Mulanje sylvan chameleon

NMZB-UM 4268, an adult male, collected ‘in the Ruo Gorge Forest on Mlanje Mountain’, Malawi by D.G. Broadley and L. Balarin on 20 December 1962.

A small, slender, long-tailed chameleon (SVL = 46.4–81.7 mm; TL = 51.0–95.6 mm) with the tail longer than the body (TL/SVL = 1.0–1.3). Head short and narrow, with a slightly elevated casque, rounded posteriorly. Parietal crest reduced, comprising 3–5 moderately enlarged, keeled scales. Temporal crest absent. Lateral crest formed by 9–12 irregularly raised tubercles curving postero-superiorly from the orbital rim to the tip of the casque. Supraorbital crest of 12–15 elongate scales extending forward as a strongly tuberculated canthus rostralis. Canthus rostralis weakly developed or absent, occasionally forming small protrusions or ‘horns’ in males. Nostril rhombic, posteriorly directed, situated midway between the tip of the snout and the orbit. Upper labials 13–17, lower labials 16–18. Gular crest absent. Gular grooves fine and inconspicuous. Dorsal crest short, consisting of 0–12 low conical tubercles, more pronounced in males. Ventral crest absent. Body scales relatively homogeneous, flattened, and often arranged in irregular rosettes on the lower flanks; scales on the limbs rounded and separated by minute granules. Tail smooth, dorso-ventrally flattened and tapering distally to a fine tip.

Hemipenis (Fig. 5A). The everted hemipenis is stout, with a short, unadorned pedicel. The asulcal surface is covered in shallow, non-papillate calyces that are slightly transversely enlarged and arranged in eight longitudinal series. The sulcal surface is more complex, with the sulcus spermaticus draining into a wide, shallow, sub-triangular distal cup. Sulcal folds are smooth and unadorned proximally. A pair of subconical bulges is present at the proximal edge of the apex. The apex lacks rotulae and bears two large papillae lobes with scalloped distal edges that partially enclose a prominent midline mucosal hump. Each lobe carries smaller superimposed papillae on its surface, and a cluster of large papillae is present at the base of the lobes. The lateral surface of the truncus is smooth.

Colouration in life (Fig. 6A). Males exhibit a light greyish-green background colour, with the ventrum and base of the tail paler than the anterior two-thirds of the body. The flanks bear two pale brown, triangular bars, and a thin, dark grey-brown broken band extending posteriorly from below the lateral crest of the head, terminating approximately two-thirds along the body length. The interstitial skin between the rosette-like body scales is dark grey along the flanks and beige elsewhere. Scales on the crown are predominantly orange-brown, interspersed with scattered bluish scales. The area below the lateral crest, including the snout below the canthus rostralis and the labial margin, is light yellow-orange, while the gular region is light green with white interstitial skin. The tail is a mixture of light green and yellow, becoming progressively browner anteriorly. Females are bright green overall, with the dorsal surface of the head and crown above the lateral crest brick red. The flanks are uniformly green, occasionally retaining faint brown, wineglass-shaped blotches and scattered yellow scales. The ventral surfaces are off-white, with a distinct white mid-ventral line extending from the chest to the cloaca.

Sexual dimorphism subtle; males exhibit a slightly higher casque, more pronounced rugose canthal tubercles, and slightly longer tail (TL/SVL = 1.16) compared to females (TL/SVL = 1.06). Largest male (PEM R18445): 81.7 (SVL) mm + 85.8 (TL) mm = 167.5 mm total length; largest female (PEM R18445): 84 mm + 94 mm = 178 mm (Stevens 1968). Summary of all measurements can be found in Table 3.

Oviparous. Gravid females were collected in the months of January (this study) and February (Stevens 1968; Tilbury 2018). Clutch sizes ranged from 7 to 10 eggs (8.0 mean), in our specimens, but was reported as low as six by Tilbury (2018). Oviductal eggs measured 14.1–17.0 (15.4 ± 0.8 standard deviation) mm in length and 6.7–8.2 (7.6 ± 0.5) mm in width. In the month of November, a presumed hatchling measuring 27.5 mm SVL was collected. In November 2009, an adult male (PEM R18445) was observed falling from a height into the leaf litter below during broad daylight. Upon capture, a clear bite mark was noted on the dorsal crest, likely inflicted by either a bird or another Nadzikambia. It is presumed that, during the attack, the chameleon dislodged itself and fell to the forest floor for safety. The orbital crest in this specimen and other males examined (PEM R18143–44) shows clear damage consistent with male–male combat in the form of bite marks which appear as V shaped scars (see Petford et al. 2024).

The original species description (Broadley 1965) states only: “The holotype was six feet from the ground, clinging to one of the lianas festooning a tree beside the path from the Lujeri Estate power house through the forest to the Ruo Falls. The Ruo Gorge has an annual rainfall of well over 100 inches.” Since 1965, additional records and observations have been made suggesting the species once occupied lowland, mid-elevation, sub-montane and montane evergreen forest between 600 and 2000 m a.s.l. above which was probably high elevation grassland (Dowsett-Lemaire and Dowsett 1988). It is assumed that chameleons use the entire canopy strata, despite all observations being within a few metres above the observer’s standpoint from the ground. The habitat can be generalised to closed-canopy forest reaching tens of meters high, with occasional small gaps in the canopy particularly along stream gullies.

Occurs in the remaining forest fragments on Mount Mulanje (Figs 1, 7). A single holotype was collected “in the Ruo Gorge Forest” (Broadley 1965) which lies on the southern slope of Mount Mulanje above the ‘power house’, a hydro-electric station for the Lujeri tea estate, which covers the southern foothills of Mount Mulanje. The elevation for the holotype was not provided nor was the distance from the power house (–15.991; 35.6608, 760 m a.s.l.) nor from the Ruo Gorge falls (–15.933; 35.647, 1650 m a.s.l.), which lies approximately 6.7 km above the power house, so a more accurate location is difficult to ascertain.

The first estimate of distribution size was 61 km² with an elevation range of 1100 to 1900 m a.s.l., occurring in Ruo Gorge and a forest patch near Lichenya hut (Tolley 2014). However, with new museum and public database records, it can be assumed that the species occurs in additional forest patches from about 600 m to nearly 2000 m elevation (Fig. 7). Notably, there have been several recent records from tiny remnant patches of lowland forest supporting the argument that the forest was once more extensive and contiguous down to about 630 m elevation (Dowsett-Lemaire and Dowsett 1988) but has since been transformed by agriculture. There are also several larger patches of forest higher on the mountain slopes where it is assumed, but not yet confirmed, the species should occur (Fig. 7). In addition, Mount Mchese lies about 5 km to the north of Mount Mulanje and with the original forest likely being more extensive, the forest may have been contiguous between them, supporting a widespread population. However, this requires confirmation and therefore Mount Mchese is not currently included within the range of N. mlanjensis.

Recent photos from Ruo Gorge on the south slopes of Mulanje show that the entire area is currently under active deforestation from illegal logging (Fig. 8), with the forest patch essentially no longer ecologically intact (T. Brammer pers. comm. 2025). Some recent satellite images from public databases (e.g., ­Google Earth, Global Forest Change) do not yet record this loss (see Hansen et al. 2013; https://glad.earthengine.app/view/glo­bal-forest-change). With the loss of the Ruo Gorge forest, the remaining patches total a maximum of ~15.8 km² but is likely lower given that other areas appear to be under deforestation at present. Anecdotal reports suggest that the patch at Lichenya Hut is also being logged at present (T. Brammer pers. comm. 2025), and recent Google Earth imagery shows that the eastern slopes of the mountain have been denuded of vegetation (Fig. 7). Some of mapped patches are very small, ranging from 0.012 to 7.2 km², and the patches are disconnected, with most separated by transformed habitat. The combination of extremely small patches (with presumably small populations) and lack of connectivity among patches point to a very high medium-term extinction risk for this species.

As of the 1980’s the estimated forest remaining at Mount Mulanje was ca. 70 km² across all forest types from lowland (from 650 m) to montane (up to 2300 m; Dowsett-Lemaire 1988; Dowsett-Lemaire and Dowsett 1988). However, the forest was already party transformed at that time and probably was more extensive historically given the tiny remaining patches that occur in the lowlands, up to several kilometres distant from the mountain (Fig. 7). Assuming these outlying forest patches were historically connected to the patches on the mountain itself, and assuming a fairly even distribution of forest down to about 650 m a.s.l. on the southern side of the mountain, the historical forest extent could have been as great as ca. 380 km².

The convex hull around the mapped forest patches suggests an upper limit for the Extent of Occurrence (EOO: convex polygon surrounding all forest patches) at about 299 km², with a lower estimate of 135 km² if the non-surveyed patches and recently deforested Ruo Gorge are excluded. The Area of Occupancy (AOO) mapped as the total summed area of the number of 2x2 km grid cells of forest, assumed to be occupied (IUCN Standards and Petitions Committee 2022) but excluding Ruo Gorge is estimated as 64 km². The AOO could be as great as 136 km² if Ruo Gorge and the unsurvey forest patches on the eastern and the western slopes are included. The global population is considered severely fragmented given the small remaining habitat patches and the severe disconnect between them. Currently, this species is considered Endangered (Tolley 2014) due to the small extent of occurrence, the severe population fragmentation and the ongoing decline in habitat quality and extent. The recent intensive habitat loss at Ruo Gorge and the presumed impacts at other forest patches necessitates an urgent IUCN reassessment for this species.

Mabu sylvan chameleon

PEM R18055, an adult male, collected form the Mount Mabu forest base camp (–16.2861; 36.4006, 967 m a.s.l.), Zambézia Province, Mozambique by W.R. Branch and W. Conradie on 30 May 2009.

A small, slender, long-tailed chameleon (SVL = 61.1–74.2 mm; TL = 65.3–94.4 mm) with the tail longer than the body (TL/SVL = 1.0–1.3). Head short and narrow, with a slightly elevated casque, rounded posteriorly. Parietal crest reduced, comprising 3–4 moderately enlarged, keeled scales. Temporal crest absent. Lateral crest formed by 8–13 irregularly raised tubercles curving postero-superiorly from the orbital rim to the tip of the casque. Supraorbital crest of 13–15 elongate scales extending forward as a strongly tuberculated canthus rostralis. Canthus rostralis weakly developed or absent, occasionally forming small protrusions or ‘horns’ in males. Nostril rhombic, posteriorly directed, situated midway between the tip of the snout and the orbit. Upper labials 16–19, lower labials 16–20. Gular crest absent. Gular grooves fine and inconspicuous. Dorsal crest short, consisting of 0–9 low conical tubercles, more pronounced in males. Ventral crest absent. Body scales relatively homogeneous, flattened, and often arranged in irregular rosettes on the lower flanks; scales on the limbs rounded and separated by minute granules. Tail smooth, dorsolaterally flattened, and tapering distally.

Hemipenis (Fig. 5B). Similar in general structure to N. mlanjensis, differing only in having the asulcal calyces arranged in 8–9 longitudinal series and sulcal folds more prominent with a central groove.

Colouration in life (Fig. 6B). Males exhibit a light greyish-green background colour, with the ventrum and base of the tail paler than the anterior two-thirds of the body. The flanks bear two pale orange-brown, triangular bars, and a thin dark red-brown stripe extending posteriorly from below the lateral crest of the head, terminating approximately two-thirds along its length. The interstitial skin between the rosette-like body scales is a darker green. Scales on the crown are predominantly pale greenish-blue with scattered smaller orange-brown scales, diffused borders. The area below the lateral crest, including the snout below the canthus rostralis and the labial margin, is bright green, while the gular region is light green and the labials are brown. The tail bears a mixture of light green and brown scales. Females are bright green overall, with the dorsal surface of the head and crown above the lateral crest dark orange-red. The flanks are uniformly green, occasionally retaining faint brown, wineglass-shaped blotches and scattered yellow scales. The ventral surfaces are off-white, with a distinct white mid-ventral line extending from the chest to the cloaca.

Sexual dimorphism subtle; males exhibit a slightly higher casque, more rugose canthal tubercles, and slightly longer tail (TL/SVL = 1.21) compared to females (TL/SVL = 1.17). Largest male (PEM R21128): 74.2 mm SVL + 94.4 mm TL = 168.6 mm total length; largest female (PEM R21130): 66.6 mm + 75.3 mm = 141.9 mm total length. Summary of all measurements can be found in Table 3.

No ova or eggs were present in specimens collected in the months of November (Conradie et al. 2016) or May–June (Branch and Tolley 2010). The absence of ova or eggs in November may indicate that the females had already deposited their eggs, as other females collected at the same time of the year from other inselbergs were gravid (see species accounts below). The orbital crest in male specimens examined (PEM R18055, PEM R21131, PEM R21128) shows clear damage consistent with male–male combat.

Occurs in the mid-elevation wet forest on Mount Mabu (Figs 1, 9; Branch and Tolley 2010). The forest habitat on Mount Mabu occurs between ca. 1000 to 1400 m a.s.l, characterised by a thick, closed-canopy reaching tens of meters high (Branch and Tolley 2010; Bayliss et al. 2014), with occasional small gaps in the canopy particularly along stream gullies.

Occurs only in Afrotemperate forest at medium to high elevations of Mount Mabu, Mozambique (Fig. 9).

Although most of the forest is still intact (Fig. 9), this chameleon is currently considered Near Threatened (Tolley et al. 2019). It is at risk due to encroachment from slash and burn clearing along the forest edges which has an impact on the ecological integrity of the forest and could cause directly mortality of chameleons through injury during tree felling and bush clearing by fire. Currently, the forest is about 48 km² in total area (Fig. 9), but it is not known if it was once more extensive as the available satellite imagery is not sufficient to assess this. The present-day Area of Occupancy (AOO) mapped as the total summed area of the number of 2x2 km grid cells assumed to be occupied (IUCN Standards and Petitions Committee 2022) is 64 km², and Extent of Occurrence (EOO: convex polygon surrounding the forest edges but upscaled to match the AOO as per the IUCN guidelines) is also 64 km². Given this species is at risk due to encroachment causing forest loss, and there is a future plausible threat of rapid declines with increased deforestation rates, a reassessment of extinction risk is needed.

Namuli sylvan chameleon

PEM R21165, an adult male, collected downstream from below the Mahno Forest satellite camp, Mount Namuli (–15.3987; 37.0183, 1632 m a.s.l.), Zambézia Province, Mozambique by S. Loader, M. Menegon and K.A. Tolley on 27 November 2014.

6 specimens: PEM R21164 and PEM R21667, gravid females, same collection details as holotype. PEM R21188, PEM R21189 and PEM R21190, gravid females, collected from near the stream of Ukalini Forest, Mount Namuli (–15.3694; 37.0614, 1618 m a.s.l.), Zambézia Province, Mozambique by G. Bittencourt-Silva, S. Loader and M. Menegon on 29 November 2014.

The new species is named after the British chemist Rosalind Franklin (1920–1958) whose work on X-Ray crystallography, particularly her legendary “photo 51”, revealed the structure of DNA (Franklin and Gosling 1953). Her ground-breaking work subsequently allowed for the field of phylogenetics to develop, decades later. Today, nearly all modern taxonomy is phylogenetically informed, including the description of N. franklinaesp. nov.

The new species is assigned to the genus Nadzikambia based on several distinctive characteristics, including a short stout hemipenis with no apical rotulae, the absence of gular and ventral crests, a weakly developed dorsal crest, a low casque, and heterogeneous body scales that form rosettes of tubercles on the lower flanks (Tilbury et al. 2006). This assignment is further supported by monophyly observed in both mitochondrial and nuclear genes (Tilbury et al. 2006; Branch and Tolley 2010).

The new species can be distinguished from other species of Nadzikambia by a combination of the following characters: lower average number upper labials (15.5 versus 16.3–17.3 in other populations, except from N. mlanjensis which have 15.2), lower average number of lower labials (16.4 versus 17.0–18.1); casque in adult males higher and rounder – similar to N. evanescenssp. nov. (versus flat and extending laterally in N. mlanjensis and N. goodallaesp. nov. and marginally raised and rounded in N. baylissi); scales on posterior and lateral crown of head smooth – similar to N. evanescenssp. nov. and N. baylissi (versus rugose in N. mlanjensis and N. goodallaesp. nov.).

Additionally, the new species occurs in allopatry from all congeneric species, with the closest geographical relative being N. evanescenssp. nov. (approx. 45 km away) and differs genetically from other Nadzikambia species by: 0.4–4.5% 16S, 3.2–12.3% ND2, and 4.2–13.6% ND4 uncorrected net p distances (Table 2).

Adult male (70.3 mm SVL + 88.1 TL = 158.1 mm total length; mass: 8.6 g), with fully everted hemipenis. The specimen has a single ventral incision in the chest region for tissue sample extraction and a small incision on the anterior ventral side of the tail to sever the retractor muscle.

Head short (HL/SVL = 0.32) and narrow (HW = 11.1 mm), distance from tip of snout to the superior edge of the casque is twice the width of the head (HL/HW = 2.03). The casque is slightly elevated above the nape, elongated, and its posterior apex points backward to a rounded rim. The crown of the head is flattened posteriorly, is centrally and anteriorly depressed, and features rugose raised scales (except along the parietal crest), abutting with a few interspersed smaller granules. The largest scales are located along the midline of the snout, adjacent to the parietal crest, and alongside the lateral and supraorbital crests, with six scales present between the orbital crests at mid-orbit. A reduced parietal crest is present on the crown, composed of three moderately raised, enlarged, keeled scales that increase in size posteriorly. Anteriorly, the parietal crest splits into one right and two left weakly raised, keeled scales, forming a V-shape. The crown is bordered on both sides by raised tubercles of the lateral and supraorbital crests and at the rear by the casque. The temporal crest is absent. The lateral crest consists of 10/11 (right/left) raised, irregular tubercular scales that arise from the mid-upper rim of the orbit and curve upward around the casque. The supraorbital crest extends from the lateral crest at the upper posterior rim of the orbit, continuing forward over the eye as a rugose canthus rostralis onto the snout. It is composed of a series of 12/13 raised scales, with the three anterior scales most conical tubercles of the canthus on the snout being much more elevated and prominent than the supraorbitals. The canthus rostralis consists of three elevated scales. The first scale is significantly higher than the adjacent scales, giving the impression of a small horn, while the second scale is much wider than the surrounding ones. A single row of scales separates the anterior canthus rostralis from the upper labials on the snout. The orbit is separated from the upper labials by a single row of small granules. Scales around the eye are elongated, longer than wide, with nine scales above and nine below the orbit.

The nostril is rhombic, posteriorly directed, and positioned midway between the tip of the snout and the front of the orbit. It is separated from the upper labials by two rows of granules, with the scales in direct contact with the nostril much smaller than the larger ones in contact with the upper labials. The nostril is also separated from the orbit and the supraorbital crest by two rows of scales each. No enlarged rostral or mental scales are present. The upper labials (13/14) are sub-hexagonal and subequal in size, with the largest scales located from the tip of the snout to the level of the posterior rim of the eye; scales at the rictus are smaller and more rounded. The lower labials (15/15) are similarly sub-hexagonal and subequal in size, with the largest scales extending from the snout to the posterior rim of the eye; scales at the rictus are also smaller and more rounded. Scales bordering the lower labials are irregularly shaped and subequal in size to the lower labials. Gular grooves are fine and inconspicuous. There is no gular crest. The tubercles of the gular region are round and raised, with the smallest scales located centrally and increasing in size toward the lower labials. The dorsal crest is weakly developed and represented by eight to ten slightly enlarged conical tubercles, largest anteriorly. A ventral crest is absent.

Body scales are relatively homogeneous and flattened, abutting one another and forming irregularly scattered rosettes on the lower flanks, which consist of clusters of two to four scales. Scales on the flanks and toward the vertebral ridge are separated by fine granules on the upper flanks. The smallest scales are located on the belly and are more conical in shape, while the largest, squarish scales are found along the flanks and the paravertebral zone. Scales on the limbs are homogeneous and rounded, separated by minute granules, with the largest scales located on the outer surfaces of the lower leg and forearm.

Soles of hindfeet and forefeet are covered with closely fit smooth round tubercles. Scales above hands and feet heterogenous, larger posteriorly getting smaller anteriorly. Scales under digits slightly larger than those on the soles and more elongated. The scales on the digits directly adjacent to the claws are enlarged, followed by a similarly sized scale that is occasionally split into two.

The tail is longer than the body (TL/SVL = 1.25), dorso-ventrally flattened and tapering distally to a fine tip. The dorsal and lateral scales are squarish and larger than the more rounded ventral scales.

Hemipenis (Fig. 5E, F). Similar in general structure to N. mlanjensis, differing only in having a more prominent flattened midline mucosal hump (similar to N. evanescenssp. nov.).

Colouration in life (Fig. 6C). Dorsal surface predominantly olive to yellowish-green, often suffused with diffuse orange-brown tones. Flanks mottled with irregular dark green to brown blotches forming two indistinct, vertically oriented triangular bars. A dark green lateral band extends posteriorly from behind the eye, below the lateral crest, terminating near the mid-flank; a paler light brownish-green stripe occurs immediately below this band. Interstitial skin between the lateral flank scales dark green. The area below the lateral crest and along the temporal region light yellow-green. Labials distinctly pinkish-brown, contrasting sharply with the surrounding yellow-green of the head. Limbs light green with scattered, irregular darker green bands. Ventral surfaces pale greenish to whitish, the gular region slightly yellowish. The ventral surface has a narrow, pale white line extending between the limbs. Tail yellow-green proximally, grading to light brown distally. Iris pale greyish-brown with fine dark reticulations.

Colouration in preservation (Fig. 10). Predominantly dark purplish-black colouration across the body, with faded patches of lighter scales. The head is mostly purplish-black, with slightly paler regions visible on the crown, around the eyes, and behind the corners of the mouth. A few scattered brownish scales are present behind the head. The gular region is dark purplish along the sides, transitioning to a lighter purplish tone along the midline, with the inner scale skin appearing white. The flanks are uniformly dark purplish, interrupted by a faint pale bar above the forelimbs that extends from the spine to the area behind the arm. A second, less distinct pale bar is visible anterior to the hind limbs, spanning from the spine to the lower flank. A row of mid-lateral scales shows a faint brownish tint. The ventral surface has a narrow, pale white line extending between the limbs. Inner limb surface lighter purple with white inner skin. Sole of feet white. The tail is dark purple with two faint, paler transverse bars near its base.

Measurements for the type series of Nadzikambia franklinaesp. nov. are summarised in Table 5. All paratypes are adult females. The size of the females ranges from 70.7–79.3 (mean: 75) mm SVL and 75.1–86.3 (79.4) mm TL. The tail is longer than the body, with an SVL/TL ratio of 1.00–1.12 (1.05). While the type series includes only a single adult male, the females, on average, exhibit a shorter tail (mean TL/SVL: 1.05 in females versus 1.25 in the male). Gravid females weighed between 13.5–18.7 g (15.7 ± 1.9 g).

The general scalation of the females is very similar to that of the holotype, differing only in minor details. The cranial crests are not as rugose, and the canthus rostralis is not elevated to form protrusions or ‘horns’. The upper labials number 15–17 (15.8), while the lower labials range from 15–17 (16.4). The lateral crest consists of 9–12 scales (10.3), the supraorbital crest has 10–14 scales (12.6), and the dorsal crest includes 3–8 scales (5). The body scalation on the flanks varies from closely fitted scales to scales separated by smaller granules dorsally.

In life, the paratype females display a lime green colouration on the body, legs, gular, and lower surfaces of the head. The ventrum, inner surfaces of the limbs, and soles of the feet are light blue-grey, with a thin white mid-ventral line. The flank bars are reduced to irregular, diffuse light brown blotches. The tail is suffused with a dirty red-orange hue. The scales covering the eyes are blue-white, while the crown of the head is orange-red posteriorly and green anteriorly. In preservative, the head and body become very dark, almost purple-black, with all normal colouration lost except for the thin white mid-ventral line.

All females, except one, were heavily gravid with mature oviductal eggs nearly ready for laying when collected in the month of November. Clutch sizes ranged from 7 to 10 eggs (8.8), with oviductal eggs measuring 13.6–15.2 (14.2 ± 0.4) mm in length and 7.5–8.8 (7.8 ± 0.3) mm in width. The eggs (n = 8) of the one exception were not fully developed and measured, on average 8.2 × 7.0 mm. Gravid females weighed between 13.5–18.7 (15.7 ± 1.9) g.

All specimens were collected in an evergreen mid-elevation wet forest between 1618–1632 m a.s.l. They were found at night perched on tree branches at heights of 5–8 m above the ground but are likely to also occur higher in the canopy. The habitat consists of a closed-canopy forest reaching tens of meters high, with occasional small gaps in the canopy along stream gullies.

Currently only known from the small remaining fragmented patches of mid-elevation wet forest of Mount Namuli (Figs 1, 11) in northern Mozambique. Individuals have been recorded from small forest patches – Ukalini Forest (~ 0.47 km²) near the Namuli (northernmost) granite dome associated with the Napuesa and Nanchili rivers (headwater tributaries of the Malema River) and Manho Forest, a once larger patch of mid-elevation wet forest to the south that is associated to the eastern tributary of the headwaters of the Licungo River, running below (eastern side) the Pilani and Peseni granite domes. The remaining habitat at Manho Forest consists of two tiny patches, each approximately 0.1 km² in size.

The mountain is heavily impacted by small scale commercial agriculture and subsistence farming for crops such as maize and potato. The forest destruction has been continuous over the last several decades and the impact has not ceased. Historical satellite images from Google Earth shows that as of 2009 the mid-elevation wet forest occurred in two main areas at high elevations, for a total of approximately 10.8 km² (Fig. 11). It is not known how widespread the forest was prior to the available satellite imagery but the natural habitat at Mount Namuli does consist of a matrix of forest and grassy savanna so it is possible that forest has been limited on the mountain historically. Nevertheless, historical surveys of the mountain (Last 1887; Vincent 1933) recorded far more widespread forest. Ukalini Forest extended 3–4 km eastwards from its current extent (Vincent 1933), and a now extinct forest patch, Ukasini, once occurred to the east of Ukalini (Vincent 1933). At present, most of the forest on Namuli has been destroyed in favour of agriculture, and the estimated remaining habitat is in three very small patches totalling ca. 0.67 km². This represents a loss of > 90% of the entire distribution of this species over a period of about 15 years. The present-day estimate of the Extent of Occurrence (EOO: convex polygon surrounding the forest edges but upscaled to match the AOO as per the IUCN guidelines) is 8 km² and the Area of Occupancy (AOO) mapped as the total summed area of the number of 2x2 km grid cells assumed to be occupied (IUCN Standards and Petitions Committee 2022) is 8 km². The population can be considered severely fragmented with no habitat patch holding > 50% of the entire population, and there is an ongoing decline in forest extent.

Ribáuè sylvan chameleon

PEM R24394, adult male, collected from the stream margin on Mount Ribáuè (–14.8841; 38.2498, 1055 m a.s.l.), Nampula Province, Mozambique by W. Conradie and K.A. Tolley on 4 December 2018.

5 specimens: PEM R24395, gravid female, collected from the stream margin at Mount Ribáuè (–14.8841; 38.2497, 1066 m a.s.l.), Nampula Province, Mozambique by W. Conradie and K.A. Tolley, on 3 December 2018. PEM R24396, gravid female, collected near a stream on Mount Ribáuè (–14.8850; 38.2491, 1056 m a.s.l.), Nampula Province, Mozambique by W. Conradie and K.A. Tolley, on 4 December 2018. PEM R24253, subadult female, collected from the forested mountain slopes on Mount Ribáuè (–14.8809; 38.2462, 1142 m a.s.l.), Nampula Province, Mozambique by W. Conradie and K.A. Tolley on 14 April 2017. NHMUK 2025.3273, adult female, collected from the stream margin on Mount Ribáuè (–14.8846; 38.2495, 1052 m a.s.l.), Nampula Province, Mozambique by W. Conradie and K.A. Tolley on 3 December 2018. NHMUK 2025.3274, adult male, same details as previous specimen.

This species is named after Jane Goodall (1934–2025), an inspirational scientist who lived and worked in Africa throughout her lengthy career. Although her work was dedicated to the study of Pan troglodytes, the Chimpanzee, she spent much of her life living and working in tropical forest, in particular at Gombe National Park, Tanzania. Like her own study species, this chameleon is a forest endemic and the destruction of forest, and other habitats, both at Mount Ribáuè as well as within the home range of P. troglodytes in Central and West Africa is causing forest-living species to decline to the brink of extinction.

The new species is assigned to the genus Nadzikambia based on several distinctive characteristics, including a short stout hemipenis with no apical rotulae, the absence of gular and ventral crests, a weakly developed dorsal crest, a low casque, and heterogeneous body scales that form rosettes of tubercles on the lower flanks (Tilbury et al. 2006). This assignment is further supported by monophyly observed in both mitochondrial and nuclear genes (Tilbury et al. 2006; Branch and Tolley 2010).

The new species can be distinguished from other species of Nadzikambia by a combination of the following characters: average number upper labials (16.6 versus 15.2 in N. mlanjensis and 15.5 in N. franklinaesp. nov., similar to other species); highest average number lower labials (18.1 versus 16.4–17.7 in other populations), casque in adult males flat and extending laterally (versus higher and rounder in N. franklinaesp. nov. and N. eva­nescenssp. nov., marginally raised and rounded in N. baylissi, similar to N. mlanjensis); scales on posterior and lateral crown of head rugose (versus smooth in N. franklinaesp. nov., N. evanescenssp. nov., N. baylissi, similar to N. mlanjensis); higher number of raised lateral crest and suborbital tubercles (11.2 and 15.6 versus 10.0–10.8 and 12.5–14.5); anterior of crown of head (on snout) narrower between supraorbital crests and more conclave (versus broader, raised and flatter in other species).

Additionally, the new species occurs in allopatry from all congeneric species, with the closest geographical relative being N. evanescenssp. nov. (approx. 85 km away) and differs genetically from other Nadzikambia species by: 0.9–4.9% 16S, 3.8–13.3% ND2, and 4.4–13.2% ND4 uncorrected net p distances (Table 2).

Adult male (75.6 mm SVL + 99.6 TL = 175.0 mm total length), with a fully everted hemipenis. The specimen has a single ventral incision in the chest region for tissue sample extraction and a small incision on the anterior ventral side of the tail to sever the retractor muscle.

Head short (HL/SVL = 0.33) and narrow (HW = 10.9 mm), distance from tip of snout to the superior tip of the casque is twice the width of the head (HL/HW = 2.26). The casque is slightly elevated above the nape, elongated, and its posterior apex points backward to a rounded rim. The crown of the head and casque is flattened posteriorly, is centrally and anteriorly depressed, and features are covered with raised rugose scales (except along the parietal crest), abutting with a few interspersed smaller granules. The largest scales are adjacent to the parietal crest, and alongside the lateral and supraorbital crests above the eyes, with five scales between the two supraorbital crests at the mid-orbital point. A reduced parietal crest is present on the casque, composed of four moderately raised, enlarged, keeled scales that increase in size posteriorly. Anteriorly, the parietal crest splits into two right and one left weakly raised, keeled scales, forming a V-shape. The crown is bordered by raised tubercles of the lateral crests. The temporal crest is absent. The lateral crest consists of 11/10 (right/left) raised, irregular tubercular scales that arise from the mid-upper rim of the orbit and curve upward around the casque. The supraorbital crest extends from the lateral crest at the upper posterior rim of the orbit, continuing forward over the eye and onto the snout as a rugose canthal crest. It is composed of a series of 15/17 tubercles, with the three anterior tubercles on the snout much larger, more elevated and elongated than those above the orbit. The canthus rostralis consists of three large elevated subconical tubercles, the two anterior tubercles are significantly larger than the third, giving the impression of small horns. Three rows of small scales separate the canthi from the upper labials on the snout. The orbit is separated from the upper labials by a single row of small granules. Scales around the eye are elongated, longer than wide, with 10 scales above and 11 below the orbit.

The nostril is rhombic, posteriorly directed, and positioned midway between the tip of the snout and the front of the orbit. It is separated from the upper labials by two rows of granules, with the scales in direct contact with the nostril much smaller than the larger ones in contact with the upper labials. The nostril is separated from the orbit by two rows of scales and three rows from the supraorbital crest. No enlarged rostral or mental scales are present. The upper labials (16/15) are sub-hexagonal and subequal in size, with the largest scales located from the tip of the snout to the posterior rim of the eye; scales at the rictus are smaller and more rounded. The lower labials (17/16) are similarly sub-hexagonal and subequal in size, with the largest scales extending from the snout to the posterior rim of the eye; scales at the rictus are also smaller and more rounded. Scales bordering the lower labials are irregularly shaped and subequal in size to the lower labials. Gular grooves are fine and inconspicuous. Gular scales are round and raised, with the smallest scales located centrally and increasing in size toward the lower labials. The dorsal crest is weakly developed with eight slightly enlarged subconical tubercles anteriorly. A ventral crest is absent.

Body scales are relatively homogeneous and flattened, abutting one another and forming irregularly scattered rosettes on the lower flanks. Scales on the flanks and toward the vertebral ridge are separated by fine granules on the upper flanks. The smallest scales are located on the belly and are more conical in shape, while the largest, squarish scales are found along the flanks and the paravertebral zone. Scales on the limbs are homogeneous and rounded, with no smaller scales separating them, with the largest scales located on the outer surfaces of the hindlimb and forelimb.

Soles of hindfeet and forefeet with closely fit smooth round scales. Scales above forefeet and hindfeet heterogenous, larger posteriorly getting smaller anteriorly. Scales under digits slightly larger than those on the soles and more elongated. The scales on the digits directly adjacent to the claws are enlarged, followed by a similarly sized scale that is occasionally split into two.

The tail is longer than the body (TL/SVL = 1.32), dorso-ventrally flattened, and tapering distally to a fine tip. The dorsal and lateral scales are squarish and larger than the more rounded ventral scales.

Hemipenis (Fig. 5D). Similar in general structure to N. mlanjensis, differing only in that the two subconical bulges is more dorso-ventrally flattened as opposed to the more rounded bulges seen in the other species.

Colouration in life (Fig. 6D). Dorsal ground colour light olive-green, overlaid with a diffuse reddish-brown suffusion most pronounced on the parietal region, casque, and anterior dorsum. Flanks with faint darker green to brownish mottling forming two indistinct triangle vertical bars. Interstitial skin between the flank scales pale green to beige. Casque and temporal regions suffused with maroon to russet tones, contrasting with the lateral surfaces of the head, which are bright green. Labials pale pinkish-brown, grading posteriorly into yellowish-green. Limbs light green with scattered darker tubercles and a faint brown hue proximally. Gular region pale greenish-yellow; ventral surfaces silvery-white with slight yellow suffusion anteriorly. Tail green proximally, becoming light brown distally, with fine darker mottling. Iris reddish-brown with fine dark reticulation.

Colouration in preservation (Fig. 12). Predominantly dark purplish-black colouration across the body, with faded patches of lighter scales. Scales on the crown a mixture of light purplish and white scales. The head is mostly purplish-black, with slightly paler regions visible on the crown, around the eyes, and behind the corners of the mouth. The gular region is dark purplish, lighter anteriorly and to the lower labials. The flanks are uniformly dark purplish. The belly has a faint narrow, pale white mid-ventral line extending from the forelimbs to the vent. Inner limb surface paler purple, with white inner scale skin. Sole of feet white. The tail is dark purple.

Measurements for the type series of Nadzikambia goodallaesp. nov. are presented in ­Table 6. The size of the females ranges from 50.8–76.4 (mean: 68.0) mm SVL and 58.8–84.6 (75.7) mm TL. The tail is longer than the body, with an SVL/TL ratio of 1.08–1.16 (1.12). While the type series includes only two adult males, the females, on average, exhibit a shorter tail (mean TL/SVL: 1.12 in females versus 1.28 in the male).

The paratype male (NHMUK 2025.3273) agrees with the holotype, except for minor differences in scale counts. The general scalation of the paratype females is very similar to that of the holotype and paratype male, differing only in minor details. The cranial crests are not as prominent and the canthus rostralis is not elevated to form small protrusions or ‘horns’. Parietal crest varies from three to four elevated scales. The upper labials number 15–19 (15.9), while the lower labials range from 16–21 (18.5). The lateral crest consists of 10–12 scales (11.1), the supraorbital crest has 14–18 scales (15.6), and the dorsal crest includes 0–7 scales (3.8). The body scalation on the flanks varies from closely opposed tubercles to scales separated by smaller granules dorsally.

In life, the paratype females displays a lime green colouration on the body, legs, gular, and lower surfaces of the head. The belly, inner surfaces of the limbs, and soles of the feet are a light blue-grey, with a prominent white mid-ventral line visible extending onto the ventral tail. The flank bars are either absent or appear as irregular, diffuse light blue-grey blotches. The tail maintains a consistent green colouration anteriorly, becoming more red-orange posteriorly to the tail tip. The scales covering the eyes are lime green, blending seamlessly with the body, while the large scales on the crown of the head have a distinct blue-grey to light brown colouration. In preservative, the paratype female’s head and body become very dark, almost purple-black, with all normal colouration lost except for the thin white mid-ventral line.

All males have damage to their orbital crest which we assume is due to male-male combat. One adult female (PEM R24296) collected in November was not gravid and might have already laid her clutch of eggs. The remaining two adult females (PEM R24395 and NHMUK 2025.3274) were heavily gravid with mature oviductal eggs nearly ready for laying when collected in November. Clutch sizes for both females were eight eggs, with oviductal eggs measuring 13.4–14.2 (13.8 ± 0.3) mm in length and 7.1–7.8 (7.4 ± 0.3 mm) mm in width.

All specimens were collected in an evergreen mid-elevation wet forest between 1052–1142 m a.s.l. on the western massif of Mount Ribáuè. They were found at night perched on tree branches at heights of 2–8 m above the ground but presumably also occur higher in the canopy. The habitat consists of a closed-canopy forest reaching tens of meters high, with occasional small gaps in the canopy along stream gullies.

Currently only known from the forest patches remaining on Mount Ribáuè in northern Mozambique (Figs 1, 13).

Mount Ribáuè is heavily impacted by small scale commercial agriculture and subsistence farming for crops such as maize and potato. The forest destruction has been continuous over the last several decades and the impact has not ceased. Mount Ribáuè consists of two adjacent massifs separated by ~0.5 km of lower valley. Historical satellite images from Google Earth shows that as of circa 2016 the eastern massif, which is closer to the town of Ribáuè, had 3.7 km² mid-elevation wet forest remaining at the highest elevations, with the lower slopes heavily impacted (Fig. 13). This forest has been reduced to ~2 km² at present. The western massif has been more slowly impacted but is currently subjected to slash and burn clearing and has been reduced from 7.2 to 3.1 km² since 2013. In total, the total distribution is estimated as 4.8 km². The Area of Occupancy (AOO) mapped as the total summed area of the number of 2x2 km grid cells assumed to be occupied is 20 km² and EOO (the convex polygon surrounding both remaining forest patches) is estimated at 31 km². The population can be considered severely fragmented with no habitat patch holding > 50% of the entire population, and there is an ongoing decline in forest extent.

Inago sylvan chameleon

PEM R24372, adult male, collected by a stream on Mount Inago (–15.1534; 37.4309, 1280 m a.s.l.), Nampula Province, Mozambique by W. Conradie, G. Bittencourt-Silva, A. Raimundo-Miguel and K.A. Tolley on 8 December 2018.

6 specimens: PEM R24362, gravid female, collected near a stream on Mount Inago (–15.1527; 37.4322, 1274 m a.s.l.), Nampula Province, Mozambique by W. Conradie, A. Raimundo-Miguel, G. Bittencourt-Silva and K.A. Tolley on 8 December 2018. PEM R24376, gravid female, collected near a stream on Mount Inago (–15.1536; 37.4296, 1235 m a.s.l.), Nampula Province, Mozambique by W. Conradie, A. Raimundo-Miguel and K.A. Tolley on 11 December 2018. PEM R24373, gravid female, collected from near a stream on Mount Inago (–15.1534; 37.4308, 1281 m a.s.l.), Nampula Province, Mozambique by W. Conradie, K.A. Tolley and G. Bittencourt-Silva on 8 December 2018. PEM R24260, subadult female, collected from a single standing tree within an area of otherwise felled forest, Mount Inago (–15.1531; 37.4269, 1238 m a.s.l.), Nampula Province, Mozambique by W. Conradie, K.A. Tolley and G. Bittencourt-Silva on 19 April 2017. NHMUK 2025.3275, gravid female, collected near a stream on Mount Inago (–15.1534; 37.4301, 1269 m a.s.l.), Nampula Province, Mozambique by W. Conradie, A. Raimundo-Miguel and K.A. Tolley on 10 December 2018. NHMUK 2025.3276, gravid female collected near a stream on Mount Inago (–15.1536; 37.4296, 1239 m a.s.l.), Nampula Province, Mozambique by W. Conradie, A. Raimundo-Miguel and K.A. Tolley on 11 December 2018.

This species is named Nadzikambia evanescens with the specific epithet from the Latin ‘evanescens’ meaning ‘vanishing’. The name is a present participle that can be used as an adjective or a noun in apposition, and the specific epithet is the same for all genders. The etymology is to highlight the rapidly vanishing forest on Mount Inago and the peril that this species is currently under. The forest has already been reduced to a few small patches, and the uncontrolled conversion of forest to agriculture is continuing. The consequence could be the demise of this endemic forest species, if action is not taken to stop the forest destruction.

The new species is assigned to the genus Nadzikambia based on several distinctive characteristics, including a short stout hemipenis with no apical rotulae, the absence of gular and ventral crests, a weakly developed dorsal crest, a low casque, and heterogeneous body scales that form rosettes of tubercles on the lower flanks (Tilbury et al. 2006). This assignment is further supported by monophyly observed in both mitochondrial and nuclear genes (Tilbury et al. 2006; Branch and Tolley 2010).

The new species can be distinguished from other species of Nadzikambia by a combination of the following characters: higher average number upper labials (16.3 versus 15.2 in N. mlanjensis and 15.5 in N. franklinaesp. nov., similar to other species), lower average number of lower labials (17.4 versus 18.1 in N. goodallaesp. nov., similar to other species); casque in adult males higher and rounder – similar to N. franklinaesp. nov. (versus flat and extending laterally in N. mlanjensis and N. goodallaesp. nov. and marginally raised and rounded in N. baylissi); scales on posterior and lateral crown of head smooth – similar to N. franklinaesp. nov. and N. baylissi (versus rugose in N. mlanjensis and N. goodallaesp. nov.).

Additionally, the new species occurs in allopatry from all congeneric species, with the closest geographical relative being N. goodallaesp. nov. (approx. 85 km away) and differs genetically from other Nadzikambia species by: 0.4–4.7% 16S, 3.22–12.7% ND2, and 4.2–12.4% ND4 uncorrected net p distances (Table 2).

Adult male (78.1 mm SVL + 97.8 TL = 158.1 mm total length), with a fully everted hemipenis. The specimen has a single ventral incision in the chest region for tissue sample extraction and a small incision on the anterior ventral side of the tail to sever the retractor muscle.

Head short (HL/SVL = 0.34) and narrow (HW = 12.4 mm), distance from tip of snout to the superior tip of the casque is twice the width of the head (HL/HW = 2.11). The casque is elevated above the nape, elongated, and its posterior apex points backward to a rounded rim. The casque bulges posteriorly, is centrally and anteriorly depressed, and features smooth flattened scales (except along the parietal crest), abutting with a few interspersed smaller granules. The largest scales are located along the midline of the snout, adjacent to the parietal crest, and alongside the lateral and orbital crests, with six scales present between the orbital crests at mid-orbit. A reduced parietal crest is present on the crown, composed of four moderately raised, enlarged, keeled scales that increase in size posteriorly. Anteriorly, the parietal crest splits into two right and two left weakly raised, keeled scales, forming a V-shape. The crown is bordered on both sides by raised tubercles of the lateral and supraorbital crests and at the rear by the casque. The temporal crest is absent. The lateral crest consists of 9/10 (right/left) raised, irregular tubercular scales that arise from the mid-upper rim of the orbit and curve upward around the casque. The supraorbital crest extends from the lateral crest at the upper posterior rim of the orbit, continuing forward over the eye and onto the snout. It is composed of a series of 14/14 raised scales, with the three anterior scales on the snout being much more elevated and elongated than those above the orbit. The canthus rostralis consists of three prominent subconical tubercles. The first scale is significantly higher than the adjacent scales, giving the impression of a small horn. Three rows of scales separate the anterior edge of the canthi from the upper labials on the snout. The orbit is separated from the upper labials by two rows of small granules. Scales around the eye are elongated, longer than wide, with 10 scales above and 11 below the orbit.

The nostril is rhombic, posteriorly directed, and positioned midway between the tip of the snout and the front of the orbit. It is separated from the upper labials by three rows of granules, with the scales in direct contact with the nostril much smaller than the larger ones in contact with the upper labials. The nostril is also separated from the orbit by three rows of scales and from the supraorbital crest by two rows of scales. No enlarged rostral or mental scales are present. The upper labials (17/18) are sub-hexagonal and subequal in size, with the largest scales located from the snout to the posterior rim of the eye; scales at the rictus are smaller and more rounded. The lower labials (18/18) are similarly sub-hexagonal and subequal in size, with the largest scales extending from the snout to the posterior rim of the eye; scales at the rictus are also smaller and more rounded. Scales bordering the lower labials are irregularly shaped and subequal in size to the lower labials. Gular grooves are fine and inconspicuous. Gular scales are round and raised, with the smallest scales located centrally and increasing in size toward the lower labials. The dorsal crest is mostly smooth, except for four slightly enlarged scales anteriorly. A ventral crest is absent.

Body scales are relatively homogeneous and flattened, abutting one another and forming rosettes on the lower flanks. The smallest scales are located on the belly and are more conical in shape, while the largest, squarish scales are found along the flanks and the paravertebral zone. Scales on the limbs are homogeneous and rounded, separated by minute granules, with the largest scales located on the outer surfaces of the forelimb and hindlimb.

Soles of hindfeet and forefeet with closely fit smooth round scales. Scales above the hindfeet and forefeet heterogenous, larger posteriorly getting smaller anteriorly. Scales under digits slightly larger than those on the soles and more elongated. The scales on the digits directly adjacent to the claws are enlarged, followed by a similarly sized scale that is occasionally split into two.

The tail is longer than the body (TL/SVL = 1.25), dorso-ventrally flattened, and tapering distally to a fine tip. The dorsal and lateral scales are squarish and larger than the more rounded ventral scales.

Hemipenis (Fig. 5C). Similar in general structure to N. mlanjensis, differing only in having a more prominent flattened midline mucosal hump (similar to N. franklinaesp. nov.).

Colouration in life (Fig. 6E). Dorsal ground colour light olive-green, overlaid with a diffuse brown suffusion most pronounced on the parietal region, casque, and anterior dorsum. Flanks with faint darker green to brownish mottling forming two distinct light brown triangular vertical bars. Interstitial skin between the flank scales pale green to beige. Dark green lateral band from lateral crest to midbody. Casque and temporal regions brown, contrasting with the lateral surfaces of the head, which are bright green. Labials yellowish-green. Limbs light green. Gular region silvery-white. Tail green proximally, becoming light brown distally. Iris reddish-brown with fine dark reticulation.

Colouration in preservation (Fig. 14). Predominantly dark purplish-black colouration across the body, with faded patches of lighter brown scales. The head is mostly purplish-black, with slightly paler white-purple scales visible on the crown. A few scattered brownish scales are present behind the head and on the flanks. The gular region is dark purplish, with the inner scale skin appearing white. The ventral surface has a narrow, pale white line extending between the limbs. Inner limbs lighter purple scales with white inner scale skin. Sole of feet white. The tail is dark purple.

Measurements for the type series of Nadzikambia evanescenssp. nov. are summarised in Table 7. All paratypes are adult females, except PEM R24260 which is a subadult. The size of the females ranges from 54.1–78.8 (67.6) mm SVL and 57.5– 9.9 (77.1) mm TL. The tail is longer than the body, with an SVL/TL ratio of 1.06–1.24 (1.14). While the type series includes only a single adult male, the females, on average, exhibit a shorter tail (mean TL/SVL: 1.14 in females versus 1.25 in the male).

The general scalation of the females is very similar to that of the holotype, differing only in minor details. The cranial crests are not as developed, and the canthus rostralis not as prominent. The parietal crest comprises of 3–4 raised keeled scales. The upper labials number 15–17 (15.8), while the lower labials range from 15–17 (16.4). The lateral crest consists of 9– 2 scales (10.7), the supraorbital crest has 11–16 scales (13.5), and the dorsal crest includes 4–9 scales (5.8). The body scalation on the flanks varies from closely fitted scales to scales separated by smaller granules dorsally.

In life, the paratype females displays a lime green colouration on the body, legs, gular, and lower surfaces of the head. The belly, inner surfaces of the limbs, and soles of the feet are a light white, with a prominent white mid-ventral line visible extending onto the ventral tail. The flank bars are absent, except for faint paler white blotches. The tail maintains a consistent green colouration anteriorly, withe red-orange bars posteriorly to the tail tip. The scales covering the eyes are lime green, blending seamlessly with the body, while the large scales on the crown of the head have a distinct blue-grey to light brown colouration. The vertebral crest scales are yellow. In preservative, the paratype female’s head and body become very dark, almost purple-black, with all normal colouration lost except for the thin white mid-ventral line.

Four of the females (PEM R24362, PEM R24376, PEM R24373, NHMUK 2025.3276) collected in November were gravid. Clutch sizes comprised nine eggs, with oviductal eggs measuring 12.3–15.1 (13.6 ± 0.8) mm in length and 6.1–7.8 (6.8 ± 0.4) mm in width.

All specimens were collected in an evergreen mid-elevation wet forest between 1235–1281 m a.s.l. They were found at night perched at heights of 2.5–8 m above the ground. The habitat consists of a closed-canopy forest, with occasional small gaps along stream gullies.

Occurs in small, fragmented and isolated mid-elevation wet forest patches at high elevation (ca. 1300 m a.s.l.) on Mount Inago in northern Mozambique (Figs 1, 15).

The area is heavily impacted by continuing small scale commercial agriculture and subsistence farming for crops such as maize and potato. The forest destruction has been continuous over the last several decades. Historical satellite images from Google Earth shows that as of 2009 the forest occurred in four main patches at high elevations, for a total of approximately 14.1 km² (Fig. 15). It is not known how widespread the forest was prior to the available satellite imagery but the natural habitat at Mount Inago does consists of a matrix of forest and grassy savanna so it is possible that forest has been limited on the mountain historically. Since 2009, most of the forest has been destroyed by slash and burn and subsequently used for agriculture (Fig. 16). At present, the Google Earth imagery shows approximately 2.3 km² remains. In just 15 years, there has been an 84% loss in forest extent. The present-day estimate of the Extent of Occurrence (EOO: convex polygon surrounding the forest edges but upscaled to equal the AOO as per the IUCN guidelines) is 16 km² and the Area of Occupancy (AOO) mapped as the total summed area of the number of 2x2 km grid cells assumed to be occupied (IUCN Standards and Petitions Committee 2022 but upscaled to match the AOO as per the IUCN guidelines) is 16 km². The population can be considered severely fragmented with no habitat patch holding > 50% of the entire population, and there is an ongoing decline in forest extent.

Chiperone sylvan chameleon

PEM R24249, adult female, collected by a stream on Mount Chiperone (–16.5070; 35.7258, 1045 m a.s.l.), Zambézia Province, Mozambique by W. Conradie, G. Bittencourt-Silva, S. Loader and K.A. Tolley on 7 April 2017.

(4 specimens). PEM R24245, adult female, collected by a stream on Mount Chiperone (–16.5072; 35.7258, 1017 m a.s.l.), Zambézia Province, Mozambique by W. Conradie, G. Bittencourt-Silva, S. Loader, and K.A. Tolley on 5 April 2017. PEM R24250, gravid female, collected by a stream on Mount Chiperone (–16.5080; 35.7248, 1021 m a.s.l.), Zambézia Province, Mozambique by W. Conradie, G. Bittencourt-Silva, S. Loader and K.A. Tolley on 8 April 2017. NHMUK 2025.3277, gravid female, collected near a stream on Mount Chiperone (–16.5066; 35.7260, 1053 m a.s.l.), Zambézia Province, Mozambique by W. Conradie, G. Bittencourt-Silva, S. Loader and K.A. Tolley on 7 April 2017. NHMUK 2025.3278, adult female, collected by a stream on Mount Chiperone (–16.508; 35.7248, 1020 m a.s.l.), Zambézia Province, Mozambique by W. Conradie, G. Bittencourt-Silva, S. Loader and K.A. Tolley on 7 April 2017.

This species is named after the “Ciperoni” – the term used locally for the weather that brings heavy clouds and orographic rainfall to the area. The cloud sustains the mid-elevation wet forest on this mountain. The epithet ‘nubila’ is derived from the Latin ‘nubilus’ meaning “cloudy,” and is modified to the feminine form to agree with the feminine gender of the genus Nadzikambia.

The new species is assigned to the genus Nadzikambia based on several distinctive characteristics, including the absence of gular and ventral crests, a weakly developed dorsal crest, a low casque, and heterogeneous body scales that form rosettes of tubercles on the lower flanks (Tilbury et al. 2006). This assignment is further supported by monophyly observed in both mitochondrial and nuclear genes (Tilbury et al. 2006; Branch and Tolley 2010).

Due to the lack of males in the type series this species cannot be compared to the male holotypes of other species. The new species can be distinguished from other species of Nadzikambia by a combination of the following characters: higher average number upper labials (16.6 versus 15.2 in N. mlanjensis and 15.5 in N. franklinaesp. nov., similar to other species), and lower average number of lower labials (17.4 versus 18.1 in N. goodallaesp. nov., similar to other species).

Additionally, the new species occurs in allopatry from all congeneric species, with the closest geographical relative being N. mlanjensis (approx. 65 km away) and differs genetically from other Nadzikambia species by: 0.7–4.7% 16S, 1.7–12.7% ND2, and 4.7–11.5% ND4 uncorrected net p distances (Table 2).

Adult female male (64.7 mm SVL + 78.0 TL = 142.7 mm total length). The specimen has a single ventral incision in the chest region for tissue sample extraction.

Head short (HL/SVL = 0.29) and narrow (HW = 8.6 mm), distance from tip of snout to the superior edge of the casque is twice the width of the head (HL/HW = 2.21). The casque is slightly elevated above the nape, elongated, and its posterior apex points backward to a rounded rim. The crown of the head is flattened both anteriorly and posteriorly. It is depressed centrally and laterally relative to the parietal crest. The dorsal surface bears flattened, smooth scales, except along the parietal crest itself, and these abut one another with a few smaller interspersed granules. The largest scales occur along the midline of the snout, adjacent to the parietal crest, and alongside the lateral and supraorbital crests. Six scales are present between the supraorbital crests at mid-orbit. A reduced parietal crest is present on the crown, composed of three moderately raised, enlarged, smooth scales that increase in size posteriorly. The crown is bordered on both sides by raised tubercles of the lateral and supraorbital crests and at the rear by the casque. The temporal crest is absent. The lateral crest consists of 8/9 (right/left) raised, irregular tubercular scales that arise from the mid-upper rim of the orbit and curve upward around the casque. The supraorbital crest extends from the lateral crest at the upper posterior rim of the orbit, continuing forward over the eye and onto the snout. It is composed of a series of 14/13 raised scales. The canthus rostralis not elevated to form any protrusions or ‘horns’. Three to four rows of smaller scales separate the canthi from the upper labials on the snout. The orbit is separated from the upper labials by one to two rows of small granules. Scales around the eye are elongated, longer than wide, with nine scales above and 10 below the orbit.

The nostril is rhombic, posteriorly directed, and positioned midway between the tip of the snout and the front of the orbit. It is separated from the upper labials by three rows of granules, with the scales in direct contact with the nostril much smaller than the larger ones in contact with the upper labials. The nostril is also separated from the orbit by three rows of scales and from the supraorbital crest by two rows of scales. No enlarged rostral or mental scales are present. The upper labials (17/17) are sub-hexagonal and subequal in size, with the largest scales located from the snout to the posterior rim of the eye; scales at the rictus are smaller and more rounded. The lower labials (19/18) are similarly sub-hexagonal and subequal in size, with the largest scales extending from the snout to the posterior rim of the eye; scales at the rictus are also smaller and more rounded. Scales bordering the lower labials are irregularly shaped and subequal in size to the lower labials. Gular grooves are fine and inconspicuous. Gular scales are round and raised, with the smallest scales located centrally and increasing in size toward the lower labials. The dorsal crest is mostly absent, except for five slightly raised scales anteriorly. A ventral crest is absent.

Body scales are relatively homogeneous and flattened, abutting one another and forming rosettes on the lower flanks. No smaller granules between scales on the flanks. The smallest scales are located on the belly and are more conical in shape, while the largest, squarish scales are found along the flanks and the paravertebral zone. Scales on the limbs are homogeneous and rounded, with the largest scales located on the outer surfaces of the hindlimbs and forelimbs.

Soles of hindfeet and forefeet with closely fit smooth round scales. Scales above digits heterogenous, larger posteriorly getting smaller anteriorly. Scales under digits slightly larger than those on the soles and more elongated. The scales on the digits directly adjacent to the claws are enlarged, followed by a similarly sized scale that is occasionally split into two.

The tail is longer than the body (TL/SVL = 1.21), dorso-ventrally flattened, and tapering distally to a fine tip. The dorsal and lateral scales are squarish and larger than the more rounded ventral scales.

Colouration in life (Fig. 6E). Green body with intricate brown blotches. The head is a pale greenish laterally, with light orange-brown labials. The scales on the crown of the head are red-brown. The eyes are surrounded by lighter, greenish-blue scales. The flanks of the body feature a mixture of green and orange-brown scales, transitioning into irregular darker brown patches. The limbs are green-brown, with a slightly paler tone along the ventral surfaces and inner surface of limbs. The chest region blueish-grey. The ventral surface is predominantly pale greenish-yellow, with a white mid-ventral line. The tail exhibits a gradient of green-brown tones, with faint banding patterns and a slight orange tinge near the base.

Colouration in preservation (Fig. 17). Dark purplish-­black colouration across the whole body. The head is purplish-black. The gular region is dark purplish, with the inner scale skin appearing white. The ventral surface has a narrow, pale white line extending between the limbs. Inner surface of limbs lighter purple than outer surface, with white inner skin. Sole of feet white. The tail is dark purple.

Measurements for the type series of Nadzikambia nubilasp. nov. are presented in Table 8. All paratypes are adult females. The size of the females (including holotype) ranges from 59.6–64.7 (62.0) mm SVL and 71.2–78.0 (72.9) mm TL. The tail is longer than the body, with an SVL/TL ratio of 1.12–1.28 (1.19).

The general scalation of the paratypes are very similar to that of the holotype, differing only in minor details. The upper labials number 16–17 (16.6), while the lower labials range from 16–20 (17.8). The lateral crest consists of 9–12 scales (10.0), the supraorbital crest has 12–16 scales (14.4), and the dorsal crest includes 4–6 scales (5.2).

Colouration in life and preservative of the paratypes are in agreement with the holotype. Male colouration currently unknown.

Three of the females (PEM R24250, NHMUK 2025.3277, NHMUK 2025.3278) collected in early April were gravid. Clutch sizes ranged from 4 to 5 eggs (4.3), with oviductal eggs measuring 12.9–14.1 (13.5 ± 0.4) mm in length and 6.4–7.1 (6.7 ± 0.3) mm in width.

All specimens were collected in an evergreen mid-elevation wet forest between 1017–1053 m a.s.l. They were found at night perched in tree branches at heights of 4–7 m above the ground but are likely to also occur higher in the canopy. The habitat consists of a closed-canopy forest reaching tens of meters high, with occasional small gaps in the canopy along stream gullies.

Occurs only in the mid-elevation wet forest of Mount Chiperone in northern Mozambique (Figs 1, 18).

Mount Chiperone is known to the local community as “the cloud maker” as the mountain catches atmospheric moisture, creating cloud cover and forming the rain that not only supplies the forest, but the surrounding landscape. The mountain is also considered holy to most of the communities surrounding it, which results in a measure of community protection in terms of over-­exploitation. Rainforest originally occurred above ca. 1000 m elevation below which was Brachstegia woodland (savanna). Forest extent as of 2024 is about 7 km² from ca. 15 km² in 2006 (Fig. 18) for a 53% decrease in forest cover. Most of the forest loss appears to be on the lower south-west slopes of the mountain and while some of this clearly shows the scars of agricultural use, other areas show only what appears to be attrition of forest on the edges, being replaced by scrubby vegetation (Fig. 18). Similarly, there appears to be a recent change in vegetation at the summit where Google Earth satellite imagery between 2006–2016 shows trees covering the summit, but beginning in 2018 and progressing to present, the area appears to be progressively devoid of vegetation, having shifted to what appears to be open granite patches. It is unclear if this change was due to slash and burn practices which seems unlikely at such high elevation, leaving the lower slopes unaffected. It is possible that orographic rainfall that would normally have provided the moisture needed to maintain forest has been reduced, and the forest extent is being impacted. Thus, while the mountain shows long-term impacts from agriculture along the lower slopes, the forest at higher elevations is nevertheless contracting, and this may be due to climatic changes.

The Area of Occupancy (AOO) mapped as the total summed area of the number of 2x2 km grid cells assumed to be occupied is 16 km² and EOO (the convex polygon surrounding both remaining forest patches but upscaled to match the AOO as per the IUCN guidelines) is estimated at 16 km².