Introduction

Species of the genus Mola (family Molidae; Tetraodontiformes) are highly derived, large-sized bony fishes characterized by an orbicular and strongly compressed body, attaining more than 3 m total length (TL) and 2,000 kg, with members widely distributed in the open ocean of tropical and temperate seas (Pope et al. 2010; Yamanoue and Sawai 2012). They completely lack a caudal fin, which is replaced by a broad rudder-like lobe called the clavus (Fraser-Brunner 1951; Nakae and Sasaki 2006). Ocean sunfishes have attracted interest for centuries because of their unique shape and large size, and they are described and illustrated in old literature dating from at least the 16th century (cf. Rondelet 1554; Salviani 1554). The published literature on species of Mola is extensive and exceeds 1,000 works; nonetheless, numerous aspects of morphology of Mola and biology remain mysterious (Fraser-Brunner 1951; Yamanoue and Sawai 2012; Matsuura 2015). Taxonomic confusion remains within the Molidae (56 nominal species, including unavailable names: Eschmeyer et al. 2017), with 34 nominal species in Mola alone. Unfortunately, most type specimens of the nominal species are lost (Parenti 2003; Eschmeyer et al. 2017).

The first comprehensive review of the Molidae, by Fraser-Brunner (1951), recognized five species in three genera: Masturus lanceolatus (Liénard 1840), Masturus oxyuropterus (Bleeker 1873), Mola mola (Linnaeus 1758), Mola ramsayi (Giglioli 1883), and Ranzania laevis (Pennant 1776). Fraser-Brunner (1951) proposed that the genus Mola is characterized by having ossicles along the rear margin of the clavus and that Mola mola and Mola ramsayi are distinguishable from each other using the following features: 1) number of clavus fin rays (16 in Mola ramsayi vs. 12 in Mola mola); 2) number of clavus ossicles (12 in Mola ramsayi vs. 8 or 9 in Mola mola); 3) size of the clavus ossicles (broader than the space between them in Mola ramsayi vs. narrower than the space between them in Mola mola); 4) shape of the clavus margin (not wavy in Mola ramsayi vs. wavy or lobed with age in Mola mola); 5) clavus ossicles on the paraxial rays [“paraxial ossicles” (PO) in Fraser-Brunner 1951] (these ossicles are separate and smaller than the other ossicles in Mola ramsayi vs. these ossicles are united to form a single ossicle and are larger than the other ossicles in Mola mola); and 6) presence or absence of a band of reduced denticles along the base of the clavus from the dorsal to anal fins [“naked area of reduced denticles” (NA) in Fraser-Brunner 1951; called a ‘smooth band’ in this study] (absent in Mola ramsayi vs. present in Mola mola). In addition, Fraser-Brunner (1951) postulated that Mola mola is widely distributed in the world’s oceans, while Mola ramsayi is distributed only in the South Pacific (New Zealand, Australia and Chile).

Presently, the morphology-based classification of Fraser-Brunner (1951) is generally accepted, yet recent studies of molecular phylogeny have suggested different opinions on the classification of species. Bass et al. (2005) found four clades in Mola (namely: Mola mola Atlantic group, Mola mola Pacific group, Mola ramsayi Atlantic group, and Mola ramsayi Pacific group) based on partial D-loop sequences of mitochondrial DNA (mtDNA), although the morphological characters were not examined. Yoshita et al. (2009) analyzed both partial and complete D-loop sequences and proposed that Mola contains at least three species, which the authors temporarily assigned to Mola sp. A, Mola sp. B, and Mola sp. C. To date, although subclades (or subgroups) were found in Mola sp. A and Mola sp. B, the three clades have been regarded as representing separate species (Yamanoue and Sawai 2012; Ahuir-Baraja et al. 2017; Sawai et al. 2017a; Nyegaard et al. 2017).

Yoshita et al. (2009) proposed that adult Mola sp. A and Mola sp. B are distinguishable using the following five characters (Fig. 1): 1) number of clavus fin rays (14–17 in Mola sp. A vs. 10–13 in Mola sp. B); 2) number of clavus ossicles (8–15 in Mola sp. A vs. 8 or 9 in Mola sp. B); 3) shape of the clavus margin [not wavy in Mola sp. A vs. wavy or lobed in Mola sp. B (although not wavy in small specimens of Mola sp. B)]; 4) presence or absence of a head bump (present in Mola sp. A vs. absent in Mola sp. B); and 5) proportion of the body length to body depth (the anterior part of Mola sp. A is larger than Mola sp. B). Further studies have provided two additional diagnostic characters: shape of the body scales (called denticles in Fraser-Brunner 1951) [rectangular (line shaped when viewed from above) in Mola sp. A vs. conical with branching tips (dot shaped when viewed from above) in Mola sp. B] (Sawai et al. 2015b; Sawai 2016b; Sawai and Yamanoue 2016b) and presence or absence of a chin bump (present in Mola sp. A vs. absent in Mola sp. B) (Sawai et al. 2017a) (Fig. 1).

Fig. 1
figure 1

Generalized body forms (side and front view) and body scale morphology of adult Mola alexandrini, Mola mola, and Mola tecta. a M. alexandrini [= Mola sp. A sensu Yoshita et al. (2009)], b M. mola [= Mola sp. B sensu Yoshita et al. (2009)], c M. tecta [= Mola sp. C sensu Yoshita et al. (2009)]. Large arrows indicate points of morphological differences (shape of head, clavus, and chin). Small arrows indicate ossicles (on snout, chin, and clavus). Broken arrows indicate the smooth band. Broken squares indicate scales on the middle region of the body (left, transverse view; right, upper view; photographs of dried skin). For further details, see ‘Materials and methods’

Full size image

Nyegaard et al. (2017) recently described Mola sp. C sensu Yoshita et al. (2009) as Mola tecta, based on the holotype and 11 paratypes collected from New Zealand and Australia, and included a review of original descriptions of the genus. Mola tecta (>65 cm TL) is distinguishable by the following combination of 11 characters: 1) 15–17 clavus fin rays; 2) 5–7 clavus ossicles; 3) anterior profile laterally tapered (no swollen dorso-lateral and ventro-lateral ridges on body); 4) head profile rounded (without protruding snout); 5) no head bump; 6) no chin bump; 7) smooth band present at base of clavus from the dorsal to anal fins; 8) smooth band with a ‘back-fold’; 9) clavus rounded with an indent; 10) ossicles on paraxial fin rays separate; and 11) body scales conical and without branching tips (simple and dot shaped when viewed from above) (Nyegaard et al. 2017). Nyegaard et al. (2017) also suggested that Mola sp. A and Mola sp. B proposed by Yoshita et al. (2009) presumably apply to Mola ramsayi and Mola mola, respectively. Although they did not conduct a detailed review about the scientific names of Mola sp. A and Mola sp. B, they confirmed that the three genetically identified species in Mola proposed by Yoshita et al. (2009) are distinct species based on morphology.

The present study aimed to resolve the nomenclatural status of the remaining two genetically identified species (i.e. Mola sp. A and Mola sp. B), based on both molecular and morphological investigation, by examining the available type material and revaluating the diagnostic characters proposed by previous studies (i.e. Fraser-Brunner 1951; Yoshita et al. 2009; Sawai et al. 2015b; Nyegaard et al. 2017).

Materials and methods

Morphological data were collected from 30 specimens of Mola: 14 fresh or preserved specimens of Mola alexandrini, plus the holotypes of Orthragoriscus alexandrini (mounted skin) and Orthragoriscus ramsayi (mounted skin); and 13 fresh or preserved specimens of Mola mola, plus the holotype of Ozodura orsini (mounted skin). Some data were obtained from photographs, and data from previous studies were also included (i.e. Yoshita et al. 2009; Sawai et al. 2009, 2015b, 2017a; Yamanoue et al. 2010; Sawai 2016a; Ahuir-Baraja et al. 2017; Nyegaard et al. 2017).

Counts generally follow the methods of Yoshita et al. (2009). Fin rays were macroscopically counted when discernible externally, and those of some specimens were confirmed by dissection or from X-rays. Fin rays of the dorsal + clavus + anal complex (D + C + A fin rays) were counted following Gudger (1937). Ossicles were counted by feeling along the rear margin of the clavus (Fig. 1; see also Sawai et al. 2015b: fig. 1, right).

Measurements were according to Yoshita et al. (2009: fig. 2) and included total length (TL), head bump length (HBL), pre-clavus band length (PCBL), body depth (BD), and total body depth (TBD). Measurements were made with a tape measure, ruler, or calipers to the nearest 0.1 cm. For comparison of the morphometrics of M. alexandrini and M. mola, the specimens examined were divided into large (>180 cm TL, presumably adult stage) and small (<70 cm TL, presumably young stage) since the body proportions of molids significantly change with age (see Table 1). Unfortunately, moderately sized specimens (70–180 cm TL) were not obtained. Assessment of the developmental stages was based on descriptions of the external morphology provided by Iwai (2005). In Iwai (2005), the young or adolescent stage is defined as the transition period from juvenile to adult where the morphological characteristics of the fish body are developing and thus the morphological characteristics of the species first appear, yet the relative body proportions are different from the final adult form.

Table 1 Measurements for Mola alexandrini and M. mola
Full size table

Morphological observations here generally follow Fraser-Brunner (1951), Yoshita et al. (2009), Sawai (2016b), Sawai et al. (2017a), and Nyegaard et al. (2017) (see Table 2). Specifically, the following external assessments were made: establishing the shape of the clavus margin as either wavy (i.e. lobed or scalloped), rounded, or rounded with an indent (Fig. 1; see also Yoshita et al. 2009: fig. 5; Nyegaard et al. 2017: fig. 5); ascertaining the presence or absence of a smooth band at the base of the clavus between the dorsal and anal fins (Fig. 1; see also Fraser-Brunner 1951: figs. 13, 14, 16, 17), the presence or absence of a head bump (Fig. 1; see also Sawai et al. 2015b: fig. 2), the presence or absence of a chin bump (Fig. 1; see also Sawai et al. 2017a: fig. 4), and the presence or absence of a smooth band back-fold (Fig. 1; see also Nyegaard et al. 2017: fig. 5). Furthermore, the shape of the body scales on the middle region of the body was observed both macroscopically and microscopically to discern the shape of the scales as rectangular (line shaped when viewed from above), conical with branching tips (dot shaped when viewed from above), or conical without branching tips (simple and dot shaped when viewed from above) (Fig. 1; see also Sawai 2016b: figs. 2–4; Nyegaard et al. 2017: fig. 9). In this study, paraxial fin rays were defined as the pair of fin rays in the middle of the clavus, nearest to the end of the vertebral column (see also Fraser-Brunner 1951).

Table 2 Morphological characters for Mola alexandrini, M. mola, and M. tecta
Full size table

Institutional abbreviations follow Fricke and Eschmeyer (2017), with the addition of the following: Ibaraki Nature Museum, Ibaraki, Japan (INM); Ibaraki Prefectural Oarai Aquarium, Ibaraki, Japan (OA); Kitakyushu Museum of Natural History and Human History, Fukuoka, Japan (KMNH); Oman Marine Science and Fisheries Centre, Muscat, Sultanate of Oman (OMMSFC); and Sea and Life Museum, Tottori, Japan (SLM).

The type specimens in MZUB, holotype of M. alexandrini, MZUB (unnumbered), and neotype of M. mola, MZUB (unnumbered), have a special position in the oldest collections and could not have a unique number ID by museum rule (Bruno Sabelli, personal communication). However, their location is particular in special well-identified positions with appropiate label as type (ICZN 1999, Art. Recommendation 72D–72F).

Mola alexandrini (Ranzani 1839)

(Japanese name: Ushi-manbo; New English name: Bump-head sunfish) [Figs. 1a, 2, 3a, c (left), 4, 5a, 6a–d, 7; Tables 1–2; Electronic Supplementary Material (ESM) Table S1]

Fig. 2
figure 2

Holotype of Orthragoriscus ramsayi (junior synonym of Mola alexandrini) deposited in BMNH. BMNH 1883.11.29.22, 229.1 cm TL (mounted skin), Darling Harbour, New South Wales, Australia. Red square, 10 cm × 10 cm

Full size image
Fig. 3
figure 3

Type specimens of Mola alexandrini and Mola mola. a Holotype of M. alexandrini (holotype of Orthragoriscus alexandrini), MZUB (unnumbered), 190.0 cm TL (mounted skin), Adriatic Sea; b neotype of M. mola (holotype of Ozodura orsini), MZUB (unnumbered), 59.7 cm TL (mounted skin), Adriatic Sea; c drawing of holotypes of Orthragoriscus alexandrini (left) and Ozodura orsini (right) from Ranzani (1839). Red square 10 cm × 10 cm

Full size image
Fig. 4
figure 4

Specimens of Mola alexandrini. a AM-54, 332.0 cm TL (photographed by K. Sagara); b AIM MA30934, 211.0 cm TL (cast); c TaHMo-2, 162.5 cm TL (photographed by Z. T. Lin); d NMNZ P.034449, 51.5 cm TL; e BMNH 1870.12.27.41, 42.5 cm TL; f NMNZ P.056071, 29.3 cm TL. Red square 10 cm × 10 cm

Full size image
Fig. 5
figure 5

Distribution of Mola as confirmed by this study and from the literature (ESM Table S1). a Mola alexandrini; b Mola mola. Star type locality (black, holotype; gray, neotype), circles records of genetically identified specimens, triangles records of morphologically identified specimens

Full size image
Fig. 6
figure 6

Smooth band and body scales of Mola. a, b the holotype of Orthragoriscus ramsayi (junior synonym of Mola alexandrini), BMNH 1883.11.29.22, 229.1 cm TL, mounted skin), smooth band (a), body scales (b); c, d holotype of M. alexandrini [MZUB (unnumbered), 190.0 cm TL, mounted skin], smooth band (c), body scales (d); e, f neotype of Mola mola [holotype of Ozodura orsini, MZUB (unnumbered), 59.7 cm TL, mounted skin], smooth band (e), body scales (f). Asterisks indicate the smooth band at the base of the clavus between the dorsal and anal fins

Full size image
Fig. 7
figure 7

Relationship between total length and body proportions. Red Mola alexandrini, blue Mola mola, open symbols, non-type material, solid symbols, type specimens, circles measurements from specimens, triangles, measurements obtained from photographs

Full size image

Orthragoriscus alexandrini Ranzani 1839: 75, foldout table, pl. 6 (left) (type locality: Adriatic Sea); Alessandrini 1839: 359, pls. 31–34; Steenstrup and Lütken 1898: 61.

Orthragoriscus ramsayi Giglioli 1883: 315 (type locality: New South Wales, Australia); Ramsay 1883: 46.

Orthagoriscus mola: Williams 1893: 110, pl. 8A; Parker 1897: 627; Wilton and Brown 1908: 5, pl. 4, fig. 7.

Orthagoriscus alessandrini: Steenstrup and Lütken 1898: 20 (misspelling of O. alexandrini).

Ozodura alessandrinii: Steenstrup and Lütken 1898: 21 (misspelling of O. alexandrini).

Orthagoriscus alexandrinii: Steenstrup and Lütken 1898: 34, fig. 15 (misspelling of O. alexandrini).

Mola mola (not of Linnaeus): Waite 1913: 223 (in part), pl. 9; Waite 1921: 198 (in part), fig. 332; Waite 1923: 230, unnumbered fig. on p. 231; Phillipps 1926: 171 (in part), fig. 2; Gudger 1928: 257 (in part), unnumbered figs. on pp. 258–259, 261; Barnard 1935: 654 (in part), figs. 5a–b; Smith 1950: 422 (in part), pl. 95, fig. 1213; Smith 1965: 422 (in part), pl. 95, fig. 1213; Khan 1975: 295, fig. 1; Anderson et al. 1998: 29 (in part), pl. 4, fig. 44; Roach 2003: 1; Yamanoue et al. 2004: 269; Barreiros and Teves 2005: 4 (in part), unnumbered fig. on p. 4; Bass et al. 2005: 405 (in part); Sagara et al. 2005: 35 (in part); Konow et al. 2006: 208, fig. 1a; Nakatsubo et al. 2007a: 259 (in part); Nakatsubo et al. 2007b: 613 (in part); Soichi 2009: 200 (in part), fig. 2B; Pope et al. 2010: 472 (in part), fig. 1; Jawad 2013: 287 (in part), fig. 1; Phillips et al. 2015: 1118 (in part), fig. 1, nos. 2, 6, 13; Pan et al. 2016: 2 (in part), fig. 1a.

Mola ramsayi: Whitley 1931: 126 (in part), pl. 16, figs. 3–4; Fraser-Brunner 1951: 90 (in part), figs. 13–14; Tyler 1980: 377; Last et al. 1983: 521, fig. 33.24; Heemstra 1986: 908, fig. 270.3; Hutchins and Swainston 1986: 114, fig. 686, Appendix I; Al-Ghais 1994: 19, unnumbered figs. on pp. 19, 30; Glover 1994: 917, fig. 809; Hutchins 2001b: 3966, unnumbered fig. on p. 3968; Brito 2003: 77 (in part), fig. 2; Parenti 2003: 3 (in part); Heemstra and Heemstra 2004: 458; Bass et al. 2005: 405 (in part); Bray and Hoese 2006: 1937; Mohan et al. 2006: 23, unnumbered fig. on p. 24; Bray 2008: 859, unnumbered fig. on p. 860; Pope et al. 2010: 473 (in part); Jawad et al. 2012: 1, figs. 1–2; Saunders 2012: 133; Kishore et al. 2013: 9, fig. 1; Thys et al. 2013: 1, fig. 1A; Yasemi and Nazari Bejgan 2014: 242, fig. 1; Matsuura 2015: 96, fig. 6; Phillips et al. 2015: 1118, fig. 1, nos. 4a–b; Stewart and Struthers 2015: 1745 (in part), fig. 250.2; Thys et al. 2016: 1, fig. 2; Matsuura 2017: 87, figs. 13-3, 15-1, unnumbered figs. on p. 95; Thys et al. 2017: 1, fig. 3.

Masturus lanceolatus (not of Liénard): Gudger 1937: 375 (in part), pl. 4, fig. 13.

Mola alexandrini: Barnard 1947: 212, pl. 25, fig. 7; Barnard 1948: 402; Smith 1950: 422; Smith 1965: 422.

Pseudomola lassarati Cadenat 1959: 1115, figs. 9–11 (type locality: Vridi, Ivory Coast); Blache et al. 1970: 198, fig. 539; Tyler 1980: 377; Tortonese 1990: 1079.

Mola ramsayi Atlantic group: Bass et al. 2005: 405.

Mola mola group A: Yoshita et al. 2005: 171.

Mola sp. group A: Yoshita et al. 2009: 232, fig. 5 (left).

Mola sp. A: Sawai et al. 2009: 9, fig. 5; Yamanoue et al. 2010: 27; Sawai et al. 2011: 181; Yamanoue and Sawai, 2012: 167, figs. 9.1 (right)–9.2 (left); Hatooka and Hagiwara 2013: 1747, unnumbered fig. on p. 1747 (bottom); Sawai et al. 2014: 127, fig. 1; Sawai et al. 2015a: 201, fig. 1; Sawai et al. 2015b: 65, figs. 1A, 2A–C, 3A; Sawai 2016b: 349, fig. 4; Sawai and Yamanoue 2016a: 54, fig. 1; Sawai and Yamanoue 2016b: 451, figs. 2A, 3A; Ahuir-Baraja et al. 2017: 1133, fig. 1; Sawai et al. 2017a: 99, figs. 3–4; Sawai et al. 2017b: fig. 1; Nyegaard et al. 2017.

?Mola ramsayi: Whitley 1931: 127 (in part), text fig. 2, pl. 16, fig. 1; Atria 1967: 3; Villalba and Fernández 1985: 71; Gauldie 1990: 193, fig. 4B; Gauldie 1992: 263, figs. 4–6; Luque and Oliva 1993: 273; Hutchins 2001a: 48; Putra et al. 2015: 545.

Holotype. MZUB (unnumbered), sex unknown, 190.0 cm TL, Adriatic Sea (mounted skin; presumably adult stage).

Other materials examined ( n = 13). AIM MA30934, 211.0 cm TL, New Zealand(?) (cast); BMNH 1870.12.27.41, 42.5 cm TL, South Australia(?); BMNH 1883.11.29.22, 229.1 cm TL, New South Wales, Australia (holotype of Orthragoriscus ramsayi: mounted skin); INM-1-000568, 280.0 cm TL (measured when fresh), off Hirakata, Ibaraki, Japan (36°51’N, 140°48’E) (sample code KIMo-1 in Sawai et al. 2015b; mounted skin); KMNH VR 100,123, female, 325.0 cm TL (measured when fresh), off Ohse, Ibaraki, Japan (36°34’N, 140°39’E) (sample code OI-1 in Yoshita et al. 2009 and Sawai et al. 2015b; mounted skin); NMNZ P.006345, male, 38.8 cm TL, off Te Kaha, New Zealand (37º39’S, 177º 31’E); NMNZ P.009887, male, 38.5 cm TL, off North Cape, New Zealand (35º13’S, 172º 28’E); NMNZ P.034449, female, 51.5 cm TL, Southern Colville Ridge, New Zealand (36º24’S, 176º 51’E); NMNZ P.036964, female, 45.3 cm TL, off Mahia Peninsula, New Zealand (39º05’S, 178º53’E); NMNZ P.056071, male, 29.3 cm TL, off Raglan Harbour, New Zealand (37º44’S, 174º10’E); OA-Pi-1, 300.0 cm TL (measured when fresh), off Ohse, Ibaraki, Japan (36°34’N, 140°39’E) (sample code OIMo-1 in Sawai et al. 2015b; mounted skin); OMMSFC1085, 91.6 cm TL, coast of Sur, Sultanate of Oman (22°35’N, 59°30’E); OMMSTC1097, 135.0 cm TL, coast of Quriat, Sultanate of Oman (22°16’N, 58°59’E).

Additional specimens not preserved (n = 8). Almsp-1 (sample code in Ahuir-Baraja et al. 2017), female, 240.0 cm TL, Almazora, Spain (39°56’N, 00°03’W); AM-54 (sample code in Yoshita et al. 2009), female, 332.0 cm TL, Aji Island, Miyagi, Japan (38°16’N, 141°28’E); FI-19 (sample code in Yamanoue et al. 2010), female, 265.1 cm TL, Funakoshi Bay, Iwate, Japan (39°23’N, 141°58’E); IO-1 (sample code in Yoshita et al. 2009), female, 181.3 cm TL, Ie Island, Okinawa, Japan (26°45’N, 127°45’E); TaHMo-2 (this study), 162.5 cm TL, off Hualian, Taiwan (24°03’N, 121°37’E); YI-9, 26, 27 (sample code in Yoshita et al. 2009), 3 specimens (3 females), 249.0–269.0 cm TL, Yamada Bay, Iwate, Japan (39°29’N, 142°04’E).

Diagnosis. Mola alexandrini is distinguishable from other species of Mola by the following combination of characters: 14–24 (mode 17) clavus fin rays; 8–15 (12) clavus ossicles (specimens >60.0 cm TL); rounded clavus (usually not wavy and without indents); smooth band on clavus without a back-fold; head with high protuberance (bump) (specimens >162.5 cm TL); scales rectangular on middle region of body, shape developing with age (final shape not established on specimens <70.0 cm TL, but established on specimens >162.5 cm TL); chin with enlarged bump (specimens >135.0 cm TL).

Description. Measurements, counts and observations are shown in Tables 1–2. Body orbicular, deep, and laterally compressed; body surface covered with small rectangular scales (line shaped when viewed from above; Fig. 1a), smooth to the touch; whole body with thick white subcutaneous gelatinous layer, except for pectoral, dorsal and anal fins; mouth small and terminal (snout projecting beyond mouth in some large specimens); teeth in both jaws fused and beak-like; pair of small nostrils in front of eyes; eyes small; gill openings small, oval, located in front of pectoral fins, and covered by soft gill membrane; gill rakers concealed under subcutaneous gelatinous layer; lateral lines present at least on head; many small, white, rounded otoconia instead of otoliths; all fins spineless; caudal fin and pelvic fins absent; pectoral fins small, rounded, located midlaterally and fitting into shallow groove on sides of body; dorsal fin located opposite anal fin, and both fins similarly triangular (fin tips change from acute angle to obtuse angle with age); caudal fin replaced with broad clavus, comprising highly modified elements of the dorsal and anal fins; clavus margin rounded in individuals of all sizes; smooth band at base of clavus between dorsal and anal fins (Fig. 1a), and body scales in smooth band smaller than in surrounding area; urogenital opening immediately in front of anal fin, and anal opening immediately in front of urogenital opening; external differences between the sexes not apparent, but the shape of gonads differs in males and females: ovary single and ball shaped, and testis paired, elongated, and rod-like.

Head bump forming with age, from above eyes to front of dorsal-fin base; chin bump evident with age, from beneath lower jaw to beneath pectoral fins (profile gently curved to bulbous in some large specimens); lateral ridges from head, above and below eyes, to beyond pectoral fins (Fig. 1a), developing with age; clavus ossicles, a snout ossicle, and a chin ossicle developing with age (Fig. 1a).

Colour of fresh specimens. When alive or freshly collected, generally gray or dark reddish brown dorsally, dusky white ventrally, all fins gray or dark reddish brown (similar to dorsal region), and many large or small paler spots and irregular pattern over the body (Figs. 4a, c). Body colour of live bump-head sunfish can instantaneously change to dark, pale and spotted pattern.

Colour of preserved specimens. Body reddish brown or yellow, and live body colours, spots, and patterns faded (Fig. 4d–f).

Distribution and ecological notes. Mola alexandrini is widely distributed in the world’s oceans except for the polar regions (Fig. 5a; ESM Table S1); collected from waters off Japan, Taiwan, Galápagos Islands, New Zealand, Australia, Turkey, Oman, and Spain (Yamanoue et al. 2004, 2010; Bass et al. 2005; Sagara et al. 2005; Yoshita et al. 2005, 2009; Sawai et al. 2009, 2011, 2017a; Yamanoue and Sawai 2012; Thys et al. 2013, Ahuir-Baraja et al. 2017; Nyegaard et al. 2017). Mola alexandrini presumably prefers warmer water temperatures than inhabited by Mola mola; in waters off the Sanriku coast of Japan, sea surface temperatures during the occurrence of M. alexandrini (16.8–25.6°C, average 19.9°C) were on average higher than during the occurrence of M. mola (11.5–25.6°C, average 17.7°C) (Sawai et al. 2011).

Remarks. A large mounted specimen, exhibited on a wall at MZUB, appears identical to the figure in the original description of Orthragoriscus alexandrini by Ranzani (1839). Also, two of the characters used in the original description [i.e. “in parte postica fere ovatum” (= the posterior part is almost oval) and “fronte altissima prominenti” (= forehead projected high)] are consistent with our observations of the holotype (Fig. 3a; Table 2). In addition, the total length of the whole specimen was given as “6. 2. 3” in the original description (Ranzani 1839), which can be interpreted as 6 feet, 2 inches, 3 lines, which is equivalent to 189 cm in English units (metric units), and is very near our measurement of 190.0 cm TL for the holotype. Furthermore, because the original description of O. alexandrini appeared in a journal published by the University of Bologna, the mounted specimen can be regarded as the legitimate holotype of this species.

In Fraser-Brunner’s (1951) comprehensive review of the Molidae, M. alexandrini was regarded as a junior synonym of M. mola, and this synonymization was later accepted by both Parenti (2003) and Eschmeyer et al. (2017). However, Fraser-Brunner (1951) examined the holotype of Orthragoriscus ramsayi, but not the holotype of O. alexandrini. Since he believed the former species was distributed only in the South Pacific, he appears to have automatically treated specimens found in the Northern Hemisphere as M. mola. Furthermore, the holotype of O. ramsayi only possesses two of the six key characters of Mola ramsayi (i.e. a rounded clavus and 12 ossicles) that were proposed by Fraser-Brunner (1951) (Table 2), and two other characters (16 clavus fin rays, and the ossicles on the paraxial rays are separate and smaller than the other ossicles) were not observed on the specimen, likely owing to the thickness of the skin. Regarding the remaining specimen, BMNH 1883.11.29.22, the clavus fin rays and state of the paraxial ossicles are not visible. Thus, we consider it likely that Fraser-Brunner obtained these two diagnostic characters solely from the latter specimen (Fig. 4e). The diagnostic character (clavus ossicles being broader than the spaces between them) is also largely congruent with our observations. However, the character ‘absence of smooth band’ contradicts our finding of a smooth band present on the holotype of O. ramsayi (Fig. 6a; cf. Strom 2016). The smooth band was probably simply overlooked by Fraser-Brunner (1951), since this feature is not as clear in M. alexandrini as in M. mola, owing to different morphology of the body scales (Sawai et al. 2017a). Furthermore, the smooth band tends to become indistinct on a mounted skin (Sawai et al. 2015b), possibly owing to the application of chemical coatings used for preservation. Thus, although the absence of a smooth band has generally been regarded as a diagnostic character for distinguishing between M. ramsayi and M. mola (cf. Jawad et al. 2012; Thys et al. 2013), it is not a valid diagnostic character.

The holotype of O. ramsayi is a mounted skin (poor condition) in BMNH (Fig. 2). The catalogue number of the former was given as BMNH 1888.11.29.22 in some literature (e.g. Bray and Hoese 2006; Matsuura 2017), but the correct catalogue number is BMNH 1883.11.29.22 (The Natural History Museum 2017). The original description of O. ramsayi by Giglioli (1883: 315) was brief and lacked detail, but four salient characters were mentioned. These were the overall shape and size of the fish, the form of the clavus, and a fourth character that is somewhat unclear: “small carinate horny scales, which appear to cover the osseous granulation of the dermis”, which may pertain to the morphology of the body scales or alternatively refer to the presence of clavus ossicles. Our examination of the holotype revealed that the state of several of its features (clavus ossicles, rounded clavus, smooth band, rectangular body scales, head bump, chin bump) are consistent with those of Mola sp. A sensu Yoshita et al. (2009) (Figs. 6a–b; Table 2). Four of these characters (rounded clavus, smooth band, head bump, chin bump) are also evident on a photograph of the recently restored holotype of O. ramsayi (Strom 2016).

Therefore, we propose that the holotypes of M. alexandrini and M. ramsayi represent the same species and furthermore match genetically with Mola sp. A sensu Yoshita et al. (2009). This conclusion is strongly supported by the reported distributions of these species, evidenced by their type localities (Mediterranean Sea and Australia, respectively), which have since been confirmed by additional specimens identified as these purported species in molecular and morphological studies (i.e. Yoshita et al. 2009; Ahuir-Baraja et al. 2017; Nyegaard, unpublished data).

Although the scientific name of this species has been accepted as M. ramsayi based on Fraser-Brunner (1951), M. alexandrini was used as a valid name by others (e.g. Barnard 1935, 1947, 1948; Smith 1950, 1965). Therefore, we conclude that M. alexandrini is a valid species according to Article 23.9.1.1 of ICZN (1999). Moreover, to date, the maximum recorded (not estimated) weight of M. alexandrini is 2,300 kg for a 272 cm TL female caught by set net off the coast of Kamogawa, Japan, on 16 August 1996 (Roach 2003; Nakatsubo et al. 2007b; Pope et al. 2010; Sawai 2017). This specimen, which is currently regarded as the world’s heaviest bony fish specimen, was previously identified as M. mola (Roach 2003; Pope et al. 2010). Here, we re-identify it as M. alexandrini—from the head bump, chin bump, and rounded clavus (see also Nakatsubo et al. 2007b: table 2; Soichi 2009: fig. 2B; Pope et al. 2010: fig. 1). Therefore, the world’s heaviest bony fish that has been actually weighed and recorded to date is a specimen of M. alexandrini, and not M. mola. However, this species may grow even larger; we could confirm a maximal length of 332.0 cm TL for a female specimen (not weighed) caught off Aji Island, Miyagi, Japan, on 11 August 2004 (Fig. 4a; see also Yoshita et al. 2009).

Neotype designation for Tetraodon mola Linnaeus 1758

Tetraodon mola Linnaeus 1758 is considered valid as one of two species of Mola, that is, Mola mola by Fraser-Brunner (1951), which has since been widely accepted (e.g. Martin and Drewry 1978; Parenti 2003; Bray 2008; Eschmeyer et al. 2017). Yoshita et al. (2009) and Matsuura (2015) suggested that genetically identified Mola sp. B has the morphological characteristics of this species that were proposed by Fraser-Brunner (1951). Although the presence of a smooth band at the base of the clavus is not a valid diagnostic character, the other morphological characters are consistent with those found in Mola sp. B, and distinguish M. mola from Mola alexandrini and Mola tecta (Fig. 1b; Table 2). Therefore, we conclude that Fraser-Brunner’s (1951) M. mola and Mola sp. B (of later investigators) represent the same species.

However, there is no known type material of Tetraodon mola (see Tortonese 1990; Eschmeyer et al. 2017). For comparison with M. alexandrini, we investigated the description of Tetraodon mola by Linnaeus (1758). The original Latin description was brief: “T. laevis compressus, cauda truncata: pinna brevissima dorsali analique annexa” (= Tetraodon smooth compressed, tail truncated: connected to short-based dorsal and anal fins). It is likely that Linnaeus (1758) described Ranzania laevis owing to the terms ‘smooth’ and ‘short dorsal and anal fins,’ which are in contrast to the coarse body scales of M. mola. Retzius (1785) indicated that the skin of M. mola is rough and that of R. laevis is smooth, and also pointed out that figures of R. laevis in Bianchi (1746), cited by Linnaeus, represent R. laevis, and not M. mola. Accordingly, Retzius (1785) believed that Linnaeus described the species without examining any specimens, but only based on the literature. Indeed, three references, Artedi (1738), Bianchi (1746) and Gronovius (1754), were used for the description of this species by Linnaeus (1758). Artedi (1738) did not include any illustrations but the following descriptions, which indicate species of Mola: “cutis dura and aspera” (= skin hardened and rough); “pondus 100 librarum” (= 100 pounds weight); “cauda semicircularis” (= semicircular tail); “ossicula in pinnis” (= ossicle in fin). Most of the eleven references cited by Artedi (1738) also indicated species of Mola (i.e. Rondelet 1554: 424, unnumbered fig.; Salviani 1554: 154, fig. P55; Gesner 1563: 85 left, unnumbered lower fig.; Gesner 1604: 640, unnumbered fig.; Plinius Secundus et al. 1608: 1365, 1391; Aldrovandi 1613: 412, unnumbered fig.; Jonston 1632: 419; Jonston 1650: 30, pl. 9, fig. 2; Charleton 1668: 129; Willughby 1686: 151; Ray 1713: 51), but exceptionally Gesner (1604: 635, unnumbered upper fig.) described a fish with a caudal fin, which cannot be interpreted as a species of Mola. Presumably, Artedi (1738) cited that reference by mistake. The description and illustration of Bianchi (1746) also represent R. laevis (Wheeler 1979). The description in Gronovius (1754) is brief and without illustrations, and does not have information enough to specify the species, but rather two references are cited: Jonston (1660) and unpublished data of Peter Artedi (Arted. mss. ad Sebam). However, the description of ‘Mola peregrina’ in Jonston (1660: pl. 9, fig. 1) clearly represents R. laevis, as pointed out by Gudger (1936). Steenstrup and Lütken (1898: 59) mentioned that Linnaeus (1758) described T. mola based on descriptions of species of Mola and R. laevis. The type locality of T. mola is the Mediterranean Sea, and M. alexandrini, M. mola, and R. laevis do occur in this area, but not M. tecta. Therefore, we consider that Linnaeus (1758) described T. mola only based on the previous descriptions of R. laevis and M. alexandrini and/or M. mola. This is consistent with the fact that the type material of T. mola cannot be found to date, despite searches for the holotype by previous and present investigators at the Linnaean fish collections of the Swedish Museum of Natural History (Fernholm and Wheeler 1983; Sven Oscar Kullander, personal communication), at the Linnean Society of London (Wheeler 1985; The Linnean Society of London 2017), and at the Museum of Evolution (formerly, the Zoological Museum of the University of Uppsala, Sweden) (Wheeler 1991; Erica Mejlon, personal communication).

Tetraodon mola was described based on multiple species, and a name-bearing type is necessary to define the nominal taxon objectively (ICZN 1999, Art. 75.1). Tetraodon mola, however, should not be regarded as a synonym of R. laevis, because the descriptions that Linnaeus (1758) based his designation on included species of Mola; thus confusion should be avoided by changing the current status, which has long been accepted (ICZN 1999, Art. 23.2). The type locality of T. mola is the Mediterranean Sea, and M. mola is more dominant in the Mediterranean than M. alexandrini, based on our genetic and morphological examinations of specimens from this area (Sawai and Yamanoue, unpublished data). Accordingly then, the current status of M. mola is appropriate and should not be changed.

Ozodura orsini was regarded as a synonym of T. mola because the holotype has the following characters, diagnostic for this species: 13 clavus fin rays; 7 clavus ossicles; conical body scales (branching tips lost owing to chemical coating) [see below “Comparisons”; Figs. 3b, c (right), 6f; Table 2]. This holotype was obtained from the Adriatic Sea in the Mediterranean (Ranzani 1839); MZUB maintains a research collection with proper facilities for preserving name-bearing types, and this specimen is herein designated as the neotype of T. mola, in accordance with Article 75.3 of ICZN (1999).

Additionally, previous genetic studies suggested the possibility that two clades of M. mola (Atlantic vs. Pacific) can be divided into the distinct species/subspecies (Bass et al. 2005; Yoshita et al. 2009; Ahuir-Baraja et al. 2017; Sawai et al. 2017a; Nyegaard et al. 2017). However, our data are not sufficient and further study is needed.

Comparisons

The total counts of the rays of three fins (pectoral fin rays, dorsal fin rays, and anal fin rays) among three species of Mola are similar to each other (Table 2), and thus cannot be a useful diagnostic character. Concerning average clavus fin rays and average D + C + A fin rays, those of M. alexandrini and M. tecta are similar to each other, but higher than those counts for M. mola (Table 2). Concerning specimens in which the final number (the constant number) of clavus ossicles has been formed (specimens >65 cm TL), the average number of clavus ossicles among the three species of Mola clearly differs: 11.8 in M. alexandrini vs. 8.6 in M. mola vs. 5.8 in M. tecta (Table 2). The shape of the clavus rear margin in large specimens is clearly distinguishable among the three species of Mola: rounded in M. alexandrini vs. wavy (or lobed) in M. mola vs. rounded with an indent in M. tecta (Fig. 1). Although the presence or absence of a smooth band at the base of the clavus is not a useful character for distinguishing the three species of Mola (Table 2), the presence of a smooth band back-fold is likely to be a useful key character for M. tecta (Fig. 1; Table 2). However, meticulous examination is needed because the back-fold in small preserved specimens of M. tecta may be difficult to detect (see Table 2; Nyegaard et al. 2017). In addition, a back-fold is occasionally found in some specimens of both M. alexandrini and M. mola, although they appear faint (Table 2; Sawai and Nyegaard, personal observation). The presence of a head bump and a chin bump is a valid diagnostic character for M. alexandrini only in large specimens (Figs. 1, 4; Table 2), but a small head bump and a small chin bump are occasionally found in some specimens of M. mola (Sawai, personal observation). The shape of the body scales is clearly different among the three species of Mola: rectangular or line shaped in M. alexandrini vs. conical with branching tips or dot shaped in M. mola vs. conical without branching tips or simple and dot shaped in M. tecta (Fig. 1).

Two diagnostic characters of Mola, as proposed by Fraser-Brunner (1951), were found only on a small number of specimens of M. alexandrini, as they require dissection or X-ray examinations for accurate assessment. The clavus ossicles in M. alexandrini tend to be broader than the spaces between them, while the clavus ossicles in M. mola tend to be narrower than the spaces between them (see Fraser-Brunner 1951: figs. 14, 16; Sawai et al. 2017a; Sawai, personal observation). The clavus ossicles on the paraxial rays in M. alexandrini are separate and smaller than the remaining ossicles, whereas the clavus ossicles on the paraxial rays in M. mola are united to form a single ossicle which is larger than the other ossicles (see Fraser-Brunner 1951: figs. 14, 16; Sawai et al. 2017a: fig. 3A; Sawai, personal observation). In M. tecta, the clavus ossicles on the paraxial rays are clearly separate and similar in size to the adjacent clavus ossicles, and a smooth band back-fold exists between the paraxial clavus ossicles (Nyegaard et al. 2017).

Furthermore, four diagnostic characters (number of clavus ossicles, shape of the clavus margin, head bump, chin bump) change as individuals grow larger, becoming well evident in large specimens but not useful as diagnostic characters in small specimens (Table 2). Macroscopic observations of body scale morphology on larger specimens revealed differences among the three species of Mola (Fig. 1); yet, based on examinations of small specimens, scale morphology may appear superficially similar among the species when the scales are less developed (Sawai, personal observation). However, scale morphology in M. tecta appears to be established at >50 cm TL (Nyegaard et al. 2017). Similarly, Yoshita et al. (2009) proposed four morphometric features that differ between M. alexandrini and M. mola, with large specimens of both species displaying extreme differences (especially, the BD/TL and TBD/TL of M. alexandrini were higher by 7.5 % than those of M. mola), while small specimens of the two species were similarly proportioned (Fig. 7; Table 1). This also appears to be the case for M. tecta (Nyegaard et al. 2017), although future investigators should take care to examine the body proportions of numerous individuals.

The smallest specimens of M. alexandrini (i.e. NMNZ P.056071, 29.3 cm TL, Fig. 4f) and M. mola (i.e. NSMT-P75065, 28.2 cm TL, Fig. 8f) examined in this study were distinguished molecularly (Yoshita et al. 2009; Nyegaard et al. 2017), but were similar in appearance. Presumably, small specimens of M. tecta are morphologically similar to the other two species of Mola. This points to the difficulty of identifying small specimens of Mola and highlights the need to include small specimens of Mola in future investigations combining morphology and molecular analyses.

Fig. 8
figure 8

Specimens of Mola mola. a YI-16, 277.0 cm TL; b NMNZ P.002629, 253.6 cm TL (cast); c YI-18, 193.7 cm TL (photographed by Y. Yoshita); d KAUM–I. 19082, 126.4 cm TL; e NSMT-P 76575, 45.5 cm TL; f NSMT-P 75065, 28.2 cm TL. Red square, 10 cm × 10 cm

Full size image

Several diagnostic characters of the three species of Mola develop with age. The head and chin profiles in M. alexandrini acquire bumps with age (Figs. 1a, 4). The shape of the clavus margin in M. mola becomes wavy or lobed with age, as the clavus edge grows beyond the ossicles (Figs. 1b, 8). In contrast, the external morphology of M. tecta >50 cm TL does not change significantly with age (Fig. 1c; Nyegaard et al. 2017). The clavus ossicles of M. mola develop and increase in number in specimens up to approximately <60 cm TL, when they reach a stable number (Sawai, unpublished data). In M. tecta, the final number of clavus ossicles is reached at >65 cm TL (Nyegaard et al. 2017); in M. alexandrini, the final number of clavus ossicles is established in specimens by at least 60 cm TL. Lateral ridges above and below the eyes develop with age in M. alexandrini and M. mola (Figs. 1, 4, 8); however, the dorso-lateral ridge in M. tecta remains slight and short with age (Nyegaard et al. 2017). The tips of the dorsal and anal fins change from an acute angle to an obtuse angle with age in all three species of Mola (Figs. 4, 8; Nyegaard et al. 2017). Clavus fin rays were not readily visible in large specimens owing to the thickness of the subcutaneous gelatinous layer (Table 2); thus, dissection was needed to count the rays in large specimens. A study of these morphological characters with ontogenetic development is still needed, since this study did not obtain data for medium-sized specimens.

Despite that the holotype of M. alexandrini and the neotype of M. mola are of remarkably different sizes, the diagnostic characters clearly differentiate them as different species. Specimens SLM-1 (M. mola; sample code YS-2 in Yoshita et al. 2009) and KMNH VR 100,123 (M. alexandrini; sample code OI-1 in Yoshita et al. 2009), both very large mounted skins, are valuable because the species identities, were confirmed genetically (Yoshita et al. 2009). In future studies, consideration of these large specimens may still inform our understanding of the morphology of the genus Mola.

Based on morphologically and genetically verified M. alexandrini from this and other studies, this species is widely distributed in the world’s oceans (Fig. 5a; ESM Table S1). In comparison, while M. tecta appears to be distributed predominantly in the temperate waters of the Southern Hemisphere (Nyegaard et al. 2017: fig. 10). Based on M. mola specimens examined in this study as well as the genetically identified specimens in previous studies, the distribution of M. mola is biased to the northern hemisphere with no records currently from the Indian Ocean (Fig. 5b; ESM Table S1). However, the distribution of each species of Mola needs to be reconsidered in future studies due to widespread species confusion and taxonomic problems, especially in M. mola.

Discussion

In the present study, Mola alexandrini, Mola mola and Mola tecta are found to be valid species names for Mola sp. A, Mola sp. B, and Mola sp. C, respectively (cf. Yoshita et al. 2009; Nyegaard et al. 2017). Mola alexandrini, senior synonym of Mola ramsayi, is herein redescribed based on 21 specimens together with the rediscovered holotype, and a neotype of Mola mola is designated to compare between these species, which have long been confused with each other.

The only molid genus and species newly described since Fraser-Brunner’s (1951) revision was Pseudomola lassarati. This description was based on one specimen (109 cm TL) from the Ivory Coast (Cadenat 1959), but the whereabouts of the holotype are unknown (Parenti 2003; Eschmeyer et al. 2017). Cadenat (1959: figs. 9–11) proposed this new genus distinguishable from Mola and Masturus by the following combination of characters: body white on the abdominal region, and numerous, nonsymmetrical (not on both sides), white spots and reticulations dorsally; rectangular body scales; and clavus rounded, without median projection, ossicles, and supported by 20 fin rays. Although Blache et al. (1970) treated P. lassarati as valid, Tortonese (1990) treated it as a nomen nudum, although Tyler (1980), Parenti (2003), and Eschmeyer et al. (2017) treated it as a junior synonym of Masturus lanceolatus. However, the aforementioned characters of P. lassarati are fairly consistent with those of M. alexandrini (Fig. 4; Table 2), except for the reported lack of ossicles in P. lassarati. However, it is possible that the ossicles of P. lassarati were overlooked because they are relatively small compared with the large body. In addition, some specimens of Mola alexandrini have numerous (large or small) white spots scattered over the body (Fig. 4c), very similar to those of Masturus lanceolatus. Therefore, Gudger (1937) and Tyler (1980) regarded this species as a synonym of Masturus lanceolatus, based on the body colour. For these reasons, we conclude that P. lassarati is a junior synonym of Mola alexandrini.

Cephalus ortagoriscus Risso 1827 may possibly be another senior synonym of Mola alexandrini as its 18 clavus fin rays closely match the number in Mola alexandrini (Table 2). However, we regard this species as a nomen dubium because other morphological characters are unknown, as no type material can be found.

Orthagoriscus eurypterus Philippi 1892 (type locality, Chile; 222 cm TL) was regarded as a junior synonym of Mola ramsayi by Parenti (2003) and Eschmeyer et al. (2017). Fraser-Brunner (1951) regarded O. eurypterus as a questionable synonym of Mola ramsayi, while Oliver Schneider (1930) and Nyegaard et al. (2017) reviewed O. eurypterus and concluded it was possibly a junior synonym of Mola mola. Based on the current data, it is likely that O. eurypterus is a junior synonym of Mola mola, rather than a junior synonym of Mola alexandrini.

Nine nominal species (Mola aculeata Koelreuter 1766; Diodon mola Pallas 1770; Diodon nummularis Walbaum 1792; Orthragoriscus hispidus Bloch and Schneider 1801; Cephalus pallasianus Shaw 1804; Mola hispida Nardo 1827; Diodon carinatus Mitchill 1828; Molacanthus pallasii Swainson 1839; Pallasia pallasi Nardo 1840) were described based on the multi-pointed star forms of the prejuvenile stage of Mola. Thus far, prejuvenile specimens of Mola collected have generally been identified as Mola mola (Martin and Drewry 1978), while no prejuvenile specimens of the other species have ever been identified. Therefore, we could not determine whether any of the nine nominal species that were described based on prejuvenile specimens represent Mola alexandrini; hence, further studies are needed to clarify the status of those species.

Ranzani (1839) counted the fin rays of the holotype of Mola alexandrini as 10 pectoral fin rays, 15 dorsal fin rays, and 13 anal fin rays, and the neotype of Mola mola as 12 pectoral fin rays, 16 dorsal fin rays, 14 anal fin rays, and 14 clavus fin rays. These counts differ from our findings (Table 2), but Ranzani (1839) possibly miscalculated the counts since it is difficult to count fin rays in mounted specimens (Fig. 3). Especially, the clavus fin rays, as well as the state of the clavus ossicles on the paraxial rays, are generally not visible in large specimens owing to the thickness of the subcutaneous gelatinous layer. Moreover, some clavus fin rays may have been included as dorsal and/or anal fin rays.

Many of the articles reviewed by Fraser-Brunner (1951) were based on cursory examinations of relatively few specimens (Matsuura 2015). Therefore, he could not fully appreciate intra- and inter-specific variations in the characters of Mola alexandrini and Mola mola. For instance, the intra-specific variation in the counts of fin rays and ossicles is higher here than would be anticipated by reading Fraser-Brunner (1951) (Table 2).

Until now, the common name ‘southern ocean sunfish’ has been used for M. alexandrini. This common name stemmed from the belief that this species occurs only in the Southern Hemisphere, as discussed by Fraser-Brunner (1951) (Thys et al. 2013). Matsubara (1955) similarly named this species ‘Goshu-manbo’ in Japanese (‘Goshu’ meaning ‘Australia,’ and ‘manbo’ meaning ‘ocean sunfish’), based on the geographical distribution proposed by Fraser-Brunner (1951). It seems likely therefore that Mola alexandrini in the Northern Hemisphere has been widely confused with Mola mola. Based on the species’ much wider distribution as revealed in recent studies, Thys et al. (2013) pointed out that the common name southern ocean sunfish is no longer appropriate. Therefore, we propose the common name bump-head sunfish for Mola alexandrini, based on its most prominent morphological character; for the same reason, we adopt the Japanese common name Ushi-manbo (‘Ushi’ meaning cow, referring to the head profile), as already proposed by Yamanoue et al. (2010).

Comparative materials. Mola mola (n = 13): BMNH2006.12.15.1, 70.1 cm TL, Whitstable, United Kingdom (51°22’N, 01°03’E); KAUM–I. 19082, 27983, 2 specimens, 109.9 and 126.4 cm TL, Ibusuki, Kagoshima, Japan (31°10’N, 130°32’E); MZUB (unnumbered) (neotype), 59.7 cm TL, Adriatic Sea (mounted skin); NMNZ P.002629, 253.6 cm TL, Palliser Bay, New Zealand (42°34’S, 175°05’E) (cast); NSMT-P 75037, 75065–75067, 4 specimens, 28.2–49.1 cm TL, North Pacific Ocean (42°46’N, 170°19’E) (sample code NP-8, 11–13 in Yoshita et al. 2009); NSMT-P 76575, 45.5 cm TL, Ryouri, Iwate, Japan (39°02’N, 141°52’E); NSMT-P 111410–111411, 2 specimens (1 male and 1 female), 28.6 and 29.6 cm TL (measured when fresh), North Pacific Ocean (38°00–04’N, 157°00–04’W); SLM-1, female, 275.0 cm TL (measured when fresh), Yunotsu, Shimane, Japan (35°05’N, 132°22’E) (sample code YS-2 in Yoshita et al. 2009 and Sawai et al. 2015b; mounted skin). Mola tecta (n = 26): NMNZ P.057679 (holotype), male, 101.1 cm TL (measured when fresh), North Taranaki Bight, New Zealand (38°26’S 174°09’E); 25 other specimens, listed in Nyegaard et al. (2017).

Additional data obtained in the field (not preserved whole body specimens) (Mola mola, n = 22). FI-86 (sample code in Sawai et al. 2009), female, 139.9 cm TL, Funakoshi Bay, Iwate, Japan (39°23’N, 141°58’E); KI-8 (sample code in Yoshita et al. 2009), female, 260.1 cm TL, off Kamaishi, Iwate, Japan (39°16’N, 141°53’E); NP-2–7 (sample code in Yoshita et al. 2009), 6 specimens, 28.4–33.0 cm TL, North Pacific Ocean (42°46’N, 170°19’E); OtI-8–10 (sample code in Yoshita et al. 2009), 3 specimens, 35.6–40.0 cm TL, off Otsuchi, Iwate, Japan (39°20’N, 141°55’E); YI-1, 5, 8, 16–21, 22, 24 (sample code in Yoshita et al. 2009), 11 specimens (including at least 3 males, 4 females), 49.0–277.0 cm TL, Yamada Bay, Iwate, Japan (39°29’N, 142°04’E).