Journal of Fish Biology (2005) 67, 849–865

doi:10.1111/j.1095-8649.2005.00791.x, available online at http://www.blackwell-synergy.com

Geographic variation in Austrolebias luteoflamulatus

Vaz-Ferreira, Sierra & Scaglia (Cyprinodontiformes,
Rivulidae)
A. D’A N A T R O * AND M. L O U R E I R O †‡
*Laboratorio de Evolución, Facultad de Ciencias, Iguá 4225, Montevideo 11400,
Uruguay and †Sección Vertebrados, Facultad de Ciencias, Iguá 4225, Montevideo
11400, Uruguay

(Received 18 June 2003, Accepted 31 March 2005)

The morphological variation in the annual killifish Austrolebias luteoflamulatus was studied
using meristic and pigmentation characters analysed with univariate and multivariate analysis,
and morphometric characters analysed with a geometric morphometry approach (thin plate
spline). The results showed that morphological variation in this species was associated with a
geographical pattern. In the case of meristic and pigmentation characters, the strongest
distinction occured in males, while females were more similar among different drainages.
Males from Rocha (eastern Uruguay) exhibited the greatest differences from the others. The
number of dorsal fin rays and bars on the flanks were the characters that contributed most to
this segregation. When morphometrics were considered, the three populations analysed could
be distinguished well from each other in both sexes. Some of the characters that contributed to
this pattern were dorsal fin origin, position of eyes and length of snout. Since pigmentation
characteristics and some characters of the dorsal fin in males were involved in the differences
observed, sexual selection could have played an important role in the differentiation of popula-
tions of this species. Since Austrolebias luteoflamulatus has a restricted geographical distribu-
tion, conservation of this species should include consideration of these populations individually
since they might represent different evolutionary units. # 2005 The Fisheries Society of the British Isles
Key words: annual fish; geographic variation; morphology; Neotropics.

INTRODUCTION
The analysis of intraspecific variation is fundamental to the study of evolution-
ary processes since this variation is acknowledged to be the raw material of
evolutionary change (Schaefer & Cavender, 1986). Patterns of phenotypic varia-
tion are often interpreted within an historical framework in order to elucidate
the causal factors and processes involved (Bailey & Smith, 1981; Thorpe, 1984).
For this reason, the study of morphological variability is of great importance
when inferences about the evolutionary history of a species are proposed.

‡Author to whom correspondence should be addressed. Tel.: þ5982 5258618; fax: þ5982 5258617;
email: [email protected]

849
# 2005 The Fisheries Society of the British Isles
850 A. D’ANATRO AND M. LOUREIRO

The order Cyprinodontiformes is among the most diverse taxa of the

Neotropical fish fauna (Vari & Malabarba, 1998). Within it, the family
Rivulidae is comprised of nearly 30 genera (Costa, 1998) including Austrolebias
Costa, which possesses a characteristic annual life cycle with diapausing embryos
developing inside drought resistant eggs (Wourms, 1972). Austrolebias luteofla-
mulatus Vaz-Ferreira, Sierra & Scaglia, together with at least six other
Austrolebias species, is an annual fish endemic to the eastern wetlands of
Uruguay. These species are distributed along the lowlands associated with
Atlantic coastal systems and the drainage system of the Laguna Merı́n (Garcı́a
et al., 2000). High morphological, chromosomal, and genetic variability seems to
be characteristic of this and other species of the genus restricted to that area
(Vaz-Ferreira & Melgarejo, 1984; Garcı́a et al., 1993, 1995, 2000; Reichert, 1994;
Wildekamp, 1995; Loureiro, 1999). This variability, however, has not previously
been well described for A. luteoflamulatus.
The natural habitats of these fish are temporary ponds that can be isolated
from each other fairly easily, thus interruption of gene flow among populations
may occur quite often. In addition, annual fishes have short generation times
(<1 year in the wild) and high metabolic rates (Cardozo, 1999) in order to attain
adulthood before their ponds dry out. Strong sexual dimorphism in shape and
pigmentation of this and other annual fishes suggest that sexual selection plays
an important role in the evolution of these taxa. In many localities within the
eastern wetlands of Uruguay as many as four annual fish species (Cynopoecilus
melanotaenia Regan, Megalebias cheradophilus Vaz-Ferreira, Sierra & Scaglia,
A. luteoflamulatus and Austrolebias viarius Vaz-Ferreira, Sierra & Scaglia) can be
found inhabiting the same ponds. Thus, ecological interactions, such as competi-
tion and predation, are also other possible mechanisms responsible for creating
variability and diversity in this system. These conditions make annual fishes a
good vertebrate model to study intra-specific differentiation and speciation.
Considering the low vagility of annual fishes, the hypothesis that was tested in
this work was that morphological variation is structured according to the
geography of drainage systems. The objective was to characterize the morpho-
logical variation of A. luteoflamulatus within this geographical context.

MATERIALS AND METHODS

Specimens used in this study are deposited in the Vertebrate Collection of the Facultad
de Ciencias, Universidad de la República, Uruguay (Institutional code ZVC-P). A
complete list of specimens and their localities is given in the Appendix. Fish were
clustered in three groups which correspond to the following drainages: Rocha, Castillos
and Merı́n Lagoons (Fig. 1). Due to a pronounced sexual dimorphism, data of males and
females were analysed separately.
Meristic variables analysed and their correspondent abbreviations are: dorsal fin rays (D),
anal fin rays (A), caudal fin rays (C), pectoral fin rays (PC), pelvic fin rays (PV), lateral series
of scales (LS), scales around the caudal peduncle (PS), predorsal series of scales (PRS),
supraorbital scales (SS) and number of neuromasts in the supraorbital series (SN).
The pigmentation pattern analysed in males was the number of dark black bars (B). Females
present a highly variable pigmentation pattern (FP). For analysis purposes this character was
divided in three different categories: 1) dots (d), 2) dots and bars (db) and 3) bars (b).
Meristic variables plus the number of bars on the flank of males were analysed with
discriminant function analysis (DFA). The pigmentation pattern of females, however,

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 851

Merín Lagoon

34°S
Castillos Lagoon

Rocha Lagoon
25 km
58°W 56°W 54°W

FIG. 1. Geographical locations of Austrolebias luteoflamulatus populations sampled: Rocha lagoon drai-
nage (&); Castillos lagoon drainage (~); Merı́n lagoon drainage (^).

were not included in the DFA due to its categorical nature; this characteristic was
analysed separately. Raw data were examined for outliers that could indicate the pre-
sence of errors in measurements and were excluded from the data set (Lee, 1990). Prior to
the analysis, variables were standardized to a mean of zero and S.D. of one to approx-
imate multivariate normality and homocedasticity. Since data could not be normalized,
the DFA was used as an exploratory technique (James & McCulloch, 1990); thus
character differences between pairs of groups were tested with the Mann–Whitney U test.
Shape variation was analysed through morphometric data acquired as homologous
landmark co-ordinates along the lateral side of specimens. Landmarks assumed to be
homologous were recorded on paper using a camera lucida mounted on a Wild dissecting
microscope. Specimens were digitized as image files with a HP Scanjet 5200C scanner and
landmark co-ordinate data acquired using the software tpsDig (version 1.31) (http://
life.bio.sunysb.edu/morph/index.html). Fourteen landmarks were digitized per specimen
in lateral view (Fig. 2). Landmark configurations for all specimens were aligned by the
generalized Procustes superimposition procedure (Bookstein, 1991; Monteiro & dos Reis,
2000). The thin plate spline and uniform component approximation were used to project
the specimens into a linear tangent space in order to perform linear multivariate analysis
of shape variation and covariation (Bookstein, 1991). A DFA was performed over the
partial warp matrix and the uniform component scores generated by the software tpsRegr

5
6 7
(a) (b)
14 8
1
2 10
4 13
3 9
12 11
1

FIG. 2. The 14 landmarks used in the geometric morphometric analyses. (a) Ventral view of head and (b)
lateral view of whole body. 1, ventral junction of the anterior ceratohyal and basihyal; 2, snout tip;
3, anterior margin of eye orbit; 4, posterior margin of eye orbit; 5, origin of dorsal fin base; 6, end of
dorsal fin base; 7, dorsal origin of caudal fin; 8, end of caudal peduncle; 9, ventral origin of caudal
fin; 10, end of anal fin base; 11, origin of anal fin base; 12, origin of pelvic fin; 13, dorsal insertion of
pectoral fin on body; 14, insertion of the opercular membrane in the head.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
852 A. D’ANATRO AND M. LOUREIRO

(version 1.24), to examine the extent to which landmarks reveal phenetic clusters of
individuals of different shape. Canonical axes were evidenced in shape space by multi-
variate regression of partial warps, and uniform component scores on canonical scores of
the two roots obtained with tpsRegr. Influence of size on the discrimination obtained was
checked by a multivariate regression of the partial warps on centroid size, a robust
measure of general size (Bookstein, 1991), using the same software.

RESULTS

M E R IS T I C A ND P I G M E N T A T I O N C H A R A C T E R S
Tables I, II and III show the descriptive statistics for males and females of the
three groups analysed.
In the case of males, the DFA shows a high percentage of total individuals
(80
5%) correctly assigned to the a priori defined groups (Table IV). The analy-
sis indicates differentiation of males of Rocha from the other drainages [Fig. 3(a)
and Table V]. The discrimination occurs along root 1 [Fig. 3(a)] where the most
important variables involved are bars along the flanks (Fig. 4) and dorsal fin
rays. The highest squared Mahalanobis distance occurs between Rocha and
Castillos populations (Table V). Furthermore, most variables analysed were
significantly different among groups (Mann–Whitney U test, see Table VI).
Individuals grouped in the Rocha drainage present the highest number of
significantly different variables from the others (Tables I and VI).
On the other hand, females appear to be more similar to each other [Fig. 3(b)].
The DFA shows a low percentage of total individuals (59
8%) correctly assigned
(Table IV) and the squared Mahalanobis distances are almost the same for the
three groups (Table V). In this case, few variables present significant differences.
Rocha specimens, however, still appear distinct from the other two (Table VI).
The pigmentation patterns in females (Table III) do not show significant
differences among basins (Table VI).

MORPHOMETRIC CHARACTERS
The discriminant function analysis for the partial warps and uniform compo-
nents combined in males results in a clear differentiation between Rocha and
Merı́n specimens along root 1 (75
0% of explained variation), whereas the
individuals from Castillos are discriminated from the other populations along
root 2 (25
0% of explained variation) [Fig. 5(a)]. Shape changes in male popula-
tions with negative scores of root 1 (Rocha) involve anterior displacement of the
dorsal fin, deepening of the body, posterior expansion of the head, dorsal
displacement of the eyes and anterior displacement of the throat region
[Fig. 5(b)]. Merı́n individuals show positive scores of root 1 which represent
opposite changes to those described for Rocha. Castillos specimens present an
intermediate shape. Shape changes associated with positive scores of root 2
mainly involve elongation of the snout [Fig. 5(c)].
The squared Mahalanobis distances between group centroids corroborate this
segregation with significant differences in all cases (Table VII). The highest
proportion of correctly classified males corresponds to Merı́n individuals

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 #
TABLE I. Meristic characters and number of pigmentation bars of males of Austrolebias luteoflamulatus in the three different drainages studied

Rocha Castillos Merı́n

N Mean Minimum Maximum S.D. N Mean Minimum Maximum S.D. N Mean Minimum Maximum S.D.

D 45 24
0 22 28 1
2 46 22
9 18 25 1
6 37 22
4 20 25 1
2

A 45 20
9 18 25 1
5 46 20
8 15 23 1
3 37 20
3 16 23 1
4
C 45 23
4 21 27 1
3 46 22
6 20 27 1
7 37 22
9 20 27 1
4
PC 45 11
5 11 12 0
5 46 11
2 10 12 0
7 37 11
1 10 12 0
4
PV 45 5
4 4 7 0
7 46 4
9 4 6 0
4 37 5
1 4 6 0
5
LS 45 31
2 28 35 1
6 46 30
6 28 34 1
5 37 30
6 28 35 1
7
PS 45 16
9 13 20 1
5 46 15
9 13 20 1
6 37 15
9 14 18 1
2
PRS 45 19
2 15 24 2
3 46 19
6 15 26 2
3 37 17
9 14 23 2
0
SS 45 1
5 0 3 0
6 46 0
8 0 3 1
0 37 0
2 0 3 0
6
SN 45 16
5 13 21 1
8 46 15
4 12 20 1
8 37 15
8 12 22 2
2
B 45 9
2 7 12 1
1 46 6
9 4 10 1
1 37 7
4 6 10 1
1
D, dorsal fin rays; A, anal fin rays; C, caudal fin rays; PC, pectoral fin rays; PV, pelvic fin rays; LS, lateral series of scales; PS, scales around the caudal
peduncle; PRS, predorsal series of scales; SS, supraorbital scales; SN, number of neuromasts in the supraorbital series; B, number of black bars.
GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH

2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
853
 854

#
TABLE II. Meristic characters (see Table I) of females of Austrolebias luteoflamulatus in the three different drainages studied

Rocha Castillos Merı́n

N Mean Minimum Maximum S.D. N Mean Minimum Maximum S.D. N Mean Minimum Maximum S.D.

D 44 17
4 16 19 1
0 37 17
6 16 23 1
7 41 17
4 13 20 1
3

A 44 17
2 16 20 1
1 37 17
2 14 22 1
9 41 17
4 14 21 1
3
C 44 23
0 20 27 1
9 37 22
5 19 27 1
8 41 22
6 20 26 1
4
PC 44 11
3 10 13 0
6 37 11
1 10 13 0
8 41 10
9 10 12 0
6
PV 44 5
2 4 7 0
8 37 5
0 4 6 0
5 41 5
4 4 7 0
6
LS 44 31
2 27 36 2
2 37 29
5 27 32 1
6 41 30
2 25 35 2
2
PS 44 16
2 14 19 1
6 37 15
2 12 19 1
4 41 15
7 11 25 2
2
PRS 44 24
7 21 30 2
2 37 23
9 15 30 3
2 41 22
5 18 31 3
1
SS 44 1
2 0 3 0
9 37 0
8 0 3 1
0 41 0
9 0 3 1
0
SN 44 16
2 13 20 1
7 37 15
8 12 21 2
2 41 15
0 11 22 2
6
A. D’ANATRO AND M. LOUREIRO

2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 855

TABLE III. Frequencies of the different pigmentation pattern categories (FP) in females
of Austrolebias luteoflammulatus from three different diainages

Rocha Castillos Merı́n

d 6 2 9
db 29 24 21
b 9 3 3
N 44 29 33
d, dots; db, dots and bars; b, bars.

(92
1%), whereas the total percentage of males correctly classified is 83
7%.

(Table VIII).
In females, the highest differentiation occurs along root 1 (75
0% of explained
variation) and between the Rocha population and the others, with no over-
lapping of individuals [Fig. 6(a)]. Shape changes associated with positive scores
of root 1 (Rocha populations) involve posterior elongation and deepening of the
caudal peduncle, posterior displacement of the pelvic fin, shortening of the anal
fin, dorsal displacement of the eyes and shortening of the snout [Fig. 6(b)].
Opposite changes occur in negative scores of root 1, corresponding to Castillos
and Merı́n populations [Fig. 6(c)]. These populations present shape differences
between each other along root 2 (25
0% of explained variation). Positive scores
for Merı́n populations are mainly associated with posterior displacement of the
dorsal and pelvic fin origins [Fig. 6(c)].
The squared Mahalanobis distances between group centroids corroborate
this segregation. Castillos and Merin populations, however, are not signifi-
cantly different from each other (Table VII). The highest proportion of cor-
rectly classified females corresponds to Merı́n and Rocha populations (100%),
whereas the total percentage of females correctly classified is 93
5%.
(Table VIII).
In both, males and females the discrimination of populations is independent of
centroid size. All regression coefficients of partial warps on size are <0
3.

TABLE IV. Classification matrix of DFA performed on meristic characters of females,

and meristic plus pigmentation characters of males for Austrolebias luteoflamulatus
populations

Males Females
Per cent correctly assigned R C M Per cent correctly assigned R C M
R 86
7 39 2 4 61
7 27 10 7
C 78
3 2 36 8 51
3 7 19 11
M 75
7 2 7 28 65
8 6 8 27
Total 80
5 43 45 40 59
8 40 37 45
R, Rocha; C, Castillos; M, Merı́n.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
856 A. D’ANATRO AND M. LOUREIRO

5
(a)
4

–1
Second discriminant axis

–2

–3
–5 –4 –3 –2 –1 0 1 2 3 4 5

4
(b)
3

–1

–2

–3

–4
–4 –3 –2 –1 0 1 2 3
First discriminant axis

FIG. 3. Specimen co-ordinates of Austrolebias luteoflamulatus populations [Rocha Lagoon (&), Castillos
Lagoon (n) and Merin Lagoon (*)] on the two discriminant axes determined from meristic data: (a)
males and (b) females. Ellipses represent 95% CI.

TABLE V. Squared Mahalanobis distances among centroid of Austrolebias luteoflamulatus

populations based on merisitc characters of females, and meristic plus pigmentation
characters of males for A. luteoflamulatus populations. Male values above diagonal,
females values below diagonal

R C M
R 0 8
2 7
3
C 1
5 0 2
2
M 1
7 1
6 0
R, Rocha; C, Castillos; M, Merı́n.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 857

(a)

(b)

(c)

FIG. 4. Pigmentation patterns in males of Austrolebias luteoflamulatus from (a) Rocha, (b) Castillos and
(c) Merı́n Lagoons. Scale bar ¼ 1 cm.

DISCUSSION
The results of this study show that populations of A. luteoflamulatus are
structured according to a geographical pattern corresponding to different drai-
nages, thus the null hypothesis is confirmed. This differentiation occurs in
meristics, pigmentation and morphometric characters, but to different degrees
in male and female fish.
In the case of meristics and pigmentation pattern, the most pronounced differ-
ences occur between males, while females are more alike among drainages. These

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
858 A. D’ANATRO AND M. LOUREIRO

TABLE VI. Values of the Mann–Whitney U test for meristic characters and pigmentation
variables (see Table I) of Austrolebias luteoflamulatus

R v. C R v. M M v. C
Males Females Males Females Males Females
D 672 792 304 894 606 753
A 1026 738 647 800 601 627
C 712 680 667 791 730 712
PC 789* 690 497 657 752 690
PV 677 694 628* 808 751 545
LS 781 445 625* 664 833 621
PS 603 554 485 734 818 653
PRS 939* 699 578 470 484 534
SS 592 597 127 704 583 732
SN 677 685 645 578 777 595
B 150 – 225 – 650* –
FP – 616 – 576* – 385
FP, pigmentation pattern.
Values in bold, significant differences at P < 0
05.
*, Marginal values (P > 0
05 < 0
08).
R, Rocha; C, Castillos; M, Merı́n.

results agree with those of Loureiro (1996) who compared individuals from three
localities of these species. The variables that contribute most to the differentiation
in males are the number of dorsal fin rays and the number of vertical bars on the
flanks. On the contrary, when morphometry was considered, the level of differ-
entiation among drainages in both sexes was found to be similar.
These results indicate that meristic and morphometric variables are affected in
different ways in both sexes. At first glance it can be asserted that allopatric
differentiation has played an important role in the evolution of populations of
this species. This can be observed in the case of morphometric characters, in
both males and females. The naturally fragmented habitat of annual fishes and
their low vagility could favour this kind of phenomenon. Meristics and pigmen-
tation characters, however, suggest that other mechanisms could be acting in the
evolution of these population. Austrolebias luteoflamulatus is the only annual fish
in the Rocha lagoon drainage, whereas in Castillos and Merı́n drainages this
species is sympatric, and even syntopic in some ponds, with up to three other
annual species (Vaz-Ferreira & Melgarejo, 1984). Thus, the absence of competi-
tion with other annuals in Rocha could also be influencing the differentiation
observed. In line with Competition Theory (MacArthur & Levins, 1964, 1967;
Jaksic, 2001), differences between Rocha and the other drainage populations
could be a consequence of character displacement due to the interactions of A.
luteoflamulatus with other species of annual fishes which share habitat and
resources.
Lowlands associated with Rocha Lagoon present higher salinity levels than
the other drainages of the region (Conde & Sommaruga, 1999). Environmental
variables such as temperature, light, dissolved oxygen and salinity (Barlow, 1961;

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 859

5
(a)

4
Second discriminant axis (25% of variation)

–1

–2

–3
–4 –3 –2 –1 0 1 2 3 4 5

First discriminant axis (75% of variation)

(b)
5
7
14
6
2 3 4 8

13
1 12 11 10
9

(c) 5
7
14
6
2 3 4 8
13
12 10
1 11 9

FIG. 5. (a) Specimen co-ordinates of Austrolebias luteoflamulatus male populations [Rocha (&), Castillos
(n) and Merı́n (*) Lagoons) on the two discriminant axes determined from morphometric data. (b)
Shape deformation associated with first axis negative scores. (c) Shape deformation associated with
second axis positive scores. Magnitude of arrows are multiplied by a factor of three for a better
representation of landmarks (see Fig. 2) relative displacement. Ellipses represent 95% CI.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
860 A. D’ANATRO AND M. LOUREIRO

TABLE VII. Squared Mahalanobis distances among centroids of Austrolebias luteoflamulatus

populations based on morphometric characters, all significantly different at
P < 0
05, except * marginal value. Male values above diagonal, female values below
diagonal

R C M
R 0 5
9 11
8
C 22
1 0 6
7
M 14
9 7
1* 0
R, Rocha; C, Castillos; M, Merı́n.

Lindsey, 1988) may affect meristic characters and thus could cause the observed
differentiation among populations. The fact that females are highly homoge-
neous and males are well discriminated, however, is contrary to this argument.
Furthermore, this difference between sexes suggests that other evolutionary
forces such as sexual selection, (e.g. mate choice) may have an important
influence on the differentiation patterns found in this and other species of the
genus. Strong sexual dimorphism in the family Rivulidae supports this view
(Parenti, 1981). Furthermore, in a preliminary study of two other annual fish
species, Loureiro (1996) found that females preferred males which exhibited
more courtship displays. Theoretical models have demonstrated that female
preferences can alter male characteristics considerably (Lande, 1981; Turner &
Burrow, 1995). In this sense, the sensory drive hypothesis (Boughman, 2002)
could be considered in relation to the variation pattern found in A. luteoflamu-
latus. According to this hypothesis, the biotic and abiotic habitat differences
mentioned for Rocha populations could have shifted the communication system
of this species with the consequent change in male morphology.
The Neotropical ichthyofauna is the most diverse freshwater fauna (Vari &
Malabarba, 1998) and one of the major reasons advocated to explain this is the
complex geomorphological history of the region (Lundberg et al., 1998).
Austrolebias luteoflamulatus localities are associated with coastal systems influ-
enced by marine transgressions (5 m above present sea level, c. 5000 years before
present) and regressions (120 m below sea level, c. 35 000 years before present)

TABLE VIII. Classification matrix of DFA performed on morphometric characters of

Austrolebias luteoflamulatus populations

Males Females
Per cent correctly assigned R C M Per cent correctly assigned R C M
R 78
3 18 4 1 100 17 0 0
C 76
0 2 19 4 73
3 0 11 4
M 92
1 0 3 35 100 0 0 30
Total 83
7 20 26 40 93
5 17 11 34
R, Rocha; C, Castillos; M, Merı́n.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 861

5
(a)
4
Second discriminant axis (25% of variation)

–1

–2

–3

–4

–5
–5 –4 –3 –2 –1 0 1 2 3 4 5 6
First discriminant axis 75% of variation

(b) 5

6 7
14
3 4
2 13 8
10
9
1 12 11

(c) 5

6 7
14
3 4
8
2
13 9
10
1 11
12

FIG. 6. (a) Specimen co-ordinates of Austrolebias luteoflamulatus female populations [Rocha (&),
Castillos (n) and Merı́n (*) Lagoons] on the two discriminant axes determined from morphometric
data. (b) Shape deformation associated with first axis positive scores. (c) Shape deformation
associated with second axis positive scores. Magnitude of arrows are multiplied by a factor of
three for a better representation of landmarks (see Fig. 2) relative displacement. Ellipses represent
95% CI.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
862 A. D’ANATRO AND M. LOUREIRO

that have occurred at least since the upper Miocene (Alonso, 1978;
Sprechmann, 1978; Garcı́a-Rodriguez et al., 2001, 2002; Lambeck et al.,
2002). These changes have probably affected the evolutionary history of this
and other species of annual fishes of this region, as hypothesized by Garcı́a
et al. (2000, 2001), as well as other fish species such as the atheriniform genus
Odontesthes (Beheregaray et al., 2001). Further research on the genetic struc-
ture of these populations is necessary to associate population changes, such as
demographic expansions and bottlenecks, with the geomorphological history of
the area.
Could Rocha populations be considered a new species? It is still not known if
the populations identified are monophyletic units. A phylogeographical
approach would be very useful to assess this question. Among all species defini-
tions proclaimed, there are two currently widely accepted: the biological species
concept (Mayr, 1963) and the phylogenetic species concept (Cracraft, 1983).
While the former is somehow difficult to test, the latter seems more readily
applicable in this case. In these analyses, however, no diagnostic character
could be found for any of the populations studied. A definition of geographical
isolates as different species is, in many occasions, an arbitrary decision (Mallet,
1995) particularly when characters present high variability and their ranges
overlap. In this particular case the differences between Rocha and the other
populations are not discrete, and thus delimitation of a new taxon seems
inappropriate.
Austrolebias luteoflamulatus is restricted to the eastern wetlands of Uruguay
which have been declared a Reserve of the Biosphere by UNESCO (PROBIDES,
1999). Many localities, including most of Rocha drainage populations, are found
in close proximity to urban settlements. Thus, conservation efforts for this
species should consider the populations identified in this study because they
have diverged morphologically and may represent different evolutionary units
(Ryder, 1986; Moritz, 1994, 2002). In terms of conservation management it will
also be important to consider the geographic situations which have promoted
rapid differentiation among populations.

We are grateful to E. Lessa and R.O. de Sá for the helpful revision of the manuscript,
R. Fariña for the suggestions made along this work, M. Arı́m for the statistical advice
and two anonymous reviewers who improved the quality of this article. This work was
partially funded by PEDECIBA, Universidad de la República, Uruguay.

References
Alonso, C. (1978). La fauna de moluscos del yacimiento de Playa Pascual con referencia
a otros yacimientos estuaricos y marinos del cuaternario de Uruguay.
Comunicaciones de la Sociedad de Malacologı´a del Uruguay 4, 365–383.
Bailey, R. M. & Smith, G. R. (1981). Origin and geography of the fish fauna of the
Laurentian Great Lakes drainage. Canadian Journal of Fisheries and Aquatic
Sciences 38, 1539–1561.
Barlow, G. W. (1961). Causes and significance of morphological variation in fishes.
Systematic Zoology 10, 105–117.
Beheregaray, L. B., Sunnucks, P. & Briscoe, D. A. (2001). A rapid fish radiation
associated with the last sea-level changes in southern Brazil: the silverside

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 863

Odontesthes perugiae complex. Proceedings of the Royal Society of London B 269,

65–73.
Bookstein, F. L. (1991). Morphometric Tools for Landmark Data: Geometry and Biology.
New York: Cambridge University Press.
Boughman, J. W. (2002). How sensory drive can promote speciation. Trends in Ecology
and Evolution 17, 571–577.
Cardozo, V. (1999). Tasa Metabólica y excreción del nitrógeno en peces anuales
Cynolebias viarius (Cyprinodontiformes). Tesis de Maestrı́a, PEDECIBA,
Universidad de la República, Montevideo.
Conde, D. & Sommaruga, R. (1999). A review of the state of limnology in Uruguay. In
Limnology in Developing Countries, Vol. 2 (Wetzel, R. G. & Gopal, B., eds), pp.
1–31. Nueva Delhi: International Association for Limnology (SIL).
Costa, W. J. E. M. (1998). Phylogeny and Classification of Rivulidae revisited: origin
and evolution of annualism and miniaturization in Rivulid fishes
(Cyprinodontiformes: Aplocheiloidei). Journal of Comparative Biology 3, 33–92.
Cracraft, J. (1983). Species concepts and speciation analysis. Current Ornithology 1,
159–187.
Garcı́a, G., Scvortzoff, E., Maspoli, M. C. & Vaz-Ferreira, R. (1993). Analysis of
karyotipic evolution in natural populations of Cynolebias (Pisces:
Cyprinodontiformes, Rivulidae). Cytologia 58, 85–94.
Garcı́a, G., Scvortzoff, E. & Hernández, A. (1995). Karyotypic heterogeneity in South
American annual killifishes of the genus Cynolebias (Pises, Cyprinodontiformes,
Rivulidae). Cytologia 60, 103–110.
Garcı́a, G., Wlasiuk, G. & Lessa, E. (2000). High levels of mitochondrial cytocrome b
divergence in annual killifish of the genus Cynolebias (Cyprinidontiformes,
Rivulidae). Zoological Journal of the Linnean Society 129, 93–110.
Garcı́a, G., Lalanne, A. I., Aguirre, G. & Cappetta, M. (2001). Chromosome evolution in
the annual killifish genus Cynolebias and mitochondrial phylogenetic analysis.
Chromosome Research 9, 437–448.
Garcı́a-Rodriguez, F., Del Puerto, L., Castiñeira, C., Inda, H., Bracco, R., Sprechmann,
P. & Scharf, B. W. (2001). Preliminary paleolimnological study of Rocha Lagoon,
SE Uruguay. Limnologica 31, 221–228.
Garcı́a-Rodriguez, F., Castiñeira, C., Scharf, B. W. & Sprechmann, P. (2002). The
relationship between trophic states and sea level variation in the Rocha Lagoon,
Uruguay. Neues Jahrbuch für Geologie und Paläontologie, Monatshefte 1, 27–47.
Jaksic, F. A. (2001). Ecologı´a de Comunidades. Santiago de Chile: Universidad Católica
de Chile.
James, F. & McCulloch, Ch. (1990). Multivarite analysis in ecology and systematics:
Panacea or Pandora’s box? Annual Review of Ecology and Systematics 21, 129–166.
Lambeck, K., Esat, T. M. & Potter, E. K. (2002). Links between climate and sea levels
for the past three million years. Nature 419, 199–206.
Lande, R. (1981). Models of speciation by sexual selection on phylogenic traits.
Proceedings of the National Academy of Sciences (USA) 78, 3721–3725.
Lee, J. C. (1990). Source of extraneous variation in the study of meristic characters: the
effect of size and of inter-observer variability. Systematic Zoology 39, 31–39.
Lindsey, C. C. (1988). Factors controlling meristic variations. In Fish Physiology,
Vol. XIB (Hoar, W. S. & Randall, D. J., eds), 197–274. San Diego, CA:
Academic Press.
Loureiro, M. (1996). Osteology, morphology and female choice in species of the genus
Cynolebias (Cyprinidontiformes, Rivulidae). MS Thesis dissertation, University of
Richmond, Richmond, VA.
Loureiro, M. (1999). Variación morfológica en Cynolebias adloffi y Cynolebias viarius
(Cyprinidontiformes, Rivulidae). Boletı´n de la Sociedad Zoológica del Uruguay
(2a. e´poca) (Actas V Jornadas de Zoologı´a del Uruguay) 11, 27.
Lundberg, J. G., Marshall, L. G., Guerrero, J., Horton, B., Malabarba, M. C. S. L. &
Wesselingh, F. (1998). The stage for Neotropical fish diversification: a history of
tropical South American rivers. In Phylogeny and Classification of Neotropical

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
864 A. D’ANATRO AND M. LOUREIRO

Fishes (Malabarba, L. R., Reis, R. E., Vari, R. P., Lucena, Z. M. S. & Lucena,
C. A. S., eds), pp. 13–48. Porto Alegre: EDIPUCRS.
MacArthur, R. H. & Levins, R. (1964). Competition, habitat selection, and character
displacement in a patchy environment. Proceedings of the National Academy of
Sciences (USA) 51, 1207–1210.
MacArthur, R. H. & Levins, R. (1967). The limiting similarity, convergence, and diver-
gence of coexisting species. American Naturalist 101, 377–385.
Mallet, J. (1995). A species definition for the modern synthesis. Trends in Ecology and
Evolution 10, 294–299.
Mayr, E. (1963). Animal Species and Evolution. Cambridge, MA: Harvard University
Press.
Monteiro, L. R. & dos Reis, S. F. (2000). Principios de Morfometria geometrica. Ribeirão
Preto: Holos.
Moritz, C. (1994). Defining ‘Evolutionarily Significant Units’ for conservation. Trends in
Ecology and Evolution 9, 373–375.
Moritz, C. (2002). Strategies to protect biological diversity and the evolutionary
processes that sustain it. Systematic Biology 51, 238–254.
Parenti, L. R. (1981). A phylogenetic and biogeographic analysis of cyprinodontiform
Fishes (Teleostei: Atherinomorpha). Bulletin of the American Museum of Natural
History 168, 341–557.
PROBIDES (1999). Plan Director. Reserva de Biósfera, Bañados del Este, Uruguay.
Rocha: PROBIDES.
Reichert, J. (1994). Láminas y datos complementarios sobre las especies del género
Cynolebias Steindachner 1876, halladas y descritas para Uruguay, hasta el año
1986. Acuariologı´a -Comunicaciones Ictiológicas 2, 1–38.
Ryder, O. A. (1986). Species conservation and systematics: The dilemma of subspecies.
Trends in Ecology and Evolution 1, 9–10.
Schaefer, S. A. & Cavender, T. M. (1986). Geographic variation and subspecific status of
Notropis spilopterus (Pisces: Cyprinidae). Copeia 1986, 122–130.
Sprechmann, P. (1978). The paleoecology and paleogeography of the Uruguayan coastal
area during the Neogene and Quaternary. Zitteliana 4, 3–72.
Thorpe, R. S. (1984). Primary and secondary transition zones in speciation and
population differentiation: a phylogenetic analysis of range expansion. Evolution 38,
233–243.
Turner, G. F. & Burrow, M. T. (1995). A model of sympatric speciation by sexual
selection. Proceedings of the Royal Society of London B 260, 287–292.
Vari, R. P. & Malabarba, L. R. (1998). Neotropical ichthyology: an overview. In
Phylogeny and Classification of Neotropical Fishes (Malabarba, L. R., Reis, R. E.,
Vari, R. P., Lucena, Z. M. & Lucena, C. A. S., eds), pp. 1–11. Porto Alegre:
EDIPUCRS.
Vaz-Ferreira, R. & Melgarejo, A. R. (1984). La distribución de las especies del género
Cynolebias Steindachner, 1876, en el Uruguay, con notas sobre C. alexandri
Castello y López, 1974. Boletı´n de la Sociedad Zoológica del Uruguay 2, 41–46.
Wildekamp, R. H. (1995). A World of Killies. Atlas of the Oviparous Cyprinidontiforms
Fishes of the World, Vol. II. Mishawaka, IN: American Killifish Association.
Wourms, J. P. (1972). The developmental biology of the annual fishes III. Pre-embryonic
and embryonic diapause of variable duration in the eggs of annual fishes. Journal
of Experimental Zoology 182, 389–414.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865
 GEOGRAPHIC VARIATION IN AN ANNUAL KILLIFISH 865

APPENDIX. Specimens examined in this study. Vertebrate Collection, Facultad de

Ciencias (Montevideo, Uruguay), Institutional Code ZVC-P.
Rocha Lagoon drainage: Route 10 km 225, ZVC-P 4077, ZVC-P 4101; Road to
Puerto de los Botes ZVC- P 4085, ZVC-P 4100, ZVC-P 4103, ZVC-P 4104, ZVC-
P 4105; Route 9 km 205 ZVC-P 4078.
Castillos Lagoon drainage: Route 10 Los Sauces, ZVC-P 784, ZVC-P 786,
ZVC-P 893, ZVC-P 902; Route 9 km 228
5, ZVC-P 4082; Route 9 Km 230,
ZVC-P 4084; Route 10 next to Valizas Road, ZVC-P 4093; Route 10 Vivero
Forestal Cabo Polonio, ZVC-P 4099; Route 10 Palmeras Gemelas, ZVC-P 4107,
ZVC-P 4130.
Merı́n Lagoon drainage: Route 18 km 369
5, ZVC-P 4114, ZVC-P 4112,
ZVC-P 4349, ZVC-P 4354, ZVC-P 4355; La Coronilla, ZVC-P 3892, ZVC-P
4083; Route 19 San Luis creek, ZVC-P 3942, ZVC-P 4089, Route 9 km 272,
ZVC-P 4092, ZVC-P 4249; Route 19 San Miguel creek, ZVC-P 529, ZVC-P
3819, ZVC-P 4096, ZVC-P 4116; Route 14 km 269
2, ZVC-P 4115, ZVC-P 4356;
Route 91, ZVC-P 4248; Route 15 km, 156 ZVC-P 4098.

# 2005 The Fisheries Society of the British Isles, Journal of Fish Biology 2005, 67, 849–865