The logarithmic series and the lognormal distribution applied to benthic infauna from Puget Sound, Washington, U.S.A.
Abstract
1. Benthic infauna collected at eight stations in Puget Sound during
1963-1964 was tested for agreement with the logarithmic series and
the lognormal distribution.
2. The data agreed only moderately well with the logarithmic series;
seven of sixteen cases showed statistically significant differences
between observed frequencies and frequencies expected from the
logarithmic series. Furthermore, there was an inverse relationship
between goodness of fit to the logarithmic series and the level of
numerical dominance of the faunal assemblages. This relationship,
and the fact that the underlying assumptions for the logarithmic series
are not applicable in benthic communities, makes the series a doubtful
model for the relationship between species and specimens in a
benthic community.
3. The lognormal distribution fitted the data reasonably well, but the
curves were always strongly truncated. Increasing the sample size
reduced the percentage of species in the first class somewhat, but only
when the data from all the seasons and all the stations were combined
was a significant part of the distribution "unveiled". The lognormal
distribution therefore represents the geographical region reasonably
well, but for each station the curves are too truncated to represent
better models than the logarithmic series.
Publisher
[Fiskeridirektoratets havforskningsinstitutt]Series
Fiskeridirektoratets skrifter, Serie Havundersøkelservol 15 no 3