99 56 (13.5) 0 1 1
employed two modifications that
5-3 10 11/17/99 61 (16) 12 0 0
may have adversely affected its per-
08/04/99 54 (12) 1 0 0
formance (see the analysis section)
7-1 7 10/25/00 52 (11) 1 0 0
(Hsu et al, 1998). Sampling was lim-
NC-3 1 02/21/01 50 (10) 1 0§ 0**
ited in a few cases. Pillai and
*TC—total coliforms
†All indicators absent in the samples not shown
Nwachuku (2000) and Dutka et al
‡Not done (1990) observed coliphage in the
§Per 500 mL
**Per 1,000 mL absence of bacterial indicators, but
sampling was limited in these studies
as well. The observed coliphage were
not the result of the conversion of
the municipality (Figures 2 and 3) (station 3029; NOAA, lysogenic to lytic phage during growth of the E. coli
2000; NOAA, 1999) show no apparent short-term influ- hosts as hypothesized by Francy et al (2000) as one of
ence of rainfall on TC detections, but interestingly, there two possible explanations (the other being laboratory
was a dry period for four and a half months from April contamination) for the coliphage they detected in con-
to mid-August of 1999 and a two-month dry period in trol samples. Somatic coliphage were observed in six of
October and November of 2000, and no TC was detected the 104 source samples (including duplicate samples)
during these periods. The operators should examine source that were free of all bacteria except TC, whereas F+
5-1 further using other indicators in addition to TC. coliphage were observed in four of these samples, rep-
Septic tanks are the likely contamination source in one resenting a 6% and 4% lytic conversion rate, respec-
case. Source NC-1 is classified as GWUDI with a known tively. Seventeen and 25 control samples would be needed
history of fecal contamination. This source was only to detect a 6% and 4% conversion rate, respectively,
sampled once, but TC, FC, E. coli, and somatic col- assuming that such events, if they occur, occur ran-
iphage were detected (Table 3). The sample was not domly. Only 12 control samples were analyzed. Although
tested for CP. As is source 6-1, this source is located in routine method blanks were not analyzed and not
limestone that is subject to solution cavities. Fecal con- enough field control samples were analyzed to detect
tamination is most likely from septic tank effluent from random conversion events if that was the cause, detec-
nearby houses and businesses. Septic tanks are well tion of the coliphage was not random. For example,
known sources of groundwater contamination in these five of the eight coliphage-positive samples (63%) (Table
geological areas (e.g., Paul et al, 1997). 4) were collected between Aug. 19, 1999, and Sept. 9,
The sanitary status of sources containing only TC or col- 1999, whereas only 10 of the 128 (8%) project sam-
iphage, but no other indicators, is uncertain. TC bacteria, but ples were collected during this period.
no other indicators except coliphage, were observed in five There are four reasons why the sources with coliphage
sources (Table 4). Because there may be significant nonfe- but no other indicators except TC bacteria could con-
cal sources of TC in environmental waters, it is not clear tain fecal contamination:
if the TC-positive sources are subject to fecal contami- • F+ coliphage, like their F+ bacterial hosts, are
nation. Although all TC may not be of fecal origin, derived exclusively from fecal wastes (Calci et al, 1998;
untreated fecal pollution sources typically contain high Stewart, 1998).
concentrations of these bacteria (Table 5), and untreated • Although not all somatic coliphage are necessarily
wastes would be more likely to affect groundwater than of fecal origin (Leclerc, 2000; Stewart, 1998), sewage

2003 © American Water Works Association

126 OCTOBER 2003 | JOURNAL AWWA • 95:10 | PEER-REVIEWED | ATHERHOLT ET AL

 FIGURE 2 Untreated source 5-1 TC monitoring results, water temperature, and daily precipitation data for 1999. Water temperature points
without a corresponding TC concentration bar mean that no TC bacteria were detected in that sample.

7 (2.76) TC concentration 20
TC sample temperature
Precipitation 5
6 (2.36) 18

TC Concentration—cfu/100 mL

TC Sample Temperature— C
o
5 (1.97) 4 16
Precipitation—cm (in.)

4 (1.57) 14
3

3 (1.18) 12
2

2 (0.79) 10

1
1 (0.39) 8

0 (0.0) 0 0
Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.

Date—1999

cfu—colony-forming units, TC—total coliform

and septage usually contain considerable quantities of albeit at opposite ends of the room and not always at
both types of coliphage (Chauret et al, 1999; Gantzer the same time that the groundwater samples were
et al, 1998). processed. Few samples were collected the second sum-
• Although coliphage are present in fecal pollution mer and none the third summer because of laboratory
sources at lower concentrations than indicator bacteria, logistical constraints. Aerosol contamination by col-
coliphage may survive longer and travel farther than indi- iphage could have occurred during sample filtration
cator bacteria in some groundwaters (Sinton et al, 1997) or media pouring, mostly during the initial summer
and thus could be present in some sources in the absence period, but again, there were not enough control sam-
of fecal indicator bacteria. ples, and they were not collected at the appropriate
• Coliphage were not detected in the field control time period to detect it. Coliphage were not detected in
samples. any sample after March 2000 unless multiple fecal
On the other hand, there are three reasons why some indicators were also observed, and control samples
sources with coliphage but no other indicators except were not analyzed before June 2000. Again, routine
TC bacteria may not be contaminated: method blanks were not part of the coliphage assay, but
• Only two of the six wells (sources 3-1 and 4-4) are even if included they may not detect infrequent aerosol
classified as GWUDI (Table 2). contamination events. In support of the aerosol cont-
• As stated, somatic coliphage may not be of fecal amination hypothesis, coliphage counts in these sources
origin. were 1 pfu/100 mL in all but one case. Coliphage
• The occasional detection of coliphage might be the counts higher than this were always observed in the
result of aerosol contamination (Carducci et al, 2000; sources with multiple indicators (Table 3). It may be
Medema, et al, 2000; Fannin et al, 1977). It should be that waterborne coliphage were present in these sources
noted that non-GWUDI status does not preclude the pos- at the lowest detectable concentration, but such a
sibility of contamination and that the wells may not be prevalence of singular counts was not expected given
accurately classified. the observed bacterial data. Pillai and Nwachuku
Aerosol contamination could have occurred during (2000) also observed singular counts for F+ coliphage
this study because considerable pipetting of shellfish using method 1602 (USEPA, 1999), which is superior
harvest water samples and media for most probable to the method used in this study at 100-mL volumes
number (MPN) TC tests took place, mostly during the (Hsu et al, 1998). Additionally, coliphage were occa-
summer, in the same room as the coliphage analyses, sionally observed in negative control samples during

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 127
TABLE 5 Indicator concentrations at a large wastewater treatment plant*

CP Som F+
TC† FC E. coli E cfu/ Phage Phage Coprostanol
Sample Date cfu/100 mL cfu/100 mL cfu/100 mL cfu/100 mL 100 mL pfu/100 mL pfu/100 mL µg/L

Untreated 09/25/01 7,100,000 5,400,000 4,000,000 975,000 78,000 ND‡ 47,000 744
01/15/02 42,600,000 7,200,000 7,000,000 4,200,000 ND ND 120,000 2136
Treated, before 06/14/00 600,000 100,000 200,000 3,200 2,360 >2,000 2,100 23§
chlorination 09/25/01 3,600,000 1,300,000 740,000 75,000 2,100 ND 500 6
01/15/02 2,980,000 1,080,000 720,000 250,000 ND ND 20,320 690
Treated, after 06/14/00 2,380 41 64 0 3,600 1,750 2,000 11§
chlorination 09/25/01 223 8 3 0 1,560 ND 200 4
01/15/02 323 10 2 1 ND ND 16,020 747

*Primary clarification, aeration with recycled sludge, secondary clarification, chlorination; capacity: 40 mgd (151 ML/d)
†TC—total coliforms; FC—fecal coliforms; E. coli—Escherichia coli; CP—Clostridium prefringens; Som Phage—Somatic coliphage; F+ Phage—F+ coliphage
‡Not done
§Delivered to analytical laboratory 22 h following collection and storage at 4oC

validation of coliphage Method 1601 and method 1602 preferable to analyze volumes greater than 100 mL.
(USEPA, 1999). Processing and testing samples within However, increasing the volume for regulatory purposes
laminar flow hoods with high-efficiency particle-arrest- may necessitate modifying standard methods that spec-
ing filters would preclude aerosol contamination, if it ify 100-mL volumes (Standard Methods, 1998; USEPA,
occurs, but such equipment is not available in many test 1997a). In addition, clogging was encountered when
laboratories. 500-mL or 1,000-mL volumes of some samples were
Most sources were free of all indicators. Fourteen sources filtered (using 47-mm-diameter filters). This problem
were free of all indicators (Table 2; 1-1 through 1-5, 2-1, could be circumvented with the use of larger filters or by
2-2, 3-3, 3-5, 4-1, 4-2, 4-3, 5-4, and NC-2). Even though performing replicate filtrations, but volumes greater
no indicators were observed, the number of samples col- than 100 mL should not be specified in a regulation
lected at most of these sources was limited. Also, five of without method validation studies using a variety of
these sources were classified as GWUDI. Thus, further source samples.
study of most of these sources is indicated. Four-L containers were used to collect 54 samples for
High HPC counts may indicate potential contamination, coprostanol analysis, but these containers were cumber-
but further study is required. HPC bacteria counts in the some because of their size and weight. Most 4-L con-
67 samples from the 22 sources that were sampled for tainers do not fit within most coolers, and glass contain-
HPC averaged 180 cfu/mL (median = 20, range 0–3,800; ers this size are more prone to breakage than smaller
duplicate data averaged). HPC concentrations were sig- containers. If shipping 4-L containers to a laboratory for
nificantly higher in the eight samples analyzed from analysis was required, it would also be expensive. Mul-
the three sources (Table 3) in which multiple fecal indi- tiple 1-L samples can be collected, but the shipping-cost
cators were observed (median 373; mean 1,027; range issue remains, and the number of bottles that would accu-
65–3,800 cfu/mL) than in the 59 samples from the other mulate when sampling multiple sources at one time could
19 sources (median 15; mean 65; range 0–700 cfu/mL) become unwieldy. Therefore, for groundwater sources,
(p = 2 × 10–5). Also, HPC concentrations were signifi- the maximum reasonable sample volume is 1 L and this
cantly higher in the 26 samples analyzed from the 10 volume may be the preferred minimum as well. One-L vol-
sources designated as GWUDI (median 68; mean 359; umes have been advocated by others (Fujioka &
range 0–3,800 cfu/mL) than in the 41 samples from 12 Yoneyama, 2001).
non-GWUDI sources (median 11; mean 66; range 0–782 Water system operators should consider using small-
cfu/mL) (p = 0.001). An individual source could not diameter sampling pipes. The diameter of the sampling
be accurately predicted to be GWUDI or non-GWUDI pipe at source 1-1 was large compared with most of the
based on comparative HPC concentration data, but it other sampling pipes. An insect constructed a cocoon
is possible that some sources are improperly classified inside this pipe between the second and third sam-
as to GWUDI status. pling dates during the six-month period when this
The sample volume for groundwater sources should be well was not in use. It is possible that the insides of
greater than 100 mL. To reliably detect fecal pollution in large-diameter taps could be occasionally contami-
groundwater, the data in Table 3 indicate that it may be nated by these or other insects such as houseflies or by

2003 © American Water Works Association

128 OCTOBER 2003 | JOURNAL AWWA • 95:10 | PEER-REVIEWED | ATHERHOLT ET AL

FIGURE 3 Untreated source 5-1 TC monitoring results, water temperature, and daily precipitation data for 2000. Water temperature points
without a corresponding TC concentration bar mean that no TC bacteria were detected in that sample.

7 (2.76) TC concentration 20
TC sample temperature
Precipitation 5
6 (2.36) 18

TC Concentration—cfu/100 mL

TC Sample Temperature— C
o
5 (1.97) 4 16
Precipitation—cm (in.)

4 (1.57) 14
3

3 (1.18) 12
2

2 (0.79) 10

1
1 (0.39) 8

0 (0.0) 0 0
Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.

Date—2000

cfu—colony-forming units, TC—total coliform

airborne microbes from nearby sources (Trest et al, tests would cost $400–$600 at a reasonable cost estimate
1999). Thus, the ends of large-diameter sampling pipes of $40–$50 per test. This expense should be kept in per-
should be capped or the pipes replaced with small- spective, especially for sources not using filtration or dis-
diameter pipes. infection, with the cost of a disease outbreak. The cost of
The proposed GWR treats any coliphage detection as a the May 2000 E. coli disease outbreak in Walkerton,
true positive with potentially significant financial conse- Ontario, was seven deaths, numerous people with chronic
quences for the owner of the source. A single positive E. health problems, a total expense of $150 million (Cana-
coli, enterococci, or coliphage test result would require dian), careers destroyed, and the loss of public faith in the
one of four alternative corrective actions (or a plan of water system as well as in local and provincial government
action) within 90 days: elimination of the source of the (O’Connor, 2002).
contamination, correction of the well deficiency (e.g., a
cracked casing), provision of an alternative source of CONCLUSIONS
water, or provision of 99.99% removal/inactivation of • Although limited, the data show that the single most
viruses (USEPA, 2000). (Compliance with these actions reliable indicator to detect fecal pollution in a groundwater
would be waived if five followup samples are all nega- source known to be contaminated was TC bacteria, fol-
tive and if there have been no positive samples within the lowed by FC bacteria. Because many TC bacteria in the
past five years.) Thus, it is imperative that a positive environment may not be derived from fecal sources, a
sample be a true-positive due to fecal contamination positive TC test result should be followed up with testing
and not a false-positive due to a source other than the for either FC bacteria, E. coli, or enterococci, perhaps in
groundwater. With regard to the GWR and coliphage- a manner similar to that used in the Total Coliform Rule
positive samples, the data in this study, along with other (USEPA, 1989b).
evidence, suggest that the potential false-positive issue • FC bacteria should be included among the choice
warrants further study. of fecal indicators in the GWR, along with E. coli and
A source should be sampled multiple times to confidently enterococci.
determine its sanitary status. This recommendation is sup- • Coliphage tests using the filter adsorption/elu-
ported by the data in Table 3 (source 5-1). Sampling mul- tion procedure and the CP test were less able to detect
tiple times entails added expenses compared with sampling fecal contamination than the TC, FC, E. coli, or ente-
just one or a few times. For example, 10 or 12 indicator rococci tests at equal volumes, and analysis of 1-L or

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 129
4-L volumes for coprostanol was also a less suitable • A source should be sampled multiple times (i.e., 10
method. or more) over an extended period of time to confidently
• The sanitary significance of the occasional presence determine its sanitary status.
of TC bacteria and/or coliphage in the absence of other
fecal pollution indicators was uncertain. ACKNOWLEDGMENT
• To maximize the possibility of detecting fecal conta- This study was funded by the New Jersey Department
mination (beyond testing for TC or FC bacteria), methods of Environmental Protection. Complete name-censored
capable of analyzing volumes greater than 100 mL should data for this project are available in electronic format
be studied, keeping in mind standardized test requirements upon request ([email protected]). The assis-
and potential clogging problems with filtration methods. tance and cooperation of the participating public water
• HPC bacteria testing was useful for quality control systems are greatly appreciated.
purposes. HPC concentrations in a source significantly
above the average for all sources could serve as a trigger ABOUT THE AUTHORS:
for fecal indicator monitoring, but the data are limited, and Tom Atherholt6 is a research scientist in the Division of
further study is indicated. Science, Research, and Technology for the New Jersey
• Large-diameter sampling pipes should be capped Department of Environmental Protection (NJDEP)
or replaced with small-diameter pipes. 401 E. State St., Trenton, NJ 08625; e-mail tom.ather-

REFERENCES Escherichia coli. Appl. & Environ. Micro- tary Quality of New Jersey Waters. Final
biol., 41:1152. Rept. Contract SR00-049, New Jersey
Armon, R. & Kott, Y., 1996. Bacteriophages as Dept. of Envir. Protection, Trenton, N.J.
Indicators of Pollution. Crit. Rev. Environ. Dutka, B.J. et al, 1990. The Presence of Bacter-
Sci. & Technol., 26:4:299. ial Virus in Groundwater and Treated Leclerc, H. et al, 2001. Advances in the Bacteri-
Drinking Water. Envir. Poll., 63:293. ology of the Coliform Group: Their Suit-
Atherholt, T. & Korn, L., 1999. ICR Protocol: ability as Markers of Microbial Water
Alternative Treatment of Parasite Sample Fannin, K.F. et al, 1977. Field Studies on Col-
Safety. Ann. Rev. Microbiol., 55:201.
Data. Jour. AWWA, 91:3:95. iphages and Coliforms as Indicators of
Airborne Animal Viral Contamination From Leclerc, H. et al, 2000. Bacteriophages as Indi-
AWWA, 1999. M48: Waterborne Pathogens. cators of Enteric and Public Health Risk in
Wastewater Treatment Facilities. Wat.
AWWA, Denver. Groundwaters. Jour. Appl. Microbiol., 88:5.
Res., 11:181.
Bisson, J.W. & Cabelli, V.J., 1979. Membrane Leclerc, H. et al, 1996. Taxonomical Changes in
Francy, D.S. et al, 2000. Occurrence and Distri-
Filter Enumeration Method for Clostridium Intestinal (Faecal) Enterococci and Strep-
bution of Microbiological Indicators in
perfringens. Appl. & Environ. Microbiol., tococci: Consequences on Their Use as
Groundwater and Stream Water. Water
37:55. Indicators of Faecal Contamination in
Environ. Res., 72:2:152.
Borchardt, M.A. et al, 2003. Incidence of Drinking Water. Jour. Appl. Bacteriol.,
Enteric Viruses in Groundwater From Fujioka, R.S. & Yoneyama, B.S., 2001. Assessing 81:459.
Household Wells in Wisconsin. Appl. & Vulnerability of Groundwater Sources to
Fecal Contamination. Jour. AWWA, 93:8:62. Leeming, R. et al, 1996. Using Faecal Sterols
Environ. Microbiol., 69:2:1172. From Humans and Animals to Distinguish
Calci, K.R. et al, 1998. Occurrence of Male-spe- Gantzer, C. et al, 1998. Detection of Infectious Faecal Pollution in Receiving Waters.
cific Bacteriophage in Feral and Domestic Enteroviruses, Enterovirus Genomes, Water Res., 30:12:2893.
Animal Wastes, Human Feces, and Somatic Coliphages, and Bacteriodes
fragilis Phages in Treated Wastewater. Medema, G.J. et al, 2000. Aerosolization of
Human-associated Wastewaters. Appl. & Microorganisms in Household Applica-
Environ. Microbiol., 64:12:5027. Appl. & Environ. Microbiol., 64:11:4307.
tions of Non-drinking Water. Proc. 2000
Carducci, A. et al, 2000. Assessing Airborne Havelaar, A.H., 1993. Bacteriophages as Mod- AWWA WQTC, Salt Lake City. (Note: man-
Biological Hazard From Urban Waste- els of Human Enteric Viruses in the Envi- uscript not in Proceedings; presentation
water Treatment. Water Res., 34:4:1173. ronment. ASM News, 59:12:614. received from Medema via personal com-
Hsu, F.-C. et al, 1998. Improved Methods to munication.)
Chauret, C. et al, 1999. Fate of Cryptosporidium
Oocysts, Giardia Cysts, and Microbial Detect Low Levels of Coliphage in Water Moe, C., 1997. Waterborne Transmission of
Indicators During Wastewater Treatment by Enrichment Presence–absence and Infectious Agents. Manual of Environmen-
and Anaerobic Sludge Digestion. Can. Membrane Filter Methods. Proc. 1998 tal Microbiology (C.J. Hurst et al, eds.).
Jour. Microbiol., 45:257. AWWA WQTC, San Diego. American Society for Microbiology Press,
Kolpin, D.W. et al, 2002. Pharmaceuticals, Hor- Washington.
Debartolomeis, J. & Cabelli, V.J., 1991. Evalu-
ation of an Escherichia coli Host Strain mones, and Other Organic Wastewater Murtaugh, J.J. & Bunch, R.L., 1967. Sterols as a
for Enumeration of F Male-specific Bac- Contaminants in US Streams, 1999–2000: Measure of Fecal Pollution. Jour. WPCF,
teriophages. Appl. Environ. Microbiol, A National Reconnaissance. Environ. Sci. 39:404.
57:1301. & Technol., 36:6:1202.
NJDEP (New Jersey Department of Environmen-
Dufour, A.P. et al, 1981. Membrane Filter Kwak, J. & Rosen, J.D., 2002. The Use of tal Protection), 1999. New Jersey Source
Method for the Enumeration of Coprostanol as an Indicator of the Sani- Water Assessment Plan. Trenton, N.J.

2003 © American Water Works Association

130 OCTOBER 2003 | JOURNAL AWWA • 95:10 | PEER-REVIEWED | ATHERHOLT ET AL

 [email protected]. He has spent the Water Monitoring Laboratory, Division of Watershed
past 12 years conducting drinking Management, NJDEP. Joseph Rosen is a professor in
water-related projects at the depart- the Department of Food Science, Cook College, Rut-
ment and has also spent 10 years con- gers University. Jae Kwak is a PhD candidate in the
ducting drinking water and other Cook College Department of Food Science.
research at the Coriell Institute for
Medical Research. In 2001 he FOOTNOTES
1I-CHEM, Nalge Nunc Intl., Newcastle, Del.
received the Professional Achieve- 2Tween80, Sigma Chemical Co., St. Louis, Mo.
ment Award from the state of New Jersey. Atherholt is 3C-18 Speedisk®, J.T. Baker, Phillipsburg, N.J.
4DB-17, J&W Scientific, Agilent Technologies, Folsom, Calif.
a member of AWWA, the American Society for Micro- 5Magnum Ion Trap, Finnigan MAT, Thermo Finnigan LLC, San Jose,
biology, and the Genetic Toxicology Assoc. He previ- Calif.
6To whom correspondence should be addressed
ously has had articles published in JOURNAL AWWA,
Water Research, Cancer Research, and other publica-
tions. He received his PhD in microbiology from Rut-
gers University (New Jersey). Eric Feerst is a supervis- If you have a comment about this article, please contact
ing environmental specialist and Bruce Hovendon is a us at [email protected].
principal environmental specialist with the Marine

NJGS (New Jersey Geological Survey), 2001. Stewart, J.R., 1998. Detection and Charac- USEPA, 1998. National Primary Drinking Water
Draft Guidelines for Delineation of Well terization of Coliphages in a Ground Regulations: Interim Enhanced Surface
Head Protection Areas in New Jersey. Water Aquifer Recharged With Fecally Water Treatment; Final Rule. Fed. Reg.,
NJDEP, Trenton, N.J. Contaminated Surface Water. Master’s 63:241:69478.
thesis, University of North Carolina,
NOAA (National Oceanic and Atmospheric USEPA, 1997a. Method 1600: Membrane Filter
Chapel Hill.
Administration), 2000. Climatological Data, Test Method for Enterococci in Water.
New Jersey, 105:1. National Climate Data Toranzos, G.A. & McFeters, G.A., 1997. Detec- EPA-821-R-97-004, Washington.
Center, Asheville, N.C. tion of Indicator Microorganisms in Envi-
USEPA, 1997b. State Source Water Assess-
ronmental Freshwaters and Drinking
NOAA, 1999. Climatological Data, New Jersey, ment and Protection Programs, Final
Waters. Manual of Environmental Micro-
104:1. National Climate Data Center, Guidance. EPA 816-R-97-009,
Asheville, N.C.
biology (C.J. Hurst et al, eds.). American Washington.
Society for Microbiology Press,
O’Connor, D.R., 2002. Report of the Walkerton Washington. USEPA. 1992. Consensus Method for Determin-
Inquiry. Ontario Ministry of the Attorney ing Groundwaters Under the Direct Influ-
Trest, M.T. et al, 1999. A Study of the Role of
General, Queens Printer for Ontario, Toronto. ence of Surface Water Using Micro-
Airborne Particulates as the Cause of
scopic Particulate Analysis (MPA). EPA
Paul, J.H. et al, 1997. Evidence for Groundwa- Unexplained Coliform Contamination in
910/9-92-029, Port Orchard, Wash.
ter and Surface Marine Water Contami- Drilled Wells. Proc. 1999 AWWA WQTC,
nation by Waste Disposal Wells in the Tampa, Fla. USEPA, 1991. Guidance Manual for Compli-
Florida Keys. Water Res., 31:6:1448. ance With the Filtration and Disinfection
USEPA, 2002. National Primary Drinking Water
Payment, P., 1991. Fate of Human Enteric Regulations: Long-term 1 Enhanced Sur- Requirements for Public Water Systems
Viruses, Coliphages, and Clostridium per- face Water Treatment Rule; Final Rule. Using Surface Water Sources. EPA 68-01-
fringens During Drinking Water Treat- Fed. Reg. 67:9:1812. 6989, Washington.
ment. Can. Jour. Microbiol., 37:154. USEPA, 1989a. National Primary Drinking
USEPA, 2001a. Method 1601: Male-specific (F+)
Pillai, S.D. & Nwachuku, N., 2000. Field Testing and Somatic Coliphage in Water by Two- Water Regulations: Filtration, Disin-
of Coliphage Methods for Screening step Enrichment Procedure. EPA-821-R- fection, Turbidity, Giardia lamblia,
Groundwater for Fecal Contamination. 01-030, Washington. Viruses, Legionella, and Heterotrophic
Proc. 2000 AWWA WQTC, Salt Lake City. Bacteria; Final Rule. Fed. Reg.,
USEPA, 2001b. Method 1602: Male-specific (F+) 54:124:27486.
Sinton, L.W. et al, 1997. Transport of Bacteria and Somatic Coliphage in Water by Single
and Bacteriophages in Irrigated Effluent Agar Layer (SAL) Procedure. EPA-821-R- USEPA, 1989b. National Primary Drinking
Into and Through an Alluvial Gravel 01-029, Washington. Water Regulations; Total Coliforms
Aquifer. Water, Air, & Soil Poll., 98:17. (Including Fecal Coliforms and E. coli );
USEPA, 2000. National Primary Drinking Water Final Rule. Fed. Reg., 54:124:27544.
Snowdon, J.A. & Cliver, D.O., 1989. Coliphage Regulations: Ground Water Rule; Pro-
as Indicators of Human Enteric Viruses in posed Rules. Fed. Reg., 65:91:30194. USEPA, 1986. Ambient Water Quality Criteria
Groundwater. Crit. Rev. Environ. Control, for Bacteria—1986. EPA 440/5-84-002,
USEPA, 1999. Summary of 100-mL and 1-L Two-
19:3:231. Washington.
step Enrichment P/A Procedure Data and
Standard Methods for the Examination of Summary Single Agar Layer (SAL) Enu- Walker, R.W. et al, 1982. Coprostanol as an
Water and Wastewater, 1998 (20th ed.). meration Procedure Data [internal let- Indicator of Fecal Pollution. CRC Crit. Rev.
APHA, AWWA, and WEF, Washington. ters]. Washington. Environ. Control, 12:91.

2003 © American Water Works Association

ATHERHOLT ET AL | PEER-REVIEWED | 95:10 • JOURNAL AWWA | OCTOBER 2003 131