Diversity of benthic marine mollusks of the Strait of Magellan, Chile (Polyplacophora, Gastropoda, Bivalvia): a historical review of natural history

Abstract An increase in richness of benthic marine mollusks towards high latitudes has been described on the Pacific coast of Chile in recent decades. This considerable increase in diversity occurs specifically at the beginning of the Magellanic Biogeographic Province. Within this province lies the Strait of Magellan, considered the most important channel because it connects the South Pacific and Atlantic Oceans. These characteristics make it an interesting area for marine research; thus, the Strait of Magellan has historically been the area with the greatest research effort within the province. However, despite efforts there is no comprehensive and updated list of the diversity of mollusks within the Strait of Magellan up to now. This study consisted of a complete bibliographic review of all available literature that included samples of mollusks in the Strait of Magellan. More than 300 articles were reviewed, covering 200 years of scientific knowledge. There were 2579 records belonging to 412 taxa, of which 347 are valid species. Of the total valid species, 44 (~13%) are considered of doubtful presence in the Strait. This work increases the known richness of mollusks of the Strait of Magellan by 228%; it is also the first report that integrates all available diversity studies of the three most speciose classes of benthic mollusks (Gastropoda, Bivalvia and Polyplacophora) from the Strait of Magellan.


Introduction
It has been described that mollusks show an increase in diversity towards high latitudes in the Chilean southeastern Pacific coast (Valdovinos et al. 2003). This increase in mollusk richness occurs around 42°S, coinciding with the beginning of the Magellanic Biogeographic Province (Spalding et al. 2007). The Magellanic Province has been the focus of study of several scientific expeditions that contributed to the knowledge of marine mollusks. The first reports were made by King and Broderip (1832), d'Orbigny (1835-1846 and Philippi (1845). Other reports that contributed considerably to the knowledge of mollusks of the Magellanic Province were Smith (1881), Rochebrune and Mabille (1889), Strebel (1904Strebel ( , 1905aStrebel ( , b, 1906Strebel ( , 1907Strebel ( , 1908, Odhner (1926), Marcus (1959) and Soot-Ryen (1959). Carcelles and Williamson (1951) published the first checklist of species of marine mollusks of the Magellanic Province in the 1950s, defining the province from around 37°S in the Pacific coast and 43°S in the Atlantic coast, to 56°S. In their checklist 614 species were reported. Many taxonomic revisions of specific groups have been published (e.g., McLean 1984a;Castellanos 1988;Castellanos and Landoni 1988, 1989, Castellanos 1990, 1992aCastellanos and Landoni 1993a, b;Castellanos et al. 1993;Ponder and Worsfold 1994;Schrödl 1996), therefore the checklist of Carcelles and Williamson (1951) had to be updated, for species synonyms and newly found species. Linse (1999) presented a new checklist of mollusks of the Magellanic Province, defining the province from around 41°S in the Pacific and Atlantic coasts to 56°S. However, the classes Polyplacophora and Cephalopoda were excluded from this checklist, which included 397 species of mollusks.
One of the most important channels in the Magellanic Province is the Strait of Magellan, where most historical reports of mollusks are focused. This extensive channel connects the Pacific and Atlantic Oceans and is considered the most important one of the province. It is influenced by water masses of the Pacific, Atlantic and Southern Oceans, and it possess several geological characteristics derived from the last glaciation (Antezana 1999). For these reasons the Strait of Magellan offers unique characteristics for the study of biodiversity and related aspects of the biogeography of mollusks (Linse et al. 2006). Linse et al. (2006) presented the only report of mollusk richness in the Strait of Magellan, which contains 116 species. However, a list of species is not provided and only the classes Gastropoda and Bivalvia are included. Between the year 2000 and the present there have been several studies that have provided more information about the diversity of mollusks in the Strait of Magellan (e.g., Ríos et al. 2003;Ríos et al. 2005;Ríos et al. 2007;Thatje and Brown 2009;Rosenfeld et al. 2013;Rosenfeld et al. 2015), presenting new records of species. Several taxonomic revisions of specific groups have been published in recent years, where erroneous records, changes in nomenclature, synonymized species and descriptions of new species have been made (e.g., Sirenko 2006a; Zelaya and Geiger 2007;Aranzamendi et al. 2009;Zelaya 2009;González-Wevar et al. 2011;Güller et al. 2016;Pastorino 2016;Güller and Zelaya 2017;Korshunova et al. 2017). In order to have a comprehensive list of species in the most important channel of the Magellanic Province it is necessary to provide an updated list of records of the malacofauna of the Strait of Magellan. The objective of this study is to provide the first list of species of benthic marine mollusks of the three most speciose and best documented classes (Polyplacophora, Gastropoda, Bivalvia) of the Strait of Magellan, integrating all studies throughout history.

Materials and methods
To make the list of mollusks as complete as possible, information was gathered from all the available scientific publications that have sampled or reviewed benthic marine mollusks in the Magellanic Province, from the expedition of the HMS Beagle in the 19 th century (King and Broderip 1832) to the present. A total of 323 articles were reviewed, of which 146 contained species within the Magellanic Province. The records and their respective geographical positions were entered into a spreadsheet structured with the Darwin Core Standard (Wieczorek et al. 2012), adjusted taxonomically according to the MolluscaBase (2019) and the revisions of classification and systematics of gastropods (Bouchet et al. 2017), bivalves (Nevesskaja 2009) and polyplacophorans (Sirenko 2006b). The Strait of Magellan was divided into 420 quadrants of 6×6 minutes of latitude and longitude. The records located within this area were analyzed (Fig. 1), taking into account their georeference or approximate location. This analysis was developed using tools for Google Earth (http://www.earthpoint.us), which transforms XLS extension files (Excel format) to KML (files that contains geographic data). In total, 108 articles provided records for the Strait of Magellan.
Dubious records were counted as were species that were recorded only once in history. Criteria were followed to determine doubtful species records, as follows: species that were cited once and later questioned in taxonomic revisions or never reported again; species that greatly exceed their distribution limit and do not appear in taxonomic revisions or alpha diversity studies; and species that have a huge geographical discontinuity and are not explained or figured in the article.
A new matrix was elaborated with the Darwin Core standard from the database, with presence-absence data of each taxon per quadrant entered as 1 or 0, respectively. The quadrants with no species were removed from the matrix and species/ taxa considered doubtful and/or with imprecise locations were not included in the matrix. However, the above cases were considered in the quantification of total richness. On the other hand, the records up to or above genus level (registered as "indet." or "sp.") were not considered as valid species for both species richness values and estimation models, except for those in which the author commented that it could be a new species. Finally, to detect whether the historial sampling effort was able to estimate all the species of mollusks in the Strait of Magellan, the non-parametric species accumulation models Chao 2 and Jacknife 1 (Burnham and Overton 1978;Burnham and Overton 1979;Chao 1987;Colwell and Coddington 1994) were used to evaluate the sampling effort spatially and estimate the number of species expected theoretically in the Strait of Magellan. These methods require only presence-absence data; Chao 2 is calculated with the species that occur in only one sample (single or singleton species) and those that occur exactly in two samples (doubletons). Jacknife 1 is a more accurate and less biased estimator, since it only uses the number of singletons and the number of samples (Moreno 2001). Complementarily, parametric accumulation models were used to detect whether the historical sampling effort was able to estimate the total species of mollusks (Soberón and Llorente 1993); the linear dependence and Clench models were used. All samples were randomized so as not to affect the shape of the curve (Colwell and Coddington 1994;Moreno and Halffter 2000). The estimation of the coefficients of each nonlinear regression model was done using the Simplex and Quasi-Newton estimation methods of the statistical package STATISTICA 7. For all models, species with imprecise locations were not included.

Results
A total of 134 articles summarizing two centuries of study were entered in the spreadsheet, representing 2579 records corresponding to 412 taxa distributed in the three classes studied (Table 1, Appendix I). Of the total taxa, 65 were reported up to or above genus level (i.e., "indet." or "sp."), finding no evidence that they may correspond to new species. On the other hand, 44 species were considered doubtful. Of the total 303 validated species with effective distribution in the Strait of Magellan (Fig. 2), 57.1% belong to the class Gastropoda (173 species); 24.1% of these correspond to the most diverse families: Buccinidae, Muricidae, Calliostomatidae, Fissurellidae, Eatoniellidae, Nassariidae, Rissoidae and Naticidae. The family Buccinidae was the most diverse in the class, with 15 species. The class Bivalvia was represented by 35.3% of the species (107 in total); 17.5% of these correspond to the most diverse families: Mytilidae, Philobryidae, Lasaeidae, Mactridae, Veneridae, Cyamiidae, Neoleptonidae, Nuculidae and Thyasiridae. The remaining 7.6% correspond to the class Polyplacophora (23 species). The family Chitonidae was the most diverse, with 2.0% of the species. In total, 106 families were recorded.
There has been a constant increase since the decade of the 1980s in the number of studies (Fig. 3a) and records (Fig. 3b). The largest number of records in history were incorporated for the Strait of Magellan in the last decade (2007-2018) (Fig. 3b).
Of the 420 quadrants proposed, 163 presented species (Fig. 4, Appendix II). Ordering the matrix of absence and presence of species according to these quadrants, 1229 mollusk records were counted. The eastern microbasin had 35 quadrants with records, while the central microbasin had 104. The western microbasin proved to be the least historically sampled, with only 24 quadrants with records. The total richness of the Strait of Magellan was 303 species. However, 47 species had imprecise locations, as they were described as inhabitants of the Strait of Magellan, but the site of their habitat was not defined with geographical accuracy. These species include three polyplacophorans (Leptochiton sp., Notoplax magellanica and Hemiarthrum setulosum), 25 gastropods ( The quadrants that had species records cover ~37% of the total area of the Strait of Magellan; most of the studies are concentrated in the central microbasin. The quad- Table 1. Species checklist of benthic marine mollusks of the Strait of Magellan (Polyplacophora, Gastropoda and Bivalvia). Those species with a single record are marked with an asterisk (*) and those which are dubious with a square (▪). Their presence is indicated (+) in the eastern (E), central (C) and western (W) microbasins. References provided at the end of the list.  Rochebrune, 1884 bm, cq, ej + + + Tonicia smithi Leloup, 1980 b, cc, bu, bv, b, a, am, cc + Mopaliidae Nuttallochiton hyadesi▪ (Rochebrune, 1884) p + Nuttallochiton martiali (Rochebrune, 1884) b, cq, br, t, bv + + Plaxiphora aurata (Spalowsky, 1795) bu, bv, cq, bm, bo, e, j, am, a, br, ar, bk, ba, bl, t, i, b + + + Acanthochitonidae Notoplax magellanica * Thiele, 1909 am Hemiarthridae Hemiarthrum setulosum Carpenter in Dall, 1876 br, cc Class Gastropoda Gastropoda indet.

Discussion
According to Valdovinos (1999), the Chilean coast has about 959 species of the three most diverse classes of benthic marine mollusks (671 gastropods, 226 bivalves and 62 polyplacophorans), including Antarctic and oceanic island species. The Magellan Biogeographic Province (41°S to 56°S) is one of the geographical areas with the highest diversity of mollusks on the Chilean coast (Valdovinos et al. 2003). Taking into account this database, the 303 mollusk species recorded in this study correspond to ~31.6% of the species cited for the Chilean coast (Fig. 6). About 400 species of marine mollusks, 250 gastropods, 131 bivalves (Linse 1999) and 19 polyplacophorans (Sirenko 2006a) have been reported for the Magellan Province. Therefore, the 303 species recorded for the Strait of Magellan represent 75% of the mollusks reported for the MBP. However, comparing the value of richness found in this study (303 species) to the 116 species of gastropods and bivalves reported for the Strait of Magellan by Linse et al. (2006), plus 17 species of polyplacophorans by Sirenko (2006a), the richness of mollusks for the Strait of Magellan was increased by 228% (Fig. 6). Most of the records were reported in the last 70 years. However, records of the late 19 th century and early 20 th century  greatly increased the knowledge of the zone, surpassing previous reports (see Fig. 3). This is mainly due to the publications of Rochebrune and Mabille (1889) and Strebel (1904Strebel ( , 1905aStrebel ( , b, 1906Strebel ( , 1907 which reported 267 records in the Strait. The number of studies has increased in the last 40 years, and therefore the records (see Fig. 3). However, some of these records belong to reviews of biological collections and older studies.
One criterion was followed to determine doubtful species; those records that were cited in the past and have been questioned in taxonomic reviews. Species such as Carditella exulata or Pandora cistula were identified as dubious according to these criteria (Güller and Zelaya 2013;Güller and Zelaya 2016b). Other criteria included records in which the same taxonomist discussed the species described such as the case of Doris magellanica (Cunningham, 1871), records that considerably exceed their distribution limit and do not appear in taxonomic revisions or alpha diversity studies or are simply dismissed, such as Lottia orbignyi, Leptochiton smirnovi, Falsilunatia falklandica, etc. (Espoz et al. 2004;Pastorino 2005b;Sirenko 2016), and records that have a huge biogeographical discontinuity and are not explained or figured in the article, is the case of Ischnochiton striolatus, Puncturella noachina and Acteon delicatus (Rochebrune and Mabille 1889;Strebel 1907;Ramírez 2000). On the other hand, of the taxa reported up to or above genus level ("indet." or "sp."), only two could correspond to new species, according to the authors' remarks: Leptochiton sp. (Sirenko 2006a) and Crepipatella sp. (Nuñez et al. 2012).
Recent studies using molecular tools have observed that several species co-distributed in the Antarctic Peninsula and South America actually belong to different lineages, with evolutionary units separated by millions of years (Poulin et al. 2014). This has been mainly observed in species of the genus Aequiyoldia Soot- Ryen, 1951(González-Wevar et al. 2019). Finally, there are species in the list that do not qualify as doubtful, but which have been classified as unknown species due to their low number of records or due to its small body size, which makes it difficult to identify the species, with poor ecological or descriptive information (Castellanos 1979;Geiger 2012;Rosenfeld et al. 2017), e.g., Notoplax magellanica, Lissotesta impervia, Onoba sulcula, Onoba georgiana, Microglyphis curtula, Cylichna gelida, Turbonilla sanmatiensis, Philobrya atlantica. In this sense, it should be noted that much of the mollusk information that was collected in this work comes from manual collections and various types of sampling gears, trawl and grabs (e.g., Watson 1886; Rochebrune and Mabille 1889;Strebel 1907;Linse 2002;Ríos et al. 2003). However, taxonomic works on specific groups have allowed a good representation of unknown micromollusks (Ponder and Worsfold 1994;Geiger 2012;Pastorino 2016;Di Luca and Pastorino 2018). Despite the aforementioned contributions, micromollusks could continue to be underestimated, since the comparative morphology of various species is only beginning to be illustrated and described in detail (Di Luca and Pastorino 2018).
This historical compilation of the richness of benthic mollusks of the Strait of Magellan promotes the need and urgency for the management of coastal environments. Despite the historical sampling effort and about 192 years of records, the Strait of Magellan has a high diversity of mollusk species which is not yet fully known. The richness estimated by the parametric models was greater than that observed. Two reasons may explain this: i) the sampling effort along the Strait of Magellan has been low (only about 36% of the total area is recorded), and ii) there is still a lack of knowledge about the taxonomy of many mollusk groups, since many species remain undetermined and are not included in the listings or are not recognized in the field. According to Soberón and Llorente (1993), the probability of finding a new species in the Clench model will increase according to experience in the field. Therefore, the Clench model suggests increasing the sampling effort but at a broader spatial and temporal scale to reach the asymptote in the estimation of mollusk species from the Strait of Magellan.
The richness estimated by non-parametric models was higher than the observed. These non-parametric models work based on the number of unique (number of species that occur only in one sample) and duplicate (number of species that occur in exactly two samples). This is based on the assumption that individuals of a species do not live alone in ecosystems, but in populations (Magurran 1988), therefore many unique species in a sample may be indicating that a sufficient number of sampling units has not been used. This historical compilation showed that there are many places in the Strait of Magellan that only have one or two records, which was reflected in both estimators.
However, it is important to consider that in order to evaluate the behavior of the different estimators, it is necessary to know the number of species in the community (Walther and Moore 2005;González-Oreja et al. 2010). Unless the community has been thoroughly sampled, these curves may not work properly (Magurran 2004). Therefore, some authors recommend not working with only one estimator, but testing several models to see how they behave with the data (González-Oreja et al. 2010), since these may vary depending on the situation or for a specific group of organisms (Walther and Moore 2005). The results of the four models used in this study allows us to infer that greater sampling effort is needed in the Strait of Magellan, mainly because the largest number of records and species richness are concentrated at the same points within the Strait of Magellan, in the central microbasin.

Conclusion
This study provides a clearer idea of the diversity of mollusks in the Strait of Magellan, identifying erroneous records and those that need verification, encouraging other researchers to sample less-studied areas of the strait. This will update knowledge of the diversity of mollusks of the Strait of Magellan, contributing to Chile's biodiversity heritage and future studies of biogeographical models that are currently based on the 116 species of gastropods and bivalves cited by Linse et al. (2006) and the 17 species of polyplacophorans cited by Sirenko (2006a) for the Strait of Magellan. Finally, with this information of all the records, it will be possible to identify the hotspots of diversity for study and gaps in knowledge, among other things.