Philippine Journal of Science 150 (4): 743-752, August 2021 ISSN 0031 - 7683 Date Received: 25 Nov 2020

Determining Shell Shape Differences in the Horse Mussels philippinarum (Hanley 1843) and Modiolus modulaides (Röding 1798) by Morphometric Analysis

Kaent Immanuel N. Uba*

Department of Fisheries Science and Technology School of Marine Fisheries and Technology Mindanao State University at Naawan Naawan, Misamis Oriental 9023 Philippines

In the present study, shell morphological variations of the horse mussel Modiolus philippinarum and Modiolus modulaides have been explored by the means of linear morphometrics for size and landmark-based geometric morphometrics for shape. Linear morphometrics revealed significant differences in shell length [t(298) = –6.29, p = 1.08 x 10–9), shell height [t(298) = 10.60, p = 1.74 x 10–22], shell width [t(298) = 2.13, p = 0.034], PAMS or posterior adductor muscle scar length [t(298) = 2.16, p = 0.032], hinge length [t(298) = 2.26, p = 0.025], umbo length [t(298) = –5.54, p = 6.73 x 10–8], and anterior length [t(298) = –5.59, p = 5.16 x 10–8] between species. However, upon the use of these morphometric characters to develop an index that will easily discriminate the species, only the relationships of shell length vs. shell width and hinge length were significant (ANCOVA, width/ length F = 18.45, p = 0.0001; hinge/length F = 7.76, p = 0.005) but an invariably high overlap between species was observed resulting in a 23.3% misclassification rate. Contrastingly, the analysis of shape variables through landmark-based geometric morphometrics revealed significant differences in shell shape between the two species [MANOVA, Wilk’s λ = 0.01, F(24,335) = 941.4, p = 3.47 x 10–291] with a 0.0% misclassification rate. Generally, M. modulaides were found to have an elongated shell while M. philippinarum was compressed and convex. Also, visualization through thin-plate spline expansion factor plots revealed that variability in shell shape between the species occurred mostly in the posterior and ventral region of the shell and was attributed to the species ecology. The findings of the present study encourage the use of geometric morphometric methods in species delineation, especially when ecotypes or sibling species are present.

Keywords: horse mussels, morphometrics, relative warp analysis, thin-plate spline

INTRODUCTION bottoms of sheltered bays while M. philippinarum is found on muddy and gravely mudflats and seagrass The horse mussels Modiolus modulaides (Röding beds (Poutiers 1998; Ozawa 2001; Uba and Monteclaro 1798) and Modiolus philippinarum (Hanley 1843) are 2020). In the Philippines, these species are commercially widespread in the Indo-West Pacific region. Both species exploited for food and aquaculture feed (i.e. freshly fed to inhabit the littoral and sublittoral up to a depth of 40 shrimps and crabs) (Napata and Andalecio 2011; Uba et al. m. However, M. modulaides are often found on muddy 2020). Specimens of M. modulaides measure up to 8 cm shell length whereas M. philippinarum is usually bigger, *Corresponding Author: [email protected] measuring up to 13.5 cm shell length (Poutiers 1998).

743 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

There has been confusion on the identification of these morphometrics to investigate shell morphological horse mussel species, as demonstrated by conflicting variations between species and the feasibility of these reports (Laureta 2008; Napata and Andalecio 2011; del techniques in distinguishing these species. Norte-Campos et al. 2020; Uba and Monteclaro 2020). Most morphological differences between species are only identifiable by experienced observers (e.g. a careful comparison of the anterior margin, posterodorsal margin, MATERIALS AND METHODS presence of periostracum hairs, etc.). Accurate species identification is the first and necessary step when one Collection of Specimens studies species biology and population and community A total of 300 individual horse mussels were randomly ecology. Unfortunately, such identification is not always collected from January–March 2018 at Looc, Romblon easy to perform. Obviously, molecular methods may be (M. philippinarum, n = 150) and Dumangas, Iloilo (M. of great help, but they are costly and may not be always modulaides, n = 150), respectively (Figure 1). The horse on hand. In many cases, researchers have to rely heavily mussel species were identified carefully by their gross on morphological traits. Traditionally, shell morphometric morphology, as described by Poutiers (1998). characters have been used to distinguish bivalve species. However, recent analyses showed that the use of indices (i.e. the arc of maximal convexity of the valve’s outline Linear Morphometric Analysis and the index of convexity ratio) for species delimitation For the linear morphometric analysis, seven inflates the number of species as different ecophenotypes morphometric measurements were recorded for each are identified as separate species (Bolotov et al. individual, which generally followed the description 2015; Klishko et al. 2014, 2016, 2017a and b, 2018). of Maas et al. (1999) (Table 1). Linear morphometric Furthermore, traditional morphometric measurements measurements were taken from the left valve of the (length, width, height) are rarely sufficient because of the shell and measured to the nearest 0.01 cm using a high morphological plasticity of bivalve shells and linear digital caliper. measurements tend to be autocorrelated with the size of Descriptive statistics including mean (x̄ ), range (Zelditch et al. 2012). (minimum and maximum value), standard deviation At present, the detection of subtle differences in shape (SD), and coefficient of variation (CV = SD/x̄ ) of each can be done through geometric morphometric analysis. morphological character were calculated using EXCEL Because no deformation of shape occurs during 2019 (Microsoft Corporation, Redmond, WA, USA). manipulation, bivalves are excellent candidates for shape The relationships between these variables were analysis using geometric morphometrics which can be obtained by regression analysis fitting the linearized easily done using landmarks. A landmark is a homologous form [log10Y = log10(a) + blog10(X)], where X is shell point in the shell, which can be identified in all specimens. length in cm and Y is the shell height, shell width, It can be further classified into anatomical landmarks PAMS length, hinge length, umbo length, or anterior (points that correspond between organisms in some length in cm of the power function (Y = aXb). In each biologically meaningful way), mathematical landmarks relationship, differences between the estimated slopes (located on an object according to some mathematical or for each species were compared through ANCOVA geometrical property, e.g. high curvature or an extreme (analysis of covariance). Whenever the results point), and pseudo-landmarks (constructed points on indicated significant differences between the slopes the an object either on the outline or between landmarks) regressions were estimated separately for each species. (Dryden and Mardia 1998). Geometric morphometrics provides statistically robust and visual methods for To test the effectiveness of these characters in predicting shape analysis as it can remove the factor of size and different group locations, which clarify the relative symmetry which enables the study of its variation and importance of such traits as discriminators between relationships (Zelditch et al. 2012). Thus, it has been a priori for groups, a linear discriminant analysis successfully used to discriminate species of aquatic was performed using morphometric measurements organisms (Kennedy and Haag 2005; Rufino et al. 2006; previously transformed with Burnaby’s method for Torres and Santos 2009; Leyva-Valencia et al. 2012; size correction. The classification success rate was Fang et al. 2018; Santos et al. 2019). evaluated based on the percentage of individuals correctly assigned into the original sample by a The present work aimed to determine differences in shape confusion matrix (Kong et al. 2007; Konan et al. 2010; between M. philippinarum and M. modulaides based Gou et al. 2017). on linear measurements and landmark-based geometric

744 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

Figure 1. Location of the horse mussel collection sites.

Table 1. Definitions of linear morphometric characters used in a measurement reference. The images were compiled, the study. scaled, and digitized using the tpsDig version 2.12 (Rohlf Morphometric 2008a) and tpsUtil version 1.44 (Rohlf 2009) software. Description characters Twelve homologous anatomical landmarks in the inner left Shell length Widest part across the shell at 90° to the height valves were landmarked, of which seven are anatomical Shell height Distance from the hinge line to the shell margin and five are mathematical landmarks (Bookstein 1991) Shell width The thickest part of the two shell valves as illustrated in Figure 2. Landmark positions were selected according to Valladares et al. (2010) with some PAMS length Length of the posterior adductor muscle scar modifications. Landmarking per specimen was done in Hinge length Distance between the anterior and the posterior triplicates on different days to minimize measurement end of the ligament error and/or biases. Umbo length Distance from the anterior edge of the umbo to The raw coordinate configuration of all specimens was the posterior shell margin aligned (i.e. translated, rotated, and scaled to match one Anterior length Distance from the anterior shell margin to the another) through the generalized Procrustes analysis anterior edge of the umbo (GPA) procedure using the tpsRelw version 1.46 (Rohlf 2008b) to eliminate variations due to differences in scale and orientation, which establishes an average Landmark-based Geometric Morphometric Analysis configuration by minimizing the sum of squared distances For the landmark-based geometric morphometric analysis, between homologous landmarks from different specimens photographs of the inner left valve of the shell of 60 (Rohlf and Slice 1990). individuals per species were taken using a digital camera in the same position and orientation and with a ruler as

745 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

Figure 2. Photograph of the inner left valve of the specimen showing the position of the 12 homologous landmarks: 1) umbo, 2) start of the ligament, 3) end of the ligament, 4) dorsal margin maxima, 5) posterior adductor I, 6) posterior adductor II, 7) posterior adductor III, 8) posterior adductor IV, 9) posterior margin maxima, 10) projection 90° of landmark 3, 11) projection of landmark 4, and 12) anterior margin maxima.

Table 2. Morphometric characters of Modiolus philippinarum (n = 150) and Modiolus modulaides (n = 150). Range (min-max), mean ± standard deviation (SD), and coefficient variation (% CV). A p-value less than 0.05 indicates a significant difference between species based on independent sample t-test. Morphometric Modiolus philippinarum Modiolus modulaides p-value characters Range Mean ± SD % CV Range Mean ± SD % CV Shell length 5.75–8.72 7.43 ± 0.62 8.65 3.95–7.99 5.65 ± 0.95 10.05 1.08 x 10–9 Shell height 2.90–4.56 3.59 ± 0.31 15.81 1.75–3.61 2.56 ± 0.39 19.39 1.74 x 10–22 Shell width 2.45–4.33 3.54 ± 0.31 8.72 1.34–2.91 1.99 ± 0.38 18.84 0.034 PAMS length 0.90–1.72 1.31 ± 0.16 15.42 0.55–1.42 0.97 ± 0.19 14.03 0.032 Hinge length 2.00-3.75 2.98 ± 0.32 4.33 1.30–3.48 2.17 ± 0.46 4.68 0.025 Umbo length 5.51–8.54 7.20 ± 0.59 8.22 3.86–7.76 5.52 ± 0.91 16.52 6.73 x 10–8 Anterior length 0.06–0.57 0.23 ± 0.08 36.57 0.02–0.29 0.13 ± 0.06 44.31 5.16 x 10–8

After the GPA, the relative warps – which are the principal RESULTS components (PCs) of the covariance matrix of the partial warp scores – were computed using the unit centroid size Linear Morphometrics as the alignment-scaling method (Adams et al. 2004) Comparison of the morphometric characters between species using tpsRelw version 1.46 (Rohlf 2008b). Thin-plate revealed that M. philippinarum is significantly bigger than M. spline deformation grids were generated to visualize the modulaides, as shown in Table 2. Furthermore, it has been degree of shape difference. The geometric morphometric found that there are significant differences in the slopes of data from the x- and y-coordinates of the 12 landmarks the relationships of shell width and hinge length vs. shell were tested for significant differences between species length between M. philippinarum and M. modulaides (width/ through the multivariate analysis of variance (MANOVA, length F = 18.45, p = 0.0001; hinge/length F = 7.76, p = p < 0.05). Also, a discriminant analysis was conducted to 0.005). Therefore, only the relationships of shell width and determine the correct classification of species using the hinge length vs. shell length and the respective indices were relative warps. All statistical analyses were performed estimated for each species (Figures 3A–D). Nevertheless, using the freely available software Paleontological the overlap between species was invariably high in these Statistics version 3.18 (Hammer et al. 2001). morphometric relationships, as reflected in the indices. This indicates that the indices were not sufficient for distinguishing

746 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

Figure 3. Morphometric relationships of the shell width vs. shell length A) hinge length vs. shell length C) of Modiolus philippinarum (red dots) and Modiolus modulaides (black dots). Box plots of the shell width/shell length index B) and hinge length/shell length index D) for each bivalve species (Modiolus philippinarum = Mod phi; Modiolus modulaides = Mod mod).

the species. Linear discriminant analysis using these linear orientation of the average shape used in this study measurements, previously transformed using Burnaby’s (Figure 5). In PC1, the highest positive loading was in the method, gave a 23.3% misclassification rate. x-coordinate of landmark (LM) 5 at 0.27 and y-coordinate of LM 10 at 0.29 while the highest negative loading was in the y-coordinate of LM 6 and x-coordinate of LM 7 both Geometric Morphometrics at –0.51. On the other hand, in PC2, the highest positive Differences in the shell shape between species were loading was in the x-coordinate of LM 10 at 0.61 while the significant as revealed by the multivariate analysis of highest negative loading was in the x-coordinate of LM 11 variance [Wilk’s λ = 0.01, F(24,335) = 941.4, p = 3.47 at –0.54. This indicates that variations in the shell shape x 10–291], and linear discriminant analysis revealed between species were mostly occurring in the posterior a 0.0% misclassification rate. The results of the PC and ventral regions of the shell, as supported by the thin- analysis, wherein the first two PCs accounted for 74.8% plate spline deformation grid of the mean configuration of the cumulative variances, clearly separated the shape in both species, as shown in Figure 4. variables between the species, as shown in Figure 4. Furthermore, the degree of shape differences between The landmarks with the highest contribution to the species was visualized in the thin-plate spline expansion observed variation were elucidated in the loadings on the factor plots of the first two relative warps of the shell shape eigenvectors of the landmarks for the first two PCs, with of M. philippinarum and M. modulaides (Figure 6). In M. the x- and y-coordinates corresponding to the particular

747 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

Figure 4. PC analysis showing a significant difference in the shell shape of Modiolus philippinarum (circle) and Modiolus modulaides (square) with the corresponding thin-plate spline deformation of the mean landmark configuration.

Figure 5. Component loadings for the first two PCs of the coordinates of the 12 landmarks used in this study.

748 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

Figure 6. Thin-plate spline expansion factor plots of the first two relative warp showing the expansion and compression of the shell shape of Modiolus modulaides (A–B) and Modiolus philippinarum (C–D). Color hues represent relative proportional variation in areas of the shell (yellow to red: expansions and green to blue: compression).

modulaides (Figures 6A–B), pronounced expansion was modulaides, while using the geometric morphometric observed throughout the shell except in the mid-ventral analysis 100.0% of samples will be correctly classified. region where pronounced compression was observed, This is particularly important in the correct identification as shown in the first and second relative warps, which of the two horse mussel species. accounted for 61.0% of the variation in its shell shape. In contrast, in M. philippinarum, there are individuals with Through technological advancement, geometric pronounced compression throughout the shell except in morphometric analyses have become simpler and faster. the mid-ventral region, where pronounced expansion was Its power to define and visualize subtle features of observed (Figure 6C) while there are also individuals with shape might discriminate species or populations. Among pronounced expansion throughout the shell except in the geometric morphometric methods, landmark-based anterior and mid-lateral, and -ventral regions (Figure 6D). analysis is more preferred due to its ability to detect Both relative warps accounted for 54.5% of the variation changes in certain structures or parts of the specimen in the shell shape. under study. It has been used to determine shell shape variation in other bivalves. High discriminating power had been obtained by Rufino et al. (2006) for distinguishing Chamelea gallina and C. striatula. Moreover, species DISCUSSION of the Littorina saxatilis complex (Conde-Padín et al. 2007), Chione cancellata, and C. elevata (Roopnarine et The linear morphometric characters were found to be al. 2008), and Panopea generosa and P. globosa (Leyva- significantly different between species, but the occurrence Valencia et al. 2012) were successfully discriminated. of an extensive data overlap prevented the separation of Moreover, Innes and Bates (1999) found significant these species using either the linear measurements or the differences in the shell shape of Mytilus edulis and M. indices considered. A misclassification rate of 23.3% was trossulus using outline-based geometric morphometrics. obtained using the linear morphometric variables. On the other hand, the use of shape variables in the landmark- In terms of size, M. modulaides are smaller and slender based geometric morphometric analysis successfully compared to the big and wide M. philippinarum, as distinguished both species with a 0.0% misclassification revealed in the linear morphometric analysis. On the rate. This indicates that using the linear morphometric other hand, more meaningful information was revealed characters used in this study only 76.7% of samples in the analysis of shapes through landmark-based will be correctly classified as M. philippinarum and M. geometric morphometric analysis. It was found that M.

749 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides modulaides assumed a shorter but expanded shell shape ACKNOWLEDGMENTS (i.e. observed expansion throughout the shell) than M. philippinarum, which assumed a relatively narrow but The author is thankful to the Commission on Higher elongated shell shape (i.e. observed compression in Education K to 12 Transition Program Scholarship for the posterior and anterior regions of the shell). Shell Graduate Studies and the University of the Philippines morphological variability in mytilids is largely influenced Visayas (UPV) Office of the Vice-Chancellor for Research by environmental variables such as trophic conditions, and Extension for the scholarship grant and thesis grant. crowding, water depth, and wave impact. Moreover, it The support of Dr. Harold M. Monteclaro and the UPV- may also vary between sexes and different geographic DOST (Department of Science and Technology) Invasive locations (Steffani and Branch 2003; Krapivka et al. 2007; Mussel Project is greatly appreciated. Special thanks to Fitzer et al. 2015; Uba et al. 2019). Mr. Godwin O. Marcelino for the help in the collection of samples in Looc, Romblon. Lastly, the support of the In the natural environment, M. modulaides were observed local governments of Dumangas and Looc in facilitating to form dense mats of closely-knit individuals linked the collection of the samples is also appreciated. together with byssal threads in a silt-clay soft-bottom habitat while M. philippinarum were not observed to form mats rather individuals were detached and grew in sandy seagrass beds. This difference in species’ ecology REFERENCES influenced the differences in shell size between species. ADAMS DC, ROHLF FJ, SLICE DE. 2004. Geometric Slender shells in bivalves are reported to be a result of morphometrics: ten years of progress following the crowding (Alunno-Bruscia et al. 2001; Lauzon-Guay et ‘revolution’. Ital J Zool 7(1): 5–16. al. 2005; Caill-Milly et al. 2014; Brash et al. 2017). Also, it can be a result of wave exposure to reduce the impact ALUNNO-BRUSCIA M, BOURGET E, FRECHETTE of hydrodynamic forces (Steffani and Branch 2003). In M. 2001. Shell allometry and length-mass-density terms of shape, an elongated shell shape was observed in relationship for Mytilus edulis in an experimental food- M. philippinarum while short but expanded shells were regulated situation. Mar Ecol Prog Ser 219: 177–188. observed in M. modulaides. This may be attributed to BOLOTOV IN, BESPALAYA YV, VIKHREV IV, phenotypic differences and may not be fully attributable AKSENOVA OV, ASPHOLM PE, GOFAROV MY, as a response to environmental variables. KLISHKO OK, KOLOSOVA YS, KONDAKOV In the context of species identification, the results of the AV, LYUBAS AA, PALTSER IS, KONOPLEVA ES, present study are primarily important for the study of the TUMPEESUWAN S, BOLOTOV NI, VOROSHILOVA biology and ecology of the two horse mussel species. IS. 2015. and distribution of freshwater More importantly, it is important for the proper assessment pearl mussels (Unionoida: Margaritiferidae) of the and management of the horse mussel species and the Russian Far East. PLoS One 10(5): e0122408. comparison between horse mussel species from different BOOKSTEIN FL. 1991. Morphometric tools for geographical areas. landmark data geometric and biology. Cambrige University, Cambrige. 435p. BRASH JM, COOK RL, MACKENZIE CL, CONCLUSION SANDERSON WG. 2017. The demographics and morphometries of biogenic reefs: important This study aimed to investigate if linear shell characters considerations in conservation management. J Mar and homologous landmarks allow distinguishing M. Biol Assoc UK 98(6): 1231–1240. philippinarum and M. modulaides. Linear characters were significantly different between species but the occurrence CAILL-MILLY N, BRU N, BARRANGER M, GALLON of an extensive data overlap prevented the separation of L, D’AMICO F. 2014. Morphological trends of four these species using the indices considered. Differently, Manila clam populations (Venerupis philippinarum) landmark-based geometric morphometric analysis on the French Atlantic coast: identified spatial patterns successfully distinguished both species. The result of and their relationship to environmental variability. J this study encourages the use of geometric morphometric Shellfish Res 33(2): 355–372. analysis as it deals with subtle shape variations by CONDE-PADÍN P, GRAHAME JW, ROLÁN-ALVAREZ quantifying and visualizing shape variations, which can E. 2007. Detecting shape differences in species of the help in species delineation especially when ecotypes or Littorina saxatilis complex by morphometric analysis. sibling species are present. J Mollusc Stud 73(2): 147–154.

750 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

DEL NORTE-CAMPOS AGC, BURGOS LA, SANCHEZ (: Unionidae) in Russia and Ukraine based on KA. 2020. A field guide to the commercially-important morphological and molecular data. Zootaxa 4286(1): mollusks of Panay, Philippines. University of the 93–112. Philippines Visayas, Iloilo, Philippines. 213p. KONAN KM, ADEPO-GOURENE AB, OUATTARA A, DRYDEN IL, MARDIA KV. 1998. Statistical shape NYINGY WD, GOURENEA G. 2010. Morphometric analysis. John Wiley and Sons, Chichester. 347p. variation among male populations of freshwater shrimp Macrobrachium vollenhovenii Herklots, 1851 from FANG Z, FAN J, CHEN X, CHEN Y. 2018. Beak Cote d’Ivoire River. Fish Res 103: 1–8. identification of four dominant octopus species in the East China Sea based on traditional measurements and KONG L, LI Q, QIU Z. 2007. Genetic and morphological geometric morphometrics. Fish Sci 84(6): 975–985. differentiation in the clam Coelomactra antiquata (Bivalvia: Veneroida) along the coast of China. J Exp FITZER SC, VITTERT L, BOWMAN A, KAMENOS Mar Biol Ecol 343: 110–117. NA, PHOENIX VR, CUSACK M. 2015. Ocean acidification and temperature increase impact mussel KRAPIVKA S, TORO JE, ALCAPAN AC, ASTORGA M, shell shape and thickness: problematic for protection? PRESA P, PEREZ M, GUIÑEZ R. 2007. Shell‐shape Ecol Evol 5(21): 4875–4884. variation along the latitudinal range of the Chilean blue mussel Mytilus chilensis (Hupe 1854). Aquac Res GOU H, ZHANG D, JIANG S, LI Y, ZHANG N, WANG 38(16): 1770–1777. Y, MA Z. 2017. Morphological variations of the winged pearl oyster Pteria penguin (Roding, 1798) from South LAURETA LV. 2008. Compendium of the economically China Sea. Indian J Fish 64(1): 9–17. important seashells in Panay, Philippines. The University of the Philippines Press, Quezon City. 162p. HAMMER Ø, HARPER DAT, RYAN PD. 2001. PAST: Paleontological Statistics software package for LAUZON-GUAY JS, HAMILTON DJ, BARBEAU education and data analysis. Paleontol Electronica MA. 2005. Effect of mussel density and size on the 4(1): 1–9. morphology of blue mussels (Mytilus edulis) grown in suspended culture in Prince Edward Island, Canada. INNES DJ, BATES JA. 1999. Morphological variation Aquaculture 249: 265–274. of Mytilus edulis and Mytilus trossulus in eastern Newfoundland. Mar Biol 133: 691–699. LEYVA-VALENCIA I, ÁLVAREZ-CASTAÑEDA ST, LLUCH-COTA DB, GONZÁLEZ-PELÁEZ S, KENNEDY TB, HAAG WR. 2005. Using morphometrics PÉREZ-VALENCIA S, VADOPALAS B, RAMÍREZ- to identify glochidia from a diverse freshwater mussel PÉREZ S, CRUZ-HERNÁNDEZ P. 2012. Shell community. J North Am Benthol Soc 24(4): 880–889. shape differences between two Panopea species and KLISHKO OK, LOPES-LIMA M, FROUFE E, BOGAN phenotypic variation among P. globosa at different AE. 2014. Are Cristaria herculea (Middendorff, sites using two geometric morphometrics approaches. 1847) and Cristaria plicata (Leach, 1815) (Bivalvia, Malacologia 55(1): 1–13. Unionidae) separate species? Zookeys 438: 1–15. MAAS PAY, O’MULLAN GD, LUTZ RA, VRIJENHOEK KLISHKO OK, LOPES-LIMA M, FROUFE E, BOGAN RC. 1999. Genetic and morphometric characterization AE, ABAKUMOVA VY. 2016. Systematics and of mussels (Bivalvia: ) from mid-Atlantic distribution of Cristaria plicata (Bivalvia, Unionidae) hydrothermal vents. Biol Bull 196(3): 265–272. from the Russian Far East. Zookeys 580: 13–27. NAPATA RP, ANDALECIO MN. 2011. Exploitation KLISHKO OK, LOPES-LIMA M, FROUFE E, BOGAN and management of brown mussel (Modiolus AE, ABAKUMOVA VY. 2017a. Unravelling the philippinarum) resources in Iloilo, Philippines. Phil J systematics of Nodularia (Bivalvia, Unionidae) species Soc Sci Human 16(2): 22–34. from eastern Russia. Syst Biodivers 16(3): 287–301. OZAWA H. 2001. Reproductive cycle and spawning of KLISHKO OK, LOPES-LIMA M, BOGAN AE, Modiolus philippinarum (Bivalvia: Mytilidae) in a MATAFONOV DV, FROUFE E. 2018. Morphological seagrass bed in Kin Bay, Okinawa Island, Southern and molecular analyses of Anodontinae species Japan. Venus 60(3): 173–181. (Bivalvia, Unionidae) of Lake Baikal and Transbaikalia. POUTIERS JM. 1998. Bivalves. Acephala, lamellibranchia, PLoS One 13(4): e0194944. pelecypoda. In: The living marine resources of the KLISHKO OK, LOPES-LIMA M, FROUFE E, BOGAN Western Central Pacific Volume 1 Seaweeds, corals, AE, VASILIEV L, YANOVICH L. 2017b. Taxonomic bivalves, and gastropods. Carpenter KE, Niem VH eds. reassessment of the freshwater mussel genus Unio FAO, Rome, Italy. p. 123–362.

751 Philippine Journal of Science Uba: Determining Shell Shape Differences between Vol. 150 No. 4, August 2021 M. philippinarum and modulaides

ROHLF FJ. 2008a. tpsDig version 2.12. Department of VALLADARES A, MANRIQUEZ G, SUAREZ-ISLA Ecology and Evolution, State University of New York BA. 2010. Shell shape variation in populations of at Stony Brook, New York. Mytilus chilensis (Hupe 1854) from southern Chile: a geometric morphometric approach. Mar Biol 157(12): ROHLF FJ. 2008b. tpsRelw version 1.46. Department of 2731–2738. Ecology and Evolution, State University of New York at Stony Brook, New York. ZELDITCH M, SWIDERSKI D, SHEETS H. 2012. Geometric morphometrics for biologists: a primer. ROHLF FJ. 2009. tpsUtil version 1.44. Department of Elsevier Academic Press, California, USA. 489p. Ecology and Evolution, State University of New York at Stony Brook, New York. ROHLF FJ, SLICE D. 1990. Extensions of the Procrustes method for the optimal superimposition of landmarks. Syst Zool 39(1): 40–59. ROOPNARINE P, SIGNORELLI J, LAUMER C. 2008. Systematic, biogeographic and microhabitat-based morphometric variation of the bivalve Anomalocardia squamosa (Bivalvia: Veneridae: Chioninae) in Thailand. Raffles Bull Zool 18: 95–102. RUFINO MM, GASPAR MB, PEREIRA AM, VASCONCELOS P. 2006. Use of shape to distinguish Chamelea gallina and Chamelea striatula (Bivalvia: Veneridae): linear and geometric morphometric methods. J Morphol 267(12): 1433–1440. SANTOS SR, PESSÔA LM, VIANNA M. 2019. Geometric morphometrics as a tool to identify species in multispecific flatfish landings in the tropical Southwestern Atlantic. Fish Res 213: 190–195. STEFFANI CN, BRANCH GM. 2003. Growth rate, condition, and shell shape of Mytilus galloprovincialis: responses to wave exposure. Mar Ecol Prog Ser 246: 197–209. TORRES SKM, SANTOS BS. 2009. Species identification among morphologically-similar Caranx species. Turk J Fish Aquat Sci 20(2): 159–169. UBA KIN, MONTECLARO HM. 2020. Habitat characteristics of the horse mussel Modiolus modulaides (Röding 1798) in Iloilo, Philippines. Phil J Sci 149(3–a): 977–987. UBA KIN, MONTECLARO HM, NOBLEZADA-PAYNE MM, QUINITIO GF. 2019. Sexual dimorphism, asymmetry, and allometry in the shell shape of Modiolus modulaides (Hanley, 1843) collected from Dumangas, Iloilo, Philippines: a geometric morphometric approach. Comput Ecol Soft 9(3): 107–120. UBA KIN, MONTECLARO HM, NOBLEZADA- PAYNE MM, QUINITIO GF, ALTAMIRANO J. 2020. Value chain analysis of the horse mussel Modiolus metcalfei (Hanley, 1843) fishery in Iloilo, Philippines. Asian Fish Sci 33(2): 106–117.

752