Introduction

There is a large consensus today that the human-mediated introduction of species outside their natural range is one of the main threats to biodiversity and the second leading cause of animal extinctions (Millennium Ecosystem Assessment 2005). Indeed, several non-indigenous species (NIS) have been beneficial to humans (e.g. corn, wheat, rice, plantation forests, and others) (Ewel et al. 1999) and many cause minimal environmental impacts, as predicted by the often cited “tens rule” (Williamson and Fitter 1996; but see Jerscke and Strayer 2005). So, the fraction of NIS that may yield problems is small, but these although few species have had catastrophic impacts. Following their introduction into the wild, they soon turned out to be “invasive” (for definitions of “invasiveness” see Gherardi 2006a), becoming numerically and ecologically prominent, spreading from the point of introduction and being often capable to dominate indigenous populations and communities (Kolar and Lodge 2001); ultimately, they had a profound effect on indigenous species, ecosystem processes, economic interests, and public health (e.g. Ricciardi et al. 1998). The economic costs produced by the various attempts to control NIS and to mitigate their impact may be high, although seldom assessed (for a short review see McNeely 2004).

Inland waters have been the theatre of spectacular invasions. Well-known examples are the introduction of the Nile perch Lates niloticus (Linnaeus) into Lake Victoria followed by the elimination of about 200 species of haplochromine cichlids (Craig 1992) and the alteration of the Laurentian Great Lakes communities and ecosystems by the sea lamprey Petromyzon marinus Linnaeus and the zebra mussel Dreissena polymorpha (MacIsaac et al. 2001). Other less celebrated dramas are ongoing in many freshwater systems of the world with the intervention of previously unsuspected actors, such as Lepomis gibbosus among fish, Dikerogammarus villosus among crustaceans, or Rana catesbeiana among amphibians (Gherardi 2007a).

These and other examples (cfr. Leppäkoski et al. 2002a; Gherardi 2007b) confirm that, compared to terrestrial systems, inland waters are highly vulnerable to either inadvertent or deliberate introductions of species and to their subsequent spread. This vulnerability is the effect of the intensive human uses (Ricciardi 2001), on the one hand, and of the natural linkages among streams and lakes, the effects of water flow, and the dispersal capability of aquatic organisms, on the other.

As a result, several NIS dominate some vast waterscapes of the world (e.g. the red swamp crayfish, Procambarus clarkii, dominating many waterbodies of southern Europe; Gherardi 2006b). Xenodiversity (sensu Leppäkoski et al. 2002b) may be extraordinarily high in, for example, large rivers of developed countries (e.g. the Hudson; Strayer et al. 2005) and largely affects many taxa (e.g. fish; Lehtonen 2002).

Species originating from diverse bio-geographical areas now coexist in several basins. In the Rhine, for example, indigenous crustaceans [Gammarus pulex (Linnaeus)] occur together with North American species (e.g. Gammarus tigrinus Sexton), Mediterranean species (the freshwater shrimp Atyaephyra desmaresti Millet), and Ponto-Caspian species (e.g. D. villosus) (Beisel 2001). Biotic homogenization—i.e. the ecological process leading to an increased similarity of formerly disparate biota over time (Olden and Rooney 2006)—is constantly accelerating (Clavero and García-Berthou 2006; Rahel 2007) and some freshwater systems, such as the Great Lakes, function as “hotspots” of xenodiversity (e.g. MacIsaac et al. 2001).

Finally, many freshwater invaders are moved within and among continents in association with economic activity and trade globalization that benefit millions worldwide (Lodge and Shrader-Frechette 2003). The inevitable tension between two often competing goals—increasing economic activity and protecting the environment from invasive species—makes it difficult to justify the need for decision makers to contain the spread of these species and to mitigate the environmental risks they pose (Gherardi 2006a).

All this implies the need to improve our understanding of the dynamics of species introductions with the purpose to prevent or control future invasions and to predict and reduce their effects (Shea and Chesson 2002). When engaged in this effort, the main steps to undertake are, first, to identify the NIS occurring in a given area and, second, to individuate the vectors of their introduction and the pathways they followed to enter that area. Inventories of NIS have had the potential to lay the basis for describing patterns of invasion at both global (e.g. Lonsdale 1999) and regional scales (e.g. Rabitsch and Essl 2006; Gollasch and Nehring 2006). An accurate knowledge of xenodiversity is also needed to develop policies to cope with the problems associated with biological invasions. Ultimately, an understanding of the mechanisms of previous invasions may help protect aquatic ecosystems from the impacts of future invaders.

This paper is the first attempt to draw a list of the animal NIS occurring today in the inland waters of Italy. Here, we also aimed at identifying the current distribution of these species in Italy, and the times and modes of their introduction. Finally, we analyzed the data to extract more general information that might provide suggestions about the approach to undertake for managing freshwater NIS in Italy and for setting priorities.

Materials and methods

The list has been compiled gathering the information available so far from the scientific literature and processing it after their validation and implementation. Here we report only species introduced by humans but cases of possible natural dispersal are also analyzed. Following the “Guidelines for the reintroduction and restocking of animal species of Community concern” (Italian Ministry of Environment, February 14, 2006), NIS are here regarded as those species that entered Italian waterbodies following the discovery of America by Columbus in 1492 (cfr. Copp et al. 2005). Some cases of single and sporadic occurrences are cited separately. We analyzed here amphibians, reptiles, birds, and mammals that need freshwater systems to complete their life cycle, and invertebrates (free-living and parasitic) and fish inhabiting inland water systems. Inland waters are here defined as all standing or flowing water on the surface of the land (Directive 2000/60/EC of the European Parliament and of the Council establishing a framework for the Community action in the field of water policy). The analyzed NIS are “foreign” (i.e. species non native to Italy; Copp et al. 2005) and not “translocated” species (i.e. species introduced from other basins).

NIS have been classified according to several categories, i.e. their native range, the date of their (first) introduction into the wild, their current distribution in Italy, mode/s of their arrival in Italy (unintentional or intentional introduction), and vector/s and pathway/s of their first introduction. Dates of NIS first introduction refer to either the exact or the approximate year reported in scientific publications or, when this was not available, to the year of the first record. To reduce the bias due to both approximation and the obvious delay between species introduction, data gathering, and publication of these data, the analysis has been done on 10-year intervals. To describe the distribution of NIS, Italy has been divided into five regions, North-west (NW, including: Aosta Valley, Liguria, Lombardy, and Piedmont), North-east (NE, including: Emilia-Romagna, Friuli-Venezia Giulia, Trentino-Alto Adige, and Veneto), Center (C, including: Abruzzo, Latium, Marches, Molise, Tuscany, and Umbria), South (S, including: Apulia, Basilicata, Calabria, and Campania), and Islands (I, including: Sardinia and Sicily). Vectors of introductions have been classified into: dispersal, if NIS entered Italy as the result of range expansion by active or passive means from populations of neighboring countries; escape, if they escaped from captivity; release, if they were deliberately released into the wild; and transport, if they were transported accidentally by human means.

Finally, pathways of NIS introductions, independently of their being target or non-target species for that pathway, have been classified into: biocontrol, when NIS have been released in the wild as control agents of other species; culture, when they have been originally imported in association with aquaculture and farming, this category also including specie inadvertently introduced with crops; ornamental, when they have been imported in association with aquariophily or for ornamental purposes; stock enhancement (or stocking), when they have been introduced to increase wild production in association with professional or sport fishing, this category also including species used as fish food or fish bait; shipping, when they have been introduced via ship hulls and in ballast water or sediments. When no documentation is available for a given category or it is dubious or anecdotal we deemed it as “unknown”.

Statistical comparisons were made using Wilks’ test after Williams’ correction (statistic: G). The level of significance at which the null hypothesis was rejected is α =  0.05.

Results

The invertebrate and vertebrate species introduced into Italian inland waters since the XVI century reach a total of 112 (Appendices 1 and 2). Our list also includes Cyprinus carpio, since its often cited introduction in Roman ages is still under debate (S. Zerunian, pers. comm.). The occurrence of several other species is reported as sporadic, such as the octochaetid earthworm Dichogaster modiglianii (Rosa) at Abano spa in Veneto (Omodeo et al. 2005), the mollusc gastropod Helisoma anceps (Menke) (Henrard 1968) in Tuscany (Cianfanelli et al. 2007), the Chinese mitten crab Eriocheir sinensis H. Milne Edwards in the Lagoon of Venice (Mizzan 2005), and the reptiles Chelydra serpentina (Linnaeus), Mauremys caspica (Gmelin), M. leprosa (Schweigger), Pseudemys concinna Le Conte, and Graptemys versa Stejniger in several Italian regions (Andreone and Sindaco 1998; Bernini et al. 2004; Bologna et al. 2000, 2003; Scalera 2001, 2003; Sindaco et al. 2006). The crayfish Cherax destructor Clark (with the associated ectocommensal Platyhelminthes Temnocephala minor Haswell) is currently confined in aquaculture facilities (Quaglio et al. 1999) and the marble crayfish Procambarus sp. (Souty-Grosset et al. 2006) in pet shops (N. Crigna, pers. comm.). The taxonomic status of some populations is still uncertain; for instance, some frog populations in North-west Italy were first assigned to Rana ridibunda Pallas and now to R. kurtmuelleri Gayda or R. balcanica Schneider and Sinsch (Sindaco et al. 2006). Finally, the bird Aix galericulata (Linnaeus) is not yet established (i.e. with self-sustaining populations), although a few couples have been recently found to breed in Piedmont (GPSO 2007).

Among the 64 invertebrates listed, 3 species are parasites (the Platyhelminthes Gyrodactylus salaris, the Nematoda Anguillicola crassus, and the Crustacea Argulus japonicus, all parasites of fish, i.e. salmonids, eels, and cyprinids/salmonids, respectively), whereas the Annelida Cambarincola mesochoreus and Xironogiton victoriensis are ectocommensals of non-indigenous crayfish, P. clarkii and Pacifastacus leniusculus, respectively. The remaining 59 invertebrates of our list are free-living species, mostly crustaceans (32 species) and molluscs (11 species); all of them are today established, except the Coleoptera Sternolophus solieri. The majority of the 48 vertebrates recorded here are fish (38 species). Nine fish species (labeled with an asterisk in Appendix 2) require a constant reintroduction by man to sustain their populations. Similarly, Trachemys scripta is considered not yet established in Italy, although some populations are know to reproduce in several sites (Di Cerbo and Di Tizio 2006).

The contribution of each division to the xenodiversity of Italian inland waters is shown in Fig. 1 and the fraction of NIS per taxon is given in Table 1. Both fish and arthropods are the most frequent non-indigenous taxa, but by far Mammalia (50% NIS) and bony fish (more than 46% NIS) communities seem to be the most affected, followed by Crustacea Ostracoda, Annelida Branchiobdellea, Crustacea Branchiura, and Crustacea Decapoda. Invertebrates and vertebrates share a similar number of NIS (G = 2.283, df = 1, ns) but the fraction of xenodiversity is significantly higher in the latter (48/204, 23.53%) than in the former (64/5453, 1.17%; G = 174.393, df = 1, P < 0.001). Overall NIS contribute for 1.98% to the whole inland-water fauna (estimated to amount to over 5,600 species; Ruffo and Stoch 2005; Anonym. 2007; pers. comm. of several experts).

Fig. 1
figure 1

Frequency (in %) of species of each phylum/division within the 112 non-indigenous animal species recorded in Italian inland waters. Ann = Annelida, Amp = Amphibia, Ave = Aves, Cni = Cnidaria, Cru = Crustacea, Hex = Hexapoda, Mol = Mollusca, Ost = Osteichthyes, Pla = Plathyhelmintes, Mam = Mammalia, Nem = Nematoda, Rep = Reptilia, Rot = Rotifera

Full size image
Table 1 Number of non-indigenous (NIS) and indigenous (IS) species recorded in Italian inland waters and percentage of NIS per each phylum/division
Full size table

Asia, North America, and the rest of Europe are the prevalent donor continents for both invertebrates and vertebrates (G = 10.533, df = 5, ns; Fig. 2), but invertebrates originate from extra-European countries more often than vertebrates (81.71% vs. 65.52%, G = 4.228, df = 1, P < 0.05), particularly from Asia. A difference was also found for the times of NIS introduction or of their first record in the wild, the entrance into Italy of vertebrate species having apparently started long before that of invertebrates (G = 25.999, df = 7, P < 0.001) (Fig. 3). Interestingly, the rate of invertebrate introductions seems to increase since the 1970s (after 1970: 82.46%), whereas vertebrate introductions appear to be relatively constant with time (after 1970: 50%; invertebrates vs. vertebrates compared before and after the 1970s: G = 11.049, df = 1, P < 0.001). Northern (NW: 29.45%, NE: 28.77%) and central Italy (19.86%) are significantly more affected by animal NIS than southern Italy (11.64%) and the islands (10.27%) (G = 49.917, df = 4, P < 0.001; Fig. 4a), without any significant difference between invertebrates and vertebrates (G = 2.239, df = 4, ns). However, non-indigenous vertebrates seem to be more widely diffused in Italy, as suggested by the larger number of Italian regions they have colonized (G = 19.206, df = 4, P < 0.001; Fig. 4b).

Fig. 2
figure 2

Frequency distribution of the non-indigenous animal species (NIS) recorded in Italian inland waters per donor continent. Species whose native range includes two or more continents were tallied more than once

Full size image
Fig. 3
figure 3

Increase with time in the frequency of the non-indigenous animal species (NIS) recorded in Italian inland waters. Dates refer to the exact or approximate year of introduction into the wild or, when this datum is absent, to the year of the first record in the published literature. The year of the first introduction/record is missing for 7 invertebrates and 10 vertebrates of the list

Full size image
Fig. 4
figure 4

Frequency distributions of inland-water non-indigenous animal species (NIS) (a) per Italian region (NW = North-west, NE = North-east, C = Center, S = South, and I = Islands) and (b) per number of regions they have colonized. Species that occur in two or more regions have been tallied more than once in (a). The distribution of 4 invertebrates is unknown

Full size image

As obvious, vertebrates were introduced by intentional means significantly more often than invertebrates (excluding unknown cases, G = 36.372, df = 1, P < 0.001; Fig. 5a) as the result of the more frequent instances of releases of animals for stocking purposes (in the case of fish) or of abandonment by pet amateurs (in the case of reptiles and birds) (excluding unknown cases, G = 87.084, df = 3, P < 0.001) (Fig. 5b). However, both invertebrates and vertebrates used the same pathways of introduction (Fig. 5c; excluding unknown cases, G = 8.282, df = 4, ns), stock enhancement (47.92%) and culture (37.5%) largely prevailing over the other pathways (ornamental: 8.33%, biocontrol: 4.17%, shipping: 2.08%; excluding unknown cases, G = 89.111, df = 4, P < 0.001).

Fig. 5
figure 5

Frequency distributions of inland-water non-indigenous animal species (NIS) per (a) mode of arrival, (b) vector, and (c) pathway used to enter Italian inland waters. Species that show two or more modes of arrival, vectors, or pathways have been tallied more than once. Data of 13, 15, and 23 invertebrates and of 4, 9, and 10 vertebrates are unknown for mode of arrival, vector, and pathway, respectively

Full size image

Discussion

Given the inevitable future increase of species introductions (e.g. Gherardi 2007a), this study provides only a snapshot of the number and distribution of animal NIS in Italian inland waters. It also offers suggestions about their prevalent modes of arrival, vectors, and pathways. According to our data, the current xenodiversity of Italian inland waters amounts to 112 species that contribute to about 2% to the Italian freshwater fauna. Northern and central regions are most affected by freshwater NIS and Asia, North America, and the rest of Europe are the main donor continents.

The total NIS we recorded is certainly an underestimate of the xenodiversity in Italian inland waters. First, notwithstanding the increased scientific interest for biological invasions in the last decade and the surge of researches focused on the identification of freshwater NIS (Gherardi 2007a), there is still a gap of knowledge about invertebrate taxa and some functional groups. For instance, we have recorded three parasitic and two ectocommensal invertebrates, which is a relatively low number if compared to the list, although provisional, of eight parasites reported by García-Berthou et al. (2007) for the Iberian Peninsula. Second, we excluded from our list the species whose “cryptogenic” status (sensu Carlton 1996) has not been fully solved, such as Lota lota (Linnaeus), Salvelinus alpinus (Linnaeus) (Betti 2004; Piccinini et al. 2004), and Tinca tinca Linnaeus (Bianco 1998). On the contrary, our list includes the gastropod Haitia acuta: its origin from North America, and not from the continental Europe as previously suggested (e.g. Burch 1989; Smith 1989), has been confirmed by fossil records, morphological studies, and historical data (Taylor 2003; García-Berthou et al. 2007). Third, we have not analyzed the many events of translocations from one to another region or from the mainland to the islands (Scalera 2001), as well as those of fish between separated ichthyo-geographic districts (Zerunian 2002).

Although provisional, our list of NIS is relatively long if compared to the inventories recently compiled for other European countries, such as Austria (92 animal NIS; Füreder and Pöckl 2007), British Isles (about 60 NIS, plants included; Minchin and Eno 2002), Germany (82 NIS, fungi and plants included; Gollasch and Nehring 2006), and the Iberian Peninsula (73 animal NIS, species from estuaries and coastal lagoons included; García-Berthou et al. 2007). Only three were the species that seem to have entered Italy by the means of natural dispersal, i.e. Daphnia parvula that possibly moved in association with migrating waterfowl (but its transport with recreational boats cannot be excluded; Panov et al. 2004), Ondatra zibeticus that expanded its range from the neighboring Slovenia (Lapini and Scaravelli 1993), and Cygnus olor (but it has been also released for ornamental purposes; Andreotti et al. 2001). Conversely, the large majority of NIS arrived into Italy as a direct or indirect effect of human intervention.

Vertebrates in general and bony fish in particular seem to be the most affected taxon by species introductions. This is not surprising: comprising the most visible and attractive species, on the one hand, and because of the perceived ecological role they play in aquatic food webs and their economic importance, on the other, vertebrates have received the greatest scientific attention. Indeed, due to sport and commercial fishing, aquaculture practices, pet trade, and fur farming, vertebrates have been also subject for more than a century to recurrent and extensive introductions into Italy, most often intentional. The voluntary release of fish into the wild reflects a general phenomenon occurring across the entire Europe (Copp et al. 2005). However, this practice seems to be more diffused in Italy, where controls on legal and illegal stocking have been, at least in the past, generally ineffective or absent. As a consequence, non-indigenous fish in Italy are relatively more numerous than in the rest of Europe (Copp et al. 2005), representing the fourth cause of threat to indigenous fish (Zerunian 2002). As a side effect, fry production centers for sport-fish restocking have probably contributed to the introduction and spread of frogs. Among the other types of intentional introductions, Bertolino and Genovesi (2007) recently reported the release of more than 30,000 individuals of Neovison vison in the last 6 years by animal liberation activists and Scalera (2007a,b) showed that pet amateurs are responsible for the recurrent, voluntary release in the wild of Trachemys scripta in the last two decades.

Our list includes 64 invertebrate NIS, which exceed the total of 50 species recently reported by Zapparoli (2005). Their introduction is certainly a long-lasting phenomenon in Italy. The increased rate recorded here since the 1970s, also apparent in German fresh waters (Gollasch and Nehring 2006), might be a by-product of the augmented scientific interest in freshwater invertebrate communities but also the result of the success of more tolerant species in systems subject to an augmented degradation and pollution (Meier-Brook 2002). Accidental transport, in association with both fish (for aquaculture or stock enhancement) and crops, especially rice, has been the main vector of invertebrate introductions. Only the crayfish Pacifastacus leniusculus seems to have been intentionally released in the wild by fishermen (Capurro et al. 2007), whereas Artemia franciscana, first introduced as fish food, subsequently spread at a wide regional scale using waterfowl as vectors (Mura et al. 2006).

Obviously, a realistic assessment of the pathways followed to introduce species should be made on a case-by-case basis. For example, in Germany the most frequent pathway for freshwater introductions is the diffusion through artificial canals (31%) (Gollasch and Nehring 2006), which seems to be obviously less important in Italy (the Alps impede the construction of canals and navigation from the ports in the Mediterranean to inland waters is scarce or absent), whereas stock enhancement has a lower impact in Germany than in Italy (23% vs. 47.92%) and aquaculture or farming has none (0% vs. 37.5%). Shipping seems to be responsible for the arrival into Italy of only two species (D. villosus among Crustacea and D. polymorpha among Mollusca; 2.08%); on the contrary, it has caused 15% of the recorded freshwater introductions in Germany and 25% in the whole Europe (Gollasch 2007).

While some NIS seem to remain insignificant additions to the native biota, 17 and seven species of our list are included among the 100 worst invasive species of Europe (DAISIE consortium, “Delivering Alien Invasive Species Inventories for Europe”) and of the world (IUCN; Lowe et al. 2000), respectively. For some (but not all) species recorded the multilevel impact exerted on the recipient communities and ecosystems is known in Italy and elsewhere, even if seldom quantified (Gherardi 2007c), including (1) competitive superiority over indigenous species, possibly leading to local extinction or extirpation—e.g. P. clarkii (Gherardi 2007d, Gherardi and Acquistapace 2007); (2) hybridization with indigenous species with the consequent reduction of genetic diversity—e.g. Barbus barbus (S. Zerunian, pers. comm.); (3) disruption of the pristine interactions between species and of the existing food web links—e.g. D. villosus (e.g. Dick and Platvoet 2000; Casellato et al. 2007) (4) habitat modification and alteration of ecosystem functioning—e.g. the coypu M. coypus (Bertolino and Genovesi 2007, Panzacchi et al. 2007); (5) introduction of parasites and disease agents—e.g. Anguillicola crassus (Kirk 2003), Aedes albopictus (Craig 1993), and Pseudorasbora parva (Gozlan et al. 2005); and (6) damages to socio-economics, recreation, human health and well-being—e.g. D. polymorpha (e.g. Karatayev et al. 1997; Strayer et al. 1999; Lancioni and Gaino 2006). Conversely, we are still ignorant about the long-term ecological and evolutionary feedbacks between invasive species and the invaded communities and ecosystems (Strayer et al. 2006). This suggests that additional research is needed to provide valuable criteria for prioritizing interventions against well established invaders, evaluate alternative management approaches, and finally identify which new potential invader should be eventually targeted as “unwanted”.