The enumeration of bacteria by the microscopic method

The method described by Breed (1911) for estimating the number of bacteria in milk has been widely applied, including the counting of bacteria in other fluids. Despite most careful technic there are two important sources of error: losses of bacteria from the slide during staining and faulty selection of areas on the film for microscopic observations. Since "slide loss " is often influenced by fixatives involved in the suspending fluids them-selves and by the complexity of the staining process, it must be admitted that absolute values are seldom attainable. However, reliable, reproducible estimates in successive samples of the same material are more readily assured if appropriate amounts of an effective fixative are used and if the method of sampling smears is improved. The two sampling methods which appear to have been used most widely are: random movements of the objective over the smears or, possibly, orderly movements along parallel lines so that the sampled areas are arranged in "checkerboard " fashion. These methods of sampling are compatible with the concept of films which are square, "flat " and of haphazard numerical density throughout. This view does not consider the possibility that distribution may be influenced consistently by the surface tension of the suspending fluid (convexity of the drop before drying), migration of fluid and bacteria during drying, or the improbability that slides will be dried in a horizontal position. The data presented deal with: (a) preparation of smears shaped in accordance with the natural distribution of bacteri