Long term changes in the fish fauna of Lago de Pátzcuaro in Central México

Fish assemblages of lakes from central México have been altered as a result of water overexploitation, pollution, habitat fragmentation, and introduction of exotic species. Patterns and the extent of change in most of these systems are poorly documented. This paper presents a long-term study of changes in the fish fauna of Lago de Pátzcuaro (State of Michoacán, México). Long-term information (years 1900-2010) of fish communities was used to explore trends in the composition of the fish community and their ecological attributes. Trends across time for each fish guild were evaluated through a multivariate analysis of variance. Native species dominated total species richness over the entire study period, but exotic fish comprised 46% of the community in the 2010s. During the last 110 years, there has been a decline in the number of carnivore species and an increase of omnivore species in the Lago de Pátzcuaro fish community. Sensitive species have disappeared from the lake and species tolerant to environmental degradation have increased from 13 to 61%. The data suggest that anthropic effects on the environmental condition of Lago de Pátzcuaro are associated with the loss of ecosystem elements (fish species) and ecosystem processes (species interactions). Ecological restoration and conservation programs must include information of these long term studies to be more effective in their efforts.


INTRODUCTION
Mexican freshwater ecosystems have been deteriorating since pre-Hispanic times, however this process has accelerated at the beginning of the 20th century, and worsened in the last 30 years (Lyons et al., 1998;Soto-Galera et al., 1999).The main causes of freshwater ecosystem degradation are deforestation, excessive sedimentation, the introduction of exotic species and discharge of urban and industrial waste water (Peters et al., 2005).These activities have altered the biotic community structure and function, decreased food web length, and favoured reduction of organism's maximum size and species diversity (Karr, 1981;Lyons et al., 1998).Other changes experienced by fish communities include the increase of exotic species, local extinction or near-extinction of several native species, and the collapse of many fisheries (Aloo, 2003).
Lago de Pátzcuaro is located in the highlands of central México and has been subjected to several hydrological modifications: erosion and runoff from deforested uplands, loss of surface area, reduction of the water column (approximately 6 m and 166 Mm 3 in the past 70 years), lower water transparency (approximately 2 m in the past 40 years), hyper-eutrophication, and point source contamination from local communities throughout the lake (Alvarado-Díaz et al., 1985;Chacón-Torres, 1993;Berry et al., 2011).Lago de Pátzcuaro supports one of the best-studied, for the last three decades, inland fisheries in México; many of its endemic species have been intensively fished since long before the colonial times (De Buen, 1941;Orbe-Mendoza & Acevedo-García, 1995;Orbe-Mendoza et al., 2002).However, fishery production from this lake has shown major fluctuations over the last 20 years and it is currently in sharp decline (Berry et al., 2011).Many of its fish populations are on the verge of collapse (Orbe-Mendoza et al., 2002).
Lago de Pátzcuaro is one of the few lakes in México with a relatively long series of fish sampling records.Earliest data of the fish community from this lake are available from the 1900 decade and sampling efforts have continued to present times, allowing for a long-term study.To date there has not been an analysis of the long-term changes of the fish communities in the lake.Historical analyses of changes in fish communities offer information about the current conservation status of aquatic ecosystems and surrounding watershed conditions (Karr, 1981;Fausch et al., 1990).Analyses of this kind can detect the anthropic impacts that influence the fish community, and are useful to identify elements and ecosystem processes lost as a consequence of human activities (Scott & Hall, 1997;Lyons & Mercado-Silva, 1999).Further information on the nature of the impacts can be obtained from the ecological attributes of individual species (Karr, 1981;Lyons et al., 1995).Ecological studies on fish assemblages can integrate and evaluate a variety of impacts (water quality, habitat structure, species interactions) on biotic systems (Hobbie et al., 2003;Mercado-Silva et al., 2006).Species origin, feeding preferences and species tolerance to environmental degradation are fish attributes to consider when analysing community change and understanding cumulative anthropic impacts (Lyons & Mercado-Silva, 1999;Mercado-Silva et al., 2006).
This paper presents a long-term study of changes in the number of species (as indicators of the condition of ecosystem elements) and three ecological attributes (as indicators of the status of ecosystems processes) of the fish community of Lago de Pátzcuaro.The study provides a broad perspective that can help our understanding of general trends in fish communities of tropical lakes with similar conditions of environmental degradation.

Study area
Lago de Pátzcuaro (Fig. 1) is part of an endorheic drainage basin in the highlands of central México (19°32′N-19°42′N, 101°32′W-101°42′W) in the State of Michoacán.With a maximum surface area of 116 km 2 , Lago de Pátzcuaro is the third largest natural lake in México (Gomez-Tagle et al., 2002).It has a maximum depth of 12 m at maximum capacity and an average depth of 5 m.Water volume is determined by precipitation, evapotranspiration, and inflows from springs at the bottom of the lake and flows from small creeks in the watershed (Bernal-Brooks et al., 2002).The lake has experienced a 6 m water level decline since 1939 (Berry et al., 2011).Lago de Pátzcuaro is circadiomictic and is generally considered eutrophic ( Berry et al., 2011).There are 26 urban areas surrounding the lake including one densely populated island (Isla Janitzio) in the southern part of the lake.The lake watershed covers 9,340 km 2 and includes land used for agriculture and livestock production (~40% of the watershed area), forestry (~30%), and urban activities (~22%) (Bravo-Espinosa et al., 2006).Erosion caused by agriculture and ranching is high in some parts of the basin.High erosion rates are reflected in the rates of sedimentation (100,000 m 3 each year) in the lake (Gomez-Tagle et al., 2002).

Study stages
This study was divided into two stages.First, we carried out a review of the historical data.Second, we made sampling efforts at the lake.

Long term data from collections
Historical fish collection information was obtained from scientific collection records at the University of Michigan Museum of Zoology (UMMZ), Universidad Michoacana de San Nicolás de Hidalgo (UMSNH), Instituto Politécnico Nacional (IPN) and data from personal collections by J. Lyons (University of Wisconsin-Madison Department of Zoology, pers.comm.).These sources comprise the majority of sampling records for Lago de Pátzcuaro.We also gathered information from records in Berlanga-Robles et al. (2002).We used species composition data for one year in the first decade of the 1900s (UMMZ) and for at least three years per decade from 1930s (UMMZ), 1950s (UMMZ), 1990s (J.Lyons 1920s, 1940s, 1960s, 1970s and 1980s.From each collection record, the taxonomic identity of the species and the number of species were obtained.Data from any location within the Lago de Pátzcuaro were considered.For most records no information on sampling effort or gear type was available, and we had to assume that sampling efforts adequately sampled the entire fish community.Because of this limitation, only presence/absence data were considered for analysis following Mercado-Silva et al. (2006).
We include the Lago de Pátzcuaro salamander Ambystoma dumerilli (Ambystomatidae) in the analysis because it is functionally equivalent to a fish in the ecosystem (Mercado-Silva et al., 2002).A. dumerilli is neotenic and completely aquatic throughout its life.It is an environmentally sensitive and a carnivorous species (Huacuz, 2002).The information about the long-term presence of this amphibian species in Lago de Pátzcuaro was taken from Huacuz (2002).

Recent sampling efforts
To assess the current condition of the fish community in Lago de Pátzcuaro we sampled six sites containing different types of fish habitat in four lake zones defined by Alcocer & Bernal-Brooks ( 2002): (1) San Jerónimo (SAJ) is a littoral zone (max.10 m from the point of fish capture to the shore) located in the northern portion of the lake (mean depth = 1.4 m); (2) Pacanda Island (PAC) is a limnetic zone (more than 30 m from the point of fish capture to the closest shore) adjacent to an island in the center of the lake (mean depth = 6 m); (3) Ucasanastacua (UCA) is a littoral zone located halfway between the northern and southern most sections of the lake (mean depth = 2 m); ( 4) Napízaro (NAP) is a limnetic zone located in the southern portion of the lake (mean depth = 2.5 m); (5) Ihuatzio (IHU) is a littoral zone located in the southern portion of the lake (mean depth = 1.9 m); ( 6) Embarcadero (EMB) is a littoral zone located in the southern portion of the lake, adjacent to the Pátzcuaro City (mean depth = 1 m) (Fig. 1).
Samples from each site were collected in September and November 2009 (wet season), and February and June 2010 (dry season).Two different seines were used to obtain representative samples of the fish community, increasing the likelihood of collecting species of small and large size.Seining has been traditionally used by local fishermen to capture fish and Ambystoma dumerilli in the same sites used in this study (Huacuz, 2002;R. Quirino, Lago de Pátzcuaro fisherman, pers. comm.).A seine 150 m long, 9 m deep, with 4 cm mesh was used to capture relatively large fishes.A smaller seine 75 m length, 7 m deep, with 1 cm mesh was used to capture small fishes.Each net was launched three times at each site for a total of 144 sets in the study.Each seine launch was carried out in a different area in each site.Captured specimens were counted to assess the absolute abundance (number of individuals), the abundance per species, and frequency (percentage of sets where a species was encountered).Fish were identified to species except individuals in the genera Oreochromis and Chirostoma because of difficulties in their identification.At least two species of Oreochromis were introduced to Lago de Pátzcuaro, blue tilapia O. aureus and Nile tilapia O. niloticus (Berlanga-Robles et al., 1997, 2002;Gaspar-Dillanes et al., 2006).Hybridization has occurred between Oreochromis species, making positive identification at species level difficult.Among Chirostoma, taxonomic differences between four native species, bigeye silverside Ch. grandocule, Patzcuaro silverside Ch. patzcuaro, slender silverside Ch. attenuatum, and Pike silverside Ch. estor and one introduced species, shortfin silverside Ch. humboldtianum are currently in dispute (Barriga-Sosa et al., 2002;Bloom et al., 2009).Many studies argue that relatively recent hybridization among Chirostoma species has occurred in Lago de Pátzcuaro, making species identification difficult and obscuring taxonomic differences among the four native and one introduced species (Ledesma-Ayala & García de León, 1991; Oseguera-Figueroa & García de León, 1991;Soria-Barreto & Paulo-Maya, 2005).In the remainder of the text and analyses however, we consider all individuals in Chirostoma as native.

Analysis
According to historical records and after identification in the field (for our field efforts), each fish and amphibian were assigned to their correspondent guild based on their origin (native or exotic to the basin), feeding habits (carnivorous, omnivorous) and tolerance to pollution or habitat deterioration (tolerant, sensitive), following published classifications (Lyons et al., 2000;Mercado-Silva et al., 2002) (Table 1).These attributes were selected because they can inform about several aspects of the conservation of freshwater ecosystems.The persistence of natives and the presence of exotics are related with the conservation status of geologic, biogeographic and evolutionary processes (Harig & Bain, 1998;Aparicio et al., 2000;Olden & Poff, 2005).The maintenance of certain proportions of carnivores and omnivorous over a long term indicates that ecological processes associated with ecosystems functional conditions are preserved (Karr, 1993;Scott & Hall, 1997;Dobson et al., 2006).The proportion of tolerant and sensitive species can be related to the levels of degradation in a lake (Karr, 1981;Whittier & Hughes, 1998;Meador & Goldstein, 2003;Seilheimer & Chow-Fraser, 2006).
Long term data samples  often had relatively vague geographical descriptions.We were able to identify that they were obtained within the lake, but rarely was there a within-lake specific location established.Therefore, samples were assumed to be representative of the entire lake.Information from collections taken in any year within a decade (regardless of the specific site) was summarized to produce an average value for each fish guild per decade between 1900 and 2010.To explore the variation of fish guilds over time, we used the number of species and average value for each guild gathered in a decade.The analysis combines species information for each of the six guilds for the following decades: 1900s, 1930s, 1950s, 1990s, 2000s and 2010s.Data were standardized using a square-root transformation (√n) following (Clarke & Gorley, 2006), to balance the contributions of dominant and rare groups.Trends across decades for each fish guild were analysed through a multivariate analysis of variance.A Table 1.Fish guilds, collection decades and presence-absence of each taxon in Lago de Pátzcuaro.For trophic guild (TG): O: omnivore, C: carnivore, H: herbivore.For origin (OR): N: native, E: exotic.For tolerance (TOL): S: sensitive, M: moderately tolerant, T: tolerant, *Amphibian included in the study, 1: presence, 0: absence.

Species OR TG TOL 1900s 1930s 1950s 1990s 2000s 2010s Cyprinidae
Algansea lacustris posteriori pairwise comparison analysis was used to search for differences among decades.Because our data do not fulfill the assumptions of normality, homogeneity of variance, or independence, we decided to use an analysis based on permutation techniques in PERMANOVA (Anderson, 2001) from PRIMER 6 (Clarke & Warwick, 2001).Ctenopharyngodon idella) species that had been historically collected were not found.Species tolerant to environmental degradation comprised most of the samples; no sensitive species were captured.Moderately tolerant fish captured include A. lacustris, A. diazi, Chirostoma spp., A. robustus, and the exotic M. salmoides.Five species were carnivorous: four native (A.lacustris, A. robustus, A. diazi, Chirostoma spp.) and one exotic (M.salmoides).All other species were omnivorous (Table 1).

Comparisons along time
All taxa encountered in samples taken in the decade of 1900s were native to Lago de Pátzcuaro.By the 1930s one species, Micropterus salmoides, had been introduced.By the 1990s, four other fish taxa were introduced (Cyprinus carpio, Ctenopharyngodon idella, Poeciliopsis infans, and Orechromis spp.).By the 2000s, two native species -one fish (Skiffia lermae) and one amphibian (Ambystoma dumerilii), and one exotic species (Micropterus salmoides) were not recorded (Table 1).
From 1900 to 2010, the number of native species changed from 8 to 3.3 ± 1.5, a reduction of 46% in native species richness.Exotic species increased from 0 to 3.6 ± 1.1 in the same period.During the same time frame, the number of carnivorous species went from 6 to 2.6 ± 1.2, a decrease from 75 to 39% of the total number of collected species.The number of omnivorous species increased from 1 to 4 (13 to 61% of the total number of collected species) and sensitive species disappeared (38 to 0% of the total number of collected species).The number of tolerant species increased from 1 to 4 (13 to 61% of the total number of collected species) in the same period (Fig. 2).

DISCUSSION
Shallow lakes are considered the most threatened ecosystems as a consequence of human fragmentation of habitats, introduction of exotic species, and changes in land use (Miller et al., 1989;Dudgeon et al., 2006).Historical research on changes in ecosystem elements and processes help understand the effects of anthropogenic threats and provide specific information that can be used to reduce further ecosystem degradation (Karr, 1981;Mercado-Silva et al., 2006).Data about long-term changes can aid in developing monitoring strategies and future conservation efforts (Olden & Poff, 2005;Mercado-Silva et al., 2006).Although anthropic effects on the watershed and the water quality of Lago de Pátzcuaro are relatively well documented, this study represents the first effort to understand the long-term changes of the fish assemblages in this lake.
The origin of fish fauna of Lago de Pátzcuaro is biogeographically associated with tectonic and volcanic events that occurred in the late Tertiary and begining of the Pleistocene in Central region of México, during which a great number of endemic species were generated in families such as Petromizontidae, Cyprinidae, Ictaluridae, Poecilidae, Goodeidae and Atherinopsidae (De Buen, 1947;Barbour, 1973aBarbour, , 1973b;;Barbour & Miller, 1976).The fish fauna and ecosystem of Lago de Pátzcuaro have experienced changes in their ecological characteristics coinci-dentally with the increase of cumulative and combined human exploitation of the lake and the surrounding drainage basin.Native, sensitive carnivorous species have decreased and the number of exotic, tolerant, omnivorous species has increased.These changes suggest that Lago de Pátzcuaro has lost ecosystem elements (fish species), ecosystems processes (species interactions), and perhaps habitat diversity.
Lago de Pátzcuaro, like many others in Central México, shows ecosystem changes related to long period of cumulative anthropic effects.The surface area of the lake has decreased by more than 30 km 2 since 1970s.Water depth has decreased >5 m, water volume has lost >150 Mm 3 and water transparency has decreased >3 m over the same period (Bernal-Brooks et al., 2002;Orbe-Mendoza et al., 2002).Fisheries were one of the most important economic activities during the earlier (1900s to 1960s) decades (Berry et al., 2011) but can no longer sustain fishermen livelihood.Currently there are no full time active fishing communities in the lake.Most local fishermen now occasionally fish for supplemental food during certain seasons and complement their diets with farm animals (R. Quirino, Lago de Pátzcuaro fisherman, pers.comm.).These landscape and human dimension changes in the history of Lago de Pátzcuaro fishery reflect our description of decline in the fish communities over 110 years, between 1900 and 2010.The yield of all fisheries in the lake have declined from ~2500 ton in 1988 to <50 ton in 2007 (Diario Oficial de la Federación, 2010).
Species richness is widely employed as a tool in ecological assessments because it has positive correlations with measurements of habitat and water quality (Scott & Hall, 1997;Gotelli & Colwell, 2001).In Lago de Pátzcuaro several native species have declined in numbers and this could be associated with alterations to the ecosystem.The absence of Ambystoma Skiffia lermae and Allotoca dugesii in our collections and the few individuals captured of Allotoca diazi, Algansea lacustris and Allophoorus robustus may indicate ecosystem degradation via the loss of their reproductive habitat and alterations to their food resources.This coincides with Soto-Galera et al. (1999) and Lyons et al. (2000) who argued that species of subfamily Goodeinae are sensitive to changes in trophic webs and habitat degradation.According to Anderson (1972) species of Ambystoma genus are also highly sensitive to habitat loss.Some authors argue that Skiffia lermae and Allotoca dugesii are locally extinct in this lake (Berlanga-Robles et al., 2002).The trends we found in Lago de Pátzcuaro resemble those found by Mercado-Silva et al. (2006) who showed that, after serious degradation, the number of native species declined greatly in lentic and lotic aquatic ecosystems of the Lerma-Chapala Basin in central México, over a period of approximately 50 years.
Untreated sewage inputs to the lake and water hyacinth trituration in parts of Lago de Pátzcuaro have caused tremendous input of oxygen-demanding wastes that degrade water quality, changing productivity (Xie et al., 2004).This productivity changes affect biotic interactions, which in turn provides advantages to omnivorous fishes (Karr, 1993;Vanni et al., 2005).For example, exotic, small-sized, and short-lived species (Oreochromis spp.) were introduced to Lago de Pátzcuaro in the 1970s (Rosas, 1976), and their persistence through the years is due in part to its capacity to resist alterations of aquatic habitats and water quality (Harig & Bain, 1998).Additionally, the common carp (C.carpio) was introduced in 1974 for commercial aquaculture (Rosas, 1976) and has subsisted in the lake probably because it can use highly diverse sources of food (Zambrano & Macías-García, 2000).In addition, these two fish taxa (Oreochromis spp.and C. carpio) can modify entire habitats and compete for feeding resources with native fish species (Zambrano et al., 2006).
A somewhat different situation has occurred with another exotic species, M. salmoides, which requires relatively good water and habitat quality and can be useful as an indicator of environmental degradation.Micropterus salmoides is a top predator in Lago de Pátzcuaro.Before 1990, the lake contained two pelagic piscivores, one native (Ch.estor) and one exotic (M.salmoides) (Solorzano-Preciado, 1963;García de León & Pérez-Velazco, 1996).Today, fisheries of both species have collapsed as a consequence of habitat degradation and overfishing (Gaspar-Dillanes et al., 2006;Berlanga-Robles et al., 2002).Fish at the top of the food chain are usually the first to disappear since they depend on a stable food web (Domínguez-Domínguez et al., 2008), and relatively high water transparency for selecting food (Fausch et al., 1900;Simon & Lyons, 1995;Lyons et al., 2000) and abundant fish and benthic invertebrates as feeding resources (Godinho & Ferreira, 1994).Both species also use the littoral zone as reproductive habitat in Lago de Pátzcuaro (García de León, 1984), and these areas have been lost in more than 70% (Alcocer-Durand & Bernal-Brooks, 2002).
Chirostoma estor is not currently extinct in this lake possibly because there are still remnant wetland areas where this fish species can breed and the contribution of terrestrial insects that may result from agriculture or livestock in areas surrounding the lake.Terrestrial insects are part of the current diet of Chirostoma spp.(Vital-Rodríguez, 2011).
In the past, top predators used to be large sized (>170 mm SL) ichthyophagous fish (García de León & Pérez-Velazco, 1996) and now they are smaller in size (<120 mm SL) planktivorous fish (Chirostoma spp.).The decrease in the sizes of large carnivores has occurred also in Lake Chapala (Moncayo-Estrada et al., 2012).According to Seilheimer & Chow-Fraser (2006), top predators (typically long-lived species) are susceptible to long-term physical and chemical habitat alterations.Other studies have argued that lakes with high clay turbidity are more beneficial to planktivorous fish than piscivorous predators for successful feeding (De Robertis et al., 2003;Moncayo-Estrada et al., 2010).In our study, small silversides (Chirostoma spp.) were captured in every seine pull and at all collection sites.Small silversides are typically open water fish using pelagic zones as feeding and reproductive habitat (Lyons et al., 1995;Mercado-Silva et al., 2002).These small fishes feed mainly on pelagic zooplankton and use a variety of reproductive substrates (gill nets, buoys, floating branches) that can be located far from the littoral zone (Rojas-Carrillo, 2006;Moncayo-Estrada et al., 2010).Silverside fishes represented more than 90% of the fish collected in our sampling efforts, which could indicate that species able to utilize non-littoral habitats have an advantage over species requiring now degraded littoral zones.These results are similar to those reported by Domínguez-Domínguez et al. (2008), where carnivorous species, such as A. robustus and A. dugesii declined in the drainage basins of the Río Morelia and Río Laja, two rivers located in central México, associated with changes in the trophic web or destruction of the habitat of prey.
Some studies have argued that it is possible to distinguish different functional groups of Chirostoma spp.by separating them into at least two different sizes: small silversides (<170 mm SL) and large silversides (from >170 mm to 340 mm SL) (Hernández-Montaño, 2006;Moncayo-Estrada et al., 2012).Further, it has been argued that small and large silversides can been classified in different trophic categories, where small silversides are secondary consumers and big silversides tertiary consumers and top predators (Moncayo-Estrada et al., 2012).In Lago de Pátzcuaro, silversides could be classified into small and large fishes before the decade of 2000s.However, in the present study, the largest captured silversides showed <130 mm SL, rendering the differentiation between large and small-sized fish ineffective.It is possible that this size reduction has resulted from intensive fishing efforts.Interestingly there is no diet variation among Chirostoma spp.30-90 mm SL, and some small silversides (<90 mm SL) can feed on juvenile silversides (Vital-Rodriguez, 2011).Some fishes in the Goodeidae family and the achoque (Ambystoma dumerilli) are known to live only in conserved lakes because they are extremely sensitive to water quality, vulnerable to predation, and have narrow habitat requirements (Huacuz, 2002;Domínguez-Domínguez et al., 2008).The presence of some goodeid fishes and achoque serve as indicators of water quality and relatively pristine habitats (Harig & Bain, 1998;Domiguez-Domínguez et al., 2008).We did not collect typically sensitive goodeid species, such as Skiffia lermae and Allotoca dugesii, or any amphibians, such as Ambystoma dumerilli.Their absence could indicate considerable degradation in water and habitat quality in the lake (Anderson, 1972;Soto-Galera et al., 1999;Lyons et al., 2000).
The native and carnivore fishes Allotoca dugesii, Allotoca diazi, Alloophorus robustus, big Chirostoma (C.estor >150 mm SL), Algansea lacustris and the exotic Micropterus salmoides depend on littoral zones for reproductive habitat in the lake (Gaspar-Dillanes et al., 2006).Severe alteration of the littoral zone could explain the decrease or absence of these species.We hypothesize that, if environmental degradation in Lago de Pátzcuaro continues, the fish community will in the future be dominated by four species: Chirostoma spp.(mostly native, carnivorous, tolerant) and Cyprinus carpio, Oreochromis spp.and Poeciliopsis infans (exotic, omnivorous, tolerant), with perhaps just a few areas of the lake where other species may persist.Since habitat structure plays an essential role in the maintenance of fish assemblages, and fish ecological guilds change when loss of habitat complexity ocurrs (Jones & Sayer, 2003;Bond & Lake, 2005), restoration of the littoral zone of Lago de Pátzcuaro will need to be a priority to achieve survival of its historical fish fauna.The loss or extinction of endemic species in the lake would represent alarming changes in the elements and ecological processes and also in geologic and biogeographic processes that allowed the origin of this fauna, and it would represent a greater chance of losing ecosystem services in the future.
Our study is based on a review of the historical collections recorded at zoological museums at a variety of institutions.We recognize that sampling equipment, time, location and objectives among many other variables may have changed from one sampling event to another.While these sampling inconsistencies certainly throw uncertainty into our analysis and forced our analyses to be relatively basic, the described trends on the fish community are strong and give a gross picture of the changes that have ocurred throughout the lake.The long period  covered by the analysis represents an advantage difficult to find in similar studies.We also recognize that "species-lumping" may have reduced the specificity of our findings.As stated above, we could not consistently separate the numerous individuals of Chirostoma we captured, and we analyzed all individuals as single taxa.This undoubtedly reduced the number of elements (species) that we could use in our comparisons.We believe however, that this aspect of data management does not obscure the overall trends we found in the Lago de Pátzcuaro fish community.
This study is a first attempt to characterize the long-term changes in the fish community associated with environmental degradation at Lago de Pátzcuaro.It could be used to predict the future status of the fish community and ecological processes of this ecosystem (Soto-Galera et al., 1999;Contreras-MacBeath et al., 1998).It could further offer the base-line to create new indices for environmental quality such as the index of biotic integrity (Karr, 1981) and other ecosystem monitoring protocols.Currently, the conservation activities have focused on improving water quality, the eradication of exotic species and the cultivation of some native species to enhance their populations.However, the dramatic changes in the ecological processes of the lake cannot ensure a successful species re-introduction.Our results suggest that the restoration of littoral habitat (wetland zones) represent a key management need to preserve ecosystem processes in Lago de Pátzcuaro.This would support the recommendations that the restoration of vegetated littoral zones could provide habitat regeneration for native small fish species and improve water quality in Lake Chapala (Moncayo-Estrada et al., 2010).We support the suggestion made by Orbe-Mendoza et al. (2002) that it is urgent to implement practices that offer fishermen a subsistence and sustainable fishery in the short term.

Figure 2 .
Figure 2. Long-term trends (1900-2010) of fish guilds in Lago de Pátzcuaro.Columns show mean and standard deviation.Average percentage change of each fish guilds by decade is presented above each bar.Letters a, b, c, and y, z indicate groups of decades for which no difference in the number of species was found in a posteriori pairwise comparisons obtained by MANOVA.