A Revision of the Bitaeniatus Chameleon Group (Hillenius 1959) with Reference to Biogeographical Origins and Radiation

From: An M.Sc. thesis carried out during the summer 2000 under the supervision of Dr E. N. Arnold, at the Department of Zoology, The Natural History Museum, Cromwell Road, London SW7 7BD.
Awaiting publication, All Rights Reserved Rob Pilley 2000.
No part of this manuscript, whether in whole or in part, may be copied and/or used without the prior permission of the author. Correspondence: [email protected]

ABSTRACT
The phylogenetic relationships of the bitaeniatus chameleons (Hillenius 1959) are assessed using both independent external and hemipenal data sets as well un-weighted and weighted combined data sets (total evidence). The analysis of the data matrix based upon weighted total evidence characters yielded a consensus tree with the bitaeniatus group species forming 3 clades. Additionally literature review of all available information concerning this group was undertaken in order to establish a complete group data base. The geochronology and climate change in East Africa since the Pre-Miocene period was also taken into consideration. This information was found to substantiate the 3 clades proposed by the weighted total evidence data set, hence these clades are proposed as sub-groups of the bitaeniatus group.

Key words: Chamaeleo, speciation, bitaeniatus group, East African Highlands.

INTRODUCTION
The bitaeniatus chameleons are a group of live-bearing chameleons of the genus Chamaeleo Laurenti 1768 (Class Reptilia; Order Squamata; Suborder Sauria; Family Chamaeleonidae) found in the highlands of east Africa (Uganda, Kenya, Tanzania, southern Ethiopia and Sudan). The group name was first applied by the Dutch herpetologist Dirk Hillenius (1959) in his paper "The Differentiation within the genus Chamaeleo Laurenti 1768" under the heading "the ovoviviparous chameleons II: Group of bitaeniatus" Hillenius (1959) states that the group consists of the species C. bitaeniatus Fischer 1884 (sensu lato), C. jacksoni Boulenger 1896, C. tempeli Tornier 1899, C. werneri Tornier 1899 and C. fuelleborni Tornier 1900.

At the time of Hillenius' writing any reference to Chamaeleo bitaeniatus also encompassed all of its (then) described subspecies, these are: bitaeniatus, ellioti Gunther 1895, rudis Boulenger 1906, schubotzi Sternfeld 1912, hoehnelii Steindachner 1891, altaeelgonis Loveridge 1957, schoutedeni Laurent 1952 and kinetensis Schmidt 1943. All of these now separately described species can be included in Hillenius' original description of the bitaeniatus group. Since his work several other recently described Chamaeleo species have been allocated to the bitaeniatus group (Tilbury 1995; 1998). In 1986 Klaver and Bohme subdivided the Chamaeleonidae into two subfamilies, the Chamaeleoninae (true chameleons) and Brookesiinae (stub-tailed and pygmy Chameleons). The subfamily Chamaeleoninae contains four genera: Calumma, Furcifer, Bradypodion and Chamaeleo. The genus Chamaeleo is divided further into the subgenera Chamaeleo and Trioceros. It is into the latter subgenus Trioceros that Klaver and Bohme (1986) placed all of the members of the bitaeniatus group. The bitaeniatus group is of particular systematic and evolutionary interest for the following reasons:

1. They are the only live-bearing chameleon group aside from the South African genus Bradypodion Fitzinger 1843.
2. They are endemic to the East African Highlands and found only above 5000'. They appear to have evolved independently and convergently to the West African montane forms C. (Trioceros) quadricornis Tornier 1899, C. (Trioceros) montium Bucholz 1874, C. (Trioceros) eisentrauti Mertens 1968, C. (Trioceros) weidersheimi Nieden 1910.
3. Most members of the group have a very limited range, for example Chamaeleo schubotzi, is only found on Mount Kenya above 9000'.
4. Some members of the group, especially those found in Tanzania (but previously Kenya also), are subject to export for the pet trade in Europe, Japan and the US. The problems associated with this can be attributed to the observation that most of the group's species appear very similar and are thus are hard to distinguish from one another (especially those associated with the south Tanzanian highlands). The result of this is that misidentification by exporters occurs readily; thus a species with a limited range or unknown conservation status may be confused with a wider ranging form and thus legally exported (eg. Chamaeleo werneri vs. Chamaeleo fulleborni). Hence there is a conservation need.
5. The members of the bitaeniatus group appear to be separated into several sub-groups dependent upon the East African geochronology and resulting habitat/climate formation.
   The speciation of the extant species appears closely tied to the formation of the East African rift valleys- the formation of the latter could be considered a major factor in the separation of the ancestral populations that gave rise to present day species.

MATERIALS AND METHODS
Initially a literature search covering all cited members of the bitaeniatus group plus other East African Chamaeleo species (+ species linked or allocated to the group by other authors) was conducted. This was necessary in order to establish all that was known (to date) of the bitaeniatus group and its natural history, range, inter-specific relationships and apparent relationships to the other members of the Chamaeleonidae. This search also revealed several other Chamaeleo species found within or around the range of the bitaeniatus group which have not received allocation to any particular species group.

All of the descriptive information found for the species deemed relevant to this study can be found in the first part of the Results section.

A problem noted by Hillenius (1959) was that many of the original type descriptions or type specimens of the chameleons he studied were based on only a few or even a single specimen. This led to the associated problems of chameleon taxonomy and meant that they were (easily) over classified if only a single specimen was available from any given area or species.

To overcome such a problem several specimens of each species (where possible) were obtained in order to account for apparent in-species variation; both within populations and from different localities. The specimens used plus collection locality etc may be found within the Appendix. All of the specimens used during this work may be located in the Zoology Department at the Natural History Museum, London, UK or were obtained privately.

Eighteen species (and subspecies) were selected as operational taxonomic units for the morphological aspect of this study. All of the following 18 species (and subspecies) were assumed at this stage to represent the members of the bitaeniatus group (or derived from it) based upon all reviewed sources. These were: C. affinis Ruppell 1845, C. bitaeniatus Fischer 1884, C. ellioti Gunther 1895, C. fulleborni Loveridge 1932, C. goetzei goetzei Tornier 1899, C. goetzei nyikae Loveridge 1942, C. hoehnelii hoehnelii Steindachner 1891, C. hoehnelii alteelgonis Loveridge 1957, C. incornutus Loveridge 1932, C. laterispinus Loveridge 1932, C. jacksoni Boulenger 1896, C. marsabitensis Tilbury 1991, C. rudis rudis Boulenger 1906, C. rudis sternfeldi Rand 1963, C. rudis ssp Rand 1963, C. schubotzi Sternfeld 1912, C. tempeli Tornier 1899, s Tornier 1899.

A further 4 species were selected as outgroups: C. deremensis Matschie 1892, C. dilepis Leach 1819, C. fischeri fischeri Reichnow 1887, C. johnstoni Boulenger 1901. The 4 outgroup species were selected as they occupy a similar range to the bitaeniatus group but are not directly related to it (Klaver 1973, 1977; Klaver and Bohme 1986). In addition, the males of deremensis and johnstoni exhibit cranial horns which they have in common with several members of the bitaeniatus group hence suggesting an apomorphic or plesiomorphic character.

The relationships of the 22 species (18 ingroup species + 4 outgroups) were assessed using 21 external characters. In addition, 10 hemipenal characters were assessed from 19 of the selected species. The remaining 3 species (of the 22 species in total) did not receive hemipenal examination for the following reasons:

• C. goetzei nyikae- only a single female specimen was available; C. marsabitensis- the only male specimen available represented the paratype of this species and thus was not dissected for this study; C. incornutus- no male specimens were available. The method employed for hemipenal removal and observation from preserved specimens is taken from Arnold (1986) (a.) pages 1224-1225. The terminology for the external anatomical features follows the terms employed by Hillenius (1959), Rand (1963) and Martin (1992). The hemipenal terminology is derived from Bohme and Klaver (1986).

Parsimony method was used to formulate phylogenetic hypotheses' in this study (Kluge and Farris, 1969; Churchill et. al., 1985; Farris and Kluge, 1985, 1986). Character state transformations were polarized using outgroup comparison. Chamaeleo dilepis / fischeri and Chamaeleo deremensis / johnstoni were included as the first and second outgroups respectively for determining the evolutionary direction of the character state changes.

A hypothetical ancestor was reconstructed to represent the outgroup node and was included in the analysis in order to root the evolutionary tree of the ingroup relationships. The ingroup was stated as monophyletic with respect to the outgroup; the latter was assumed to be paraphyletic (after Hillenius 1959).

The two separate data matrices (external morphology, hemipenal morphology) were subjected to analyses using the computer program PAUP* 4.0b4a PPC (Phylogenetic Analysis Using Parsimony)(Swofford 1991). Within PAUP the individual data matrices were searched via Branch and Bound searches using random addition sequences (TBR). This was followed by Bootstrap analysis (1000 replications) and consensus tree construction for each of the separate data sets. In each case all characters were unweighted.

The external/ ecological character set was then combined with the hemipenal character set (total = 31 characters) in order to give an overall, morphological total evidence hypothesis of relationships (see Kluge (1989), Kluge and Wolfe (1993) and De Queiroz et al. (1995)).

Bootstrap analysis (1000 replicates) was also performed on this combined total evidence reconstruction.

As mentioned above, all characters up to this stage are unweighted.

Once unweighted character assessment had been completed, the following character weighting was applied: A weighting of 1 (as opposed to 0 for all non-weighted characters) was applied to group members that possessed occipital lobes (External character 1 in the appendix). This was based upon the following; Hillenius (1959) separates members of the bitaeniatus group into two sub-groups; those which possess occipital lobes and those in which they are absent. This observation combined with the observations by this study that species that DO possess occipital lobes are only found within the southern Tanzanian mountain ranges (and never cross the present day ranges' of other members of the bitaeniatus group mentioned in the text) led to a weighting of 1 being applied to this character during the weighted character assessment. Similarly, a weighting of 1 is also applied to external character 14; the arrangement of dorsal squamation and in particular the presence of enlarged lateral shields arranged in rows.

In addition, weighting of 1 was applied to hemipenal characters 6 and 7 (see appendix for details)(sulcal papillary arrangement a and b). This was based upon the observations of Bohme and Klaver (1980) and Klaver and Bohme (1986) in which they reference the relevance of sulcal papillary field arrangement and formation within the Chamaeleoninae as a whole.

Once weighting had been applied to the above four characters they were then added back to the combined external/ hemipenal morphology data set (31 characters) and subjected to Bootstrap analysis (1000 replicates) and consensus tree construction.

RESULTS
Current literary knowledge of the bitaeniatus chameleons

a) Chamaeleo bitaeniatus FISCHER 1884. Type specimen from Lake Naivasha, Kenya.

• This represents the type species of the Hillenius' bitaeniatus group.
• It was the first member of the group to be discovered and named, it was given nominate sub-specific rank of the group by Loveridge in 1942.
• Loveridge also went on to rename several other members of the group as subspecies of bitaeniatus:
• C. ellioti GUNTHER 1895 became C. bitaeniatus ellioti (1942), C. hoehnelli STEINDACHNER 1891 became C. bitaeniatus alteelgonis (1935) and was later split again into C. bitaeniatus hoehnelli LOVERIDGE 1957 and C. bitaeniatus altaeelgonis LOVERIDGE 1957. At the same time Loveridge also renamed C. rudis BOULENGER 1906 as C. bitaeniatus rudis LOVERIDGE 1957.
• When first discovered, C. rudis schoutedeni LAURENT 1952 was named C. bitaeniatus schoutedeni LAURENT 1952 but this was later changed by Rand (1963). Similarly, C. kinetensis SCHMIDT 1943 was named as C. bitaeniatus kinetensis and C. schubotzi syn. C. bitaeniatus schubotzi STERNFIELD 1912.
• Again it was Rand (1963) who recognised them as full species and separated them from the bitaeniatus sub-specificity. Such sub-specific allocation has no doubt added to the many range and altitude inconsistencies and misnomers associated with both this species and the species group as a whole. This is especially true where some of the cited authors continue to talk of bitaeniatus and its subspecies under the general heading "bitaeniatus" whereas others separate them as full species
• Chamaeleo bitaeniatus is found throughout Kenya, northern Tanzania, eastern Uganda, Ethiopian central Highlands and Somalia (Rand 1963).
• Its range in Ethiopia and Somalia however seems sketchy and somewhat vague, Bohme and Klaver (1980) also place it in Sudan at the Imatong Mountains where they found it living sympatrically with C. ellioti and C. kinetensis. Rand (1963) states that it is a species from open plains and valleys to lowland mountain slopes up to 7000'. He also goes on to state that it can be found on the plains north of Mount Kenya, the plains of Guaso Nyira (Kedong Valley, Kenya) and further north to the plains of Addis Ababa in Ethiopia. In addition Rand (1971) places it in the Nguru Mountains, Tanzania and also the lower slopes of Kilimanjaro (Rand 1963) but again no altitude or specific data is given and can only be summised. I myself have received specimens collected from the latter locality along with specimens of C. rudis sternfeldi but again no specific collection data was available.
• All the specimens examined during this study support this and suggest that bitaeniatus is indeed a low montane species that does not exceed 7000' in altitude. However, its range and ceiling below this seems sketchy and is not stated in the literature, for example, its apparent range in lowland Ethiopia and Somalia is not listed. Above 7000' it is replaced by both C. hoehnelii and C. schubotzi on Mount Kenya. Rand (1963) states that bitaeniatus and the former species occur in the same general area and have been collected at the same locality (Loita Plains, Mau Escarpment, Lukenya, Lukenya Hills, Kijabe, Aberdare range). However, it seems that in the places where they are collected together bitaeniatus is at its maximum ceiling of 7000' and that this also represents the lowest altitude at which hoehnelli maybe found. Hence this altitude represents a boundary zone between the two species. The altitudinal relationship between bitaeniatus and schubotzi remains unknown.
• It is predicted by this author that as the lowest altitude that hoehnelii may be found at is on the upper limits of the bitaeniatus range then it is assumed that the even higher reaching schubotzi has a minimum altitude far higher than that of bitaeniatus. This suggests that the boundaries of the schubotzi and bitaeniatus do not overlap.

b) Chamaeleo ellioti GUNTHER 1895. Type specimen from the eastern foot of the Ruwenzori Mountains, Uganda.

• Replaces bitaeniatus west of Lake Victoria, across Uganda to Rwanda, Burundi and south to Northern Tanzania (Rand 1963). It is recorded up to a maximum ceiling of 6-7000' and thus appears to represent the lowland montane form of the group in the western part of the groups' range (C. bitaeniatus across Kenya and southern Ethiopia and C. ellioti across a similar altitude in Uganda). A transisiton zone between the two species appears to be present around Lake Victoria and into eastern and northern Uganda although no evidence of sympatry has been represented to date in this particular region. However, Rand (1963) points out that despite the lack of apparent sympatry in this zone members of ellioti found within the area possess striking black gular grooves that are absent in individuals of populations found to the west of Lake Victoria and south to northern Tanzania. He goes on to state that the gular groove change from black in the east and north to less distinct in the south and west is a gradual one and interprets this as what Brown and Wilson (1956) have called "character displacement". Thus he believed the throat grooving to be a species recognition character to enable eastern ellioti to recognise their own species from the similarly proportioned and closely related bitaeniatus which he assumes to also be living in the same area. Rand (1963) noted populations of ellioti from the west and south of Uganda and northern Tanzania exhibit the blackened gular grooves to a lesser extent than their northern counterparts and thus look very similar to bitaeniatus in the east.
• Bohme and Klaver (1980) noted that ellioti DOES live sympatrically with bitaeniatus along with kinetensis SCHMIDT 1943 on Mount Kineti in the Imatong Mountains of southern Sudan. Specimens of ellioti from this area also possessed the black gular grooving attributed to specimens found in the Rand transition zone to the south-east. Personal observation of specimens of ellioti obtained from northern Tanzania (actual locality and altitude data unavailable) over the past two years suggests that specimens from this region also possess the blackened gular grooves attributed to the more northern transition zone populations. This would seem to tie in with Rand's hypothesis and observations that bitaeniatus also occurs within this area.
• On Mount Elgon, Uganda ellioti is replaced above 6000' by C. hoehnelli altaeelgonis but no apparent relationship or transition zone data is given.
• On the mountains that edge the western side of the Rift valley ellioti is replaced by C. rudis above 6000'. Witte (1933) found that there was a large zone of overlap between the two species throughout the Ruwenzori's and the mountains west of Lake Kivu but that ellioti was never found over 6000'. Laurent (1952) noted that this range separation was not associated with the altitude specifically but the ecological niches inhabited by the two; ellioti being an animal of open dryer areas whereas rudis is associated with damper mountain forests. He noted that ellioti reaches higher altitude where the forest had been cut for cultivation.

c) Chamaeleo hoehnelii STEINDACHNER 1891, Type specimen from Laikipia, Kenya at 6000'.

• This species has had a variable taxonomic history along with the other species closely allied to bitaeniatus. Initially named C. hoehnelii by Steindachner in 1891 it was renamed C. bitaeniatus bergeri by Sternfeld in 1912. In 1935 its name was again changed by Loveridge and became C. bitaeniatus altaeelgonis. It was split into two forms by Loveridge in 1957: C. bitaeniatus hoehnelii on Mount Kenya and C. bitaeniatus altaeelgonis on Mount Elgon in Uganda. The latter form was described on account of its overall smaller size and brighter colouration compared to the Mount Kenya form. In his 1963 paper Rand refers to it as the full species Chamaeleo hoehnelii Steindachner from Mount Kenya. The Elgon form syn. C. bitaeniatus altaeelgonis was later renamed C. hoehnelii altaeelgonis Mertens 1966.
• Much of this taxonomic confusion can be attributed to Loveridge's (1957) naming both hoehnelii and altaeelgonis as sub-specific to bitaeniatus rather than to one another. Personal observation of all specimens suggests that both hoehnelii and altaeelgonis are indeed closer to one another than either is to any other member of the bitaeniatus group. This is further substantiated by the similar altitudes and habitats occupied by both subspecies compared to the morphologically and ecologically distinct bitaeniatus.
• Chamaeleo hoehnelli ssp. is found on the upper slopes of Mount Kenya, the Aberdares, Kenya and Mount Elgon, Uganda above 6000'. Unsubstantiated reports also place it in the Cherangani Hills to the north of Mount Kenya and Kabete near Nairobi. C. hoehnelii hoehnelii replaces bitaeniatus above 7000' on Mount Kenya and the Aberdares in Kenya and C. h. altaeelgonis replaces ellioti on Mt Elgon, Uganda above 6000'. The nominate form is found on Mount Kenya to a maximum altitude of 10000' whereas C. h. altaeelgonis on Elgon reaches a maximum of 11000'. On Mount Kenya this is surpassed by C. schubotzi which attains a maximum altitude of 14000'. The ecological relationship of hoehnelii and schubotzi remains unknown in the cited literature but it may be assumed that at altitudes from 7000' to 10000' on Mount Kenya (where both forms occur) they live sympatrically. Above 10000' only schubotzi may be found (Rand 1963).

d) Chamaeleo kinetensis SCHMIDT 1943, Type specimen from Mount Kin(y)eti, Sudan at 10500'.

• Originally named as C. bitaeniatus kinetensis Schmidt 1943 and described as an intermediate between bitaeniatus and ellioti, it was later given full species status by Rand (1963). It has only been found in the Ima(n)tong Mountains of southern Sudan to an altitude in excess of 10500'.
• Klaver and Bohme (1980) received several specimens of this species and found they agreed with the original type description by Schmidt. The specimens they received were found in areas where bitaeniatus and ellioti were also found, thus in the Imantong mountains it appears that all three species are sympatric.
• This still remains the least studied of the original Hillenius' bitaeniatus group. Further work by Bohme and Klaver (1980) examined both the hemipenal and lung morphology of this species and compared it to other members of the bitaeniatus group.

e) Chamaeleo rudis rudis BOULENGER 1906, Type specimen from the Ruwenzori Mountains, Uganda at 10000'.

• Again, another member of the group with a colourful taxonomic history.
• Described as C. bitaeniatus graueri Sternfeld 1912 from the Rugege and Bugoie forests, Uganda at 10000' and C. bitaeniatus tornieri Sternfeld 1912 from the Ituri District, Belgian Congo. Later it was named C. burgeoni Witte 1933 from the Kivu District, Belgian Congo and later still Chamaeleo bitaeniatus rudis Loveridge 1957. In 1963 Rand resurrected the original name by Boulenger to cover all of the above as he noted that the broad-headed chameleons of the western rift highlands were closer to one another than they were to any other bitaeniatus chameleon in the surrounding areas. Loveridge (1957) recorded the range of this species as "mountains above 9000' only, from western Uganda, Rwanda and Burundi to Belgian Congo". The specimens examined by Rand (1963) range from altitudes of 9200' on Mt Ninagongo (= Mt Niragongo) Congo, 12000' on Mt Karisimba, Rwanda and up to 12400' on the western slope of Mount Ruwenzori, Uganda.
• Rand goes on to discuss the variability of species across its large highland range: In some specimens from the Lake Kivu area the junction of the canthal ridges extends forward slightly forming a small dorso-ventrally flattened projection. The length of the gular crest is also highly variable; specimens from the Ruwenzoris, Uganda possessing a crest far longer than that of the same species found in the Congo. Indeed it seems the latter character variability is all that Sternfeld (1912) used for his description of C. bitaeniatus graueri from Bugoie forest, Congo. A correspondence from a Mr Battersby to Rand in 1959 as cited in Rand (1963) states that "the longest gular scales are a little smaller than the size of the eye opening or about x to & the orbital ring in the type of rudis but about 2/3 to 1/5 the eye opening in graueri".
• Rand (1963) felt that since all of the characters Battersby and Sternfeld (1912) mentioned showed considerable variation both with in and in other populations, graueri should be considered conspecific with rudis.
• Witte (1933) described C. burgeoni from the mountains north of Lake Kivu and on the basis of the presence of lateral protuberances at the end of the snout. His later improved experience with the group and discovery of C. schoutedeni to the south led him to re-examine the validity of burgeoni. He decided that "burgeoni was indistinguishable from the other broad-headed chameleons (graueri= rudis) that occur alongside of it".
• Rand (1963) concludes that some, but not all, specimens of rudis DO possess small nasal protuberances in the Kivu area and that SOME individuals from the Ruwenzori's have gular crests far longer than material collected from elsewhere. This author would agree with the above taxonomic decisions based upon all studied material in the museum. Further observation and conclusion shall be made in both the illustration and discussion sections respectively.

f) Chamaeleo rudis schoutedeni LAURENT 1952, Type specimen taken from Kakobo Mountains, Albertville Territory, Tanganyike Province, Belgian Congo at 8000'.

• Originally described as C. bitaeniatus schoutedeni Laurent 1952, recognised as belonging to the rudis species group by Rand (1963), it differs from C. rudis rudis in having 1 to 2 scales between the junction of the canthal ridges and the supralabials rather than 3-4 in the nominate form. In addition it also possesses very short gular and ventral crests.
• Little is known of this form apart from the mention by Bohme and Klaver (1980) of its recent discovery from Kissenje (= Gisenyi) in Rwanda. This implies sympatry with C. rudis rudis and thus full specific status for schoutedeni. They also go on to discuss how the full specific allocation of this species away from rudis is supported by lung morphology (see later morphological section).

g) Chamaeleo rudis sternfeldi RAND (1963), Type specimen from Laikinae, Mt Meru, Arusha District, Tanzania at 7500'.

• Description as for C. rudis rudis, differs only in having the posterior part of parietal crest always swollen (knob-like in some individuals) and the gular/ ventral crest are shorter (longest x vertical eye diameter)(Rand 1963).
• This sub-species is interesting for two reasons:

1. Its range is distinct and separated from the other subspecies to the west and
2. specimens of this sub-species were originally listed as the co-types of C. schubotzi from Mt Kilimanjaro, Tanzania. The latter also appears to tie in with the range description of schubotzi given by Loveridge (1957) as including Mt Kilimanjaro and Nguru, Tanzania. This shall be discussed under the heading of schubotzi later in this section.

h) Chamaeleo rudis ssp. RAND (1963). From various locations as discussed by Rand (1963). Rand mentions several other specimens of C. rudis type material that he cannot allocate to any of the known forms or locales for the rudis species group. Three of these are from Embagi, Tanzania and are described as similar to C. schoutedeni in that they too have very weakly formed gular and ventral crests but the canthals are more widely separated from the labials, they also lack the swollen parietal attributable to the sternfeldi sub-species. I, myself, have received C. rudis ssp. material from Tanzania and also observed two apparent morphological varieties: One type matches that of the description of sternfeldi whereas the other was considerably smaller and matches the description of the above unknown form. The type localities of both recent Tanzanian forms is unknown and requires further material and collecting data in order to establish true relationships and identification.

i) Chamaeleo schubotzi STERNFELD 1912, Type locality Mount Kenya at 14000'.

• Originally described as Chamaeleo bitaeniatus schubotzi Sternfeld 1912, this sub-specific status was later acknowledged by Parker (1932) and Loveridge (1957).
• Rand in 1963 noted that it was closer to the rudis species group rather to the typical bitaeniatus (but that the known rudis members were closer to one another than any one was to schubotzi) and thus recognised it as a full species. C. schubotzi has a very limited range as it is only found on Mount Kenya above 9000', the only other Chamaeleo species found in this area is C. hoehnelii which reaches a maximum ceiling of 10000' on Mt Kenya. The ecological relationship between these two species is not mentioned in any of the cited literature.
• As mentioned above Loveridge (1957) suggests the range of schubotzi as Mount Kenya and Mount Kinangop in Kenya and Mt Kilimanjaro and Nguru in northern Tanzania. On examining the described Tanzanian specimens of schubotzi Rand (1963) dismissed their allocation to the above but instead included them in his newly described C. rudis sternfeldi, thus limiting the range of schubotzi to Mt Kenya (and Mt Kinangop?) only.
• Chamaeleo schubotzi with C. kinetensis in the Imatongs are the highest reaching members of the bitaeniatus species group reaching 14000' and 10500' respectively.

j) Chamaeleo jacksoni jacksoni BOULENGER 1896, Type locality Nairobi at 6000'.

• Hillenius (1959) placed this species within the bitaeniatus group despite its larger size (total length to 14 inches/ 35cm) and presence of three cranial horns in male specimens. His inclusion of jacksonii was based upon it possessing the Y shaped parietal crest (see morphology section for details) which all members of the bitaeniatus group possess (Hillenius' original inclusion character. This inclusion was substantiated by Klaver (1973, 1977) based upon external and lung morphology and also by Bohme and Klaver (1980) who state that hemipenal morphology also confirms this view (the presence of papillary fields on the sulcal apex emphasizing its relationship to bitaeniatus). The type specimen originally came from Kikyu near Nairobi. In 1903 Tornier described the subspecies C. j. vauerescecae also from Nairobi. He based this on its smaller body size and the presence of fully developed horns on females. Hillenius (1959) dismissed this subspecies stating that female specimens he had seen from Nairobi had preorbital horns ranging from completely absent to fully developed.
• In 1958 Rand described a distinct form from Mount Meru, Arusha District, Tanzania. This was named C. jacksoni merumontana RAND 1958.
• In 1988 Eason et al. described a new subspecies from the southern and eastern sides of Mount Kenya only. This was considered distinct from the other forms and named C. jacksoni xantolophus EASON, FERGUSON AND HEBRARD 1988. The three subspecies of Chamaeleo jacksoni can be distinguished by the variation of the number of interorbital scales, body size, parietal crest structure and colour and dorsal crest scale clusters (Eason et al.1988). C.j.jacksoni is found at mid elevation (altitude not available) from the Aberdares (to the west of Mount Kenya) to Nairobi and up onto the southern and south-western slopes of Mount Kenya itself. According to Eason et al. 1988 there is a transition zone between C. j. jacksoni and C. j. xantholophus on the southern slope of Mount Kenya, this extending for some 12.5 km. his has also been suggested by Le Berre (pers.comm 1999) who states that such an area on Mount Kenya represents a hybrid zone between the two sub-specific populations. C. j. xantolophus range extends from Kagumo on the south side of Mount Kenya to Meru on the northeastern side of the mountain. Two specimens from Marsabit to the north of Mount Kenya were also assigned to this subspecies. Whether this represents an isolated population or is just a continuation of the Mount Kenya population remains to be examined. C. j. merumontana has only been described from Mount Meru and Arusha District, Tanzania. Its relationship to the other forms has not been discussed although a morphological similarity dendrogram by Eason et al. (1988) suggests that it is morphologically closer to xantolophus rather than to jacksoni. It is not known whether the range of merumontana extends northwards to reach the boundaries of the other subspecies' ranges or whether it represents a geographically isolated population.

k) Chamaeleo tempeli TORNIER 1899, Type specimen from Uzungwe Mountains, Tanzania

• Loveridge (1957) also records this species from Ubena and Ukinga Mountains, Tanzania.

l) Chamaeleo werneri TORNIER 1899, Type specimen from Uluguru Mountains, Tanzania.

• Loveridge (1957) also records this species from Uzungwe Mountains, Tanzania.

m) Chamaeleo fulleborni (incorrectly spelt fuelleborni) TORNIER 1900, Type specimen from Ngosi Volcano, Poroto Mountains, Ukonde Tanzania at 7000'.

• Rand (1957) also places this species in the Rungwe Mountains, Tanzania.
• (The above three species are endemic to the southern highlands of Tanzania. All appear closely related both in range and morphology, this shall be discussed later. Male specimens of fulleborni and werneri possess three annulated horns, thus making them appear similar to jacksoni in the north.)

The following represent Chamaeleo species which have since been added to the group by various authors.

n) Chamaeleo incornutus LOVERIDGE 1932, Type specimen from Ukinga Mountains, Tanzania.

• Hillenius (1959) discusses this species as perhaps belonging to the group but does not include it as he did not have specimens to hand to examine for himself.

o) Chamaeleo marsabitensis TILBURY 1991, Type locality Marsabit mountain forest, Kenya at 1250 metres.

• Males of this species possess a single, annulated rostral horn. Tilbury (1991) has suggested that such a horn bridges the morphological gap between the hornless bitaeniatus group members and the three horned jacksoni.

p) Chamaeleo conirostratum TILBURY 1998, Type location Lomoriti, south west Imatong Mountains, Sudan at 3500'.

• Tilbury (1998) closely allies this species to the bitaeniatus group members discussed by Rand in 1963 as evidenced by its overall similarity to bitaeniatus and the presence of a Y-shaped parietal crest.
• Conirostratum, like marsabitensis, also possesses a rostral horn/projection, however, in the case of the former species it is flexible and pliable not a solid, annulated horn as in marsabitensis.

Chamaeleo species found within the range of the bitaeniatus group but of uncertain taxonomic status:

q) Chamaeleo laterispinus LOVERIDGE 1932, Type locality Uzungwe Mountains at 6000'.

• Loveridge (1957) also places it in the Uluguru Mountains, Tanzania.

r) Chamaeleo goetzei goetzei TORNIER 1899, Type locality Uzungwe Mountains, Tanzania. Loveridge (1957) also records this species from the Ubena, Ukinga, Rungwe and Porto Mountains, Tanzania.

• The subspecies C. g. nyikae is recorded from the Nyika Plateau, Malawi.
• Hillenius (1959) placed this species with no particular chameleon group but noted that it had characters in common with Chamaeleo affinis and Chamaeleo weidersheimi from Cameroon. Klaver (1977) also leaves its taxonomy open for discussion and placed it with affinis as a "miscellaneous"species based upon lung morphology, BUT he does note the similarity between the lungs of goetzei ssp. and ellioti. Klaver (1977) suggests that goetzei ssp. (with affinis) represents an eastern evolution that radiated to the west African highlands.

s) Chamaeleo affinis RUPPELL 1845, Type locality Ethiopian Plateau.

• The range of this species encompasses that of C. harennae LARGEN 1995, Largen (1995) states that the two species appear closely related.
• C. affinis is described from the mountain forests of the Ethiopian plateau down to the Somalia escarpment (where it is described as a desert species by Martin, 1992).
• It is not known whether it may be found down to the lowlands or whether it is simply a mid-altitude species. For further details on its range relationship with harennae and balebicornutus TILBURY 1998 please see notes under respective species below. For details of taxonomic status see heading under C. goetzei.

t) Chamaeleo harennae LARGEN 1995, Type locality Harenna forest, Katcha, Bale Mountains, Ethiopia at 8000'.

• Largen (1995) notes both the morphological and range similarity between this and the preceding species. He does however point out that harennae is a species associated with forests on the south facing mountain slopes only whereas affinis in the same area are found on the drier, northern facing slopes.
• Lagen separates this species from the bitaeniatus group on account of its parietal crest formation, the same can be said of affinis. Egg layer?

u) Chamaeleo balebicornutus TILBURY 1998, Type locality Harenna forest, Bale Mountains, Ethiopia at 8000'.

• Tilbury (1998) states that this species occurs sympatrically with the preceding species around the 2400m zone but above this only harennae may be found (to 3300m). He also records that neither species lives sympatrically with affinis, the latter species being confined to the northern, drier slopes of the Bale Mountains (to 2700m). He notes the unusual form of the paired rostral horns in balebicornutus in that all other species with paired rostral horns may only be found in west Africa (C. montium and quadricornis). This coupled with the absence of parietal and temporal crests (also in common with the West African forms) supports the hypothesis of Klaver (1977) and Klaver and Bohme (1986) of a West African splinter group (see above reference to goetzei). However, Tilbury (1998) also stated that this is not borne out in any of the other herptiles associated with the Ethiopian Highlands and suggests again that this is simply an area with a high rate of endemnism. Egg layer?

Other Chamaeleo species from within the range of the bitaeniatus group but do not exhibit close affinity to the group (to date).

v) Chamaeleo fischeri REICHNOW 1887, Type locality virgin forest of Eastern Usambara Mountains, Tanzania.

• Now split into 6 subspecies, this species complex is associated with the mountain forests of southern Kenya, the Usambara and Uluguru Mountains and Mounts Kilimanjaro and Hanang, Tanzania. In 1977 Klaver placed the complex into the genus Bradypodion based upon lung morphology. This has since been disputed by Raw (pers. comm 2000), he concludes that the fischeri complex has greater affinity to certain Madagascan Chamaeleo species rather than to those of mainland Africa.
• Similar to this species complex are C. tenuis (tenue) Matschie 1892, C. spinosus Matschie 1892 and C. oxyrhinum Klaver and Bohme 1986(b), all of which occupy a similar range in the Usambara mountains, Tanzania. They were also placed into the genera Bradypodion by Klaver and Bohme (1986)

1. All are egglayers.

w) Chamaeleo dilepis LEACH 1819, Type specimen from French Congo.

• An extremely wide-ranging species found from sub-Saharan Africa to South Africa.
• It now comprises 7 subspecies. Found from near-desert to mountain forest, over most of its range it is mainly associated with savanna-type habitat.
• Taxonomically it is part of Hillenius (1959) "Chamaeleo chameleon group" and considered close to the ancestral chameleon form. This is also substantiated by Klaver (1973, 1977) in his work on lung morphology. All are egg layers.

x) Chamaeleo johnstoni BOULENGER 1901, Type location Virgin forests of Uganda at 6000'.

• Listed also by Loveridge (1957) as being found in the mountain forests of Rwanda and Burundi. This species on first appearance looks remarkably like C. jacksoni but the morphological work of Hillenius (1959) placed it in its own group with C. ituriensis Schmidt 1919 (Johnstoni group Hillenius 1959). Egg layers.

y) Chamaeleo deremensis MATSCHIE 1892, Type locality forests near Derema, Usambara Mountains.

• Again this species bears a resemblance to C. jacksoni in that it too possesses three annulated horns in the male. However, beyond this character it has little resemblance to the bitaeniatus group or any other Chamaeleo species recorded within the area.
• Hillenius (1959) recorded it of "miscellaneous" taxonomic status.
• Klaver (1977) suggests that its complex lung structure indicates that it is one of the most highly evolved of the extant Chamaeleo species. An egg layer.

New data set results and Discussion etc to follow.