Tree snails in the family Partulidae are widespread across the tropical Pacific, with endemic species occurring on most high islands. Partulid species have faced catastrophic range reductions and extinctions due primarily to introduced predators. Consequently, most extant species are threatened with imminent extinction. The U.S. administered Mariana Islands, consisting of Guam in the South and the Commonwealth of the Northern Mariana Islands (CNMI) in the north, historically harbored six endemic partulid species, half of which are thought to be extinct. While conducting a phylogenetic assessment of Partula gibba, an extant tree-snail with a range spanning at least seven islands within the archipelago, it was discovered that what has been identified as P. gibba on the island of Rota is a misidentified cryptic species. Here we use molecular phylogenetics, shell morphometrics and reproductive anatomy to describe it as a new species, Partula lutaensis sp. nov.. Because the new species has suffered population declines and has a restricted range, consisting solely of the small island of Rota, we highlight the urgent need for conservation measures.

Cryptic species, Mariana Islands, Micronesia, Mollusca, narrow-range endemic, systematics

The tree-snail family Partulidae is known from islands across the tropical Pacific (Cowie 1992). While most of the known diversity occurs in the eastern portion of the family’s range, particularly the Society Islands for the genus Partula and the Samoan Islands for the genus Samoana, the progenitors of both genera likely came from further west (Lee et al. 2014). The 15 islands of the Mariana Archipelago in the western Pacific (Fig. 1) historically harbored six described species. Two of them, Partula radiolata (Pfeiffer, 1846), and Partula salifana Crampton, 1925, were known only from the island of Guam, while Partula desolata Bauman & Kerr, 2013, described from sub-fossil shells, and Partula langfordi Kondo, 1970, were described as single-island endemics from Rota and Aguigan, respectively. The only two Mariana species with multi-island distributions are Samoana fragilis (Férussac, 1821), known from both Guam and Rota, and Partula gibba Férussac, 1821, known from seven islands, from Guam, in the south, to Pagan Island in the north. In a recent publication, we described the discovery of a cryptic species of Partula on the island of Rota, based on molecular evidence (Sischo and Hadfield 2017). Due to greatly similar shell shape, this species had been identified as Partula gibba in prior surveys and publications (Kondo 1970; Bauman 1996; Bauman and Kerr 2013; Hadfield 2015). Adding further taxonomic confusion, a colony of Partula once maintained at the Invertebrate Conservation Center, Zoological Society of London in London, was labeled Partula langfordi, although it was originally collected on Rota (Goodacre and Wade 2001; Sischo and Hadfield 2017). Because P. langfordi was described as endemic to the island of Aguigan, snails bearing that name in the London collection were likely not P. langfordi (Kondo 1970). Unfortunately, P. langfordi is now thought to be extinct (Smith 2013; J. Liske-Clark, Northern Mariana Department of Fish and Wildlife, personal communication). The name of Aguigan Island is variously spelled on different maps and in different resources, including: Agiguan, Agijuan and Aguijan. We apply here the spelling currently in use by CNMI bureaus, the NOAA and elected officials in the CNMI.

Figure 1. 

Map of the Mariana Archipelago with a star indicating the location of Rota Island.

To distinguish the new Rota Partula species from other extant species in the Mariana Archipelago, we paired our previously published phylogeny (Sischo and Hadfield 2017) with anatomy of the male reproductive tract, the latter having been used extensively as a diagnostic trait for Partula species (Pilsbry 1909; Pilsbry and Cooke 1934; Kondo 1955, 1968, 1970; Gerlach 2016; Slapcinsky and Kraus 2016). Because we were not able to extract useful DNA from preserved tissue of Partula langfordi, we were unable to carry out molecular phylogenetics with that species. In addition to anatomy of the male reproductive tract, we further distinguished P. langfordi from the species on Rota by replicating the shell morphometric analysis originally conducted by Kondo (1970) in his study of P. langfordi and co-occurring Partula gibba on the island of Aguigan. Here we describe Partula lutaensis sp. nov. and designate type material.

Due to the endangered status of all partulid species from Guam and the Commonwealth of the Northern Mariana Islands, we have elected to not include specific location information of any of the extant colonies or type specimen material in this paper. However, this information has been deposited in the Bernice Pauahi Bishop Museum with each specimen.

Shell morphological assessment

Kondo (1970) found no differences in conchology, aside from non-overlapping shell sizes, between P. langfordi and sympatric P. gibba on the island of Aguigan. Because Partula from Rota had been mistakenly identified as P. langfordi, we carried out a similar comparison. Shells of P. langfordi from the mollusc collection housed at the Bernice Pauahi Bishop Museum (BPBM) in Honolulu, Hawai‘i were used for analysis. BPBM lot numbers 213092, 21309, 213104, 213012, 213024, originally collected by Y. Kondo from Aguigan Island were compared with shells from what was then labeled as P. gibba (BPBM lot numbers, 217155, 213251, 213248, 213151, 213241) originally collected on Rota (Fig. 2). Lengths and widths were measured to 0.01 mm from 48 shells of P. langfordi and 47 shells of P. gibba with precision calipers. Shell length was measured parallel to the shell axis from the apex to the base of the aperture, and width was measured perpendicular to the shell axis across the widest portion of the shell. Means (M) and standard deviations (s.d.) are reported. Adult Partula spp. stop growing and become sexually mature when they form a characteristic thickened flare around the aperture of their shells, here referred to as a lip. All shells measured were lipped, indicating they were mature adults (Cowie 1992). Mean shell length and width were compared with independent-samples, two-sided t-tests, assuming unequal variances, using Microsoft Excel (version 16.44). Extensive shell metrics and comparisons for all partulid species of the Mariana Islands are found in historical publications (Crampton 1925; Kondo 1955, 1970). More recently Gerlach (2016) carried out extensive morphometric analyses of partulid shells, including P. lutaensis sp. nov., there referred to as Partula sp. (Rota). We therefore have made no other shell size comparisons. Lacking phylogenetic data for P. langfordi, our objective here was to replicate Kondo’s shell size comparison between P. langfordi and sympatric P. gibba to determine whether P. langfordi and P. lutaensis sp. nov. are similarly distinct.

Figure 2. 

A typical shell of P. langfordi from Aguigan (A) compared to light and dark shell morphs of P. lutaensis sp. nov. from Rota (B and C). The three shells are shown in abapertural view in the upper row and apertural view below.

The phylogeographic assessment of the extant partulids in the Mariana Islands reported by Sischo and Hadfield (2017) strongly supports the presence of a cryptic species on Rota and is concordant with further analyses of the male reproductive tracts, with P. lutaensis sp. nov. and P. radiolata sharing a lower attachment of the vas deferens in relation to the retractor muscle. Because this cryptic species was not found on any other islands, we conclude it is endemic to the island of Rota and have given it the name Partula lutaensis sp. nov. to recognize the indigenous Chamorro name for the island.

Available data indicate that all known populations of the genus Partula on Rota are P. lutaensis sp. nov.. This does not rule out the possibility that P. gibba once was, or currently is, on the island. Further surveys for extant partulid populations and analysis of sub-fossil shell remains on Rota may provide further evidence as to the present and historical distribution of these two species on the island. Moving forward, we strongly encourage that DNA barcoding be employed to determine species identification of any new living populations of Partula spp. discovered on Rota and elsewhere in the Mariana Islands. Furthermore, should P. gibba be located on Rota it should be attempted to find shell or body characters that might aid in distinguishing the two species without dissection or tissue sample collection.

Partula lutaensis sp. nov. was observed in locally high abundance on Rota, similar to observations of P. radiolata on Guam (Sischo and Hadfield 2017). Unfortunately, P. gibba, once the most abundant partulid on Guam, is now almost entirely extirpated (Hopper and Smith 1992; Sischo and Hadfield 2017; C. Fiedler personal communication June 2018). Persistence despite depredation by introduced predators may be further evidence of the shared ancestry between P. lutaensis sp. nov. and P. radiolata. Possibly, the P. lutaensis – P. radiolata clade shares behavioral and or life-history traits that have allowed the species to persist despite significant threats. For example, a recent study found that P. radiolata has a higher reproductive rate than P. gibba (Bick et al. 2018).

Across the Pacific, partulid species have been driven to extinction by introduced predators, most notably North American carnivorous snail species in the genus Euglandina, and the New Guinea flatworm Platydemus manokwari (Clarke, Murray and Johnson 1984; Murray et al. 1988; Hopper and Smith 1992; Bauman 1996; Coote et al. 1999; Cowie and Cook 2001; Régnier, Fontaine and Bouchet 2009; Pelep and Hadfield 2011; Meyer et al. 2017; Sischo and Hadfield 2017; Hadfield 2020; Gerlach et al. 2020). This has been particularly true in the Mariana Islands where half of the described partulid species are thought to be extinct, and the remaining species are imperiled across their ranges. While we observed P. lutaensis sp. nov. in locally high numbers in 2010, known populations are few and geographically discrete. Sub-fossil shells of partulids are ubiquitous on Rota (Bauman 1996) and Saipan (personal observations, October 2014), suggesting severe range reductions (Bauman 1996). Additionally, we observed Platydemus manokwari depredating P. lutaensis sp. nov. while collecting tissue samples on Rota (Fig. 8). Because P. lutaensis sp. nov. is an island-endemic species with a very restricted range and is clearly under significant predation pressure from introduced species, its existence is imperiled. When the Rota populations were considered to be Partula gibba, they were protected by a federal Endangered Species declaration. For these reasons, it is imperative that Partula lutaensis sp. nov. be listed as Endangered as soon as possible.

Figure 8. 

A Partula lutaensis sp. nov. freshly depredated by Platydemus manokwari observed by authors on Rota.

This project was supported by the US Fish and Wildlife Service and by the University of Hawai‘i at Mānoa. We would like to thank Ms Regina Kawamoto and Dr Norine Yeung with the Bishop Museum of Honolulu for providing access to the mollusc collection and for allowing us to borrow and dissect historically collected specimens for anatomical analyses. We would like to thank Paul Pearce-Kelly, Angus Davison, Chris Wade, Justin Gerlach, and Barry Smith for helping us sort out the mystery of a possible captive colony of P. langfordi in England and for mailing us specimens. Additionally, we would like to thank Jill Liske-Clark of the Commonwealth of the Northern Mariana Islands – Division of Fish and Wildlife for her support in our efforts and for collecting voucher specimens and additional tissue samples from Rota.