Shore fishes of the Marquesas Islands, an updated checklist with new records and new percentage of endemic species

Expedition Pakaihi I Te Moana was conducted in 2011 to the Marquesas Islands, lying between 07°50ʹ S and 10°35ʹ S latitude and 138°25ʹ W and 140°50ʹ W longitude. The expedition combined extensive collections and visual censuses of the shore fish fauna. A total of 74 species are added as new records for the Marquesas Islands; the coastal fish fauna of the Marquesas Islands is increased from 415 to 495 species and the number of endemic species is increased from 48 to 68 species. This increases the percentage of species-level endemism for the Marquesas Islands to 13.7%, ranking as the third highest region of endemism for coral reef fishes in the Indo-Pacific. Only two other peripheral regions, the Hawai’ian Islands and Easter Island, have higher values.


INTRODUCTION
Tropical reefs represent a high priority for conservation action among marine ecosystems (Roberts et al. 2002). They are known to host some of the most diverse communities in the world with nearly 6,500 species of coral reef fishes (Kulbicki et al. 2013), a diversity that reaches its maximum in the Indo-Malay-Philippine archipelago (Bellwood and Hughes 2001;Roberts et al. 2002;Mora et al. 2003;Reaka et al. 2008;Bellwood and Meyer 2009;Hubert et al. 2012), while peripheral areas of the Indo-Pacific basin host high percentages of endemism. Percentages of endemism change as our knowledge of the reef fish fauna improves and may also vary according to the sizes of the regions considered. The highest percentages of endemism for reef fish in the Indo-Pacific are: 1) Hawai'i with 25% endemism according to Randall (2007); 2) Easter Island with 21.7% (Randall and Cea 2011); 3) Red Sea third with 13.6% (Eschmeyer et al. 2010).
The census of biodiversity constitutes the primary basis for conservation efforts and the establishment of protection measures, as indicated for instance by IUCN red lists. Endemism is often perceived as an important characteristic in species conservation (e.g., Parravicini et al. 2014). Although extensive databases for coral reef fishes are being developed with more accurate taxonomy and geographical distributions (e.g., Randall 2005Randall , 2007Randall and Cea 2011;Kulbicki et al. 2013), our knowledge of the reef fish fauna still has major gaps due to the isolation of some remote regions or the technical difficulty of studying the reef fish fauna in certain regions of the world. The Marquesas Islands are a prime example. Located in Northeastern French Polynesia between 07°50ʹ S and 10°35ʹ S latitude and 138°25ʹ W and 140°50ʹ W longitude, they are geographically isolated. To the east, the closest islands are the Galapagos (5 300 km), to the north-west they are 2,200 km from the Line Islands and 3,500 km from Hawai'i. The closest island is in the Tuamotu archipelago some 500 km away. The South Equatorial Current, flowing between 04° N and 17° S from east to west (Wyrtki and Kilonsky 1984;Bonjean and Lagerloef 2002;Gaither et al. 2010), seems to constitute a hydrographical barrier to dispersal leading to the genetic differentiation of some Marquesan populations of otherwise widespread species (Planes and Fauvelot 2002;Winters et al. 2010; Motu Iti, Nuku-Hiva, Ua-Huka, Ua-Pou, Fatu-Huku, Hiva-Oa, Tahuata, Fatu-Hiva; Figure 1) during a threeweek expedition in 2011, aboard the M.V. Braveheart. A diversity of habitats was explored with shallow and deep air dives (down to 50-55 m) for a total of 54 sampled sites. Extensive collections and visual censuses were combined to establish the species composition of shore fishes of the Marquesas Islands. The complementarity of these sampling methods (Williams et al. 2006) allowed us to target different components of the ichthyofauna. Rotenone (powdered root of the Derris plant) allowed us to sample the cryptic and small fish fauna while spear guns and visual censuses allowed us to sample and record larger specimens of species not susceptible to rotenone collecting.
Fishes were identified using identification keys and taxonomic references (Randall 2005;Bacchet et al. 2006) and representative specimens of all species collected were photographed while they had their fresh coloration, sampled for tissues, labeled, and preserved as voucher specimens for the sequences made for a COI Barcode library. Voucher specimens were preserved in 10% formalin (3.7% formaldehyde solution) and later transferred into 75% ethanol. Preserved specimens were cataloged into the fish collection at the Museum Support Center, National Museum of Natural History, Smithsonian Institution, Suitland, Maryland, USA. Underwater visual censuses and underwater photographs allowed us to complete the list of new records. Nomenclature follows Randall (2005) and we followed recent taxonomic changes tracked using the Gaither et al. 2010). In addition to the isolation of the Marquesas, their environmental conditions are unique compared to the remainder of French Polynesia. The high islands are not surrounded by lagoons, and coral cover is minimal compared with other parts of French Polynesia. Sea temperatures are unusually variable (26-30°C) for a locality this close to the equator (Randall and Earle 2000) and upwelling of cold enriched waters leads to a general low coral cover and major production of phyto-and zooplankton (Martinez and Maamaatuaiahutapu 2004). All these features give to the archipelago its uniqueness not only within the French Polynesian landscape, but also among Pacific tropical reefs. Selection processes in such a contrasting environment have already been highlighted as a driver of speciation for a Marquesan endemic reef fish species (Gaither et al. 2015).
In 2000, Randall and Earle identified the archipelago as a major hotspot of endemism with 11.6% endemic fish species (Randall and Earle 2000) with only a relatively small portion of the islands having been explored. The present work is based on a compilation of all previous fish records in addition to a preliminary reef fish survey on one island (Mohotani) in 2008 and to the first reef fish survey, which explored all islands of the archipelago in 2011 (expedition Pakai I Te Moana -Nov. 2011).

MATERIALS AND METHODS
Sampling was carried out in 2008 in Mohotani and in 2011 for the first time at every island in the Marquesas (Clark Bank, Motu One, Hatutaa, Eiao,

DISCUSSION
Percentage of endemism is dependant on our level of knowledge of the locality in question but also of the surrounding islands and archipelagos. While remote places like Hawaii tend to lose percentage of endemism due to new records of widespread fishes in their waters, the Marquesas constitutes one of the rare places where both the number of widespread and endemic fishes have grown despite continuous explorations of islands around it; with 10% of endemism in 1976, 1978 (Randall 1976(Randall , 1998; 11.6% in 2000 (Randall and Earle 2000), 12.3% (8.3% identified endemics + 4% unidentified) in 2007 (Kulbicki 2007), while Williams et al. (2013) evoked 12.9% of endemism based on a preliminary analysis of the collection presented in this study.
Fifteen years ago, the Marquesas were reported to be one of the few hotspots of endemism in the Indo-Pacific for shore reef fishes (Randall and Earle 2000). This percentage is presently at its highest level ever with 13.7%. This census of the Marquesan fish biodiversity constitutes a baseline that is essential to ensure the future protection of these islands. A plan to set up a  Marine Protected Area is in progress for the Marquesas Islands. Our study highlights the uniqueness of the Marquesan reef fish fauna and emphasizes the necessity to preserve the reef fish fauna of the archipelago, possessing the third highest percentage of endemism in the Indo-Pacific.