|Title||Quantitative estimations of plankton from small samples of Sedgewick-Rafter-cell mounts of concentrate samples|
|Publication Type||Journal Article|
|Year of Publication||1949|
|Journal||Transactions of the American Microscopical Society|
1. The distribution of plankton in a Sedgewick-Rafter cell filled by standard methods is not uniform. The degree of inequality of distribution varies with the species, the concentration of plankton, and the filling of the counting chamber. 2. The distribution of plankton in the counting chamber may differ from the expected distribution of a Poisson function by varying amounts. The departure varies with the species, the concentration of plankton, and also from one preparation to another. Several species, counted as one, show a greater departure than anyone of the species studied individually. 3. The distribution of sampling units, each formed of several random fields, approaches the distribution of the Poisson function as the number of fields included in the sampling unit is increased. Ten to twenty fields must be included in the unit before approach to the Poisson distribution becomes satisfactory for some species and also for mixed species. This number is too large to be of practical value in reducing variance in counts in the Sedgewick-Rafter cell, but it can be of value in other biological and limnological problems. 4. Experimental studies were made of the limitations of Fisher's "t", and of Lord's "u" test, which employs the range as an approximation to the standard deviation, in testing significance of differences between means of small random samples drawn from Poisson distribution and from the observed Sedgewick-Rafter-cell distribution. When the mean of the distribution was 5.0, or larger, conformity with theoretical values was obtained for absolute values of deviations. 5. A criterion is introduced by use of which the unit counting area may be increased for a given counting cell preparation until reasonable certainty is attained that the mean of the unit counting areas will be at least 5.0.