Abstract The Andalusian toothcarp, Aphanius

baeticus, is a critically endangered cyprinodontid

species, with only nine known extant populations.

Although not yet studied in the field, the distri-

bution and abundance of Andalusian toothcarp

are thought to be strongly influenced by inter-

specific interactions. We analysed the abundance

and microhabitat use of Andalusian toothcarp in

two water courses, one in which several other fish

species occurred (sympatric site) and one hy-

persaline stream in which toothcarp was the only

species present (allopatric site). Fish were sam-

pled using plastic minnow traps and results were

analysed separately for three size categories.

Toothcarps were clearly more abundant in the

allopatric population than in the sympatric one,

though the difference was less apparent in the

smallest size category. In coexistence with other

species, toothcarp occupied shallower microhab-

itats, but in both sites in the absence of shelter fish

selected deeper positions than in its absence.

While in the sympatric site sheltered microhabi-

tats were used predominately by small individu-

als, in the allopatric ones they were used by larger

ones. Observed patterns strongly suggest that

predation is the main mechanism involved in the

differences in abundance and microhabitat use

between sites. Our results confirm that the pres-

ence or absence of coexisting species is an

important habitat feature for Andalusian tooth-

carp populations.

Keywords Cyprinodontidae Æ Mediterranean

streams Æ Predation Æ Competitive interactions ÆHabitat shifts

Introduction

Interspecific competitive and predatory interac-

tions are important determinants of organism (e.g.

behaviour) and population (e.g. density) character

(Sih et al. 1985). Numerous studies have shown

that among fishes these interactions can modify

abundance (Baltz et al. 1982; Perssons 1983), sex-

ratio (Belk and Lydeard 1994), growth (Bystrom

and Garcıa-Berthou 1999), diet (Brabrand 1985)

or habitat use (Power 1984; Harvey 1991; Prenda

et al. 1996). When dealing with endangered

species, it is crucial to take into account the

possible effects of co-occurring species that could

reduce population viability of the former (Fuselier

2001; Rincon et al. 2002).

M. Clavero (&)Institut d’Ecologia Aquatica, Universitat de Girona,17071 Girona, Spaine-mail: miguel.clavero@udg.es

F. Blanco-Garrido Æ J. PrendaDepartamento de Biologıa Ambiental y SaludPublica, Universidad de Huelva. Avd/Andalucıa s/n,21007 Huelva, Spain

Environ Biol Fish (2007) 78:173–182

DOI 10.1007/s10641-006-9088-2

123

ORIGINAL PAPER

Population and microhabitat effects of interspecificinteractions on the endangered Andalusian toothcarp(Aphanius baeticus)

Miguel Clavero Æ Francisco Blanco-Garrido ÆJose Prenda

Received: 24 March 2006 / Accepted: 15 June 2006 / Published online: 19 July 2006� Springer Science+Business Media B.V. 2006

The Andalusian toothcarp, Aphanius baeticus, is

a small (usually less than 50 mm) euryhaline cyp-

rinodontid species (Sanz 1985; Oltra and Todolı

2000) that inhabits streams and lagoons in the

Atlantic slope of southern Spain. It has been re-

cently described as a separate species from A. ibe-

rus (Doadrio et al. 2002), which is distributed along

the Iberian Mediterranean coast (Doadrio 2001).

Though Andalusian toothcarp is not yet officially

listed as a threatened species, according to Doadrio

et al. (2002) it should be considered Critically

Endangered (CR), due to its extremely reduced

distribution with only nine known locations.

Cyprinodontid fishes typically inhabit envi-

ronments that are marginal for other fishes, such

as small streams from arid or semiarid regions,

which often lack competitors and major predators

(McMahon and Tash 1988; Bennett and Beittin-

ger 1997). They often occur at high population

densities only in habitats supporting few or no

other species (Echelle et al. 1972). Both Iberian

Aphanius species have been shown to be very

sensitive to the presence of other fish species.

However, although special interest has been de-

voted to remark the impact of exotic species on

Iberian cyprinodontids (Rincon et al. 2002 and

references therein), direct field work on the issue

is very scarce (e.g. Garcia-Berthou and Moreno-

Amich 1991). Information on the interspecific

relations between Iberian toothcarps and sym-

patric native species is even scarcer. In spite of

this, Prenda et al. (2003) reported a negative

relationship between the abundance of Andalu-

sian toothcarp and that of other fish species

(including both native and introduced). Tooth-

carp achieved its highest abundances in locations

where it is the only species present.

Clavero et al. (2005) analysed microhabitat use

of Andalusian toothcarp and other small-bodied

fishes in a species-rich stream stretch. Due to

spatio-temporal segregation of fish species, that

work suggested that toothcarp microhabitat use

could be influenced by biotic interactions. In the

present study we compare results from this

Andalusian toothcarp population with those from

another one inhabiting a hypersaline stream in

which no other fish species is present. Our aim was

to asses the potential influences of interspecific

interactions on Andalusian toothcarp abundance

and microhabitat use. A better understanding of

these interactions could have important implica-

tions for conservation of the species. We limited

our study to the summer dry season, when most

Mediterranean-regime streams are reduced to

isolated pools (Gasith and Resh 1999) and inter-

actions between fish species are most acute

(Magalhaes et al. 2002; Magoulick and Kobza

2003).

Study sites

Tarifa-sympatric site

In the lower reaches of the La Vega River near

Tarifa (Cadiz, S Spain) Andalusian toothcarps

(henceforth toothcarp) occupies a stretch of about

600 m length just upstream of the stream’s tidal

section (Clavero et al. 2005). Flow ceases during

summer, when only some pools retain freshwater.

The occurrence of both estuarine and freshwater

species favours the presence of a rich ichthyofa-

una in the site (Clavero et al. 2006). Among fish

species occupying this stream stretch at least eel,

Anguilla anguilla, and European sea bass,

Dicentrarchus labrax, the later being quite

uncommon, can predate upon the whole size

range of toothcarp. Other species, such as Iberian

chub, Squalius pyrenaicus, large sand smelt,

Atherina boyeri, individuals and even large

invertebrates (e.g. red swamp crayfish, Procamb-

arus clarkii, or water scorpions, Nepa cinerea),

could also be able to predate upon small tooth-

carps (e.g. Rosecchi and Crivelli 1992; Blanco-

Garrido et al. 2003). Viperine snake, Natrix

maura, is abundant in the area (Clavero et al.

2005) and shore birds are also present.

Montellano-allopatric site

El Montero Stream, located near the village of

Montellano (Sevilla, S Spain), is a tributary of El

Salado River, which flows into the Guadalquivir

River’s marshes. El Montero is a hypersaline

stream, with conductivity in summer pools

reaching more than four times that of sea water

(ranging 110–180 mS cm–1). Toothcarp is the only

fish species present in this stream. The main fish

174 Environ Biol Fish (2007) 78:173–182

123

predators observed in the area were shore birds

(kingfisher, Alcedo atthis, or egrets, Ardeidae),

since viperine snake and large predatory inver-

tebrates were absent.

Methods

Fish captures

Fish were captured using cylindrical (240 mm

length, 95 mm width, 21 mm mouth) plastic

minnow traps (Clavero et al. 2006), henceforth

traps, used in pairs, with one trap touching the

stream bottom and another placed just below the

water surface (Clavero et al. 2005). Traps were

set during the day for 3.9 ± 2.2 h (mean ± SD).

The depth of each trap and the presence of sur-

rounding shelter (submerged or emergent mac-

rophytes or among riverine vegetation entering

the water) were noted. Captured fish were mea-

sured for total length (TL) and released. The

sympatric site was surveyed in five occasions

(October 2002, July and September 2003, July

and August 2004), while two surveys were per-

formed in the allopatric one (October 2003 and

August 2005). Sampling effort in each site is

shown in Table 1.

Data analyses

Toothcarp densities were expressed as catch per

unit of effort (CPUE), which was defined as the

number of fish captured per trap per hour

(f · t · h). To account for ontogenetic differ-

ences in microhabitat use, captured individuals

were classified into three size categories: small

(£22 mm), medium (23–28 mm) and large

(>28 mm) (Clavero et al. 2005) (Fig. 1). CPUE

values were calculated for each size category. We

used site (allopatric–sympatric), trap position

(bottom–surface), shelter (present–absent) and

depth as explanatory factor to analyse variations

in fish density and fish size. Depth measurements

were converted to 4 categories ( < 20 cm, 20–

39 cm, 40–59 cm and >59 cm) and were intro-

duced as a categorical factor in the analyses.

Variations in fish density were analysed

through generalised linear models (GLMs). Since

CPUE values derive from count data, we used a

Poisson error distribution and log link function.

Significance of explanatory factors’ effects was

assessed through the F-statistic instead of using

v2-tests, in order to avoid inflated Type-I error

probability due to overdispersion (i.e. dispersion

parameter being larger than 1) (Crawley 2002).

Different models were run for data from each size

Table 1 Total sampling effort and total sample sizes in for the different levels of the explanatory variables used in theanalyses

Position Shelter Depth (cm) Number of traps Number of fish

Allopatric Sympatric Allopatric Sympatric

Bottom Yes < 20 4 8 61 219Bottom Yes 20–39 12 48 1052 724Bottom Yes 40–59 4 29 253 56Bottom Yes ‡60 5 8 0 9Bottom No < 20 11 20 110 184Bottom No 20–39 20 37 840 552Bottom No 40–59 21 17 1115 73Bottom No ‡60 8 9 865 18Surface Yes < 20 3 5 0 37Surface Yes 20–39 12 20 101 211Surface Yes 40–59 4 9 0 29Surface Yes ‡60 5 3 0 0Surface No < 20 10 3 0 0Surface No 20–39 20 15 0 58Surface No 40–59 21 12 0 28Surface No ‡60 8 9 0 3

Total 168 252 4397 2201

Environ Biol Fish (2007) 78:173–182 175

123

category. A first set of analyses used included as

explanatory factors trap position and site. A sec-

ond set of analyses focused exclusively on bottom

traps and included site, depth and shelter as fac-

tors. Full factorial models were performed in ev-

ery occasion.

Factorial analysis of variance (ANOVA) was

used to analyse variations in fish size. Only fish

captured in bottom trap were used in this analysis,

which included site, depth and shelter as factors.

To compare effect sizes (i.e. importance of fac-

tors) we used partial g2 (partial eta squared),

which is the proportion of variation explained by

a certain effect (effect SS/effect SS + error SS).

Partial g2 has the advantage of not depending on

the number of sources variation in the ANOVA

design used or the number of levels (i.e. degrees

of freedom) of each factor (Tabachnick and Fidell

2001).

Results

We deployed 420 traps to capture 6598 toothcarps

(Table 1). Overall mean CPUE was 3.88 f · t · h

with a maximum value of 124 f · t · h. Tooth-

carp abundance was almost five times greater in

the allopatric site (7.3 f · t · h) than in the sym-

patric one (1.6 f · t · h). When only bottom traps

were considered, inter-site differences became

even more important (13.2 vs. 2.0 f · t · h), since

very few toothcarps were caught in surface posi-

tions in the allopatric site (Table 1). However, the

effect of site on toothcarp abundance was

dependent on fish size. There were no or only

slight difference in the abundance of small toot-

carp between sites, while medium and large-sized

individuals were much more abundant in the

allopatric site (Tables 2, 3). In the different

models that were run the factor site accounted for

between 3 and 6 times more deviance (i.e. vari-

ance) in fish density in the two largest size cate-

gories than in the smallest one.

Toothcarp tended to occupy positions near the

stream bottom, a behaviour that was consistent

between sites in the two largest size categories.

However, while in the allopatric site toothcarps

across the whole size range were equally scarce

on surface positions, small individuals in the

sympatric site were as abundant on the surface as

at the bottom (Fig. 2).

The densities of all toothcarp size categories

were influenced by water depth (Table 3). This

influence changed with fish size, larger fish occu-

pying deeper waters (Fig. 3). The significant ef-

fect of the interaction depth · shelter (Table 3)

accounted for the fact that fish across all sizes

0

200

400

600

800

<14 20 32 >40

Num

ber

of in

divi

dual

s

Fish size (mm)

Fig. 1 Size frequency distribution of captured toothcarpin the sympatric (black bars) and allopatric (white bars)sites. Size limits separating the three size categories (small,medium, large) are also shown

Table 2 Results of the GLM on the influence of site (sympatric–allopatric) and trap position (surface–bottom) on densities(CPUE) of the different Andalusian toothcarp size categories

d.f. Size category

Small Medium Large

Site 1 5.2 16.1*** 18.5***Position 1 19.5* 14.2*** 22.5***Site · position 1 7.0 < 0.01 < 0.01Error 416% explained deviance 7.1 30.3 40.5Dispersion parameter 3.6 4.5 4.3

Numbers are % deviance (i.e. variance) explained by each factor in the model. Significance levels (as assessed by the Fstatistic): *P < 0.05, **P < 0.01, ***P < 0.001

176 Environ Biol Fish (2007) 78:173–182

123

occupied shallower waters when being protected

by shelter. Medium and large toothcarp changed

their depth preferences between sites, tending to

use deeper microhabitats in the allopatric site,

clearly avoiding the shallowest positions. Though

large toothcarp’s density was greater in exposed

microhabitat their use of sheltered microhabitats

also changed among sites (Table 3). In the sym-

patric site exposed microhabitat were clearly

preferred while in the allopatric one shelter

preference was dependent on water depth

(Fig. 3).

Although there was a significant effect of site

on fish size, toothcarp being larger in the allo-

patric site (Fig. 1), the most influential factor on

fish size was the interaction between site and

shelter (Table 4). While in the sympatric site

sheltered microhabitats were used by smaller

toothcarps than exposed ones, in the allopatric

site sheltered microhabitats were used by larger

fish (Fig. 4). Superimposed to this change in fish

size across microhabitats and sites, there was a

general increase of mean fish size with increasing

water depth.

Discussion

Toothcarp populations

There were important differences in toothcarp

abundance between sites, with higher fish density

Table 3 Results of the GLMs on the influence of site, depth and shelter on densities (CPUE) of the different Andalusiantoothcarp size categories

d.f. Size category

Small Medium Large

Site 1 5.2** 28.3*** 29.1***Depth 3 7.1** 2.6* 6.3***Shelter 1 0.1 0.2 1.9*Site · depth 3 4.8* 4.9** 2.6*Site · shelter 1 0.5 0.0 2.9**Depth · shelter 3 1.8 5.9*** 7.8***Site · depth · shelter 3 0.1 0.5 0.7Error 245% explained deviance 19.7 42.3 51.4Dispersion parameter 3.5 4.3 4.9

Numbers are % deviance (i.e. variance) explained by each factor in the model. Only bottom traps were used in theseanalyses. Significance levels (as assessed by the F statistic): *P < 0.05, **P < 0.01, ***P < 0.001

Allopatric Sympatric

0

0.1

0.2

0.3

0.4

0.5

Fis

h de

nsity

- L

og

(CP

UE

+1)

10

22 mm 23-28 mm >29 mm 22 mm 23-28 mm >29 mm≥ ≥

Fig. 2 Mean CPUEs(+SE) of the differenttoothcarp size classes inbottom (black bars) andsurface (white bars) traps,shown separately for echsampling site

Environ Biol Fish (2007) 78:173–182 177

123

0

0.2

0.4

0.6

0.8

1

0

0.2

0.4

0.6

0.8

1

0

0.2

0.4

0.6

0.8

1

Allopatric Sympatric

22 mm

23–28 mm

>29 mm

<20 30 50 60<20 30 50 60

Depth (cm)

Fis

h de

nsity

- L

og

(CP

UE

+1)

10

≥

> >

Fig. 3 Mean CPUEs(±SE) of the differenttoothcarp size classes inrelation to depth andpresence of shelter:sheltered (filled circles),unsheltered (emptycircles), shown separatelyfor ech sampling site

Table 4 Results from the factorial ANOVA assessing the influence of site, depth and shelter on toothcarp size

d.f. F Partial g2 P

Site 1 54.8 0.009 ***Depth 3 32.6 0.016 ***Shelter 1 0.2 0.000 0.62Site · depth 3 7.0 0.003 ***Site · shelter 1 136.7 0.022 ***Depth · shelter 3 11.5 0.006 ***Site · depth · shelter 3 6.2 0.002 **Error 6116Adjusted R2 0.16

Only bottom traps were used in these analyses. Significance levels: *P < 0.05, **P < 0.01, ***P < 0.001

178 Environ Biol Fish (2007) 78:173–182

123

in the allopatric population. This result is con-

gruent with the previously observed negative

relationship between sympatric fish abundance

and that of toothcarp (Prenda et al. 2003). Do-

adrio et al. (2002) also noted that toothcarp oc-

curred at very low densities when in coexistence

with invasive mosquitofish (Gambusia holbrooki)

or mummichog (Fundulus heteroclitus). In fact,

the only extant dense toothcarp populations are

reported from the extremely harsh Mediterra-

nean hypersaline streams, which during summer

are reduced to isolated pools where the effects of

high salinity, temperature and low oxygen content

are combined. Apart from toothcarp no other

species can occupy these stream reaches, allowing

the development of high density competitor and

predator-free populations, as has been previously

noted for North American cyprinodontids

(Echelle et al. 1972; Gido et al. 1999; Martin and

Saiki 2005). Since high salinity per se does not

provide toothcarp any reproductive advantage

(Oltra and Todolı 2000), even reducing individual

survival at extreme values (Sanz 1985), the ab-

sence of interspecific interactions would appear as

an important benefit in hypersaline streams.

Other advantages of populations occupying these

harsh environments could involve parasite exclu-

sion (Rogowski and Stockwell 2006). Similar

competitor or predator-free refugia have been

also described associated to other extreme envi-

ronmental conditions, such as thermal waters (e.g.

Ortubay and Cussac 2000).

Abundance differences were not as evident for

small toothcarp as for large size categories, sug-

gesting that the mechanism responsible for

abundance differences would involve an in-

creased mortality in the sympatric site (e.g. see

Rogowski and Stockwell 2006). Kodric-Brown

and Mazzolini (1992) proposed that salinity reg-

ulated interactions in two lake populations of

pupfish Cyprinodon pecosensis and killifish

Fundulus zebrinus. As in our study sites, pupfish

was more abundant and larger at high salinity.

Pou-Rovira et al. (2004) monitored two Aphanius

iberus populations in coastal lagoons with low

abundances of other fish species and found that

toothcarp abundances increased in early summer

but reached their maxima at early autumn. In

contrast, intense monitoring of the sympatric site

population showed that toothcarp density peaked

in early summer, due to a massive emergence of

young of the year, and then suffered a pro-

nounced decline (Clavero et al. 2006), again sug-

gesting high juvenile mortality rates. In fact,

predation upon juvenile Aphanius toothcarp by

mosquitofish and sharp reproduction failure in

the presence mosquitofish was demostrated by

Rincon et al. (2002) in microcosm experiments.

Microhabitat use

Toothcarp’s microhabitat preferences showed

clear differences in the two studied populations.

When in coexistence with other fish species,

21

23

25

27

29

31

33

<20 30 50 >60 <20 30 50 >60

Depth (cm)

Fis

h si

ze (

mm

)

Allopatric SympatricFig. 4 Mean tootcarpsize (±SE) in relation todepth and presence ofshelter: sheltered (filledcircles), unsheltered(empty circles), shownseparately for echsampling site

Environ Biol Fish (2007) 78:173–182 179

123

toothcarp positively selected shallow areas, while

sheltered microhabitats were used by smaller

individuals than exposed ones. Contrarily, in the

allopatric site the shallowest depth class was not

the preferred one for any size category, there was

a clear preference for deep microhabitats in the

absence of shelter (which increased with fish size)

and sheltered microhabitats were used by larger

fish (see Figs. 3, 4).

It is commonly assumed that the combined

effects of aquatic predators (mainly piscivorous

fishes) and terrestrial ones (wading birds and

mammals) are the main exogenous determinants

of microhabitat use by stream fish. In order to

avoid piscivorous fish, small fish usually occupy

shallow microhabitats, which constitute effective

refuges increased vulnerability of large fish to

wading predators (Power 1984, 1987). Thus, in the

absence of aquatic predators a shift towards

deeper microhabitats should be expected (Mit-

telbach 1986; Harvey 1991). This pattern was

evident in the allopatric site and in exposed mi-

crohabitats, with an almost linear increase of fish

density with water depth. The occupance of

shallower positions in the sympatric site toothcarp

would therefore lead to a higher vulnerability to

predation by wading animals. However, since

shelter can provide effective protection from

terrestrial predators (e.g. Valdimarsson and

Metcalfe 1998), the presence of shelter allowed

fish to occupy shallower microhabitats (see also

Clavero et al. 2005), as has been observed in both

study sites.

The simultaneous action of terrestrial and

aquatic predators is also claimed to be the

determinant of the larger-fish-deeper-water pat-

tern (Harvey and Stewart 1991). This pattern was

very clear in unsheltered microhabitats in both

sites, suggesting that the predation pressure from

terrestrial animals may be strong enough to in-

duce a size-related depth use. Depth-size rela-

tionships were less clear when fish used sheltered

microhabitat, and, more importantly, size of fish

using them strongly differed between sites (see

Fig. 4). These opposite patterns can be linked to

differences in predation risk. Shelter provided by

aquatic plants has been shown to reduce foraging

success of piscivorous fish (Orth et al. 1984;

Gotceitas 1990). Since small individuals are more

vulnerable to gape-limited predators such as fish

(Tonn and Paskowski 1986), small fish would tend

to occupy sheltered microhabitats in the presence

of predatory fish. In the other hand, if, as in the

allopatric site, only terrestrial predators are

present, sheltered microhabitats should be se-

lected by larger individuals, since wading birds

are size-specific predators, consuming preferen-

tially larger fish (Britton and Moser 1982). The

different patterns of size related use of sheltered

microhabitats in the two study sites fit well with

these predictions.

Although it had been previously suggested that

introduced species was the most important factor

implied in the decline of Iberian toothcarp species

(Doadrio 2001; Doadrio et al. 2002), this paper

shows that the presence of coexisting native fish

species is also an important habitat feature

determining toothcarp’s ecology. We do not have

direct evidence on the mechanisms generating the

observed patterns, but differences in abundance,

population size structure and microhabitat use

could be related to the effect of predation. Extant

populations where toothcarp is the only fish spe-

cies present should have priority consideration for

conservation purposes. This should also be taken

into account when selecting possible areas for

reintroduction of the species (e.g. Pintos et al.

1999), since interspecific interactions could re-

duce the viability of newly established popula-

tions.

Acknowledgements We greatly acknowledge the assis-tance an company during the field work provided byLeonardo Fernandez, Marta Narvaez, Luis Barrios, JoseAlvarez, Eduard, Juan Cesar, Matu, Laura and Eli. Thanksalso to Miguel Delibes, Gordon H. Copp, Emili Garcia-Berthou and two anonymous referees, whose comments onearly drafts of the manuscript really improved it. Thisstudy is part of the project ‘‘Biotic integrity and environ-mental factors of watersheds in south-western Spain.Application to the management and conservation of Med-iterranean streams’’ (Ministerio de Ciencia y Tecnologıa,REN2002–03513/HID).

References

Baltz DM, Moyle PB, Knight NJ (1982) Competitiveinteractions between benthic stream fishes, rifflesculpin, Cottus gulosus, and speckled dace, Rhinich-thys osculus. Can J Fish Aquat Sci 39:1502–1511

180 Environ Biol Fish (2007) 78:173–182

123

Belk MC, Lydeard C (1994) Effect of Gambusia holbrookion a similar-sized, syntopic poeciliid, Heterandria for-mosa: competitor or predator? Copeia 1994:296–302

Bennett WA, Beittinger TL (1997) Temperature toleranceof the sheepshead minnow, Cyprinodon variegatus.Copeia 1997:77–87

Blanco-Garrido F, Sanchez-Polaina FJ, Prenda J (2003)Summer diet of iberian chub (Squalius pyrenaicus) ina Mediterranean stream in Sierra Morena (Yeguasstream, Cordoba, Spain). Limnetica 22:99–106

Brabrand A (1985) Food of roach Rutilus rutilus and ideLeuciscus idus: significance of diet shift for interspe-cific competition in omnivorous fishes. Oecologia66:461–467

Britton RH, Moser ME (1982) Size specific predation byherons and its effect on the sex-ratio of natural pop-ulations of the mosquito fish Gambusia affinis Bairdand Girard. Oecologia 53:146–151

Bystrom P, Garcıa-Berthou E (1999) Density dependentgrowth and size specific competitive interactions inyoung fish. Oikos 86:217–232

Clavero M, Blanco-Garrido F, Prenda J (2005). Size-re-lated and diel variations in microhabitat use of threeendangered small fishes in a Mediterranean coastalstream. J Fish Biol 67:72–85

Clavero M, Blanco-Garrido F, Prenda J (2006) Monitoringsmall fish populations in streams: a comparison of fourpassive methods. Fish Res 78:243–251

Crawley MJ (2002) Statistical computing. An introductionto data analysis using SPlus. John Wiley & Sons,Chichester

Doadrio I (ed) (2001) Atlas y libro rojo de los pecescontinentales de Espana. Ministerio de Medio Am-biente, Madrid

Doadrio I, Carmona JA, Fernandez-Delgado C (2002)Morphometric study of the Iberian Aphanius (Actin-opterygii, Cyprinodontiformes), with description of anew species. Folia Zoologica 51:67–79

Echelle AA, Echelle AF, Hill LG (1972) Interespecificinteractions and limiting factors of abundance anddistribution in the Red River pupfish (Cyprinodonrubrofluviatilis). Am Midl Nat 88:109–130

Fuselier L (2001) Impacts of Oreochromis mossambicus(Perciformes: Cichlidae) upon habitat segregationamong cyprinodontids (Cyprinodontiformes) of aspecies flock in Mexico. Revista de Biologia Tropical49:647–655

Garcıa-Berthou E, Moreno-Amich R (1991) New recordsof Aphanius iberus (Pisces: Cyprinodontidae) andreview of the geographical distribution of Cyprin-odontiform fishes in Catalonia (NE-Spain). ScientiaGerundensis 17:69–76

Gasith A, Resh VH (1999) Streams in Mediterranean cli-mate regions—abiotic influences and biotic responsesto predictable seasonal events. Annu Rev Ecol Sys-temat 30:51–81

Gido KB, Schaefer JF, Work K, Lienesch PW, Marsh-Matthews E, Matthews WJ (1999) Effects of red shi-ner (Cyprinella lutrensis) on Red River pupfish(Cyprinodon rubrofluviatilis). Southwestern Nat44:287–295

Gotceitas V (1990) Plant stem density as cue in patchchoice by foraging juvenile bluegill sunfish. EnvironBiol Fish 29:227–232

Harvey BC (1991) Interactions among stream fishes:predator-induced habitat shifts and larval survival.Oecologia 87:29–36

Harvey BC, Stewart AJ (1991) Fish size and depth rela-tionships in headwater streams. Oecologia 87:336–342

Kodric-Brown A, Mazzolini P (1992) The breeding systemof pupfish, Cyprinodon pecosensis: effects of densityand interspecific interactions with the killifish, Fund-ulus zebrinus. Environ Biol Fish 35:169–176

Magalhaes MF, Beja P, Canas C, Collares-Pereira MJ(2002) Functional heterogeneity of dry-season fishrefugia across a Mediterranean catchment: the role ofhabitat and predation. Freshwater Biol 47:1919–1934

Magoulick DD, Kobza RM (2003) The role of refugia forfishes during drought: a review and synthesis. Fresh-water Biol 48:1186–1198

Martin BA, Saiki MK (2005) Relation of desert pupfishabundance to selected environmental variables innatural and man-made habitats in the Salton Sea ba-sin. Environ Biol Fishes 73:97–107

McMahon TE, Tash JC (1988) Experimental analysis ofthe role of emigration in population regulation ofdesert pupfish. Ecology 69:1871–1883

Mittelbach GG (1986) Predator-mediated habitat use:same consequences for species interactions. EnvironBiol Fish 16:159–169

Oltra R, Todolı R (2000) Reproduction of the endangeredkillifish Aphanius iberus at different salinities. Envi-ron Biol Fish 57:113–115

Orth RJ, Heck KL, van Montfrans J (1984) Faunal com-munities in seagrass beds: a review of the influence ofplant structure and prey characteristics on predator–prey relationships. Estuaries 7:339–350

Ortubay S, Cussac V (2000). Threatened fishes of theworld: Gymnocharacinus bergi Steindachner, 1903(Characidae). Environ Biol Fish 58:144

Persson L (1983) Effects of intra- and interspecific com-petition on dynamics and size structure of a perchPerca fluviatilis and a roach Rutilus rutilus population.Oikos 41:126–132

Pintos R, Gutierrez-Estrada JC, Torralva M, Oliva FJ,Fernandez-Delgado C (1999) El plan de recuperaciondel fartet (Lebias ibera, Valenciennes, 1846) enAndalucıa. In: Planelles-Gomis M (ed), Peces Cip-rinodontidos Ibericos. Fartet y Samaruc. GeneralitatValenciana, Valencia

Prenda J, Armitage PD, Grayston A (1996) Habitat use bythe fish assemblages of two chalk streams. J Fish Biol51:64–79

Prenda J, Clavero M, Blanco-Garrido F (2003) Los pecescontinentales de la provincia de Cadiz. Revista dela Sociedad Gaditana de Historia Natural 3:217–232

Pou-Rovira Q, Alcaraz C, Feo C, Zamora L, Moreno-Amich R (2004) Recuperacio de l’habitat del fartet(Aphanius iberus) a la Pletera:Seguiment de la co-munitat de peixos del Ter Vell. Unpublished final

Environ Biol Fish (2007) 78:173–182 181

123

report. LIFE 99 NAT/E/006386. (http://www.torroel-la.org/siteResources/data/files/plife/Estudio_pe-ces.pdf)

Power ME (1984) Depth distribution of armored catfish:predator induced resource avoidance. Ecology65:523–528

Power ME (1987) Predator avoidance by grazing fishes intemperate and tropical streams: importance of streamdepth ands prey size. In: Kerfoot WC, Sih A (eds)Predation: direct and indirect impacts on aquaticcommunities. University Press of New England,Hannover, NH

Rincon PA, Correas AM, Morcillo F, Risueno P, Lobon-Cervia J (2002) Interaction between the introducedeastern mosquitofish and two autochthonous Spanishtoothcarps. J Fish Biol 61:1560–1585

Rogowski DL, Stockwell CA (2006) Parasites and salinity:costly tradeoffs in a threatened species. Oecologia146:615–622

Rosecchi E, Crivelli AJ (1992) Study of sand smelt(Atherina boyeri Risso 1810) population reproducingin fresh water. Ecol Freshwater Fish 1:77–85

Sanz A (1985). Lımites de hipersalinidad de los cip-rinodontidos ibericos. Donana Acta Vertebrata12:166–170

Sih A, Crowley P, McPeek M, Petranka J, Strohmeier K(1985) Predation, competition and prey comunities: areview of field experiments. Annu Rev Ecol Systemat16:269–311

Tabachnick BG, Fidell LS (2001) Computer-assisted re-search design and analysis. Allyn and Bacon, Boston

Tonn WM, Paskowski CA (1986) Size-limited predation,winterkill and the organization of Umbra-Perca fishassemblages. Can J Fish Aquat Sci 43:194–200

Valdimarsson SK, Metcalfe NB (1998). Shelter selection injuvenile Atlantic salmon, or why do salmon seekshelter in winter? J Fish Biol 52:42–49

182 Environ Biol Fish (2007) 78:173–182

123