The earliest studies of the soil bacteria consisted of such estimations of numbers, and showed that the soil contained a very numerous population of bacteria. About 20,000,000 bacteria per gram of soil is now considered a fair average number. The number and variety of bacteria existing in the soil is so enormous that the method of separating out all the different forms, and of discovering their characters and functions, has proved impracticable. In practice, therefore, the problem has been approached from the biochemical standpoint. That is to say, the special chemical changes that the bacteria produce in the soil have first been investigated, and this has been followed by the isolation and study of the various groups of bacteria that bring about the changes under investigation. The method commonly employed in isolating the organisms that produce a given chemical change in the soil is called the “elective” method. The soil is inoculated into a culture medium that will especially favour the group of bacteria to be isolated, to the exclusion of others. For example, if it is desired to isolate the organisms that attack cellulose, a medium is made up containing no other organic carbon compounds except cellulose. Such a selective medium encourages the growth of the group of organisms to be investigated, so that after several transfers to fresh medium a culture is obtained containing only two or three different types of organisms. These are separated by plating and pure cultures obtained. Another difficulty which has not yet been completely overcome is that of adequately describing an organism when it is isolated. The morphology of bacteria is not the constant thing that is seen in the more stable higher organisms. In many cases the appearance of a single strain is entirely different on different media, and may be quite altered by such conditions as changes in acidity of the medium or temperature of incubation. Even on a single medium remarkable changes in morphology occur, at any rate, in some bacteria. This is well seen in a cresol-decomposing organism under investigation at Rothamsted. In cultures a few days old this organism develops as bent and branching rods; these rods then break up into chains of cocci and short rods, which separate, and in old cultures all the organisms may be in the coccoid form (Fig. 1). It is claimed by Löhnis[47_b_] that the possession of a complex life-cycle of changing forms is a universal character in the bacteria. The instability of shape in many bacteria makes it necessary to standardise very carefully the cultural conditions under which they are kept when their appearance is described. [Illustration: Culture 15 hours old. Culture 3 days old. FIG. 1.--Change in appearance, in culture, of a cresol decomposing bacterium.]