Introduction

Tellinidae is one of the most diverse families found in the Brazilian continental shelf. Huber (2015) recognizes nine subfamilies of Tellinidae: Tellininae, Gastraninae, Phyllodainae, Moerellinae, Aenigmotellinae, Arcopaginae, Apolymetiinae, Strigillinae, and Macominae. Macominae is characterized by rounded, inequilateral shells, with strong to weak hinges, containing two cardinal teeth, being one bifid, and lateral teeth absent. The sculpture is mostly commarginal and weak, and the periostracum is usually strong. The adductor muscle scars are subequal, and there is a deep pallial sinus (Mikkelsen and Bieler 2008; Huber 2015). Macominae includes seventeen living genera divided into five groups (Huber 2015): Psammotreta, Salmacoma, Psammacoma, Austromacoma, and Macoma. On the Brazilian coast, six species have been registered to these groups: Austromacoma biota Arruda & Domaneschi, 2005, Austromacoma constricta (Bruguière, 1792), Psammotreta cleryana (d’ Orbigny, 1846 in d’ Orbigny 1834–1847), Psammotreta brevifrons (Say, 1834 in Say 1830–1834), Macoploma tenta (Say, 1834 in Say 1830–1834), and M. tageliformis (Dall, 1900) (Rios 2009). The separation of Macominae from the other subfamilies based only on the absence of lateral teeth has been considered a mistake, and the subfamily has been considered paraphyletic (Huber 2015) and even polyphyletic (Coan and Valentich-Scott 2012).

There is little information about Macominae species from the Western Atlantic, except for the surveys carried out by Tenório et al. (1986), Narchi (2003), Arruda and Domaneschi (2005), Piffer et al. (2011), and Huber (2015). The identification of Macominae and other Tellinidae species are based on shell morphology, which have diverse and homoplastic characteristics (Simone and Wilkinson 2008; Huber 2015). Despite the many limitations of using shell morphological data for taxonomy, the method remains widely used due to the need to identify through observable characters.

Therefore, taxonomists have sought different methods and tools to help them minimize difficulties with identification, one of which is morphometric analysis. This method provides statistical support to identify differences in continuous characteristics, which are often observed subjectively and not measurable. For example, Marko and Jackson (2001) studied the shell morphometry of five pairs of cryptic bivalve species of Arcidae separated by the Isthmus of Panama, while Rufino et al. (2006) separated two commercially fished Venus clams (Veneridae family) using both contour- and landmark-based methods. Since Bookstein (1991), the geometric morphometric method has been used as a tool to quantify the morphological modifications of an organism throughout its evolution processes. Unlike linear morphometry, the geometric morphometric approach separates shape and size and represents deformations through graphical structures (Bookstein 1991; Rohlf and Marcus 1993; Reis 1998; Rohlf 2000a, b, c; Jensen 2003; Moraes 2003; Adams et al. 2004; Sheets et al. 2006; Rocha 2011). Although this methodology has not been explored as much as other approaches in the study of mollusks, some studies have demonstrated that it may be applied for different purposes (Anderson 2001; Anderson and Roopnarine 2005; Inoue et al. 2014; Kosnik et al. 2006; Marquez et al. 2017; Monnet et al. 2009; Vuolo et al. 2011).

Therefore, the goal of is the present study was to identify morphometric characters using Psammotreta, Austromacoma, and Macoploma species shells, in order to improve taxonomic identification and increase knowledge of the life habits of Tellinidae.

Materials and methods

Samples from the Zoology Museum of the Sao Paulo University (MZUSP) and the Adão José Cardoso Zoology Museum of the State University of Campinas (ZUEC) were reviewed and underwent species recognition, following which five species, which were well represented in a number of individuals, were selected. Approximately thirty adults of each of the following species were selected—Austromacoma biota, 26; A. constricta, 28; Psammotreta brevifrons, 11; P. cleryana, 26; and Macoploma tenta, 28 (Fig. 1).

Fig. 1
figure 1

Macominae species studied. a–b Austromacoma biota (Arruda & Domaneschi, 2005); a internal surface ZUEC-BIV 2192, Cidade Beach, Caraguatatuba Municipality SP; b external surface, holotype MZUSP 41183, Cidade Beach, Caraguatatuba Municipality, SP; c–d Psammotreta brevifrons (Say, 1834 in Say 1830–1834) MZUSP 21974, Ubatuba Municipality, SP; c internal surface; d external surface; e–f Austromacoma constricta (Bruguière, 1792), UFScar Collection, Araçá Bay, São Sebastião Municipality, SP; e internal surface; f external surface; g–f Psammotreta cleryana (d’Orbigny, 1846 in d’Orbigny 1834–1847) MZUSP 43893, Bertioga Municipality, SP; g internal surface; h external surface; i–j Macoploma tenta (Say, 1834 in Say 1830–1834) MZUSP 21949, Andorinhas Beach, Ubatuba Municipality, SP; i internal surface; j external surface

Full size image

Fourteen (14) landmarks were selected using the internal region of the right valve for each individual (Fig. 2). The internal region was chosen due to having the most similar points among the species, including the intersection of the muscle scars and the extreme points of the shell curvature. Using the Bookstein (1991) classification of landmarks as type one, two, and three, those selected were classed as follows: landmarks 1, 3, 4, and 5 were type one; landmarks 2, 6, 7, 9, and 10 were type two; and landmarks 8, 11, 12, 13, and 14 were type three (Fig. 2). The images were taken by a 12.1 megapixel Canon Power Shot SX50 HS58 digital camera fixed to a tripod and from the same distance for all the valves.

Fig. 2
figure 2

Landmark or homologous points configuration. Landmarks: 1, extreme dorsal point; 2, beak of umbo; 3, extreme anterior point; 4, extreme ventral point; 5, extreme posterior point; 6, extreme ventral point of posterior muscle scar; 7, extreme dorsal point of posterior muscle scar; 8, intersection between pallial sinus and posterior muscle scar; 9, extreme dorsal point of pallial sinus; 10, extreme anterior point of pallial sinus; 11, intersection between pallial sinus and pallial line; 12, intersection between pallial line and anterior muscle scar; 13, anterior cruciform muscle scar; and 14, posterior cruciform muscle scar

Full size image

The x- and y-coordinates of each landmark were obtained using the TpsDig software package, version 2.30 (Rohlf et al. 2004), and general Procrustes analysis (GPA) was used to quantify shape and size variation. This analysis applies a least-square superposition algorithm to a set of Cartesian coordinates of anatomical landmark configurations, which broadly represent the shape of the specimen (Rohlf 1990). This procedure fits each configuration to the mean shape of the sample as closely as possible, so that size, position, and orientation remain constant according to the criterion for the least-squares fit. Consequently, the remaining variation in landmark positions is due to the variation of shape (Klingenberg 2013).

The GPA, in addition to most statistical analyses, were carried out using the MorphoJ software package, version 1.06d, which provides a user-friendly platform for a broad range of morphometric analyses (Klingenberg 2011). MorphoJ uses a full Procrustes fit and projects the data to the tangent linear Euclidean space by orthogonal projection, which produces a new set of shape variables, a variation/covariation matrix of shape coordinates (Klingenberg 2011), which were used in multivariate statistical analyses. The point of tangency between the two spaces, the Procrustes shape space and the Euclidean space, is the mean or consensus shape of the sample (Viscosi and Cardini 2011).

The relative displacement from one average configuration to another, or from one shape to another in the shape space, can be used to analyze the differences between shapes (Klingenberg 2013). Thus, shape changes were described using two wireframes based on landmark displacement: one for the starting shape, usually a consensus, and another for the target shape to be analyzed.

Procrustes Anova was used to assess the relative amount of variation among species. However, as MorphoJ does not allow permutation procedures, it was also performed in R software, using the geomorph package (Adams et al. 2021). For this, the matrix with the Cartesian coordinates of the landmarks was exported from MorphoJ, and GPA was performed again in R software, providing the variation matrix of GPA shape coordinates used in Procrustes Anova. Principal components analysis (PCA) was carried out to identify and account for the redundancy of the variation/covariation matrix of the GPA shape coordinates. In addition, PCA displays the major shape variation features. Finally, to test the separation of shell shapes between the Macominae species, linear discriminant analysis (LDA) using the leave-one-out cross-validation procedure was performed to estimate the reliability of the discrimination. In such validation, an individual is left out of the analysis, one-by-one, and prediction is performed using data from all the other specimens. In this way, the cross-validation predictions avoid “circular reasoning,” where a specimen is classified using functions calculated with samples that included the specimen itself. MorphoJ automatically includes a parametric T-square test for the difference between group means in LDA and provides both Procrustes distance and the Mahalanobis distance, comparing the shape of species pair by pair. The LDA performed in MorphoJ considers the dataset to be parametric. As the data only seldom comply with all the assumptions of the parametric tests, the species were compared pair by pair using Goodall’s F test with bootstrap in R software, using the shapes’ package (Dryden 2019), and the results obtained from the LDA and Goodall’s F test were compared. Two other permutation tests with the same objective of comparing the shape of the species pair by pair were also performed, the Hotelling test with permutations in the R software (Dryden 2019) and two groups with permutations in the Past software (Hammer et al. 2001). The results of these tests are available in the supplementary material.

To study the relationship between the variation of geometric shape and size of individuals from the Macominae species, a multivariate regression of the Procrustes coordinates onto a measure of size was used. For each independent variable, MorphoJ computes a shape score by projecting the shape data onto a line in the direction of the regression vector. According to Drake and Klingenberg (2008), this score is the shape variable associated with the shape changes predicted by the regression model and includes the residual variation in the direction of the shape space. The centroid size was used as a dependent variable. Centroid size is the most used size measure in the geometric morphometric method; it is the square root of the sum of squared distances of all the landmarks of an object from their centroid (Klingenberg 2016). Therefore, regression plot scores versus centroid size can give a visual impression how closely the data point fits a straight line (Klingenberg 2016).

The vector graphics and wireframes were generated by MorphoJ, and the artwork was created using Photoshop version CS6. Further details of these methods are described extensively in Bookstein (1991), Monteiro and Reis (1999), Zelditch et al. (2004), Viscosi and Cardini (2011), and Klingenberg (2011, 2013, 2016).

Results

As the number of individuals of P. brevifrons is much smaller than the other species studied, we initially performed a morphometric analysis without this species and then included it to compare the results. Procrustes Anova analysis indicated significant differences among species in both analyses with and without P. brevifrons (Table 1). The analysis carried out with permutations indicated the same results (Table 1).

Table 1 Procrustes Anova performed for differences in shape with and without P. brevifrons and with and without permutations
Full size table

The graphical representations of principal components analysis (PCA) with and without P. brevifrons show differentiations, based, mostly, on the first two axes (Figs. 3 and 4). In PCA analysis without P. brevifrons, axis 1 explains 58.7% of the variation and axis 2 16.5% of the variation (Fig. 3), while in PCA with P. brevifrons, axis 1 is responsible for 56.2% of the variation, axis 2 for 18.0%, and axis 3 for 7.68% (Figs. 4 and 5). As there were no differences between the initial analyses with and without P. brevifrons, only results obtained with P. brevifrons will be provided below.

Fig. 3
figure 3

Graphical representation of the first and second principal component axes without Psammotreta brevifrons in analysis. The wireframes of four specimens (in gray) in extreme positions of the axes compared to the consensus shape (in black) are illustrated to show variation along the first and second principal components axes. The arrows indicate the specimens represented. AB, Austromacoma biota; AC, Austromacoma constricta; PC, Psammotreta cleryana; MT, Macoploma tenta

Full size image
Fig. 4
figure 4

Graphical representation of the first and second principal components axes including Psammotreta brevifrons in analysis. The wireframes of four specimens (in gray) in extreme positions of the axes compared to the consensus shape (in black) are illustrated to show variation along the first and second principal components axes. The arrows indicate the specimens represented. AB, Austromacoma biota; PB, Psammotreta brevifrons; AC, Austromacoma constricta; PC, Psammotreta cleryana; MT, Macoploma tenta

Full size image
Fig. 5
figure 5

Graphical representation of the first and third principal components axes including Psammotreta brevifrons in the analysis. The wireframes of four specimens (in gray) in extreme positions of the axes compared to the consensus shape (in black) are illustrated to show variation along the first and second principal components axes. The arrows indicate the specimens represented. AB, Austromacoma biota; PB, Psammotreta brevifrons; AC, Austromacoma constricta; PC, Psammotreta cleryana; MT, Macoploma tenta

Full size image

When comparing the consensus with the individual settings in the PCA extreme position, it was observed that axis 1 (Figs. 4 and 5) demonstrates variation on the pallial sinus, specifically in the position of landmark 11, separating Austromacoma biota from the other species. In this species, landmark 11 is in a much more anterior position, compared to the consensus shape. Consequently, the pallial sinus is only slightly detached from the pallial line in this species, while in A. constricta and M. tenta, the pallial sinus is notably detached from the pallial line. Austromacoma biota also differs from all the other species through variation in landmarks 13 and 14, which is positioned more posteriorly than the consensus. Axis 2 separates the species into two different groups based on the external shape of their shells. The first group, situated at the most positive end of axis 2, contains three species with elongated shells—P. brevifrons, P. cleryana, and M. tenta, while the second group, situated at the most negative end, contains two species with an oval shell shape—A. biota and A. constricta. The variation in the external shape of the shell occurs mainly due to variations in landmarks 1, 3, 4, and 5. On the positive side of axis 2, landmark 1 tends to be in a more ventral position, landmark 3 in a more anterior position, and landmark 4 in a more dorsal and anterior position, causing the species to have a lower and more elongated shape, compared to the consensus shape. On the negative side of axis 2, landmark 1 tends to be more dorsal, landmarks 3 to 4 are almost in the same position as the consensus shape, and landmark 5 is more posterior; consequently, the shells of A. biota and A. constricta are taller and have an oval shape. Axis 3 (Fig. 5) also exhibited variation on the pallial sinus and the anterior adductor muscle, due to changes in landmarks 10 and 12, respectively. On the positive side of axis 3, landmark 10 tends to be more posterior, so that the pallial sinus is shallower, while landmark 12 tends to be more ventral and posterior, meaning that the posterior adductor muscle has a more ventral-posterior position. On the negative side of axis 3, meanwhile, landmark 10 tends to be more anterior, and consequently the pallial sinus is deeper, while landmark 12 tends to be more dorsal and anterior, so that the anterior adductor muscle has a more dorsal-anterior position. Thus, A. biota and M. tenta tend to have a deeper pallial sinus and anterior adductor muscle more dorsal than the other species.

The pair-by-pair analysis of the shell shape agreed with the PCA, separating the species into two groups according to the external shape of their shells. Mahalanobis distance compares the species by showing the distance among them, so, in Table 2, the greatest distance was between the oval and elongated shells, supported by the distances between A. biota and P. brevifrons (32.31), A. constricta and P. cleryana (22.19), A. constricta and P. brevifrons (21.88), and A. biota and P. cleryana (21.49), with A. biota and A. constricta having the same external shell shape, as well as P. cleryana and P. brevifrons. The shortest distance was between species with the same shell shape, P. brevifrons and P. cleryana (5.10), P. brevifrons and M. tenta (7.97), P. cleryana and M. tenta (8.13), and A. constricta and A. biota (11.67). Procrustes distance and the Goodall’s F test also provided similar results but also showed a great difference between A. biota and M. tenta and A. biota and A. constricta, which, despite having the same external shell shape, are quite different in pallial sinus shape and in the position of the anterior adductor muscle (Table 2). The results obtained with the Hotelling and two groups test with permutations were similar to the Goodall’s F statistic (supplementary material).

Table 2 Differences in shape between the five Macominae species, pair-by-pair: Procrustes distance, Mahalanobis distance, the T-square statistic and associated parametric P-value, permutation test and associated bootstrap and parametric P-values AB Austromacoma biota, PB Psammotreta brevifrons, AC Austromacoma constricta, PC Psammotreta cleryana, MT Macoploma tenta
Full size table

Table 3 presents the cross-validation matrix classification of linear discriminant analysis, with comparison between group 1 and 2 above the diagonal line and comparison of group 2 and group 1 below it. Most species were identified 100% correctly, except for P. brevifrons, P. cleryana, and M. tenta, with the lowest percentage between P. cleryana and P. brevifrons (72.73%), P. brevifrons and P. cleryana (73.07%), and M. tenta and P. brevifrons (78.60%).

Table 3 Cross-validation matrix classification of discriminant function among the five Macominae species: AB Austromacoma biota, PB Psammotreta brevifrons, AC Austromacoma constricta, PC Psammotreta cleryana, MT Macoploma tenta
Full size table

The LDA cross-validation scores (Fig. 6) of P. brevifrons and P. cleryana revealed many overlapped individuals. Compared to the consensus wireframes of these two species, the main difference is between landmarks 3 and 12, with landmark 3 in P. brevifrons (black line) dorsally dislocated and landmark 12 ventrally dislocated. The scores of M. tenta with P. brevifrons and M. tenta with P. cleryana are shown in Figs. 7 and 8, respectively. Besides landmarks 10, 11, and 12 previously mentioned, M. tenta (gray line) also differs from P. brevifrons (black line, Fig. 7) and P. cleryana (black line, Fig. 8) in landmark 1, positioned more ventrally, landmark 3, more anteriorly, landmark 5, more posteriorly, landmark 6 more anteriorly and dorsally, landmark 7 more anteriorly and ventrally, and landmarks 8 and 9, positioned more anteriorly, so that M. tenta has a more elongated shape and lower posterior adductor muscle than P. brevifrons and P. cleryana, in addition to a deeper pallial sinus.

Fig. 6
figure 6

Cross-validation analysis between Psammotreta brevifrons (PB) (black line) and Psammotreta cleryana (PC) (gray line). a Lollipop graph showing the variation of each landmark comparing the analyzed species and b consensus wireframe graphs comparing the analyzed species

Full size image
Fig. 7
figure 7

Cross-validation analysis between Psammotreta brevifrons (PB) (black line) and Macoploma tenta (MT) (gray line). a Lollipop graph showing the variation of each landmark comparing the analyzed species and b consensus wireframe graphs comparing the analyzed species

Full size image
Fig. 8
figure 8

Cross-validation analysis between Psammotreta cleryana (PC) (black line) and Macoploma tenta (MT) (gray line). a Lollipop graph showing the variation of each landmark comparing the analyzed species and b consensus wireframe graphs comparing the analyzed species

Full size image

Regression analysis did not reveal a clear relationship between shape and size but did show that the elongated shells tend to have smaller centroid sizes than oval shells (Fig. 9). Procrustes Anova analysis confirmed that there was a significant difference between the centroid size of species (F= 349.32 and p < 0.0001). However, this result should be considered carefully, as the individuals used in the present study have a restricted size range (Table 4). Table 4 summarizes the information on the shape and size of each species, associating it with information on their life habits.

Fig. 9
figure 9

Relationship between the regression score, representing the shape variation of the five studied species and the centroid size. AB, Austromacoma biota; PB, Psammotreta brevifrons; AC, Austromacoma constricta; PC, Psammotreta cleryana; MT, Macoploma tenta

Full size image
Table 4 Shape and size characteristics associated with the life habits of the studied Macominae species
Full size table

Discussion

Macominae species have shells with a similar external shape, making their identification by taxonomists difficult (Arruda and Domaneschi 2005; Piffer et al. 2011). The results obtained in the present study identify three main characters that allow differentiation of the analyzed species, supported by geometric morphometric data: the external shape of the shell, the shape of the pallial sinus, and the position of the adductor muscle scars.

The landmarks that identify the differences in the external shape of the shell are 1, 3, 4, and 5, points which allow the separation of the studied species into two groups, oval-trigonal (A. biota and A. constricta) and elongated shells (P. cleryana, P. brevifrons, and M. tenta). These landmarks were chosen as the most extreme points in the shell where variations mark a change in the external shape and consequently life habits. The Tellinoidea superfamily is known for its infaunal burrow specialists, for which elongated shells are a benefit due to their elongated and streamlined outline, which is generally unsculptured, as well as their extensive siphons and pronounced pallial sinus and central umbo (Watters 1993). Some species lie on the left valve with the long axes horizontal, while others lie with their long axes at angles of up to 30° or 45° (Stanley 1970). Horizontal, or close to horizontal, positioning can offer an advantage to siphonate deposit feeders, facilitating horizontal migrations for better substrate exploitation (Stanley 1970). In the case of the studied species, A. biota and A. constricta lie horizontally, with prominent posterior curvature to the right, showing that siphons emerge perpendicular to the orientation of the shell (Arruda et al. 2003; Piffer et al. 2011). Deposit feeders, oval shell species, and species which acquire a horizontal life position are common in Tellinidae. Eurytellina lineata (Turton, 1819), for example, has a striking similarity to A. constricta and A. biota and occupies the same type of dissipative or sheltered intertidal environment (Arruda et al. 2003). According to Stanley (1970), the life position of M. tenta is similar to that of A. biota and A. constricta, lying on the left valve, but is slightly inclined, with the posterior region upwards. This position may explain the posterior flexure to the right seen in M. tenta, which is less prominent than in A. biota and A. constricta. There are no studies that demonstrate the position in sediment of P. brevifrons and P. cleryana under wild conditions, but their more elongated and streamlined shells, with a weak posterior flexure to the right, suggest they remain inclined, like M. tenta (Table 4).

Austromacoma biota differs from A. constricta in landmark 11, the intersection point of the pallial sinus with the pallial line, which is more anterior than in A. constricta (Fig. 4), a characteristic also supported by Arruda and Domaneschi (2005) and Piffer et al. (2011). This shows that the pallial sinus is only slightly detached from the pallial line in A. biota and very detached in A. constricta. It also demonstrates that, anatomically, the retracting muscles of the siphons in these two species are different. Hypothesizing that the siphon retracting musculature in A. biota is more robust and closer to the mantle margin than in A. constricta, its siphons can most likely be retracted more quickly than those of A. constricta. This rapid retraction is likely to result in more particles entering the mantle cavity, obstructing the gills. One finding that confirms that A. biota handles larger amounts of material in the mantle cavity than A. constricta is the presence of paired siphonal organs in A. biota, while A. constricta has only a single unilateral left siphonal organ (Piffer et al. 2011). The particles caught by the siphonal organs in A. biota are conveyed to the posterior extremity of the waste channel and eliminated as pseudofeces (Piffer et al. 2011).

The cruciform muscle, localized at the base of the siphonal internal apertures, connecting the ventral mantle lobes, characterizes species of Tellinoidea. The function of the cruciform muscle is still poorly understood, but it appears to have the function of anchoring the siphons during retraction and protraction (Yonge 1949) and also of working as a ventral pivot during the rocking movements involved in burrowing (Watters 1993). The position of this muscle was represented in this study by landmarks 13 and 14. Generally, the species studied showed little variation in these landmarks, except for A. biota, in which the landmarks 13 and 14 are in a more anterior position than the consensus wireframe. In addition to the cruciform muscle, A. biota has an accessory adductor muscle, which to date has not been found in any other species of bivalve, located between the anterior and the posterior insertions of the cruciform muscle (Arruda and Domaneschi 2005). It is likely that this accessory adductor muscle interferes in the position of the cruciform muscle in A. biota. This muscle has the same function as the adductors (Piffer et al. 2011), but it is likely that it also reinforces the functions of the cruciform muscle, serving as a more robust anchorage during the rapid movement of the siphons.

In addition to the differences between A. biota and A. constricta described above, morphometric analysis highlighted the variation in the position of the anterior adductor muscle scar, represented by landmark 12, which in A. biota is more dorsal than in A. constricta. This means that the anterior adductor muscle of A. constricta has a larger dorsoventral axis than the same muscle in A. biota.

Comparing the elongated group, P. brevifrons, P. cleryana, and M. tenta, the latter is that most differs (Fig. 6, 7, and 8), as its anterior adductor muscle is positioned more anteriorly, illustrated by landmark 12, and its specimens are longer and shorter than the others. Landmark 11 varies among the individuals of M. tenta, which explains the reason for the misidentification of some individuals during the analysis (see Figs. 4 and 5). This more elongated form of M. tenta suggests that its burrowing behavior differs from other species. The physiology of bivalve burrowing is well known, with a rocking movement occurring when the foot is anchored in the substratum and the contraction of the pedal retractor muscles (first the anterior and then the posterior) pulls the bivalve into the sediment (Stanley 1970). Stanley (1970) demonstrated that there is a relationship between the angle of rotation during the rocking movement and the degree of elongation, measured by the relationship between length and height (L/H), and that highly elongated species tend to use little or no rocking movement during burrowing, penetrating the substratum in an almost parallel direction to the long axis of the shell, while moderately elongated forms tend to use larger angles of rotation. Macoploma tenta exhibits little or no rocking movement during burrowing (Stanley 1970), while based on the external shape of the shells of P. brevifrons and P. cleryana, which are less elongated than M. tenta, we can assume that these species exhibit larger rocking movements, as do A. biota and A. constricta (Table 4). It is likely, however, that these species have even smaller rotation angles than highly rounded species, such as those of the Veneridae family.

In addition to the external shape of the shell and the position of the anterior adductor muscle scar, M. tenta differs from the other elongated species by the shape of its posterior adductor muscle scar, represented by landmarks 6 and 7, and that of the pallial sinus, represented by landmarks 8, 9, and 10. Landmarks 6 and 7 delimit the ventral and dorsal extremity of the posterior adductor muscle scar, respectively. The more dorsal position of landmark 6 and the more ventral position of landmark 7 demonstrate that the posterior adductor muscle of M. tenta has a smaller dorsoventral axis, which is more rounded than that of the other species. The more anterior position of landmarks 8, 9, and 10 demonstrates that M. tenta has a deeper pallial sinus.

Based on morphological similarity, Psammotreta brevifrons resembles P. cleryana (Tenório et al. 1986), supported by the results obtained in this study, and the occurrence of both species overlaps on the Brazilian coast (Rios 1970, 1974, 1985, 1994, 2009) and Western Atlantic coast (Huber 2015). Based on this analysis, P. cleryana should be considered a synonym of P. brevifrons. Huber (2015) has stated that the holotype of P. cleryana does not differ from P. brevifrons but also considered the latter to be variable along the Western Atlantic coast. Further detailed populational studies are therefore necessary to provide a better characterization of this species.

Despite the significant environmental effects suffered by shell morphology, and its disjointed characters, which often make taxonomic description abstract, detailed morphological analysis can be extremely useful in distinguishing bivalves. The results of the present study demonstrate that the geometric morphometric approach is especially useful when validating these characteristics of continuous variation and can provide important information for comparing the life habits of species. The external shape of the shell, the pallial sinus, and the adductor muscles proved to be efficient features for differentiating Macominae species.