The specimens were collected throughout the entire geographic distribution range of the South American razor clam, Ensis, from San Matías Gulf (41°34’S) in the Atlantic Ocean to Trujillo in the Pacific Ocean (08°06’S; Table 1). This study is compliant with the CBD and Nagoya protocols. All material was examined and compared with the specimens assigned to the new species. An updated distribution range for each valid species is illustrated (Fig. 1). The examined specimens were deposited in the Invertebrate Collection of IBIOMAR (CNP-Inv).

A total of 370 specimens, 274 from the southern clade and 96 from the northern clade were photographed with the inner side upwards. Following Márquez et al. (2017), the inner shell shape variation was studied using landmark-based GM and multivariate statistics. Twelve landmarks (2D configuration) were captured to examine shape variation of the internal muscle scars and pallial lines (Fig. 2). Procrustes analysis (Rohlf and Slice 1990) was used to remove rotation, translation and scale effects of the raw landmark coordinates. Centroid size (CS) was calculated as the square root of the sum of the square distances from the landmarks to the CS which they defined (Zelditch et al. 2004) and was used as a proxy for shell size. Multivariate regression (pooled within-site) between shape variables and CS values was calculated to evaluate and correct the positive allometry. To capture the shape components that maximized the separation between species, a discriminant function was calculated. Finally, the differences in mean shell shape between species were analysed by a T2 Hotelling test with 1000 permutations, and a re-sampling procedure (leave-one-out-cross-validation) was performed to estimate the percentage of misclassification to each species (Johnson and Wichern 2007). All GM analyses were conducted in MorphoJ, version 1.07a (Klingenberg 2011).

The independence between the Procrustes coordinates and the CS values was rejected (P<0.0001). Allometry explained 2.62% of the variation, so the residuals were used as the new allometric-free shell shape variables for subsequent analyses. Discriminant analysis showed that the shell shape range was different between the southern and northern clades. The mean inner shell shapes of the clades were significantly different (T2=761.07, p<0.0001, Mahalanobis distances= 3.27) (Fig. 3). The mean shell shape of the southern clade shows an anterior retraction and an expansion of the posterior part with a reduction of the anterior adductor muscle scar and the anterior and posterior retractor muscle scar of the foot in comparison with the mean shell shape of the northern clade (Fig. 3). The cross-validation function, based on the Mahalanobis distance of each individual from group means, indicated a low number of allocation errors in the species assignation (3.1% of northern individuals were placed in the southern group and 5.5% of the southern individuals were placed within the northern group; Table 2).

After the molecular data published by Márquez et al. (2020) and the GM results presented in this study, two species of Ensis distributed along the Pacific coast of South America are now recognized. They are Ensis loboi n. sp., distributed from Tubul, Chile (37°S) to Trujillo, Peru (8°S); and the Magellan E. macha, with a distribution range from San Matías Gulf (41°S) in the Atlantic Ocean to Niebla (40°S) in the Pacific Ocean. All records of E. macha mentioned northward of Tubul, Chile, must be considered synonyms of E. loboi (Soot-Ryen 1959, Peña 1971, Basly-Santa Maria 1983, Guzmán et al. 1998, Ramírez et al. 2003, Uribe et al. 2013, Paredes et al. 2016).

Family Pharidae H. Adams and A. Adams, 1856

Genus Ensis Schumacher, 1817

[= Ensatella Swainson, 1840: 365. Type species Ensatella europaea Swainson, 1840 (=Ensis ensis (Linnaeus, 1758), OD); Hypogaea Poli, 1791 (in Poli, 1791-1795): 29 (in part), name given for soft parts of species included in Solenoidea, Tellinidae and Pholadidae ].

Type species: Ensis magnus Schumacher, 1817 by monotypy.

Distribution. Widely distributed along the Pacific coast of South America (Osorio and Reid 2002), southwestern USA to western Central America (Coan and Valentich-Scott 2012, Valentich-Scott et al. 2020), western Atlantic coast of South America (Rios 2009, Scarabino et al. 2016, Giacomino and Signorelli 2021), east coast of North America (Mikkelsen and Bieler 2007), Europe (Costello et al. 2001, Cosel 2009) and tropical west Africa (Cosel 2009).

Remarks. The genus Ensis includes 14 extant species (Cosel 2009, Coan and Valentich-Scott 2012, Valentich-Scott et al. 2020) widely distributed along both coasts of America, Europe and Western Africa. Ensis loboi n. sp. constitutes the fifteenth species of the genus.

Ensis loboi n. sp.

(Fig. 4A-R)

Ensis californicus of authors (not Dall, 1899) Alamo-Vázquez and Valdivieso-Milla, 1997: 141-142, 144, fig. 378.

Ensis macha of authors (not Molina, 1782) Stempell, 1899: 239-240 (in part); Soot-Ryen, 1959: 67; Peña, 1971: 137; Basly-Santa Maria, 1983: 26, pl. 10, fig. 83; Guzmán et al. 1998: 70-71; Ramírez et al. 2003: 268; Uribe et al. 2013: 230; Paredes et al. 2016: 149.

Type material: CNP-Inv 875, holotype. Paratypes: CNP-Inv 844, 1 specimen, CNP-Inv 872, 1 left valve, CNP-Inv 874, 1 specimen, Tubul, Talcahuano, Concepción, Chile (36.726202°S 73.132905°W).

Type locality. Tubul, Talcahuano, Concepción, Chile (36.726202°S 73.132905°W).

Etymology. This species is dedicated to José María “Lobo” Orensanz in recognition of his significant contribution to the field of fisheries and his invaluable advice to the authors of this manuscript over several years.

Description. Shell (Fig. 4A-R): large, thick, length up to 230 mm, cylindrical in cross section, elongate; dorsal and ventral margins almost straight, posterior end truncate; anterior end rounded; yellowish to dark brown periostracum, thin to moderately thick, partially eroded along the dorsal slope in larger specimens; exterior surface smooth with three sculptured zones, usually observed within Ensis; external ligament long and brown, over the dorsal edge, anteriorly placed behind the beaks; interior surface white to purple; hinge plate with two cardinal teeth in the right valve, one anterior and vertically oriented and one posterior and horizontally oriented; four cardinal teeth in the left valve, two vertically placed and two horizontally oriented, parallel to dorsal margin (Fig. 4), the horizontally oriented teeth are four times longer than the vertically oriented ones; ligament dark, opisthodetic (Fig. 4).

Additional material examined. Peru: CNP-Inv 3903, three specimens collected at Trujillo by José M. Lobo Orensanz. Chile: Bahía Coliumo, Dichato, Concepción (36.720555°S 73.150833°W), CNP-Inv 292, 19 valves, Tubul, Talcahuano, Concepción (36.726202°S 73.132905°W), Cnp-Inv-281, 26 specimens.

Distribution. From Trujillo, Peru (8°8’S 79°3’W) to Tubul, Chile (37°13’S 73°26’W).

Habitat. E. macha and E. loboi inhabit superficial sandy bottoms, from the shallow subtidal to about 30 m depth, forming fishing grounds or banks (Osorio and Bahamonde 1970, Márquez and Van der Molen 2011). Razor clams can be found in a substrate characterized by a mixture of grain sizes, where fine and very fine sand with a low level of organic matter generally predominate (Aracena et al. 1998, Jaramillo 1998, Márquez and Van der Molen 2011).

Remarks. None of the synonyms of E. macha mentioned by Giacomino and Signorelli (2021) apply to the northern clade. Solen poirieri Mabille and Rochebrune, 1889 and Solen scalprum P. P. King, 1832 were described from the Magellan region. The registered type locality of Ensis luzonicus Dunker, 1862 (Luzon Island, Philippines) was erroneous. And the type of Solen gladiolus, described from “S. America”, has not been found in the NHMUK. The synonymy with E. macha is currently open awaiting further analysis (Giacomino and Signorelli 2021). For these reasons, a new name is proposed for the northern clade. The shell morphology of E. loboi n. sp. and E. macha (Fig. 5) is similar. Slight differences are illustrated in Figure 6 and listed in Table 3. Ensis loboi n. sp. was found to possess a stronger and thicker shell than E. macha. In addition, the GM revealed slight differences between the two species. The main shape differences were the distance between the posterior adductor muscle and the postero-ventral mantle projection (landmarks 5-6, Fig. 3); the distance between ventral and dorsal mantle projections (landmarks 1-5, Fig. 3); and the length of the anterior adductor muscle (landmarks 10-8, Fig. 3). Ensis californicus was reported from southern Peru (Alamo-Vázquez and Valdivieso-Milla 1997: 141-142, 144, fig. 378). However, the illustrated specimen clearly differs from the type material of Dall (USNM 158891), which has a well-defined concave dorsal margin. This was also noticed by Valentich-Scott et al. (2020). In addition, the distribution of E. californicus is well documented from Sonora to Jalisco, Mexico (Coan and Valentich-Scott 2012). After the analysis of new material from Trujillo, Peru, the specimens illustrated by Alamo-Vázquez and Valdivieso-Milla (1997) must be considered E. loboi n. sp.