16.1 Bacteriological Analyses: Indicator Organisms, Membrane Filtration & Geometric Mean Reporting
Key Takeaways
- Bacteriological limits are written for indicator organisms such as fecal coliform or E. coli rather than for pathogens directly, because indicators are inexpensive and reliable to measure and their presence signals fecal contamination.
- In the membrane filtration method a measured sample volume is drawn through a 0.45 micron membrane, the membrane is incubated on a selective medium, and colonies are counted, with fecal coliform incubated at 44.5 degrees C plus or minus 0.2 degrees C.
- Membrane filtration results are calculated as colonies counted multiplied by 100 and divided by the milliliters of sample filtered, giving a result in colonies per 100 mL, and the count is only statistically valid within the method’s ideal counting range.
- Bacteriological sample bottles must be sterile and must contain sodium thiosulfate to neutralize any chlorine residual, because residual chlorine continues killing organisms in the bottle and produces a falsely low result.
- Bacteriological permit limits are almost always expressed as a geometric mean rather than an arithmetic average, because bacterial counts span orders of magnitude and a single high value would otherwise dominate the reported result.
16.1 Bacteriological Analyses: Indicator Organisms, Membrane Filtration & Geometric Mean Reporting
Exam Focus: "Collect samples for bacteriological analyses," "Conduct bacteriological analyses," and "Interpret data from bacteriological analyses" are three separate lines in the Laboratory Analysis content area. Disinfection performance is judged entirely by this test.
1. Why Indicator Organisms
Testing directly for every pathogen that might be present in wastewater — Salmonella, Shigella, hepatitis A, Giardia, Cryptosporidium — would be slow, expensive, and unreliable, because pathogens are present intermittently and in low numbers even when the risk is real.
Instead, regulators use indicator organisms: bacteria that are abundant in the intestines of warm-blooded animals, easy and cheap to culture, and present in far greater numbers than pathogens. Finding them means fecal contamination is present; not finding them means disinfection is working.
| Indicator | What It Indicates | Typical Use |
|---|---|---|
| Total coliform | Broad group including environmental organisms; not specific to fecal contamination | Drinking water; less useful in wastewater |
| Fecal coliform | The thermotolerant subset that grows at elevated temperature; strongly associated with fecal contamination | The traditional and still very common wastewater permit parameter |
| E. coli | A specific member of the fecal coliform group; the most direct indicator of fecal contamination | Increasingly the parameter of choice for freshwater discharges |
| Enterococci | Survives better in saline water | Marine and estuarine discharges |
2. Membrane Filtration
The membrane filtration (MF) method is the workhorse of the wastewater laboratory.
- Sterilize the filtration apparatus and use aseptic technique throughout.
- Filter a measured volume of sample through a sterile 0.45 micron membrane filter. Bacteria are retained on the surface of the membrane; water passes through.
- Transfer the membrane to a pad or plate saturated with a selective growth medium.
- Incubate at the temperature and for the time the method specifies. For fecal coliform this is 44.5 degrees C plus or minus 0.2 degrees C, a tight tolerance that requires a properly calibrated and monitored incubator or water bath.
- Count the characteristic colonies — for fecal coliform on mFC medium, the blue colonies.
The Calculation
Colonies per 100 mL = (Number of colonies counted x 100) / mL of sample filtered
Worked example. An analyst filters 25 mL of final effluent and counts 32 blue colonies.
(32 x 100) / 25 = 3,200 / 25 = 128 colonies per 100 mL
Valid Counting Range
A count is only statistically meaningful within the method's ideal counting range — for fecal coliform by membrane filtration this is conventionally 20 to 60 colonies per membrane.
- Too few colonies and the count is dominated by random variation.
- Too many colonies and they crowd, merge, and compete for nutrients, so the count under-reports.
- Because the number of colonies depends on how much sample you filter, and effluent quality varies, filter several different volumes (a dilution series) on every sample so at least one plate lands in the ideal range.
- A plate with excessive growth is reported as TNTC (too numerous to count) and is not used to calculate a result.
3. Other Approved Approaches
- Multiple tube fermentation (MTF) inoculates a series of tubes at several dilutions and reports a statistical Most Probable Number (MPN) based on the pattern of positive tubes. Slower and more labor-intensive, but it works on turbid samples that would clog a membrane.
- Enzyme substrate (defined substrate) methods use a medium containing substrates that specific enzymes in coliforms and E. coli cleave, producing a color change or fluorescence. Results are reported as MPN. These methods are simple and widely used, but the reagent lot and incubation conditions must follow the manufacturer's instructions exactly.
- Approved methods only. Compliance results must come from a method approved for the parameter under 40 CFR Part 136. A method that is convenient but unapproved produces a number you cannot report.
4. Sampling and Handling — Where Results Get Ruined
| Requirement | Reason for It |
|---|---|
| Sterile bottle | Any organism already in the bottle is counted as though it came from the effluent |
| Sodium thiosulfate in the bottle | Neutralizes chlorine residual. Without it, residual chlorine keeps killing organisms during transport and the reported count is falsely low — the plant looks like it is disinfecting better than it is |
| Do not rinse the bottle | Rinsing washes out the thiosulfate |
| Do not touch the bottle lip or the inside of the cap | Contamination from hands is the most common laboratory error in this test |
| Fill without overflowing; leave headspace | Air space is needed to mix the sample before analysis |
| Ice immediately and keep cold | Bacteria multiply or die off at ambient temperature, changing the result in either direction |
| Analyze promptly | Holding time is short — bacteriological samples must be processed as quickly as practical, within the holding time the approved method and the permit specify |
| Grab sample only | Bacteriological samples are never composited, because organisms would grow or die during the compositing period |
| Sample at the permitted monitoring point | Usually after disinfection and, where dechlorination is practiced, at the location the permit designates |
5. Geometric Mean Reporting
Bacterial counts range over orders of magnitude, so an arithmetic average is dominated by a single high value. Permits therefore express bacteriological limits as a geometric mean.
Method A — the root of the product. Multiply the n results together and take the nth root.
Method B — logarithms (easier with a calculator). Take the base-10 logarithm of each result, average the logarithms, and take the antilogarithm of that average.
Worked example. Seven weekly fecal coliform results: 10, 40, 100, 200, 800, 1,000, and 5,000 colonies per 100 mL.
- Arithmetic mean = (10 + 40 + 100 + 200 + 800 + 1,000 + 5,000) / 7 = 7,150 / 7 = 1,021 per 100 mL — a value larger than five of the seven results, pulled up almost entirely by the single 5,000.
- Geometric mean: logs are 1.000, 1.602, 2.000, 2.301, 2.903, 3.000, and 3.699. Their sum is 16.505 and the average is 2.358. The antilog of 2.358 is approximately 228 per 100 mL.
The geometric mean of 228 represents the central tendency of that data set far better than 1,021 does, which is exactly why permits use it.
Practical points. A result reported as zero cannot be used directly in a logarithm, so approved conventions substitute a value such as 1 for a non-detect. And because the geometric mean is resistant to a single high value, a single excursion rarely violates a geometric mean limit by itself — but permits usually pair the geometric mean limit with a separate maximum single-sample limit that a single excursion violates immediately.
An analyst filters 20 mL of final effluent through a 0.45 micron membrane, incubates it on mFC medium at 44.5 degrees C, and counts 45 blue colonies. What is the reported fecal coliform result?
A bacteriological sample bottle is prepared sterile but the sodium thiosulfate is inadvertently omitted, and the sample is collected downstream of chlorination. How will the reported result be affected?
Why are bacteriological effluent limits expressed as a geometric mean rather than an arithmetic average?