INTRODUCTION
Manta and devil rays (family Mobulidae) are iconic, endangered animals, yet their large size and elusive behaviour has caused a scarcity of representation in ichthyological collections and historical records. Descriptions are often based on a single specimen at an early stage of development or in poor state of preservation (Notarbartolo-di-Sciara, Reference Notarbartolo-di-Sciara1987). In addition, their external resemblance led to numerous taxonomic ambiguities and misidentifications even in recent literature, especially in the Mobula genus (Couturier et al., Reference Couturier, Marshall, Jaine, Kashiwagi, Pierce, Townsend, Weeks, Bennett and Richardson2012). Updated information on mobulid occurrences and distribution ranges is relevant as it might have significant implications for the conservation of their populations (Jaine et al., Reference Jaine, Couturier, Weeks, Townsend, Bennett, Fiora and Richardson2012). The available information on the occurrence of mobulids in oceanic regions is particularly short, in spite of their largely pelagic behaviour. The Azores, an isolated archipelago located between the warm temperate and subtropical North Atlantic provinces, is one such region (Figure 1).
Location of the Azores Archipelago in the central North Atlantic and updated global distribution of the four mobulid rays possibly occurring in the region based on IUCN distribution maps (www.iucnredlist.org). Mobula mobular is retained as the most plausible species of the M. mobular/M. japanica species complex, but both species may occur in the region.
The first reference of mobulids for the region dates from 1777, when Duhamel du Monceau described a fish named ‘Raie cornue’ from the Azores and ‘Mobular’ from the Caribbean. Thereafter, two species, giant manta Manta birostris (Walbaum, 1792) and giant devil ray Mobula mobular (Bonnaterre, 1788), were cited by various authors, and a third, Chilean devil ray Mobula tarapacana (Philippi, 1892), was recently added (Afonso, Reference Afonso2001). However, very few references are associated with preserved specimens or detailed descriptions and misidentifications may well have occurred.
The aim of this work was to clarify which mobulid species occur in the Azores exclusive economic zone (EEZ) and to provide an updated checklist of the family for the region.
MATERIALS AND METHODS
A total of 2252 photographs and 188 videos of mobulids were analysed from the 1990–2013 image database of IMAR-UAz (ImagDop) and from a web-based effort launched in 2012 (http://mantacatalogazores.wix.com/mobulaid). The pictures were taken by scientific, professional and recreational underwater photographers, mostly at the shallow banks of Princesa Alice, Formigas and Ambrósio where targeted mobulid diving operations concentrate their effort. The identification of the species in the images was then conducted based on external distinctive features as per Marshall et al. (Reference Marshall, Compagno and Bennett2009) for Manta species, and Notarbartolo-di-Sciara (Reference Notarbartolo-di-Sciara1987) and Stevens (Reference Stevens2010) for Mobula species. The occurrences in the image database were then assembled with previous citations of mobulids for the Azores EEZ in scientific journals, underwater guides and public articles, after a re-analysis. Personal information (communications and old photographs) was also incorporated. An annotated checklist for the Mobulidae family, Order Myliobatiformes (Compagno, Reference Compagno and Hamlett1999), was produced based on this information. The rules adopted for species nomenclature and synonymy were based on Eschmeyer (Reference Eschmeyer2013). Primary order records, i.e. whenever specimens or individuals were directly studied or observed by the author, are denoted by the symbol ‘°’ (e.g. Afonso, Reference Afonso2001°). Otherwise, the reference is of secondary order and based on another citation.
RESULTS
The database was dominated by two species: M. birostris and M. tarapacana (Figure 2). A third species appears in one photograph from 2004 and a video from 2013. Although an absolute identification of the species could not be reached in either case, the morphological features suggest they correspond to the M. mobular/spinetail devil ray Mobula japanica (Müller & Henle, 1841) species complex (hereafter called Mobula cf. mobular). The annotated checklist is presented bellow.
Mobulids of the Azores and their major external distinctive features: (A) ventral view of Manta birostris showing dark coloration around mouth extending to first gill openings, dark spots between the fifth gill openings and the opening of the cloaca, and dark coloured posterior edges of pectoral fins; (B) dorsal view showing black surface extending to dark T-shape separating large white shoulder patches, and blazes of white on the tips of the pectoral fins; (C) ventral view of Mobula tarapacana showing irregular margin between white and grey areas, and grey shaded gill areas; (D) dorsal view showing brown to olive green surface with pronounced dorsal ridge; (E) ventral view of Mobula cf. mobular showing white surface; (F) dorsal view showing dark blue dorsal surface with a crescent-shaped black band on the head base, triangular light areas on shoulders and white tip of the dorsal fin.
Manta birostris (Walbaum, 1792), giant manta ray, manta, manta gigante.
REFERENCES FOR THE AZORES
Collins, Reference Collins1954°; Ré, Reference Ré1979°; Wirtz, Reference Wirtz1994, Reference Wirtz2011; Saldanha, Reference Saldanha1995°; Santos et al., Reference Santos, Porteiro and Barreiros1997; Garcia, Reference Garcia2008; Barreiros & Gadig, Reference Barreiros and Gadig2011; Gallagher et al., Reference Gallagher, Porteiro and Santos2013.
REMARKS
Collins (Reference Collins1954), the first confirmed citation of this species for the archipelago, reports a specimen harpooned and photographed in Horta harbour in 1951. A total of 128 underwater photographs of M. birostris individuals were identified in the database. Noticeably, all individuals analysed from the photographs showed the external morphological characteristics distinctive of the M. birostris form, namely the dark coloration of the mouth, and not those of the form which has been hypothesized to correspond to a putative third species occurring in the western Atlantic (Marshall et al., Reference Marshall, Compagno and Bennett2009).
MISIDENTIFICATIONS
M. mobular by Saldanha, Reference Saldanha1995.
Mobula cf. mobular (Bonnaterre, 1788), giant devil ray, jamanta.
REFERENCES FOR THE AZORES
Sampaio, Reference Sampaio1904°; Ferreira, Reference Ferreira1939°; Santos et al., Reference Santos, Porteiro and Barreiros1997.
REMARKS
Duhamel du Monceau (Reference Duhamel du Monceau1777) describes a fish named ‘Horned ray from the Azores’ but the single individual described is from the Mediterranean. Therefore, although he knew about the term being used in the Azores or referring to fish from the region, this is not considered here as a valid reference for the region as well as the various subsequent second order citations (see Santos et al., Reference Santos, Porteiro and Barreiros1997). Sampaio (Reference Sampaio1904) refers to the infrequent occurrence of a Ceratopterus sp. but he does not provide any other information. The original species identification as considered by Santos et al. (Reference Santos, Porteiro and Barreiros1997) is retained for this and Ferreira's (Reference Ferreira1939) citations, although these could very well be misidentifications of M. tarapacana. There were two confirmed records of M. cf. mobular in the image database: a photograph of an adult individual (~2 m disc width) accidentally caught in a commercial surface longliner in 2004 and an underwater video showing five adult individuals (~2 m disc width) swimming in blue water off Pico Island in 2013.
SYNONYMS AND MISSPELLINGS
Ceratopterus sp. by Sampaio, Reference Sampaio1904; Cephaloptera giorna by Ferreira, Reference Ferreira1939.
Mobula tarapacana (Philippi, 1892), Chilean devil ray, jamanta.
REFERENCES FOR THE AZORES
Wirtz, Reference Wirtz1990°, Reference Wirtz1991°, Reference Wirtz1994°, Reference Wirtz2011; Santos et al., Reference Santos, Monteiro, Alves, Isidro, Martins and Hawkins1994, Reference Santos, Hawkins, Monteiro, Alves and Isidro1995; Afonso, Reference Afonso2001°; Garcia, Reference Garcia2008; Barreiros & Gadig, Reference Barreiros and Gadig2011; Gallagher et al., Reference Gallagher, Porteiro and Santos2013.
REMARKS
Afonso (Reference Afonso2001) is the first original reference of M. tarapacana for the Azores as all of Wirtz's original references were confirmed by the author to be misidentifications. A total of 2125 underwater images of this species were identified in the database.
MISIDENTIFICATIONS
M. mobular by Wirtz, Reference Wirtz1990, Reference Wirtz1991, Reference Wirtz1994; Santos et al., Reference Santos, Monteiro, Alves, Isidro, Martins and Hawkins1994, Reference Santos, Hawkins, Monteiro, Alves and Isidro1995; Garcia, Reference Garcia2008.
DISCUSSION
This study confirms that two species of mobulid rays (M. birostris and M. tarapacana) frequently occur in Azorean waters. Since M. tarapacana represents by far the most abundant species (94%) in the 23-year photographic database we assume that it must have previously occurred in the region but was repeatedly misidentified as M. mobular. Although this assumption was impossible to ascertain in some older references due to the lack of descriptions and specimens preserved (e.g. Sampaio, Reference Sampaio1904; Ferreira, Reference Ferreira1939), we could confirm it in all recent references through photographs and positive confirmation of the authors.
Considering that (1) the image database is based on a minimum of ~100 events per year spread over 23 years across the archipelago, and that (2) this period covers the advent of touristic diving in the region and therefore benefits from the large effort in the areas where mobulids are actively looked for, it appears that M. tarapacana occurs frequently in the region (>1 occurrence per week during the summer aggregation season across all years) while M. birostris occurs occasionally (several encounters per year). Mobula tarapacana was mostly photographed and filmed in aggregations of several to tens of individuals in the offshore reefs and banks, whereas M. birostris occurred solitarily in these same habitats and also on more coastal reefs. These facts are consistent with the behaviour, habitat preferences and distributions of these two species described in the literature (Clark et al., Reference Clark, Smith and Bizzarro2006; Marshall et al., Reference Marshall, Compagno and Bennett2009, Reference Marshall, Bennett, Kodja, Hinojosa-Alvarez, Galvan-Magana, Harding, Stevens and Kashiwagi2011), although information on M. tarapacana is very limited. They also confirm the previous notion that the seasonal occurrence of these species at remote seamounts and oceanic, isolated islands suggest a strong seasonal migratory behaviour (e.g. Marshall et al., Reference Marshall, Compagno and Bennett2009, Reference Marshall, Bennett, Kodja, Hinojosa-Alvarez, Galvan-Magana, Harding, Stevens and Kashiwagi2011). In this regard, the Azores seems to be one of the few locations in the world where M. tarapacana is known to aggregate in large numbers. This species is considered to be the most oceanic and least known of the mobulids (Clark et al., Reference Clark, Smith and Bizzarro2006), and such fundamental aspect of their ecology deserves urgent research.
In addition, we provide the first confirmed occurrences for the region of either M. mobular or M. japanica, or both. In fact, there is an ongoing debate on the taxonomic status of these two species because of their close resemblance and separation based on the examination of a single M. mobular specimen. According to Notarbartolo-di-Sciara (Reference Notarbartolo-di-Sciara1987), the two species may be distinguished by their maximum size (M. mobular being larger) and morphometrics (M. mobular reaching greater disc width relative to body size) but observing these features in the wild is particularly difficult. Furthermore, the author noted ‘the overlap in their morphological features is potentially indicative of single species status’. Mobula japanica is probably circumgobal in inshore and possibly oceanic environments of warm temperate and tropical seas (Notarbartolo-di-Sciara, Reference Notarbartolo-di-Sciara1987; White et al., Reference White, Clark, Smith and Bizzarro2006; Couturier et al., Reference Couturier, Marshall, Jaine, Kashiwagi, Pierce, Townsend, Weeks, Bennett and Richardson2012), but detailed information on its distribution and large scale movements patterns remains scarce (White et al., Reference White, Clark, Smith and Bizzarro2006; Bustamante et al., Reference Bustamante, Couturier and Bennett2012). Mobula mobular is an offshore, epipelagic ray that occasionally frequents shallow coastal waters of the Mediterranean Sea (Notarbartolo-di-Sciara et al., Reference Notarbartolo-di-Sciara, Serena and Mancusi2006) and, possibly, the north-east Atlantic (Notarbartolo-di-Sciara, Reference Notarbartolo-di-Sciara1987; Notarbartolo-di-Sciara et al., Reference Notarbartolo-di-Sciara, Serena and Mancusi2006; Couturier et al., Reference Couturier, Marshall, Jaine, Kashiwagi, Pierce, Townsend, Weeks, Bennett and Richardson2012). The possible occurrence of both species in the Azores has been suggested before (Wirtz, Reference Wirtz2011). Genetic and morphological analysis of additional specimens from the Atlantic is needed in order to clarify the taxonomic status and distribution ranges of these two species.
The Azores constitutes the known northernmost distribution limit for all of these three, maybe four mobulid rays in the Atlantic and globally, with the exception of a couple of citations of vagrant isolated individuals (Notarbartolo-di-Sciara, Reference Notarbartolo-di-Sciara1987; Tomita et al., Reference Tomita, Kawai, Matsubara and Nagata2013). This fact demonstrates the role of the region as an important habitat within the normal distribution range of migratory fish in the North Atlantic (Afonso et al., Reference Afonso, Porteiro, Fontes, Tempera, Morato, Cardigos and Santos2013).
ACKNOWLEDGEMENTS
The authors would like to thank all the dive operators, divers and photographers who kindly collaborated by providing photographs and videos. Special thanks to N. Sá for sharing his photographs, to M. Santos and R. Santos for providing the M. mobular records and to F. Porteiro for sharing old photographs and ideas. This work has been conducted in agreement with national laws and animal care protocols of the University of the Azores.
FINANCIAL SUPPORT
This research was partially supported by the Azores Regional Government (A.F.S., grant number 135104); the Portuguese Foundation for Science and Technology/Ministry of Education and Science (FCT/MCTES-MEC) (P.A., grant number Ciência 2008/POPH/QREN); and the LARSyS Strategic Project 2011–2012 (grant number PEst/OE/EEI/LA00009/2011).
