Search

Species distributions are influenced by processes occurring at multiple spatial scales. It is therefore insufficient to model species distribution at a single geographic scale, as this does not provide the necessary understanding of determining factors. Instead, multiple approaches are needed, each differing in spatial extent, grain, and research objective. Here, we present the first attempt to model continent-wide great ape density distribution. We used site-level estimates of African great ape abundance to (1) identify socioeconomic and environmental factors that drive densities at the continental scale, and (2) predict range-wide great ape density. We collated great ape abundance estimates from 156 sites and defined 134 pseudo-absence sites to represent additional absence locations. The latter were based on locations of unsuitable environmental conditions for great apes, and on existing literature. We compiled seven socioeconomic and environmental covariate layers and fitted a generalized linear model to investigate their influence on great ape abundance. We used an Akaike-weighted average of full and subset models to predict the range-wide density distribution of African great apes for the year 2015. Great ape densities were lowest where there were high Human Footprint and Gross Domestic Product values; the highest predicted densities were in Central Africa, and the lowest in West Africa. Only 10.7% of the total predicted population was found in the International Union for Conservation of Nature Category I and II protected areas. For 16 out of 20 countries, our estimated abundances were largely in line with those from previous studies. For four countries, Central African Republic, Democratic Republic of the Congo, Liberia, and South Sudan, the estimated populations were excessively high. We propose further improvements to the model to overcome survey and predictor data limitations, which would enable a temporally dynamic approach for monitoring great apes across their range based on key indicators.

Numerous protected areas (PAs) have been created in Africa to safeguard wildlife and other natural resources. However, significant threats from anthropogenic activities and decline of wildlife populations persist, while conservation efforts in most PAs are still minimal. We assessed the impact level of the most common threats to wildlife within PAs in tropical Africa and the relationship of conservation activities with threat impact level. We collated data on 98 PAs with tropical forest cover from 15 countries across West, Central and East Africa. For this, we assembled information about local threats as well as conservation activities from published and unpublished literature, and questionnaires sent to long-term field workers. We constructed general linear models to test the significance of specific conservation activities in relation to the threat impact level. Subsistence and commercial hunting were identified as the most common direct threats to wildlife and found to be most prevalent in West and Central Africa. Agriculture and logging represented the most common indirect threats, and were most prevalent in West Africa. We found that the long-term presence of conservation activities (such as law enforcement, research and tourism) was associated with lower threat impact levels. Our results highlight deficiencies in the management effectiveness of several PAs across tropical Africa, and conclude that PA management should invest more into conservation activities with long-term duration.

Atheris matildae Menegon, Davenport & Howell sp. nov. (Fig. 1 –3) Holotype. Adult male, MTSN 9344, collected in a forest fragment in Southern Tanzania, at about 1995 m by Omari Kibure and Obadia Mwaipungu in February 2009; fixed in 70 % EtOH, tissue fixed in 90 % EtOH. Paratypes. 2 adult males, MTSN 9399 and MTSN 9418 and an immature MTSN 9417 collected in February 2011 at the same locality as the holotype, by Michele Menegon, Tim Davenport and Sophy Machaga. Additional material. 10 specimens collected at the type locality between March and April 2011 and being kept alive for conservation purposes. 4 are males, 5 are females and 2 are immatures. Among them there is the adult female individual shown in Fig. 3. Type locality. Remote fragmented montane forest in Tanzania's Southern Highlands. Precise locality withheld until conservation insurance population secure. Additional information on the locality can be obtained for scientific purposes from the authors upon request. (www.atherismatildae.org) Diagnosis. Atheris matildae sp. nov. is distinguished from all other members of the genus except A. ceratophora by the presence of two to three very enlarged erect, hornlike, supraciliary scales. It is distinguished from A. ceratophora by the combination of the following morphological and molecular features, based on the data from 69 specimens from all over the known range of A. ceratophora: (1) its larger size, TL of A. matildae type is 643mm (the biggest male A. ceratophora ever recorded does not exceed 510mm TL), (2) higher count of maximum transverse head scales (max. 20 in A. ceratophora, 28 in A. matildae), (3) four subequal suprarostral scales in A. matildae, the two central ones of the same size and the outer ones double in size in A. ceratophora, (4) marked difference in dorsal scale microdermatoglyphic pattern (irregular smooth surface in A. ceratophora, presence of papillae-like ridges in A. matildae), (5) in A. matildae, an extensive black marking across the frontal part of the mouth, including part of nasal, rostral, mental and few infralabial scales is often present, the above described colour feature has not been recorded in the examined A. ceratophora specimens and photographs. Genetic divergence of mitochondrial gene cytochrome b between A. ceratophora collected at type locality and A. matildae expressed as actual substitution difference is 3.18 % based on uncorrected p-distance of 0.03180. Paratypes and additional material variation. Details and meristics for the type series are summarized in Table 1. A total of 13 specimens have been observed, three of them are paratypes. The most significant differences between holotype and paratype specimens are in body colouration (see Fig. 2). The young specimen MTSN 9417 tends to be more greenish, the zigzag ornamentation is more conspicuous and the top of the head is marbled in green/yellow. The two adult individuals, both males, are similar in colouration to holotype, with a back dorsum and a bright yellow zigzag dorsolateral pattern. Ten additional specimens have been recently collected and are being kept alive for conservation purposes. 5 are males, 5 are females, and 2 are young. A black patch around nasal, rostral, mental and first infralabials is present in most of the observed males but also in few females and immature individuals. Males in general tend to be darker with belly suffused with black. Adult females tend to be more yellow, in some cases with immaculate throat and belly; horn-like scales are yellow with black outer edges. Side of the head can be completely yellow or with black patches at the tip of the scales. In preservation the specimen retains the original colouration (see Fig. 1). FIGURE 2. A. matildae paratypes showing body shape, colouration and head details. Top - down: MTSN 9399; MTSN 9418 and MTSN 9417. FIGURE 3. Illustration showing the variation in colour of the rostral and chin area, The black patches are present in many individuals of both sexes. A. matildae A. matildae A. matildae A. matildae A. ceratophora Specimen number MTSN 9344 MTSN 9417 MTSN 9418 MTSN 9399 n = 53 Description of holotype. Adult male preserved in 70 % EtOH. Snout-vent length (SVL) 540.7 mm, tail 96.0 mm, rostral width 2.9 mm, rostral height 0.8 mm; eye diameter (vertical) 3.2 mm; snout to eye 3.3 mm. A heavybodied forest viper, sub-quadrangular in cross-section, with a rather thick prehensile tail (SVL/Tail approximately 5.7 times); head pear-shaped, with a very distinct neck, rounded snout and swollen supraorbital region that does bear two/three elongated, horn-like scales; eyes relatively large, laterally placed, and with a horizontal diameter approximately 3 / 4 of the snout length. Crown of head covered in small scales, slightly larger over the temporal region (maximum transverse head scales— 28); they bear a prominent keel and become mucronate over the head; the rostral is flattened, rectangular, about 3.5 times broader than high, contacting first supralabials and four small; unkeeled, roundish, subequal suprarostrals, nasal wider than high, with raised, embossed posterior edge, nostril circular and approximately in the centre of the nasal; internasals 5, all strongly keeled; interorbitals 9, keeled; circumoculars 16 – 16, not keeled but terminating in black blunt knobs; 1 row of suboculars present; circumoculars separated from nasals by two to three rows of feebly knobbed scales; a row of three irregular scales, bordering supralabials between nasal and lower circumoculars; supralabials 10 – 10, the first three smaller, and 6–8 with a swollen lower edge; infralabials 10 – 10, posteriormost with swollen upper edge and first in contact at the midline behind the mental; mental triangular, approximately twice as wide as deep; gulars bordering chin shields feeblykeeled, but prominently keeled towards the rictus; 2 preventrals, first largest; 150 ventrals; 49 undivided subcaudals (including spine); anal entire; 25 rows dorsal scales anteriorly, 26 rows at midbody, 19 rows posteriorly. FIGURE 4. Dorsal scale microdermatoglyphic pattern of A. matildae (A–B) and A. ceratophora: (C–D). Note in B the papillaelike ridges covering the keratin layers on the scale of A. matildae. Hemipenes. Both hemipenes are only partially everted. They resemble the A. ceratophora one as described by Emmrich (1997). The sulcus is bifurcate on a typically divided organ. The extreme basal area is naked, followed by an area with enlarged proximal spines, most prominently on the outer side of each lobe. Towards the apex, on the inner side of the sulcus, there are a few smaller scattered spines, while the distal area seems to be characterized by the lack of clearly differentiated ornamentation. A more detailed description of hemipenial morphology will be possible when a fully everted hemipenis becomes available for examination. FIGURE 6. The forest fragment in the Southern Highlands of Tanzania where A. matildae was collected, and a detail of the forest canopy. Colouration. Dorsally it appears as a black snake with bright yellow dorso-lateral zig-zag lines. Flanks are marbled in yellow. Dorsum of the head is almost entirely black with scattered yellow scales or groups of scales, sides of the head are mainly yellow with an irregular longitudinal black marking. An extensive black marking across the frontal part of the mouth, including part of nasal, rostral, mental and few infralabial scales is present, by contrast it delimitates an inverted pale triangle. Eyes are light olive green (in life). Throat is yellow; belly is pale yellow to greyish-green, suffused by black speckling; horn-like scales are yellow with black outer edges. In preservation the specimen retains the original colouration (see Fig. 1). Dorsal scale microdermatoglyphics. The surface microstructure of several scales from midbody and the last third of the body of two specimens of A. matildae (MTSN 9344, 9417) and three specimens of A. ceratophora from Usambara Mountains (MTSN 5117, 5118, 5121) were examined by scanning electron microscopy, in order to evaluate the intra- and inter-specific differences. Two-dimensional classes of microdermatoglyphics were identified; the coarser one consists of juxtaposed or imbricated layers of keratin with a raised edge, forming a ‘scaly background’. This layer is shared by both A. ceratophora and A. matildae. At greater magnification (4000 x) a further pattern of microdermatoglyphics is visible in A. matildae, where papillae-like ridges cover the entire surface of the keratin layer. The latter ornamentation is absent in A. ceratophora specimens from type locality. Distribution and conservation. Atheris matildae is currently known only from the type series and a few other individuals of both sexes, all collected in a remote montane forest fragment in the Southern Highlands. The site probably represents the remnants of a wider forested landscape, interspersed with plateau grasslands and possibly naturally isolated from other Southern Highland forest blocks. For this reason the forests are of great bio- logical value and now the focus of further exploration and conservation intervention. During the last decade the Southern Highlands have been the subject of extensive biological investigation by the Wildlife Conservation Society. However, this species has not been detected in any other areas. It is therefore probable that A. matildae is a rangerestricted forest species, now relying on just a few forest fragments. A. matildae has an extent of occurrence smaller than 100 km 2 with extent of occurrence, area of occupancy and quality of habitat in continuing decline. According to IUCN guidelines (IUCN 2010) therefore, we propose to list A. matildae as ‘Critically Endangered’ CR B 1 b(i,ii,iii). Further investigations are being carried out in order to collect more information on this magnificent snake, and a small breeding programme has been established (see www.atherismatildae.org). Etymology. Atheris matildae is named for TRBD's daughter Matilda Davenport, one of the next generation of herpetologists. We suggest the common name 'Matilda's Horned Viper'.

Pseudophryne vivipara Diagnosis: tympanum visible at least in adults, tip of forth toe exceeds tip of snout, fifth toe shorter than third, viviparous Description: snout pointed; canthus rostralis distinct; loreal region slightly concave, snout slightly longer than eye diameter; nares positioned below canthus rostralis, directed a bit caudally; interorbital space as wide as upper eyelid; pupil horizontal; tympanum present but not very distinct, small about 1 / 3 of eye diameter, coloured dark brown; head comparatively large, markedly separated from body; finger considerably long, widely splayed from each other and in life certainly not laying together; first finger much shorter than second, second only half the length of third finger, fourths finger shorter than third bit longer than second; skin below fingers without prominent bulges, like swollen; first toe much shorter than second, second much shorter than third, third shorter than fourth, fifths shorter than forth and third; below toes and on sole of foot the tubercles are only almost plain, roundish bulges, as are the outer and inner metatarsal tubercle; when hind legs pulled forward, hock already extending snout tip, toes extending snout by far; skin smooth; parietal glands remarkably large but not very prominent; no breast fold that would reach from one armpit to the other; as juveniles uniform blackish brown; they fade much later on; first the belly gets branching clear yellow spots, which extend in size in a way that finally a white-yellowish venter is covered with numerous blackish brown points; dorsal parts of back and legs fade as well to a colour that is clear chocolate brown; finally the animals have dorsal surfaces that are irregularly blackish brown and clear chocolate brown. The species is viviparous. The lower parts of both oviducts are enlarged; internal walls smooth without diaphragms or protrusions where eggs or larvae could lie or adhere, at the lowest ends both oviducts fuse to one large cavity. In both oviducts a large number of mature frog-larvae lay loosely and independently from each other. In an especially large female the right oviduct had 37, the left oviduct 30, a total of 67 larvae. Accordingly a younger specimen carried less. The young have no ���amnion��� and apparently are far developed; i.e. they have the head and mouth shapes of adult animals, the eyes are large and fully developed, and the deep black iris distinctly contrasts from the only sparsely pigmented and hence still colourless individual. The vertebrae sit on the bulbously bloated body like a just closed medulla on a cleavage egg. The abdominal cavity is filled with a huge amount of unused yolk. The vertebrae end with a long, round tail. Both hind and front extremities are already visible as buds. The posterior buds seem to be a bit further developed. The vent is positioned between the hind legs and originates a bit further dorsally than them. The facts that these larvae do not have fins on the tail, that the extremities are well developed, the vent is positioned above the hind legs, and head and mouth already have the definite shape, makes it certain that this frog species is not only viviparous but gives birth to even young frogs that are fully metamorphosed., Published as part of Loader, Simon P., Poynton, John C., Davenport, Tim R. B. & R��del, Mark-Oliver, 2009, Re-description of the type series of Nectophrynoides viviparus (Bufonidae), with a taxonomic reassessment, pp. 41-50 in Zootaxa 2304 on page 50, DOI: 10.5281/zenodo.275339

Nectophrynoides viviparus (Tornier, 1905) Pseudophryne vivipara Tornier, 1905, Sitzungsber. Preuss. Akad. Wiss. Berlin, 39: 855. Nectophrynoides werthi Nieden, 1910, Sitzungsber. Gess. naturf. Freunde Berlin, 10: 439. Tornierobates vivipara Miranda-Ribeiro, 1926, Arq. Mus. Nac., Rio de Janeiro, 27: 19. Nectophrynoides vivipara Noble, 1926, Am. Mus. Novit., 212: 15. Barbour and Loveridge, 1928, Mem. Mus. Comp. Zool., 50: 191. Nectophrynoides viviparus Perret, 1971, Ann. Fac. Sci. Cameroun, 6: 99. Nectophrynoides viviparus Anon., 1996, Bull. Zool. Nomencl., 53: 229. Type locality Kratersee des Nyisilvulkans (Ngosi Crater Lake, Fig. 1). Translation from original text: Species collected from various localities in German East Africa, i.e. Dar es Salam (collected by Werth and Emin Pascha), in Rungwe and the Kinga mountains (collected by S. F��lleborn). The specimens in which viviparity was detected originate from the collection donated to the museum by medical officer Dr. S. F��lleborn with the support of the Academy (of Sciences). Description of the lectotype (measures in mm) ZMB 21775, adult female (Fig. 2). Dissected at lower mid ventral region. Embryos presumably removed and separately stored in small jar (18 embryos, 2.0- 2.2 diameter= head and body including yolk; total length variable: 4.8-6.4). Distance from tip of snout to urostyle 32.9, width of head at jaw articulation 12.6, length of tibia 15.0, length of foot 14.4; tympanum and tympanic annulus easily discernible, ovoid shape 1.5 in diameter; parotoid glands present as a smooth glandular mass, 7.8 in length and 5.4 at widest point starting on the dorsal surface just behind the eye, directed posteriorly; snout relatively short, nostrils closer (2.2) to the snout tip than to each other (2.7) and also to the corner of the eyes (3.5); interorbital distance 6.7; eyes sunk and just about visible ventrally; eye diameter 3.2 right and left, width of eyelid, right 2.7 and left 2.3; interorbital distance 3.8 (difficult to measure as eyes are almost sunken into skull); canthus rostralis concave; arms and legs relatively slender; continuous glandular mass on both arms and legs; on arms, glands are continuous on dorsal, lateral and ventral surfaces of radius/ulna, not present on humerus; on legs, glands are continuous on dorsal, lateral and ventral surfaces of tibiofibulae and tarsal joints, not present on femur; continuous glandular mass on lateral and dorsal margin of the foot, attenuating to the margin of the phalange; rudimentary webbing on hands and feet; first finger shortest with second and fourth equal in size shorter than third finger which is the longest (6.4); first and second toe equal in size, third and fifth toes equal and larger, with fourth toe longest (10.5); small inner metatarsal tubercle, and large expanded outer metatarsal tubercle (1.8), shortest toe 3.0; the body skin is smooth both on glandular and non-glandular areas. Colour pattern of lectotype in preservative Dorsal ground colour light brown with lighter cream patches on head region and parotoid glands; dark brown patches on posterior lateral margins, femoral and humerus regions of the specimen; glands present on lower arm, tibia and foot; ventral colour pattern is cream coloured, slightly white on the centre of ventral region. Variation in material The material from Berlin comprises a range of embryos and adult specimens (SUL: 10.8-32.9). The addition of SHCP material includes adult specimens (SUL: 32.0-40.0). The larger size reported by Channing & Howell (2006) for N. viviparus is based on 166 specimens collected by A. Loveridge from the Ulugurus (e.g. males reach max SUL 56, and females c. 60; Barbour & Loveridge 1928; Channing & Howell 2006). It is unclear whether there are significant geographical differences in SUL between populations. This will require examination of specimens across all the assumed distribution range of the species (e.g. Ulugurus, Udzungwa and Rubeho; Channing & Howell 2006). We present morphometric data and SUL ratios for key measures for the specimens of N. viviparus from the Southern Highlands (Table 2). Barbour & Loveridge (1928) noted colour variation in Uluguru material of Nectophrynoides viviparus and hinted at a correlation with the habitats they were found ��� or ���very slowly��� changing their colouration to suit their habitat. Colour range from jet-black dorsal and ventral regions with glands marked rusty brown. Variations of this include dorsal regions coloured greenish, rufous-brown, bright yellow, in patches. The colour of the glands vary from rusty brown, to a lighter cream colour. In topotypic material colours from the original type series are all discoloured but Tornier (1905) provided some brief comments on their colour at the time of their descriptions (see Appendix 1) and remarked that juveniles are uniform blackish brown. Dorsal parts of back and legs fade as well to a colour that is clear chocolate brown. Dorsal surfaces are irregularly blackish brown and clear chocolate brown. From a recent collection in the Southern Highlands we noted colour variation (Fig. 3). For dorsal and ventral surfaces we observed specimens jet black with cream light brown coloured glands. The chin region was often also lighter coloured being a cream light brown. Differences to this pattern included specimens with a mid dorsal reddish-brown colouration and ventral region reddish brown. We also observed specimens with a light brown/yellow colouration completely covering dorsal and ventral aspects (Fig. 3). Habitat information Nectophrynoides viviparus has been collected in montane forest from approx. 1800 to 2700 m a.s.l. This includes wet, open, secondary and disturbed montane forest, including logged Hagenia forest. In addition specimens have been collected in montane grassland and ericaceous heathland near the forest edge., Published as part of Loader, Simon P., Poynton, John C., Davenport, Tim R. B. & R��del, Mark-Oliver, 2009, Re-description of the type series of Nectophrynoides viviparus (Bufonidae), with a taxonomic reassessment, pp. 41-50 in Zootaxa 2304 on pages 44-46, DOI: 10.5281/zenodo.275339, {"references":["Tornier, G. (1905) Pseudophryne vivipara n. sp. ein lebendig gebarender Frosch. Sitzungsberichte der koniglich preussischen Akademie der Wissenschaften, 39, 1 - 3.","Nieden, F. (1910) Neue ostafrikanische Frosche. Sitzungsberichte der Gesellschaft naturforschender Freunde Berlin, 10, 436 - 452.","Barbour, T. & Loveridge, A. (1928) A comparative study of the herpetological faunae of the Uluguru and Usambara mountains, Tanganyika Territory, with descriptions of new species. Memoirs of the Museum of Comparative Zoology, 50, 87 - 265.","Perret, J. L. (1971) Les especes du genre Nectophrynoides d'Afrique (Batraciens Bufonides). Annales de la Faculte des Sciences du Cameroun, 6, 99 - 109.","Loader, S. P., Channing, A., Menegon, M. & Davenport, T. (2006) A new species of Probreviceps from the Eastern Arc Mountains, Tanzania. Zootaxa, 1237, 45 - 60."]}

Loader, Simon P., Poynton, John C., Davenport, Tim R. B., Rödel, Mark-Oliver (2009): Re-description of the type series of Nectophrynoides viviparus (Bufonidae), with a taxonomic reassessment. Zootaxa 2304: 41-50, DOI: 10.5281/zenodo.275339

Key to species and subspecies of Probreviceps 1 a Tympanum visible......................................................................................................... 2 1 b Tympanum hidden ................................................................................... P. uluguruensis 2 a Snout pointed. Males (> 9 % of SUL) and females (> 4 % of SUL) with a relatively large tympanum...................................................................................................................... 3 2 b Snout rounded. Males (................. P. durirostris 3 b Canthus with no ridges or hardened white .......................................... P. m. rungwensis 4 a Limbs short, tibio��tarsal articulation not reaching elbow ........................ P. rhodesianus 4 b Limbs long, tibio��tarsal articulation reaching shoulder or beyond............................... 5 5 a Tympanum half the diameter of eye. Known from the Uluguru and Udzungwa Mountains........................................................................................................................ ................................................................................................................ P. m. loveridgei 5 b Tympanum 2 / 3 the diameter of the eye. Known from the Usambara and Nguru Moun�� tains .................................................................................................P. m. macrodactylus, Published as part of Loader, Simon P., Channing, Alan, Menegon, Michele & Davenport, Tim R. B., 2006, A new species of Probreviceps (Amphibia: Anura) from the Eastern Arc Mountains, Tanzania, pp. 45-60 in Zootaxa 1237 on page 57, DOI: 10.5281/zenodo.172816

Probreviceps durirostris Loader, Channing, Menegon & Davenport sp. nov. Figures 1���5 Holotype ��� A large female, snout��urostyle length (SUL) 40.1, BMNH 2005.1564 (field number MW 3058) deposited at the Natural History Museum, London (NHM). Collected 5 May 2002, by R. Hinde, S. P. Loader and J. Mariaux from Ikwamba Forest Reserve, Morogoro Region, Kilosa District, Ukaguru Mountains, Tanzania, 06�� 21 ��� 31.5 ��� S; 36 �� 58 ��� 57.8 ��� E, 1500 m. Tissue has been preserved under the field numbers MW 3059 and 3060. Paratypes ��� Eleven specimens, MTSN 5545 �� 51, collected by M. Menegon and Wilirk Ngalason from Mamiwa��Kisara Forest Reserve, 1850 m, 06�� 22 ��� 48 ������S, 36 �� 56 ��� 02������ E, deposited at Museo Tridentino di Scienze Naturali, Trento, Italy (MTSN 5547 is now accessioned in the NHM, as BMNH 2005.1573 and is cleared and stained. Tissue samples MW 4089 ��� 90 were taken from MTSN 5546 ��� 47 respectively); and four specimens MHNG 2670.32, MHNG 2670.33, MHNG 2670.31 and UDSM 2326 were collected by William T. Stanley from ���Camp A���, which is in slightly degraded primary forest, and has the locality: Tanzania, Morogoro Region, Kilosa District, Ukaguru Mts, Mamiwa��Kisara Forest Reserve, 1 km E, 0.75 km S Mt. Munyera, 1900 m, 06�� 22 ��� 45 ��� S, 36 �� 56 ��� 10 ��� E. Diagnosis The species is a brevicipitid on account of the following combination of characters, mainly associated with the retention of the shoulder girdle (Parker, 1934, see Fig. 3): (1) Omosternum cartilaginous (2) Fully developed clavicles and procoracoid cartilages (3) Dilated coracoids (4) Cartilaginous sternum (5) Expanded vomer. The species is assigned to the genus Probreviceps based on the following characters, which separate it from other brevicipitines: (1) Fusion of the urostyle and the sacral vertebra (unfused in Callulina and Spelaeophryne). (2) Palate smooth with two serrated denticulate ridges sensu Parker, (1934); Largen and Drewes, (1989) (Fig. 4) (absent in Breviceps, single row in Callulina, Spelaeophryne, Balebreviceps). (3) Omosternum moderately large (reduced/vestigal in Breviceps and Spelaeophryne) (4) Moderately dilated coracoids (as in Spelaeophryne and Callulina but strongly in Balebreviceps, Breviceps) (5) Terminal phalange simple (expanded in Callulina). P. durirostris is a medium sized Probreviceps which can be easily distinguished from all other species: (1) Snout morphology: P. durirostris has a markedly pointed snout, which is hardened white with ridges along the canthus to the apex of the snout (see Fig. 1). Probreviceps m. rungwensis has a pointed snout, but is not hardened, white in colour, nor does it have a ridged canthus. All other species of Probreviceps have relatively rounded snouts. (2) Tympanum size: P. durirostris is sexually dimorphic with respect to tympanum size, which varies from 13 to 16 % of SUL in males, and 4 to 7 % in females. This is similar to P. m. rungwensis, which varies from 9 to 14 % in males, and 4 to 7 % in females. Both P. m. loveridgei and P. m. macrodactylus males do not possess such large tympani, varying from 4 to 7 % in males, and 3 to 5 % in females. Four P. rhodesianus specimens measured vary in typmanum size 3���5 % of SUL. P. uluguruensis does not have a visible tympanum. (3) Finger ridging: Parker reported a fleshy ridge along the inner side of the second and third finger in P. uluguruensis. Finger ridging is present in P. durirostris, though in some specimen only faintly and not as strongly as P. uluguruensis. Similarly, in P. m. rungwensis and P. rhodesianus finger ridging is present, though very slight in the latter. The ridge is almost entirely absent in P. m. macrodactylus and P. m. loveridgei. (4) Position of vent: The vent opens posteroventrally in P. durirostris (as in P. m. macrodactylus, P. m. loveridgei, and P. m. rungwensis) (opens ventrally in P. rhodesianus, and P. uluguruensis). (5) Advertisement call: the new species can be distinguished from all other species recorded. The call of P. durirostris is a very slow series of clicks, slightly faster than one every two seconds (Fig. 5). This contrasts with the calls of three taxa of Probreviceps in Tanzania that are rapidly pulsed chirps (Mkonyi et al 2004), varying from a mean of 20 /sec (P. uluguruensis); 38 /sec (P. m. macrodactylus) to 90 /sec (P. m. loveridgei). The calls of P. m. rungwensis from Mount Rungwe and the Udzungwa Mountains in Tanzania, and P. rhodesianus from Zimbabwe, are unknown. Description of holotype The SUL is 34.3mm. The head is flat on top with a protruding snout, 5.0mm, the canthal ridges, tip of snout and the upper lip are white and hardened (Fig 1). The canthal ridges, which form an apex at the anterior of the snout are continuous with the ridged skin of the eyelids. The apex of the snout is ridged and slightly raised and pointed in profile. Loreal region is almost vertical, slightly concave. Snout tip to mouth 3.0mm. Maximum head width 14.0mm. Nostrils are round, situated midway between lip and top of snout, with the openings visible laterally. Tongue longer than wider, posterior half and lateral edges not adhered to the lower jaw. Choanae are small and round. Internostril distance 3.3. Snout tip to nostril 2.2mm, and nostril to eye 2.5mm. The eyelid (4.7mm) extends back as a strong supratympanic ridge to above the arm as a fold in the dorsal skin. The eyes are visible from below, eye 3.6mm. Interorbital distance measured across the top of the head 4.3mm. The anterior corners of the eyes are 6.2mm apart. The tympanum is round, 2.8mm, placed 2.5mm from eye, and aligned vertically when viewed from the front. The tubercles of the hand are pale, large with faint skin ridges (fig 1). First finger is the shortest, with second and fourth toe slightly larger, the third being the longest. Length of third finger to include basal tubercle 7.25mm. The legs are relatively short, with a tibia length 17.2mm. First toe is the shortest, with second and fifth toes slightly longer, roughly similar in length, third toe is nearly twice as long as second and fifth, with the fourth the longest. The length of the fourth toe (Fig 1), including the outer metatarsal tubercle 18.6mm. The subarticular tubercles of the toes are rounded, pale, with faint skin ridges. Inner metatarsal tubercle well developed, rounded, length 2.9mm, while outer is large but flattened. Skin rough with small distinct glands that tend to merge on the lower back and upper legs. They are small anteriorly, becoming bigger posteriorly. The only parts not covered with glands are the plantar surfaces of the hands and feet. The vent opens posteroventrally. The following are the body proportions: Maximum head width/SUL 0.35, snout/eye 1.39, eye/ SUL 0.09, eye��tympanum distance/horizontal tympanum 0.89, tympanum/SUL 0.07, anterior corners of eyes/head width 0.44, snout tip to mouth/internostril distance 0.91, internostril/nostril��eye 1.32, third finger/head width 0.52, tibia/head width 1.23, tibia/SUL 0.43, fourth toe/SUL 0.40, inner metatarsal tubercle/fourth toe 0.16. Colour In life: The back and sides have dark warts on a tan brown background. Five faint transverse pairs of dark��grey narrow bands are evenly spaced on the posterior dorsum, angled diagonally and posteriorly from the midline. Small white��tipped warts are irregularly present on dorsum and upper limb surfaces. The snout tip is pale blue��grey, with the side of the head dark grey to black from side of the snout to the arm insertion, running below the upper eyelid and the supratympanic fold. Pupil horizontal with a gold iris. The tympanum is dark grey to black. The upper surfaces of the limbs are dark grey (Fig 2). In alcohol: The background dorsal colour is brown, with darker warts. Four pairs of darker transverse lines are just visible, from midway down the back to the legs. Glands are present on the belly. Back of ankle, and underside of forearm are white. Paratype variation The paratypes are similar to the holotype in proportions and colour. Males vary in SUL from 25.8 to 26.8, with the females larger at 34.3 to 43.1. The males have large tympani, 13 to 16 % of SUL, but 4 to 7 % in females. Ranges of body proportions for the 11 paratypes include: Maximum head width/SUL 0.34���0.49, snout/eye 0.88���1.35, eye/SUL 0.08���0.13, eye��tympanum distance/horizontal tympanum 0.20���0.31 (males), 1.11���1.86 (females), tympanum/SUL 0.13���0.16 (males), 0.04���0.07 (females), anterior corners of eyes/head width 0.39���0.47, snout tip to mouth/internostril distance 0.65���0.94, internostril/ nostril���eye 1.05���1.45, third finger/head width 0.36���0.73, tibia/head width 0.88���1.30, Probreviceps durirostristibia/SUL 0.39���0.46, fourth toe/SUL 0.37���0.46, inner metatarsal tubercle/fourth toe 0.10���0.17. Finger ridges vary from prominent to faint. Advertisement call Males call actively during the day, with some residual calling at night. Calling takes place from within leaf litter or while the animals are concealed under logs or other forest debris. Choruses form rapidly after an individual starts calling. Males were recorded in the field at the Mamiwa��Kisara Forest Reserve. The call is a slow series of clicks (Fig 5), produced at a rate of approximately 32 /minute. Each click consists of a single high��energy pulse with a duration of 6 ms. No real harmonics are present, although significant energy is present at 1.2, 1.6, 2.6, 4.3 and 5.4 kHz. Phylogenetic analyses Alignment of sixteen DNA sequences (concatenated 12 S and 16 S sequences) resulted in a matrix of 717 unambiguously aligned characters, of which 557 were constant and 160 variable; of these, 81 were informative under parsimony. Fourteen Probreviceps were included in the analysis, one brevicipitine Callulina, and the microhylid Hoplophryne uluguruensis was used as an outgroup, based on its position outside of a brevicipitine clade in previous analyses (Loader et al. 2004). Maximum likelihood (heuristic search using 10 random��addition sequence replicates and TBR swapping) and Bayesian analyses were carried out under a GTR I + G model as calculated by Modeltest 3.04 (Posada and Crandall, 1998). Base frequencies were estimated as 0.3143, 0.2341, 0.2241, and 0.2275 for A, C, G and T respectively, substitution rates = 8.6657 26.7726 19.7332 3.2566 77.3197 with the proportion of invariant sites set at 0.4201 and a gamma distribution shape parameter 0.6367. The phylogeny recovered is shown in Fig. 6 a. An exhaustive search option using parsimony yielded eight best trees (231 steps), identical to the phylogeny recovered in ML and Bayesian analyses, differing only in the resolution of P. m. macrodactylus, and P. m. loveridgei from Uluguru, Usambara and Udzungwa which collapse into a polytomy in the consensus tree. Support for clades was measured with P. durirostrisbootstrap proportions for 1000 pseudoreplicates (Felsenstein, 1985), Bayesian posterior probability, and decay indices (Bremer, 1988). In summary, the tree demonstrates that Probreviceps durirostris is monophyletic, and is the sister species to Probreviceps m. rungwensis. Probreviceps m. rungwensis is also shown to have two divergent lineages, which each correspond to distinct areas�� the Udzungwa and Southern Highlands. For further details of the analysis of species and intrageneric relationships in brevicipitines, see Loader et al. (2004). One alternative tree arrangement was investigated to evaluate a previous taxonomic hypothesis based on morphology (Parker, 1934), i.e. the monophyly of P. macrodactylus. Using parsimony, a Templeton test (Templeton, 1983) compared the optimal tree (Fig. 6 b) with constrained suboptimal topology (Fig. 6 c). Likelihood tests (Shimadairo��Hasegawa Test) were similarly applied to compare optimal and suboptimal trees (Fig. 6 c). The best suboptimal trees including a constrained monophyletic P. macrodactylus subspecies complex (see Fig. 6 c), are a significantly worse fit to the data than the optimal arrangements. This result indicates, as shown by Loader et al. (2004), that P. macrodactylus requires re��evaluation, the subject of a paper in preparation. Distribution This new species is presently only known from two forest reserves in the Ukaguru Mountains; Mamiwa��Kisara Forest Reserve and the Ikwamba Forest Reserve (Fig 7). Etymology The name durirostris is derived from the latin ���dura��� meaning hardened and ���rostris��� meaning nose. The name is used in reference to the distinctive hardened nose of the new species. Remarks The new species P. durirostris resembles P. m. rungwensis in general body shape, including the enlarged snout. Their morphological similarity is mirrored in their close phylogenetic relationship as recovered in the optimal trees. It is unclear whether the pointed snout has any specific function, for example, the snout might be used to facilitate excavation of burrows. Probreviceps are believed to primarily burrow using their shovel like hind limbs, being backward burrowers (Van Dijk, 2001). Investigation of the behavioural ecology of each Probreviceps species (with and without pointed noses), and their associated habitats will need to be carried out to infer any functional associations. The distribution of Tanzanian Probreviceps species and subspecies is shown in Fig. 8., Published as part of Loader, Simon P., Channing, Alan, Menegon, Michele & Davenport, Tim R. B., 2006, A new species of Probreviceps (Amphibia: Anura) from the Eastern Arc Mountains, Tanzania, pp. 45-60 in Zootaxa 1237 on pages 49-55, DOI: 10.5281/zenodo.172816, {"references":["Parker, H. W. (1934) A monograph of the frogs of the family Microhylidae. London: Trustees of the British Museum (Natural History), London, i - viii, 1 - 208.","Largen, M. J. & Drewes, R. C. (1989) A new genus and species of brevicipitine frog (Amphibia Anura Microhylidae) from high altitude in the mountains of Ethiopia. Tropical Zoology, 2, 13 - 30.","Mkonyi, F. J., Ngalason, W., Msuya, C. A., Howell, K. M. & Channing, A. (2004) Advertisement calls: Probreviceps loveridgei, P. uluguruensis, and P. macrodactylus. Herpetological Review, 35 (3), 2004.","Posada, D. & Crandall, K. (1998) Modeltest: testing the model of DNA substitution. Bioinformatics, 14, 817 - 818.","Felsenstein, J. (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution, 39 (4), 783 - 791.","Bremer, K. (1988) The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution, 42, 795 - 803.","Templeton, A. R. (1983) Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution, 37, 221 - 244.","Van Dijk, D. E. (2001) Osteology of the ranoid burrowing African anurans Breviceps and Hemisus. African Zoology, 36, 137 - 141."]}