An important part of the study of plant communities is the analysis, by statistical methods, of the distribution of each species within the community. A complete census of all the individual plants for each species would be impracticable because the area occupied by a plant community is usually large. The community may be sampled by selecting at random small sample areas called quadrats. These quadrats are marked out by placing on the ground a suitably sized frame, which may be rectangular or square in shape. The size of the quadrat to be used for sampling a specific community is decided subjectively by the ecologist. The techniques employed to obtain random sampling vary. Frequently a line is chosen within the community, and quadrats are selected at regular intervals along this line and along lines marked out parallel with it. Or, sometimes, a skewer may be thrown into the ground some distance away from the observer, and the quadrat frame is then located with reference to it in a predetermined manner. Or, as a final example, relatively large plots forming a systematic pattern over the whole community may be marked out, and quadrats are selected at random within these plots. In practise randomly located sets of contiguous quadrats are widely used instead of single quadrats. The method of sampling has also been used in the study of insect populations. The sampling unit is usually a unit of vegetation, such as a potato plant, rather than a unit of area, although this distinction is mainly a change of emphasis in the case of crops that have been planted in a regular pattern. The results are very similar to those obtained for plant populations as far as divergence from the Poisson form of distribution is concerned, and it has been suggested that insects may also be contagiously distributed.