In the fluctuation tests (Fig. III.7), we also determine whether the mutation providing resistance against a certain toxic agent occurred before addition of the selection agent or after its addition. The experiment can be performed by growing a larger innoculum of genetically identical bacteria from one cell, adding this innoculum in the same amount to a series of test tubes with fresh nutrient medium and leaving the bacteria to further multiply here, e.g., for 24 hours. Then a sample of bacteria from each test tube is seeded on one Petri dish containing an antibiotic to which all the bacteria were sensitive at the beginning of the experiment. After a certain time, e.g., after two days, we count the colonies of resistant bacteria on the individual dishes (unmutated bacteria, i.e. those sensitive to the antibiotic, do not grow on the dishes). If the mutations occurred only as a reaction to the presence of the antibiotic, then there should be approximately the same number of colonies on all the dishes or, to be more exact, the number of colonies should have Poisson distribution (Fig. III.8). If the mutations occurred spontaneously, i.e. before the bacteria came into contact with the antibiotic, the numbers of colonies on the individual dishes should differ substantially, i.e. should substantially fluctuate. This is a result of the fact that a mutation can occur in the test tube prior to seeding on the dish at any moment; the mutant could multiply exponentially in the particular test tube up to the moment of seeding on the dish. If approximately the same number of mutations occurred in each test tube, the resultant numbers of colonies in the dishes would differ according to when the mutations occurred. If the particular mutation occurred just before seeding on the dish, then a single colony is obtained; however, if the same mutation occurred 3 hours prior to seeding then, for a generation time of the bacteria of, e.g., 30 minutes, the particular mutation can appear as the presence of up to 64 colonies on the dish. As the number of colonies on the dishes actually differed substantially in the laboratory experiments, fluctuation tests were considered until recently to be the strongest proof that mutations always occur randomly, spontaneously, and are not environmentally directed, i.e. induced by the presence of an antibiotic.