Alien fish, amphibian and reptile species in Romania and their invasive status: a review with new data

The paper presents a review of literature data, supplemented with original observations, on the presence, establishment, distribution and invasive status of alien fish, amphibian and reptile species in Romania. Consistent criteria were followed in defining alien species records, establishment and invasive status. From the 48 alien fish species, 1 fish hybrid, 1 amphibian and 18 reptile species recorded, only 16 fishes and 3 reptiles can be regarded as established. Of these we consider the criteria for invasive status as being probably fulfilled by one fish species (Perccottus glenii), and less likely by six more fish species. The presence and the alien status of the one amphibian are debatable. No reptile species can be considered invasive at present.


Introduction
The alien, or non-native, species are of increasing concern for conservation-oriented management all over the world. Indeed, some introduced species have wrought havoc across ecosystems in various places and caused significant reduction in the native ones, even extinction, especially in vulnerable insular situations (see, e.g., BirdLife International 2016; Wyatt et al. 2008). However, introduction of species is a common phenomenon and there is a considerable range of potential results, in terms of ecological significance and impact. Firstly, not all introductions succeed in establishing viable populations, and even established populations of non-native species are not inherently destructive or "invasive", there being numerous instances where their ecological impact is neutral or negligible (Copp et al. 2005). There is considerable discussion about the average rate of success of alien species (i.e. species introduced to a non-native setting) but it may range from ca. 10% (Williamson and Fitter 1996) to ca. 50% (Jeschke and Strayer 2005) for each adaptation/expansion "step" -being usually considered three major such "steps": "escaping", "establishing" and "becoming a pest" (Williamson and Fitter 1996). For Romania, there are numerous efforts to list and discuss non-native/invasive species (e.g. Anastasiu et al. 2016;Anastasiu et al. 2017;Stănescu et al. 2020), some focused upon fishes as a group of particular interest, with numerous introduced and acclimatized species (Manea 1985;Gavriloaie et al. 2003;Gavriloaie 2007;Iacob and Petrescu-Mag 2008;Gavriloaie and Berkesy 2013;Gavriloaie et al. 2015 etc.). However, the criteria used for listing species are variable, some, for instance, including (e. g. Iacob and Petrescu-Mag 2008) and some excluding (e.g. Stănescu et al. 2020) non-native species that only survive in thermal waters; variable therefore being also the contents of the national lists of introduced species. We have endeavoured to introduce a consistent set of criteria, and to discuss the current distribution and situation of non-native fish, amphibian and reptile species in Romania, using both literature references and the result of personal observations.

Materials and methods
We have searched the literature pertaining to introduced fish, amphibian and reptile species for Romania, including non-technical (especially inasmuch as they present imagery that can be checked) and online publications (introductions of alien forms/ subspecies of species natively occurring in Romania have not been considered 1 ). Species considered are those that have either been found free-ranging, even as a single specimen, without known ownership, or are/were being kept under open, pond-or basin-based farming conditions (from which escape or release is theoretically possible) or are known from immediately outside Romania from ecosystems (such as a river channel) which are contiguous to Romanian ones and are likely, or could be expected to, occur on the Romanian side also. Transient introductions that cannot survive over the winter, or introductions limited to thermal waters, have also been considered. Practically we have chosen to treat all introductions outside of zoological collections and indoors farming, thus being able to cover species such as Parabramis pekinensis which are treated by most authors (e. g. Manea 1985;Gavriloaie et al. 2003;Gavriloaie 2007;Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015;Stănescu et al. 2020) although they were only transiently present in Romania, in pond-based aquaculture (Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015). However, we introduce a discussion of the current status and spread that attempts to clarify which species are still present and where, precisely. Personal records are integrated into this discussion, with mention of observation year. Also, data from the collection of the "Grigore Antipa" National Museum of Natural History were used. Classification is after Fricke et al. 2021and van der Laan et al. 2021, for fishes, Frost 2020, Uetz et al. 2020, and Speybroeck et al. 2020 for amphibians and reptiles -except as otherwise stated.

Fam. Polyodontidae
Polyodon spathula (Walbaum, 1792). Common name(s): English: Paddlefish, Mississippi Paddlefish (Froese and Pauly 2019); Romanian: "pește spatulă" 2 (Iacob and Petrescu-Mag 2008). The species is native to North America (Froese and Pauly 2019). It was introduced in aquaculture starting with 1992 (at Nucet fish farms, Dâmbovița county), it was successfully reproduced and then populated into some lakes along the Argeș river (Simeanu et al. 2015). Escapes from farms (probably both Romanian and Bulgarian) have led to specimens being found in the Argeș ("Grigore Antipa" National Museum of Natural History collection) and the Danube (Simonović et al. 2006;Bănăduc et al. 2014). These are most likely not reproducing (Nalbant 2003;Lenhardt et al. 2011), though some juveniles (which may have escaped from farms) were found in the lower Danube river (Jarić et al. 2019). There is little information on the potential of Polyodon for impact upon native species; still, as there is no firm evidence for its reproduction in natural waters anywhere it was stocked, it is considered of low risk of becoming invasive (Jarić et al. 2019).

Fam. Acipenseridae
Acipenser baerii Brandt, 1860. Common name(s): English: Siberian Sturgeon (Froese and Pauly 2019); Romanian: sturion siberian (CE regulation no. 506/2008), nisetru siberian (WWF Internațional/Programul Dunăre-Carpați 2014). The species is native to Siberia (Froese and Pauly 2019). This species was introduced in aquaculture in Romania in the early 2000s (Williot and Chebanov 2018). There are no records of this species escaping from Romanian farms so far, but upon floods in 2008 numerous specimens from Ukrainian farms have reached into the Prut river and the species was found down to the Stânca-Costești reservoir, on the Romanian- Moldavian border, in 2008(Khudyi and Khuda 2013) -therefore, A. baerii can be said to have reached free Romanian waters. However, it was not caught in that lake in a 2014 study, but neither was the native sterlet (Acipenser ruthenus), known to be present from fishermen's reports (Bulat et al. 2016). A. baerii has also escaped in the Middle Danube region, being present in free waters of Hungary (Weiperth et al. 2013) and has already hybridized with the native species Acipenser ruthenus (Ludwig et al. 2009). Hybrids of Acipenser baerii and Acipenser gueldenstaedtii are also kept in aquaculture in Romania (Popa et al. 2017 (Bănărescu 1964). The species is native to Eastern Asia (Froese and Pauly 2019). Goldfishes were introduced into ornamental and angling fishponds, sometimes escaping into rivers such as the Jiu (Bănărescu 1964;Oțel 2019). The "goldfish" is distinct from the "gibel carp" or "Prussian carp", Carassius (auratus) gibelio (Bloch, 1782) 3 . The former is an Asian 3 Rylková et al. 2013, Oțel 2019and Fricke et al. 2021 consider gibelio a valid species. However, it appears to be a (mostly) polyploid, alternately sexual and clonal (gynogenetic) form of C. auratus (Xiao et al. 2011;Luo et al. 2014;Li et al. 2018). Morphologically gibelio and auratus overlap (Oțel 2019), and gibelio appears to be but one of many recurrently/fluctuatingly polyploid auratus strains (Qin et al. 2016;Liu et al. 2017). Apparently sexual individuals of all forms of the C. auratus group (auratus, gibelio and even the Japanese cuvieri) interbreed without ill effects (Xiao et al. 2011;Kokodiy 2018;Liu et al. 2019;Mezhzherin et al. 2020) and the sex determinism is partly genetic and partly temperature dependent (Li et al. 2018); this and parasite resistance dynamics (Pakosta et al. 2018) determine frequent reversal of gynogenetic condition towards the sexual one (Xiao et al. 2011;Rylková et al. 2013;Luo et al. 2014;Li et al. 2018;Przybył et al. 2020). Therefore, gibelio is best considered a subspecies/form of C. auratus, wider spread across Eurasia (including Eastern Europe) than the typical C. a. auratus (native to Eastern Asia). Sexual (and even gynogenetic) forms of this fish, the latter (despite being frequently treated as alien, e.g. Bănărescu 1964;Iacob and Petrescu-Mag 2008) is arguably native not only in Siberia but also in Central and Eastern Europe (see, e.g., Rylková et al. 2013;Fricke et al. 2021) including the Romanian Danube and its large tributaries (as convincingly argued on basis of historical but accurate information by Oțel 2019). The expansion of the gibel carp to the detriment of the crucian carp, Carassius carassius, appears to be primarily a consequence of anthropogenic ecological change favouring the first species (shift from clear to turbid water, see Năvodaru et al. 2002) and but secondarily the result of competition (following artificial translocations of C. (a.) gibelio, as stated by some authors -e.g. Bănărescu 1993;Bănărescu 2005). As for the goldfish, the impact of its introduction is unclear for now; it might hybridize with the gibel carp (Oțel, pers. comm.), which is to be expected given their conspecific condition.

Subfam. Smiliogastrinae
Pethia conchonius (Hamilton, 1822). Common name(s): English: Rosy Barb (Froese and Pauly 2019). The species is native to South Asia (Froese and Pauly 2019). Temporarily introduced to a small basin in the Botanical Garden in Bucharest by aquarists (Gavriloaie 2008); they reproduced during summer but died with the cold of late autumn (Iftime and Iftime, pers. obs., 2001).
Pethia ticto (Hamilton, 1822). Common name(s): English: Ticto Barb (Froese and Pauly 2019). The species is native to South and South-East Asia (Froese and Pauly 2019). Temporarily introduced to the same small basin in Bucharest by aquarists (Gavriloaie 2008); they reproduced during summer but died with the cold of late autumn (Iftime and Iftime, pers. obs., 2001).

Fam. Xenocyprididae
Hypophthalmichthys molitrix (Valenciennes, 1844). Common name(s): English: Silver Carp (Froese and Pauly 2019); Romanian: sânger, hipo, crap chinezesc (Oțel 2007). The species is native to Eastern Asia (Froese and Pauly 2019). This species was introduced together with other "Asiatic cyprinids" (Hypophthalmichthys nobilis, Ctenopharyngodon idella, Mylopharyngodon piceus, Parabramis pekinensis, Megalobrama terminalis) in 1960-1962 at the fish farms Nucet, Caraorman (Tulcea county) and Cefa (Bihor county); it was also deliberately populated in lakes, fishponds and reservoirs (Bușniță 1964;Manea 1985;Oțel 2007;Gavriloaie et al. 2015). Apparently numerous specimens escaped from culture in the 1970s, aided by species do interbreed with C. carassius , but the scarcity of such hybrids (Mezhzherin et al. 2017), their growth rate, initially high supposedely due to heterosis, but afterwards much decreased (Kokodiy 2018) and their ploidy aberrations (Knytl et al. 2018) would indicate that C. carassius is a distinct species from C. auratus/gibelio. flooding (Ciolac 2004). It has acclimated and reproduces freely in natural waters in the Danube and its Delta (Jankovic 1998;Staraș and Oțel 1999;Nalbant 2003;Ciolac 2004;Oțel 2007;Iacob and Petrescu-Mag 2008;Lenhardt et al. 2011;Zorić et al. 2014;Gavriloaie et al. 2015), though there are opinions that it cannot as yet reproduce in Bulgarian waters, including the Danube (Yankova 2016). Reproduction is dependent upon "long unregulated water courses in which the pelagic eggs can incubate" (Copp et al. 2005), conditions that are only met in some areas and periods (Staraș and Oțel 1999;Oțel 2007), therefore reproduction, while probably sufficient to propagate the Danube population independently of re-stocking from farms, is limited. The species is widespread in the Danube and Delta, but also found in most large rivers, though it does not reproduce there (Oțel 2007). It was particularly abundant during the mid-to-late 1990s in the lower Danube and its Delta but has declined later (Ciolac 2004;Oțel 2007). Its spread appears to have coincided with the anthropogenic (but independent from alien species introductions) shift from clear water to turbid water in most of the lower Danube and Delta (Năvodaru et al. 2002;Copp et al. 2005). Thus, this species, a phytoplankton filter-feeder, likely profited from the increased nutrient load (and associated algal blooming), hence probably its abundance and later decline. Some studies claim that H. molitrix can compete with native planktivorous species, e.g. in lake Balaton (Boros et al. 2015), where, however, the stock is derived from farming, not reproducing in free waters (Vitál et al. 2017). Other authors estimate that its impact upon native species is limited, as it has reached an ecological equilibrium state (Lenhardt et al. 2011) or is insufficiently known (Soes et al. 2011;Povž 2017), as is its capacity of controlling algal blooms (Vörös 2000;Oțel 2007). The introduction and transport of H. molitrix and other Asian carps has probably contributed to the establishment of the Asian bivalve Sinanodonta woodiana in Romania and Europe (Popa et al. 2011;Konečny et al. 2018). It has been conjectured that H. molitrix, may, through spread of S. woodiana, negatively impact the native species Rhodeus amarus (Soes et al. 2011). However, the impact of S. woodiana upon R. amarus was estimated as ranging "from almost neutral to highly negative" (Konečny et al. 2018, and literature quoted therein), with Danubian R. amarus in the "almost neutral" part of the spectrum (Reichard et al. 2015), all the more that in Romania R. amarus remains relatively abundant (Oțel 2007).Therefore, and inasmuch as its free-range reproduction is probably discontinuous and its free-ranging stock has apparently declined after a peak, its actual impact as an established non-native species in Romania is likelier to be limited than otherwise.
It is a zooplankton filter-feeder (Oțel 2007). The impact could be similar to that of H. molitrix, but proportional to the lesser abundance of H. nobilis; some authors even comment about a mainly positive effect of H. nobilis upon native fishes (see Lenhardt et al. 2011 and literature quoted). Hybrids between the two Hypophthalmichthys are known, e. g. in Hungary (Boros et al. 2015).
Ochetobius elongatus (Kner, 1867). Common name(s): none in either English or Romanian (Froese and Pauly 2019). The species is native to Eastern Asia (Froese and Pauly 2019). It has been introduced unwittingly together with the other Asiatic cyprinids; it lived in fishponds but did not escape to free waters and has doubtlessly been eliminated along time (Bușniță 1964).
Mylopharyngodon piceus (Richardson, 1846). Common name(s): English: Black Carp (Froese and Pauly 2019); Romanian: scoicar (Oțel 2007). The species is native to Eastern Asia (Froese and Pauly 2019). In Romania it was introduced in aquaculture/ fishponds together with other Asiatic cyprinids (with the intent of making use of the abundant mollusk resource, as it is an eater of mollusks - Manea,1985) but it does not reproduce unassisted here and only persists in culture, in small numbers (Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015). Apparently, it occurs in the Hungarian stretch of the Crișul Repede/Sebes Körös river (Telcean et al. 2005;Telcean and Cupșa 2009), however, it is unclear whether it reproduces there.
Squaliobarbus curriculus (Richardson, 1846). Common name(s): English: Barbel Chub (Froese and Pauly 2019). The species is native to Eastern Asia (Froese and Pauly 2019). It has been introduced unwittingly together with the other Asiatic cyprinids; it lived in fishponds but did not escape to free waters and has doubtlessly been eliminated along time (Bușniță 1964 (Iacob and Petrescu-Mag 2008). The species is native to Eastern Asia (Froese and Pauly 2019). It was introduced in aquaculture/fishponds together with other Asiatic cyprinids, but apparently the attempt was only transiently successful, and nowadays the species is no longer present in Romania (Bușniță 1964;Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015).
Hemiculter leucisculus (Basilewsky, 1855). Common name(s): English: Sharpbelly (Froese and Pauly 2019). The species is native to Eastern Asia (Froese and Pauly 2019). It has been introduced unwittingly together with the other Asiatic cyprinids; it lived in fishponds but did not escape to free waters and has doubtlessly been eliminated along time (Bușniță 1964). Abbott, 1901. Common name(s): none in either English or Romanian (Froese and Pauly 2019). The species is native to Eastern Asia (Froese and Pauly 2019). Same introduction history as H. leucisculus (Bănărescu 1963;Bușniță 1964).

Fam. Gobionidae
Pseudorasbora parva (Temminck et Schlegel, 1846). Common name(s): English: Stone Moroko (Froese and Pauly 2019); Romanian: murgoi bălțat (Oțel 2007). The species is native to Eastern Asia (Froese and Pauly 2019). It has been introduced unwittingly together with the other Asiatic cyprinids, and has escaped from the fishponds spreading through the Danube watershed practically all over the country, in the plains and lower hills, occasionally penetrating the sub-montane areas. It was also vehiculated by fishermen as bait (Bușniță 1964;Bănărescu 1964;Bănărescu et al. 1999;Oțel 2007;Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015;Telcean et al. 2015). According to some opinions it did not cause a decline in any native fish species (Bănărescu 1993); others claim a negative impact upon native fishes and cyprinid aquaculture (e.g. Giurcă and Angelescu 1971;Oțel 2007). Foreign free-range studies are not unanimous as to the impact upon native species abundance, through competition and/or disease vehiculation (see, e.g., Rosecchi et al. 1993;Carpentier et al. 2007;Gozlan et al. 2010;Didenko and Kruzhylina 2015;Povž 2017;Lojkásek and Lusk 2018;Spikmans et al. 2020). This species can feed upon the mucus of larger fish species, causing wounds and infections (Oțel 2007). P. parva preys upon eggs and larvae of amphibians when introduced to ponds where these reproduce, but so do many native fishes (Hartel et al. 2007). Therefore, non-native fishes are more of a threat for native amphibians than native fishes only inasmuch they are likelier to penetrate habitats previously devoid of fish, where non-fish-tolerating amphibians do reproduce, or inasmuch as they are more effective predators of amphibians than native species and it is unclear whether P. parva meets such criteria. Occasionally some amphibians -i. e. Triturus cristatus -can turn the tables and feed upon the introduced P. parva, if only on dead ones (Iftime and Iftime 2011).

Fam. Serrasalmidae
Piaractus brachypomus (Cuvier, 1818). Common name(s): English: Pacu, Pirapitinga (Froese and Pauly 2019); Romanian: pacu (GherlaInfo 2019a). The species is native to South America (Froese and Pauly 2019). Specimens were recently (2017-2019) captured in the Danube and Lake Firiza, Maramureș county, as well as in the Someș, at Gherla; they were at first taken for the similar-looking, feared piranha fishes, from the same family (Tarnovschi 2019; GherlaInfo 2019a). They are probably released from aquaria as has happened in other European countries (Piria et al. 2017) -at least in one case this was confirmed (GherlaInfo 2019a).

Fam. Ictaluridae
Ameiurus nebulosus (Lesueur, 1819). Common name(s): English: Brown Bullhead (Froese and Pauly 2019); Romanian: somn pitic, somn american (Wilhelm 2013; Froese and Pauly 2019). The species is native to North America (Froese and Pauly 2019). It was introduced in Lake St. Ana (Harghita county) in 1908, but also appeared in Romania through secondary dispersion from other European introductions, especially in the western-draining rivers (Tisa catchement, Timiș) but also in the lower Danube, Argeș watershed and the lakes in Bucharest (Bănărescu 1964;Ardelean and Bere, 2000;Stănescu and Gavriloaie 2011;Gavriloaie and Berkesy 2013;Wilhelm 2013; Iftime and Iftime, pers. obs., 2020 - Fig. 1; identification after the morphological criteria as given by Nowak et al. 2010, Rutyaková et al. 2012, and Wilhelm 2013. Romanian studies indicate that it has no negative impact upon native fishes, at least not through direct consumption (Gavriloaie and Berkesy 2013, and literature quoted therein). In neighbouring countries it has been repeatedly pointed as detrimental to other fishes (Lenhardt et al. 2011;Movchan et al. 2014, and literature quoted therein), though there are opinions that its impact is unknown (Povž 2017).  (Popa et al. 2006) and Bucharest lakes (Iftime and Iftime, pers. obs., 2021 - Fig. 2; identification after the morphological criteria as given by Nowak et al. 2010, Rutyaková et al. 2012, and Wilhelm 2013. According to some studies it might displace A. nebulosus, at least partly (Wilhelm 2013;Movchan et al. 2014) but in Romania the two appear to coexist widely (Telcean and Cupșa 2009).

Ameiurus melas
Ictalurus punctatus (Rafinesque, 1814). Common name(s): English: Channel Catfish (Froese and Pauly 2019); Romanian: "pește-pisică de canal" (Iacob and Petrescu-Mag 2008). The species is native to North America (Froese and Pauly 2019). It was introduced unsuccessfully in Nucet fishponds in 1978; re-introduced in 1981, it still persists in reduced number in the same fishponds (Gavriloaie and Berkesy 2013). It apparently occurs in the Prut river basin, probably including Romanian waters ( Moshu et al. 2006); it is also found in the Hungarian stretch of the Crișul Repede/ Sebes Körös river, but not in the Romanian stretch (Telcean et al. 2005;Telcean and Cupșa 2009).

Fam. Clariidae
Clarias gariepinus Burchell, 1822. Common name(s): English: Sharptooth Catfish, North African Catfish (Froese and Pauly 2019); Romanian: somn african (Iacob and Petrescu-Mag, 2008). The species is native to Africa and the Near East (Froese and Pauly 2019). This species was introduced in aquaculture in Romania in 2002, together with Clarias gariepinus X Heterobranchus longifilis Valenciennes, 1840 hybrids (Iacob and Petrescu-Mag 2008), mostly making use of thermal waters (Dynavit Impex 2021; Oțel, V., pers. comm.). Nowadays C. gariepinus occurs in fishponds in the Crișul Repede/Sebes Körös river drainage; it was found in the Hungarian stretch of the same river, but not free-ranging in Romania (Telcean et al. 2005); it is doubtful whether it can survive overwinter (especially in harder winters) in European free waters (Piria et al. 2019). The similar, south-central African species Clarias ngamensis Castelnau, 1861 was published as a new introduction record after a single specimen was found dead at the edge of a lake in Bucharest, in November 2004 (Gavriloaie and Chișamera 2005). As the animal was not found alive and it is unclear how it ended up there, this being the first record for Europe (! -Iacob and Petrescu-Mag 2008) and C. ngamensis is not a widely-kept aquaculture species (Na-Nakorn and Brummett 2009), we express doubts that this species can be considered as introduced alive to any Romanian water; at best it could have been released from an aquarium and died of cold (Iacob and Petrescu-Mag 2008), but it may just as well have been discarded after purchase from a supermarket, as imported Clarias fishes were widely available as such at the time (Iftime andIftime, pers. obs., 2004-2005).

Subfam. Coregoninae
Coregonus (albula) ladogae (Pravdin, Golubev et Belyaeva, 1938). Common name(s): English: Vendace (for C. albula; C. (a.) ladogae has no specific common name) (Froese and Pauly 2019); Romanian: coregon mic (Bănărescu 1964). The species is native to Lake Ladoga, Russia (Froese and Pauly 2019). This species was introduced in 1956, together with C. maraenoides (at that time the introduced stock was labeled as C. albula ladogensis, a replacement subspecific name now considered unnecessary as the earlier ladogae was shown to be available; ladogae is now frequently considered a distinct species 4 -Fricke et al. 2021, and literature quoted therein). These were first brought into aquaculture in the fish farms at Nucet and Tarcău (Neamț county) and then introduced into some lakes and reservoirs, including Lake Roșu, a natural dam lake in Harghita county, the alpine lake Călțun in Făgăraș mountains, and the reservoirs of Bicaz, Vidraru, Vidra (Lotru) and Făerag (Bănărescu 1964;Decei 1981;Manea 1985;Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015). A Coregonus species, possibly one of these two, was also introduced to Tarnița reservoir (Cluj county) but specimens were not available in 1992-1996 (Bănărescu et al. 1999). Some authors treat them as no longer spawning in free waters (Nalbant 2003), others consider their current fate unknown (Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015). The management plan of the Cheile Bicazului-Hășmaș National Park, which includes Lake Roșu, does not mention any Coregonus species being found in that waterbody (Romanian Environment Ministry 2016).
Coregonus (lavaretus) maraenoides (Polyakov, 1874). Common name(s): English: Peipsi Whitefish (Froese and Pauly 2019); Romanian: coregon (Bănărescu 1964). The species is native to Lake Peipus (Peipsi) and associated waters between Russia and Estonia (Froese and Pauly 2019). The introduction history is the same as for C. albula; at the time the introduced stock was labeled as C. lavaretus maraenoides (Bănărescu 1964;Decei 1981;Manea 1985;Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015), a subspecies now frequently treated as a valid species (Fricke et al. 2021). Like C. albula, it was not found in Lake Roșu during the elaboration of the management plan of the Cheile Bicazului-Hășmaș National Park (Romanian Environment Ministry 2016). However, some specimens of "Coregonus lavaretus" (probably of this C. (l.) maraenoides stock) were found in 2010 in the Vidraru reservoir (Argeș county) as per the data of the "Apele Române" (Romanian Waters) National Administration for 2010 (Csaba et al. 2011). The impact of such a population upon native fishes should be minimal, especially as they, if still present, are confined to one artificial lake. Coregonus maraena is mentioned in the DAISIE database as introduced to Romania (Roy et al. 2020); this unsourced mention was probably a confusion with C. (l.) maraenoides (Stănescu et al. 2020).
Coregonus peled (Gmelin, 1789). Common name(s): English: Peled (Froese and Pauly 2019), Romanian: coregon (Iacob and Petrescu-Mag 2008). The species is native to northern Russia (Froese and Pauly 2019). It was introduced to four alpine lakes in the Retezat massif (1967)(1968)(1969)(1970)(1971)(1972)(1973) and subsequently in 1980 at the Podu Iloaiei fish farms (Iași county) (Decei 1981;Manea 1985;Cocoș 2007;Iacob and Petrescu-Mag 2008;Gavriloaie et al. 2015), and it is unclear whether it survived. At least some introductions were unsuccessful (Decei 1981;Demeter and Mori 2004;Cocoș 2007), and of the four lakes stocked with C. peled only one (Zănoaga) had recently any fish (Cogălniceanu et al. 2012). However, the lake was stocked with other species as well, and the survival of C. peled was considered unlikely even in 1981 (Decei 1981). C. peled was considered to have no negative impact in Serbia (Lenhardt et al. 2011). There is a negative correlation between fish stocking and the survival of crustacean and amphibian populations in Retezat alpine lakes, but this is valid for both native and non-native species (Demeter and Mori 2004;Cocoș 2007;Cogălniceanu et al. 2012). C. peled is doubtful to be a major factor in the local extinction of amphibian (meta) populations as it is mainly planktivorous (Malbrouck et al. 2005;Salonen and Mutenia 2007); translocations of the native Salmo trutta, which is known to occasionally eat amphibians (Bănărescu 1964) might have been more effective in that respect.

Subfam. Salmoninae
Oncorhynchus mykiss (Walbaum, 1792). Common name(s): English: Rainbow trout, Steelhead (Froese and Pauly 2019), Romanian: păstrăv curcubeu (Bănărescu 1964;Iacob and Petrescu-Mag 2008). The species is native to the Pacific coast of North America and to Kamchatka, as both freshwater and anadromous forms (Froese and Pauly 2019). It was introduced in Romania around 1885, in specialized trout farms (Bănărescu 1964;Manea 1985;Iacob and Petrescu-Mag 2008;Gavriloaie and Berkesy 2013). It has been released in various montane rivers and lakes (Bănărescu 1964;Decei 1981); however, despite repeated and massive stocking and escapes from farms, only one population in a stream in Maramureș county appears to be successfully reproducing (Ardelean and Beres 2000;Iacob and Petrescu-Mag 2008;Gavriloaie and Berkesy 2013). Whatever ecological impact it has must therefore be extremely limited; apparently the large-scale release of (eventually non-reproducing) individuals did not cause a reduction in native species (Bănărescu 1993).
Salvelinus fontinalis (Mitchill, 1814). Common name(s): English: Brook trout, Brook Char (Froese and Pauly 2019), Romanian: păstrăv fântânel (Bănăresc, 1964;Iacob and Petrescu-Mag 2008). The species is native to North America (Froese and Pauly 2019). It was introduced in Romania in 1900-1906(Manea 1985Iacob and Petrescu-Mag 2008), being bred in fish farms (Nistor et al. 2012) but also stocked into upper reaches of mountain brooks, upstream from the range of native salmonids (Bănărescu 1964;Manea 1985) or into alpine lakes (Decei 1981). It survives in small numbers in some locations (Iacob and Petrescu-Mag 2008;Dudu et al. 2010;Gavriloaie and Berkesy 2013;Bănăduc et al. 2013;Imecs et al. 2014;Imecs and Nagy 2016). Apparently, it has no negative impact upon native salmonids (Bănărescu 1964;Bănărescu 1993;Gavriloaie and Berkesy 2013).  (Iacob and Petrescu-Mag 2008). The species is native to Eastern Asia (Froese and Pauly 2019). This species has expanded into Romania from introductions in western Ukraine in the late 1960s (Reshetnikov 2013); the first Romanian record was in 2001 (Nalbant 2003;Nalbant et al. 2004). Subsequently it has spread throughout the Tisa catchment, the Timiș and Aranca, the Siret and Prut, the Balasan, the Danube and most of the Danube Delta (Năstase 2007;Reshetnikov 2013;Ureche and Ureche 2016;Covaciu-Marcov et al. 2017a;Covaciu-Marcov et al. 2017 b;Năstase et al. 2019, and literature quoted therein). It was recently found in the Black Sea as well, at the Ukrainian coast, suggesting a higher tolerance for salinity (and capacity for spreading) than previously believed (Kvach et al. 2021). It appears to be quite damaging to amphibian populations and even to some fishes (Reshetnikov 2013, and literature quoted therein); in Romania it was found to feed on newts, namely on Lissotriton vulgaris (Telcean and Cicort-Lucaciu 2016), although it can coexist with this species, probably on a short term, before it eliminates it (Covaciu-Marcov et al. 2018). It presents a more severe threat than other introduced fishes as it can induce local extinctions of Triturus dobrogicus (Reshetnikov 2013) a species which otherwise is, unlike other European large newts, generally tolerant of fishes in its reproduction waterbodies (Arntzen and Wallis 1999).

Subfam. Macropodusinae
Macropodus opercularis (Linnaeus, 1758). Common name(s): English: Paradisefish (Froese and Pauly 2019), Romanian: pește paradis (Iacob and Petrescu-Mag 2008). The species is native to Eastern Asia (Froese and Pauly 2019). It is found in thermal waters at Băile Felix (Bihor county), probably as a release from aquaria, date unknown (Iacob and Petrescu-Mag 2008). The quoted authors stated that the species did not adapt very well to the new conditions and was reproducing poorly, suffering from the effects of the cold air (Iacob and Petrescu-Mag 2008). It was also found in one of the thermal lakes on the course of the thermal brook Pețea, close to Băile Felix, Bihor county (Iacob and Petrescu-Mag 2008;; here the species probably died out with the massive dessication, brought about by the excessive diverting of thermal water, that devastated the native species in Pețea in 2012-2014 (Müller et al. 2018). Still, we have found specimens persisting in some small basins in one public park in Băile Felix through 2012-2015 (Iftime and Iftime, pers. obs.).  (Lozinsky, 2009). The species is native to South-East Asia (Froese and Pauly 2019). It was supposed to be present (but not actually found) as an introduction in one of the thermal lakes on the course of the thermal brook Pețea, close to Băile Felix, Bihor county (Lozinsky 2009). Other authors state that it was released frequently there, but doubt its overwinter persistence . The grounds for surmising its introduction are unclear, since the species was not actually found (Lozinsky 2009;); maybe the authors had knowledge of releases from the aquarists' community. Whatever the situation, even if it was present it could not have persisted after the dessication (Müller et al. 2018).

Subfam. Trichogastrinae
Trichopodus trichopterus (Pallas, 1770). Common name(s): English: Threespot Gourami (Froese and Pauly 2019), Romanian: gurami (Iacob and Petrescu-Mag 2008). The species is native to South-East Asia (Froese and Pauly 2019). It is found in thermal waters at Băile Felix (Bihor county), probably as a release from aquaria, date unknown (Iacob and Petrescu-Mag 2008). The species is probably reproducing there, specimens from Băile Felix apparently being crosses between two colour forms commonly bred in aquaria (Iacob and Petrescu-Mag 2008). We have found them persisting through 2012-2015 (Iftime and Iftime, pers. obs.). It was also found in one of the thermal lakes on the course of the thermal brook Pețea, close to Băile Two more species of this family were tentatively identified by us in a basin in Bucharest: Chindongo demasoni (Konings, 1994) (or C. demasoni X Labidochromis caeruleus Fryer, 1956 hybrid). Common name(s): none in either English or Romanian (Froese and Pauly 2019). The species is native to Lake Malawi in Africa (Froese and Pauly 2019). Temporarily introduced from aquaria to the same small basin in Bucharest as the Pethia species mentioned above; in September 2019 they were present but did not survive (or were captured and rescued) as the basin was emptied for the cold season. They were few in numbers, with no signs of reproduction (Iftime and Iftime, pers. obs.).

Lithochromis rufus Seehausen et Lippitsch, 1998. Common name(s): none in either
English or Romanian (Froese and Pauly 2019). The species is native to Lake Victoria in Africa (Froese and Pauly 2019). Temporarily introduced from aquaria to the same small basin in Bucharest as the Pethia species and Chindongo; in September 2019 they were relatively numerous but did not survive (or were captured and rescued) as the basin was emptied for the cold season. There were no indications of reproduction (Iftime and Iftime, pers. obs - Fig. 3).
Ord. Cyprinodontiformes  mosquito larvae and thus contribute to the control of malaria. The species was populated in Lake Mangalia, Constanța county (Bănărescu 1964) and/or Mlaștina Hergheliei/Balta Mangalia, where it survives (Popovici and Jianu 2006), also in other littoral lakes and in waterbodies in and nearby Bucharest and Oradea (Călinescu 1938;Bănărescu 1964;Papadopol et al. 2012;Gavriloaie and Berkesy 2013), is present in the lakes in Nicolae Romanescu Park in Craiova (Iftime and Iftime, pers. obs., 2020 - Fig.  4) and was even found in the main channel of the Danube (Polačik et al. 2008). However, at least in some places it did not persist; it was present in the late 1990s-early 2000s in the lake in Tineretului Park, Bucharest, but was no longer found by us in 2020-2021 (Iftime andIftime, pers. obs., 1990-2021). It appears to be constrained by temperature to some extent, and is not tolerant of cold winters, long-term populations being thus mainly limited to Mediterranean and sub-Mediterranean areas, or urban settings with a favourable microclimate (see Landeka et al. 2015;Harka and Szepesi 2016); in Romania it appears, indeed, to have succeeded in the milder climate of the littoral, in thermal or thermal-influenced waters (e.g. at Mangalia - Popovici and Jianu 2006) and in some city parks. It is still anthropogenically bred and vehiculated, and reproduces well, even explosively on occasions (Papadopol et al. 2012). This species can have a significant detrimental impact upon native fish and amphibian species in some settings, e.g. in the Mediterranean area, as they are apparently capable of penetrating more amphibian breeding-grounds than native fishes (see, e.g.  Xiphophorus maculatus (Günther, 1866).  (Oțel 2007). It reproduces freely (Oțel 2007). Its spread has coincided with, and may have caused, a reduction in the populations of native mugilid species (Yankova 2016, and literature quoted therein).

Fam. Centrarchidae
Lepomis gibbosus (Linnaeus, 1758). Common name(s): English: Pumpkinseed (Oțel 2007; Froese and Pauly 2019), Romanian: biban soare, sorete, regină (Oțel 2007). The species is native to North America (Froese and Pauly 2019). It was introduced as an ornamental fish to Western and Central Europe (France and Germany) in the late XIXth century, whence it has spread eastward through the catchements of great rivers, notably the Danube. In Romania it seems to have appeared in 1918 (Iacob and Petrescu-Mag 2008;Gavriloaie and Berkesy 2013;Bănăduc et al. 2014, and literature quoted therein). It reproduces freely and has spread throughout the country, in most waters except those of the montane areas (Fig. 5), aided by fishermen and other enthusiasts who transported the species from one waterbody to another, as an ornamental fish or for the purpose of angling (Bănărescu 1964;Oțel 2007;Iacob and Petrescu-Mag 2008;Gavriloaie and Berkesy 2013). Some data from Romania suggest no negative influence of L. gibbosus on native species, at least not as regards the trophic spectrum (Bănăduc, unpubl., ap. Gavriloaie and Berkesy 2013), though others consider it damaging to native species (Bănărescu 1964) -but not a competitor of the native, ecologically similar Perca fluviatilis (Bănărescu 1993). Detrimental effects upon native species are reported in other countries, mostly in the Iberian peninsula (Yankova 2016, and literature quoted therein; Zięba et al. 2019, and literature quoted therein). It is among the species that are negatively associated with reproductive success in amphibians when introduced to ponds where these reproduce (Hartel et al. 2007).

Fam. Ranidae
Pelophylax (ridibundus) kurtmuelleri (Gayda, 1940). Common name(s): English: Balkan Water Frog (e.g., Kolenda et al. 2017). This taxon is considered native to most of Greece, Albania, Northern Macedonia, Montenegro, southern Serbia, south-western Bulgaria (Lukanov et al. 2015;Kolenda et al. 2017;Vucić et al. 2018), Figure 6. Micropterus salmoides, Bucharest, lake in National Stadium Park, 2020, photo A. Iftime. but widely introduced across Europe, probably mainly together with fish stocks (see, e.g., Kolenda et al. 2017;Bisconti et al. 2019). In Romania, a few specimens were determined through molecular protocols to pertain to Pelophylax kurtmuelleri (as opposed to the widespread Pelophylax ridibundus) (Marosi et al. 2014;Vucić et al. 2018). For some of these the collecting locality is not given (Marosi et al. 2014), and another was identified by Plötner et al. (2008) from the Sfântu Gheorghe arm of the Danube near Murighiol (Tulcea county), determined as P. ridibundus (see Plötner et al. 2008, Appendix S1) and later as P. kurtmuelleri (Vucić et al. 2018;see Litvinchuk et al. 2020 -and here it is worth mentioning that the results differ from Plötner et al. 2008 for one of the markers and in some of the specimens, which raises certain doubts with respect to the relevance of such analysis). Thus, P. kurtmuelleri (or its allelic markers, naturally encompassing at least part of the same polymorphism as P. ridibundus -i. e. not very specific -or derived from backcross hybridization with P. ridibundus) may be native, naturally spread in Eastern Europe, north of the "classic" areal (Herczeg et al. 2016;Kolenda et al. 2017;Litvinchuk et al. 2020). Moreover, P. kurtmuelleri and P. ridibundus (and European Pelophylax bedriagae) are sometimes seen as biologically conspecific, with no reproductive barrier (Hotz et al. 2013, on a molecular basis only) and sometimes as distinct, with P. kurtmuelleri -P. ridibundus F2 hybrids unviable and backcross individuals (F1 x parental species) sterile (Plötner et al. 2010). Clearly, hybridization is possible (Dufresnes et al. 2017) but this is most likely impaired by strong (but not total) hybrid dysfunction, as per Plötner et al. 2010, while the results of Hotz et al. 2013, may be explained by low specificity of the markers used, coupled with a limited amount of gene flow through the few fertile hybrids (see also the discussion by Papežík et al. 2021, including the gene flow hypothesis). P. kurtmuelleri and P. ridibundus widely overlap in morphological and morphometric characteristics (Papežík et al. 2021). The calls of P. kurtmuelleri are distinct, especially when in sympatry, from those of P. ridibundus -but whatever the song differences, even valid species do not segregate by song type and hybridize despite differences in calls (Lukanov et al. 2015). Therefore, P. kurtmuelleri, even if often treated as part of P. ridibundus (Speybroeck et al. 2016;Speybroeck et al. 2020) is most probably a valid species, but its Romanian records may represent either misidentified (due to low marker specificity) P. ridibundus, or naturally occurring recombinants. Hence the doubtful status of P. kurtmuelleri both as a Romanian record and as a non-native species in Romania.

Fam. Gekkonidae
Mediodactylus danilewskii (Strauch, 1887). Common name(s): English: Bulgarian Bent-toed Gecko (Speybroeck et al. 2020), Romanian: "gecko cu degete subțiri" (Panaitescu 2019a). The species is native to the Eastern Balkans, Anatolia, Crimea and some Greek islands (Kostakiozi et al. 2018). It was introduced (possibly vehiculated with merchandise and/or transit minibuses) in Bucharest, on dilapidated buildings and adjacent trees; a single specimen was recorded in 2018, then a population was found in 2019 (Panaitescu 2019a 5 ; Cioflec 2019a; see also discussion at Gherghel andTedrow 2019, andKoynova et al. 2020) which survived the next winter, and predation by numerous free-ranging cats (Cioflec 2020); adults and subadults were observed by us (Iftime and Iftime pers. obs., 2019-2020 - Fig. 7). Morphological characteristics (Tzankov 2007;Ajtić 2014) are sufficient for the identification of M. danilewskii in the context of treating forms in the M. kotschyi complex as full species (Kostakiozi et al. 2018). Its introduction has no detrimental consequences, and it has not been observed to spread beyond urban settings after introductions (Urošević et al. 2016).
It is considered a potential threat for other lizards if introduced into their habitat, but examples of serious impact have so far been from Mediterranean islands with endemic native Podarcis species with close ecological niches to that of P. siculus (see Adamopoulou and Pafilis 2019, and literature quoted therein). In our opinion it should be further studied whether on Mediterranean islands poor habitat bioproductivity contributes to increased competition and therefore increased impact of P. siculus introduction. It is vulnerable to cat predation, an introduced population in Athens being depleted as much from stray cat predation as from a human eradication project (Adamopoulou and Pafilis 2019).
7 Elaphe (quatuorlineata) sauromates is native to Romania. The separate species status for E. quatuorlineata and E. sauromates, as opposed to both being subspecies of E. quatuorlineata, has been long argued (Helfenberger 2001;Lenk et al. 2001;Utiger et al. 2002) and widely accepted (e.g. Uetz et al. 2020, Speybroeck et al. 2020) but, as also discussed by Iftime, 2010, reproductive isolation/ incompatibility between the two forms has not been actually investigated, quoted authors advocating for the split on the grounds of genetic distance and parapatric distribution with limited contact. Therefore, our position is that, pending data that would demonstrate such isolation/incompatibility, the split should be taken under reserve.
In the G. pseudogeographica species complex, of interest for us, a G. geographica X G. ouachitensis F1 hybrid female was apparently sterile (Vogt 2018) and there appears to be a decrease in fitness in G. geographica X G. pseudogeographica hybrid females (Freedberg and Myers 2012;Mitchell et al. 2016), but there is no indication (so far) for any incompatibility between G. pseudogeographica and G. ouachitensis (Freedberg and Myers 2012;Vogt 2018 but also in Ploiești and Piteşti (Iftime and Iftime, pers. obs., 2020 -Fig. 11). There is no evidence of reproduction so far.
Pseudemys floridana (Le Conte, 1830) 9 .  Spinks et al. 2013, presenting a molecular phylogeny failing to produce well-defined Pseudemys species clusters (which, however, may reflect either hybridization or poor marker resolution). Following the biological species concept might, in the light of the few data on this subject, favour the b) option, as some degree of reproductive incompatibility was inferred between P. concinna and P. floridana (Crenshaw 1965) and there appears to be extensive intergradation between introduced (possibly ca. 2016) in the same way as Trachemys scripta, and is found in Bucharest (Iftime and Iftime, pers. obs., 2021 -Fig. 13). There is no evidence of reproduction so far.
P. floridana and P. peninsularis (Aresco 2004) -see also Thomas and Jansen 2006, who follow variant b), but also conclude that more data are needed for a definite resolution of this problem.
There is no evidence of reproduction so far.
There is no evidence of reproduction so far.

Discussion
Our tally differs from that of Stănescu et al. 2020, which use two cathegories of alien species: "casual", i.e. those "that are surviving in the new habitat but are not spreading", and "established", i. e. "reproducing and spreading, with some becoming invasive" (Stănescu et al. 2020), and list 29 fish species, 18 of them "casual" and 11 "established", and two reptiles, both "casual". However, they include as "casual" species no longer present in Romania (Parabramis pekinensis, Megalobrama terminalis) or for which there is no indication of even transient survival in free-ranging condition (Clarias ngamensis) while neglecting species which were at least found alive together with Parabramis and Megalobrama (Ochetobius elongatus, Xenocypris macrolepis and other Asiatic species with the same introduction history, see above), had/still have self-sustaining populations for some time (Mediodactylus danilewskii) or which did survive for some time (and may still be found) free-ranging in Romanian waters (Acipenser baerii), and rejecting those definitely established and reproducing, but only in a thermal environment (e.g. Macropodus opercularis and other aquarium fishes, see above). Also, they overlook significant evidence towards the native status of Carassius (auratus) gibelio (see, e.g., Oțel 2019) and treat it as alien without even mentioning the alternative perspective, which we consider to be better argued. Clearly, there is need of more detailed discussion and at the same time more consistent criteria 11 . By Table 1. Alien species (fishes, amphibians, reptiles) found in Romania, with discussion of established and invasive status. our definitions ("introduced", even if known by only one record, and "established/ acclimatized", i.e. showing free-range reproduction) the list of non-native species is much more comprehensive (49 fishes, including one hybrid, 18 reptiles and one amphibian [ Table 1], Pelophylax (ridibundus) kurtmuelleri, the presence and status of which are, however, doubtful, for the reasons discussed above and was therefore excluded from statistics) but only a limited number can be consider acclimatized, i.e. as reproducing unassisted in free ecosystems and persisting over the whole seasonal climatic spectrum (16 fishes and 3 reptiles -including cases of probable reproduction and species limited to thermal waters). Clear-cut empirical evidence for detrimental impacts upon native species, as required for the scientific (e.g., IUCN 2017) or legal (see OUG 57/2007, L49/2011 for Romanian law on the subject) definition of invasive species, exists only for one species (Perccottus glenii) and that is (for now) from outside Romania. Gambusia holbrooki is clearly detrimental for native species in other settings but is probably thermally constrained in Romania, for now at least, while a few other species (Pseudorasbora parva, Ameiurus nebulosus, A. melas, Planiliza haematocheila, Lepomis gibbosus) may have a detrimental impact upon some native species but this is debated or as yet unclear (see Fig. 18); moreover, fish impact upon amphibians is not limited to non-native fishes and is a recurrent factor in amphibian ecology (see, e.g., Hartel et al. 2007;Semlitsch et al. 2015). Other species are widely suspected of a detrimental impact, probably because of their "alien" status (Hypophthalmichthys molitrix, Trachemys scripta) but so far this has not been substantiated by a clearly associated decline of any native species. As for species known to have a detrimental impact in other areas, the uniqueness of local ecological conditions in Romania may have its consequences and has to be taken into account. Thorough ecological studies of the impact of non-native species should therefore be performed before any non-native species be considered "invasive" or "a pest", and therefore a target of active management measures. However, even before such studies, maximal caution would suggest a "defensive" management that would ensure keeping local "insurance" populations of endangered species outside of the potential invasion range of non-native species potentially detrimental to them -and of course when such detrimental effects can be suspected by reasonable scientific presumtions correlating with data obtained in comparable ecological conditions (e.g. for Triturus dobrogicus as highly vulnerable to Perccottus glenii).