Introduction

The biogeography of Patagonian fishes has been marked by the Andes uplift, marine incursions, and glaciations (Moyle and Cech 1982; Nelson 1994; Menni 2004; Hubert and Renno 2006). After the glacial retreat during the Pleistocene, Patagonian fishes’ ability to colonise postglacial water bodies determined their present distribution (Cussac et al. 2004; Ruzzante et al. 2006), clearly constrained by climate and, in particular, by temperature. Temperature has been recognised as one of the cues for the understanding of the biogeography of fish in Southern South America (Ringuelet 1975; Gómez 1988; 1996; Menni and Gómez 1995; Menni et al. 1996; 1998). Simultaneously and consistent with historical changes occurring in the South American transition zone (Lopretto and Menni 2003; Morrone 2004), the northern border of the Patagonian Province (Ringuelet 1975) was shifted southward by Arratia et al. (1983) and Almirón et al. (1997, Fig. 1).

Fig. 1
figure 1

Austral Subregion (shaded area) and northern limit of the Patagonian Province. This limit is indicated according to Ringuelet (1975, dotted line), Arratia et al. (1983, dashed line) and the southern limit of the transition zone of Almirón et al. (1997, solid line). Numbers indicate the main basins, Atlantic (1: Colorado, 2: Negro, 3: Chubut, 4: Santa Cruz, 5: Gallegos) Pacific (6: Hua Hum, 7: Manso, 8: Futaleufu, 9: Corcovado, 10: Engaño), Intermitent (11: Senguerr, 12: Deseado) and Beagle channel (13: Pipo)

Full size image

In a comprehensive survey, Quirós et al. (1986) and Quirós (1991) related the abundances of fish species to annual mean air temperatures. Shuter and Post (1990) discussed the potential effects of climate warming on the zoogeography of temperate freshwater fishes, assuming that the limit of distribution towards high latitudes depends on the size of the young-of-the-year necessary to minimize specific metabolic rates and maximize stored energy for the fish to endure periods of resource scarcity.

The localities for native fishes in Patagonia show a clear pattern (for example in Baigún and Ferriz (2003) and Liotta (2006)), where diversity exhibits a similar declining trend toward high latitudes, already reported for the Brazilic Subregion (Lopretto and Menni 2003). From north to south, it is possible to note the progressive disappearance of Diplomystes cuyanus Ringuelet 1965, Diplomystes viedmensis MacDonagh, 1931, Trichomycterus areolatus Valenciennes, 1846, D. mesembrinus, H. macraei, O. hatcheri and finally P. trucha. Only species of the family Galaxiidae are found in Tierra del Fuego (Cussac et al. 2004).

The invasive capacity of introduced fish is well documented (Marchetti et al. 2004a; b). Fish introductions (Welcomme 1988; Cambray 2003) are frequent and usually elicit changes in the trophic web (McDowall 2003; Reissig et al. 2006), predation on amphibians (Fox et al. 2005; Ortubay et al. 2006), and negative interactions with other fishes (Macchi et al. 1999; McDowall et al. 2001; Milano et al. 2002; McDowall 2006). The interaction between native fishes and the salmonids introduced into Patagonia (Table 1) at the beginning of the Twentieth Century as environmental (Pascual et al. 2002; Macchi et al. 1999; Milano et al. 2002; 2006) developed at the same time as environmental change (Raven 1987; Gille 2002; Munn 1996; Jansen and Hesslein 2004; Rahel 2002).

Table 1 Salmonid and native fish species present in Patagonia
Full size table

The widely introduced salmonids show a complex pattern. In northern Patagonia, a loss of diversity can be seen eastward (Pascual et al. 2007). Macchi et al. (2007) point out that stocking policies, dispersal capabilities of each salmonid species and interactions among them produced changes in local and regional abundance and distribution throughout the last 100 years. Whereas S. fontinalis was dominant until the mid-1940s (Bruno Videla 1944; Gonzáles Regalado 1945), O. mykiss became the most important salmonid species in the 1950s (Fuster de Plaza 1950). Today O. mykiss, S. trutta and S. fontinalis are the most commonly found salmonid species (Pascual et al. 2002). Another source of salmonid diversity is the recent immigration of O. kisutch and O. tshawystcha through Pacific drainages. Today, S. namaycush is exclusively located at high latitude and longitude, S. fontinalis is restricted to the Andes (higher longitude) and O. mykiss and in less extent S. trutta, are scattered throughout the Patagonian Province.

The aim of the present work includes three main objectives: (a) to compose a general and updated picture of the latitudinal distribution range of native and alien fish species, (b) to analyze the historical changes in the relative abundance of Percichthys trucha (sensu Ruzzante et al. 2006), Odontesthes sp., and salmonids in lakes and reservoirs, and (c) to relate the diversity and relative abundance of native and salmonid fishes to the environmental variables of lakes and reservoirs, in order to improve our knowledge of habitat use and our criteria for management and conservation.

Materials and methods

To characterize the fish assemblages in streams and lakes, we took information about presence/absence of species. Information for streams was limited to recent presence/absence data recorded in our own samplings and data obtained from the literature. In the same way, information about lakes came from data obtained recently, some by us. For both streams and lakes, we calculated the “zoogeographic integrity coefficient” (ZIC, Elvira 1995), which refers to the number of native species  ×  (total number currently recorded)−1, as an index of the degree to which fish populations have been invaded by introduced species. This index ranges from “1”, which is equivalent to pristine conditions, to “0”, showing the highest degree of alteration. Differences of integrity (ZIC) between rivers and lakes were analysed through the Mann–Whitney test. The different distributions of ZIC values were analysed with the Kolmogorov–Smirnov test. All statistical analyses were conducted with Statistical Package for Social Sciences (SPSS; Norusis 1986). Presence of native and alien species in Patagonian basins was visualised using the frequency of occurrence FO (%) = 100 · number of streams with presence · (number of streams sampled within the basin)−1.

The changes in the northern border of the Patagonian Province (sensu Ringuelet 1975) mainly involved lotic systems of the basins of the rivers Colorado and Negro. A set of isolated references of new localities for Brazilian fish species was considered in the Patagonian Province (Cazzaniga 1978; Ferriz and López 1987; Almirón et al. 1997; Ortubay et al. 1997; Baigún et al. 2002).

To analyze the historical changes in the relative abundance of native fishes and salmonids in lakes and reservoirs we used Quirós’ (1991) database, which included relative abundances, as proportions of salmonids, P. trucha, and Odontesthes sp. in captures for lakes sampled between 1984 and 1987. Quirós (1991) treated salmonids (including O. mykiss, S. trutta, S. fontinalis and S. salar), Percichthys (including all the nominal species of the genus) and Odontesthes (including O. bonariensis and O. hatcheri) together as single categories. Considering the results of Ruzzante et al. (2006), we considered all the nominal species of Percichthys as P. trucha. Regarding Odontesthes, the only reference to O. bonariensis southward the river Negro is that of the Ramos Mexia reservoir. In consequence, we considered that all the Odontesthes were O. hatcheri for the subsequent analysis.

We compared Quirós’ findings with data obtained recently (Table 2), some of them by us. All past and present samplings considered were performed during spring-summer surveys and include data on relative abundance (Table 2) from littoral gillnet captures using low selective mesh arrangements. Initially, we only considered lakes of Quirós’ (1991) database included within the geographic range of the most recent studies (38 to 54°S). We visualised past and present values of relative abundance by constructing bubble plots (Sigmaplot (R)). In a second step we kept only the lakes that coincided in both databases, constructed the bubble plots for relative abundances, and tested the median differences between them (Wilcoxon test on two related samples).

Table 2 Latitude, longitude, altitude and morphometry-area, perimeter, perimeter area−1 ratio (PAR) and line coast development (DL)- of Patagonian lakes (N = 99). Zoogeographic Integrity Coefficient (ZIC), presence (number of species) of nativesa, aliensb and relative abundance of P.trucha (P), salmonids (S), and Odontesthes (O) are indicated. Comparisons with abundance data of Quirós (1991)c are indicated in partitioned cases
Full size table

In order to relate the zoological integrity, diversity and relative abundance of native and salmonid fishes with the environmental variables of lakes and reservoirs, we considered the ZIC, the number of native and alien species, and the relative abundance of P. trucha, Odontesthes sp. and salmonids. The altitude, geographic position, area and perimeter were obtained from Google Earth images (http://www.earth.google.com/) processed with an image analyzer (Image Pro Plus). Areas and perimeters were also considered as line coast development (DL = perimeter · [2 · (π area)1/2]−1, Wetzel 1981) and as perimeter · area−1 ratio (PAR). PAR and DL reflect the development of the littoral zone, nutrient input, macrophyte abundance and shelter availability. The association between fish assemblage characteristics (ZIC, diversity, and abundance) and geographic and environmental variables was treated using Canonical Correspondence Analysis (CANOCO 4.5, ter Braak and Smilauer 1998).

Results

River and lake assemblages

The ZIC data (Tables 2 and 3) revealed that many more lakes than streams were sampled. In addition, there are basins whose streams have been better sampled than others due to geographic or human constraints.

Table 3 Patagonian streams (N = 56). Basin, Zoogeographic Integrity Coefficient (ZIC) and presence of nativesa and aliensb fishes
Full size table

Rivers showed lower integrity than lakes (Mann–Whitney test, n = 154, P < 0.002) and a different distribution of ZIC values (Kolmogorov–Smirnov test, n = 154, P < 0.004), unimodal in lakes and with three modes in rivers. Salmonids were always strongly present both in lakes and streams. Rainbow trout was the most frequent among salmonids. Galaxias platei and P. trucha were the most widespread native species (Fig. 2).

Fig. 2
figure 2

Frequency of occurrence (percentual, FO (%)) of native (top panel, At: A. taeniatus, Az: A. zebra, Ci: Cheirodon interruptus, Cp: Corydoras paleatus, Dm: D. mesembrinus, Dv: D. viedmensis, Ga: G. australis, Gm: G. maculatus, Gp: G. platei, Hm: H. macraei, Jm: J. multidentata, Ob: O. bonariensis, Oh: O. hatcheri, Oj: O. jenynsi, Pt: P. trucha, Ta: T. areolatus) and alien species (bottom panel, Cc: C. carpio, Om: O. mykiss, Ot: O. tshawystcha, Sf: S. fontinalis, Sn: S. namaycush, Ss: S. salar, St: S. trutta) in the Patagonian basins (ordered by increasing latitude, N = number of streams sampled within the basin)

Full size image

We observed a conspicuous overlap of specific localities for Austral, Brazilic and Marine species (Table 4) along the basins of the rivers Colorado and Negro. Before Ringuelet (1975), the following species composition existed (excluding the exotic species of Salmonidae introduced since 1904, see Pascual et al. 2002) 2 Brazilic (Gymnocharacinus bergii Steindachner, 1903, Jenynsia multidentata Jenyns, 1842), 3 Austral (D. viedmensis, P. trucha and Galaxias maculatus (Jenyns, 1842)), and 1 Andean (D. cuyanus). Since the general scheme of Ringuelet (1975), new localities for Brazilic, marine and non-salmonid exotic species in the Austral Subregion have been noted. The new records were: 7 Brazilic (Astyanax eigenmanniorum Cope, 1894, Cheirodon interruptus Jenyns, 1842, Oligosarcus jenynsii Günther, 1864, Corydoras paleatus Jenyns, 1842, Cnesterodon decemmaculatus (Jenyns, 1842), J. multidentata-a new southern record, and O. bonariensis); 4 Austral (Hatcheria macraei (Girard, 1855), T. areolatus, Galaxias platei Steindachner, 1898, O. hatcheri); 3 marine (Odontesthes argentinensis (Valenciennes, 1835), Mugil liza Valenciennes, 1836, Paralichthys brasiliensis (Ranzani, 1842)), and 1 exotic species (Cyprinus carpio Linnaeus, 1758), introduced into the south of the Brazilic Subregion and arriving at the Austral Subregion with no known means of dispersal. Thus, we considered a total of 8 Brazilic, 7 Austral, 1 Andean, 3 marine, and 1 exotic species, summing a total of 20 species (Table 4).

Table 4 Specific localities (first records) for Austral (A), Brazilic (B), Andean (AN), and Marine (M) species at the northern border of the Austral Subregion. Over 20 species, 40% are Brazilic and 37% Austral. Localities considered new records after the general scheme of Ringuelet (1975) for Brazilic (B), Andean (AN) and Marine (M) species are indicated with capture date (CD), latitude and longitude. The exotic species Cyprinus carpio (E) was also considered
Full size table

Some of the new records reveal established populations with a high number of individuals captured, such is the case of J. multidentata, A. eigenmanniorum, O. jenynsii, C. carpio, and M. liza (Almirón et al. 1997). The “new record" condition of J. multidentata deserves additional explanation. This species was already recorded in the rivers Colorado (in 1916) and Negro (in 1967). However, new records (1987 and 1997) confirm a southward displacement (from 40 to 41°S).

In addition to the new localities for Brazilic and marine species at the northern border of the Austral Subregion, new localities for Austral species already cited in the northwest of the Austral Subregion were also found southward of their known distribution range: H. macraei (at Jeinimeni and Ecker rivers) and T. areolatus (in the Negro, Tecka and Lepa rivers) (Almirón et al. 1997; Baigún and Ferriz 2003).

Historical changes in fish abundances

In lakes, the graphs for the relative abundances of salmonids in the area common (38 to 55°S) to the databases of Quirós (1991, n = 42) and our own present databases (n = 44) showed, at first view, a similar situation regarding distribution and relative abundance (Fig. 3). However, comparing these databases restricted to common lakes (n = 18, Table 2), we observed that the relative abundance of salmonids decreased (Wilcoxon signed-ranks test, n = 18, P < 0.001, Fig. 4) and P. trucha increased (Wilcoxon signed ranks test, n = 18, P < 0.001, Fig. 5). It must be noted that although the relative abundance values are linked, there is variation within native fishes since changes in silverside abundances were not significant (Wilcoxon signed ranks test, n = 18, P > 0.68). Among these 18 lakes and reservoirs, five lakes (Gutiérrez, Mascardi, Steffen, Yehuin, and Escondido) showed no changes for 100% of salmonids. However, we must note that only salmonid populations in littoral gillnet captures were considered (the small G. maculatus is not captured by gillnets and G. platei dwells in the deep bottom, Table 2).

Fig. 3
figure 3

Relative abundance (bubble size indicates the % of total capture) for salmonid populations of lakes and reservoirs (from 38 to 54° S), according to the database of Quirós (1991) for years 1984–1987 (left, blue circles, n = 42) and recent samplings (right, red circles n = 44). Crosses indicate absences

Full size image
Fig. 4
figure 4

Bubble plot (size indicates the % of total capture, top panel) and line plot (bottom panel) for relative abundance of salmonid populations of lakes and reservoirs (n = 18, ordered by latitude from 38 to 54° S) common to the database of Quirós (1991) (left, blue circles) and recent samplings (right, red circles). Big crosses indicate unchanged values. Small crosses indicate absence or values lower than 10% (see Table 2 for details)

Full size image
Fig. 5
figure 5

Bubble plot (size indicates the % of total capture, top panel) and line plot (bottom panel) for relative abundance of P. trucha populations of lakes and reservoirs (n = 18, ordered by latitude from 38 to 54° S) common to the database of Quirós (1991) (left, blue circles) and recent samplings (right, red circles). Big crosses indicate unchanged values. Small crosses indicate absence or values lower than 10% (see Table 2 for details)

Full size image

Spatial distribution patterns in abundances and diversity

The relationship between relative abundances of species and environmental variables was significant (Monte Carlo test, n = 44, F = 20.9, P < 0.001) and explained (the first two axes) the 100 % of the variance. The CCA revealed an appreciable separation among the relative abundances of P. trucha, Odontesthes sp. and salmonids in relation to the environmental variables, along the two canonical axes (λ1 = 0.193, λ2 = 0.033). Latitude, longitude and area of lakes were significant in the explanation of the gradient of relative abundances (Table 5). In Fig. 6, we could see that the high abundances of salmonids were related to high latitudes and longitudes and lakes smaller than those where the abundances of Odontesthes sp. and P. trucha were higher. Odontesthes sp. had its higher abundance at lower longitudes and P. trucha at lower latitudes.

Table 5 Forward selection of geographic and environmental variables to determine their importance (Lambda-A) in explaining the abundance (relative abundance of salmonids, P. trucha and O. hatcheri) and diversity (number of native and alien species, and ZIC) variables
Full size table
Fig. 6
figure 6

First two axes of the canonical correspondence analysis for abundances of P. trucha, Odontesthes and salmonids populations, geographical (latitude and longitude) and morphometric (area) variables in lakes and reservoirs (circles) of Patagonia. Only significant variables are indicated

Full size image

The relationship between diversity variables (number of native and alien species and ZIC) and environmental variables was significant (Monte Carlo test, n = 99, F = 3.38, P < 0.007) and explained (the first two axes) the 100% of the variance. The CCA revealed an appreciable separation among diversity variables in relation to the environmental ones, along the two canonical axes (λ1 = 0.013, λ2 = 0.001). Only latitude and PAR were significant in the explanation of the gradient of diversity variables (Table 5). In Fig. 7, we could see that the higher number of alien species was more related to lower latitudes and lower PAR than the high number of native species. The ZIC was mostly associated with high latitudes. However, the meaning of ZIC was constrained by the simultaneous change of alien and native diversity. Although not significant (P > 0.06), high DL resulted strongly associated with high number of native species and greater longitude with a high number of alien species.

Fig. 7
figure 7

First two axes of the canonical correspondence analysis for number of native (Natives) and alien species (Aliens), ZIC, geographical (Latitude, Longitude and Altitude), and morphometric variables (DL and PAR) in lakes and reservoirs (circles) of Patagonia

Full size image

Discussion

Salmonid and native assemblages

The data about the ZIC in lakes and streams are limited due to the varying sources of information. In this sense, data have been reported by sport anglers and divers; dead fish have been observed by rangers, and information has been gathered in scientific studies. While there have been multiple efforts to survey fish in lakes, river surveys have been rare and sketchy. However, the resulting ZIC has a clear consistency. The analysis points to a variable impact of salmonids on lakes, ameliorated by the availability of littoral refuges (Cussac et al. 1992; Barriga et al. 2002, Buria et al. 2007), and a major impact on streams, where salmonids (in particular O. mykiss) seem to have displaced the native fishes almost completely. Stream records with significant captures of H. macraei, D. viedmensis, G. maculatus or P. trucha nowadays seldom occur (Barriga et al. 2007). The causes involved in the generation of a salmonid-rich or -poor stream (Allouche 2002), together with the role of rising temperature (Dunham et al. 2003; Wehrly et al. 2003), have just begun to be studied in Patagonia (Habit et al. 2007). In all cases, the impact is notorious when comparing the situation in Patagonia with that of heavily populated areas such as Greece (ZIC = 88), Italy (ZIC = 56), Portugal (ZIC = 65), and Spain (ZIC = 63, Elvira 1995).

Changes in fish distribution

A dispersion of Brazilic, Andean and marine populations into the Austral Subregion was observed, as well as a southward movement of northernmost Austral species. While the movement of northern species into Patagonia appears as a likely scenario, the comparison of historical and modern records has the weakness of comparing poor historical records and more intensive recent sampling. However, it should be noted that no new record for Austral species within the Brazilic Subregion was found in the literature and data reviewed. In addition, the observed increase (300%, from 2 to 8 species, excluding J. multidentata) in the number of Brazilic species is far greater than the increase in Austral species (133%, from 3 to 7 species) which is an expected increase from better sampling.

In addition to the introduction of salmonids, the last century witnessed major artificial changes involving damming, canal construction, water extraction (Almirón et al. 1997; Gómez et al. 2004b), deforestation, and the consequently increased rainfall (Hoffmann 1989; Dyer 2000). Currently, we have the first evidence of a complex environmental change, with multiple causes, contemporary with native-exotic interactions. Artificial changes to the landscape (canal construction and weirs) obviously facilitate the movement of biota out of their natural range. In addition, an obvious man made fish transport could be observed in the sale of bite fish (Alvear et al. 2007). However, the potential for such landscape changes and transport to cause range expansion in the absence of climatic change is not clear. For example, Dyer (2000) noted that the Atacama Desert area of northern Chile and southern Perú between the rivers Loa and Rimac, previously considered “empty" (Ringuelet 1975; Arratia et al. 1983; Arratia 1997), is at present inhabited by the Atherinopsidae Basilichthys semotilus (Cope, 1874) and the Trichomycteridae Trichomycterus punctulatus Valenciennes, 1846. Similarly, Hoffmann (1989) reported an important change in the position of the 800 mm isohyets before and after 1959 in the south of the Brazilian Subregion. During 2000, new wetlands with nine species of Brazilian fishes were recorded there, in the formerly called “pampeana” dry zone (sensu Canevari et al. 1998). These new locations were the consequence of an increase in average annual rainfall and the construction of new artificial drainage channels, allowing the rapid dispersion of fish into an ecophysiologically suitable range (Gómez et al. 2004a, 2004b). The southern limits of the distribution of two Brazilian species—O. bonariensis and the Pimelodidae Rhamdia quelen (Quoy and Gaimard, 1824)—are clearly related to their tolerance to low temperature (Gómez 1988, 1990, 1996). In addition, Gómez et al. (2004b) observed new southernmost localities for these Brazilian fishes. New records of the Serrasalmidae Serrasalmus spilopleura Kner, 1858, found southwards of its known distribution range, have been published by Gómez et al. (2004a), and new records of two Brazilic species (from a total of 12) in the southern Brazilic Subregion (38°S) have been reported by Casciotta et al. (1999).

Regarding abundance of native fishes and salmonids in lakes and reservoirs, the link established by the relative abundance data between the different species cannot be eliminated, however some punctual data could improve our comprehension. For example, in the Lake Laguna Blanca the records of Quiros (1991) showed near 50% of salmonids and 50% of P. trucha in 1984–1987 samplings. After 20 years, capture of P. trucha was the highest recorded in all Patagonian lakes and reservoirs, and salmonids were nearly undetectable (Ortubay et al. 2006). In the same way, the results of Alonso (2003) and Vigliano and Alonso (2007), expressed as caught per unit effort, signaled a significant decrease in the abundance of wild salmonid populations in three reservoirs in the Limay river basin.

The decrease of salmonid abundance in lakes and reservoirs could have different causes. One is a pioneer effect and its consequent stabilization (Macchi et al. 2007). Another possibility is that, considering we are working with littoral captures, the decrease of relative abundance of salmonids could be another example of the exclusion of salmonids from the littoral zone observed by Jansen and Hesslein (2004) in relation to an increase in water temperature at lake shores.

The knowledge about the responses of fish species to habitat heterogeneity in multiple scales can be used for management purposes, conservation and restoration (Ferreira et al. 2007). We can expect that the intralacustrine and between-lakes distributions of fish populations change even at spatial and geographical scales. Our results agree with the pattern found by Quirós (1991) regarding the relationship between abundance, latitude and temperature. Most of the geographic and morphometric variables explained fish abundance and diversity. Particularly, abundance showed mainly geographical cues and the diversity relied largely on morphometric characteristics. The cues of abundance and diversity seem to have a common point in the lake area, included into the PAR concept. Following Quirós (1991), the coexistence of salmonids and native populations mainly depends on the existence of multiple habitats, allowing negative interactions to be minimised. Native abundance and alien diversity were negatively related with latitude. The PAR, and to a less extent the DL, showed greater native diversity in lakes with high PAR.

Diversity seems to have a strong relationship with the morphometry of the lake. Pascual et al. (2007) found that abundance, diversity and even the existence of fish populations are related with the lake and shallow water bodies connected to deeper lakes. Most of the literature concerning Patagonian fishes suggests that the interaction between salmonids and native species mostly takes place in the littoral zone (Macchi et al. 1999; Quirós 1991; Ruzzante et al. 1998, 2003; Milano et al. 2002, 2006). The coexistence between salmonids and native fishes has mainly benefited from the spatial and temporal segregation of breeding habitats; streams during autumn-winter for salmonids, and lake’s littoral zone during spring-summer for native fishes (Cussac et al. 1992; Cervellini et al. 1993; Barriga et al. 2002, 2007; Buria et al. 2007). Macchi et al. (1999) showed that salmonids and P. trucha share benthic food resources and also predation on Galaxiidae species. These shared roles have been confirmed in several studies addressing fish diets in Patagonia (Cussac et al. 1998; Ruzzante et al. 1998, 2003; Logan et al. 2000; Milano et al. 2002, 2006; Ferriz 1984, 1987, 1988, 1989, 1993/94, 1994).

Climatic relationships

The climate trends regarding southern South America provide some relevant data. One is the two-degree (Celsius) increase in the mean annual air temperature over the last century in the South Orcadas Islands (60°45′ S, 44°43′ W, Servicio Meteorológico Nacional 2007). In the last decade, the increase has been 0.2°C (Servicio Meteorológico Nacional 2007). The exclusion of salmonids from the littoral zone due to an increase in water temperature at lake shores (Jansen and Hesslein 2004) could benefit P. trucha and could adversely affect salmonids (Elliot 1981), at least according to preliminary data on thermal tolerances and preferences (Ortubay et al. 2004, Cussac et al. 2005, Aigo et al. 2006) and the data of Quirós (1991) and Quirós et al. (1986).

The present situation features an Austral fish fauna (Ringuelet 1975; Arratia et al. 1983; Almirón et al. 1997) interacting with salmonids from the beginning of the 20th Century, and suggests that major artificial changes plus a detectable climate change, are probably at the root of a change in the composition and relative abundances of fishes in the assemblages. The result of the new interactions is a highly dynamic situation, hardly predictable and one that should be carefully observed in the future. Particularly, the importance of the heterogeneity of the littoral zone (Wei et al. 2004; Lewin et al. 2004) is awaiting further studies in Patagonia in relation to the relative abundance of Salmonidae and native fishes.

Conclusion

Although other factors like geological history, population dynamics, and interspecific interactions could affect native and alien fish distribution (Ferreira et al. 2007), we could find patterns for abundance and diversity clearly related with the development of the littoral zone. Our results agreed with previous literature regarding the geographical pattern of native and alien fish abundances and with the importance of the lake littoral zone for the conservation of native diversity. Description of geographical patterns for abundance and diversity and historical changes, like southward dispersion and abundance changes, is a useful tool not only for research but also for future management design of Patagonian fish populations.