![Chimanimani vista with Dissotis pulchra [BW] Figure 2.5. View over upper Mufomodzi valley [JT]](https://www.researchgate.net/profile/Jonathan-Timberlake/publication/333458183/figure/fig4/AS:856997964558343@1581335645449/Chimanimani-vista-with-Dissotis-pulchra-BW-Figure-25-View-over-upper-Mufomodzi-valley_Q320.jpg)
Content uploaded by Jonathan Timberlake
Author content
All content in this area was uploaded by Jonathan Timberlake on Feb 10, 2020
Content may be subject to copyright.
CEPF Grant 63512: In From the Cold: Providing the Knowledge Base for
Comprehensive Biodiversity Conservation in the Chimanimani Mountains,
Mozambique
CHIMANIMANI MOUNTAINS:
BOTANY AND CONSERVATION
Revised, November 2016
Jonathan Timberlake, Iain Darbyshire, Bart Wursten, Jeneen Hadj-Hammou, Petra
Ballings, Anthony Mapaura, Hermenegildo Matimele, Aurelio Banze, Hercilia
Chipanga, Daglasse Muassinar, João Massunde, Inês Chelene, Jo Osborne & Toral Shah
Chimanimani Mountains: Botany & Conservation, page 2
Front cover: Mt Messurussero & Brachystegia woodland, N Chimanimani Mts (JT).
Frontispiece: Sunrise with Mt Messurussero (JT, top); view over upper Mufomodzi valley (JO, middle L);
weathered quartzite by Mt Nhamadimo (JT, middle R); gold diggers, Mufomodzi tributary (JT, bottom L);
Disa fragrans (JT, bottom R).
Photo credits: BW ‒ Bart Wursten, JO – Jo Osborne, JT – Jonathan Timberlake, TS ‒ Toral Shah
Suggested citation: Timberlake, J.R., Darbyshire, I., Wursten, B., Hadj-Hammou,
J., Ballings, P., Mapaura, A., Matimele, H., Banze, A., Chipanga, H., Muassinar,
D., Massunde, M., Chelene, I., Osborne, J. & Shah, T. (2016). Chimanimani
Mountains: Botany and Conservation. Report produced under CEPF Grant 63512.
Royal Botanic Gardens, Kew, London. 95 pp.
Chimanimani Mountains: Botany & Conservation, page 3
LIST OF CONTENTS
List of Contents .......................................................................................................................... 3
List of Tables .............................................................................................................................. 4
SUMMARY ............................................................................................................................... 5
1.INTRODUCTION ............................................................................................................... 7
2.DESCRIPTION OF THE AREA ........................................................................................ 9
2.1Administration ................................................................................................................ 9
2.2Physical Features and Geomorphology ........................................................................ 11
2.3Geology & Soils ........................................................................................................... 13
2.4Climate ......................................................................................................................... 15
3.HISTORICAL BACKGROUND ...................................................................................... 17
3.1Prehistory & Archaeology ............................................................................................ 17
3.2Colonial & Recent History ........................................................................................... 17
3.3Conservation Situation and Initiatives.......................................................................... 21
3.4Chimanimani Trans-Frontier Conservation Area ......................................................... 23
4.PREVIOUS BIOLOGICAL WORK ................................................................................. 26
4.1Previous Botanical and Ecological Studies .................................................................. 26
4.2Moist Forest Studies ..................................................................................................... 28
4.3Botanical Collectors ..................................................................................................... 29
4.4Phytogeography ............................................................................................................ 32
4.5Wildlife Studies ............................................................................................................ 34
5.VEGETATION ................................................................................................................. 36
5.1Regional Level.............................................................................................................. 36
5.2District Level ................................................................................................................ 36
6.BOTANY AND ECOLOGY ............................................................................................ 45
6.1Species and Collections ................................................................................................ 45
6.2New Species and Records ............................................................................................ 46
6.3Endemics ...................................................................................................................... 49
6.4Assessment of Conservation Status .............................................................................. 57
6.5Ecology ......................................................................................................................... 58
7.THREATS AND CONSERVATION ............................................................................... 62
7.1Gold Panning ................................................................................................................ 62
7.2Environmental Effects of Artisanal Mining ................................................................. 64
7.3Alien Invasive Plants .................................................................................................... 69
7.4Climate Change ............................................................................................................ 70
7.5Unprotected Habitat and Corridors .............................................................................. 70
7.6Wildlife and Wildlife Habitats ..................................................................................... 71
8.MAIN FINDINGS ............................................................................................................ 72
9.RECOMMENDATIONS .................................................................................................. 74
10.ACKNOWLEDGEMENTS .............................................................................................. 76
11.REFERENCES AND BIBLIOGRAPHY ......................................................................... 77
ANNEX 1. Participants in the CEPF-funded botanical expeditions, 2014 & 2016 ................. 84
ANNEX 2. List of range-restricted species associated with the Chimanimani Mountains. ... 85
ANNEX 3. Georeferenced localities on or related to the Chimanimani Mountains. .............. 92
Chimanimani Mountains: Botany & Conservation, page 4
LIST OF TABLES
Table 2.1. Rainfall figures for Chimanimani area, Zimbabwe .............................................. 16
Table 4.1. Botanical collections and collectors in the Chimanimani area 1906
to 2016 ................................................................................................................. 30
Table 5.1. Summarized relationships of Chimanimani plant communities and
ecological factors ................................................................................................. 44
Table 6.1. Species newly recorded for Mozambique during the 2014 and 2016
Chimanimani botanical expeditions .................................................................... 47
Table 6.2. Chimanimani endemic and near-endemic species from lower altitudes .............. 50
Table 6.3. Chimanimani endemic species – Umkondo sandstones ....................................... 51
Table 6.4. List of species endemic to the Chimanimani massif at montane and lower
altitudes, along with Red List assessments ......................................................... 53
Table 6.5. Number of endemic or near-endemic taxa in the Chimanimani area ................... 56
Table 6.6. Summary of conservation assessments of Chimanimani montane and
quartzite endemics and near-endemics ................................................................ 58
Table 7.1. Estimated density of gold-panners and impacts across Mozambique portion
of the Chimanimani Mountains, Dec 2006 ......................................................... 64
Chimanimani Mountains: Botany & Conservation, page 5
SUMMARY
Straddling the Mozambique‒Zimbabwe border at around 20oS, the Chimanimani Mountains
have long been known as an area of high plant diversity and endemism. With almost three-
quarters of the montane massif in Mozambique, the upland area covers around 530 km2 and
ranges in altitude from about 500 m in the south to the highest peak of Mt Binga at 2436 m.
Most of the main plateau lies at around 1000 to 1800 m. The craggy appearance of the
mountains results from the underlying geology of resistant Precambrian Umkondo quartzites,
the weathering of which results in very nutrient-deficient soils. It is this that is believed to be
the evolutionary driver for the exceptional levels of plant endemism found here. The
mountains are protected on both sides of the border as a National Park or National Reserve,
and together form they part of a Trans-Frontier Conservation Area (TFCA).
Despite being a protected area, the mountains were invaded in the 2004‒2006 period by many
thousands of illegal small-scale miners, who dug into stream beds looking for alluvial gold;
perhaps a thousand still remain, nearly all on the Mozambique side. At the time, fears were
expressed by conservationists over the possible impacts of these miners on the flora,
vegetation and wildlife.
In order to evaluate any conservation impact, and to address the comparative lack of
knowledge of plant distribution on the Mozambique side, the Royal Botanic Gardens, Kew in
UK and the Micaia Foundation (an NGO in Mozambique), in conjunction with the National
Herbarium in Maputo (Instituto de Investigação Agrária de Moçambique, IIAM) and the
National Herbarium and Botanic Garden in Harare, received a grant in 2014 from the Critical
Ecosystem Partnership Fund. This was to look at restricted-range and endemic species on the
Mozambique side of the mountains in order to assess their distribution, the threats to them and
their conservation status, and to provide appropriate conservation recommendations for the
national authorities. Three field trips were carried out in different areas and in different
seasons. This report describes and discusses the findings, and also provides some historical
background to botany and plant collecting across the mountains.
A database of herbarium specimens of many range-restricted species known from the
Chimanimani area was compiled, and a list of true endemics determined. Out of an estimated
920 taxa recorded from the mountains above 1200 m altitude, 78 taxa (species, subspecies and
described varieties), or 8.5%, are believed to be endemic, including 9 taxa thought to be new
to science. Nearly all these 78 endemics appear to be confined to quartzite sandstone
substrates, with just two confined to schist grassland. The main habitats for the endemics are
Ericoid scrub on quartzite outcrops or boulder slopes ("rock gardens"), quartzite rock faces
and ledges, and some quartzite grasslands, with many of the species preferring relatively
open, non-shaded habitats with some bare soil. Such habitats are widely dispersed across the
whole massif and on both sides of the border.
Chimanimani Mountains: Botany & Conservation, page 6
IUCN Red List assessments of 66 endemic or near-endemic taxa were carried out, with 27
considered to be threatened (21 Vulnerable, 5 Endangered, 1 Critically Endangered) and 34 of
the remainder being Least Concern, with 1 Near-threatened and 4 Data Deficient. However, a
few of the threatened taxa are known only from lower altitudes in or adjacent to the
threatened moist forests on the footslopes, or are near-endemics found also on nearby farms
or Forest Land.
Unlike on the forested eastern footslopes where clearance for agriculture is common, followed
by frequent fires and subsequent invasion by the introduced shrub Vernonanthura, habitat loss
does not appear to be a major threat to the endemic flora of the Chimanimani Mountains. In
general, populations of montane endemic species were found to be relatively intact and
healthy in the core zone of the TFCA; they are less under threat than had at first been feared.
This is primarily because rocky sandstone habitats are little used by the small-scale miners
and mining activities are mostly confined to streams and stream margins, habitats that are not
significant for the range-restricted species. However, gold mining activities are having very
deleterious effects on aquatic ecosystems, particularly on flow regimes and probably on
populations of aquatic invertebrates and vertebrates. There also appears to have been a
significant reduction in the populations of larger mammals on the grasslands, populations for
which the mountains were known in the past.
Owing to the large numbers of people now living or passing through the mountains, it is
assumed there has been an increase in the incidence of fire, but the impacts of fire and fire
frequency are still not clear. It is probable that Ericoid scrub and the margins of small patches
of moist Afromontane forest are being affected, but to what extent is not known.
Ten recommendations are given concerning management actions and possible further
research. The major management issue for the core zone of the mountains is to control the
numbers of small-scale miners active there, while on the lower slopes and in the buffer zone
the main issues are to control the increase in forest clearing for agriculture, wildfires and the
spread of the invasive Vernonanthura.
Chimanimani Mountains: Botany & Conservation, page 7
1. INTRODUCTION
The Chimanimani Mountains, lying on the border between Zimbabwe and Mozambique at
around 19o50' South, have been recognised as an important area for plant biodiversity for at
least 50 years (Wild 1964, van Wyk & Smith 2001, Mapaura 2002). In addition, the
mountains have been recognised as one of the main Key Biodiversity Areas (KBAs) in the
Eastern Afromontane Hotspot (Eastern Afromontane Ecosystem Profile, CEPF 2012: 174).
However, although the smaller Zimbabwe portion of this montane massif has been relatively
well studied, much less is known about the Mozambique portion. This disparity has become
more relevant over the last 15 years as there has been a significant influx of illegal small-scale
gold miners (gariemperos) and significant environmental damage, particularly on the
Mozambique side. It was not known what impact this might be having on plant diversity, in
particular on the more than 75 endemic species.
With this conservation issue in mind, an application was made to the Critical Ecosystem
Partnership Fund (CEPF) by the Royal Botanic Gardens, Kew (Kew) and the Mozambican
NGO the Micaia Foundation (Micaia) for a grant to both address the comparative imbalance
in botanical knowledge across the border and to look at the impacts on plants of small-scale
artisanal mining. Other partners in this project included the National Herbarium at the
Instituto de Investigação Agrária de Moçambique (IIAM) in Maputo and the National
Herbarium (SRGH) in Harare.
After discussions, the agreed grant was divided into two components, one dealing with
botanical and conservation field studies in the montane parts in Mozambique (carried out
primarily by Kew) and an awareness-raising and management-orientated component to be
carried out by Micaia. This report covers just the first component (for the Micaia-led
component see www.cepf.net/SiteCollectionDocuments/eastern_afromontane/FinalReport-
MICAIA-62603.pdf), although activities in most cases were carried out jointly. The stated
activities were:
a) Produce a clear and detailed assessment of the conservation threats to the range-restricted
or threatened species and their habitats, with particular reference to the threats arising from
artisanal mining.
b) Produce a series of preliminary conservation assessments for the recorded endemic/near-
endemic or threatened species for submission to the IUCN Red List Authority, further
justifying the importance of the KBA [Key Biodiversity Area].
c) Produce a set of evidence-based recommendations for presentation to the Mozambique and
TFCA [Trans-Frontier Conservation Area] authorities on appropriate conservation actions
that could be taken.
d) Collect some hundreds of fully labelled and georeferenced plant specimens collected for
the Maputo, Kew and Harare herbaria, which will also provide a basis for a preliminary
listing of species and locations for future targeted seed collection under a joint
Micaia/community/Kew Millennium Seed Bank project aimed at ex situ propagation and
conservation.
e) Produce a comprehensive report documenting all findings, which will be presented at a
joint workshop held with Birdlife Zimbabwe to assist in dissemination of findings and
conclusions.
Chimanimani Mountains: Botany & Conservation, page 8
f) As a result of training from RGB Kew, staff members from IIAM, Harare Herbarium and
Micaia will be more capable of conducting botanical surveys.
Included among the outputs of the botanical component were (a) increased and better-
documented knowledge of the endemic and range-restricted plant species found in the
Chimanimani Mountains, (b) possible new species identified, (c) IUCN Red List assessments
carried out for the majority of these species of interest, (d) a clearer identification and
justification for its status as an Important Plant Area (IPA) and Key Biodiversity Area (KBA),
and (e) an assessment of the conservation threats posed by artisanal mining to those species.
After preliminary documentation work and selection of the species of interest, herbarium
specimens of these species at both the Kew Herbarium (K) and National Herbarium in Harare
(SRGH) were databased and georeferenced where possible. Owing to historical collecting
patterns, hardly any specimens of these species were thought to be in the National Herbarium
at IIAM (LMA) or the University Herbarium (LMU) in Maputo. Three fieldtrips were
undertaken by staff from Kew, Micaia, the Maputo Herbarium, the Harare Herbarium and
other botanists (Meise, Belgium and the Natural History Museum, London) in April 2014,
September 2014 and April/May 2016, resulting in over 1000 specimens being collected and
identified. During fieldwork a number of artisanal mining areas where extraction of gold from
streams and rivers was taking place were visited to determine impacts.
This report covers these activities and presents the main findings and recommendations, along
with a background to the area, its botany, vegetation and ecology.
Along the base of Mt Binga, N Chimanimani [BW]
Chimanimani Mountains: Botany & Conservation, page 9
2. DESCRIPTION OF THE AREA
The Chimanimani Mountains ‒ the name is said to be derived from a Ndau word
T'shimanimani meaning a gorge or gap (Dutton & Dutton 1975) ‒ form part of the
Mozambique‒Zimbabwe border between 19o36' and 20o04' South. They extend for around 50
km north to south and are about 20 km wide at their widest (Figures 2.1, 2.2, 2.3). Of the
rocky montane area ‒ roughly defined as that area above 500‒700 m altitude in the south and
east and 1000‒1200 m in the west and north ‒ the largest portion, perhaps three-quarters, lies
in Mozambique.
The mountains characteristically comprise quartzite or white sandstone crags, interspersed
with more gentle grasslands, forming a plateau that slopes eastwards into Mozambique. An
accurate figure for the total extent of the massif is not available, but from polygons drawn
using Google Earth Pro an estimate of the area on both sides of the border above where the
land starts to rise steeply from the surrounding pediment and hills was made (Timberlake,
unpublished). This was around 530 km2, of which 380 km2 is underlain by quartzite and 150
km2 comprises schist grassland.
The principal features of the Chimanimani Mountains, the main montane massif covering just
the Core Zone of the Trans-Frontier Conservation Area (TFCA) ‒ the Chimanimani National
Park in Zimbabwe and the Reserva Nacional de Chimanimani in Mozambique ‒ are given
below. Other upland areas included within the broader TFCA (i.e. the Buffer Zone) ‒ such as
Rotanda and Tsetserra in Mozambique, the Mussapa Gap, Forest Land, Musapa Mountain, the
Himalayas and Banti Forest in Zimbabwe ‒ are not specifically covered here. Details on
vegetation are given later in Section 5.2 and previous studies are summarised in Section 4.
2.1 Administration
Administratively, in Zimbabwe the Chimanimani Mountains fall entirely in Chimanimani
District, with the District Administration based in the small farming and tourism town of
Chimanimani (formerly Melsetter) around 20 km away to the west at an altitude of 1200 m.
The entire montane area is protected as the Chimanimani National Park. Apart from some
former commercial farms lying west of the Haroni River but adjacent to the National Park,
there are a number of Forest Areas (Martin I & II, Lionhills, Chisengu, Tarka and Glencoe
Forest Lands) administered by the parastatal Zimbabwe Forestry Commission. These are used
primarily for commercial plantation of Pinus species, although they do contain some pockets
of natural vegetation. At the far south of the mountains lies the Rusitu Valley (the Rusitu
River becomes the Rio Lucite when it enters Mozambique) at an altitude of 320 m in which
there are two Botanic Reserves ‒ Rusitu and Haroni ‒ protecting lowland forest, as well as the
lowland Mukurupini forest at the southern end of the National Park.
The Mozambique portion of the mountains lies in Sussundenga District of Manica Province,
with the District Administration at the small town of Sussundenga some 40 km to the north-
east. In the lowlands (150‒350 m altitude), on the footslopes and pediplain to the east of the
mountains, numerous communities engaging in subsistence agriculture are found, such as at
Maronga, Zomba and Mahate. People have been living in these lowland areas for many years,
although probably at a lower population density than at present (Hughes 2006, Bannermann
2010). Many people moved out during the civil war from 1978 to 1992, but have since moved
back. Expansion of clearance for agriculture has been especially noticeable over the last 10
years. Natural resource management, conservation and livelihood issues arising from this
expansion are being addressed, amongst others, by Micaia and RBG Kew through a Darwin
Initiative project (Balancing Conservation and Livelihoods in the Chimanimani Forest Belt,
Mozambique; see Timberlake et al. 2016).
Chimanimani Mountains: Botany & Conservation, page 10
Figure 2.1. Location of the Chimanimani Mountains on the Mozambique‒Zimbabwe border.
Figure 2.2. Detailed Google Earth image of the Chimanimani massif. The quartzite rock appears paler.
Chimanimani Mountains: Botany & Conservation, page 11
The Chimanimani Transfrontier Conservation Area, which should be a jointly-managed
conservation area straddling the international border, includes a core non-use zone and a
buffer zone in which habitation and natural resource use is allowed. This is discussed later
(Section 3.4).
2.2 Physical Features and Geomorphology
The Chimanimani Mountains are a massif rising, often very steeply, from the Central African
plateau. There are three north‒south ridges ‒ the first in Zimbabwe in the west is fault-
determined with the Haroni River running south along its western side, the second forms a
series of high points (e.g. Mts Messurussero, Dombe, Mawenje, BB71a) along the
international border, while the third is much shorter and lies at the far eastern edge of the
massif above Zomba. In between, the massif is dissected, sometimes very deeply, by north‒
south-flowing rivers such as the Bundi in Zimbabwe, and the Rio Murera (or Morera) and Rio
Mufomodzi (or Mevomodzi) in the Mozambique part.
The majority of the massif lies at an altitude of between 1000 and 1800 m (1500‒1800 m on
the Zimbabwe side, Phipps & Goodier 1962), with the highest point being Mt Binga at 2436
m altitude on the international border. Other significant peaks are Mt Peza (2152 m), Mt
Dombe (2188 m) and Mawenje or Turret Towers (2362 m) in Zimbabwe and Mt Nhamadimo
(2144 m) in Mozambique.
On the Zimbabwe side, most of the drainage is into the Bundi River, which in turn flows into
the Haroni. In the foothills at the southern end of the mountains, the by-now lowland Haroni
joins the larger east-flowing Rusitu River, which becomes the Rio Lucite on crossing the
border. The smaller Mukurupini (or Makurupini) River, flowing south off the very southern
end of the massif, forms part of the international border before it joins the Haroni (although
this is disputed by Capela (2006), who suggests the border lies more to the west along the
Haroni). Much of the centre and north of the Mozambique part of the massif is drained by the
Rio Mufomodzi, which in its lower reaches flows through the lowland forest and agricultural
areas of Zomba to join the main south-flowing Rio Mussapa east of the main massif. The
Mussapa eventually flows into the Rio Lucite and in turn into the Rio Buzi, which enters the
Indian Ocean near Beira. The Mussapa actually rises north of Chimanimani town in
Zimbabwe then crosses eastwards through the quartzite ridge at Chikukwa in the Mussapa
Gap, after which it is called Rio Mussapa Grande. From the confluence with the Mussapa
Pequeno coming from the north, it becomes the Rio Mussapa (although on some maps this is
surprisingly still called the Mussapa Pequeno). The northern footslopes of the massif in
Mozambique drain into the Rio Nyamadzi in the centre-north, which in turn flows into the
Mussapa Grande, while the east-north-east slopes drain into the Rio Mucutucu and then to the
Mussapa. There seems to be a certain amount of confusion with river names in this area, with
different "official" maps having different names.
The largest and most extensive gorges are along the middle reaches of the Rio Mufomodzi
and Rio Muerera in the south-eastern part of the massif, and also along the middle and lower
reaches of the Haroni River in Zimbabwe. These deep gorges have effectively blocked east-
west access across the area. Perhaps the three best known waterfalls are Mufomodzi or
Martins' Falls (19o46'51"S, 33o07'05"E, 1500 m) and Gossamer Falls (19o53'10"S,
33o08'41"E, 490 m) on the upper and middle Mufomodzi in Mozambique, Mukurupini Falls
(19o59'59"S, 33o01'55"E, 870 m) on the upper Mukurupini in the far south, and Ragon Falls
(19°52'02"S, 33°01'37"E, 1226 m) where the Bundi falls off the massif towards the Haroni in
Zimbabwe.
Chimanimani Mountains: Botany & Conservation, page 12
Figure 2.3. Map of the Chimanimani Mountains (from Goodier & Phipps 1961, Kirkia).
The least rugged part of the massif ‒ and thus the area that has been best explored botanically
‒ is the central northern part, extending across the border from the Bundi Valley and Southern
Lakes area of Zimbabwe across to Mt Nhamadimo (or Nhamudima), Martin's Falls and the
Ma-Esese areas in Mozambique, as well as the readily-accessible Corner in Zimbabwe and
the Mussapa Gap (see Annex 2).
Erosion Surfaces
The landforms here are undoubtedly ancient. Some have suggested that the upland and
montane areas along the Zimbabwe‒Mozambique border, and further north into Malawi,
represent the ancient vestiges (Permo-Triassic period c.250 mya) of a Trans-Gondwana
Mountain Belt (Key, Cotterill & Moore 2015). However, the first reference to ancient
geomorphology was by King (1962), who suggested that the highest part of the Nyanga and
Chimanimani mountains forms part of Gondwanaland (Jurassic) or post-Gondwanaland
(Cretaceous) land surface.
Chimanimani Mountains: Botany & Conservation, page 13
Based on this, Lister (1987), in her study of the erosion surfaces of Zimbabwe, suggested that
the peaks of the northern Chimanimani above 2200 m (e.g. Dombe, Binga, Mawenje)
represent a residual Gondwanaland erosion surface dating from 186‒146 million years ago
(mya) (mid to end Jurassic), whereas the saddles in between and much of the remainder of the
northern Chimanimani above 1970 m represent the post-Gondwanaland erosion surface (early
Cretaceous, 145‒100 mya). Below about 1750 m is the African erosion surface (mid-
Cretaceous to end of Oligocene, 100‒23 mya), while valleys such as the lower Bundi at
around 1450 m represent the post-African surface (Miocene, 23‒5 mya), which also covers
much of the central parts of Zimbabwe. Wild (1964, quoting G. Bond pers. comm.) gives a
somewhat shorter period and states that it is believed the quartzites would have been exposed
more-or-less as they are today since at least the beginning of the Tertiary (around 65 mya).
Caves
An interesting aspect of the high Chimanimani Mountains is the presence of deep caves, very
different in formation and structure than the better-known shallow rock overhangs that people
have commonly used for shelter. After first being noted by someone from Zimbabwe's
Outward Bound Centre in the area behind Mawenje (Turret Towers), almost on the
international border, the first exploration was by a group of South African cavers and
Zimbabwe Mountain Club members in 1990 and 1991 (Howell 1990, Trueluck n.d.). This
was followed by a larger trip in 1992 by cavers from the South African Speleological
Association (Le Roux et al. 1993), and then by a larger and lengthier expedition by cavers
from South Africa and Zimbabwe in 1993 (Koliasnikoff 1994, Trueluck n.d.). Fourteen caves
were recorded, eight were explored and six were surveyed, one of which is just over the
border in Mozambique (Mozpot). The deepest shaft surveyed (Mawenge Mwena) is 305 m
deep.
All the caves have been developed by water erosion of the Frontier quartzites and seem to be
linked to doline-like features (Trueluck 1994a). As far as is known, these caves have not been
explored since, probably as they require specialist equipment and skills.
As regards biodiversity, no troglomorphic invertebrates were found inside them (Trueluck
1994b), just animals that had been washed in. A few insectivorous bats were seen, but mostly
in the upper reaches and in crevices.
2.3 Geology & Soils
It is geology that determines not only the presence of so many endemic plant species, but also
their distribution across the mountains (Phipps & Goodier 1962). The great majority of the
endemics are restricted to soils derived from the very nutrient-poor quartzitic sandstones.
Figure 2.4. Chimanimani vista with Dissotis pulchra
[BW]
Figure 2.5. View over upper Mufomodzi valley
[JT]
Chimanimani Mountains: Botany & Conservation, page 14
The rocks of the Chimanimani area are primarily of the Frontier Series of the Umkondo
Group, sediments over 1785 million years old dating from the later Precambrian period and
which lie on the older Archaean Basement Complex (Watson 1969). These ancient sediments
were deposited in a basin lying mainly in Mozambique, and have since been extensively
overfolded and piled up against younger rocks, possibly by tectonic events 400‒650 million
years ago (Stagman 1978). They have also undergone low-grade metamorphosis.
The Chimanimani massif itself comprises four groups of sediments originally laid down in
shallow water environments ‒ in decreasing order of age, the Lower Quartzites, the Lower
Chlorite Schists, the Upper Quartzites, and the Upper Chlorite Schists. These have been
described in detail for the Zimbabwe portion by both Watson (1969) and Stagman (1978).
The northern half of the Chimanimani comprises Lower and Upper Quartzites, with sandy and
often iron-rich and micaceous Lower Chlorite Schists in between. Rocks are folded into a N‒
S synclinal structure, complicated by over-folding (klippen), multiple thrusting and facies
changes. The highest ridge along the Mozambique‒Zimbabwe border (Mt Dombe, Mt Binga
and Mawenje to south of The Saddle) consists of Lower Quartzites, whereas the Upper
Quartzites predominate south of this. Although it is not explicit in Watson (1969), one can
assume that the southern parts of the Chimanimani massif, and possibly some of the eastern
parts in Mozambique, consist of Upper Quartzites. According to the Mozambique geological
map (ING 1987), much of the Chimanimani massif comprises quartzites (CGFQ), with
significant extents of mica-schists with aluminium and quartzite (CGF) to the east; these
overlie older Precambrian quartzitic gneiss (CGN) on the lower slopes.
The quartzites, or metamorphosed sandstones, are generally fine to medium-grained (Watson
1969), granular (occasionally termed "sugary"), massive and whitish, appearing at times like a
medium-textured sandstone. There is a well-developed system of joints and fractures that
generate distinctive weathering features. On some vertical faces there is an unusual
weathering phenomenon with the formation of honeycomb-like structures (tafoni, Lister
1987), sometimes passing right through a boulder. Quartzite landscapes are often rugged and
full of boulders, while the schists on the other hand are sandy, chloritic and micaceous meta-
sediments. Landforms derived from the latter generally show undulating grassy slopes with
reddish-brown soils, for example the gentler slopes of Mt Peza north of Mt Binga and much
of the Bundi Valley in Zimbabwe.
Soils
Soils found on the Chimanimani massif are primarily determined by the underlying geology, a
link that is also true for much of the surrounding lowlands, although depositional soils
become significant on the eastern footslopes. As with so much of the geology and ecology,
they are better documented from the Zimbabwe side.
As mentioned above, the main crags of the Chimanimani Mountains consist of highly-folded
quartzites and metamorphosed sandstones of the Frontier Series interspersed with more-
readily eroding schists. The latter give rise to more rounded, less rocky, grassy hills, whereas
the quartzites give rise to heavily-weathered nutrient-poor soils.
Phipps & Goodier (1962) describe four main soil types in the montane areas: (i) red soils
derived from schist, (ii) white sandy soils derived from quartzite, (iii) alluvial soils fringing
larger streams and small rivers, and (iv) forest soils which are humus-rich and underlie the
moist forest. These are briefly discussed below. Detailed analyses of some soils are given in
Phipps & Goodier (1962, Table 4).
Soils derived from schists (not from the Umkondo sediments as stated in Wild 1964) are
orange-red in colour, and range from shallow and rocky on steeper slopes to more mature
Chimanimani Mountains: Botany & Conservation, page 15
soils around 45‒60 cm deep on gently undulating grassland slopes. They have a higher water-
retaining capacity than quartzite soils owing to their 20% clay content, but are low in
exchangeable bases (less than 1 m.e.%), presumably a result of leaching over a long period
(Phipps & Goodier 1962). However, compared to quartzite soils they are significantly higher
in available phosphorous (50 ppm).
White quartzite soils, on the other hand, have less than 1% clay hence a low water-retention
capacity, and less than 1.5 m.e. % exchangeable bases (Phipps & Goodier 1962). They range
in depth from almost zero (outcrops of rock, sometimes with a lithophytic flora or just small
pockets of vegetation) to soils up to 30 cm deep in flattish areas. However, the somewhat
deeper soils here are very sandy with little humus, except where a sort of bog vegetation
occurs. Apart from colour and texture, their main differentiating characteristic is that they are
also very low in available phosphorous (less than 5 ppm), which is thought to be a significant
factor in promoting sclerophylly (i.e. with evergreen leathery leaves; Wild 1964).
Alluvial soils flanking larger watercourses can be derived from either quartzite or schist, but
tend to have similar properties to quartzite soil. Very acidic bog vegetation is also found.
Forest soils are derived from either quartzite or schist and their properties are a result of the
vegetation cover, with a significant well-drained humus layer being formed. Forests are nearly
always found on slopes or among rock outcrops.
Coupled with changes in altitude, it is this distinct difference in soil properties that gives rise
to the range of vegetation types found across the mountains.
2.4 Climate
Rising out of the low Mozambique plain the Chimanimani massif intercepts moisture-laden air
coming from the Indian Ocean, resulting in high levels of orographic precipitation. Being in the
rain shadow from the prevailing south-easterly air flows the Zimbabwe side is drier, but there
do not seem to be any specific data on this.
The main summer rainy season is from November to late March or April, but on the high
mountains rain can occur throughout the year; the dry season is not as marked here as it is on
the central plateau below. No accurate figures exist for the mountain and escarpments, but by
extrapolating from similar areas in Zimbabwe rainfall is estimated to be between 1500 and
2000 mm/year. Phipps & Goodier (1962) state that further north in Zimbabwe, Mt Nyangani
(Inyangani B) receives 2997.2 mm/year; the figure for the higher east-facing parts of Mt
Binga and Mawenje may be similar, indicating that the high mountain peaks can receive
almost 3 metres of rainfall each year. For Inyangani B the wettest month was January with
639.3 mm, while the driest was July with 37.8 mm.
The nearest available rainfall figures in Zimbabwe (Agritex/FAO 1989; Table 2.1) are for
Chimanimani District Administration (Chimanimani town in Zimbabwe, 1520 m) and for
Chisengu Forest Area (1340 m), which lies a bit further south but still within the rain shadow
of the mountains. The figures are 1074 and 1406 mm/year, respectively. However, Phipps &
Goodier (1962) give an average figure for Martin Forest on the western (rain shadow) side of
the Chimanimani as 1072.3 mm/year (Table 2.1).
There are no rainfall data available for the Mozambique portion of the Chimanimani
Mountains.
Chimanimani Mountains: Botany & Conservation, page 16
A major feature of the mountains is the frequency of mist ('guti') and overcast days during the
dry season (Phipps & Goodier 1962), which effectively decreases the length and stress of the
dry season. In the wet season a "mist belt" is often present at 1500 m and above.
Table 2.1. Rainfall figures for Chimanimani area, Zimbabwe (from Agritex/FAO 1989 1;
Phipps & Goodier 1962 2).
Chimanimani1 Chisengu1 Chipinge2 Martin Forest2
Average rainfall
(mm/yr)
1073.7 1406.1 1060.8 1072.3
wettest month (mm) 215.8 (Dec) 246.2 (Dec) 212.8 (Jan) 234.0 (Jan)
driest month (mm) 16.3 (July) 37.0 (July) 18.3 (Sept) 10.8 (July)
wettest year (mm/yr) 1704.6 (1954/5) 1766.6 (1987/8)
driest year (mm/yr) 554.3 (1963/4) 960.3 (1986/7)
period 32 years (1952-
1984)
26 years (1955-
1972, 1976-1987)
As regards temperature, the climate is generally considered humid tropical to temperate
(Ghiurghi, Dondeyne & Bannermann 2010, Annex 2) with average temperatures across the
eastern Chimanimani area of 23‒25oC in January and 17‒19oC in July. Mean average
temperature varies from 22oC in the south-eastern lowlands to less than 18
oC in the high
mountains. Frost is frequent on the plateau above 1500 m, but absent from the low valleys.
Phipps & Goodier (1962) give tabulated mean maximum and minimum temperature figures for
a number of localities in Eastern Zimbabwe, showing that frost is not uncommon from May to
August.
Waterfall (Tucker's Falls) with Afromontane forest at base, N Chimanimani [JT]
Chimanimani Mountains: Botany & Conservation, page 17
3. HISTORICAL BACKGROUND
3.1 Prehistory & Archaeology
The archaeological record for the Chimanimani area is poor compared to many other places in
the region. There are records from caves up in the mountains (1000‒1700 m) of what are
probably Late Stone Age sites with Bushman (Khoisan or San hunter-gatherers) rock
paintings. On the Zimbabwe side, in the northern extension called The Corner, at least six
rock art sites were documented (at c.1400 m altitude) during a Rhodesian Schools expedition
(Hughes & Gunns 1973), and a local informant (Chief Nyahedze) also mentioned rock
paintings in this broad area. Dutton & Dutton (1975) mention cave paintings, but it is not
clear if they were found on the Mozambique side. Rob Burrett (pers. comm., June 2014), a
Zimbabwean archaeologist, mentions rock art at Ragon Falls, St Georges Cave, Red Cave and
at a small site towards the Mussapa Gap, most of these being in Zimbabwe, while Doug van
de Ruit (pers. comm., November 2014) mentions rock paintings and an old iron smelting
facility, but without locality. Garlake (1995), in his study of the prehistoric art of Zimbabwe,
does not mention any sites in the Chimanimani area, but this may be because any such
paintings there weathered away more rapidly compared to those on the less permeable granite
or syenite of the other mountains, such as the Matopos and Nyanga. Bannermann (2010), in
his detailed account of the history of the Mozambique side of the Chimanimani area,
mentions "many of these [rock art] sites in the project area" including on the high
Chimanimani plateau, Chikukwa Ferreira (The Corner) and (outside our study area) on rock
outcrops between Mavita and Rotanda.
It is probable that the Bushman and subsequent peoples generally did not live on the high
mountain plateau but only in lower, more sheltered areas, perhaps just going into the
mountains for hunting. There is also the issue that in recent years any rock art may well have
been destroyed or severely damaged by fires and the use of caves for shelter by gold-panners.
There seems to be rather little in the way of more recent archaeological evidence from the
Chimanimani massif, although more has been found on the footslopes and at lower altitudes.
This is surprising given the number of archaeological finds on similar upland areas in Nyanga
(Soper 2002) and Mt Mulanje and the Nyika plateau in Malawi (Rangeley 1960, McCracken
2006), ranging from iron smelting to hut sites to field boundaries. In his extensive study of
agriculture and archaeology in the Nyanga area, Soper (2002) does not mention any similar
historic activity in the Chimanimani area, apart from a cryptic reference by Beach (in Soper,
p.202) to "...reported terraces and fortifications from the eastern [Mozambique] side of the
Chimanimani range..." based on comments in Octávio Roza de Oliviera (1973).
What is not clear is if this lack of archaeological record and evidence is due to an absolute
lack or because little effort has gone into looking for remains up on the massif. Or perhaps the
terrain is just not conducive to preserving or locating archaeological sites. Given the harsh
climate (high rainfall and low temperatures) and the great paucity of soil nutrients compared
to the dolerite and granite soils of Nyanga and Matopos, the area perhaps had too low a
carrying capacity for human settlement. Agriculture of any type is unlikely to have been
feasible on quartzite soils above 1500 m and the carrying capacity for any grazing animals,
domestic or wild, would also be very low.
3.2 Colonial & Recent History
As might be expected, the more recent and colonial history of the Chimanimani area is much
better documented from the Zimbabwe side than from the Mozambican, although
Bannermann (2010) gives a detailed and comprehensive account for Mozambique. Much of
this is based on Appendix 3 of the late Richard Bell's management plan, a document that was
Chimanimani Mountains: Botany & Conservation, page 18
almost lost (see Ghiurghi et al. 2010a: 7). Bannermann's account covers the community
histories on that side from the Early Iron Age (Chifumbazi), through the trading era of the
Mutapa and Zimbabwe states around 1300‒1500 to the early colonial Portuguese traders, the
Portuguese attempts to supress Chief Ngungunhane (or Ngungunyana) and the establishment
of the colonial trading companies. It ends up discussing the impact on communities of both
the liberation struggle that ended with a Frelimo government in 1975 and the subsequent civil
war with Renamo. Some relevant details from his account are given below. The communities
discussed by Bannermann are mostly those living along the rivers and in the forests on the
low-altitude areas (150‒350 m altitude); the mountains themselves seem to have played little
role in this history and have had minimal economic importance.
A detailed account of the resource politics around the Moribane Forest Reserve on the
Chimanimani foothills is given in Schafer & Bell (2002), while Hughes (2006) covers the
political and resource history on both sides of the border, with particular reference to
communities living at the southern end of the mountains in Maronga.
According to Bannermann's (2010) account (much of what is written below is based on this),
there was a significant trade in gold from the hinterland to the coast from around 500 AD
onwards, which later focused on the Great Zimbabwe area of the Monomatapa Kingdom.
Gold was also mined in these early days in areas to the north of Chimanimani, such as
Bandire east of Rotanda and at Penhalonga just north of present-day Mutare in Zimbabwe.
Trade routes, probably through the Mussapa Gap and further south along the Rusitu/Lucite
valley, went to Sofala on the Buzi estuary, which was the chief seaport for the Monomatapa
Kingdom.
A more modern history of the Mozambique side probably starts with the establishment of an
administrative zone (Circunscrição de Moribane) based at Mavita in 1894. This was initially
administered by Sr. Campos Santos, and later by José Luis Ferreira, after whom the Moribane
centre was named around 1900. It was Ferreira who, in 1899‒1907, built a road from
Macequece (now Manica town) to Mavita and on to Melsetter (now Chimanimani town in
Zimbabwe), probably through the Mussapa Gap, and it was this road that probably did most
to open up the area economically.
Much of the interest in the area seems to have focussed on the possibilities of wild rubber (sap
from lianas of Landolphia) and timber in the forests along the Chimanimani foothills. The
Companhia de Moçambique encouraged the planting of rubber vines, and the Moribane
Company established plantations along the banks of the Mussapa and Mutucutu rivers in
Zomba. Forced labour with very low wages was often imposed. A Frenchman, Paul Bindé,
also established a Landolphia plantation near the Rio Lucite in Maronga in the far south
around 1907. In addition, small quantities of bananas, papaya and citrus were planted.
However, in the 1940s the Companhia de Moçambique, which was effectively controlling this
part of Mozambique, collapsed and its administrative functions were taken over by the
Portuguese state. Soon after, administrative posts were established in Dombe and Mavita, and
then, in the 1940s to 1950s, forestry concessions and sawmills were established in the
Moribane‒Mavita area to exploit the lowland moist forest and dense woodland. The Forest
reserves of Maronga, Zomba and Moribane were gazetted in July 1953 (see Gomes e Sousa
1968, Timberlake et al. 2016), but it is not clear if these reserves were protect places from
exploitation or just to control concessions and extraction. Later, in the 1960s, cotton and
wheat production was encouraged in the lowland areas, and a Sr. Carvalho established a large
farm on Tsetserra, one of the northern, non-quartzite plateaux.
Chimanimani Mountains: Botany & Conservation, page 19
During the Zimbabwe Independence struggle, the Rhodesian forces not only mined some of
the montane passes but also attempted to disrupt communications in border areas to try and
stop incursions by ZANU fighters. For example, in 1976 to 1980 the Rhodesians destroyed
large bridges across the Lucite and Mussapa rivers, which have only recently been rebuilt. At
around the same time, the newly-independent Frelimo government established communal
villages (aldeias communais) at Rotanda and Dombe, and expected most people to live there
rather than scattered throughout the agricultural areas. This caused dislocation of rural
populations, which in turn effectively assisted destabilization by Renamo. Economic activity
in the area stagnated and this has had an effect on subsequent land use in the lowland areas.
The civil war between Renamo and Frelimo ended in 1992 with a peace accord, but even
since then there have been some disturbances.
On the Zimbabwe side, modern colonial history started with treks in 1892 from what was the
Afrikaner (Boer) Orange Free State in South Africa. The two Moodie treks, led by brothers
Dunbar and Thomas Moodie, were encouraged by Leander Star Jameson on behalf of Cecil
Rhodes' British South Africa Company to move into the territory called Gazaland to the west
and south of the Chimanimani Mountains. This was to act as a bulwark against possible
Portuguese expansion and to ensure the Company's dominance in the area. At this time the
international border was not clear and each side, the Portuguese authorities and the British
South Africa Company, was each trying to claim more territory ‒ "Ultimately, they [the treks]
kept Portugal south of the Rusitu River and east of the Chimanimani mountains" (Hughes
2001). Active possession through settlement and farming was considered to be "nine-tenths of
the law". The Portuguese authorities were also struggling at this time with insurrection led by
great Shangaan chief Ngungunyana (Mullin 1994).
The treks (there were nine in total; Hughes 2006) ended up in the area around the moderately
fertile and well-watered headwaters of the Haroni and Nyahode rivers which they called
Melsetter. Each settler could stake out land under 'Pioneer Title' granted by the British South
Africa Company, and use the local native population living there for labour. Many farms were
established at this time on higher ground, ranging from Cashel in the north southwards to
what is now Chipinge and west to Gwendingwe. However, there seem to be few references
(although see below) to land being claimed in the Chimanimani Mountains themselves,
probably as it was recognised to be of very low agricultural potential. The history and
subsequent development of the Melsetter‒Chipinga area and its agriculture is given in detail
by Sinclair (1971).
It is not known if there was any Portuguese settlement in the western area before this.
However, one interesting but not well-documented find was of the ruins of a small square
building in Mutema Communal Land some 40 km southwest of present-day Chimanimani,
where a small-bore cannon, possibly made in Scotland around 1668, was found (Martin
Sanderson, pers. comm., May 2016). This may have been a small fort on the gold-trading
route from the Great Zimbabwe area to Sofala on the Buzi estuary, and indicates the
importance of early trade to the coast through the Chimanimani area.
The international boundary itself, following the watershed between east- and west-flowing
rivers on the second, western ridge (except in the Mussapa Gap area; Bannermann 2010), was
only settled by the Anglo-Portuguese Boundary Delimitation Commission in the late 1890s.
The concrete beacons were put in place by R.S.T. Fairbridge and A.E. Wayland in the 1920s
(Sinclair 1971). However, this artificially split the Ndau-speaking communities which lived
on both sides of this new border, but probably did not act as a barrier to their regular
movement.
Chimanimani Mountains: Botany & Conservation, page 20
Hughes (2001) describes how the settlers in Melsetter used the native African inhabitants as
somewhat involuntary labour. This was not considered a good state of affairs by some of the
colonial authorities, and in 1896 the Native Commissioner J.D. Hulley, followed by his
successor L.C. Meredith, created Native Reserves in lower-altitude areas such as Ngorima in
the Rusitu valley at the southern end of the Chimanimanis. This allowed Africans to avoid
what was effectively forced labour and maintain a certain independence. The reserves, which
remain until today in terms of their boundaries, have had an effect on subsequent agricultural
development in the Chimanimani area. Their creation has also had a significant effect on the
conservation of the lowland forests at the southern end of the mountains in Zimbabwe, and
the later creation of Botanic Reserves within them is still contentious (Hughes 2001, 2006).
On the Mozambique side of the mountains land was not "alienated" to colonial settlement
(Hughes 2006), and because land was not officially demarcated or "privatised" on the
Mozambique side there has been a marked difference in the way land is used on both sides of
the mountains ‒ what Hughes (2001) terms "cadastral politics". Biodiversity conservation
thus has to be approached differently in each country, with differing possibilities, problems
and opportunities. Some of the conservation issues relating to the lowland forests on the
Mozambique side are briefly discussed in Timberlake et al. (2016) and are not repeated here.
In Zimbabwe (then Rhodesia) there were a number of issues that faced the Melsetter
community between 1892 and Zimbabwe Independence in 1980, but the biggest appear to be
the poor state of roads, and hence great difficulties in travel as well as export of produce, and
the continual recurrence of the tick-borne cattle disease, East Coast Fever, that effectively
reduced the viability of any cattle industry. It is interesting to note from Sinclair's book (1971)
that most communication and trade seems to have been with the large Rhodesian town of
Umtali (now Mutare), some 100 km to the north (and to a lesser extent Chipinge), and very
little with Mozambique. There was a Portuguese border post not far from Melsetter at the
Mussapa Gap, manned by a Portuguese official who used to visit Melsetter for recreation and
possibly supplies. However, Sinclair's book hardly refers to any use of the crossing by the
settlers, although the Gap crossing was probably extensively used by the local African
population who formed (and still form) a community on both sides of the border.
After the Second World War, there were renewed attempts to develop the Melsetter area and a
number of important initiatives started up. Firstly, much land was bought by large companies
such as the Rhodesian Wattle Company, as well as the Forestry Department (from 1954 this
became the parastatal Forestry Commission), to create extensive plantations of pine (mostly
Pinus patula), wattle (Acacia mearnsii) and gums (Eucalyptus species). Martin Forest Land
was developed in 1945, followed in the early 1950s by Tarka, Chisengu, Lionhills and
Glencoe. Later, secondary industries such as sawmills were established. It would seem that
the conservation status of many of the Umkondo sandstone endemics (see Section 6.3) may
have deteriorated at this time. Much of their somewhat open scrubby habitat would have been
cleared for plantations. A lot of what had been small-scale mixed agricultural holdings with
family-level production became more commercialised, roads became better and there was an
influx of people. In 1949 the government Agriculture Department suggested the future of the
District lay in small intensive mixed farming, mainly fruit and dairy, with commercial
afforestation on sloping land. Arising from this, a 1500 acre Government Agriculture
Research Station was established just outside Melsetter town on Lindlay North farm in 1949,
but this was closed down in 1965.
In this immediate post-War period the Melsetter community also tried to develop tourism in
the area, although it appears that this revolved primarily around local scenic and historical
attractions (see Sinclair 1971) rather than the Chimanimani Mountains themselves. To an
extent, the mountains were initially more of a backdrop to the Melsetter community than
Chimanimani Mountains: Botany & Conservation, page 21
forming part of their lives; Sinclair's book has just passing mention of brief visits to the Bundi
Plain over the first ridge, or of adventurous individuals such as the forester John Ball, an avid
orchid collector, who spent time exploring it. For a few years in the late 1940s a farmer from
Melsetter, Hendrik Olwage, used to take visitors up the mountains on ponies (Jane Browning,
pers. comm.). The only mention of actual use is by Gideon Martin (after whom Martin's Falls
were named, and probably also Martin Forest Land) who grazed his cattle up on the Bundi
Plain on the Zimbabwe side around 1944 (Sinclair 1971, Dutton & Dutton 1975).
Of much greater significance for the Chimanimani Mountains themselves was the gazetting of
the 20,213 acre (8186 ha) Chimanimani National Park in 1949, one of the first two declared in
Zimbabwe (the other being Hwange). However, the actual justification or reason for this is
not clear, nor is it clear if any land was compulsorily purchased although, for example, it
appears that the farm Rocklands (possibly belonging to Gideon Martin) originally extended to
a peak on the Mozambique border near what was then called Ben Nevis (now Mt Peza), as did
the farm Dunblane. From historical accounts, including of Rawdon Goodier's first visit in
1956 (Goodier 2009), it seems that Dead Cow Camp and Long Gully, situated at the end of a
motorable track near Charleswood Farm (just south of the present National Parks offices),
was the main entry point. The Mountain Hut, situated in the middle of the Bundi Plain, was
under construction in 1956, if not earlier. John Ball had a major role in its establishment (Jane
Browning, pers. comm.). All the materials for this were brought in by porters or by donkey,
and even at that early stage it was fully kitted out with paraffin stove and fridge, showers and
bunk beds. There was also a ram pump for the water supply.
During the 1950s there were many school trips up the mountains, including one for ecological
studies by a girl's school in 1957 (led by Jane Browning, pers. comm.). The Rhodesian
Schools Exploration Society (e.g. Quaile 1973, Shaw 2012), and other schools such as
Peterhouse, regularly used the mountains for trips, a use that continues to this day. Some of
the more adventurous of these early school visits also involved multi-day trips through the
Mozambique portion. Accommodation was primarily in the many caves found across the
plateau, which has given rise to some of their names (see Annex 2).
Later on, the Outward Bound centre was established in 1961 at the foot of the mountains,
north-east of Melsetter. This encouraged many more visits to the mountains by a wider range
of people, including school children getting outdoor and leadership training.
As can be seen, the trajectory of "official" development, and thus conservation, was very
different on each side of the mountain range with seemingly very little communication or
exchange between the national authorities. A situation that, hopefully, the formation of the
Trans-Frontier Conservation Area will help to address.
3.3 Conservation Situation and Initiatives
Straddling the border between Zimbabwe and Mozambique, with the border running down the
long axis and a significant length of it following the watershed, the Chimanimani Mountains
are effectively one ecological and conservation unit, and it is generally recognised that they
need to be managed as such.
On the Zimbabwe side, all of the quartzite areas and much of the remaining lowland forest
(which is very significant from a conservation viewpoint) lie within the 155 km2
Chimanimani National Park gazetted in 1949. Some additional forested and dense woodland
areas are protected in the Haroni and Rusitu Botanic Reserves adjacent to the Makurupini
Forest at the far southern end (Timberlake 1994) or as Forest Land under the control of the
parastatal Zimbabwe Forestry Commission (e.g. Martin and Tarka Forest Lands). Virtually all
Chimanimani Mountains: Botany & Conservation, page 22
the areas of conservation interest within the Chimanimani complex south of 19o39'S are
formally protected in Zimbabwe.
However, by far the larger part of the massif lies in Mozambique. There nearly all (but see
Section 7.5) the upland areas of conservation interest now lie within the Core Area of the
Chimanimani Trans-Frontier Conservation Area (TFCA, see next section), although a
significant extent of lowland forest, also of great conservation importance, lies within the
TFCA Buffer Zone where human settlement and land use is allowed (see Timberlake et al.
2016).
The TFCA Core Zone in Mozambique is called the Chimanimani National Reserve (Reserva
Nacional de Chimanimani) under recent legislation (República de Moçambique 2003), and is
meant to be a zone of total protection of certain species (rare, endemic, in decline) or fragile
ecosystems. This is a very similar status to a National Park (Parque Nacional) but where a
limited amount of human use is allowed under licence, for example utilization of natural
resources, as long as it is included in the management plan and is compatible with the
overriding protection objectives. However, it would appear that no people can live inside a
National Reserve, which seems to be equivalent to IUCN Protected Area Category II. It is not
clear why the Chimanimani massif was gazetted as a National Reserve and not a National
Park, but this may be because as well as a Core Zone there is an adjacent Buffer Zone within
the broader TFCA.
All protected areas in Mozambique now fall under the Conservation Directorate of the large
Lands, Environment and Rural Development Ministry (Administração Nacional das Áreas de
Conservação (ANAC), Ministerio da Terra, Ambiente e Desenvolvimento Rural).
The first mention of protection for the Mozambique side of the Chimanimani appears to have
been in 1953 by the Chef do Posto (administrator) in Mavita, four years after the
establishment of the National Park on the Zimbabwe side, but nothing was apparently done. A
bit later, at the 1966 AETFAT Congress, the botanist Grandvaux Barbosa listed what he
regarded as the botanical conservation priorities for Mozambique (Barbosa 1968). He
recognised that the Chimanimani Mountains contained "endemic montane vegetation" and
that a National Park, adjoining the one already in place in Zimbabwe, was required for its
conservation, and he also clearly brought out the necessity for conservation of the associated
moist forests on the footslopes, although some Forest Reserves (Maronga, Zomba, Moribane;
Gomes e Sousa 1968) were already in place. Although not strictly conservation, there was
also cooperation between the Rhodesian and Portuguese authorities in 1969 over wildfire and
veterinary controls (Bannermann 2010).
The first major initiative in this regard came from Paul Dutton (Dutton & Dutton 1973) in his
survey of the area for the colonial Veterinary Department (then the body responsible for
biodiversity conservation) just before Mozambican Independence. A map was produced
(Mapa 2) showing suggested boundaries for, initially, a core area (Zona de Vigilância) above
the 1000 m contour from near Mt Peza in the north to the Haroni/Mukurupini rivers in the
south and also including the separate Mt Mucota area just north of Mavita. This was to be
followed (second phase) by a broader area extending from the Rio Mussapa Pequeno, along
the Rio Mussapa and, avoiding areas that were heavily settled, across the pediplain west of
the Rio Mussapa to a point perhaps 15 km downstream of the Zimbabwe border along the Rio
Lucite. The proposed area included the gazetted Maronga and Zomba Forest Reserves, but not
Moribane Forest Reserve.
Chimanimani Mountains: Botany & Conservation, page 23
Independence was gained in 1975, followed later by political instability. Thus it was only in
the mid-1990s that attention appeared to return to the Chimanimani area and its conservation,
attention that was now focussed on the idea of a trans-frontier conservation area.
3.4 Chimanimani Trans-Frontier Conservation Area
The process towards the establishment of a Trans-Frontier Conservation Area (TFCA) was
initiated when Richard Bell was employed as a consultant for the National Directorate of
Forests and Wildlife and The World Bank/Global Environment Facility to draw up plans for a
conservation area to be called the Nakaedo Biosphere Reserve (Bell 1999). Unfortunately,
only incomplete documentation appears to remain on this (see Ghiurghi et al. 2010a, vol. 1,
p.7). As with Dutton & Dutton's proposal, there was to be a core conservation zone in the
highlands above 1000 m altitude with its boundary from Chimanimani camp at the source of
the Rio Mussapa Grande, along the foot of the high escarpment eastwards and southwards
past Mahate to Zomba, then along the foothills of the southern escarpment south-eastwards to
Maronga and the Haroni‒Rusitu confluence. However, unlike the initial suggestions from
Dutton and Dutton (1973), Mt Mucota was excluded whereas the uplands by The Corner and
the southern Rotanda area were to be included. This core zone was to be surrounded by an
outer zone of lower conservation priority in the surrounding mid- and lower altitude areas,
based on the rationale that the highlands were more-or-less uninhabited and had very high
aesthetic and biological values. The outer zone, by contrast, was relatively heavily settled and
of lower aesthetic value, although it contained high biodiversity value in the form of the
remaining lowland forests. However, nothing further seemed to happen at that time with this
initiative, possibly due to Richard Bell's untimely death.
Later, in the 2000s and with support from The World Bank, the whole Chimanimani area was
designated as the Chimanimani Transfrontier Conservation Area, approximately 4091 km2 in
extent (Anon. n.d.). In Mozambique this initiative received extensive support from the Peace
Parks Foundation and The World Bank from the late 1990s to around 2012. Although there
appears to have been little funding or activity on the Zimbabwe side over that period, existing
management and investment by the Zimbabwe Parks and Wildlife Authority has ensured that
the biodiversity values of the Chimanimani National Park have been retained there, even in
the face of an invasion of illegal gold miners (see Section 7).
The TFCA in Mozambique comprises a Core Zone, where no settlement or extractive use is
allowed, and a Buffer Zone (Figure 3.1). Natural resource management and some economic
activity is allowed in the Buffer Zone as long as it is felt to be not incompatible with
conservation objectives (equivalent to IUCN Category VI), and conservation issues should
form a significant part of land use and economic planning.
In Mozambique the core conservation zone was originally gazetted in 2003 as the Reserva
Nacional de Chimanimani (República de Moçambique 2003) with a calculated total extent of
640.6 km2 (area determined from digitised version of gazetted points) The boundaries have
recently been re-defined (República de Moçambique 2013 with a full list of UTM
coordinates) owing to land use and subsequent human intrusion, and now has an extent of
660.2 km2 (area calculated from gazetted points and following international border; this
differs from the 645 km2 in Anon. n.d. [GIZ]). In both cases (2003 and 2013) the perimeter is
approximately 181 km long. The main changes between 2003 and 2013 were that the new
boundary now lies 1‒5 km further away from the top edge of the massif to the south-east, the
included area around Nyabowa community in the north is somewhat more extensive, and the
northern extremity around The Corner is significantly smaller, being confined to higher
rockier ground. The Core Zone now covers land above 350 m altitude in the south and 600‒
Chimanimani Mountains: Botany & Conservation, page 24
800 m in the north. The initial intention was that the Core Zone would primarily consist of
land above 1000 m altitude (Dutton & Dutton 1976, Zona de Vigilância).
Figure 3.1. Boundaries of Chimanimani TFCA core zone (revised) and buffer zone in Mozambique
(Google Earth image). The main quartzite mountains on both sides of the border are seen from their
lighter colour.
To the south, east and north, the Core Zone is surrounded at lower altitudes by an extensive
Buffer Zone (Zona Tampão) covering a larger area of 1721.4 km2 (from digitised version of
gazetted points, minus extent of Core Zone). This extends eastwards from the Chimanimani
Chimanimani Mountains: Botany & Conservation, page 25
foothills to the Rio Lucite in the south, the Rio Mussapa in the east, and northwards to Mt
Tsetserra (see Ghiurghi, Dondeyne & Bannerman 2010a). Three Forest Reserves, originally
gazetted in 1953 for timber production ‒ Maronga, Zomba and Moribane ‒ lie in the Buffer
Zone (see Timberlake et al. 2016).
On the Zimbabwe side the TFCA core zone comprises the Chimanimani National Park (155
km2) and the small Eland Sanctuary (15 km2) above the town of Chimanimani (Anon. n.d.
[giz]). It is not clear how far the buffer zone extends or if any Forest Land is included.
Although in the recent past, cross-border collaboration within the TFCA as regards
conservation management between the two countries has been limited, this now appears to be
increasing (Muassinar 2015).
Plant collecting on steep quartzite slopes, N Chimanimani [JT].
Anthony Mapaura with Leucospermum saxosum [BW].
Chimanimani Mountains: Botany & Conservation, page 26
4. PREVIOUS BIOLOGICAL WORK
There are a large number of publications and studies on various aspects of the biology and
conservation of the Chimanimani Mountains and the associated lowlands. Most of these are
cited in the bibliography (Section 11). Below we outline the main botanical and ecological
studies, those on the montane and lowland forests (see Timberlake et al. 2016 for more on the
latter) and the phytogeographical ideas that the Chimanimanis have added to. Brief details on
what is known of vertebrate life are given in Section 4.5.
4.1 Previous Botanical and Ecological Studies
The Chimanimani area and the general Manica Highlands region along the Zimbabwe‒
Mozambique border (in Zimbabwe usually called the Eastern Highlands) have a relatively
long botanical history. Indeed, the first publication on the botany of Rhodesia (now
Zimbabwe) was that by Rendle et al. (1911) documenting Charles Swynnerton's botanical
collections from "Gazaland", the name given to what are now the districts of Chimanimani
and Chipinge in Zimbabwe and adjacent parts of Mozambique. At that time Swynnerton was
based at Gungunyana Farm, now part of Chirinda Forest (Timberlake & Shaw 1994). In this
publication a number of species and 11 new taxa were described from the Chimanimani
mountains and Mt Peza areas of Zimbabwe. A later publication by Gilliland (1938) on the
vegetation of Rhodesian Manicaland, despite its title, did not cover mountains south of
Mutare.
The resurgence of botanists visiting the Chimanimani mountains just after the Second World
War up until the mid-1950s, although resulting in many collections and in a few new species
being described, did not initially seem to produce any comprehensive picture of the botany or
ecology of the mountains, although its botanical richness and interest was obviously fairly
widely known (Jane Browning and Jim Phipps, both pers. comm.). Raymond Munch collected
there in the late 1940s, while in the early 1950s Paddy Crook and his wife Lucy Chippendall
(a South African grass specialist) collected there, as did Darrel Plowes and John Ball, the
latter particularly focussing on orchids (see Table 4.1). The high levels of endemism were
something that was surely recognised, but just not written about.
This started to change in 1956 when Rawdon Goodier, then a young tsetse ecologist based in
the Zambezi Valley, started visiting the mountains to rock climb (Goodier 2009, R. Goodier
pers. comm., April 2015). He later teamed up with Jim Phipps, a botanist working at the
Government Herbarium in Salisbury (now Harare), who had similar interests (Figure 4.1).
Over the next 3 to 4 years they paid many visits to the mountains and collected
systematically, building on the earlier extensive collections of Hiram Wild and R.C. Munch
(1944‒1950). Schools expeditions also went up the mountains (e.g. Mitchell et al. 1958,
Quaile 1973, Shaw 2012) and collected.
In 1958 Goodier and Phipps produced a preliminary checklist listing 573 species found above
1200 m in the area underlain by rocks of the Frontier System (quartzite and schist). Following
further plant collections this was revised in 1961, listing 859 taxa (Goodier & Phipps 1961).
The latter list was also republished as an appendix in Dutton & Dutton (1975) but,
surprisingly, no further similar lists have since been produced. Much later the National
Herbarium in Zimbabwe produced a list of 2182 species based on their databased collections
of all species recorded in Chimanimani District (Chapano & Mamuto 2003), but this covers
species both on and away from the mountains and with no means to distinguish between
them. However, building on the initial Goodier and Phipps checklists, a revised list is now
under preparation for the area above 1200 m (Wursten, Timberlake & Darbyshire, in prep.)
that will also incorporate recent collections under this CEPF project.
Chimanimani Mountains: Botany & Conservation, page 27
Figure 4.1. Jim Phipps (L) and Rawdon Goodier
(R), Harare 1961 [R. Goodier].
On the ecological side, Goodier & Phipps produced
a sketched vegetation map of the Chimanimani
Mountains, mostly of Zimbabwe (Goodier & Phipps
1962), but their main work was an account of the
ecology of the mountains in the Journal of Ecology
(Phipps & Goodier 1962). This very detailed and
comprehensive study discussed not just the
mountains and their endemic flora, but also the
possible reasons for it, and linked geology and soils
to plant distribution. Nothing published since has
superseded it.
Following on from Phipps & Goodier, the last significant studies on Chimanimani botany
until recently were by Wild (1964, 1968) on the endemic flora and phytogeography. In his
1964 paper, based upon the collections of Goodier and Phipps as well as his own, Wild
described a number of new species (some in conjunction with other experts) and discussed the
significance and origins of this endemism "hot spot". Later he looked at the phytogeography
of south-central Africa as a whole (Wild 1968) and the place of the Chimanimanis in it (see
Section 4.4).
In the later 1960s through to Zimbabwean Independence in 1980, there were quite a few
collectors visiting the mountains (James Whellan, Keith Coates Palgrave, Oliver West,
Rosemary Grosvenor, Bob Drummond, Rob Kelly, Brian Simon, Tom Müller and Blake
Goldsmith), and some expeditions (Peterhouse in 1971, Quaile 1973), but documentation is
limited. Also at this time, interest started to be focussed on the lowland forests of the Haroni‒
Rusitu area (see below), possibly as access there became better.
From a conservation viewpoint, the area in Zimbabwe was well-protected and relatively easy
of access from Chimanimani town. However this was not the case in Mozambique. Paul and
Elizabeth Dutton made an extensive study of the larger Mozambique portion of the mountains
and proposed the formation of a protected area (see Section 3.3). Their published document
(Dutton & Dutton 1973) included much information on its natural history, including plants
(much of it taken from Phipps & Goodier's publications). Written in Portuguese, this must
have been a very useful document that brought attention from the Mozambique authorities to
the area. It also appears to be the first publication to provide much information on the
vertebrates, as well as being the first to focus on the Mozambique side.
In 2002, under the Sabonet (Southern African Botanical Diversity Network) project, all
species that were believed to be endemic (or near-endemic) to Zimbabwe were listed and
provisionally assessed at a national level in a workshop (Golding 2002, Mapaura &
Timberlake 2002). This was later written up more systematically by Mapaura (2002). In the
Sabonet publication, most of the Chimanimani quartzite endemic taxa were nationally
assessed as Lower Risk near-threatened as they were believed to be found over a
comparatively large area, albeit very restricted, well-protected within a National Park and
with no real threats. If a species was thought to be restricted to a specific peak or a more
marginal habitat, it was considered Vulnerable D2. In a similar exercise for Mozambique in
the same publication, a lower number of the known Chimanimani endemics were assessed as
Chimanimani Mountains: Botany & Conservation, page 28
Vulnerable (VU D2), with a few as Data Deficient (Izidine & Bandiera 2002). There did not
appear to be any attempt to harmonise either listed species or conservation assessments.
Mapaura (2002) considered the Chimanimani mountains to be one of Zimbabwe's 13
biogeographical areas of endemism, but by far the largest in terms of number of species as
there are 33 Zimbabwe endemic and 37 near-endemic taxa (i.e. taxa known from the
Zimbabwe side of Chimanimani but also recorded in Mozambique). Overall, there were 55
taxa recorded only from the "quartzite grasslands" of the Chimanimani Mountains.
More recently, on the Mozambique side, the Chimanimani Management Plan for the TFCA
has an appendix of rare and endangered plants listing 265 taxa (Ghiurghi, Dondeyne &
Bannerman 2010b, Annex 6). It is not clear how this was compiled but it appears to have been
based on the Sabonet list for Zimbabwe. An additional output of this management plan was a
photographic guide illustrating many of the plants of interest (Dondeyne, Wursten &
Bannermann 2009), each photo with locality details. Unfortunately, this does not appear to
have been produced in hardcopy and is only available digitally.
Finally, linked in with the current CEPF study, Toral Shah wrote her MSc thesis on the
conservation of four contrasting endemic species on the Mozambique side (Shah 2016), using
habitat as a proxy in modelling their conservation status.
From a review of the literature, one question not yet fully answered is when it was first
realised that the Chimanimani Mountains supported such a significant number of endemic
species; descriptions of new taxa only really started appearing in the 1950s. There appears to
be no mention of such species until Goodier & Phipps (1958), and nothing explicit until Wild
(1964). When was it suddenly realised this was an area of high endemism, possibly with new
species still to discover?
4.2 Moist Forest Studies
As part of his comprehensive study on moist forests in Zimbabwe, Tom Müller (1999, 2006;
see also Appendix 4 in Ghiurghi et al. 2010a) maps and describes forest patches (patch
numbers 68, 78‒88, 102‒105) in the Chimanimani Mountains area in Zimbabwe, although
most of these are rather small. The altitude at which they can be found, and hence the forest
type, varies from 350 m to over 2000 m. Most of the higher altitude patches (above 1000 m)
are types 5 and 7, with smaller extents of types 4 and 8. Type 5 is Syzygium guineense subsp.
afromontanum montane forest (from 1500‒1900 m), while Type 7 is Mixed submontane
forest (with elements of medium altitude forest such as Chrysophyllum gorungosanum,
generally 1600 up to 1750 m). Type 4 is Ilex mitis‒Schefflera umbellifera‒Maesa lanceolata
montane forest, where conditions are too dry for either Syzygium species to dominate, and
Type 8 is forest dominated by Craibia brevicaudata. The latter is nationally quite scarce,
being known only from the Chimanimani and Vumba areas (generally at 1400‒1600 m).
From his Table 2 (Müller 2006) there are 190 ha of moist forest within the Chimanimani
National Park and an additional 225 ha in the broader Chimanimani area, plus 425 ha in the
lower altitude Haroni‒Rusitu‒Chisengu rivers area.
Lowland Forests
Although not part of the Chimanimani montane flora, the forested areas on the lower slopes
are of great conservation significance. In Zimbabwe the Mukurupini Forest (at the southern
tip of the Chimanimani National Park) and the Haroni and Rusitu Botanic Reserves in the
lower Rusitu valley have been the focus of many biological trips since the early 1960s. These
are almost the only remaining examples in that country of lowland moist forest (Timberlake
1994), although such forests are far more common in Mozambique along the eastern
Chimanimani Mountains: Botany & Conservation, page 29
Chimanimani foothills. Further details of the Mozambican forests can be found in a botanical
and conservation report prepared under the Darwin Initiative project implemented by Micaia
and RBG Kew in 2014‒2017 (Timberlake et al. 2016), but are not specifically looked at
under this CEPF project.
One early published account of the Haroni‒Rusitu junction area came out of a Rhodesian
School's expedition visit in April 1973 ('Vakwashi') to what was then a remote roadless area
(Magness et al. 1973), including a section on the botany. A map was produced based on aerial
photos showing the main forest patches. They found that the main and dominant tree was
Newtonia buchananii, also with numerous lowland forest tree species, with a canopy 40‒60 m
high. On steeper slopes a Brachystegia microphylla‒Uapaca kirkiana woodland was seen.
Ten new records for Zimbabwe were found, and a checklist was produced of 240 taxa
(including cultivated species and arable weeds).
Gerald Pope and Tom Müller studied the lowland forests in the Mukurupini area at the
southern end of the Chimanimani and along the foothills on the Mozambique side in April
1974, from which the checklist of woody plants published in Dutton & Dutton (1975;
Appendix A) comes, but this did not cover montane areas. Later, a checklist of all plants
recorded was compiled for the Haroni‒Rusitu junction area in Zimbabwe based on a
manuscript dating from the 1970s in the Harare Herbarium (Timberlake 2000, unpublished).
Although quite a few Chimanimani endemics are listed there, showing that their distributions
are not solely montane, it is not clear what the altitudinal limits of the original list were.
A Visitors Guide to the Haroni‒Rusitu‒Makurupini area of Zimbabwe was in the process of
being prepared in 2000 by a Zimbabwean NGO, the Biodiversity Foundation for Africa, to be
published as a small book. However, this never reached completion, partly because of
Zimbabwe's political and economic problems soon after. A number of chapters ‒ on ferns,
orchids, plants, vegetation, fish, butterflies, mammals ‒ were written or are in various stages
of completion (Timberlake, pers. comm.).
4.3 Botanical Collectors
As part of this CEPF project, all collections of the known Chimanimani endemics, near-
endemics and what were believed to be range-restricted species found in the Chimanimani
mountains and upland areas of Chimanimani District held at Kew and at the Harare
Herbarium (SRGH) were databased using the BRAHMS herbarium database program (Filer
2016). Collections of all species from the 2014 expeditions were later added in (collections
from the 2016 expedition will be added later). Once corrected, this database, now containing
2153 records with 1687 of them from the broad Chimanimani area, was used as the source of
the observations below. Although the database only covers records of the selected species, it
is believed to give a good indication of collectors and collecting patterns, especially as it is
these species of particular interest that most plant collectors would be targeting.
Over half (50.9%) of the 1158 pre-2014 collections from the broad Chimanimani area in the
database were collected between 1948 and 1960, in what could be considered the heyday of
collecting in the mountains. Almost 81% were collected in the slightly broader period of 1948
to 1970, with just 106 collections between 1905 and 1947, and a further 106 between 1971
and 2013. Surprisingly, there were only 12 collections from the period between 1990 and
2013, including just one since 2000 (in 2011).
Mozambique
As can be seen from the list of the main Chimanimani collectors in Table 4.1, there have been
significantly fewer plant collections from Mozambique over the years than from Zimbabwe,
Chimanimani Mountains: Botany & Conservation, page 30
probably due to the great majority of collectors being Zimbabwe-based. But also because the
area covered by the Kew database covers the non-montane farming areas around
Chimanimani as well as the mountains themselves, giving a bias in collection numbers.
Table 4.1. Botanical collections and collectors in the Chimanimani area 1906 to 2016
(data from RBG Kew database of endemic and near-endemic species only).
Collector/s dates country area
Johnson, W.H. 1907 M 'Moribane' (The Corner?)
*Swynnerton, C.F.M. 1906‒1908 M, Z Chim Mts, 'Melsetter', Mt Pene
Cooper, J.C. Aug 1921 Z 'Melsetter'
Cronwright, W. May 1923 Z Chim Mts
Weiste, Miss May 1942 Z Chim Mts
Finlay, R.H. Sep 1944 Z Chim Mts
*Munch, R.C. 1944‒1948 Z, M, Chim Mts, Chim foothills
Wild, H. Apr 1947 Z Orange Grove
McCosh, F.W.J. May 1947 Z Chim Mts
Chase, N.C. Dec 1948, 1953 Z Pork Pie, Chim farms
**Wild, H.; Munch, R.C. Jun 1949 Z, M Chim Mts, Bundi valley
Pedro, J.G.; Pedrogão, J. Jul 1949 M Chim Mts, Mt Messurussero
Thompson, S. Jul 1950 Z Chim NP
*Crook, A.O. 1950‒1953
(various)
Z, M Mussapa Gap, Chim farms, Corner
**Wild, H.; Chase, N.C.; Munch,
R.C.; Plowes, D.C.H.; Sturgeon,
K.E.; Panton, C.A.
Oct 1950 Z, M Mussapa Gap, Mt Peza, Corner, Pork Pie
Guy, G.L. Apr 1951 Z Chim Mts
Finch, R. Aug 1951 Z Corner
*Ball, J.S. 1953‒1963,
1973
Z Chim Mts, Chim farms (mostly orchids)
Wild, H. Aug 1954 Z Chim Mts
Leach, L.C. 1954‒1961 Z Chim Mts, Chim farms
Watmough, R. Sep 1955 Z Chim Mts
Drummond, R.B. Nov 1955 Z Glencoe FR, Mt Pene
Barrett, R.L.; Stables, J.H. Dec 1955 Z Chim Mts, Mussapa Gap
*Taylor, H.C. Sep 1956 Z, M Chim Mts, Mt Binga, Martin's Falls,
Skeleton Pass
Coates Palgrave, K. May 1956 Z Chim Mts
**Goodier, R.; Phipps, J.B. 1957 (various),
1959 (various),
1960 (various)
Z, M Bundi valley, upper Haroni, Mt Dombe,
Mt Binga, Mt Peza, Mussapa Gap,
Mawenje, Skeleton Pass, Mt Nhamadimo,
Martin's Falls
*Chase, N.C. Apr 1957, May
1958
Z Mt Pene, Chim Mts
Whellan, J.A. Apr 1957 Z Chim Mts
Garley, D.L. Aug 1957 Z Chim Mts
*Hall, A.V. Feb 1958 Z, M Bundi plain, Mt Binga, Martin's Falls
*Chase, N.C. May 1958, Nov
1965
Z Bundi plain, Outward Bound
*West, O. May 1958, Sep
1965
Z Skeleton Pass, The Corner
*Noel, A.R.A. May 1959 Z Chim lower slopes, Bundi plain, Skeleton
Pass
*Loveridge, J.P. Sep 1961, Jun
1966
Z Chim Mts, Chim farms
Chimanimani Mountains: Botany & Conservation, page 31
*Plowes, D.C.H. 1962 (various) Z Chim Mts
*Whellan, J.A. Aug 1964, Dec
1964, May 1965
M, Z Bundi valley, Mawenje, Martin's Falls, R.
Mufomodzi, Saddle
Wild, H.; Goldsmith, B.; Müller,
T.
Dec 1964 Z, M Mukurupini
*Corby, H.D.L. Jul 1965, Feb
1968
Z Chim foothills, Chim farms
Coates Palgrave, K. et al. (school
expedition?)
Aug 1965 Z, M S Chim Mts, Mt Mahoenzi
Torre, A.R.; Perreira, A. Oct 1965 M Rotanda
Perreira, A.; Sarmento, A.;
Marques, A.
Jan 1966 M Mt Binga
*Grosvenor, R.K.; Drummond,
R.B.; Simon, B.K.; Plowes,
D.C.H.
Sep 1966 Z, M Outward Bound, Bundi plain, Mt Binga,
Skeleton Pass, Camp Portage
*Grosvenor, R.K.; Drummond,
R.B.; Plowes, D.C.H.
Apr 1967 Z, M Outward Bound, Bundi plain, Skeleton
Pass, St George's cave, Saddle
Grosvenor, R.K. Sep 1967 Z Chim Mts
*Mavi, S.; Müller, T. Nov 1967 Z, M Mussapa Gap, Chikukwa, Martin FR
*Simon, B.K.; Ngoni, J.F. Nov 1967 Z, M Ngorima, Mukurupini
*Goldsmith, B. 1968‒1973 Z Tarka FR, Mt Pene, Chim Mts
Kelly, R.D.; Simon, B.K. Apr 1969 Z Bundi plain
Müller, T.; Kelly, R.D. May 1969 M Mukurupini, S Chims
Wild, T.M. Aug 1969 Z, M Mukurupini
Plowes, D.C.H. Oct 1970 Z Skeleton Pass
Müller, T.; Gordon, T.A.D.; Pope,
G.
Jun 1971 M Mukurupini
Ngoni, J.F. Apr 1973 Z Haroni-Rusitu
Dutton, P. 1973 M Chim Mts
*Bamps, P.; Symoens, J.J.; vanden
Bergen, C.
Jan 1974 Z Chim Mts, Mt Pene
Beasley, A.J. Mar 1974 M R. Mufomodzi
Pope, G.; Müller, T. Apr 1974 M Chim foothills
Müller, T. Aug 1975, Aug
1977
Z Mukurupini, Mt Pene
Philcox, D.; Leppard, M.J. et al. Mar 1981 Z Chim lower slopes, Mt Pene
Burrows, J.E. Sep 1981 Z Chim Mts
Müller, T.; Best, B. Nov 1983 Z Chim Mts
Linder, H.P. Nov 1986 Z Chim Mts
Carter, S.; Coates Palgrave, M. Aug 1988, Jun
1994
Z Corner, Chim foothills
**Timberlake, J.R.; Ballings, P.;
Wursten, B.; Hadj-Hammou, J.;
Mapaura, A.; Matimele, H.
Apr 2014 M Chim Mts, Mt Pene, Mt Nhamadimo
**Timberlake, J.R.; Wursten, B.;
Mapaura, A.; Mutasa, K.; Chelene,
I.
Oct 2014 M Chim Mts, R. Mufomodzi
**Timberlake, J.R., Mapaura, A.,
Banze, A., Osborne, J., Shah, T.,
Massunde, J., Fijamo, V.
May 2016 M Chim Mts, R. Mufomodzi, Mt Binga, Mt
Nhamadimo, Saddle
Note: Entries marked with a single asterisk are significant collections in terms of numbers (more
than 15); those marked with a double asterisk have more than 50 collections in the database.
Some collectors other than those listed here have collected in the Chimanimani area, but with
only single or very few database entries.
Chimanimani Mountains: Botany & Conservation, page 32
Of the 1687 collection records in the Kew database from the broad Chimanimani area on both
sides of the border, only 699 (41%) are from Mozambique. But of that Mozambique total,
over three-quarters (528) are from the CEPF 2014 Mozambique expeditions, which therefore
represents a four-fold increase in our knowledge. And there have been around another 400
collections since then (see Section 6.1). Up to the recent CEPF collections, there were only
171 records from the Mozambique side of the Chimanimani Mountains, just 15% of total pre-
2000 collections.
The first collections in Mozambique were by W.H. Johnson in 1907 from what he called the
"Moribani" area. Moribane is now regarded as a lowland forest area although the species he
collected included some quartzite endemics. Given the main travel routes at that time, it is
probable that his collections were actually from the Mussapa Gap area near the present-day
Zimbabwe border at around 1000‒1200 m. There are no further Mozambique entries until
R.C. Munch collected montane species in June 1948, including a number of the endemics, and
he later collected with Hiram Wild in June 1949. This was followed by a significant set of
collections from a cross-border expedition with Wild, Munch, Norman Chase, Darrell Plowes,
Kathleen Sturgeon and C.A. Panton in October 1950, the first of the larger trips with a
number of botanists participating. Slightly earlier, in July 1949, J.G. Pedro and J. Pedrógão
collected in the mountains including Mt Messurussero, but there was very little collecting
afterwards until H.C. Taylor entered Mozambique on what was primarily a Zimbabwe field
trip in September 1956, as did A.V. Hall in February 1958.
The main collectors on the Mozambique part of the mountain over the pre-Independence
period were Rawdon Goodier and Jim Phipps, both based in Zimbabwe. They collected over
much of the Chimanimani range on both sides of the border, but in Mozambique only in
December 1959 and April 1960. After that period, other significant collections from
Mozambique were made by J.A. Whellan in 1964 to 1965 in the montane part, by Hiram
Wild, Blake Goldsmith and Tom Müller in December 1964 from the lowland Haroni‒
Mukurupini area, and two cross-border trips by Rosemary Grosvenor, Bob Drummond, Brian
Simon and Darrell Plowes in September 1966 and April 1967. The only Portuguese botanists
we have records from this period were A.R. Torre & A. Perreira, collecting in the Rotanda
area in October 1965, and Perreira, A. Sarmento and A. Marques who collected on Mt Binga
in January 1966.
Although some studies were carried out on the flora (e.g. as seen in Dondeyne et al. 2009 and
Ghiurghi et al. 2010a), no significant collections appear to have been made in the mountains
from 1975 until the CEPF trips in 2014. At this time, numerous collections were also made by
Kew and IIAM botanists (Banze, Cheek, Darbyshire, Fijamo, Massunde, Timberlake) in the
forested Chimanimani foothills in 2015 as part of a Micaia‒Kew Darwin Initiative project. A
partial checklist is available (Timberlake et al. 2016, Annex 2).
4.4 Phytogeography
Surprisingly, the phytogeography of such a fascinating flora as that along the mountains of
the Zimbabwe‒Mozambique border, in particular the Chimanimanis, has been discussed very
little. Weimark (1941, cited in Wild 1964) first showed that the flora of these borderland
mountains was significantly different from that of the surrounding plateau-lands (the
Highveld), yet quite similar to that found on mountains in the Western Cape in South Africa.
Wild (1964) also considered the flora similar to that of the Drakensberg mountains in the
former Natal and Transvaal, as well as to that found in Malawi (Kirk Highlands, Zomba, Mt
Mulanje, Viphya plateau). However, two big geographic disjunctions ‒ the Zambezi Gap and
the Limpopo Gap ‒ were recognised. Later, building upon the geomorphological and
continental erosion thinking of King (1962), which was elaborated upon by Lister (1987),
Chimanimani Mountains: Botany & Conservation, page 33
Wild (1968) suggested that any species common across the mountains must not only have
evolved before the formation of the much hotter and arid gaps but also must have been part of
a more-or-less continental belt of more mesic forest vegetation in the distant past.
Wild (1964) suggested that what he termed the Inyangani Subcentre (Nyanga + Chimanimani)
separated from the Drakensberg, Cape and Mulanje subcentres more than 1 to 2 million years
ago (mya), and went on to suggest (Wild 1964, 1968) that the Cape Flora (or a direct
evolutionary precursor) covered a large continuous area of the African continent up to the
early Tertiary (c. 50 mya). But that before the beginning of the Quaternary (c. 1.5 mya) this
had become very fragmented and remained only in montane areas. He thought that the
endemism had developed over 60‒70 million years.
Broad phytogeographical patterns for the montane areas of eastern and southern Africa are
discussed at length by White (1978), with much of the detailed thinking outlined in a specific
Malawian study (Chapman & White 1970). A specific Chimanimani "island" is recognised by
White (1978) within his broader Afromontane archipelago-like regional centre of endemism,
but his focus is mostly on the trees and forest vegetation. He does mention an Ericaceous Belt,
which seems closest to the upland parts of the Chimanimani Mountains; this is discussed
further in Section 5.1. However, although Killick (1978) in the same publication does not
consider that any part of Zimbabwe or Mozambique falls in his Afro-alpine Region, the
descriptions in particular of the Drakensberg subalpine belt seem very close to what is found
above 1500 m on the Chimanimani massif.
Although the Chimanimani Mountains did not fulfil the full criteria required in their centres
of plant diversity study ‒ areas with over 1000 species and at least 10% endemism ‒ Davis,
Heywood & Hamilton (1994) provide a short description of both the Chimanimani (Af 79)
and Nyanga mountains. A few years later Van Wyk & Smith (2001) clearly identify the
Chimanimani and Nyanga area as being a single centre of endemism, and give a brief account
of the broad area and discuss some of the flora's evolutionary history and links to the
Drakensberg Mountains and the Cape Floristic Region. The long-standing nature of the
mostly grassland endemics is also pointed out, indicating that such grasslands have been
around for a long time.
Most recently, in his overview of the evolution of African plant diversity, Linder (2014) has
suggested that the Chimanimani forms part of the Austro-temperate flora, an outlier of the
Cape flora, characterised by Restionaceae, Proteaceae and taxa typical of the Drakensberg
grasslands. It is not part of what he calls the Tropic-montane flora, which White (1983) calls
Afromontane. This flora, he says, dates back to the Paleogene to Eocene (perhaps 35‒50
mya). However, Linder recognises that there are bound to be some common elements such as
Cussonia and Erica.
Given our greatly improved knowledge today on continental drift, on the comparative ages of
plant genera and on paleoclimates, some of the earlier phytogeographical interpretations
would seem to fall away. Surprisingly, Wild did not seem to accept a major role for long-
distance dispersal either. However, to date he has been the only person to publish anything on
the type of vicariant endemism (closely-related or allopatric species that have descended from
a common ancestral population and attained spatial isolation) found on the Chimanimani
mountains. He pointed out the difference between the vicariant endemics and those that were
sympatric (occurring together); in the latter the two species must be incompatible from a
breeding viewpoint.
Wild also recognised clearly (Wild 1964, 1968) that almost all of the Chimanimani endemics
were confined to quartzite substrates with very few being found on schist or on both (an
Chimanimani Mountains: Botany & Conservation, page 34
exception being Aeschynomene inyangense). He suggested that the lower available
phosphorous levels in soils derived from quartzite, and the significantly lower soil fertility,
must be major factors in causing speciation through ecological stress. Some species such as
Erica johnstonii occur on most mountains from Chimanimani to Mt Mulanje, while others are
confined to quartzite soils. He listed 10 taxa pairs known at that time where one species is
confined to the Chimanimani while a closely related one is more widespread. It is of interest
to note that the species from which each of these particular Chimanimani endemics has
probably evolved are spread across families and in their geographical distribution, and Wild
found no clear phytogeographical patterns.
Other than the discussion in Van Wyk & Smith (2001), no further work since Wild appears to
have been done on the phytogeography and origins of the Chimanimani flora. Given our new
findings, it is hoped that further research will now be stimulated.
4.5 Wildlife Studies
Compared to plants and vegetation much less has been written about the vertebrate and
invertebrate wildlife of the mountains. In the recent past the extensive grasslands and open
areas were renowned for large herds of Eland and Sable antelope, but owing to hunting by the
small-scale miners since the mid-2000s these have mostly disappeared. Phipps and Goodier
(1962) mention Klipspringer, Eland and Sable antelope as being the main grazers, with
baboons and Rock Hyrax also common. Large mammals used to roam freely across the
border but, according to the Reserve Warden (pers. comm., Nov 2014), the remaining animals
have now moved to more wooded areas at slightly lower altitudes in the east and north east
where there are fewer people. During many weeks of fieldwork in 2014 and 2016, almost no
antelope were seen, although evidence of small carnivores was noted. A species of particular
interest is the Aardvark or Ant bear (Orycteropus afer), the dug holes of which provide an
unusual habitat across the grasslands, one that is often used by Blue Swallows for their nests
(Little 2013). Numerous Aardvark holes were seen during our survey, but it was not checked
if these were still occupied, either by Aardvarks or Blue Swallows.
The main published work on wildlife is that by Dutton and Dutton (1973). In their appendices
they provide lists of 67 mammals (including rodents and bats) compiled by Reay Smithers
and José Lobão Tello from museum collections and survey work (much of which probably
later appeared in the atlas of Mozambique mammals, Smithers & Tello 1976); 167 species of
bird based on a list by Hodgson (1971) with supplementary information from Desmond
Jackson (1973); 62 reptiles and 35 amphibians from an unpublished list by Don Broadley (but
see Broadley & Blake 1973); and 49 freshwater fish compiled by Graham Bell-Cross from
museum collections. These lists cover the whole Chimanimani area in both Zimbabwe and
Mozambique and cover both montane habitats and lowland forest (effectively the whole
TFCA); a list of just montane species would be significantly shorter.
Surprisingly, there are very few endemic or range-restricted vertebrates on the mountains.
One amphibian is considered endemic (the Cave Squeaker frog Arthroleptis troglodytes,
although it has not been found since 1962; ZSL 2016) and one possibly endemic fish, but no
endemic bird species.
The Chimanimani Mountains are designated as two Important Bird Areas (IBAs), one on the
Zimbabwe side of the border (Childes & Mundy 2001) and a much larger one on the
Mozambique side (Parker 2001). Only three restricted-range bird species occur here ‒
Swynnerton's Robin Swynnertonia swynnertoni, Chirinda Apalis Apalis chirindensis and the
Briar Warbler Prinia robertsii. Together, these IBAs cover a large part of both the highlands
and surrounding lowland forests and include five globally threatened or near-threatened (A1)
Chimanimani Mountains: Botany & Conservation, page 35
species ‒ Taita Falcon Falco fasciinucha (VU), Swynnerton’s Robin (VU), Southern Banded
Snake-eagle Circaetus fasciolatus (NT), Blue Swallow Hirundo atrocaerulea (NT) and Plain-
backed Sunbird Anthreptes reichenowi (NT). Of these, only the Blue Swallow occurs only in
the highland, non-forested habitats. The status of its population on the Mozambique side
seems to be uncertain (Parker 2005, Little 2013). However, during our fieldwork there would
appear to be as much suitable habitat on the Mozambique side as on the Zimbabwe.
Despite the large number of endemic plants, some of which appear to be suitable larval food
species, there is only one species of butterfly known to be endemic to the Chimanimani
Mountains ‒ Lepidochrysops barnesii (Lycaenidae) ‒ known only from montane grassland at
1800 m in a small valley on the Zimbabwe side. In addition there are five other near-endemic
species known only from montane grasslands or Afromontane forest along the Zimbabwe‒
Mozambique borderlands. Surprisingly there is no checklist of Lepidoptera from the
mountains (Alan Gardiner and Colin Congdon, both pers. comm. October 2016).
View over upper Mufomodzi valley from slopes of Mt Binga, northern Chimanimani [TS].
Chimanimani Mountains: Botany & Conservation, page 36
5. VEGETATION
5.1 Regional Level
At a continental level, the vegetation of the Chimanimani Mountains and associated
mountains along the Zimbabwe‒Mozambique border has been described as Undifferentiated
Afromontane Vegetation (unit 19a) by White in his map of African vegetation (White 1983).
Forming part of his Afromontane archipelago-like regional centre of endemism, this type is
embedded within the Zambezian regional centre of endemism comprising woodlands and
other formations on the main continental plateau. The Afromontane archipelago extends from
Yemen through the Ethiopian Highlands, the Albertine Rift and Malawi to the Drakensberg
and Cape mountains in South Africa, with outliers in West Africa and on the Angolan scarp.
In his earlier study on the Afromontane Region, White (1978) recognised the Chimanimani
area as an 'island' in this Afromontane archipelago. Following Hedberg's work on vegetation
zonation on East African mountains (Hedberg 1951), in which he describes a Montane Forest
belt, an Ericaceous Belt and an Afro-alpine Belt, White (1978) mentions the presence of an
Ericaceous zone at higher altitudes on the Chimanimanis and on Mt Malawi (although the
Nyanga massif should certainly also be included here). However, zonation on the mountains
of south-central Africa is not as marked as on the taller East African mountains such as Mt
Kenya, Mt Kilimanjaro and the Ruwenzori. White suggests that the Ericaceous Belt, which in
the case of the Chimanimani and other similar mountains is the formation that supports most
of the narrow-range species, is best described together with the Montane Forest Belt, while
the Afro-alpine Belt is more distinctive (but see Section 4.4).
At a regional level, the main framework available is that for the whole Flora Zambesiaca area
(Wild & Barbosa 1968), which describes all the montane vegetation in this area as Themeda‒
Exotheca‒Loudetia Submontane and Montane Grassland (unit 68). In his precursor study of
the vegetation of Mozambique, Barbosa (1955) shows the upland Chimanimani area as Zona
subalpestre (complexo 39), which he divides into eight subtypes, including evergreen forest
with Podocarpus, secondary "matagais" with Widdringtonia, Phillipia (now Erica) and
sometimes Brachystegia spiciformis, stands of Strelitzia, scrub with species of Erica,
Passerina, Helichrysum and Aeschynomene, secondary grasslands, and, finally, boggy areas
with Sphagnum, Lycopodium and Drosera.
5.2 District Level
At a more local level, and with the exception of moist forest, the vegetation of the
Chimanimani Mountains is primarily determined by and follows the underlying geology
(Phipps & Goodier 1962, Goodier & Phipps 1962, Wild 1964). Altitude is also a factor with
sclerophyllous vegetation and grasslands at higher altitudes and woodland and forests lower
down. There are distinct differences within these vegetation formations depending on
substrate.
An idea of the "greenness" of the vegetation across the mountains is given in Figure 5.1, a
recent (2013) satellite image that was used in a biomass assessment study (Casey Ryan, pers.
comm.). The quartzite areas show up clearly as having significantly less vegetation
reflectance.
By far the most comprehensive account of the vegetation and ecology of the Chimanimani
Mountains, although confined primarily to the Zimbabwe side, is found in Phipps & Goodier's
1962 Journal of Ecology paper, with a map and less detailed account in Goodier & Phipps
(1962). Their accounts are simplified and paraphrased below along with some additional
observations, particularly for the Mozambique side which they did not look at in detail, and
their map is shown in Figure 5.2. The area covered is essentially above 1000 m altitude.
Chimanimani Mountains: Botany & Conservation, page 37
Figure 5.1. Satellite image of the Chimanimani TFCA area showing levels of
vegetation production (source: Casey Ryan, pers. comm.).
The main vegetation types recorded by Goodier & Phipps (1962) were:
A. Forest
i) Dry montane forest
ii) Marginal (seral) forest
B. Woodland
i) Uapaca kirkiana woodland (miombo)
ii) Brachystegia spiciformis woodland (miombo)
iii) Brachystegia tamarindoides woodland
Chimanimani Mountains: Botany & Conservation, page 38
C. Scrub
i) Ericaceous scrub
ii) Proteaceous scrub
D. Grassland (both wet and well-drained)
i) On quartzite terraces
ii) On schist slopes
iii) Hydromorphic grasslands
E. Aquatic communities
F. Lithophytic communities
Figure 5.2. Vegetation map of the Zimbabwe portion of the Chimanimani Mountains
(from Goodier & Phipps 1962).
Chimanimani Mountains: Botany & Conservation, page 39
A. Forest
The forests found above 1000 m are somewhat dry but with a 10‒15 m high closed canopy
and a sparse ground flora (type A.i). Few trees are larger than 30 cm diameter, while lianas
and epiphytes are common in the canopy with ferns and mosses in the ground layer. This is
often called Afromontane forest and is somewhat stunted compared to forests at lower
altitudes. It mostly occurs in smaller patches from 1 to 5 ha in extent, although in favourable
localities they can reach 30 ha. The largest patch seen, on the west-facing slopes above the
Rio Nyahedzi, was around 240 ha (Figure 5.3). Most patches are confined to sheltered gullies
(sometimes called kloofs) that are subject to less evapotranspiration. There does not appear to
be a direct link to underlying geology, possibly as once forest develops it creates its own
humus-rich soil. Access to year-round moisture and better protection from wild fires is also
important, although fire itself is not thought to be a major determinant of these forest's
distribution (Müller 1999). Some of the more characteristic trees are Schefflera umbellifera,
Ilex mitis, Macaranga mellifera, Maesa lanceolata, Morella pilulifera, Podocarpus
milanjianus and Syzygium cordatum. In drier areas Widdringtonia nodiflora can be found.
Surrounding many forest patches and in smaller gullies, a scrubbier vegetation type can be
found (type A.ii), transitional between Ericaceous scrub and forest. Common species here
include Erica (Phillipia) mannii, Englerophytum magalismontanum, Rapanea melanophloeos
and Myrsine africana, while bracken (Pteridium aquilinum) is found on the outer margins
where soils are somewhat deeper and with some humus. Along some sheltered rocky streams
groves of the banana-like Strelitzia caudata are often seen, while the tree fern Cyathea
capensis grows from sheltered stream banks in more open areas.
B. Woodland
Woodland is of three types, all generally termed miombo, with their distribution primarily
related to underlying geology and soil type. Confined mostly to well-drained slopes,
woodland is most extensive at medium to lower altitudes and only occasionally found above
1400 m. Using Google Earth Pro the extent of denser woodland (all types) above
approximately 1000 m altitude was estimated at 112 km2. Most woodlands are fairly open
with about 20‒60% canopy cover and trees usually only 4‒8 m high. Trees are shorter at
higher altitudes and on more shallow soils, as well as being much more open in their canopy.
Uapaca kirkiana (mzhenje) woodland (type B.i) is mostly confined to altitudes below 1200 m
covering much of the lower eastern slopes and is rarely found on quartzite. At lower altitudes
it is typically found with Brachystegia utilis, Pterocarpus angolensis and Pericopsis
angolensis and forms a typical and extensive miombo woodland, while at higher altitudes
these other species tend to disappear. The ground flora, especially higher up, tends to be
sparse, commonly with Loudetia simplex and Helichrysum kraussii. In some places closer to
Figure 5.4. Brachystegia tamarindoides
woodland on quartzite [JT].
Figure 5.3. Afromontane forest on slopes of Mt
Nhamadimo [JT].
Chimanimani Mountains: Botany & Conservation, page 40
streams, the stilt-rooted Uapaca lissopyrena is seen, which can also form a localised
woodland.
Brachystegia spiciformis (msasa) woodland (type B.ii) is primarily found on schist soils and
rarely on quartzite. Trees can be quite low (2‒4 m high) and scattered at higher elevations,
although large spreading trees can be seen in favourable sites. This type is generally found at
lower altitudes with Uapaca kirkiana and sometimes Faurea saligna, with medium-height
grasses 60‒90 cm high of Themeda triandra, Digitaria species, Loudetia simplex and
Tristachya nodiglumis, sometimes with Pteridium aquilinum and the creeper Smilax anceps.
Brachystegia tamarindoides subsp. microphylla woodland (type B.iii, Figure 5.4) is very
attractive and confined to quartzite outcrops and rocky areas where the low spreading
branches are often draped in long strands of Usnea lichen. The canopy is very open but,
owing to the generally rocky or shallow substrate, there is a sparse herbaceous layer.
Although Phipps & Goodier (1962) said it has a lower altitudinal limit than the other two
woodland types, we did not find that to be the case in the Mozambique side.
C. Scrub
This group of vegetation types is widespread across the plateau above about 1200 m altitude;
Phipps & Goodier (1962) recognised five shrubland types with various facies, but we only
cover the two main ones here.
Ericaceous scrub (type C.i) is one of the more species-rich vegetation types across the
mountains and it is in this that many of the Chimanimani endemics or near-endemics are
found. It is characterised by low (0.5‒1 m to 3 m high) shrubs of Erica (Phillipia) simii along
streamsides, Erica pleiotricha, E. johnstonii and E. lanceolifera in semi-fire protected areas
and, especially, taller shrubs of Erica (Phillipia) hexandra on rocky outcrops. All of these
species are only known from quartzite substrates. Along stream margins Ericaceous scrub can
give way to marginal evergreen forest with Podocarpus milanjianus, Widdringtonia nodiflora
and Syzygium cordatum (type A.i), sometimes with Strelitzia caudata. The Erica hexandra
thickets, forming ‘rock gardens’ or on boulder slopes, are fairly widely distributed above
1400 m, and are often associated with Schistostephium oxylobum, Anthospermum vallicola,
Myrsine africana, Phylica ericoides, Passerina montana and Aloe munchii, along with some
shrubby legumes (Figures 5.5, 5.6). This formation is said to be fire-sensitive (Phipps &
Goodier 1962), although it can readily regenerate. In addition, fires are stopped from
spreading rapidly by the large gaps between boulders, which fires do not readily jump across.
The other main scrub type is Proteaceous scrub (type C.ii), primarily found on rolling schist
grassland between 1100 and 1800 m, although it is not always easy to clearly demarcate this
Figure 5.6. Rock garden, N Chimanimani [JT].
Figure 5.5. Rock garden, N Chimanimani [JT].
Chimanimani Mountains: Botany & Conservation, page 41
type and schist grassland. It is characterised by low (1 to 2 m high) shrubs of Protea caffra
(previously P. gazensis), P. welwitschii, P. wentzeliana (previously P. crinita) and
Leucospermum saxosum, along with the small shrublet or suffrutex Morella chimanimaniana.
Associated species, especially at slightly lower altitudes, are Parinari curatellifolia, Faurea
saligna and Syzygium cordatum, all of which can become small trees. Woody herbs include
Aeschynomene grandstipulata and Diplolophium buchananii. Unlike ericaceous scrub, this
vegetation type is moderately fire-tolerant and many stems are blackened at the base.
Two of the three other scrub types described by Phipps & Goodier (1962) were noted, but we
prefer to group them under the two main types.
Mixed sclerophyll scrub seems to be intermediate between the two previous types, restricted
as they are to differing substrates. There are a number of places where quartzite-rich soils
have slumped over red schistose soils, creating a sort of hybrid environment. In our fieldwork
we did not recognise this intermediate type, regarding all scrub on rocky quartzite terrain as
ericaceous scrub or 'rock gardens'.
Bracken scrub, vegetation dominated by the bracken fern Pteridium aquilinum, is quite
clearly discernible in the field but seems to normally occur on slightly deeper, well-drained
and humus-rich soils associated with forest patches. Associated species, depending on the
proximity to forest patches, include Aeschynomene gazensis, Anthospermum vallicola,
Buddleja salviifolia, Dissotis princeps, Eriosema montanum, Kotschya thymodora,
Harungana madagascariensis, Hypericum roeperianum, Polygala virgata and Tetradenia
riparia, along with the thorny scrambler Smilax anceps. The grasses are often quite tall and
include Themeda triandra, Eragrostis species and even Hyparrhenia newtonii. As Phipps &
Goodier point out, this is essentially a seral or intermediate type, and if there was no fire it
may well thicken up to a woody scrub or even forest.
We did not specifically note any Thesium sclerophyll scrub, as described by Phipps &
Goodier, on the Mozambique side. It is said to be limited in extent on the Zimbabwe side on
quartzite and is dominated by the semi-parasitic Thesium whyteanum.
D. Grassland
This is the most extensive and characteristic vegetation formation across the plateau and is
also moderately species-rich. Generally found on level or rolling terrain, grassland is rarely
seen on steeper rocky slopes except as small pockets on more level terraces. Although Phipps
and Goodier say this is a fire sub-climax type, we suggest that it is as much a natural response
to prevailing soil and climatic conditions (as discussed in Section 6.5), although its extent
may have increased with the coming of humans while boundaries between grassland and
forest or scrub have become harder and more marked.
There are three broad grassland types based on soil type and drainage status, with the most
abundant grass ‒ Loudetia simplex ‒ being generally dominant in the non-hydromorphic
types.
Grasslands on shallow soils derived from quartzite or quartzitic sandstone (type D.i) are
mostly associated with larger broad valleys of larger rivers such as the Bundi and Mufomodzi.
Grasses are fairly low in height and somewhat tufted. Apart from the dominant Loudetia
simplex, Sporobolus festivus, Panicum brazzavillense, Elionurus muticus, Monocymbium
ceresiiforme, Panicum ecklonii, Rhytachne rottboellioides and Trachypogon spicatus are
typical. In wetter patches Eriospermum mackenii, Otiophora inyangana, Mesanthemum
africanum, Xyris spp. and Platycaulos (Restio) quartziticola are seen. A figure for its extent
Chimanimani Mountains: Botany & Conservation, page 42
(separate from quartzite scrub and bare rock) is not available, but is probably in the order of
50‒100 km2.
The other main grassland type is found on rolling hills of red soils derived from schist (type
D.ii). This occurs across the Chimanimani massif but large expanses are seen in the Bundi
valley and on Mt Peza in Zimbabwe, and in Mozambique in the middle reaches of the Rio
Murera (Ma-Esese) and the eastern portion flanking the middle reaches of the Rio Mufomodzi
(Figure 5.7). The grasses are mostly denser and taller than on the shallow and nutrient-
deficient quartzite soils. Although Loudetia simplex is common, the most characteristic grass
is Themeda triandra, along with Tristachya hispida, Monocymbium ceresiiforme and the
sedge Bulbostylis contexta. Small shrubs of Protea gazensis, Indigofera cecilii and the schist
endemic Morella chimanimaniana are also typical, as is bracken Pteridium aquilinum.
Another characteristic feature, absent from quartzite grassland, are low (50 cm high) rounded
termite mounds (Figure 5.8) as well as aardvark holes. These grasslands are readily
discernible on Google Earth imagery, from which an estimated total extent of 150 km2 was
calculated.
Hydromorphic grassland (type D.iii), the last of the three broad types, is, as the name
suggests, associated with lower parts of the soil catena near watercourses, areas of impeded
drainage and around seepages. Soils can be quite peat-rich. Phipps & Goodier separate out
those on quartzite and schist, and provide detailed and differentiated species lists, but here
they are combined. The habitats can be more open than in the better-drained grasslands, so a
number of less-competitive and low-growing herbs such as Xyris species, Mesanthemum
africanum and various orchids are found, as well as sedges such as Cyperus denudatus.
E. Aquatic communities
These are obviously very restricted in extent. Very few submerged plants such as Isolepis
fluitans have been seen, but others are associated with sand banks and stream banks that are
regularly flooded. At lower altitudes, below 900 m, submerged species such as Hydrostachys
polymorpha are found, with the palm Phoenix reclinata and screw pine Pandanus
livingstonianus along the river banks.
F. Lithophytic communities
The last vegetation type described here is that found growing on exposed rocks ‒ lithophytic ‒
nearly all of which is on quartzite (Figure 5.9). A number of crustose lichens cover the rock
but also larger plants such as the tufted sedge Coleochloa setifera, Xerophyta argentea and
Aloe hazeliana. Where there are cracks and crevices, woody plants such as Olea chimanimani
are found along with Asparagus chimanimaniensis and Plectranthus chimanimaniensis. In
Figure 5.7. Schist grassland, C Chimanimani [JT]. Figure 5.8. Termite mound in schist
grassland, C Chimanimani [JT].
Chimanimani Mountains: Botany & Conservation, page 43
shaded crevices Impatiens salpinx (also on shaded streambanks), various Streptocarpus
species and a number of small ferns are found. Small lithophytic orchids such as Polystachya
valentina can be locally common.
The relationship between vegetation type, geology and soils was described in detail by Phipps
& Goodier (1962). A modified version of their table showing these relationships, and that
with fire, is given in Table 5.1.
Figure 5.9. Quartzite crags, N Chimanimani [JT].
Chimanimani grassland [JT]
Chimanimani Mountains: Botany & Conservation, page 44
Table 5.1. Summarized relationships of Chimanimani plant communities and ecological factors (modified from Phipps & Goodier 1962).
Ecological factors
Level ground/
dee
p
er soil
Medium slope/
shallower soil
Rocky slopes,
cra
g
s
Schist soils
Quartzite soils
Permanently high
water table soils
Seasonally high
water table soils
Moist but well-
drained soils
Severe fires
Moderately hot
fires
Fires rare
Fires occasional
Aquatic
Wind-exposed
areas
Sheltered areas
Below 1500 m
1350-1650 m
Above 1650 m
A. FOREST
Ai) Dry montane forest × × × ×
× ×
× × ×
Aii) Marginal (seral) forest × × × × ×
× ×
× ×
B. WOODLAND
Bi) Uapaca kirkiana woodland × × × ×
×
× × ×
Bii) Brachystegia spiciformis woodland × × ×
×
× × ×
Biii) Br. tamarindoides woodland × ×
×
×
× × ×
C. SCRUB
Ci) Ericaceous scrub × × × × × × × × ×
Cii) Proteaceous scrub × × × × × × × ×
× ×
D. GRASSLAND
Di) On quartzite terraces × ×
× × × × × ×
Dii) On schist slopes ×
×
× ×
× × ×
Diii) Hydromorphic grasslands ×
× × ×
×
× ×
E. AQUATIC COMMUNITIES ×
× ×
F. LITHOPHYTIC COMMUNITIES × × ×
× × × × ×
Chimanimani Mountains: Botany & Conservation, page 45
6. BOTANY AND ECOLOGY
The main botanical and ecological findings of the study are described below. This is followed
by conservation assessments carried out under the project. A more detailed account of the
threats is given in the following chapter (Sections 7.1 to 7.4).
6.1 Species and Collections
Three fieldtrips were undertaken under the project, two in 2014 (April and October-
November) and one in 2016 (April-May). The total time spent up in the mountains was 37
days and a fairly extensive area was covered, particularly across the centre and northern parts.
The main places visited are shown on a Google Earth image (Figure 6.3).
The range of fieldwork conducted included general botanical survey and collecting, which
was mostly focussed on species of restricted distribution, and a survey of plant habitats and
their associated threats. To help with the field surveys, an informal illustrated field guide for
these species was compiled at Kew (Hadj-Hammou 2014).
There were 11 collectors with numbered series at different times, with an overall total of 927
numbered collections (346 in Apr 2014, 183 in Oct-Nov 2014 and 398 in Apr-May 2016). All
specimens, where possible, were collected in series of three or four, with the top duplicate
going to the National Herbarium in Maputo (LMA) and duplicates to Kew in London (K), the
National Herbarium in Harare (SRGH), and in many instances also to the Meise Herbarium in
Belgium (BR) and the Micaia herbarium at Ndzou camp. Duplicates of N. Brummitt numbers
(mostly Pteridophytes) went to the Natural History Museum in London (BM).
Provisional identifications, at least to genus, were made of all plant specimens collected in the
field, aided by the illustrated field guide mentioned above (Hadj-Hammou 2014).
Identifications were confirmed or re-determined at Kew by a range of plant family specialists
and African plant generalists through reference to relevant literature (notably Flora
Zambesiaca), online resources such as the Flora of Zimbabwe and Flora of Mozambique
websites (http://www.mozambiqueflora.com and http://www.zimbabweflora.co.zw) and
comparison with authoritatively named collections in the Kew herbarium. All specimens
collected were databased in