Density dependence tests, are they?

摘要: A large number of time series abundances insects and birds from a variety data sets were submitted to new density dependence test. The results varied enormously between sets, but the relation frequency statistically significant (SSDD) length was similar that power curve test, making consistent with hypothesis density-dependent model being universally applicable throughout used. Pest non-pest species did not differ in incidence SSDD. more sampling error present data, higher percentages This expected given test increases increasing data. Many used here, as well literature, clearly violate basic assumption organism concerned should be univoltine semelparous. Yet SSDD same bi/polyvoltine semelparous organisms are reproductively active than one year. seasonal migrant Autographa gamma Britain Czechoslovakia even rainfall found have Statistical significance, however, does automatically lead conclusion regulation. Any random variables which stochastic equilibrium, such independent, identically distributed, variables, is typically described better by alternative (density-dependent) null (density-independent) model. Significant often obtained sloppy assumptions other where an interpretation terms densitydependent regulation absurd. Given fact explanations for all these cases, mechanisms may very too entirely appropriate simply data-based choice without equilibrium. contain any information about causes fluctuation pattern, so cannot expect statistics produce series. result using suitable regulation, also hypotheses. Because generally universal applicability density-dependence model, negative only mean long enough become significant. Positive difficult interpret, results. final decision needs based much on detailed population concerned. "density-dependence test" presence mechanism because loaded, multiple meanings term "density-dependence", calling "test statistical dependence" preferable.