Use this URL to cite or link to this record in EThOS:
Title: Degrees of visible polymorphism and new methods for measuring frequency-dependent selection by predators
Author: Weale, Michael Edmund
ISNI:       0000 0001 3564 7603
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 1992
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Effective conservation management will benefit from a better understanding of how biological diversity is maintained in nature. Currently, our knowledge is still elementary. Within a single species, colour-pattern diversity can range from almost total uniformity to complex and extensive `massive polymorphisms', where even large population samples may not yield two like individuals. Frequency-dependent selection by predators, in which predators consume disproportionately more of commoner prey types, has been proposed as one mechanism that may maintain both simple and massive polymorphisms, though its general efficacy is still unknown. A simple model is presented here that affirms the ability of this mechanism to generate as well as maintain massive polymorphisms. Previous work to test for frequency-dependent selection by predators has virtually all been carried out using simple dimorphic prey populations. New conceptual and analytical techniques are developed here that allow hypotheses to be tested about the presence or absence of selective predation; the type of frequency-dependent selection that may be acting in populations with more than two prey types; its strength and constancy as numbers of prey types change; and its efficacy in cases of massive polymorphism. Specifically, rigorous model-fitting methods are given for avoiding bias in estimated selectivity indices; a family of descriptive frequency-dependent selection models (derived both from Manly's linear model and Elton & Greenwood's exponential model) is developed, together with details of how they can be fitted to prey-choice data by maximum likelihood or least squares techniques; and a 'random morph' method is developed that allows frequency-independent predator preferences to be averaged out and thus allows frequency-dependent selection to be tested even in massively polymorphic populations. The new methods, models and techniques are tested both with computer simulations and with experimental data collected from wild birds feeding on artificially-coloured pastry baits and from human 'predators' selecting `prey' objects on a computer screen.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Statistics