9336

• UCREFRP Catalog Number: 9336
• Author: Douglas, M. E., M. R. Douglas, J. M. Lynch and D. M. McElroy
• Title: Use of Geometric Morphometrics to Differentiate
• USFW Year: 2001
• Copyright Material: YES

393

DOUGLAS ET AL.-GlLA GEOMETRIC MORPHOMETRICS

CV2

TABLE 2. CLASSIFICATION RESULTS FOR Gila cypha
FROM A CANONICAL ANAL'tSIS OF SHAPE COORDINATES.
Numbers in parentheses are classification results from t[)'C\
truss analysis of McElroy and Douglas (1995). Key to V::.:::I
sample codes is given in Table 1.

DC

LC

ww

YR

BR

CC

BR
CC
DC
LC
WW
YR

24 (24) 0 (0) 0 (0) 0 (0) 1 (1) 0 (0)
0(0) 11 (11) 0 (0) 0 (0) 0 (0) 0 (0)
o (0) 0 (0) 22 (22) 0 (0) 0 (0) 0 (0)
o (0) 0 (0) 0 (0) 28 (28) 0 (0) 0 (0)
3 (1) 0 (0) 0 (0) 0 (1) 54 (55) 0 (0)
o (0) 0 (0) 0 (0) 0 (0) 0 (0) 5 (5)

97.3% of the specimens being correctly allocat-
ed to their collection site (Table 2). This value
is significantly different from that obtained
from 100 randomized canonical variate analyses
(mean of 65.5%; u = 3.50; P < 0.01). Speci-
mens from Westwater and Black Rocks, the two
sites with smallest D2 values, also group closest
to one another in canonical space (Fig. 2).

Body size and shape variation in Gila TOUusta.-
Roundtail chub from Rifle and Debeque can-
yons were significantly larger than those from
the other five populations but were indistin-
guishable from one another (Student-Neuman-
Keuls test; F6.20B = 9.34; P < 0.0001). As with G.
CYPha, all shape coordinate pairs differed signif-
icantly among populations (all P < 0.0001), and
a significant MANOVA resulted (Wilks' A =
0.002; X2288 = 1154.9; P < 0.0001). Here, vari-
ation in peduncle and head characteristics is
emphasized. Specimens from Cataract Canyon
(Appendix 2B) have a relatively shorter pedun-
cles that (again) taper in depth from anterior
to posterior, whereas those from Desolation
Canyon (Appendix 2C) are noticeably longer
and of uniform depth. Debeque (Appendix 2F)
and Rifle (not shown) Canyons have relatively
smaller heads. Indeed, the latter two popula-
tions are characterized by having all landmarks
of the head (with the exception of LMs 11, 13
and 25; Fig. 1) placed relatively more anteriorly
when compared to other populations. Addition-
al aspects of shape variation are found in rela-
tionships of body depth to fin placement. For
example, Debequeand Rifle populations have
deepest bodies yet shortest distances between
pectoral and pelvic fins, whereas Westwater Can-
yon specimens seemingly have shallowest body
depths. The development of a nuchal hump is
slight in all G. TOUusta populations.
Three (of six) statistically significant canoni-
cal variates accounted for 85.3% of among-
group variation (Fig. 3). As with G. CYPha, all

8

2

@

@)

2

CV1

e@ @

Fig. 3. Shape variation among seven Gila robusta
populations based on the first two variates derived
from a canonical analysis of shape coordinates. Pop-
ulation abbreviations follow Table 1. Circles represent
95% confidence intervals for the group mean.

Mahalanobis distances (and therefore all
arnong-group comparisons) were significant (P
< 0.0001). Mahalanobis D2 values varied greatly.
The lowest [i.e., 11.9 sdu (= standard deviation
units)] occurred between Westwater and Black
Rocks specimens, which were closest together in
CV-space (Fig. 3). The highest value (76.5 sdu)
occurred between Cataract and Rifle Canyons.
Populations from Cataract and Desolation Can-
yons were well removed from the other five,
with 96.3% of all specimens correctly allocated
to site of origin (Table 3). This value is signifi-
cantly different from that obtained from 100
randomized analyses (54.9%; u = 3.42; P <
0.01).

Body size and shape variation among all samples,-
As might be expected from the above, body size
(as measured by centroid size), varied among
all 13 samples. Two groups were detected (F12,350
= 7.94; P< 0.0001): One consisted of G. TObusta
from Rifle and Debeque Canyons and G. CYPha
from the Little Colorado; the second consisted
of all other populations.
Relative warp analysis clearly separated pop-
ulations of G. cypha from those ofG. rOUusta,
except for Desolation and Caratact Canyons
(Fig. 4). Body shapes of G. CYPha at the latter
two locations were more rOUusta-like. Of all cy-
Pha-like forms, those from the Little Colorado
River were most distinct, paralleling the results
of the CVA of shape coordinates (Fig. 2).
Operational Taxonomic Units with low values
on the first relative warp (RWl) clearly have a
cypha-like morphology (i.e., deeper body with
pronounced nuchal hump; smaller head; lon-
ger, narrower peduncle; Fig. 5A), whereas those

!:. -

i