Collecting a water sample is simple. Just use a sterile plastic container and immerse it the water. All you need is a couple of milliliters. The most important thing is to try and avoid collecting particulate: you want the water to be as pure as possible. Be sure to label your samples with the location where the water was collected.

We use pre-prepared Luria broth with agar to make plates. Making the plates is easy. Dissolve 38 grams of the Luria broth media to make 1000 milliliters of media. Autoclave. You will make two types of plates: one without antibiotics and one with antibiotics. For the plates with antibiotics, pour 500 milliliters of the liquid media into a sterile flask. When the liquid media is about 55°C, add 43 milligrams of antibiotic to the media so that the final concentration of the antibiotic is 86 µg/ml. Make sure the antibiotic is completely dissolved and the media is well mixed. Gently pour the solution onto the petri dish so that it covers the bottom to a depth of about 2-4 mm. Make sure there are no bubbles in the media. Place the top of the petri on the plate so that there is room for the hot air to exit. After the plates have cooled, cover the petri dish with the cover plate. For the plates without antibiotics, just gently pour the plates. Store the plates upside down in plastic in the refrigerator. Warm to room temperature before use.

Ideally, for each sample, you have at least two different antibiotic plates (e.g. ampicillin, streptomycin) and a plate without antibiotics. We have discovered that it is important to plate enough water to get somewhere between about 10 to 300 colonies. It is impossible to know how many colonies will grow on the plate for a given amount of water. Typically, you have to do a dilution series in which you plate increasing amount of water onto different plates, ranging from 25µl to 500 µl. We have found that relatively pristine sites require about 250 µl of water whereas sites expected to have a lot of bacteria—near municipalities or agricultural locations—require only 50 to 100 µl of water.

To plate the sample, use a pipette and a sterile tip. Squirt the liquid in the center of the plate and use a plate spreader to gently spread the fluid evenly across the plate. In between spreads, soak the spreader in ethanol and burn away the liquid to sterilize before re-submerging in the sample water. Let the fluid soak in for a couple of minutes and then invert the plate and store at an appropriate temperature (room temperature is typically good).

One way to estimate the number of bacteria is to choose a central section of the plate and count all colonies that fall mostly (75%) within a specified box. A second strategy is to overlay a grid onto the plate and count colonies that fall on a line.

The same method for counting colonies should be used for both the plate with and without antibiotics. The estimated frequency of antibiotic resistant bacteria in the sample is the number of colonies on the antibiotic + plate divided by the number of colonies on the plate without antibiotics (antibiotic -). Try to remain unbiased when selecting zones - the count should be as random as possible.

For each location, you need latitude and longitude, a good picture of each plate with a label identifying the plate, and the estimated counts of bacteria for each plate. Ideally, a photograph of the sample location and the surrounding environment is also included.

Be extremely sure to spread the samples evenly across the plates. Uneven distribution will cause errors in your count.

Sterilizing in between spreads is very important: you don't want bacteria from one site appearing on the wrong petri dish.