Katie Stanislaw


Research Experiment

Fecal Coliform Bacteria in Lake Michigan vs. Pewaukee Lake


Fecal coliform bacteria are a "group of bacteria that are aerobic and facultatively anaerobic, gram-negative, nonsporeforming, rod shaped bacteria that ferment lactose with gas and acid production" (Toranzos, 1991). The "bacteria in this coliform subgroup has been found to have an excellent positive correlation with fecal contamination from warm blooded animals" (Pourcher, 1991). The presence of fecal coliform bacteria in water is considered significant for public health (Toranzos, 1991). This is so significant that there are standards for drinking, bathing, and recreational waters, to provide optimal protection to the public (Toranzos, 1991). These standards are set so that the water does not lead to the cause of any adverse health consequences (Toranzos, 1991). The U.S. Environmental Protection Agency has set criterion for recreational water containing total coliforms. This standard is 200 total coliforms per 100 ml of water (Toranzos, 1991). When looking at the standard for recreational water containing fecal coliform bacteria there have not been any standards set (Toranzos, 1991). Based on this information samples from one of the Great Lakes and a small inland lake were compared for fecal coliform contamination. The decision to conduct this experiment is based on the hypothesis that there will be a greater amount of fecal coliform bacteria in Lake Michigan than in Pewaukee Lake.


First sixteen vials were obtained to use for sample collection. These vials carried 20 ml of water and had plastic lids. Then on October 19, 1998 eight samples of water were taken from various places along a one-mile stretch of shore on Lake Michigan (see Table 1). Next on October 25, 1998 seven more sample of water were collected from various places along a of a mile stretch of shore on Pewaukee Lake (see Table 2). These samples were not randomly chosen, they were spots where there was local access to the water. To collect the samples the vials were submerged into the water and when they were full they were capped and labeled. The samples were then used to run a presumptive test that would show if the water contained any lactose fermenting or acid-producing bacteria.

The presumptive test was done by using phenol lactose broth, which was prepared by the microbiology laboratory assistants. For each water sample nine tubes of phenol lactose broth were used. The first three tubes contained double strength lactose broth and to these tubes 10 ml of water sample was added. Then the next sets of three tubes were single strength lactose broth and 1 ml of water sample was added to these. Finally the last three tubes were single strength lactose broths and .1 ml of water sample was added. Next a set of nine tubes were used as control tubes. The tubes were filled in the exact same way as with the water samples, but distilled water was used instead of the lake water samples. In order for there to be enough water samples for each test the samples were all diluted to 50 ml with distilled water. Once all of the samples were pipetted into the phenol lactose broth the tubes were put into an incubator at 35 degrees Celsius for twenty-four hours. After the twenty-four hours the tubes were taken out of the incubator and the color change amount of gas produced were recorded (see Tables 3 and 4).

Once this information was recorded the number of tubes for each samples that produced 10% or more gas was compiled. From this information the most probable number (MPN) of bacteria per 100 ml of water could be found (see Tables 5 and 6). The "MPN analysis is a statistical method based on random dispersion of microorganisms in a given sample" (Toranzos, 1991). Once this data was collected a T-test was run to see if the null hypothesis was correct.


The results of the Most Probable Number test can be found in Tables 5 and 6. To find the MPN for each sample the tubes were looked at to determine how much gas was produced (see tables 3 and 4). Then using the table, MPN Determination from Multiple Tube Test the most probable number of bacteria could be found (Benson, 1985). Once this was done each of the samples were looked at to see their color change, if any (see tables 3 and 4). This color change is representative of acid producing bacteria. Finally a T-test was done to see if the null hypothesis could be confirmed or rejected.


Lake Michigan (y1)









Pewaukee Lake (y2)









df = 8 + 7 2

df = 13

Lake Michigan Pewaukee Lake

S y1 = 30 S y2 = 566

y1 = 30/8 y2 = 566/7

y1 = 3.75 y2 = 80.857

S y12 = 578 S y22 = 117052

s2 y1 = 1/ 8-1* (578-302/8) s2 y2 = 1/7-1 * (117052 80.8572/7)

s2 y1 = .14286* (465.5) s2 y2 = .1667 * (116118.02)

s2 y1 = 66.5 s2 y2 = 19353.003

Sp = 66.5 (8-1) + 19353.003 (7-1)


Sp = 465.5 + 116118.02


Sp = 8967.96

Sp = 94.699

t = 3.75 80.857

    1. 1/8 + 1/7

t = -77.107


t = -1.57

After getting the value of 1.57 for t, the value was looked up on the table of percentage points of the t distribution. Then from looking at the table the two a-values could be found. These values were .05 and .1. Therefore the probability is .05< P< .1.


When looking at the data to determine fecal coliform presence in Lake Michigan and Pewaukee Lake it can be seen that both of these lakes contain fecal coliform bacteria. First looking at the most probable number results it can be seen that Lake Michigan had two areas where there seemed to be fecal coliform contamination. In area 5, which was along Bradford Beach, there were approximately 23 bacteria per 100 ml of water. When looking in area 7, which was also along Bradford Beach the most probable number of bacteria in 100 ml of water was 7. These two areas both contained fecal coliform bacteria, but area 5 had a significantly larger amount. Next by looking at the color changes in the tubes the same two areas had color changes from red to yellow. This shows that acid is being produced and there are bacteria present. These areas are along a beach that is open to the public for swimming in the public months and there also seems to be a lot of sea gulls around this area.

When looking at the most probable number test for Pewaukee Lake it can be seen that there is a significant difference between this lake and Lake Michigan. In six out of the seven samples gas was produced which is a positive sign that fecal coliform bacteria is present. The six samples that did come out positive were in the low twenties, with two of them being over 200. Next looking at the color changes of the phenol lactose broth almost all of the tubes changed to yellow, which shows that acid was being produced and bacteria were present. So even though all of the tubes may not have produced gas it was shown by the color changes that there was bacteria present. These samples were all taken along a public beach that was close to some apartment buildings and there were many ducks and geese around. Next when looking at the T-test data the fact that the probability is greater than .05 shows that this data is not significant, and the null hypothesis can be rejected.

When comparing Lake Michigan to Pewaukee Lake it can be seen that Pewaukee Lake has more fecal coliform bacteria present than Lake Michigan. This would therefore refute the hypothesis that Lake Michigan has a greater amount of fecal coliform bacteria than Pewaukee Lake. One thing that needs to be remembered is that each of the samples were diluted to 50 ml so they were not full strength. A reason that the hypothesis may be wrong is because Lake Michigan may have more strict rules when it comes to fecal coliform bacteria because it is a lake that supplies the cities with drinking water. Another theory may be that the town of Pewaukee has septic systems that may empty into this lake would cause the fecal coliform bacteria to be higher. Looking at the locations of the samples, it can be seen that the Pewaukee Lake samples were taken from by an apartment building and there may be a septic tank from there that enters into the lake. Since the United States does not have a standard set concerning fecal coliform bacteria and recreational water then this lake does not have to worry that they have levels of over 200 in some areas (Toranzos, 1991).

If this experiment were to be repeated there are some things that should be changed. First the amount of water collected should at least be 40 ml. This is because in order to do the MPN test 34 ml of water sample is needed. If more water can be collected then the samples do not have to be diluted and the samples may give a more accurate reading. Another thing that needs to be done is after the twenty-four hours of incubation for the phenol lactose broth tubes if there are any positive results these need to be streaked on agar to confirm the positive result. The last thing that should be changed is that there should be more samples. This may lead to better statistical data in the future.














Toranzos, G.A. and McFeters, G.A. 1996. Water Microbiology in Public Health. Manual of Environmental Microbiology: 184-186.

Pourcher, A.M., L.A. Devriese, J.F. Hernandez, and J.M. Delattre. 1991. Enumeration by a miniaturized method of E. coli, S. bovis, and enterococci as indicators of the origin of fecal pollution of waters. Journal of Applied Bacteriology. 70: 525-530.

Benson, H.J. 1991. Bacteriological Examination of Water. Microbiological Applications: 170-172, 332.