A new morphologically cryptic species of forest frog (genus Platymantis) from New Britain Island, Bismarck Archipelago

Transcription

1 Zootaxa : (2006) Copyright 2006 Magnolia Press ISSN (print edition) ZOOTAXA ISSN (online edition) A new morphologically cryptic species of forest frog (genus Platymantis) from New Britain Island, Bismarck Archipelago RAFE M. BROWN 1, STEPHEN J. RICHARDS 2, JEET SUKUMARAN 1 & JOHANNES FOUFOPOULOS 3 1 Department of Ecology and Evolutionary Biology and Natural History Museum and Biodiversity Research Center, University of Kansas, Dyche Hall, 1345 Jayhawk Blvd, Lawrence, KS , USA. (RMB): rafe@ku.edu; (JS): jeetsukumaran@frogweb.org 2 Vertebrates Department, South Australian Museum, North Terrace, Adelaide, S.A. 5000, Australia. Richards.Steve@saugov.sa.gov.au 3 School of Natural Resources and Environment, Dana Hall, 440 Church St. University of Michigan, Ann Arbor, MI , USA. jfoufop@umich.edu Abstract We describe a new species of forest frog in the genus Platymantis from New Britain Island, Bismark Archipelago, Papua New Guinea. The new species is a morphologically cryptic form that has masqueraded for almost four decades under the name P. schmidti (formerly P. papuensis schmidti, Brown & Tyler, 1968). The new species is microsympatric with the geographically widespread P. schmidti at two known localities. We diagnose the new species on the basis of its distinctive advertisement call and slight but consistent differences in body size and proportions. Calling males of the new species appear to prefer more elevated perches than do males of P. schmidti and the new species may exhibit a greater extent of sexual size dimorphism. Key words: forest frogs, wrinkled frogs, Platymantis, Bismarck Archipelago, SW Pacific, advertisement calls, cryptic species diversity Introduction Frogs of the genus Platymantis (family Ranidae) have two major centers of diversity: one in the Philippines (26 species; Brown et al. 1997, Alcala & Brown 1998, 1999), and another in the Solomon-Bismarck archipelagos (22 species; Brown 1952, 1997, Foufopoulos & Brown 2004, Brown et al. in press). Minor centers of diversity include 2 4 species endemic to eastern Indonesia (Menzies 1982a, 1982b, Edgar & Lilley 1993), six species in mainland New Guinea (Zweifel 1969, Allison 1996, Günther 1999), and possibly as many as seven species in the Admirality Archipelago (W. C. Brown pers. comm.; SJR, unpubl. data); additionally, there are two species in Fiji (Gorham 1965, Morrison 2003), and one in Palau (Crombie & Pregill 1999). Accepted by M. Vences: 11 Sept. 2006; published: 16 Oct

2 ZOOTAXA The islands of the Bismarck Archipelago (Fig. 1) support intermediate levels of species diversity: New Ireland has four species of Platymantis (Brown & Menzies 1979, Allison 1996, Allison & Kraus 2001) while New Britain has 11 species (Brown & Tyler 1968, Zweifel 1975, Foufopoulos & Brown 2004, Brown et al. in press). Of these 11 Platymantis species, nine are considered endemic to this island (Foufopoulos & Brown 2004, Foufopoulos & Richards in press, Brown et al. in press). FIGURE 1. Map of New Britain Island in relation to the Bismarck Archipelago (inset), showing type locality of P. adiastola new species (A: Wanui), the type locality of P. schmidti (B: Talasea, Willaumez Peninsula), the foothills of the Nakanai Mountains (C: Sulu and Silali Villages), and the East New Britain capital (star: Rabaul/Kokopo). Two groups of Platymantis from the SW Pacific can be readily distinguished on the basis of external morphology. The first group includes arboreal species with widely expanded terminal disks of fingers and toes, while the second group is more variable, consisting of primarily terrestrial or scansorial forms with non-expanded or only narrowly expanded terminal disks of the digits (Brown 1952, Gorham 1965, Brown et al. 1997). The arboreal group on New Britain includes four species (P. nexipus, P. macrosceles, P. mamusiorum, and P. n. sp. [Brown et al. in press]). The terrestrial group include seven species (P. akarithyma, P. boulengeri, P. gillardi, P. mimica, P. rhipiphalca, P. magna, and P. schmidti; Foufopoulos & Brown 1994, Foufopoulos & Richards in press). Finally, in addition to these two groups P. browni is a miniature terrestrial/scansorial species (Allison & Kraus 2001). Zweifel (1960) first commented on morphological differences between New Britain and New Guinea forms of the terrestrial forest frog P. papuensis. Brown and Tyler (1968) Magnolia Press BROWN ET AL.

3 later described P. papuensis schmidti (from New Britain) as a distinct subspecies and distinguished it from P. papuensis papuensis (from New Guinea) and P. p. weberi (from the Solomon Islands) on the basis of morphology. Later, Menzies (1982) confirmed the distinctiveness of P. schmidti with a comparison of the advertisement calls of frogs from the Bismarck archipelago (P. schmidti) and those from mainland Papua New Guinea (P. papuensis). He also described an additional subspecies of P. papuensis (P. p. occidentalis) from eastern Indonesia on the basis of body proportions (Menzies 1982b). At present, P. papuensis, P. weberi, and P. schmidti are all recognized as full species; P. p. occidentalis is still considered a subspecies of P. papuensis. In this paper we describe a new morphologically cryptic species that is distinguished from the widespread P. schmidti by its distinct advertisement call, slight size and body proportions difference (including a possible greater extent of sexual size dimorphism), and a tendency towards a more elevated calling perch height. ZOOTAXA Materials and methods We recorded morphometric data from fluid-preserved specimens (see Specimens Examined section). Sex was determined by gonadal inspection when possible, by prominent secondary sexual characteristics (large body size of females), or, in the cases of males, by direct observations of vocalizing behaviour. Standardized measurements were taken (to the nearest 0.1 mm) with digital calipers and we minimized inter-observer bias (Lee 1982, Hayek et al. 2001) by analyzing only data collected by RMB. Character definitions follow Zweifel (1960, 1969, 1975) unless otherwise noted (see also Foufopoulos & Brown 2004, Brown et al. in press). These include: snout vent length (SVL), head length (HL), snout length (SNL), interorbital distance at the midpoint of the orbits (IOD), eye diameter (ED), horizontal tympanic annulus diameter (TD), head width at the widest point (HW), forearm length (FAL), femur length from the cloaca to the outer surface of the flexed knee (FL), tibia length (TBL), tarsus length (TSL), pes length (PL), manus length from tip of third digit to the base of the outer metacarpal tubercle (ML), fourth toe length (Toe4L), first finger length (Fin1L), third finger length (Fin3L), first finger disk width (Fin1DW), third finger disk width (Fin3DW), fourth toe disk width (Toe4DW), and widths of penultimate phalanges of third finger (PpFin3) and the fourth toe (PpToe4). Other morphological characteristics considered include color pattern, dermal ornamentation, and size and shape of subdigital tubercles. We explored patterns in morphometric data (following qualitative visual confirmations of normality) with box-and-whisker plots and simple correlation analyses of individual variables plotted against SVL. We applied a principal component analysis (PCA) on logtransformed morphometric data to assess whether these continuous characters could form the basis of qualitatively detectable structure in the data and to explore the relative contributions of specific variables to group separation in multivariate space. These PLATYMANTIS 2006 Magnolia Press 47

4 ZOOTAXA multivariate analyses were conducted using males only due to a lack of sufficient sample sizes for females. We successively extracted components until they cumulatively accounted for 95% of the total morphological variation. All analyses were conducted with SPSS software (SPSS Inc. 2001). We conducted multivariate analyses of morphological dimensions in order to determine if continuously varying morphological characters could contribute to the diagnosis of the new species identified on the basis of advertisement call. The data were not only analyzed for taxon-based structure, but also for possible patterns corresponding to samples (i.e., distinguishing not only between nominal Platymantis schmidti and Platymantis new species per se, but also between Platymantis schmidti from the original type series, collected from Talasea, New Britain (in 1966), and putative Platymantis schmidti collected recently from Wanui, East New Britain (in 2000). Our prediction was that, if these characters were useful in differentiating between these two species, we would see two distinct groupings in the data structure: a Platymantis schmidti group, encompassing specimens from both the old as well as new collections of Platymantis schmidti, and a Platymantis new species group. Alternatively, if time of collection or specimen preservation conditions have significantly and consistently affected continuous morphometric characters, we would expect to see a data structure in which the new collections of Platymantis schmidti grouped with the Platymantis new species, distinct from the original collections of the Platymantis schmidti type series. As a third alternative, if continuous morphological characters were not very informative in distinguishing between these species, we would expect to see no clear pattern in the data corresponding to these samples. Advertisement calls were recorded with a Sony WM DC6 Professional Walkman with a Sony ECM-Z200 microphone. Calls were recorded at distances of approximately m and ambient temperatures were taken during recordings. All calls were recorded within a range of 3 C, so no temperature corrections of the data were undertaken. Calls were digitized with Soundedit (Macromedia 1995) and analyzed with Canary (Charif et al. 1996) software. We examined oscillograms (waveforms), audiospectrograms (sonograms), and results of the Fast Fourier Transformation (FFT; frequency spectrum) for a variety of temporal and spectral characters (Foufopoulos & Brown 2004, Brown et al. in press). Results Initial exploration of the data produced mixed results with respect to our predictions, as illustrated by the box-and-whisker plots (Fig. 2). Most characters exhibited properties consistent with the taxonomically-informative morphological character hypothesis, e.g., ED, TD, IOD, FA, Fin3DW. In these cases, there was a clear tendency for recent P Magnolia Press BROWN ET AL.

5 schmidti specimens to overlap the range of variation exhibited by the original P. schmidti types, and neither P. schmidti samples overlapped the morphological range or variation in the new species. However, in the case of SVL, the pattern was more consistent with the sampling effect hypothesis, with the recent P. schmidti collections overlapping the range of P. new species specimens. Furthermore, in other characters, such as HW, HL, SNL, TSL, PL, and ML the recent P. schmidti specimens appeared to be intermediate, overlapping the more extreme ranges of the original collection of P. schmidti on one end and P. new species on the other. These results suggest that the differences in size between these two species are not isometric across all characters, and thus while some morphometric characters may be useful in distinguishing Platymantis new species from Platymantis schmidti, others may not be very informative. ZOOTAXA FIGURE 2. Box and whisker plots (median = central black bar; boxes = 25 th 75 th quartiles; whiskers = maximum and minimum values excluding outliers) of variation in three samples (1, P. schmidti type specimens from Talasea; 2, P. schmidti recent collections from Wanui; 3, P. adiastola from Wanui) of selected morphometric variables: A, Snout vent length; B, Head width; C, Eye diameter; D, Tympanum diameter; E, Interorbital distance; F, ratio of Finger III terminal disk width to penultimate phalange width. We applied PCA to determine if continuous variables of morphology could form the basis of group structure in these data and to examine the relative contributions of individual characters to group separation (and taxon identification). Although seven components were necessary to account for 95% of the total variation, we present loading of only four components (together accounting for = 90% of the total variance) because individual components V VII each accounted for only a very small fraction (= 2%) of the total variance. Most of the loadings for PC I are large, positive, and relatively homogeneous (with the exception of IOD, TD, F3PPW, and T4PPW), indicating that PC I primarily is a size-based component. PC I clearly separates the paratypes of P. schmidti from the new species, but recent collections of P. schmidti fall intermediate to these two PLATYMANTIS 2006 Magnolia Press 49

6 ZOOTAXA groups at either ends of the size spectrum. PC II loads most heavily and negatively on ED, TD, and IOD, suggesting that dimensions of the head contribute most heavily to separation between recent collections of P. schmidti and the new species. In general, PCA does not fully justify the recognition of the new species but it does confirm our impression that primarily size-based morphological variation in our three samples identifies three weakly differentiated groups of specimens. These include the older paratypes of P. schmidti, recent collections of P. schmidti, and a larger, stout-bodied form. Together with information from advertisement call and differences in microhabitat preferences, we feel justified in describing the large bodied form as a new species, below. FIGURE 3. Orthogonal bivariate ordination of principal components I versus II. Darkened circles = P. schmidti paratypes (from Talasea); open circles = P. schmidti (from Wanui); triangles = type series of P. adiastola (Wanui). See Table 2 for component loadings. Taxonomic account Platymantis adiastola, new species (Figs 4 6) Holotype: SAMA R61906, adult male, collected by S. Richards at Wanui Camp, Wanui River Valley (near Wide Bay), 310 m above sea level ( S, E), New Britain Isl., East New Britain Province, Papua New Guinea, 20 March Magnolia Press BROWN ET AL.

7 ZOOTAXA FIGURE 4. Dorsal views of P. adiastola (A: SAMA R61908; B: SAMA R60257), P. schmidti from Wanui (C: SAMA R57015; D: SAMA R57016), and P. schmidti paratypes (E: SAMA 6858; F: SAMA 6862) from Talasea, Willameuz Peninsula. Scale bars = 2 cm. PLATYMANTIS 2006 Magnolia Press 51

8 ZOOTAXA FIGURE 5. Photographs in life of P. adiastola (A: paratype SAMA R61907) and P. schmidti (B: SAMA R57015) Magnolia Press BROWN ET AL.

9 Paratypes: SAMA R , R , same data as holotype except collected 13 March 2000; R , 15 March 2000; R60259, 18 March 2000; SAMA R61907, 19 March; SAMA R61908, UPNG , 20 March 2000; SAMA R61909, 22 March Etymology: The specific epithet is chosen from the Greek adiastolos, meaning confused, mixed up, or not separated, in reference to past confusion surrounding the taxonomic status of the new species. Diagnosis: Platymantis adiastola is distinguished from congeners by (1) body size ( for 11 males; 71.7 for one female), (2) narrowly expanded terminal disks of the fingers and toes, (3) skin of dorsum with scattered longitudinal tubercular ridges, including two long (4 6 mm) medially bowed suprascapular dermal ridges, (4) upper eyelids devoid of conical tubercles, (5) subarticular tubercles of fingers, toes, hands and feet prominent and raised, (6) digits lacking all vestiges of lateral flange and interdigital webbing, and (7) unique advertisement call. The new species is morphologically nearly identical to P. schmidti but differs from this species by slight differences in body size and proportions (Table 1), an apparent tendency towards a more elevated calling perch preference, and a possible greater extent of sexual size dimorphism (P. adiastola male/female SVL = 0.57; P. schmidti = 0.69). The main distinguishing feature between these two morphologically cryptic forms is the male advertisement call. Platymantis adiastola produces a call composed of a series of shrill, high frequency (mean peak dominant frequency khz), complex, tonal notes whereas P. schmidti calls with a simple train of twin-pulse lower frequency ( khz) notes lacking tonal elements. Platymantis adiastola has a lower mean note repetition rate (4.3 notes/s) than P. schmidti (12.8 notes/s) and produces fewer mean notes per call (mean = 16.6) than P. schmidti (mean = 23.9). Finally, the call of P. adiastola exhibits a progressively decreasing internote interval across the call, whereas P. schmidti has a uniform internote interval throughout the entire call. Description of holotype: A mature, undissected male specimen, in excellent condition; habitus moderately stocky; head distinct, slightly wider in dorsal aspect than body, length 42.6% of SVL; head length 114.2% of head width; snout short, its tip bluntly rounded, protruding slightly beyond lower jaw; snout posteroventrally slanted in lateral aspect; eyes protrude laterally beyond silhouette of head in dorsal aspect, and strongly beyond dorsal surface of head in lateral aspect; labial region flat and smooth, not flared, lips not swollen, not extending past eyes when viewed in dorsal aspect; interorbital region flat; eye diameter 143.4% interorbital distance; pupil horizontally ovoid; canthus rostralis slightly concave; loreal region gently sloping, slightly concave; eye diameter 73.0% of snout length, 126.6% of eye narial distance; nostrils slightly laterally protuberant; eye narial distance much longer than distance from nostril to tip of snout; internarial region flat; tympanum distinct, its diameter 56.9% of eye diameter; dorsal edge of tympanic annulus in contact with and slightly overlapped by supratympanic fold, the latter extending from dorsoposterior corner of eye, along dorsal edge of tympanum, and ZOOTAXA PLATYMANTIS 2006 Magnolia Press 53

10 ZOOTAXA TABLE 1. Morphometric variation (in mm) in adult Platymantis adiastola n. sp. and P. schmidti (original collections, holotype and paratypes, as well as recent collections). Table entries include mean ± 1 standard deviation (with range shown below), and sample size (n). See text for character abbreviations. Platymantis adiastola n. sp. Platymantis schmidti (original collections) Platymantis schmidti (recent collections) Holotype Male (n=10) Female (n=1) Holotype Male (n=8) Female (n=8) Male (n=7) SVL ± ± ± ± ED ± ± ± ± TD ± ± ± ± HL ± ± ± ± SNL ± ± ± ± IOD ± ± ± ± HW ± ± ± ± FA ± ± ± ± TBL ± ± ± ± TSL ± ± ± ± PL ± ± ± ± ML ± ± ± ± TL ± ± ± ± Fin1L ± ± ± ± Fin3L ± ± ± ± Fin3DW ± ± ± ± Fin3PPW ± ± ± ± Toe4DW ± ± ± ± Toe4PPW ± ± ± ± Magnolia Press BROWN ET AL.

11 TABLE 2. Loadings for the first four principal components extracted from the correlation matrix. All variables were log transformed. I II III IV SVL ED TD HL SNL IOD HW FA TBL TSL PL ML T4L F1L F3L F3DW F3PPW T4DW T4PPW Eigenvalues % of Variance Cumulative % ZOOTAXA terminating nearly at supraaxillary region; tongue subcircular, with deep posterior notch and narrow anterior attachment; choanae round, minute, at anterolateral edge of palate, widely separated by a distance eight to ten times greater than their diameter, not obscured by palatal shelf; dentigerous process of vomer oblong; vomerine teeth tiny, translucent, numbering four to six; dentigerous process very slightly angled anterolaterally, with closest (posterior) points separated by a distance two times the diameter of one choana, their most distant (anterior) ends separated by distance equal to three times diameter of choanae; openings to vocal sac short slits, slightly posterior to rictus; prominent postrictal tubercle cluster present at posteroventral edge of tympanum. PLATYMANTIS 2006 Magnolia Press 55

12 ZOOTAXA FIGURE 6. Ventral views of hands and feet of P. adiastola (A, B: SAMA R61908), a recent specimen of P. schmidti from Wanui (C, D: SAMA R57016), and a paratype of P. schmidti from Talasea (E, F: SAMA R6862). Scale bars = 2 mm. Skin of dorsal surfaces of body, head, and limbs generally smooth, except for fine longitudinal ridges of variable length on dorsum; a faint pair of medially bowed suprascapular dermal folds present; several other short glandular folds present along dorsolateral edge of trunk and in temporal region; ventral surfaces of trunk, head, throat, and limbs smooth; manus length 45.9% of foot length; digits of manus narrow in Magnolia Press BROWN ET AL.

13 appearance and round in cross-section; lateral edges of digits without fleshy dermal flanges; terminal disks moderately expanded and protuberant, with short circumarginal folds on distal tips of digits; minute supraarticlar flaps present above penultimate-ultimate phalangeal articulation; decreasing digital length III, I, II=IV; subarticular tubercles extremely prominent and protuberant, rounded on ventral surfaces; one subarticular tubercle on digits I II, two tubercles under digits III IV; supernumerary tubercles very prominent and rounded, present at the base of all digits; minute palmar tubercles preset basal to supernumary tubercles; thenar (inner metacarpal), medial palmar and outer metacarpal tubercle prominent, ovoid, highly convex, and nearly equal in size; slightly raised dermal tubercles present on ventral surfaces of wrists; nuptial pads absent, forearm musculature not hypertrophied. Hindlimbs moderately well developed; foot length 91.9% of tibia length; tibia length 60.3% of snout-vent length; tarsus smooth, lacking folds, flaps, or tubercles; digits of toes moderately expanded and ovoid; slight, narrow lateral dermal flanges present along distal phalanges; circummarginal grooves limited to distal ends of digits; supraarticular cutaneous folds slight; plantar surfaces of pes smooth, with prominently protuberant subarticular tubercles but lacking supernumerary tubercles; subarticular tubercles numbering three under Toe IV, two under Toes III and V, and one each under Toes I and II; decreasing digit length IV, III, V, II, I; outer metatarsal tubercle prominent and subcircular, pointed; inner metatarsal tubercle prominent and elongate, with a sharp venterolateral edge; digits of pes completely free of web; cloacal region smooth and devoid of supracloacaal tubercles or dermal flaps. Measurements of holotype: SVL 40.6; ED 5.6; TD 3.2; HL 17.3; SNL 7.7; IOD 3.9; HW 15.1; FA 8.6; TBL 24.5; TSL 13.5; PL 22.5; ML 10.3; Toe4L 15.0; Fin1L 5.9; Fin3L 7.2; Fin3DW 1.5; Fin3PPW 0.8; Toe4DW=1.5; Toe4PPW=0.6. Color of holotype in preservative: Dorsum homogeneous dark brown, blending laterally into light and dark brown marbled flanks with scattered enlarged distinct white spots; dorsal surfaces of head grayish; lateral surfaces of snout and head darker, nearly black with light brown tympanum and distinct white spots on lips; postrictal region heterogeneous, with black and brown blotches and white-tipped postrectal tubercles; dorsal surfaces of limbs brown with darker transverse bars (two on forearm, three on humerus, three on tibia, two on tarsus); dorsal surfaces of hands, feet, and digits dark brown with white bars on phalangeal articulations; inner two digits of hands and feet distinctly lighter than outer digital surfaces; venter pale cream; underside of throat speckled with brown, coalescing into nearly solid brown ventrolateral margins of throat; ventral surfaces of hands dark brown with prominent white subarticular tubercles; ventral surfaces of feet solid dark gray with light gray, velvety digital disks. Holotype color in life unrecorded. Variation: There are three basic color patterns evident in our sample of P. adiastola from the type locality. The majority (9 of 11 specimens) exhibit solid gray dorsal ZOOTAXA PLATYMANTIS 2006 Magnolia Press 57

14 ZOOTAXA coloration, with faint brown interoribital bars and only scattered flecks of black associated with dorsal tubercular ridges. This largely pigment-free pattern continues on to dorsal surfaces of limbs in a few specimens (SAMA R61907, 61909) but all remaining specimens have distinct transverse brown bars on limbs. The other two major color patterns include a specimen with a single, thick, vertebral yellow stripe (SAMA R61908) and one with two dorsolateral pale lines (SAMA R60259). Lateral surfaces of head with dark bar from tip of snout, through loreal region, encircling orbit, and following supratympanic fold posteroventrally to rictus of jaw. In most specimens (8/11), this dark coloration covers the majority of the lateral portions of the head. In three specimens (SAMA R61907, 61909, 60259), lateral portions of the head are distinctly lighter and darker pigmentation is limited to the canthus rostralis and supratympanic fold. Flank coloration varies from a simple dark (dorsally) to light (ventrally) fade (SAMA R60257, 60259) to a more distinct stratification between dark (above) and pale (below), with numerous moderately distinct pale round spots scattered across flanks. Ventral surfaces of the body are uniform pale cream with throat similarly colored or with a diffuse dark gray ventral surface of the entire head and throat (SAMA R61906, 61908, UPNG ). In these specimens, the chin and lower lips are distinctly lighter than surrounding throat. Palmar surfaces of the hands vary from very dark gray with white subarticular tubercles (e.g., UPNG 9974) to pale cream throughout their ventral surfaces (SAMA R60257). Plantar surfaces of the feet are uniform light gray (SAMA R60258) to dark gray with lighter subarticular tubercles (SAMA R61906, R61908, UPNG 9974). Color in life: Based on color images of paratype SAMA R61907 before preservation. Dorsal surfaces are khaki to dark tan or brown (e.g. Fig. 5). The dark canthal and supratympanic bars are dark brown, tympanum is light brown, the loreal region is dark brown with gold flecks and the labial region grayish pink with gold bars. The iris is gold above and dark below below pupil and postrictal tubercles are tan with white tips. Bars on limbs are varying shades of brown on a lighter tan or brown background. The dorsal surfaces of the digits are brown with white on spots on digital articulations and the inner two digits on hands and feet and bright white. The throat is white and the venter is white anteriorly, fading to cream or yellow ventrally by groin. Ventral surfaces of limbs are yellow or pink. Size dimorphism: The single available female specimen of P. adiastola has a snoutvent length of 71.7 mm which is considerably larger than all of the nine available P. schmidti specimens examined by us (Talasea population: SVL ; Table 1). This suggests to us the possibility that P. adiastola may have a greater extent of sexual size dimorphism than P. schmidti (P. adiastola male/female SVL = 0.57; P. schmidti = 0.69) but larger sample sizes will be required to address this question in the future. Ecology and natural history: Platymantis adiastola was observed, collected, and recorded at Wanui Camp in primary forest (two hours walking distance from Sampun Village; 05 o 'S, 152 o 'E), m above sea level. This site was sampled Magnolia Press BROWN ET AL.

15 for 14 days (14 28 March 2000). At Wanui the new species was subjectively characterized as 8 10 times more abundant than P. schmidti. In contrast, P. schmidti is clearly the most common form at lower elevations. Vocalizations confirming the sole presence of P. schmidti have been recorded or noted at numerous coastal localities, e.g. Kokopo and the vicinity of Rabaul, (Gazelle Peninsula), Sulu and Silali (foothills of the Nakanai Mountains) and, most significantly, at Talasea, the type locality of P. schmidti (SJR and JF pers. obs.). It was also abundant near Sampun Village (~ 50 m above sea level) and no P. adiastola were confirmed from this lowland site. Where the two species have been observed together, we have noted slight but clear microhabitat differences between the two forms (e.g. Foufopoulos & Richards in press): P. adiastola has been consistently observed calling from elevated perches (leaves and branches) as high as 2 m above the ground. In contrast, P. schmidti called predominantly from the ground, from low, herbaceous vegetation, or from stumps or fallen logs, consistently less than 1 m above the ground (Menzies 1982a, SJR and JF pers. obs.). Distribution: Confirmed only at two localities (Fig. 1): near Sulu and Silali Villages, foothills of the Nakanai Mountains, West New Britain (Foufopoulos & Brown 2004) and Wanui, East New Britain (Foufopoulos & Richards in press). Advertisement calls: The advertisement call of P. adiastola is markedly distinct from that of P. schmidti in both temporal and spectral characteristics. It is, however, similar in general call structure (a rapidly amplitude-modulated train of notes) and so its impression on the human ear is similar. Because the call of the new species is one of the primary characteristics distinguishing the new species, we describe its vocalizations and compare them to those of P. schmidti in some detail, below. The following descriptions are based on 4 vouchered recordings of P. adiastola (2 from the Nakanai Mountains; 2 from Wanui) and 11 of P. schmidti (9 from the Nakanai Mountains; 2 from Wanui). The call of P. adiastola is a series of shrill, high-frequency tonal notes, delivered in a train of variable total duration and note repetition rate. One very distinctive characteristic of the call is the progressive shortening of inter-note intervals throughout the call, such that calls start slowly and gradually shift towards a more rapid note delivery rate towards the end of the note train. The effect on the human ear is a sound akin to that of a marble being dropped and bouncing, faster and faster, until it comes to rest after a final burst of rapid strikes (Fig. 7A). In contrast, the calls of P. schmidti consist of a much more rapid, evenly-spaced train of dull, lower frequency, stridulating, vibrational notes, like the sound produced by slowly running a finger over the teeth of a comb or running a stick down a length of wood fence (Fig. 7B). Menzies stated that the call of P. schmidti lacks musical quality and merely sounds like a stick being rubbed against wooden railings (Menzies 1982a: 2). Individual male mean note repetition rate ([total number of notes 1]/time from beginning of first note to beginning of last) was (= 4.3 ± 0.4 SD; n = 4) notes/s for P. adiastola and (= 12.8 ± 3.7 SD; n = 9) notes/s for P. schmidti. ZOOTAXA PLATYMANTIS 2006 Magnolia Press 59

16 ZOOTAXA Calls of individual males of P. adiastola contained as few as 9 15 (mean 11.6 ± 3.6 SD) to as many as (17.1 ± 8.9 SD) notes per call. The average number of notes per call for our sample of four males was 16.6 ± 3.4 ( ; n = 5). Calls of individual male P. schmidti contained as few as 8 15 (mean 13.6 ± 5.2 SD) to as many as (45.2 ± 8.1 SD) notes per call. The average number of notes per call for our sample of 13 P. schmidti recordings was 23.9 ± 8.8 SD ( ). FIGURE 7. Typical 5.0 s oscillograms from advertisement calls of P. adiastola (A: SAMA R60258) and P. schmidti (B: SAMA R57014). Note decreasing internote interval across the call of P. adiastola (A). Individual note structure between the two species was quite different. The shrill, high frequency tonal notes of the call of P. adiastola have a relatively complex internal temporal and spectral structure (Fig 8). Each note contains three distinct frequency components (Fig. 8A, C) and a complex, amplitude modulated internal note structure with pronounced differences in call energy contained in the different amplitude modulated within-note pulses (Fig. 8B). The dull, lower frequency twin-pulse notes of the call of P. schmidti (Fig. 9) are quite different in that each note contains a tightly-spaced twin-pulse structure (Fig. 9B D), with pulses only slightly differentiated in amplitude (Fig. 9B, D) and both very similar to one another in spectral structure (Fig. 9C). Menzies referred to the call of P. schmidti as "little more than a succession of double clicks (Menzies 1982a: 2) Magnolia Press BROWN ET AL.

17 ZOOTAXA FIGURE 8. The advertisement call of P. adiastola (paratype SAMA R60258) recorded at ambient temperature of 25 C on 15 March 2000: a 150 ms power spectrum (A: Fast Fourier Transformation; relative amplitude vs. frequency in khz) and an expanded waveform (B: relative amplitude vs. time in ms) of a single note. A 1500 ms audiospectrogram (C: frequency in khz vs. time in ms) and oscillogram (D: relative amplitude vs. time in ms) of six typical notes from the call depicted in Fig. 7A. In P. adiastola calls, each note contains a clear fundamental frequency component at 1.0 to 2.2 khz (mean peak frequency = khz for four males), an emphasized (=dominant) frequency between 2.5 and 3.6 khz (mean peak frequency = khz), and then an additional frequency component between 4.2 and 5.0 khz (mean peak PLATYMANTIS 2006 Magnolia Press 61

18 ZOOTAXA frequency khz). The three frequency components are stereotyped, with variation in frequency between males present (and apparently varying inversely proportionately with body size; correlations not shown) but with little or no within-male or within-call frequency modulation or energy shifts. In P. schmidti, the emphasized (= dominant) frequency lies between 2.2 and 2.9 khz (peak frequency = khz for 11 males), and then minor frequency components include a fundamental between khz (= 1.5kHz) FIGURE 9. The advertisement call of P. schmidti (SAMA R57014) recorded at ambient temperature of 28 C on 13 March 2000: a 150 ms power spectrum (A: Fast Fourier Transformation; relative amplitude vs. frequency in khz) and an expanded waveform (B: relative amplitude vs. time in ms) of a single note. A 1500 ms audiospectrogram (C: frequency in khz vs. time in ms) and oscillogram (D: relative amplitude vs. time in ms) of 12 typical twin-pulse notes from the call depicted in Fig. 7B Magnolia Press BROWN ET AL.

19 and weak harmonics between 7.4 and 8.6 khz. Unlike the new species, P. schmidti males exhibit a shift in emphasized frequency across the duration of long calls. Thus, calls typically start with the first 6 10 notes at a lower dominant frequency ( khz) and males shift the calling energy of subsequent notes to higher dominant frequencies ( khz) towards the end of the call. Individual notes of P. adiastola begin with 2 4 internal amplitude modulated pulses of 5 10 ms (Fig. 8B), followed by the intense tonal note (Fig. 8B D) of ms. Average relative intensity (in db) of each of the preceding pulses is approximately 45 55% of the subsequent tonal portion of each note. Total note duration for individual males ranged from ms (= 44.8 ± 3.8; n=13 notes) to ms (= 50.8 ± 3.1; n=17 notes); mean note duration for our entire sample of recordings was 47.3 ± 4.4 ( ; n=4). Inter-note intervals for the same males ranged from ms to ms; mean inter-note interval for the entire sample was ± 26.3 SD ms ( ; n=4). Individual notes of P. schmidti are paired pulses (Fig. 8B D) of similar amplitude (Fig. 9D). Both pulses begin with a rapid rise to peak amplitude (2 5 ms), followed by a more gradual decline to ambient levels (Fig. 9B); in the second pulse, time for amplitude to fall to ambient levels is 2 3 times that of the first pulse. The average relative intensity (in db) of the first pulses is approximately 75 90% of the second pulse (Fig. 9B, D). Total note duration for individual males ranged from ms (= 45.8 ± 2.9; n=27 notes) to ms (= 56.1 ± 4.7; n=34 notes); mean note duration for our entire sample of recordings was 47.3 ± 4.4 SD ( ; n=11 males). Duration of the first pulse was consistently briefer (approximately ms) than that of the second (25 33 ms). Individual male interpulse intervals ranged from (= 12.4 ± 1.2 SD) to (20.6 ± 2.6 SD and internote intervals varied from (= 32.2 ± 4.6 SD) to (= 58.2 ± 8.3). Mean interpulse interval for the entire sample was 16.7 ± 5.4 SD ( ; n=11) and mean internote interval was 46.3 ± 12.9 SD ( ; n=11). ZOOTAXA Discussion The description of P. adiastola brings the number of species of the genus Platymantis from New Britain Island to 12. Our impression is that additional new species almost certainly await biologists if logistical obstacles to field work can be overcome and biodiversity specialists can utilize new kinds of data (DNA sequences, advertisement calls, ecological differences) to distinguish subtle but biologically meaningful differences in phenotypes of New Britain forest frogs. Foufopoulos & Brown (1994) have previously commented on unidentified vocalizations from high elevation forested sites in the Nakanai Mountains that may represent undescribed species. Brown et al. (in press) report on distinctive vocalizations of a potentially undescribed cryptic species currently assigned to P. nexipus. We are certain that with additional work in forested regions of New Britain, additional species discoveries will be forthcoming. PLATYMANTIS 2006 Magnolia Press 63

20 ZOOTAXA The combination of morphometric, ecological, and acoustic differences between the two species addressed in our study (P. schmidti and P. adiastola) convince us of the taxonomic distinctiveness of these taxa but leave one remaining issue to be addressed in future studies. We note that in some characters and in the PCA, the older paratypes of P. schmidti were nearly as statistically distinct from recent collections of P. schmidti from Wanui and the Nakanai Mountains as they were from P. adiastola. Some of this apparent statistical distinctiveness (Fig. 3; Table 1) is probably related to the circumstances of preservation and obvious state of desiccation of the older specimens (Fig. 4E F, 6E F). Alternatively the distinctiveness of older paratypes (from Talasea) from more recent (Wanui) specimens of P. schmidti could conceivably be interpreted as evidence for an additional cryptic species in the P. schmidti complex. Lacking recently collected specimens or recordings of calls of the Talasea population, we take no action at this time. At the present time, available evidence limits us to only the two following conclusions: (1) P. schmidti is the common widespread, predominantly lowland form observed by us and others (Menzies 1982a, 1982b) at many localities; (2) P. adiastola possesses an advertisement call that has not previously been reported and may be a more geographically restricted, less common species. We know that the two species are sympatric in at least two localities (Nakanai Mountains and Wanui). Supporting evidence for our interpretation comes from numerous observations and recordings of advertisement calls matching that of P. schmidti (Fig. 9) at many localities throughout New Britain (SJR and JF pers. obs.; see also Menzies 1982a, 1982b). Platymantis schmidti is a common species around low elevation coastal localities, including Talasea. One of us (JF) has traveled to the vicinity of Talasea and noted the typical P. schmidti call at this site. If future excursions to the type locality should uncover yet another unique advertisement call, and/or confirm the finding of a morphologically distinct form at Talasea, the name P. schmidti would have to be restrictively applied to the Talasea population. This might necessitate the naming of yet another species to accommodate the widespread form, here (and in Menzies 1982a) treated as P. schmidti. Whatever the case, we have no doubt that P. adiastola is distinct from the P. schmidti types, our own recent collections of P. schmidti, and all other populations considered to be P. schmidti by earlier workers (Brown & Tyler 1968, Menzies 1982a, 1982b). This convinces us that at least two morphologically similar but acoustically distinct species occur on New Britain. At present, given the available data, recognizing only two forms is the most reasonable and conservative course of action to take. Specimens examined Platymantis acrochorda. (n=15) Solomon Islands, North Solomons, Bougainville Isl., Bougainville Province, Kunua: MCZ-A (paratype); Asesi, S. of Kunua MCZ-A , (paratypes) Magnolia Press BROWN ET AL.

21 Platymantis aculeodactyla. (n=4) Solomon Islands, Bougainville Isl., Bougainville Province, Kunua: MCZ-A Platymantis akarithyma. (n=5) Papua New Guinea, Bismarck Archipelago, New Britain Island, West New Britain Province, S coast, ca 14 km NW Pomugu, Kandrian: CAS- SU (paratype); Moramora, 3 km N, 7 km E Hoskins: MCZ-A 88823; Pomugu, SAMA R7073 (holotype), R6982 (paratype); East New Britain, Malasait (04.464oS, oE), SAMA R7066, Platymantis boulengeri. (n=4) Papua New Guinea, Bismarck Archipelago, New Britain Island, West New Britain Province, ca 40 km S of Talasea, San Remo Plantation on Willaumez Peninsula: CAS-SU 22876; New Britain Archipelago : MCZ-A 1729, 9372; Moramora, 3 km N, 7 km E Hoskins: MCZ-A Platymantis browni. (n=10) New Ireland Island, Weitin River Valley, 8 km N, 7 km W of river mouth, River Camp ( S E), 150 m: BPBM 12090, 12099, 12102, 12104, 12106l, 12109, 12113, 12115, 12188, (paratypes). Platymantis gillardi. (n=14) Papua New Guinea, Bismarck Archipelago, New Britain Island, West New Britain Province, S coast, ca 7 mi NW Pomugu, Kandrian: CAS-SU ; Papua New Guinea, West New Britain Province, northern Nakanai Mountains, ridge between the Ivule and Sigole rivers on the northern edge of the Nakanai Plateau: UWZM , Platymantis guppyi. (n=38) Solomon Islands, Bougainville Isl., Bougainville Province, Camp Torokina: USNM ; Kunua: MCZ-A 38628, , 38635, , , 38668, 38674; Melilup: MCZ-A 38629, , 38667, , ; Mutahi: CAS Platymantis macrops. (n=4) Solomon Islands, North Solomons, Bougainville Isl., Bougainville Province, Kunua: MCZ-A (paratypes); Aresi, S. of Kunua: MCZ-A (holotype); Matsiogu: MCZ-A Platymantis magna. (n=3) Papua New Guinea, New Ireland Isl., New Ireland Province, W. Coast, approx. 88 km S Kavieng: CAS (Paratype); Utu, 1 km S, 5 km E Kavieng: MCZ-A (Paratypes). Platymantis mamusiorum. (4) Papua New Guinea, West New Britain Province, northern Nakanai Mountains, ridge between the Ivule and Sigole rivers on the northern edge of the Nakanai Plateau ( 'S, 'E): UWZM (holotype), UWZM 23719, 23722, UPNG 9992 (Paratypes). Platymantis mimica. (n=6) Papua New Guinea, Bismarck Archipelago, New Britain Island, West New Britain Province, ca 18 mi S of Talasea, Numundo Plantation on Willaumez Peninsula: CAS SU (paratype), SAMA R6868 (holotype), R7064 (paratype); Kandrian: SAMA R 7069 (paratype); Moramora, 3 km N, 7 km E Hoskins: MCZ-A 88826, Platymantis myersi. (n=5) Solomon Islands, Guadalcanal Isl., river E Popomaneseu track: MCZ-A ZOOTAXA PLATYMANTIS 2006 Magnolia Press 65

22 ZOOTAXA Platymantis n. sp. (n=4) Papua New Guinea, West New Britain Province, northern Nakanai Mountains, northern edge of the Nakanai plateau, on a ridge between the Ivule and Sigole rivers 1640 m elevation; UWZM , UPNG (holotype and three paratypes; Brown et al., in press) Platymantis neckeri. (n=43) Solomon Islands, Bougainville Isl., Bougainville Province: MCZ-A (paratypes); Bougainville Isl., Kunua: USNM ; Melilup: MCZ-A , 66849, 66849, ; Mutahi: MCZ-A , , , 66893; ; CAS Platymantis nexipus. (n=10) Papua New Guinea, West New Britain Province, New Britain Island, Nakanai Mountains, ridge between the Ivule and Sigole Rivers, m above sea level: UPNG , UWZM 23893, ; Papua New Guinea, East New Britain Province, New Britain Island, Wanui Camp, Wanui River Valley (near Wide Bay), 310 m above sea level ( S, E): SAMA ; East New Britain Province, Gazelle Peninsula, Baining Mountains, St. Paul's, m above sea level, BPBM 1009 (holotype). Platymantis papuensis. (n=12) Indonesia, Irian Jaya Province, Hollandia: CAS-SU: ; Lake Sentani: CAS-SU ; Indonesia, Irian Jaya Province, Madang, Naru Village: TNHC ; Indonesia, Irian Jaya Province, Madang, Baiteta cave: TNHC 51541, 51978, Platymantis parkeri. (n=10) Solomon Islands, North Solomons, Bougainville Isl., Bougainville Province, Kunua: MCZ-A (paratypes), (holotype). Platymantis rhipiphalca. (n=3) Papua New Guinea, Bismarck Archipelago, New Britain Island, West New Britain Province,ca 40 km S of Talasea, San Remo Plantation on Willaumez Peninsula: CAS-SU (paratype), SAMA R7078 (paratype); Pomugu: SAMA R7071 (holotype). Platymantis schmidti. (n=41) Papua New Guinea, Bismarck Archipelago, New Britain Island, East New Britain Province, Karat, Cherub Plantation: CAS ; Baining Mountain Range, Gazelle Peninsula: CAS-SU (paratypes); Talasea Plantation, Willaumez Peninsula: SAMA 6762, 6764, 6784, 6786, 6791, 6795, 6813, 7093, 7097 (paratypes), 7618 (holotype); San Remo, Willaumez Peninsula: 6858, 6862, 6912, 6923 (paratypes); L.A.E.S., Karavat, Gazella Penninsula (near Rabaul): SAMA 7147, 7099 (paratypes); Wanui, Wanui River Valley (near Wide Bay), 310 m above sea level ( S, E), New Britain Isl., East New Britain Province: SAMA R , ; Papua New Guinea, West New Britain Province, northern Nakanai Mountains, ridge between the Ivule and Sigole rivers on the northern edge of the Nakanai Plateau: UWZM ; 23782, Platymantis solomonis. (n=6) Solomon Islands, North Solomons, Bougainville Isl., Bougainville Province, Topanas: CAS ; Mutahi: CAS ; , Platymantis vitiana. (n=8) Fiji, Viti Levu Isls., Viwa Isl., Viwa Village, SW side of Magnolia Press BROWN ET AL.

23 island: CAS ; Ovalau Isl., 0.5 mi N of Navuloa Village: CAS Platymantis vitiensis. (n=13) Fiji, Viti Levu Isls., Viti Levu Isl., Savura Creek Rd., ca 1 km W of Savura Creek: CAS , , , , , ; Ovalau Isl., 10 km S, of Levuka, St. John s: CAS Platymantis weberi. (n=19) Solomon Islands, North Solomons, Bougainville Isl., Bougainville Province, Mutahi: CAS , , ; MCZ-A , ZOOTAXA Acknowledgements We thank the National Research Institute (Jim Robins), the PNG Department of Environment and Conservation, and the governments of West New Britain Province and East New Britain Province for issuing research, collection, and export permits necessary for this study. Support for fieldwork on New Britain was provided by Conservation International (CI), the University of Wisconsin Natural History Museum Council, the Quixote Foundation, and Fauna and Flora International. SJR is particularly grateful to Maureen Ewai and Gai Kula of CI for their tremendous support. Assistance on New Britain was provided by S. Seeto (Mahonia na Dari/The Nature Conservancy), P. Dam (European Union-IRECDP), M. and C. Benjamin (Walindi Resort), ENBSEK, and numerous individuals and institutions acknowledged in Foufopoulos & Brown (2004). For the loan of specimens, and assistance with museum collections, we thank M. Hutchinson and C. Kovach (SAMA), A. Dassow, P. Holahan (UWZM), J. Menzies, R. Singadan (UPNG; museum abbreviations follow Leviton et al. 1985). Support for RMB s work on this project was provided by the US National Science Foundation and SJR received support from the Violet Scott Estate, The Mark Mitchell Research Foundation, the South Australian Museum Board, and Conservation International. References Alcala, A.C. & Brown, W.C. (1998) Philippine amphibians: an illustrated field guide. Bookmark Inc., Makati City, Philippines, 116 pp. Alcala, A.C. & Brown, W.C. (1999) Philippine frogs of the genus Platymantis (Amphibia: Ranidae). Philippine Journal of Science, 128, Allison, A. (1996) Zoogeography of amphibians and reptiles of New Guinea and the Pacific region, In: Keast, A. & Miller, S. E. (Eds.), The Origin and Evolution of Pacific Island Biotas, New Guinea to Eastern Polynesia: Patterns and Processes. SPB Academic Publishing, Amsterdam, pp Allison, A. & Kraus, F. (2001) New species of Platymantis (Anura: Ranidae) from New Ireland. Copeia, 2000, Brown, R.M., Foufopoulos, J. & Richards, S.J. (in press) New species of Platymantis (Amphibia; Anura; Ranidae) from New Britain and redescription of the poorly known Platymantis nexipus. Copeia. PLATYMANTIS 2006 Magnolia Press 67