Description of a new Xenorhina species (Anura, Microhylidae) from northwestern Papua New Guinea

We describe a new species of the asterophryine microhylid genus Xenorhina from the lowlands of northwestern Papua New Guinea. It is a medium-sized species (SUL of two males 29.2 and 29.9 mm; of four females 29.9–33.0 mm) that can be distinguished from congeners by having a single short, triangular odontoid spike (palatal tooth) on each vomeropalatine bone, moderately short legs (TL/ SUL 0.40–0.44) and ventral surfaces heavily spotted with reddish-brown blotches or reticula. The advertisement call comprises 7–10 loud, melodious hooting notes lasting 141–165 ms and produced at a repetition rate of 2.19–2.35 notes/s. Description of this species brings to 41 the number of Xenorhina known from New Guinea and surrounding islands.


Introduction
The asterophryine microhylid genus Xenorhina Peters, 1863 currently contains 40 named species of generally squat, short-legged frogs with narrow snouts and small eyes, all of them confined to the New Guinea region (Zweifel 1972;Blum and Menzies 1989;Menzies 2006;Frost 2021;Günther and Richards 2021). Most members of the genus are fossorial or terrestrial (Menzies and Tyler 1977), but three species (X. arboricola Allison and Kraus, 2000, X. macrodisca Günther and Richards, 2005 and X. varia Günther and Richards, 2005) are mainly arboreal. Recent studies of Xenorhina have improved knowledge of the genus in western (Günther and Richards 2005;Günther and Knop 2006;Günther et al. 2020) and southern (Günther et al. 2009;Kraus 2011;Günther and Richards 2021) New Guinea. By contrast, knowledge of the Xenorhina fauna of northern Papua New Guinea remains relatively scant, the most recent taxonomic treatment of material from the region being the description of Xenorhina zweifeli (Kraus and Allison, 2002) nearly two decades ago.
During a study of the beta diversity of frogs across the extensively forested lowlands of the Sepik River catchment in northern Papua New Guinea (Dahl et al. 2009(Dahl et al. , 2013, one of the authors (CD) encountered an unnamed Xenorhina with a striking pattern of spots and reticulations Vertebrate Zoology 71, 2021, 621-630 | DOI 10.3897/vz.71.e66954 Copyright Rainer Günther et al.. This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. across the venter. It belongs to a group of Xenorhina that exhibits a single enlarged odontoid spike on each vomeropalatine bone but differs from all congeners by its bold ventral pattern, unique advertisement call and a suite of mensural characters. The new species is known from two locations approximately 150 km apart in the vast lowlands of the Sepik River catchment in northwestern Papua New Guinea, and its description brings to 41 the number of described species in the genus.

Material and methods
Male frogs were located at night by their advertisement calls; females were encountered while digging for calling males. Specimens were euthanized in an aqueous chlorobutanol solution (Gamble 2014) and subsequently fixed in 5 % formalin. All specimens were transferred to 70 % ethanol within two days of fixation. The following measurements were taken with a digital calliper (> 10 mm) or with a binocular dissecting microscope fitted with an ocular micrometer (< 10 mm) to the nearest 0.1 mm from preserved specimens: SUL -snout-urostyle length from tip of snout to posterior tip of urostyle bone; SUL is generally slightly shorter than snout-vent length (SVL). As the measurement error is higher in the latter, we prefer to use the former. Both measurements are sufficiently simi lar (unpublished data) that, where relevant, we compare our SUL measurements with SVLs presented for members of the genus in some papers; TL -tibia length: external distance between knee and tibio-tarsal articulation (referred to herein as 'shank'); some measurements of TL from the literature use a method introduced by Zweifel (1972) that took measurements from the fold of skin on the knee to the heel, producing shorter tibia lengths; we extrapolate these to standard measurements using known ratios of differences between the two techniques; TaLlength of tarsus: external distance between tibio-tarsal and tarsal-metatarsal joints held at right angles; T4Llength of 4th toe: from tip of toe to proximal end of inner metatarsal tubercle; T4D -transversal diameter of disc of 4th toe; T1D -transversal diameter of disc of first toe; F3L -length of 3rd finger: from tip of 3 rd finger to proxi mal edge of palm; F3D -transversal diameter of disc of 3rd finger; F1D -transversal diameter of disc of first finger; HL -head length, from tip of snout to posterior margin of tympanum; HW -maximum head width, taken in the region of the tympana; SL -snout length, from an imaginary line connecting the centres of the eyes to tip of the snout; EST -distance from anterior corner of orbital opening to tip of snout; END -distance from anterior corner of orbital opening to centre of naris; IND -internarial distance between centres of nares; ED -eye diameter, from anterior to posterior corner of orbital opening; TyD -horizontal diameter of tympanum.
Absence of clavicles and procoracoids was veryfied by CT scans. Sex was determined by observations of calling, presence of vocal slits and/or testes (males), or absence of vocal slits and/or presence of eggs (females). Advertisement calls were recorded with a Sony TCM-5000 Tape Recorder and a Sennheiser ME66 shotgun microphone and analysed with Avi soft-SAS Lab Pro software. Air temperature adjacent to the calling male was recorded using a rapid-reading digital thermometer. Terminology and acoustic analysis procedures mostly follow Köhler et al. (2017). Accordingly, and following Günther et al. (2020), we consider the multi-note vocalisations of this species to be 'call series' containing multiple single calls (= notes). However, we note that these call series could equally be considered a single call containing multiple single notes. Colour of the holotype in life was described from digital photographs, and of preserved specimens from direct observations. Most colours were determined according to a colour matching system created and administrated by the German RAL GmbH (https://en.wikipedia.org/wiki/RAL_colour_standard). When it was not possible to find an exact match between observed colours and RAL colour numbers, the most similar RAL number was chosen.
Measurements are presented as arithmetic means ± Standard Deviation. Specimens are deposited in the collections of the South Australian Museum, Adelaide, Australia (SAMA) and the Museum für Naturkunde, Berlin (ZMB). One paratype (SJR3914) will be repatriated to the Papua New Guinea National Museum (PNGNM). SJR is the field number of Stephen Richards. Specimens examined for comparative purposes are listed in Appendix 1. Additional morphometric and other data were extracted from original species descriptions and/or recompiled treatises, particularly Zweifel (1972), Blum and Menzies (1989), Kraus and Allison (2002) and Menzies (2006).

Systematics
Specimens were assigned to the genus Xenorhina on the basis of the following combination of features: jaw symphygnathine (maxillae meeting in front of the premaxillae); clavicles and procoracoids absent; a conspicuous spike present on each vomeropalatine bone; body squat; head small, triangular, with small eyes; life style subterrestrial.

Diagnosis.
A species of Xenorhina characterized by the unique combination of: medium size (SUL of two males 29.2-29.9 mm; of four females 29.9-33.0 mm); vomeropalatines each with one short triangular odontoid spike; legs moderately short (TL/SUL 0.40-0.44); all fingers and toe 1 without, and toes 2-5 with, expanded terminal discs; eye-naris distance greater than internarial distance (END/IND 1.10-1.21); tympanum about same size as eye (TyD/ED 0.95-1.16); dorsal surfaces in life different tones of brown with small blackish spots; ventral surfaces light ivory heavily spotted with reddish-brown blotches or reticula; advertisement calls uttered in series containing 7-10 loud hooting calls = notes each lasting 141-165 ms and produced at a rate of 2.19-2.35 calls/s. Description of the holotype. Adult male with vocal slits, calling when collected. Measurements are summarized in Table 1, a dorsolateral view in life is shown in Fig. 1a and ventral surfaces in life in Fig.1b. Head broader than long (HL/HW 0.81); snout acuminate from above and below, distinctly protruding in profile; vomeropalatines each with one short, triangular and acuminate odontoid spike; loreal region oblique, no canthus rostralis; nostrils near tip of snout, directed dorsolaterally, visible from above but not from below; eye-naris distance greater than internarial distance (END/IND 1.21); tympanum visible in life and preservative, its diameter slightly less than that of eye (TyD/ED 0.95); supratympanic fold weakly expressed, extending from behind eye to behind tympanum; fingers moderately short, not webbed; tips of all fingers with barely detectable circum-marginal grooves, tips not wider than penultimate phalanges; subarticular and metacarpal tubercles barely visible; relative lengths of fingers 3>4>2>1 (Fig. 1c); shank short (TL/SUL 0.42); all toe tips with circum-marginal grooves and, with exception of toe 1, tips wider than penultimate phalanges; toes not webbed, most subarticular tubercles and oval inner metatarsal tubercle moderately well defined; relative lengths of toes 4>3>5>2>1 (Fig. 1d). Body laterally with numerous irregularly shaped and irregularly arranged tubercles; dorsally posterior of head with four regularly spaced longitudinal tubercle rows (two paravertebral and two dorsolateral); dorsal surfaces of head, limbs, and all ventral surfaces without tubercles; tip of snout smooth.
In life dorsal surfaces of head, body and extremities mostly ochre-brown (RAL 8001); posterior back with extended daffodil yellow (RAL 1007) flecks; tubercles on dorsum mainly black-brown (RAL 8022) with light ivory (RAL 1015) apices; a black-brown stripe runs along supratympanic ridge. Ventral surfaces mostly light ivory (RAL 1015) with beige-grey (RAL 7006) reticulation and diffuse orangebrown (RAL 8023) spots on extremities and both sides of abdomen; throat orange-brown with black-brown spots. Iris predominantly blackish.
In preservative ground colour of dorsal surfaces of head, back and extremities beige (RAL 1001) with some inconspicuous brown-beige (RAL 1011) spots. Supratympanic ridge and cutaneous tubercles partly (especially on their bases) black-brown; rear of thighs predominantly fawn-brown (RAL 8007). Basic colour of ventral surfaces light ivory; flecks on chest, abdomen, and extremities beige-brown; throat light ivory with mahogany-brown Calls are unpulsed and the first call in most series is slightly longer than subsequent ones. Call intervals tend to become longer during the course of each series, and the last interval is clearly greater than preceding ones. Unlike many other Xenorhina species, neither volume (Fig.3, upper) nor pitch of calls (Fig. 3, lower) increase during the series. Calls start abruptly at maximum amplitude, and amplitude decreases rapidly at the very beginning and then gradually until end of call (Fig. 3, upper). Frequency decreases slightly at the very end of each call (Fig. 3,  lower), where this pattern is most evident in the upper harmonics). All calls have four welldefined harmonics, with maximum energy at 0.85 kHz, 1.70 kHz, 2.55 kHz and 3.40 kHz. The first harmonic with peak at 0.85 kHz is clearly the dominant one (Fig. 4). Distribution and ecological notes. Xenorhina ventrimaculata sp. nov. is known with certainty from two locations approximately 150 km apart in the lowlands of the Sepik River basin in northwestern Papua New Guinea (Fig. 5). An additional male specimen (SAMA R71744) from Wamangu Village, about 300 km northeast of the type locality (Fig. 5), closely resembles this species in its ventral colouration but is substantially larger than the two male types (39.5 mm vs. 29.2 and 29.9 mm SUL) and further differs in a number of body ratios. We refrain from including it in the type series and consider the taxonomic status of the Wamangu population to be uncertain pending the collection of additional information. The habitat at the type locality at Utai is secondary lowland forest, where both the holotype and paratype ZMB 91632 were calling from beneath leaf litter after rain at night. At Yapsiei this species was found in primary lowland forest, where all specimens were found between 1.5-3.0 cm beneath the soil surface when attempting to locate calling males. The local name for this species at Utai is Mopepe.
Etymology. The specific epithet is an adjective compound of two Latin words. Venter is a substantive and means belly or underside of the body and maculata is a feminine adjective meaning flecked or spotted. The specific epithet refers to the conspicuously spotted ventral surfaces of most specimens of the new species.
Comparisons with other species. Xenorhina includes a group of species with one or more distinct odontoid spikes on each vomeropalatine bone (formerly allocated to the genus Xenobatrachus) and another group lacking spikes on the vomeropalatines. Xenorhina ventrimaculata sp. nov. belongs to the former group and we compare it here only with other Xenorhina species of a similar size (25-38 mm SUL) that have a single odontoid spike on the vomeropalatines. Note that the terms call and note are used synonymously.

Discussion
The Sepik River in northern Papua New Guinea is the country's largest river system, and the river and its associated floodplains and lakes bisect a vast expanse of relatively uniform lowland tropical rainforest (Novotny et al. 2007). Few studies of amphibians have been conducted in these lowland forest habitats, and a large proportion of the fauna remains undescribed. For example, Austin et al. (2008) reported that 42% of 33 species encountered during an intensive survey at Utai on the western edge of the Sepik catchment were undescribed or unidentifiable and Dahl et al. (2013) reported a similar result (nearly 40%) for 44 frog species encountered at five sites across the catchment.
Our study adds to a growing effort to better document the frog fauna of northwestern Papua New Guinea and builds on important contributions to knowledge about patterns of species diversity and distributions (Kraus and Allison 2006;Austin et al. 2008;Dahl et al. 2009;Kraus 2010) and taxonomy Kraus and Allison 2000, 2001, 2002Richards 2007;Kraus 2013). Much of this effort has focused on frogs occupying the outlying north-coast ranges including the Bewani, Hunstein, Torricelli and Prince Alexander Mountains (e.g. Allison 2002, 2006;Kraus 2013a,b). These mountain ranges are derived from offshore island arcs that have accreted onto the New Guinea mainland over the past 20 million years (Davies et al. 1997;Davies 2012) and the discovery and descriptions of numerous new frog species in these mountains, many with limited known distributions (e.g. Allison and Kraus 2000;Kraus and Allison 2000, 2001, 2002, 2009Kraus 2013a,b), suggests that these accreted island arc systems have contributed to the generation of high levels of regional amphibian endemism.
The drivers of amphibian diversity in New Guinea's lowland forests are less well understood. Dahl et al. (2013) found that the Sepik River does not represent a major barrier to frog dispersal. In the absence of major topographical barriers most frog species documented in the lowlands and foothills north of New Guinea's central cordillera have broad distributions (Kraus and Allison 2006;Dahl et al. 2009Dahl et al. , 2013Kraus 2010). Xenorhina ventrimaculata sp. nov. appears to fit this pattern, with the two known sites being about 150 km apart. However, documenting the true extent of its distribution will require further survey effort in this underdocumented region of Papua New Guinea.