7258

• UCREFRP Catalog Number: 7258
• Author: Rosenfeld, M. J.
• Title: Morphometry in Chubs of the Genus
• USFW Year: 1986.
• Copyright Material: NO

~

chubs, morphological study might best be done with the inte~t of

finding natural groups, not by predesignating them (i .e., fishes

not ~ priori 2ssigned to species).

In this way, relationships

may be objectively discovered, for corroboration by other method-

ologies (e.g., mitochondrial DNA, cytogenetics, electrophoresis,

etc.). The choice of morphometric characters should entail an

even sampling of the body (= box truss; Bookstein et al., 1985)

and, perhaps, characters diagnostic for chub species elsewhere,

to maximize the liklihood of finding relevant differences. A

.~~
large number of measurerrents can be rapidly taken with a digi-

tizer from photographs of live fish, allowing analyses without

sacrifice.

One multivariate comparison~l technique, principal components

analysis (PCA), is well suited for exploration of interpopula-

tional differences.

The method requires no prior categorization

(i .e., population, species, etc.) for a specimen, since the

algorithm sorts individuals.

peA ge~erates a set of 60mposite

characters arranged in order of decreasing variance <= diagnostic

power

Each composite character is made up from the origi~al

measurements (e.g., head length, snout length, standard length,

et c. ) .

Consequently, porulations will not only be better

compared by simultaneous consideration of numerous morphological

characters but also with the discernation of diagnostic ones. A

review of PCA can be found in Dunn and Everitt (1982).

3