Atlantic sturgeon Acipenser oxyrinchus and alien sturgeon species in Polish waters: can biometric analysis assist species discrimination and restoration?

Biometric character analyses were conducted on Atlantic sturgeon Acipenser oxyrinchus, a species included in a re-establishment programme in the Baltic Sea basin. The study sought to identify measurable and countable characters most useful for distinguishing A. oxyrinchus from three alien species found in open waters of Poland: Acipenser baerii, Acipenser gueldenstaedtii, and Acipenser stellatus. Measurable characters that contributed most to discrimination included preorbital distance, eye diameter, ventral fin base to anal fin base, and postorbital distance. Among countable characters, the number of lateral scutes made the greatest contribution. The data from the present study may be used to supplement existing systematic keys and lay the foundations for creating a catalogue or atlas of popular species and interspecific hybrids of sturgeon, including their complete systematic determination. Jesiotr ostronosy Acipenser oxyrinchus oraz obce gatunki jesiotrów w wodach Polski: czy analiza biometryczna uzupełni identyfikację i restytucję gatunków? Słowa kluczowe analiza dyskryminacyjna, cechy biometryczne, identyfikacja gatunków, przyłów, restytucja gatunku Streszczenie W niniejszej pracy przedstawiono wyniki analizy cech biometrycznych osobników jesiotra ostronosego Acipenser oxyrhinchus oxyrhinchus, gatunku objętego restytucją w basenie Morza Bałtyckiego. Przy wykorzystaniu technik aglomeracyjnych podjęto próbę wyjaśnienia, które z cech mierzalnych i policzalnych mają największą przydatność przy odróżnianiu Acipenser oxyrinchus oxyrinchus od trzech gatunków obcych, łowionych w wodach otwartych Polski, tj. jesiotra syberyjskiego (Acipenser baerii), rosyjskiego (Acipenser gueldenstaedtii) oraz siewrugi (Acipenser stellatus). Wyniki analiz wykazały, że największy udział w dyskryminacji na podstawie cech mierzalnych miały: długość rostrum (R), odległość między płetwami V i A (VA) oraz przestrzeń zaoczna (OP), natomiast wśród cech policzalnych była to liczba tarczek bocznych (SL). Wyniki uzyskane w niniejszej pracy wzbogacają aktualnie #0#


Material and methods
Fifty dead juvenile specimens of A. oxyrinchus were collected from seine fishers in three locations (N 53°43′59″; E 014°28′11″; N 53°44′22″, E 014°27′51″, N 53°43′10″, E 014°26′42″) in the Szczecin Lagoon near the town of Stepnica, from 21 May to 9 June 2008. The fish, some of which were tagged, originated from German stocking (Gessner et al., 2008) undertaken as part of a species restoration plan (Figure 1). Specimens were frozen using the method proposed by Keszka and Krzykawski (2008). Measurements were taken with calipers of 0.01 mm accuracy according to figure published by Kempter et al. (2013). Masses of individuals were measured with electronic scales with 0.01 g accuracy. Twenty-one measurable and five countable biometric characters were compared with characters of A. baerii (n = 336), A. gueldenstaedtii (n = 99), and A. stellatus (n = 69) (Keszka, 2000;Keszka, Raczyński, 2006). Information on measurable characters for these 3 species were published by Keszka and Raczyński (2006), Keszka and Krzykawski (2008), Keszka et al. (2009). All individuals of alien species originated from warm water aquaculture systems in Nowe Czarnowo and Gryfino (NW Poland). Biometric characters of the studied species are given in Table 1. Statistical analysis was performed using Statistica 10.1. Measurable and countable characters were summarized calculating minimum and maximum values, arithmetic mean, and standard deviation.
Per cent of total length (L T ) of measurable characters and absolute values for countable characters were used for conducting stepwise discriminant analysis and canonical discriminant analysis. Results were analysed taking into consideration the classification matrix and values of standardized coefficients for canonical variables and visualized as scatter plots of canonical scores.

Measurable Characters
Length and mass were comparable for all specimens (Table 1). Measurable characters of A. oxyrinchus presented as absolute values (mm) and as per cent calculated with respect to total length (L T ) are presented in Table 2. Among the five countable characters, standard deviation for the number of dorsal scutes and the number of ventral scutes were the lowest and therefore similar to the mean value (Table 3).  The characters showed a low variation range. In order to reduce the signal and the number of characters for further comparison, principal component analysis (PCA) was conducted for the 21 measurable characters of the 4 sturgeon species. The value of χ 2 was high (cumulative inertia: 0.999, χ 2 = 6,304.226, P = 0.000), which indicated that the species differed substantially with respect to measurable characters. Figure 2 shows distribution and direction of vectors on the graph depicted by the first two principal components of 21 morphological variables.  The PCA revealed that the measurable characters with the highest contribution to differentiation of the 4 sturgeon species included head characters along with distance from ventral fin base to anal fin base (VA) and length of ventral (pelvic) fin (lV). Characters with the highest contribution to the first two principal components were selected for stepwise discriminant analysis. All compared measurable characters were taken into account in the discriminant function model (Wilks' lambda: 0.00014 approx. F (27.2065) = 1,512.9 P < 0.0000). A classification matrix for the stepwise discriminant analysis generated for nine measurable characters of the species was characterized by a high level of correctness (99.5828% Table 4).
The classification matrix displayed the lowest correctness for A. oxyrinchus and A. stellatus, and the highest for A. baerii and A. gueldenstaedtii. Only individual specimens of A. oxyrinchus might be incorrectly classified as A. stellatus or A. gueldenstaedtii (Table 4). On the basis of both discriminant functions, canonical discriminant analysis clearly discriminated the four species with respect to their biometric characters (Figure 3).  The first discriminant function separated A. gueldenstaedtii, A. oxyrinchus, A. stellatus, and A. baerii. The second function discriminated between A. oxyrinchus and A. gueldenstaedtii and also between A. stellatus and A. gueldenstaedtii. According to the position of A. baerii on the graph, the second function situate this species between A. oxyrinchus and A. gueldenstaedtii. In the first function (root 1), character VA made the most significant contribution to discrimination, while, in the second function (root 2), the significantly contributing characters included horizontal eye diameter and preorbital distance. Postorbital distance had a significant contribution only in the first function (Table 5, Figure 4).

Countable Characters
The analysis of standardized coefficient values for canonical variables indicated number of lateral scutes as the character with the most significant contribution to discrimination (Table 6). Correctness of the classification matrix for countable characters was calculated as 79.92%, lower than for measurable characters (Table 7). In spite of there being relatively fewer A. oxyrinchus, 98% of A. oxyrinchus individuals were correctly discriminated. The highest possibility of an error based on countable characters was observed for A. gueldenstaedtii, which showed similarity of meristic characters with A. baerii, as well as the distinctiveness of the A. oxyrinchus from other species on the basis of having significantly fewer lateral scutes ( Figure 5).

Discussion
The data from the present study may be used to supplement existing systematic keys and lay the foundations for creating a catalogue or atlas of popular species and interspecific hybrids of sturgeon, including their complete systematic determination (Krylova, 1997). Correct identification is especially important in the context of species restoration. Stocking of juvenile European sea sturgeon and A. oxyrinchus as part of their re-establishment programme in Europe (Elvira, Gessner, 1996;Gessner et al., 2010) along with the co-occurrence of alien species, requires an informational campaign for users of the waters, focusing on distinguishing between native and non-native species. A reference book containing a simple and user-friendly key with sound taxonomic data is necessary. The data available for comparison with the results of the present biometric analyses were conducted on only a small number of living specimens (Artiukhin, Vecsei, 1999;Debus, 1999). When examining 27 museum specimens and a single live sturgeon, it was noted that variability of some characters of the sturgeon head, mainly due to scraping the substratum while feeding, precluded considering distances connected with snout structure as distinguishing characteristics. Comparison of dorsal and lateral scutes suggested that A. oxyrinchus differs from A. sturio from the Gironde (Atlantic Ocean) and the Rioni (Black Sea) rivers (Ninua, 1976). The analyses discussed in the present paper compared morphometric data of A. oxyrinchus, A. stellatus, A. gueldenstaedtii, and A. baerii. Among measurable characters expressed as per cent of L T , preorbital distance, distance from ventral fin base to anal fin base (VA), horizontal eye diameter (O), and postorbital distance (OP) had the highest contribution to discrimination (Figure 4). Mainly on the basis of VA, the first function clearly discriminated between A. baerii and those of A. oxyrinchus, A. gueldenstaedtii, and A. stellatus (Figure 3). A longer VA distance distinguished A. baerii from the other three sturgeon species (Figure 4). In the second function, preorbital distance contributed significantly to discrimination due to the positive value of the canonical variable coefficient; whereas, in the second function, horizontal eye diameter contributed significantly to discrimination (Table 5). On the basis of this analysis it can be concluded that the shorter the rostrum of an A. oxyrinchus individual, the higher the possibility of mistaking it for A. gueldenstaedtii. Distinguishing A. oxyrinchus from A. baerii may present difficulties due to considerable changes of the rostrum length during the lifetime of the latter, as well as the existence of short-rostrum and long-rostrum forms of A. baerii (Keszka et al., 2009). Thus, the risk of an erroneous classification is high in the case of juvenile A. baerii individuals, which have a longer rostrum than adults of the species. The risk of mistaking A. oxyrinchus for another species on the basis of measureable characters is the highest with respect to A. stellatus. The second classification function, with rostrum length having the major contribution to discrimination, clearly distinguished between the A. gueldenstaedtii and A. baerii grouping and the A. stellatus and A. oxyrinchus grouping while the first classification function separated the group of A. baerii from the groups of the remaining three sturgeon species.
Based on the five analysed countable characters, exotic species are clearly distinguishable from A. oxyrinchus. The number of lateral scutes had the highest contribution to discrimination (Table 6). In the sample, the number of scutes ranged from 21 to 30, with a mean of 25.22. The value was lower than that observed for A. oxyrinchus in the St Lawrence River in Canada (28.67) (Artiukhin, Vecsei, 1999) and considerably lower than found for exotic species occurring in Polish waters (Keszka, Heese, 2003;Keszka, Krzykawski, 2008;Keszka et al., 2009). Low variability of the countable characters in A. oxyrinchus specimens in the present study might be due to the limited number of broodstock used to produce stocking material as well as to an effect of artificial rearing conditions on juvenile fish, as had been noted in the case of other cultured sturgeon species (Ruban, Sokolov, 1986).
Exotic sturgeon introduction into European waters may occur from fish released by aquarium owners and hobbyists who want to dispose of large individuals, sturgeons intentionally released by fishing clubs and associations, and accidental escape from ponds and farms. Despite clear regulations applying to alien species in aquaculture under the European Commission Regulation (EC) no. 535/2008 of 13 June 2008 rules for implementation of the Council Regulation (EC) no. 708/2007 concerning use of alien and locally absent species in aquaculture, a fourth source of exotic sturgeon introduction involves artificial breeding and release (Britton, Davis, 2006). The occurrence of non-native sturgeon species in German and Polish coastal waters and river estuaries has been growing since the beginning of the 1990s (Spratte, Rosenthal, 1996;Keszka, Stepanowska, 1997;Arndt et al., 2000Arndt et al., , 2002Keszka, Heese, 2003;Keszka et al., 2011). The presence of exotic sturgeon species was also observed in the Gironde River in France in 1999 after a period of storms that caused water levels to rise and resulted in the escape of several thousand A. baerii from fish farms. This presented a serious problem for inland fishers and anglers in distinguishing between two sturgeon species, of which one (A. sturio) was strictly protected and the second (A. baerii) needs to be eliminated from open waters (Gessner et al., 2010). Alien species release, taking place in spite of the prohibitions connected with environmental protection, should be considered a potential threat, since alien species may have a negative effect on native species and their populations, especially in strongly altered ecosystems (Leppäkoski et al., 2002).
Apart from the above implications for alien species introduction, genetic implications cannot be dismissed, as sturgeons display a potential for hybridization (Kozhin, 1964). Current information suggests that the hybrids are fertile, which means that mixed genetic material is passed on to the next generation, resulting in a dispersal of the original genetic information and reducing adaptation ability to habitats. The existence of the first hybrids created as a result of release of exotic A. baerii has been confirmed by genetic analyses conducted on specimens collected from the Danube River (Ludwig et al., 2009).
Avoidance of loss connected with fisheries is an important principle behind protecting the remaining native populations in situ , Gessner et al., 2010), Key factors in effective restoration of a population include decreasing fishing-associated mortality and obtaining a high level of acceptance and support from the fishery sector. Measures that need to be taken include instituting and monitoring programs to limit accidental catch based on the ability of fishers to identify species. The example of A. sturio in France and preliminary results obtained from Germany clearly show that this is possible (Gessner et al., 2010).