Cent. Eur. J. Biol. • 5(6) • 2010 • 888-893 DOI: 10.2478/s11535-010-0069-2

Central European Journal of Biology

Morphometry of non-native black bullhead Ameiurus melas from

Research Article Andrea Novomeská1,*, Vladimír Kováč1, Stanislav Katina2,3

1Department of Ecology, Faculty of Natural Sciences, Comenius University, Bratislava 842 15, Slovakia 2Department of Probability and Mathematical Statistics, Faculty of Mathematics and Physics and Informatics, Comenius University, Bratislava 842 48, Slovakia 3Department of Statistics, University of Glasgow, Glasgow G12 8QQ, UK Received 22 January 2010; Accepted 29 June 2010

Abstract: The study on the external morphology of the non-native black bullhead (Ameiurus melas) was carried out on a population from Slovakia, based on triple regression and geometrical analysis. The breakpoints distribution in distance-based morphometric characters indicated that black bullhead reached its definitive phenotype early in ontogeny. Ontogenetic changes in external morphology occurred continuously throughout the whole size-range of the sample examined, and inter-individual morphological variation was very low. Such uniformity may reflect the possible founder effect and/or little phenotypic plasticity of this non-native population. No sexual dimorphism was observed. This is the first detailed study on external morphology of the , and as suchit contributes to the assessment of phenotypic plasticity and/or overall morphological variability of black bullhead’s invasive populations.

Keywords: Ameiurus melas • External morphology • Growth changes • Biological invasions © Versita Sp. z o.o.

1. Introduction invasive species, as they can effectively adapt their body shape according to new habitats occupied The black bullhead Ameiurus melas (Rafinesque, [5-7]. Thus, the aim of this paper is to provide the 1820) is one of the non-native freshwater fish species first detailed examination of the species´ external introduced in Europe from North America. Over morphology, including its evaluation within an one century it has spread into many countries and ontogenetic context. An underlaying objective of this still continues to establish viable populations in new study is to contribute to the assessment of phenotypic areas [1]. The black bullhead shows various traits plasticity and/or overall morphological variability of the characteristic for successful invaders (tolerance to black bullhead, especially in its invasive populations. water pollution and low oxygen amount, aggressive behavior, omnivory, parental care, prolonged reproductive period). Based on its high flexibility in 2. Experimental Procedures its life-history traits [1-4], one may assume that the black bullhead has a great potential to spread into new Specimens of black bullhead (n=231; 111 juveniles, 54 environments. However, little is known on the external females, 66 males) were collected in monthly intervals morphology of this species. Environmental conditions from March to August 2007 from an oxbow of the River are usually responsible for the morphological Váh (47°48´30´´ N, 18°07´30´´ E; river km 2.1-4.1) in variability of organisms, including fishes and this Komárno (Slovakia) using fishing rods and kiddles. is likely to be one of the attributes of successful The population examined appeared in the River

* E-mail: [email protected] 888 A. Novomeská et al.

Váh oxbow just 2-3 years before the sampling was 30 mensural characters were plotted against SL [8] done. To comply with the rules for ethical treatment of and expressed in %SL to allow future interpopulation , all specimens were killed with an overdose of comparisons [9; Table 1]. In triple-regression analysis 2-phenoxyethanol, followed by immediate fixation in 4% (TRA), the null hypothesis was tested, stating that formaldehyde. development of body proportions is isometric (linear Photographs of all 231 specimens were taken regression) and not gradually allometric (quadratic using a Nikon D200 digital camera (Nikon Corporation, regression; first alternative hypothesis) or isometric Tokyo, Japan). Subsequently, 32 mensural characters with abrupt change (split-linear regression; broken (Figure 1, Table 1), including standard length (SL) stick regression, piecewise linear regression; second and total length (TL), were measured using Impor alternative hypothesis) [8]. Using an F-test [10], the 2.31E software (Kvant s.r.o., Bratislava, Slovakia). To model that provided the best fit was assessed. In the examine the patterns of relative growth, values from case of isometric development with abrupt change,

Character Mean SE SD Min Max CV

Total length (mm) 120.21 0.176 40.79 20.82 219.73 33.9

Standard length (mm) 101.80 0.149 34.62 17.60 184.95 34.0

% of SL 1. Head length 27.25 0.009 2.25 20.19 34.41 8.2 2. Preorbital distance 7.14 0.003 0.90 4.87 9.08 12.6 3. Eye diameter 4.00 0.003 0.75 2.83 6.91 18.7 4. Postorbital distance 16.81 0.005 1.21 13.77 20.90 7.2 5. Head depth 21.53 0.006 1.51 18.32 25.90 7.0 6. Predorsal distance 41.03 0.008 1.84 35.58 45.96 4.5 7. Preventral distance 49.96 0.008 1.95 45.04 57.46 3.9 8. Preanal distance 62.18 0.007 1.78 58.82 71.27 2.8 9. Postdorsal distance 56.68 0.008 1.92 50.87 62.72 3.3 10. V - A distance 12.77 0.005 1.22 9.63 15.54 9.5 11. D - A distance 36.52 0.006 1.39 32.04 40.45 3.8 12. D - adip distance 36.32 0.008 1.99 30.39 43.12 5.4 13. Adip - A distance 26.91 0.006 1.50 22.66 30.73 5.6 14. Adip - post. A distance 23.06 0.006 1.59 18.62 27.34 6.8 15. Post. adip - C fin base 19.80 0.004 1.14 16.45 23.68 5.7 16. C peduncle length 16.42 0.004 1.00 13.78 19.51 6.1 17. C peduncle depth 14.61 0.003 0.87 11.84 17.28 5.9 18. Minimum body depth 14.15 0.003 0.84 11.25 16.47 5.9 19. Body depth 29.83 0.011 2.57 24.35 36.26 8.6 20. D-fin depth 19.36 0.006 1.52 11.79 23.32 7.8 21. V-fin depth 13.88 0.005 1.34 9.44 18.35 9.6 22. A-fin depth 13.31 0.006 1.49 9.11 17.46 11.2 23. C-fin depth 25.23 0.012 2.84 16.34 31.58 11.2 24. D-fin length 9.39 0.005 1.24 7.05 15.66 13.2 25. Adip length 11.75 0.007 1.61 8.13 22.65 13.7 26. A-fin length 23.75 0.006 1.44 16.24 27.65 6.0 27. C-fin length 18.48 0.005 1.20 14.06 21.32 6.5 28. P-fin length 15.73 0.006 1.41 10.85 21.60 8.9 29. Interorbital distance 19.57 0.005 1.32 15.79 23.33 6.7 30. Head width 27.22 0.005 1.37 24.23 33.47 5.0

Table 1. Mean, standard error (SE), standard deviation (SD), range and coefficient of variation (CV) of morphometric characteristics in the black bullhead (n = 231) from the River Váh oxbow, Slovakia. Dorsal fin (D), pinna adiposa (adip), pectoral (P), ventral (V), anal (A) and caudal (C) fins are denoted by their respective abbreviations.

889 Morphometry of non-native black bullhead Ameiurus melas from Slovakia

Figure 1. Schematic illustration of the distance-based morphometric characters (lines) taken from the lateral (a) and dorsal (b) view of black bullhead from the Váh oxbow in Komárno (Slovakia): total length (1-2); standard length (1-3); head length (1-4); preorbital distance (1-5); eye diameter (5-6); postorbital distance (6-4); head depth (7-8); predorsal distance (1-9); preventral distance (1-10); preanal distance (1-11); postdorsal distance (12-3); V - A distance (10-11); D – A distance (9-11); D – adip distance (9-13); adip – A distance (13-11); adip – post. A distance (13-14); post. adip – C fin base (15-3); C peduncle length (14-3); C peduncle depth (14-16); minimum body depth (17-18); body depth (9-19); D-fin depth (9-20); V-fin depth (10-21); A-fin depth (22-23); C-fin depth (24-25); D-fin length (9-12); adip length (13-15); A-fin length (11-14); C-fin length (2-3); P-fin length (26-27); interorbital distance (28-29); head width (30-31). The dots represent 18 landmarks used in geometrical analyses.

breakpoints (or inflection points) were used to identify RWA) on the approximate tangent coordinates (centred the saltatory pattern during ontogeny. Procrustes shape coordinates) and log CS, variability To complement the TRA, geometrical shape was decomposed into orthogonal components of analysis [analysis of landmarks, coordinates-based size-and-shape variation, which were examined. measurements in a 2-dimensional space [11] was used. Interpretation of the principal components (PCs) A total of 18 landmarks (Figure 1) in 224 specimens followed Mitteroecker et al. [16] and Katina [13], where (7 specimens contained outlying landmarks) were the effects of each PC can be visualized (i.e. back- considered using the ImporPro 3.2 software (Kvant projected into the configuration space). Usually, the PC1 s.r.o., Bratislava, Slovakia) and evaluated in R software reflects changes in external morphology associated [12] with programme routines by Katina [13]. Three with growth changes (the allometry). In all three size size intervals were assessed according to breakpoints intervals, the correlation of PC1 scores and log CS identified by TRA (Figure2 ). Subsequently, the landmark- (calculated by Pearson product moment correlation point coordinates were transformed to Procrustes shape coefficient) was higher than 0.999. This indicates that coordinates using generalized Procrustes analysis [14]. PC1 shape changes were equivalent to those modelled First, the centroid of each form was found, as well as its by MMLRM. Finally, the MMLRM of log CS on shape root mean square distance to the landmarks (i.e. centroid variables (Procrustes coordinates) were used and the size, CS). Second, the forms were rescaled to CS=1 significance was Bonferroni corrected. with the centroids superimposed and the forms rotated In the laboratory, sex was determined by dissecting around one another about the common centroid until adult specimens. To assess sex dimorphism in black the sum of squared distances between corresponding bullhead, a permutation Goodall F-test for shape landmarks is a minimum over all such rotations. This (number of permutations = 999) and a permutation resulted in Procrustes shape coordinates, which were chi-square test for size (number of permutations = 999) then averaged across the pooled sample. These shape were used. coordinates were augmented by log CS, in order to render the form space (i.e. size-and-shape space). A permutation two-sample Goodall F-test (number of 3. Results and Discussion permutations 999) was used to test for differences in Procrustes mean shapes between the three intervals The SL of black bullheads ranged from 17.6 to 184.9 mm (i.e. 1 vs. 2, 1 vs. 3 and 2 vs. 3), expressed in P-values. (juveniles 17.6-101.8 mm, females 101.7-184.9 mm, To assess the shape changes based on the size males 97.9-165.6 mm), while the range of TL was changes, a multivariate multiple linear regression model 20.8-219.7 mm (Table 1). No significant difference was (MMLRM) of shape (Procrustes shape coordinates) on found either in shape (P-value = 0.169) or size (P-value size (CS) was used [15]. In the form-space Principal = 0.200) of females and males, and thus no sexual Component Analysis (PCA, relative warps analysis, dimorphism was observed.

890 A. Novomeská et al.

The ten mensural characters were best explained The breakpoints distribution, with the first breakpoint by linear regression, indicating isometric development. found as late as at 71.4 mm SL, indicates that black Seven characters showed allometric development bullhead reaches its definitive phenotype at a relatively (quadratic regression) and 13 mensural characters small size, i.e. early in ontogeny. It was also found that developed isometrically up to a certain SL (breakpoint), the developmental changes in external shape increased where they shifted to allometric development for a short with the size of fishes – no significant variation between time (defined by the standard error of breakpoints) small specimens (juveniles) was detected. The same and then continued to develop isometrically (split- pattern was observed in pumpkinseed Lepomis gibbosus linear regression; Table 2). Three size intervals during [5,6]. Such a relatively late occurence of breakpoints, the course of ontogeny were assigned according together with their relatively even distribution up to an to breakpoints distribution over SL (Figure 2). SL of 159.9 mm (Figure 2, Table 2), suggests that the The first interval included only juvenile specimens up to 71.4 mm SL (n=38) and did not contain any breakpoint.The second interval included specimens with the size range of 71.4–112.8 mm SL (n=89) and 6 breakpoints, while the third interval included specimens ≥114.7 mm SL (n=104) and 7 breakpoints (Figure 2). The geometrical analysis indicated that in all three intervals the affine growth changes (i.e. changes expressed as horizontal or vertical stretching in overall body shape; PC1) predominated. In the first interval, PC1 represented 88.4%, in the second interval 90.1% and in the third interval 81.9%. Total growth changes (allometry) during ontogeny represented as much as 96.69% (Figure 3) of overall variability, whereas the non-allometric inter-individual variability was very small (3.31%). Main morphological changes during ontogeny occurred in the head area (head length increased, eye diminished and moved anteriorly). Apart from that, the predorsal distance increased and the dorsal fin length, as well as the dorsal fin-adiposa distance, decreased while the abdomen area grew larger. The differences between Procrustes mean shapes of the three intervals were not identical. The P-values between the intervals Figure 2. Estimates of the breakpoints and their standard errors 1–2, 1–3, and 2–3 were 0.001, 0.002 and 0.004, for the mensural characters demonstrating significant abrupt changes in slope when plotted against SL respectively (α=0.05). (according to Kováč et al., 1999, [8]). Three intervals can be distinguished. The numbering of the characters corresponds to their order in Table 1 and 2.

Figure 3. Form-space Principal Components (PC) analysis (size-and-shape space PC analysis, relative warps analysis, RWA) and thin-plate spline (TPS) grids for the black bullhead from Slovakia. (a) TPS grid illustrating the mean-to-juvenile shape change in PC1; (b) PC1 (96.69% of variability) and PC2 (1.57% of variability) scores, juvenile-to-adult growth from left to right; (c) TPS grid from mean-to-adult shape in PC1. The difference between panels a and c represents the body shape changes during ontogeny (allometry). The low remaining non-allometric variability (3.31%) indicates that the population examined is morphologically almost uniform.

891 Morphometry of non-native black bullhead Ameiurus melas from Slovakia

external shape of black bullhead does not indicate any Such uniformity may reflect a possible founder ontogenetic thresholds. effect and/or little phenotypic plasticity of this non- In conclusion, ontogenetic changes in external the native population. Inasmuch, to assess the overall morphology of black bullheads were found to occur morphological variability of invasive black bullhead, continuously, and the Slovak population studied further populations from various environments should appeared to show little morphological variation. be analysed.

F-test F-test F-test r2 L r2 Q r2 S P-value P-value P-value Best fit Breakpoint SE Q/L S/Q S/L

1. Head length 0.9335 0.9336 0.9366 0.34 NS 11.02 <0.01 5.68 <0.05 S 151.12 4.42

2. Preorbital distance 0.9032 0.905 0.8952 4.30 <0.05 -21.00 NS -8.55 NS Q

3. Eye diameter 0.8194 0.8281 0.829 11.48 <0.01 1.18 NS 6.34 <0.05 Q 4. Postorbital 0.9661 0.9661 0.9661 0.00 NS 0.10 NS 0.05 NS L distance 5. Head depth 0.9556 0.9567 0.9562 5.76 <0.05 -2.26 NS 1.71 NS Q

6. Predorsal distance 0.9864 0.9865 0.9872 1.68 NS 12.35 <0.01 7.05 <0.01 S 150.09 4.09

7. Preventral distance 0.9832 0.9837 0.9837 6.96 <0.01 0.72 NS 3.83 NS Q

8. Preanal distance 0.992 0.9924 0.9926 11.94 <0.01 7.07 <0.01 9.67 <0.01 S 88.52 7.83 9. Postdorsal 0.988 0.9883 0.9881 5.82 <0.05 -2.21 NS 1.76 NS Q distance 10. V - A distance 0.9248 0.926 0.9220 3.68 NS -11.39 NS -3.95 NS L

11. D - A distance 0.9835 0.9839 0.9842 5.63 <0.05 5.37 <0.05 5.56 <0.05 S 132.74 7.91

12. D - adip distance 0.9654 0.967 0.9676 11.00 <0.01 4.64 <0.05 7.91 <0.01 S 155.35 3.62

13. Adip - A distance 0.9713 0.9732 0.9740 16.09 <0.01 7.31 <0.01 11.92 <0.01 S 86.83 6.90 14. Adip - post. A 0.9528 0.9535 0.9540 3.41 NS 2.76 NS 3.10 NS L distance 15. Post. adip - C fin 0.9705 0.9727 0.9730 18.29 <0.01 2.97 NS 10.71 <0.01 Q base 16. C peduncle 0.9691 0.9691 0.9702 0.00 NS 8.48 <0.01 4.24 <0.05 S 153.53 4.55 length 17. C peduncle depth 0.963 0.9665 0.9676 23.71 <0.01 8.04 <0.01 16.24 <0.01 S 106.85 5.97 18. Minimum body 0.9643 0.969 0.9704 34.41 <0.01 10.99 <0.01 23.46 <0.01 S 98.27 4.99 depth 19. Body depth 0.9326 0.9326 0.9347 0.00 NS 7.47 <0.01 3.73 NS L

20. D-fin depth 0.9432 0.9439 0.9437 2.83 NS -0.56 NS 1.12 NS L

21. V-fin depth 0.9366 0.9367 0.9371 0.35 NS 1.61 NS 0.98 NS L

22. A-fin depth 0.8989 0.9048 0.9093 14.06 <0.01 11.35 <0.01 13.03 <0.01 S 120.47 5.68

23. C-fin depth 0.8853 0.8997 0.9009 32.59 <0.01 2.79 NS 17.81 <0.01 Q

24. D-fin length 0.8731 0.8804 0.8831 13.85 <0.01 5.32 <0.05 9.72 <0.01 S 77.76 6.30

25. Adip length 0.8858 0.8862 0.8851 0.79 NS -2.14 NS -0.67 NS L

26. A-fin length 0.9696 0.9697 0.9701 0.74 NS 3.31 NS 2.03 NS L

27. C-fin length 0.9568 0.9573 0.9576 2.65 NS 1.85 NS 2.26 NS L

28. P-fin length 0.9353 0.9398 0.9414 16.96 <0.01 6.33 <0.05 11.84 <0.01 S 98.20 7.06 29. Interorbital 0.9673 0.9677 0.9683 2.81 NS 4.67 <0.05 3.76 NS L distance 30. Head width 0.9848 0.9849 0.9855 1.50 NS 10.11 <0.01 5.83 <0.05 S 159.95 4.84

Table 2. Linear (L), quadratic (Q) and split linear (S) regression statistics for morphometric characters in the black bullhead from the River Váh oxbow, Slovakia. Abbreviations in morphometric characters as in Table 1.

892 A. Novomeská et al.

Acknowledgements We thank to members of the Poseidon diving club in the article. This study was supported by VEGA in Komárno for their assistance in the field, to Mária Project 1/0226/08 and Comenius University grant Plachá for her help with statistics analyses and UK/336/2009 and UK/24/2010. to Daniel Gruľa for help with graphical processing

References

[1] Ribeiro F., Elvira B., Collares-Pereira M.J., [8] Kováč V., Copp G.H., Francis M.P., Morphometry Moyle P.B., Life-history traits of non-native fishes in of the stone loach, barbatula: do Iberian watersheds across several invasion stages: mensural characters reflect the species´ life history a first approach, Biol. Inv., 2008, 10, 89-102 thresholds?, Environ. Biol. Fish., 1999, 56, 105-115 [2] Novomeská A., Kováč V., Life-history traits of [9] Snedecor G.W., Statistical methods, Iowa State non-native black bullhead Ameiurus melas with College Press, Ames, IA, 1946 comments on its invasive potential, J. Appl. [10] Sokal R.R., Rohlf F.J., The principles and practice Ichthyol., 2009, 25, 79-84 of statistics in biological research, 2nd Ed., W. H. [3] Scott W.B., Crossman E.J., Freshwater Fishes Freeman, San Francisco, CA, 1981 of Canada, B. Fish. Res. Board Can., 1973, 184, [11] Katina S., Application of allometric shape analysis 1-966 to fish growth with geometrical notes, Tatra Mt. [4] Pedicillo G., Bicchi A., Angeli V., Carosi A., Viali P., Math. Publ., 2003, 26, 323-336 Lorenzoni M., Growth of black bullhead Ameiurus [12] R Development Core Team, A language and melas (Rafinesque, 1820) in Corbara Reservoir environment for statistical computing, R Foundation (Umbria – ), Knowl. Manag. Aquat. Ecosyst., for Statistical Computing, Vienna, , 2008 2008, 1-15 [13] Katina S., Shape analysis in the light of simplical depth [5] Šumer S., Kováč V., Povž M., Slatner M., estimators, In: Barber S., Baxter P.D., Mardia K.V., External morphology of a Slovenian population of (Eds.), Systems Biology & Statistical Bioinformatics, pumpkinseed Lepomis gibbosus (L.) from a habitat Leeds University Press, Leeds, 2007 with extreme thermal conditions, J. Appl. Ichthyol., [14] Dryden I.L., Mardia K.V., Statistical shape analysis, 2005, 21, 306-311 John Wiley & Sons, New York, 1999 [6] Tomeček J., Kováč V., Katina S., The biological [15] Venables W.N., Ripley B.D., Modern Applied flexibility of the pumpkinseed: a successful Statistics with S, 4th Ed., Springer, Berlin, 2002 colonizer throughout Europe, In: Gherardi F., [16] Mitteroecker P., Gunz P., Bernhard M., Schaefer K., (Ed.), Biological invaders in inland waters: profiles, Bookstein F.L., Comparison of cranial ontogenetic distribution and threats, Springer, Dordrecht, trajectories among great apes and humans, , 2007 J. Hum. Evol., 2004, 46, 679-697 [7] Záhorská E., Kováč V., Falka I., Beyer K., Katina S., Copp G.H., et al., Morphological variability of the Asiatic cyprinid, topmouth gudgeon Pseudorasbora parva, in its introduced European range, J. Fish. Biol., 2009, 74, 167-185

893