Chapter 3: Null Expectations in Subspecies Diagnosis

• Published in: Ornithological monographs Vol. 67; no. 1; pp. 35 - 41
• Main Author: Patten, Michael A.
• Format: Journal Article
• Language: English
• Published: Washington The American Ornithologists' Union 01.01.2010; American Ornithological Society
• Subjects: Allopatric species; Animal genetics; Biodiversity; Biological taxonomies; Evolutionary genetics; Mitochondrial DNA; Monophyly; Museum exhibits; Phylogenetics; Population genetics; Species; Taxa
• ISSN: 0078-6594; 1941-2282
• DOI: 10.1525/om.2010.67.1.35

The utility of subspecies in studies of evolution and migration and in conservation planning has been debated hotly for a half-century. Inconsistent and sometimes sloppy application of the subspecies concept has led some to deem it a failure, but recent quantitative definitions of subspecies have put the concept on more rigorous footing. Nonetheless, the molecular revolution has added fuel to the fire as researchers attempt to test subspecies by genetic means. Until a sound and defensible null expectation is developed for genetic differentiation of subspecies, genetic approaches will be fraught with problems. A test for monophyly is insufficient, because parapatric subspecies interbreed by definition. Moreover, because much geographic variation may arise via natural selection, tests restricted to selectively neutral genetic data are likewise problematic. Moreover, long-standing charges of subjectivity in the naming and diagnosis of subspecies must be addressed if subspecies are to continue to be accepted as valid taxonomic entities. Statistical advances, including pairwise tests, spline regression, module identification in neural networks, Monmonier's algorithm, and unsupervised, fuzzyk-means cluster analysis offer considerable promise as means of identifying and quantifying geographic variation in an objective yet statistically rigorous manner.