The methods for estimating the population density of organisms presented in earlier chapters are applicable when the sampling unit is well defined. Examples of these sampling units include a plant, a leaf, a fruit, a nest, and a quadrat. Brock (1954) found it impractical to use these methods when studying reef fish populations because the terrain was irregular (see Seber 1982, p. 28). Further, the effective use of well-defined sampling units requires the organisms of interest to remain relatively immobile. For example, if a plant is the sampling unit and an insect the species of interest, then it is assumed that all insects are observed, including those that leave the plant due to the approach of the sampler. With more mobile animals, this assumption is clearly violated, leading to severe underestimates of population density. Deer, fish, birds, rabbits, flies, and rodents are but a few of the animal species whose mobility require alternative sampling methods. Difficulties also arise if the population is sparse. Either very large sampling units or a large number of them is required to estimate population density with an acceptable level of precision.