Biodiversity Survey of the upper

Angolan Catchment of the

Cubango -Okavango River Basin

0 | P a g e

Contents

Executive Summary 2

A. Overall Introduction 4 B. Survey sites 4 C. Overall Results 11

Taxa Reports 19

1. Fish 19 2. Dragonflies 57 3. Reptiles & Amphibians 68

Way Forward 76

Dr Chris Brooks

May 2012

This publication was prepared by Chemonics International Inc.

1 | P a g e

Executive Summary

The Southern Africa Regional Environmental Program (SAREP) has a strategic goal to improve the conservation and sustainable use of biological resources of the Cubango-Okavango River Basin. This report fulfils one of the initial steps in helping to mitigate a critical threat to biodiversity within the system, i.e.; Poor Knowledge of the Status, Extent and Factors Regulating Biodiversity within the Basin (Indirect Threat) (SAREP, 2012), providing information on some of the aquatic faunal diversity of the Cubango-Okavango basin within Angola.

This report provides an initial summary of the key findings from the aquatic biodiversity survey of the Angolan portion of the Cubango-Okavango River Basin. The findings presented within the report are provisional, interim results, which need to be validated with post-survey laboratory analysis. The post-survey analysis is on-going, with finalised results to be published in early August 2012.

The aquatic biodiversity survey formed the first of a series of planned surveys cataloging and improving the inventory of species that occur within the Cubango-Okavango basin. The survey was a collaborative venture between SAREP and the Angolan Ministry of Environment’s (MINAMB) Institute of Biodiversity and the Angolan Ministry of Agricultures National Institute of Fish Research (INIP) The survey was organized by Dr Chris Brooks, SAREP’s natural resource specialists and Dr Paula Coelho, the Director of MINAMB’s Biodiversity Directorate. The survey was undertaken by specialists from the following organizations with the assistance of Angolan counterparts from the stated Ministries;  Prof Paul Skelton & Roger Bills (fish); The Sothern African Institute for Aquatic Biodiversity, Grahamstown, SA  Werner Conradie (amphibians & reptiles); Port Elizabeth Museum; Bayworld, SA  Jens Kipping (dragonflies); BioCart, Leipzig, Germany  Vince Shacks (crocodiles); Maun, Botswana

The need to improve our knowledge of biodiversity within the upper catchment of the Cubango- Okavango basin was prioritised within Angola’s National Action Plan for the basin and the Transboundary Diagnostic Analysis for the basin. Planning for the survey was founded on principles of good ecosystem management, following the IUCN’s integrated wetland assessment concept1, which adopts the Convention of Biodiversity’s ecosystem approach. The IUCN concept defines eleven steps in three stages towards developing a good foundation for ecosystem management. Stage two of the wetland management assessment is to define the scope and status of biodiversity within the study area through surveys. Stage three is currently being undertaken, which consists of analyzing the data from the survey to present to policy makers in a practical and policy-relevant form.

The survey was undertaken at the start of the dry season, when many seasonal migratory species and wet season emergents are in low numbers or in states of aestivation / hibernation. To maximize the potential for sample collections the survey therefore focused on the aquatic ecosystems of the basin; the tributaries, floodplains, marshes, open channels, lagoons and rapids, as opposed to the terrestrial ecosystems; Miombo and mixed Brachystegia woodlands, edaphic grasslands and transformed / altered habitats. The survey team comprised of specialists in the fields of fish,

1 Springate-Baginski, O., Allen, D. and Darwall, W.R.T. (eds.) 2009. An Integrated Wetland Assessment Toolkit: A guide to good practice. Gland, Switzerland: IUCN and Cambridge, UK: IUCN Species Programme. xv+144p.

2 | P a g e amphibians, dragonflies and crocodiles, with five international specialists who were accompanied by on the survey by six Angolan counterparts from MINAMB, INIP and the Angolan Agriculture and Rural Development and Fisheries Department;  Martha Alexandre Zumbo (MINAMB)  Maria Helena Loa (MINAMB)  Julius Bravo (MINAMB)  Francisco de Almeida (INIP)  Maunuel Domingos (INIP)  Gabriel Cabinda (Agriculture and Rural Development and Fisheries)

The survey took place between the 6th and 25th of May, 2012, starting in vicinity of Menongue and moving north towards Huambo and the headwaters of the Cubango, before returning to Menongue and moving east to Cuito Cuanavale, with the final site on the Cuchi River west of Menongue. In total twelve main sites were surveyed, with anywhere between two and four sampling locations within each site. Sites were stratified across the principal habitat classes occurring within the basin and the main tributaries of the Cubango and Cuito River systems, and selected to ensure samples could be collected from a broad range of micro-habitats effectively and efficiently.

The most outstanding finding from the survey was that the Cubango and its upper tributaries have a fundamentally different fish and dragonfly assemblage to that of the Cuito and its tributaries. The Cuito catchment is more similar to the lower Okavango and Delta system in Namibia and Botswana, but having a number of species that appear to be different to the forms encountered in the Delta region. The survey found several (4 or possibly 5) un-described species of fish and a further 5 un- described species of dragonfly in the Cubango-Okavango system. The Odonata fauna found within the Cubango system shows more similarities to the upper Zambezi than to the Okavango Delta fauna, while the Odonata fauna within the Cuito system is more similar to that found in the Delta (although with 2 possible new distinct species found in the Cuito). Every record made for both the amphibians (15 species) and reptiles (10 species) found within the Angolan Okavango system is a new confirmed record for Angola. No historical or museum collections exist for this area.

During the survey records were made of the various observed threats to biodiversity. The most prominent of which was the shear scale of utilization and dependency upon the region’s natural resources (surface flowing water and timber in the Miombo woodlands) by communities within the basin. While the survey noted the current high quality of fresh water in the system, the rapid population growth, coupled with poor environmental awareness could lead to a decline in water quality and subsequently to the status of biodiversity within the system. Evidence of a large scale commercial rice development on the floodplains of the Longa River are also a matter of concern, unless appropriate environmental interventions are made.

It is planned to undertake a second survey that will again cover the aquatic ecosystems, sampling the same sites (plus some additional locations), but at a different time of the year to maximize the potential to discover additional new species for Angola, the basin and science as a whole. Ideally the survey will be undertaken during the first two weeks of November, 2012, to take advantage of the new rains, while not being restrained by the difficulties associated with operating within the region during the later part of the rainy season. A third survey has also been discussed with MINAMB to focus on the terrestrial systems of the recently declared National Parks. The National parks cover 46,000km2 and 22,600km2 respectively between Mavinga and Luengue-Luiana, in south-eastern Angola, adjoining the Cuito and lower Kuvango Rivers.

3 | P a g e

We would like to take this opportunity to thank the Halo Trust for the logistical support they provided during the survey and the extremely important advice they gave on avoiding and working safely around areas with land mines.

4 | P a g e

A. EXECUTIVE INTRODUCTION

On the 14th of November 2011, Steve Johnson, the Director of the Southern Africa Regional Environmental Program (SAREP) signed a Memorandum of Understanding (MoU) with the Angolan Minister of Environment (MINAMB), Maria de Fatima Monteiro Jardim and the National Director for Biodiversity within the Ministry of Environment, Paula Francisco Coelho. The signing of the MoU completed the procedural requirements for SAREP and MINAMB to collaborate and work together in planning for the Aquatic Biodiversity Survey.

The need for the biodiversity survey was prioritised within the MoU as the second objective, after the strengthening of cooperation between the parties. Justification for such a high level of prioritisation is clear when addressing the Angolan National Action Plan (NAP) for the Cubango- Okavango, which states under objective 4 the need, “To develop a better knowledge and understanding about the basin ecosystems, through biodiversity monitoring, environmental management and staff training programs” and under target 11 further defines the activity as; “Basin biodiversity known and protected thanks to the implementation of scientific research and biodiversity monitoring actions and staff training”. The poor status of knowledge on the nature and extent of biodiversity within the Angolan portion of the basin is well recorded, attributing to its high level of concern with the Preliminary Biodiversity Threat Assessment for the Cubango-Okavango basin2. Undertaking the aquatic survey thereby helps to mitigate this identified threat and contribute to the expressed needs of the Angolan Government.

The aquatic biodiversity survey helps SAREP to contribute towards the development of an inventory of species within the system, as well as providing a benchmark from which to monitor their distribution and status. As Ferreira & Towns (2001)3 suggest, it is important when monitoring ecosystems to measure community structure at different trophic levels. By doing so it is then possible to see if and how community structures diverge and evaluate the impacts of different drivers of change, either human or climatically driven. This first survey focused on specific, diverse components of the aquatic realms, rather than the full community structure, and has enabled us to learn from the process, build on our experiences and plan for successive surveys, which will include additional components and build on both the inventory and baseline assessment of the system.

Wetland biodiversity provides extensive direct and indirect benefits to the people of the basin. Provisioning services from the wetlands, such as food from fish and income from selling dried fish were observed throughout the surveyed area, while almost all of the communities are dependent upon the rivers for drinking water. Supporting and regulating services, such as nutrient cycling from pristine dambos and floodplains, which are critical to sustaining the ecosystem functions of the Cubango-Okavango are in evidence through the region, but are under threat from the encroaching human development. The survey helps with SAREP’s assessment of the status of these ecosystem services, while the results from this and further surveys will help to monitor water quality and indicate if there is a decline in the quality and nature of these services and the very functioning of the system.

The survey was a collaborative venture with both MINAMB and the INIP involved in the planning of the survey and its implementation. One of the objectives, apart from improving our knowledge of biodiversity in the system, was to build capacity within Angola for MINAMB and INIP to undertake

2 Brooks, C. 2012. Preliminary Biodiversity Threat Assessment for the Cubango-Okavango Basin. Chemonics International, Washington. 3 Ferreira, S.M.; Towns, D.R. 2001: An ecosystem approach to maximise conservation of indigenous biological assets of New Zealand. Auckland Conservancy, Department of Conservation, Auckland

5 | P a g e

these surveys independently in the future. These objectives were realised, with training provided during the survey and can also be built on with further surveys.

In summary, the purpose and ultimate benefit of assessing the status and distribution of the aquatic species within the Cubango-Okavango is to enable the effective conservation of biodiversity and improve livelihood values through presenting information on species diversity in a format that can be integrated into SAREP’s various decision support making processes. The data from this survey in addition to that collected from future, planned surveys will also serve as a baseline for monitoring the impacts of any development or management interventions, and will enable adaptive management and evaluation of any mitigation measures put in place.

B. SURVEY SITES

a. Survey timing The survey was initially planned to take place in February 2012, and with a more extensive team of specialists, but was delayed and reduced in size to account for logistical requirements, which included delays in getting Ministerial approval in Angola and delays in acquiring visa invitation letters for the specialists. Whilst initially frustrating, the delay enabled the survey team to capture good records of dry season species diversity in May, and at a slightly lower flood water level than would have been possible in February. Access to many of the sites would also have been more problematic in February with heavy rains and water logged roads.

Ideally, from a weather perspective, the survey would have been undertaken in April, rather than May, when warmer conditions would have improved sampling for dragonflies, reptiles and some amphibians and may have improved fish samples caught on the line.

b. Survey sites The geographical scope of the survey is, in part, defined by the watershed of the Cubango-Okavango basin. Within the watershed, sites were selected on the evaluation of four factors; i. Stratification of the main tributaries of the Cubango and Cuito ii. Stratification of the principal habitat classes within the basin iii. The location of sites known by local stakeholders for their importance for biodiversity (Biodiversity Hotspots defined through a series of consultative workshops within Luanda, Kuando Kubango, Huambo and Bie) iv. The practicality of accessing the site and the ability to re-sample the site effectively and efficiently in the future

However, to help place the results into context and understand the relationship and connectivity with adjoining river basins an attempt was made to select sites close to and within adjoining basins, namely the Kwando, Cuanza, Cuvo and Cunene (Fig. 1). Taxa specialists, especially fish specialists, were extremely keen to look at the influence of these watersheds on speciation between systems. Difficulties arouse however in accessing sites along the watershed boundaries of the Kwando, due to road access, and while a site on the watershed of the Cunene was found in the vicinity of Sambo, south of Huambo, we ran out of time to access the site. A site was also planned to be surveyed on the watershed with the Cuanza, close to Mumbué, but landmine fields prevented access to this site. A site was however sampled in the Cuvo.

6 | P a g e

Figure A, The Cubango-Okavango River basin in relation to adjoining river basins (EPSMO 2010)

Stratification of the main tributaries of the Cubango and Cuito The underlying geology of the basin differs across its principal tributaries. The Cubango and its tributaries lay to the west of the basin and are underlaid by granite outcrop and characterised by a series of rapids and some waterfalls in places. This contrasts with the Cuito and its tributaries to the eastern side of basin, characterised by kalahari sands and large slow flowing rivers which act as a ‘sponge’ slowly releasing water into the system.

The survey design incorporated a stratification of the river systems across the middle line of the basin to determine if there was also a biological diversity across these main rivers and tributaries. A poor road network, the occurrence of land mine fields and the constraints of time meant it was not possible to have a series of parallel transects cutting across these rivers at various levels up the system towards it source.

Sites were selected on all the main tributaries of the Cubango; the Cuebe (sites 1& 2), Cuelei (site 8), Cacuchi (site 3), Cuchi (site 4 & 12), Cutato (site 7) and the Cubango (site 5) itself, while on the Cuito, sites were selected on the Luassingua (site 11), Longa (site 10), Cuanvale and the Cuito (site 9) (Fig. B).

7 | P a g e

Figure B, Sites samples during the Aquatic Biodiversity Survey

8 | P a g e

Stratification of the principal habitat classes within the basin The upper catchment of the Cubango-Okavango is characterised by Miombo woodland and mixed Brachystegia woodland, with secondary grasslands on the Alto Plateau near to Huambo and Kuito, with edaphic grasslands spread throughout the upper catchment in association with large dambos. As you move southwards towards the Namibian border the habitat changes into mixed deciduous woodland and then into mixed acacia and mopane woodland in the region surrounding the Delta.

Sites were selected to ensure there was a suitable coverage within the principal habitats of the upper catchment (Fig. C). The lower mixed deciduous woodland was not represented due to a last minute change in the logistical arrangements of the trip, which restricted us sampling at Caiundo and close to Savate. These sites will be included in the next survey.

Figure C, Surveyed sites in relation to the habitat classification of the Cubango-Okavango basin

It was regrettable that the survey sites of Caiundo and Savate were not sampled, as they were also key sample sites in relation to a stratified sample of the profile of the river system (Fig. D). As a result, the upper, higher altitudes of the upper catchment were samples but not the lower sections, which might include an array of different species, where rainfall and climate differ to higher altitudes.

9 | P a g e

Figure D. Surveyed sites in relation to the altitude profile of the Cubango

The location of sites known by local stakeholders for their importance for biodiversity (Biodiversity Hotspots defined through a series of consultative workshops within Luanda, Kuando Kubango, Huambo and Bie) The main objective of the survey is to improve our knowledge and the documented evidence of biodiversity within the upper catchment of the Cubango-Okavango system. There is however already an extensive level of undocumented knowledge of biodiversity, known by the local stakeholders working and living within the region. To ensure that SAREP captured this local knowledge a series of workshops were held across the basin in Kuando Kubango, Huambo and Bie, as well as two further workshops in Luanda. Information gathered from these workshops helped SAREP to define known biodiversity hotspots; sites of high biodiversity importance. The location of these hotspots was then used in the selection criteria for the survey (Fig. E).

Figure E. Surveyed sites in relation to identified biodiversity hotspots within the upper basin

10 | P a g e

The practicality of accessing the site and the ability to re-sample the site effectively and efficiently in the future This survey is not planned to be the only biodiversity survey of the upper catchment, with subsequent surveys planned to occur in early November 2012 and in 2013. Further surveys must also be undertaken as part of the monitoring processes within the basin. These monitoring surveys replicated on an annual or tri-annual basis will help OKACOM to assess the impact of human development in the upper catchment on biodiversity and water quality.

Survey sites were selected in proximity to main, tar roads and next to bridges so that access to the rivers was quick, easy and efficient. Sampling occurred in close proximity to people and further up and downstream of human activity. While this interference could have affected sample quality, it was felt this impact would be minimal.

c. Sampled Taxa The objectives of the survey were to improve our overall knowledge of the nature, extent and status of biodiversity within the upper catchment. The area has received minimal attention prior to this survey and the inclusion of any taxa would have helped to contribute towards science and our understanding of the system.

The timing of the survey in the early dry, winter had a bearing on the selection of the sampled taxa. An aquatic based survey helps to limit the constraining influence of the colder, drier sampling period and many of the terrestrial taxa which were originally considered were left out of this survey. Those excluded taxa included; terrestrial insects (focused on ants), small mammals, birds, plants and molluscs. The decision to exclude these taxa was made at the last minute and was unavoidable. While regrettable the smaller group of specialists, all of which had quite similar site requirements and sampling period needs, helped with the logistics of this first survey. The knowledge gained and experiences learnt will help to ensure that a larger, more dynamic group of specialists are as equally successful as the first.

Taxa included within the survey were;  Fish  Dragonflies  Amphibians  Reptiles

It was anticipated however that both the diversity of species and number of specimens amphibians and reptiles would be affected y the colder, drier month. Dragonflies are also affected by this period, but we hoped to capture the latter end of the dragonflies, amphibians and reptiles. It was important to take other considerations into context with this first survey than just the season. Issues of defining g critical indicators, developing a baseline of easily recognisable species, assessing those taxa which form a provisioning service were also of important concern and of building on an existing baseline level of knowledge for the basin.

Fish were selected as a critical provisioning resource within the system and on which an extensive level of knowledge is known for the lower sections of the basin and for adjoining basins. Building on our knowledge of fish for the basin will help us understand the importance of linkages with adjoining basins and ultimately to also assess the impact of man on the system.

Dragonflies are one of the best biodiversity indicators for water quality (the nymphs depend upon clean water to survive) and as a group of species are far more easily recognisable than other

11 | P a g e freshwater invertebrates. An extensive database of dragonfly species records is also available for lower sections of the basin as well as knowledge about their ecology. Building on this knowledge will ensure a biodiversity index for the Okavango catchment is developed that can help us to monitor the quality of the system into future years.

Amphibians are well known as good indicators of ecological change within aquatic / wetland systems, due to their high degree of sensitivity, either during tadpole stage or as adults. Frogs and toads are also easily monitored through both sight and sound.

C. OVERALL RESULTS a. Biodiversity highlights from the Fisheries specialists team (Prof P. Skelton & R. Bills) This report covers the initial results from 25 sites of the survey, covering Menongue, upper Cacuchi, Cutato and Cubango around Huambo and the Cuito and its tributaries en-route to Cuito Cuanavale. A preliminary estimate is that at least 73 species (of about 96 recorded from the system) have been collected. Several (4 or possibly 5) un-described species of fish have been found in the Okavango system, including small barbine minnows, such as the Hyphen Barb (Barbus bifrenatus) that show a mixture of Okavango Delta, Kuenene and Cuanza affinities, highlighting the potential and importance of speciation across these watersheds.

There were two key outstanding findings, firstly that there was a total lack of certain groups from the Okavango Delta region found in the upper catchment. In the upper Cubango systems the Synodontis and Clariidae catfishes, as well as the Sargochromis and Serranochromis cichlids were represented by single species and these themselves seem different from anything in the lower rivers. The second key finding was that the Cubango and its upper tributaries have a fundamentally different fish assemblage to that of the Cuito and its tributaries. The Cuito catchment is most similar to the lower Okavango and Delta system in Namibia and Botswana, but also has a number of species that appear to be different to the forms encountered in the Delta region. For example the Hyphen Barb (Barbus bifrenatus) from the Cuito is distinctively marked like that in the Quanza system and unlike the markings from the Delta form. Also present within the Cuito sub-drainage are species like the spottail barb (Barbus afrovernayi) and the copperstripe barb (Barbus multilineatus) that have not been found in the Cubango sub-drainage.

Figure E, Barbus bifrenatus – spots (Cuito River) Figure F, Barbus bifrenatus – double spot (Cuito River)

Figure G, Barbus bifrenatus- long-dash (Cuito River) Figure H, Barbus multilineatus (Longa River)

12 | P a g e

Figure I, Barbus afrovernayi (Cuito River)

Several distinct varieties of the Orangefin Barb (Barbus eutaenia) were also discovered in the various tributaries of the Cuito, which differ to those found in the Cubango. There is currently confusion over the classification of this family, with many distinct species found in the upper Zambezi. Results from this survey may also show there to be several distinct species in the upper Cubango-Okavango catchment as well.

Figure J, Barbus cf. eutaenia (Cuanavale River) Figure K, Barbus cf. eutaenia (Cuebe River)

Figure L, Barbus cf. eutaenia (Cacuchi River) Figure M, Barbus cf. eutaenia (Cuanavale River)

The upper Cubango drainage also have distinctive species have not been found in the Cuito system, e.g. the Cubango Kneria (Parakneria fortuita), the topstripe barb (Barbus dorsolineatus), and a possible un-described large mouth bream (Serranochromis) species. There are many species with unusual features (colouration etc.) that need detailed laboratory examination in order to determine their taxonomic status, such as a possible new species of Okavango Rock Catlet (Chiloglansis) found on the Cacuchi, while an example of the ‘real’ Okavango Rock Catlet (Chiloglansis fasciatus), a fish adapted to rocky, rapid habitats, was found close to the source of the Cubango, that may lead to a review of many samples found in the upper Zambezi. Several samples of a longtailed spiny eel (Mastacembelus) were found on the Cuchi and Cacuchi, which may help to resurrect a species previously described from the upper Okavango system, but then placed within a different African eel species (Mastacembelus frenatus).

13 | P a g e

Figure N, Parakneria fortuita (Cubango River) Figure O, Barbus dorsolineatus (Cubango River)

Figure P, Chiloglanis sp.nov. (Cubango River) Figure Q, Chiloglanis sp.nov. (Cubango River)

Figure R, Serranochromis sp. nov. (Cacuchi River) Figure S, Mastacembelus. sp. nov. (Cacuchi River)

The Southern African Institute of Aquatic Biodiversity (SAIAB) has taken tissue samples for molecular DNA analysis (the international barcode of life gene CO1) and will be used to verify identities and indicate taxonomic relationships and issues with these species. The evidence suggests that the upper Cubango system has links with the Kunene and Cuvo de Queve and that the Cuito is clearly linked to the Quanza. In addition to these observations it is evident that practically every tributary has its own distinctive characteristic fauna to a certain degree – and there are a number of species collected in one or two sites only, sometimes in low numbers. Finally some species or species groups have not been found in any abundance, e.g. the Sargochromis group of breams, largemouth breams (Serranochromis spp.) and large head catfishes (Clarias gariepinus). Even widespread and common species like the southern mouthbrooder (Pseudocrenilabrus philander) is surprisingly rare in the collections.

A brief survey of Cuvo de Queve River to the north of Huambo found a collection of fish species unrecorded by science, highlighting the need to survey more of the diverse river systems of Angola, especially along watersheds between different river systems.

EUS disease was found on some fishes within the upper catchment. To confirm these findings specimens will be sent them off to a specialist in RSA. If confirmed then this will be the first record for Angola.

14 | P a g e b. Biodiversity highlights from the Reptile / Amphibian specialists team (Werner Conradie) The herpetofauna of Angola remains one of the most poorly documented on the continent. The only synthesis dates from the 19th Century (Bocage 18954). Due to the protracted civil war which concluded in 2002, modern biodiversity surveys in the country are just beginning, (e.g. Huntley 2009 & 2011) and large areas of the country have never been surveyed.

Most previous reports on the country’s herpetofauna have been restricted to the western or southern regions, e.g. Monard (19315, 19376), Schmidt (19337), Parker (19368), Mertens (19389), Hellmich (195710), FitzSimons (195911) and Branch & McCartney (199212). The current herpetological survey was undertaken by Werner Conradie, experienced in Africa herpetofauna and based at Bayworld (formerly Port Elizabeth Museum), South Africa. Unfortunately reptile and amphibian activity at the time of the year the survey took place was relatively quiet as breeding activity had declined and many species were entering a period of seasonal dormancy.

In total 19 amphibian and 17 reptile species were documented, the greater majority with voucher specimens. Every record made during the survey within the Angolan Okavango system represents a new confirmed record to the Angola fauna. No historical or museum collections exist for the surveyed area, except for a handful of specimens collected from Cuito Cuanavale area (Branch & McCartney 1992).

Species of special interest:  A rarely sampled aquatic skink was recorded (Trachylepis ivensi), it is a new record for the Cubango-Okavango system (previously recorded from the Zambezi and Congo drainage systems), found at the floodplains of the Cuebe River north of Menongue..  New records for a poorly known frog (Ashey reed frog - Hyperolius cinereus), previously only known from one site in southern Angola (Caluquembe) are now reported from four distinct sites within the Okavango system and becomes a new species recorded for the Okavango system. Expanding the known distribution eastwards by 350 km.  Only the second record for Ellenberger’s Long Tail Seps (Tetradactylus ellenbergeri), which was found on the Luassingua River, expanding the known distribution south by 400 km.

4 Bocage, J.V. Barboza du (1895). Herpetlogie d’Angola et du Congo. Lisbon. 203 pp. 5 Monard, A. 1931. Mission scientifique Suisse dans l'Angola. Résultats scientifiques. Reptiles. Bull. Soc. Neuchâtel Sci. Nat., vol. 33, pp. 89-111. 6 Monard, A.1937. Contribution à l'herpétologie d'Angola. Arq. Mus. Bocage, Lisbon 8:19-153. 7 Schmidt, K. P. 1933. The reptiles of the Pulitzer Angola Expedition. Annals of the Carnegie Museum 22 (1): 1-15 8 Parker,H.W. 1936. Dr. Karl Jordan's expedition to South-West Africa and Angola: herpetological collections. Novit. Zool. (London) 40: 115-146 9 Mertens, R. 1938. Amphibien und Reptilien aus Angola, gesammelt von W. Schack. Senckenbergiana, 20:425- 443 10 Hellmich, W. 1957. Herpetologische Ergebnisse einer Forschungsreise in Angola. Veröff. Zool. Staatssammlung München 5: 1-91 11 FitzSimons, V. 1959. Some new reptiles from southern Africa and southern Angola. Annals Transvaal Mus. 23:405-409 12 Branch, W.R., and McCartney, C.J. (1992) A small collection of reptiles from southern Angola. J. Herpetol. Assoc. Afr. 41: 1 3.

15 | P a g e

Figure T, Trachylepis ivensi (Soba Matias) Figure U, Tetradactylus ellenbergeri

Figure V, Hyperolius cinereus (Huambo) Figure W, Hyperolius cinereus (Menongue)

c. Biodiversity highlights from the Dragonfly specialists team (Jens Kipping) Angolan dragonflies were only poorly surveyed in the past. Of the 84,000 records of the Odonata Database of Africa (ODA) only 656 are from Angola. The majority of them were collected in the 1930s through to the 1970s. Of many old records identification of species and/or geo-referencing of the sampling localities remain doubtful. The Cubango catchment in general is less well sampled for dragonflies (only 56 records) than most of Angola, forming a gap on Odonata distribution maps. The Swiss zoologist Albert Monard collected in 1928/29 and 1932/33 some dragonflies within the region (Longfield 194713, Ris 193114). Later Pinhey (196615, 197516) added a few more records from the Cubango catchment. In contrast to this situation the dragonfly fauna of the Namibian part of the Okavango River and the Okavango Delta belong to the best known on the African continent. Also the knowledge about neighbouring Zambia is comparably good. On the African continent dragonflies in general reach their highest diversity along the watersheds of the major river systems and also the amount of endemic species is high on highlands and headwater systems. From this perspective the Odonata fauna of the upper catchments of the Okavango was expected to be rich and unique but almost unknown.

On the SAREP fieldtrip 95 species of dragonflies were recorded in total. The Odonata fauna found within the Cubango system shows more similarities to the upper Zambezi catchment than to the Okavango Delta fauna, only the Odonata fauna within the Cuito system is more similar to that found in the Delta (although with 2 possible new distinct species found in the Cuito). The relatively high content of at least 5 new species show some unique characters of the upper catchment fauna. The new results show clearly that the Cubango catchment fauna belong to the 'Zambezian' complex

13 Longfield, C. 1947. The Odonata of South Angola. Arquivos do Museo Bocage, Lisboa 16: 1-31. 14 Ris, F. 1931. Odonata aus Süd-Angola. Revue Suisse de Zoologie 38: 97-112. 15 Pinhey, E.C.G. 1966. New distributional records for African Odonata and notes on a few larvae. Arnoldia Rhodesia 2(26): 1-5. 16 Pinhey, E.C.G. 1975. A collection of Odonata from Angola. Arnoldia Rhodesia 7(23): 1-16.

16 | P a g e fauna, delineated by the Palaeo-Chambeshi-Katanga region. This fauna is extremely rich and unique and makes the region an 'overlooked biodiversity hotspot'.

 95 species in total found within the survey so far, 18 of which are new records for the whole of Angola  54 species have been found as new for the Angolan Cubango catchment, of which 26 are new to the whole Okavango system  At least 5 species are new to science belonging to the genera Chlorocypha, Zygonyx and Pseudagrion, which were found on the Cuito, Cuchi and Cubango rivers.  The survey has now shown there to be up to 175 species of dragonfly within Angola  The survey fills large gaps in our knowledge of the distribution of species within Africa.

The following records are remarkable:  From the group of "blue" Chlorocyphids the true Chlorocypha crocea could be found, a species endemic to Angola and last recorded in 1960. This species might act as a top indicator for small intact headwater streams and dambos. A new species similar to C. crocea has been found. This helps to clarify the taxonomic status of C. crocea, C.c.bamptoni and a species found in 2009 in the Serra de Chela by W. Tarboton. A possible new species belonging to the "red" Chlorocyphids could be recorded from Cuito and Keve rivers.  One new species of Pseudagrion was found at the Cubango source and headwaters. It is similar to Pseudagrion melanicterum but with very distinct male appendages.  Another new species of Pseudagrion was found from the Cuito, Longa and Luassingua rivers is similar to P. fisheri and P.greeni but with very unique coloration and male appendages.  A new species of Zygonyx was recorded from the Cuvo de Queve River north of Huambo. It is similar to Zygonyx flavicosta but with very different genitalia and colouration.  The rare Congothemis leakeyi was recorded from Cuito and Luassingua rivers. It is the first record since 1969 and outside Zambia and DRC.  A male of Phyllogomphus annulus/coloratus found from the Cuanavale River is far outside the known distribution or is a new species.  Africa's smallest damselfly, Agriocnemis bumhilli Kipping et al. 2012, recently described from the Kwando River in Namibia's Caprivi strip were found at several tributaries feeding the Cuito River.

Many species formerly only known from northern Zambia and Katanga were recorded. These are Aethiothemis bequaerti, Crocothemis brevistigma, Nesciothemis fitzgeraldi, Trithemis anomala, Orthetrum macrostigma and Orthetrum monardi. Records of range restricted and localized Okavango Delta species like Ictinogomphus dundoensis, Pseudagrion deningi, Pseudagrion rufostigma and Pinheyagrion angolensis help to understand their real distribution.

Figure X, Chlorocypha sp.nov. (Cuchi River) Figure Y, Pseudagrion sp .nov. (Cubango River)

17 | P a g e

Figure Z, Pseudagrion sp.nov. (Longa River) Figure Aa, Zygonyx sp. nov. (Cuvo de Queve River)

d. Biodiversity highlights from the Crocodile specialists team (Vince Shacks) This initial aquatic biodiversity survey was used to gather feedback from local communities on their knowledge of the distribution of crocodiles in the upper catchment, rather than undertaking extensive surveys. Local information proved to be useful, with extensive feedback and numerous concerns on crocodile related conflict gathered from across the region. Reports on the presence of crocodiles and of crocodile attacks were gathered from communities at Soba Matias, north of Menongue, around Mumbué, and also around Cuito Cuanavale. Sections of the rivers in these regions displayed suitable areas for breeding, with well defined banks and thick reedbeds for cover. However the energised feedback from communities in response to the discussion on crocodiles suggests that human-crocodile conflict is a significant problem, which could lead to persecution. There were no sightings of crocodiles during the survey and so without undertaking nocturnal surveys it will be hard to confirm this provisional community feedback. The nature of the habitat would however suggest that the upper catchment is ideally suited for crocodiles. In areas close to human settlements where many of the sampling sites for this survey took place the crocodiles may hide deep within reedbeds, emerging only at night.

A boat was not brought along on the trip for logistical reasons, but a small, three-man, inflatable semi-rigid boat was borrowed from the Ministry of Water & Energy in Menongue to assist with provisional assessments of suitable crocodile breeding habitat. The boat almost sank on the Cuito River, hindering surveys, but the Halo Trust were kind enough to lend us one of their aluminium boats, which was used to assist in fish surveys. Information collected from this survey has been very constructive in helping to plan for future surveys, which should help to determine the extent of the crocodile population in the upper catchment.

18 | P a g e

Taxa Results

1. FISH

1.1. Earlier ichthyological explorations.

The earliest collection of fishes from the Okavango reported in the scientific literature was by the French author Castelnau in 1861. The origins of this collection are somewhat vague (Barnard 1948 records that it was made by one of Castelnau’s ‘préparateurs’), and whilst part of the collection only survives in the Paris Museum (MNHN), its provenance is from Lake Ngami at the southern end of the Delta reaches in Botswana. Several prominent species derive from this earliest collection – Mormyrus lacerda, Hydrocynus vittatus, Hepsetus cuvieri, Clarias ngamensis, Oreochromis andersonii, and Serranochromis thumbergi, and several others whose names are unfortunately replaced through taxonomic uncertainty and protocol. George Boulenger (1911) working at the British Museum (Natural History) in London reported on a collection also from the Delta region and Lake Ngami made by Mr. RB Woosnam (Barnard, 1948). The comprehensive catalogues of African freshwater fishes in the BMNH published by Boulenger (1909-1916) and in the SAM by Gilchrist & Thompson (1913-1915) included species from the Okavango, again drawn from the southern reaches in Botswana. Nichols & Boulton (1927) described three species from collections drawn from the Cunene and upper Quanza Rivers, made by the Vernay Angola Expedition mounted by the American Museum of Natural History in New York. Fowler (1930) described species from the upper Quanza, and in 1935 described the fishes taken in the Delta reaches by the Vernay-Lang Expedition (Fowler, 1935). Pellegrin (1936) described the fishes collected by the Swiss Expeditions of 1928-29 and 1932- 33 and deposited in the MNHN (Paris) and the Museum at La Chaux de Fonds in Switzerland. The names of several of the species described by Pellegrin as new have been placed in synonymy but may well on re-inspection and review be resurrected. Trewavas (1936) described several new fish species collected by the British scientist Dr Karl Jordan during an expedition to Namibia (SWA) and Angola in 1935. The Angolan species were from the Cunene River and certain coastal systems with headwaters in the highlands close to the source of both the Cunene and the Okavango. Several of these species appear to be found in our collections, either in the Okavango system or the Cunene or the Cueve. Barnard (1948) described a collection made in 1939 at Rundu in Namibia and advanced the taxonomy of species extensively but without providing any new descriptions. The most significant work after WWII concerned a German project in the late 1950s with the collection lodged in the Hamburg Museum (Ladiges 1964) and the classic overview presented by Poll (1967) of the fishes of Angola. Several lesser collections were made but not publicly described by Grahams Bell- Cross in 1965 (collection deposited in the BMNH, London) and Mike Penrith (State Museum Windhoek) in the early 1970s. The State Museum Windhoek collections have been used by Penrith (1973) and Greenwood (1984) to describe new species from the Okavango and Cunene rivers. Further collections in Angola itself were disrupted by the three decade long civil conflict (1975-2005)

During this extended time, fish surveys with collections and fisheries investigations either in the Delta reaches in Botswana or along the Okavango River on the Namibian shore were made by several different parties from South Africa, including Jubb & Gaigher (1971), Skelton (et al. 1985) and Dr G Merron and colleagues from the JLB Smith Institute of Ichthyology (now SAIAB) in Grahamstown. In addition there has been a flurry of fisheries and ichthyological research in Namibia that embraces the Kavango reaches of the Okavango River and the Cunene River since the turn of

19 | P a g e the millennium (e.g. Hay et al. 1996, 1998, 1999, 2000, Kramer et al. 2003, 2007, 2011, 2012) and also a revival of interest and research on biodiversity in Angola itself, including this project. The present state of taxonomy of many species is decidedly uncertain as a result of renewed attention and investigation using new technologies (DNA molecular analysis, electric discharge analysis) and new samples.

1.2 Methods

1.2.1 Sampling areas and sites Sampling sites within the broader region were chosen in consultation with the entire scientific and logistical team. Attempts were made to sample a range of different water bodies e.g. larger rivers through to small streams, swamps and isolated pools. Even at sites this principle was adopted with multiple microhabitats being sampled where possible e.g. riffles, runs, pools, marginal vegetation, marginal waters. Important additional factors were ease of access and security. Sites sampled are shown on the map below (Figure 1.2 and Table1.1) and some sites are shown in Figure 1.3.

Figure 1.2. A map showing the sampling sites during the SAREP Angolan (Okavango) field trip (6-26th May 2012).

Table 1.1. Collection sites for the SAREP Angolan (Okavango) field trip (6-26th May 2012). Site Site name Coordinate River number South East RB12-

Ang01 Kabumbe lodge dam margin 14° 40' 32.2" 17° 44' 09.9" Cuebe Ang02 Square Ponglesia - Cubango, Menongue 14° 56' 22.9" 17° 43' 07.7" Cuebe Ang03 Stream/seepage - 5km south of Menongue 14° 44' 46" 17° 40' 06.4" Cuebe Ang04 Flood plain below Soba Matias 14° 26' 20.8" 17° 48' 53.7" Cuebe Ang06 Small dambo stream, Soba Mathias 14° 15' 25.4" 17° 46' 42.7" Cuebe Ang07 Cacuchi bridge near Mambue 13° 35' 39.6" 16° 52' 49.8" Cachuchi

20 | P a g e

Ang08 Small stream near Chungueia village, Bie' 13° 41' 37.3" 17° 03' 40.5" Cacuchi Ang09 Cacuchi tributary (culvert), Mumbue, Bie' 13° 16' 46.7" 14° 44' 46.1" Cacuchi Ang10 Pool in roadside ditch, Bie' 12° 52' 20.7" 16° 46' 02.7" Cacuchi Ang11 East of Bailundo town north of Huambo 12° 12' 11.3" 15° 35' 42.5" Qeve Ang12 Stream near Halo camp site 12° 44' 14.4" 15° 49' 05.0" Cunene Ang13 Cubango falls, agricultural station, Huambo 12° 40' 16.6" 16° 06' 41.2" Cubango Ang16 Cutato river at bridge 12° 34' 20" 16° 23' 49" Cubango Ang17 Just south of Chitembo 13° 31' 41.1" 16° 45' 31.4" Cacuchi Ang18 Satchijamba, Cuele River near Menongue 13.71337° 17.08962° Cacuchi Ang19 Luassingua River, 55 km east Menongue 14° 35' 24.2" 18° 10' 14.4" Cuito Ang20 Cuito Canavale bridge 15° 10' 16.4" 19° 11' 39.3" Cuito Ang21 Sand pits near C-C bridge 15° 10' 5.8" 19° 11' 51.8" Cuito Ang22 Confluence of Cuito and Canavale rivers 15° 08' 19.1" 19° 11' 31.4" Cuito Ang23 Canavale river just above C-C confluence 15° 08' 08.5" 19° 11' 44.6" Cuito Ang24 Washing site west of Halo trust camp, C-C 15° 05' 12.7" 19° 08' 55.4" Cuito Ang25 Longa bridge on the C-C road 14° 36' 22.6" 18° 28' 02.2" Cuito Ang26 Low level road bridge, Cuchi town 14° 38' 58.9" 16° 54' 23.7" Cubango Ang27 Cuchi tributary stream/swamp 14° 39' 07.4" 16° 54' 13.3" Cuchi Ang28 First stream east of Cuchi 14° 39' 10" 16° 55' 57" Cuchi Ang29 Stream on road Cuchi – Menongue 14° 39' 58.4" 16° 58' 42.3" Cuchi Ang31 Main road between Cuchi and Menongue 14° 42' 07.8" 17° 22' 39.9" Cacuchi Ang32 Kaisosi lodge, Rundu outskirts 17° 52' 25.3" 19° 49' 52.9" Okavango Ang33 Popa 18° 07' 18.2" 21° 35' 00.2" Okavango Ang34 Toteng bridge Ang35 Thamalakane river at back packers bridge 19° 56' 38.8" 23° 29' 22.5" Thamalakane Ang37 Rundu downstream of Water affairs 17° 54' 28.3" 19° 45' 51.3" Okavango

21 | P a g e

Figure 1.3. Selected collection sites during the SAREP Angolan (Okavango) field trip (6-26th May 2012).

RB12-Ang02 Quebe River downstream from RB12-Ang07 Cacuchi River at main road Menongue 9-5-2012. bridges 11-5-2012.

RB12-Ang11 Qeve River near Bailundo town RB12-Ang13 Cubango source stream S-E of north of Huambo 14-5-2012. Huambo 15-5-2012.

RB12-Ang23 Cuanavale River 19-05-2012. RB12-Ang25 Longa River 20-5-2012.

22 | P a g e

1.2.2 Fish capture equipment and methods. Fish were captured using the following equipment.  a 12 Volt DC back-pack electric-fisher (Samus 725G)  a 6m minnow seine net (4mm mesh) with a central bag  fyke nets (single trap)  baited traps  a variety of dip-nets.  purchase from fishermen, or pick-up remains from riverside.

Figure 1.4. Fish capture methods.

Electric-fishing at the Cubango source. Seine netting on the Luancingwa River.

Capture methods. Usually multiple methods were used at sites and ideally overnight collections allowed most of the methods to be employed. Almost all methods are selective and result in slightly different species being collected. In some instances equipment was used together e.g. seine nets with electric-fishing. Fishing effectiveness was enhanced by placing the seine net across the stream channel and electric-fishing downstream towards this “stop-net”. Where we were able to fish overnight fyke nets and traps were used and were very effective. Placement of these is important and they were set on smooth substrates out of currents, traps were always baited with either kitchen waste of dead fishes.

The numbers of fish collected at most sites were relatively few and given our uncertain field identifications we decided to preserve most specimens collected. Fishes were fixed in 10% formalin in the field by placing directly into formalin and transported back to Grahamstown in ziplock bags in drums. In the laboratory fixed specimens were transferred through an ethanol wash series (10%, 50% and 70%) to a final concentration of 70% ethanol at which time they could be easily worked with.

In the field select specimens were also sampled for genetic material, usually a piece of muscle was removed and more rarely a fin clip taken. This tissue was placed into a 2ml Ependorf tube with 95% ethanol. The tube is pre-labeled and the numbers were cross referenced with the voucher specimens on our data sheets. Select individuals were also photographed in the field to record live colouration. Small fish were placed into a white porcelain dish, immersed in clean water and photographed on high F-stop settings using flash. Larger fishes we simply photographed on appropriate backgrounds and lighting on camera auto-settings.

23 | P a g e

1.2.3 Site data collected. Basic details of collection sites were recorded on data sheets together with information concerning specimen sampling (genetic samples) and photographs. It is important to be able to link tissues and photographs with the specimens they originate from. An example of a data sheet is given in Figure 5 below. These data are transferred into the SAIAB database when specimens are formally catalogued into the SAIAB collections. On cataloguing, data and specimens become available to the general public through the SAIAB loan system (http://saiab.ac.za/loans/index.asp).

Figure 1.5. Data record sheets.

24 | P a g e

1.2.4 In the laboratory Collections are returned to SAIAB in 10% formalin. After two to three weeks of fixation these are then washed through a series of ethanol concentrations up to 70% ethanol which is the final storage preservative. This takes about a week and after this process fishes can be safely handled and sorted. Sorting involves sorting specimens into obvious species groups based on morphological characters and colouration patterns. Once done they are then keyed out and identified, counted and measured ready for cataloguing. These data gathered in the laboratory are added to the original field data sheets which thus evolve over time. As the cataloguing is completed the catalogue numbers are also added to data sheets. At this stage it is likely that species names will change from those allocated in the field based on examination of literature, keys and better examinations.

Tissues for genetic examination had their ethanol replaced with new 95% ethanol and were placed into -80C freezers for long-term storage. Some 300 specimens out of approximately 600 were sent off for bar-coding and some of the results are given below in relevant sections.

1.3 Fish survey results

1.3.1 Fish catch data - introduction The following section deals with species of interest and complex groups depending on the level of diversity within the group. Some photos and parts of the bar-code trees are given to illustrate diversity and enhance discussions of taxonomic issues. After some general comments the groups are dealt with in phylogenetic order starting with the Family Mormyridae (snout fishes).

25 | P a g e

Barcode and morphological comparisons have been made with specimens from the Kunene, the upper Zambezi in Zambia, the Zambian Congo systems and across to Malawi in order to highlight the Angolan Okavango fishes in a broader regional context.

For this expedition the majority of collections were made in the upper Okavango system with a single site each being in the Qeve (a coastal system north of Huambo) and the Kunene (in Huambo). The Qeve in particular appears to have a distinct fauna with most of the species differing significantly from those of the neighbouring Okavango. Partly because of newly recognized diversity within the Angolan Okavango and partly because our sampling crossed over into two other catchments the species count is considerably higher than for the previously known Okavango system (Skelton et al. 1985, 2001).

A summary of the fishes captured at the 37 sampling sites during the Angolan expedition is given in Table 1.2. This gives details of the numbers of each species caught, the number of sites at which they were caught, mean and catch ranges and size ranges. Full catch details are available through the SAIAB web-site (www.saiab.ac.za). Not all species have been identified to species level and some certainly are undescribed and new to science. Additional surveys, more material, more analyses and time are probably required to make a better assessment of the biodiversity of this region. The composition of families (number of genera and species) are given in Table 1.3.

Table 1.2. Fish collections summary for Angolan (Okavango) field trip (6-26th May 2012). Species name Total Composition Number count Range of sites

Amphilius sp. 1 0.5 1 Amphilius uranoscopus 92 0.9-16.3 (5.4) 9 Aplocheilichthys cf. hutereaui 1 1.9 1 Aplocheilichthys cf. johnstoni 2 3.3 1 Aplocheilichthys cf. johnstoni 'mesh' 3 1.6 1 Aplocheilichthys cf. johnstonii, side-stripe 21 6.3-8.5 (7.4) 2 Aplocheilichthys johnstoni 226 1.9-65.0 (22.9) 11 Aplocheilichthys cf. katangae 69 0.9-97.0 (20.4) 7 Aplocheilichthys sp. pygmy 97 8.3-81.0 (37.6) 3 Barbus afrovernayi 45 2.9-13.1 (8.7) 3 Barbus barnardi 11 1.4-11.6 (6.5) 2 Barbus barotseensis 9 0.8-3.1 (1.3) 4 Barbus bifrenatus 81 0.5-46.2 (9.10 10 Barbus brevidorsalis 1 0.6-5.9 (2.1) 6 Barbus eutaenia 233 1.2-58.2 (12.9) 14 Barbus eutaenia 'red eye' 5 2.0 1 Barbus eutaenia 'sharp nose' 207 1.3-25.1 (10.7) 6 Barbus eutaenia 'thin line' 190 3.1-72.2 (19.4) 6 Barbus fasciolatus 8 0.6-2.1 (1.3) 4 Barbus haasianus 7 0.6-28.6 (14.6) 2 Barbus kerstenii 14 0.4-3.8 (1.4) 4 Barbus kessleri 121 0.4-94.1 (22.5) 10 Barbus lineomaculatus 17 1.2-8.8 (4.5) 4 Barbus mocoensis 60 2.3-9.4 (5.8) 5

26 | P a g e

Barbus multilineatus 7 13.2 1 Barbus paludinosus 67 0.8-7.9 (2.7) 4 Barbus poechii 2 0.9-5.7 (3.3) 2 Barbus radiatus 96 1.7-42.9 (14.4) 4 Barbus sp. 2 2.7 1 Barbus sp. 'qeve' 2 1.0 1 Barbus sp. 'qeve 2' 28 13.7 1 Barbus tangandensis 10 0.5-5.2 (2.9) 2 Barbus thamalakaneneis 134 0.4-78.7 (14.5) 11 Barbus trimaculatus 37 0.8-4.5 (2.7) 2 Barbus unitaeniatus 16 3.1-30.0 (16.5) 2 Barbus viviparus 55 0.8-53.8 (11.1) 6 Brycinus lateralis 100 1.0-54.5 (19.6) 5 Chetia sp. 1 1.7 1 Chiloglanis fasciatus 103 1.8-8.1 (5.3) 5 Chiloglanis sp. 1 28 13.7 1 Chiloglanis sp. 2 11 5.4 1 Chiloglanis sp. 'kunene' 16 6.4 1 Clarias cf. liocephalus 9 0.1-1.5 (0.8) 6 Clarias ngamensis 3 5.1 1 Clarias stappersii 3 0.8-1.7 (1.3) 2 Clarias theodorae 1 0.1-1.0 (0.5) 4 Cyphomyrus cubangoensis 34 0.6-7.9 (3.5) 4 Hemichromis elongatus 26 0.5-18.0 (4.7) 7 Hemigrammocharax machadoi 4 1.1-5.0 (3.0) 2 Hemigrammocharax multifasciatus 79 1.9-18.6 (6.1) 7 Hepsetus cuvieri 5 05.-1.9 (1.4) 3 Hippopotamyrus ansorgii 3 2.4 1 Hippopotamyrus sp. 1 26 0.4-8.9 (2.5) 6 Hippopotamyrus sp. 2 10 12.8 1 Hydrocynus vittatus 5 0.5-1.8 (1.1) 2 Labeo cylindricus 57 0.6-7.9 (3.4) 7 Labeo sp. 'pointed head, no red eye' 36 4.5 1 Labeobarbus codringtonii 1 0.4 1 Labeobarbus sp. 'qeve' 2 1.0 1 Marcusenius altisambesi 28 0.1-13.6 (3.4) 8 Mastacembelus sp. 'cuele' 4 3.1 1 Mastacembelus frenatus 11 0.4-6.4 (2.2) 4 Mastacembelus sp. 'qeve' 5 2.5 1 Micralestes acutidens 191 0.5-38.4 (11.3) 11 Micralestes argyrotaenia 10 8.0 1 Micralestes cf. argyrotaenia 14 6.9 1 Microctenopoma intermedium 14 7.3 1 Mormyrus lacerda 1 0.1 1 Nannocharax macropterus 61 0.4-17.9 (7.1) 10 Opsaridium zambezense 216 0.5-35.9 (11.0) 9

27 | P a g e

Oreochromis andersonii 7 0.5-3.9 (2.2) 2 Oreochromis macrochir 4 0.8-5.7 (3.3) 2 Parakneria fortuita 8 1.0 1 Parauchenoglanis ngamensis 18 0.4-9.4 (3.0) 6 Petrocephalus sp. 'black spots' 49 0.4-17.9 (3.8) 10 Petrocephalus sp. 'cunene' 2 1.6 1 Petrocephalus sp. 'light' 24 3.4-9.6 (6.5) 2 Petrocephalus sp. 'qeve' 52 25.5 1 Pharyngochromis acuticeps 65 0.4-6.8 (2.9) 12 Pollimyrus castelnaui 43 1.6-14.8 (5.5) 6 Pseudocrenilabrus philander 41 0.6-11.6 (4.6) 7 Rhabdalestes maunensis 13 8.4 1 Sargochromis sp. 18 0.6-6.7 (2.6) 8 Schilbe intermedius 32 0.6-7.9 (3.2) 5 Serranochromis angusticeps 9 4.8 1 Serranochromis cf. thumbergi 3 3.5 1 Serranochromis cf. macrocephalus 24 0.6-16.7 (3.8) 8 Serranochromis robustus 8 0.5-1.7 (1.2) 5 Serranochromis sp. 'red spotted' 1 0.1 1 Serranochromis sp. 9 0.9-3.8 (2.6) 3 Synodontis leopardinus 8 15.1 1 Synodontis macrostigma 51 0.4-25.0 (5.9) 11 Synodontis nigromaculatus 15 0.3-2.9 (3.1) 6 Synodontis sp. 1 7 3.1 1 Synodontis sp. 2 7 43.8 1 Synodontis thamalakanensis 1 6.3 1 Synodontis vanderwaali 2 0.9 1 Synodontis woosnami 4 0.6-12.5 (5.3) 3 Thoracochromis cf. buysi 8 6.4 1 Thoracochromis sp. 10 4.9 1 Tilapia rendalli 22 0.4-5.7 (2.2) 10 Tilapia ruweti 1 1.9 1 Tilapia sparrmanii 122 0.4-71.4 (11.6) 17 Zaireichthys sp. 12 1.7-3.9 (2.8) 2 Zaireichthys sp. 'brown' 4 3.8 1 Zaireichthys sp. 'brown blotch' 7 4.1 1 Zaireichthys sp. 'long barbels, pallid' 5 4.7 1

Table 1.3. The composition of fish families from the Angolan (Okavango) field trip. Family Genera Species Alestiidae 4 6 Amphiliidae 2 6 Anabantidae 1 1 Cichlidae 9 18 Clariidae 1 4

28 | P a g e

Claroteidae 1 1 Cyprinidae 4 31 Distichodontidae 2 3 Hepsetidae 1 1 Kneriidae 1 1 Mastacembelidae 1 3 Mochokidae 2 12 Mormyridae 6 11 Poeciliidae 1 7 Schilbeidae 1 1 Total 37 107

1.3.2 Species and ‘group’ accounts

Mormyridae – electric fishes, elephant snout fishes

The current taxonomy of the mormyridae in the Okavango and Zambezi systems is complex and unsettled. Recent work by Kramer and co-workers (Kramer & van der Bank 2011, Kramer et al. 2003, 2012) has indicated much greater diversity than was originally recognised. This expedition’s collections appear to have at least two, possibly four species of Petrocephalus, one Mormyrus, one Cyphomyrus, three Hippopotamyrus, a Pollimyrus and one Marcusenius. Based on external morphology, colour, and barcoding most of these appear to be distinct from similar forms from the neighbouring upper Zambezi tributaries in Zambia.

Cyphomyrus cubangoensis (Pellegrin, 1936) specimens have been shown by Kramer & van der Bank (2011) to be distinct from those of the upper Zambezi in Zambia.

Hippopotamyrus sp. from the Kunene is distinct from other rivers and this agrees with previous Kunene samples from lower in that system identified and described by Kramer & Swartz (2010) as Hippopotamyrus longilateralis. Hippopotamyrus specimens from site 17, Chitembo (Cacuchi sub- system) appear to be a little different from the remaining samples from this system (sites 2, 7 & 9) and require more detailed examination in the light of possibly being H. szaboi which is present in the Upper Zambezi. We didn’t catch Hippopotamyrus at the Qeve coastal system but if present the species might well be expected to be unique to that system.

29 | P a g e

Figure 1.6. Barcode results for southern African Hippopotamyrus (upper) and Petrocephalus (lower), recent Angolan collections in blue.

Our collections include at least two distinct species of Petrocephalus, which is in tune with Kramer et al. (2012) who recognised Petrocephalus okavangoensis and P. magnitrunci from the Delta reaches of the system. From the original description P. okavangoensis has a head with a fairly acute or angular profile, and a dark spot below the base of the dorsal fin and 37-38 lateral series scales (12 caudal peduncle scale rows), whereas P. magnitrunci has an obtuse rounded head, is intensely brown without markings and has 39-41 lateral series scales and 11-12 caudal peduncle scale rows. One of our species has distinct dark markings on the dorsal and anal fins while the other is light golden in colour without such markings. Further work is necessary to identify the species.

30 | P a g e

Figure 1.7. Mormyridae collected during the Angolan (Okavango) field trip (6-26th May 2012).

Cyphomyrus cubangoensis RB12-Ang07 Cacuchi River at bridge 5596b.jpg

Hippopotamyrus longilateralis (ansorgii 'kunene') RB12-Ang12 Halo camp stream 5862b.jpg

Marcusenius altisambesi RB12-Ang07 Cacuchi River 5454b.jpg

Mormyrus lacerda RB12-Ang07 Cacuchi bridge traps 5541b.jpg

31 | P a g e

Pollimyrus castelnaui RB12-Ang23 Canavale River near Cuito confluence 6082b.jpg

Petrocephalus sp. 'light' RB12-Ang04 Soba Mathias floodplain swamp 5433b.jpg

Petrocephalus sp. 'black point' RB12-Ang02 5385b.jpg

Petrocephalus sp. RB12-Ang11 14-5-2012 Qeve River 5744b.jpg

32 | P a g e

Kneriidae – shell-ears Interestingly no Kneria were collected during the survey despite sampling in several suitable habitats. Although these have never been recorded from the Okavango system previously they were thought to have been a possibility in headwater areas. Kneria are abundant in adjacent upper Zambezi tributaries in Zambia and an isolated population of a different species also occurs in Kunene tributaries.

A small number of Parakneria fortuita were collected from the Cacuchi River near the main road bridge (site RB12-Ang07) and these represent a significant find. All the specimens were collected with electric-fishing in rocky areas downstream of very large rocks and in a small side-channel to the main current. They are probably more abundant than our catches indicate as they are fast open water swimmers and are likely to easily evade electric fishing gear once disturbed.

Parakneria fortuita was described from the Cutato River near Chitembo (Penrith 1973) and have, up till now, been known only from the type locality. The habitat described by Penrith is remarkably similar to where we collected specimens. The collection site at the Cacuchi River was a slow channel running through a jumble of large rocks (see below).

Parakneria were also recently collected in northern Zambia and thought to be a new species. However, superficial external examination of all material and barcode results suggests the two populations are in fact the same species. It seems likely that there are other populations between these known widely separated sites and perhaps their low abundance coupled with the difficulty in collecting them is a factor in their absence from previous collections. Ideally barcode results require additional samples to give a broader understanding to the group. There are species of Parakneria in the Buzi River in central Mozambique and in the southern Congo systems close to the Angolan border.

Figure 1.8. Barcode results, collection site, recent and type specimens of kneriids.

Parakneria fortuita RB12-Ang07 Cacuchi River at main road bridge (Angola) 5559b.jpg

33 | P a g e

Parakneria n. sp. RB11-Kal17 Kasanjiku River, upper Zambezi (Zambia) 4309b.jpg

Parakneria n. sp. RB10-E020 Chisolo River, upper Zambezi (Zambia) 6-11-2010 0409b.jpg

Parakneria fortuita paratype BMNH 1973.11.10.1-2 3216c.jpg

RB12-Ang07 Cacuchi River just below the main road bridge 11-5-2012 5512.jpg

34 | P a g e

Cyprinidae – barbs and minnows, barred minnows and yellowfishes. Cyprinids were numerically the dominant fishes at most sites sampled with four genera and 31 species being represented in our catches. The diversity is high and seemingly quite distinctive for the region. Many of the typical Okavango minnows were either absent or rare from the western sector (Cubango, Cuchi branches) of the Angolan area surveyed. Some species that are common in the downstream Okavango Delta reaches started occurring in the rivers connected to the Cuito as we moved south (down-stream) and east of Menongue. Examples of these ‘delta’ species are Barbus fasciolatus, B. multilineatus, B. thamalakanensis, B. afrovernayi.

The species most common in the upper sections of the western rocky rivers are not known from the Okavango Delta and examples are: Barbus trimaculatus, Barbus mocoensis, Barbus kessleri, B. brevidorsalis, B. lineomaculatus and B. kerstenii. Whilst B. trimaculatus has previously been identified from the upper reaches by a number of workers including Fowler, 1935, Pellegrin, 1936, Barnard, 1948, and Ladiges (1964) this identification has been down-played against the recognition of B. poechii in the system (see Skelton et al. 1985). However field observations undoubtedly support the recognition of B. trimaculatus and not B. poechii in the western sectors at least, and the opportunity to delve deeper into the taxonomy is available through genetic analysis of these samples.

There are points of interest with many of the species/groups and much additional research is required to determine the identities of all the fauna. Of special interest are the barbs in what is loosely termed the ‘multilineatus-eutaenia’ group. These are typically small and brightly coloured minnows with serrated dorsal spines and there appear to be several new species within our collections from Angolan Okavango waters and wider upper Zambezi fauna.

Figure 1.9. Barcode results for the small barbus, serrated spine ‘multilineatus-eutaenia’ complex for southern Africa (recent Angolan samples blue highlighted).

35 | P a g e

Figure 1.10. Some of the cyprinid diversity dominated by the small minnows (Barbus).

Barbus sp. RB12-Ang13 5837b.jpg Barbus cf. eutaenia ‘thin line’ RB12-Ang07 5481b.jpg

Barbus cf. eutaenia RB12-Ang23 6062b.jpg Barbus multilineatus RB12-Ang25 Longa River 6140b.jpg

Barbus cf. eutaenia bright orange fins RB12- Barbus cf. eutaenia 'sharp' RB12-Ang23 6060b.jpg Ang02 5388b.jpg

Barbus afrovernayi RB12-Ang22 Cuito-Cuanavale Barbus radiatus RB12-Ang22 Cuito-Cuanavale confluence 6102b.jpg confluence 6087b.jpg

36 | P a g e

Barbus cf. dorsolineatus RB12-Ang03 5417b.jpg Barbus mocoensis RB12-Ang11 5738b.jpg

Barbus sp. RB12-Ang03 5416b.jpg Barbus cf. thamalakanensis RB12-Ang22 Cuito- Cuanavale confluence 6100b.jpg

Barbus bifrenatus RB12-Ang22 Cuito-Cuanavale Barbus trimaculatus RB12-Ang07 Cacuchi bridge confluence 6098b.jpg PHS 0512_04 059b.jpg

Barbus paludinosus RB12-Ang07 Cacuchi bridge Barbus cf. barotseensis RB12-Ang22 Cuito- PHS 0512_04 034b.jpg Cuanavale confluence 6092b.jpg

Opsariduim zambezense RB12-Ang13 5824b.jpg Barbus sp. 'silver' Qeve R. RB12-Ang11 5695b.JPG

Labeo spp. – Mudfishes Labeo were reasonably common in rocky habitats with over 90, mostly mature specimens being collected at 7 sites. In the field these were identified as Labeo cylindricus but then during laboratory examinations it was clear that there were in fact two morphologically distinct species. This was later confirmed with bar-code results. The two species occur in the same habitats /sites and we were distinguishing them based on paired fin sizes, and the width of the head and mouth size. In the genetic analysis the one is clustering with Labeo ansorgii an endemic of the neighbouring Cunene River and also would seem to have representatives as far east as Malawi.

37 | P a g e

Figure 1.11. Barcode results for Labeo specimens and Labeo cylindricus RB12-Ang07 Cacuchi River at bridge 5487b.jpg.

Alestiidae – tigerfish and robbers All four species of the family Alestiidae known from the Okavango were sampled, but three of these (Hydrocynus vittatus, Brycinus lateralis and Rhabdalestes maunensis) were not common in the collections, being more-or-less restricted to the Cuito branch. In addition based on colouration it appears that there might be differences between the western upper-reach populations of Micralestes, with orange tinted fins and the eastern Cuito population that displayed a distinct greenish hue and yellowish fins.

Figure 12. Alestiidae specimens collected during the Angolan (Okavango) field trip.

Rhabdalestes maunensis RB12-Ang21 Cuito pools Hydrocynus vittatus RB12-Ang20 Cuito River at 7015b.jpg C-C bridge 5976b

38 | P a g e

Brycinus lateralis RB12-Ang22 Cuito-Cuanavale Micralestes acutidens RB12-Ang20 Cuito River at confluence 6096b C-C bridge 5961b

Distichodontidae – dwarf citharines These small citharines, superficially similar to minnows, are characterized by having oral teeth, lacking barbels on the mouth, usually having a small adipose fin and having ctenoid scales. They are also usually patterned with striking bars or stripes. They were rather uncommon particularly in the western region of the Cubango and its tributaries and no specimens were collected in the Qeve or Kunene. From our collections there appears to be a split in the Nannocharax between the eastern sandy Okavango Rivers and the western rocky Cubango tributaries. We have few photos or DNA samples from these collections and this is a gap which needs to be rectified next trip. In contrast there does not seem to be the same structuring in Hemigrammocharax although this needs to checked during future trips with additional sampling. Nannocharax are bottom dwellers/foragers and perhaps less migratory although they do move up rivers and are all caught in fishing baskets. Hemigrammocharax are open water species foraging on drift and certainly migrate up systems actively in the summer floods.

Figure 1.13. Barcode results and specimens of Hemigrammocharax and Nannocharax collected during the Angolan (Okavango) field trip.

Hemigrammocharax multifasciatus RB12-Ang20 Hemigrammocharax multifasciatus RB12-Ang01 Cuito PHS Oka_0512_12 112 Kabumbe lodge dam 5359c.jpg

Hemmigrammocharax machadoi RB12-Ang20 Nannocharax macropterus RB12-Ang19 Cuito PHS Cuito River at CC 5969b.jpg 0512_14 131b.jpg

39 | P a g e

Hepsetidae – African pike This is a widespread and common species within the lower Okavango systems and the Zambezian floodplains. It was rarely seen during the May expedition probably due to the lack of suitable habitat types sampled and our methods of fishing. The species is a lurking ambush predator found in marginal swamps and lagoons of large floodplain rivers. We collected the species and saw it in fishermen’s catches at the Soba Mathias sites and around Cuito-Canavale sites. It was not abundant and it is either out-competed in this upper region by tigerfish or simply prefers more marshy habitats. Both are likely susceptible to overfishing with gill nets as they are easily caught with this gear at night and dusk.

Figure 1.14. Hepsetus cuvieri RB12-Ang21 Cuito River sand pits PHS 0512_13 003b.jpg

Poeciliidae – Topminnows The known species from the Okavango delta are Aplocheilichthys johnstoni, A. katangae, A. hutereaui and an undescribed species Aplocheilichthys sp. ‘pygmy’. All are abundant in suitable habitats of the delta and each species appears to have clear habitat preferences. Unexpectedly Aplocheilichthys were rare during the May 2012 Angolan survey. They were not collected at all sites and generally when they were collected they were not abundant. The species/forms encountered were also different from those known from lower parts of the system. Aplocheilichthys hutereaui was absent, Aplocheilichthys sp. ‘pygmy’ was collected at only 1-2 swampy floodplain sites (typical habitat). The two most common species were similar to A. johnstoni and A. katangae but both seemed to differ in colour patterns to specimens from the Okavango delta.

Not all bar-code results are back yet but results so far plus our morphological examinations indicate there is greater diversity than previously recognized within the Okavango system. New species include the known, undescribed Aplocheilichthys sp. ‘pygmy’ and then one Aplocheilichthys katangae-like species and one A. johnstoni-like species. Additional collections of these with more careful field examination and targeted genetic sampling will help to hone our determinations within this group. Interestingly there appear to be considerable genetic differences between morphologically similar groups across the broader Zambezian region indicating a complex of species.

Figure 1.15. Barcode results and specimens of Aplocheilichthys collected.

Aplocheilichthys johnstoni RB12-Ang20 Cuito River PHS 0512_12 059b.jpg

40 | P a g e

Aplocheilichthys cf. johnstoni 'side-stripe' Liucingua R. RB12-Ang19 6146c.JPG

Aplocheilichthys cf. katangae RB12-Ang08 5575c.jpg

Aplocheilichthys katangae RB12-Ang25 Longa River 20-5-2012 6142b.jpg

41 | P a g e

Claroteidae – grunters This small predatory catfish was not collected in high numbers (18 specimens from 6 sites). Although superficially similar to squeaker catfishes grunters can be distinguished by their unbranched barbels, the positioning of the nostrils on the upper lip, upper jaw teeth in a single tooth pad and spots roughly arranged into a series of vertical bars. Morphologically specimens appear to be the same as those found in lower reaches of the Okavango system although additional samples are awaiting genetic analysis. Specimens were typically caught electric fishing, in traps and fyke nets all around weed beds and marginal cover.

Figure 1.16. Parauchenoglanis ngamensis RB12-Ang19 Cuito PHS 0512_14 115b.jpg

Amphiliidae – mountain and torrent catfishes and sand catlets Amphiliids are represented by two distinct genera in southern Africa, Amphilius mountain or torrent catfishes and Zaireichthys sand catlets. Amphilius are characterized by their compressed body form, fan-like pectoral and pelvic fins, the lack of fin spines and their colour patterns typically vary considerably between populations. Zaireichthys are typically small (less than 5cm SL), their fins are

42 | P a g e proportionately smaller and they have barbed spines, their colour patterns are also variable between populations/species. Amphilius habitats are complex rocky riffles and waterfalls, Zaireichthys are found in sand substrates and weed-beds in flowing waters but not as fast as with Amphilius.

Amphilius were abundant in the western rivers with 93 specimens being collected at 10 sites. We observed differences in Amphilius populations between our sampling sites during the Angolan 2012 expedition. Interestingly the Okavango populations are coming out as genetically homogenous despite these pattern differences. The coastal Qeve River population north of Huambo is clearly distinct though, and is of a completely different lineage to the Okavango species (Amphilius uranoscopus complex), more closely allied to and possibly the same as Amphilius lentiginosus.

Zaireichthys sand catlets were rarely collected (28 specimens, between 2-4 species at 4 sites). The barcode results are rather odd indicating a very wide split and the results need to be re-examined. Certainly it would appear that there are at least two species and these need to be compared with new taxonomic descriptions (Eccles et al. 2011). More material is certainly required.

Figure 1.17. Barcode results and specimens of Amphilius and Zaireichthys spp. collected.

RB12-A121 Amphilius uranoscopus from the Cacuchi River north of Menongue.

RB12-A248 Amphilius sp. ‘qeve’ from the Qeve River north of Huambo.

43 | P a g e

Zaireichthys sp. dark granular Luancingwa RB12-Ang19 6168b.jpg

Zaireichthys sp. palid Luancingwa RB12-Ang19 6165b.jpg

Mochokidae – squeakers and suckermouth catfishes Mochokid catfishes in southern Africa are represented by two distinctly different genera which both have strong sharp spines in their pectoral and dorsal fins and complex tooth patterns. Synodontis or squeaker catfishes are usually larger river species often found in main channels and in complex habitats, they have complex branched barbels around the mouth, they are usually strikingly patterned and as their common name suggests they make squeaking sounds when disturbed. Chiloglanis or suckermouth catfishes are usually smaller than squeakers, their mouths are in the form of suction discs allowing them to hold fast onto rock in fast flows, their barbels are simple and

44 | P a g e not branched and although pigmentation is quite variable it is usually cryptic allowing animals to blend in with substrates.

Squeaker catfishes (Synodontis spp.) from the upper Zambezian region are extremely difficult to identify – a situation that has been obvious for several decades. We are beginning to interrogate this clade using genetic methods and it appears that the radiation has evolved recently. Morphological variation is considerable and seemingly species external pigmentation and colour patterns and morphological feature ranges overlap which makes character based keys difficult to construct. As a first impression our collections in Angola revealed a lower level of diversity than in the Okavango River in Namibia and in the delta in Botswana. Given that the radiation probably occurred in the larger Magkadigadi Lake (now centered around the Okavango Delta) this isn’t entirely surprising. This pattern is shared with other peripheral Zambezian systems. Barcode results suggest there are two species present in the upper river reaches which correspond to our concepts of S. nigromaculatus and S. macrostigma. Synodontis macrostigma showed particularly high colour pattern variation ranging from almost unspotted through spots to large lines (see below images). A third species similar to S. thamalakanensis was sampled near Rundu and morphological examinations suggest that S. woosnami is also present in our samples although we didn’t get bar- code results for that species. The situation clearly requires further attention and probably more genetic and morphological sampling on the next trip to sort out to a reasonable level.

Figure 1.18. Barcode results and specimens of Synodontis spp. collected.

Syndontis nigromaculatus RB12-Ang07 Cacuchi Synodontis sp. juvenile RB12-Ang26 6194b.jpg bridge PHS 0512_05-6 119b.jpg

Synodontis cf. macrostigma ventral head RB12- Synodontis cf. macrostigma RB12-Ang25 Longa Ang25 Longa River 6138b.jpg River 6135b.jpg

Synodontis cf. macrostigma RB12-Ang19 Luacingwa Synodontis cf. macrostigma RB12-Ang32 Okavango River 6170b.jpg River, Rundu 6237b.jpg

45 | P a g e

Synodontis cf. thamalakanensis RB12-Ang20 Cuito bridge 5954b.jpg Synodontis cf. thamalakanensis RB12-Ang20 Cuito Bridge 5957c.jpg

Suckermouth catfishes (Chiloglanis spp.). Chiloglanis were common in most areas where fast flowing rocky habitats occurred. The current taxonomy recognizes a single widespread species C. fasciatus throughout the upper Zambezi systems. The group however exhibits considerable morphological variation in body forms, fin and spine sizes and colour patterns all indicating that there are in fact several more species in the system. Genetic results from the May expedition confirms this for the upper Okavango and neighbouring systems. From the cladogram below there seem to be five or more species with two being present at site 7 (Cacuchi River bridge) and two different species at site 11 (Qeve River):  A379 (site 12) = Top Halo Camp – Kunene  A134 (site 7) & A008 (site 2) = Qebe near Menongue and Cacuchi north of Menongue – sister in this cladogram to Malawi specimens – probably this represents the common form in the lower Okavango River and what is widely considered as C. fasciatus.  A220 (site 11) Qeve River  A240 (site 11) = both from the coastal Qeve – seems like two species present in samples need to check specimens as I think they were lumped under one species.  A299 & A277 (site 13) = Cubango headwater falls, A102 (site 7) = Cacuchi at bridge, D066 (a site in the upper Kabompo River in Zambia), A262 & A257 (site 13) = Cubango falls. There is some structuring even within this cluster and certainly there are several colour patterns (see below) and morphological variation.

46 | P a g e

The level of differentiation between clades in the Angolan Okavango and neighbouring systems is considerably greater than in either the Quanza or in Malawi. In some instances two species are present at sites and this will require us to re-examine samples based on our bar-code results.

Figure 1.19. Barcode results and specimens of Chiloglanis spp. collected.

Chiloglanis sp. 'qeve' lateral view RB12-Ang11 Qeve River 5743b.jpg

Chiloglanis sp. RB12-Ang12 Cunene in Huambo 5793b.jpg

Chiloglanis sp. 'cubango gold' lateral view RB12-Ang13 Cubango source 5848b.jpg

Chiloglanis sp. 'cubango dark' lateral view RB12-Ang13 Cubango source

47 | P a g e

5842b.jpg

Chiloglanis sp. 'cubango dark' lateral view RB12-Ang07 Cacuchi River 5469b.jpg

Chiloglanis sp. RB11-Kal21 West Lumwana River, Zambia 4358b.jpg

Clariidae – airbreathing catfishes Clarias catfishes are one of the characteristic and dominant groups in the lower Okavango systems and form a major portion of artisanal fisheries catches. There are four species in the lower river, all

48 | P a g e are common but the most common is the largest species Clarias gariepinus. Remarkably during our Angolan survey Clarias were rare and C. gariepinus was absent from our own catches although we saw a few in fishermen’s catches. This certainly is a significant feature of the upper Okavango fauna in Angola as clariids are top predators and have a major impact on fish communities. Clarias we caught were known upper Zambezian species but more typical of smaller sub-systems and higher altitudes, most commonly caught was C. liocephalus characterized by its small adult size and high-lighted fin pattern. The barcode results suggest a split between what is recognized as C. stappersii between the Zambezi/Okavango and the Cunene systems. Interestingly the Qeve River which for most other groups has a unique fauna has C. liocephalus which clusters with upper Zambezia/Okavango specimens.

Figure 1.20. Clarias specimens and barcode results.

Clarias theodorae RB12-Ang07 Cacuchi River at bridge 5641b.jpg

Clarias liocephalus RB12-Ang07 Cacuchi River at bridge 5595b.jpg

Clarias stappersii RB12-Ang18 Cuchi River - rabbit bridge PHS 0512_11 030b.jpg

49 | P a g e

Cichlidae – cichlids, bream As with clariids, the cichlid assemblage is a characteristic and dominant feature of the lower river and delta and similarly cichlids were remarkably rare and less diverse in the Angolan upper catchments. Barcode results need to be examined with care as the cichlid ‘flock’ is recently evolved (Joyce et al. 2005) and also the CO1 mitochondrial gene is probably a poor marker for examining their diversity. Despite their low diversity and numbers there were a few species of taxonomic interest which we have struggled to identify with certainty and these may be new species. We encountered one clearly distinct ‘red spotted’ Serranochromis species only at the Cacuchi Bridge site which we consider to be a new species at this stage. The Thoracochromis collected in the upper Kunene catchment in Huambo was also morphologically and genetically distinct from T. buysi and T. albolabris, the species known from that system. The Psuedocrenilabrus philander exhibited a distinct colour pattern and colouration being deep red, something not seen in the lower river. Pseudocrenilabrus are highly variable between populations and at this stage we are uncertain about the significance of this feature.

Figure 1.21. Cichlid diversity in the upper Okavango in Angola with barcode results.

Hemichromis fasciatus RB12-Ang23 6064b.jpg Oreochromis macrochir RB12-A029 Qebe River, Menongue PHS 0512_02 035

50 | P a g e

Pseudocrenilabrus philander RB12-Ang23 6071b.jpg Pharyngochromis acuticeps RB12-Ang02 Qebe River, Menongue 5391b

Serranochromis angusticeps RB12-Ang20 5967b.jpg

Sargochromis sp. Canavale R. RB12-Ang23 6067b

Serranochromis sp. ‘red spot’ RB12-Ang07 5444b sp. Serranochromis robustus RB12-Ang01 5364b RB12-Ang07 5444b

Thoracochromis cf. buysi RB12-Ang09 Upper Cunene at Huambo 13-5-2012 5794b

Tilapia sparrmanii RB12-Ang25 Longa River PHS 0512_14 084b

51 | P a g e

52 | P a g e

Mastacembelidae – spiny eels Mastacembelid eels are widespread through the upper Zambezi region. Currently two species are recognised, M. frenatus and M. vanderwaali. We did not collect M. vanderwaali and so all our specimens, following the current taxonomy are assigned to M. frenatus. However, the taxonomy of the genus has been actively researched during the last few years and the diversity appears to be much higher than the current taxonomy would suggest. Firstly, M. frenatus was described from the northern part of Lake Tanganyika and thus it seems unlikely to be the correct name for species in the Zambezian region. Secondly, several species have been previously described and synonymised with M. frenatus and these are possibly more appropriate names. Further taxonomic work is needed.

Mastacembelus frenatus, Boulenger G.A. 1901. Northern end of Lake Tanganyika, Burundi. Holotype (unique): BMNH 1906.9.6.161. Mastacembelus mellandi, Boulenger G.A. 1914. Solwezi River, tributary of Chifulowa River, Zambia. Holotype (unique): BMNH 1914.9.21.13. Mastacembelus mutombotombo, Pellegrin J. 1936. Cubango [Okavango] River near Vila da Ponte, Angola. Holotype (unique): MNHN 1936-0157. Mastacembelus thompsoni, Boulenger G.A. 1918.Sesheke, Zambia. Holotype (unique): SAM 14542 [now at SAIAB].

Interestingly Pellegrin (1936) described the species from the upper Okavango as M. mutombotombo and all local people during the May expedition referred to eels by the local name ‘mutombotombo’.

The specimens we have collected vary considerably in colour patterns which are difficult to interpret. There are few other good external characters to help with species determinations as scales are minute and fin spines numerous and deeply embedded. From our examinations so far (based on bar-code results and colour patterns) there seems to be three species/groups in the expedition’s collections:  A211 & A381 – both from the coastal Qeve River system (site 11) and genetically very different to the Okavango specimens.  A278 – Cuela (site 18, upper Cacuchi) ‘rabbit bridge’ – swampy, sandy headwater river - unpatterned and a little different genetically.  A387 (site 17, Chitembo stream), A147 (site 7, Cacuchi bridge), A245 (site 9, Mumbue tributary) – all Cacuchi sub-system

Figure 1.22. Mastacembelid eel specimens and barcode results.

Mastacembelus frenatus RB12-Ang07 Cacuchi River (Okavango system) 5603b.jpg

53 | P a g e

Mastacembelus sp. RB12-Ang17 Chitambo (Okavango system) 5886b.jpg

Mastacembelus sp. RB12-A278 Cuela River – tributary (Cacuchi, Okavango system)

Mastacembelus sp. RB12-A211 and A381 Qeve River near Huambo (coastal system)

Anabantidae – climbing perches or labyrinth fishes Anabantids are small percoid fishes typically occurring in floodplain swamps and possessing accessory air-breathing organs to deal with the low oxygen levels in these environments. During the May Angola (Okavango) expedition we rarely sampled suitable swampy habitats and consequently collected very few specimens and only one species of anabantid. We collected a few specimens of Microctenopoma intermedium at a single site and none of the larger Ctenopoma multispine. The only site where M. intermedium was recorded was the grassy margins of the Canavale River. They were collected using a D-net which was raked through marginal flooded grasses and overhanging root

54 | P a g e stocks in the bank. The group is probably more typical of the southern and eastern rivers of the Angolan sector of the Okavango system.

Figure 23. Microctenopoma intermedium RB12-Ang23 Canavale River 6075b.jpg

1.4 Artisanal fisheries

We observed relatively few, low-intensity fishing activities during the survey. However they were widespread and during the dry season would probably increase in intensity as other sources of food decline. Increased pressure on local fisheries resources due to population growth and easier access to rivers could result in over-fishing and depleting local fish populations in the future. This could be exacerbated by varied habitat degradation and pollution. We suggest a dedicated programme is needed to examine the current fishery and develop a long-term fisheries management plan ideally this should be trans-boundary in scope.

Aquaculture and dam fisheries. With the construction of dams and the growth of urban populations the likelihood of both enhancing dam fisheries with introduced stocks and aquaculture projects is high. Neighbouring regions have suffered from the impacts of alien introductions for aquaculture purposes. We strongly recommend the use of locally caught indigenous species in all aquaculture projects and dam fisheries. Obvious candidates would be Oreochromis andersonii and O. macrochir collected from the system.

Figure 1.24. Artisanal fisheries activities observed during the May 2012 expedition.

Fishing baskets – typically used by groups of Hook and line fishermen – usually young women to scoop fish up boys use this method

55 | P a g e

Fish weir – this had open (non-valved) Valved fish trap – typically this will be set in constriction traps in and was in use. a well-constructed un-passable fence

56 | P a g e

2. DRAGONFLIES – ORDER ODONATA

2.1 Aims of the project from Odonata experts perspective

Apart from the overall SAREP project objectivs there are more specified aims regarding the Odonata research.

In terms of dragonfly distribution Angola remains as one of the poorest studied countries on the African continent.

One aim of the study was to get an up to date overview about Odonata distribution in the Angolan part of the Okavango catchment that helps to understand distribution patterns on the whole southern sub-continent. Another one is to obtain information and knowledge about ecological preferations of selected species. This should help to calibrate parameters that are needed to use dragonflies as indicators of freshwater health also in this part of Africa. This also will contribute valuable bckground knowledge for the development of a DBI (Dragonfly Biotic Index, see SIMAIKA & SAMWAYS 2008) for the Okavango catchment as well as for the whole continent. A task for the near future is to publish an updated checklist and a Red List of the Odonata of Angola in collaboration of other dragonfly experts (Viola Clausnitzer, K.-D.B. Dijkstra), also for this the gathered informations will be more than valuable. The obtained knowledge, photographs and specimens will also be used in the preparation of an illustrated fieldguide to the dragonflies of south-central Africa.

The next important aim is to get information's about existing or potential threats to the dragonfly fauna of the Okavango catchment in Angola and potential threats that can arise from this to the downstream environments. The knowledge potentially gathered can be applied throughout Angola's Okavango river systems to help minimize or mitigate the impact of future development actions, thus allowing dragonflies to fulfill their role as “guardians of the watershed”. For example, when selecting sites for development, the private sector can be advised to safeguard sensitive habitats and species. Key sites and species for monitoring can be selected by consulting the distribution ranges of threatened species presented at www.iucnredlist.org. If dragonflies are lost, we not only lose an important component of aquatic ecosystems, but also reliable indicators of external impacts. Without such sensitive monitors, changes may be detected too late to take remedial measures (see also SIMAIKA et al. 2013).

2.2 Methods

A first SAREP field survey to assess freshwater biodiversity in the Angolan part of the Okavango catchment was conducted between 05 February and 22 February 2012.

On Angolan territory altogether 30 sites were sampled for Odonata, 28 within the Okavango catchment, 2 outside in the Cuanza catchment. All sampling points are shown in Fig. 2.1 and listed with details in Table 2.1. The first survey focused in the middle and upper sections of the tributaries. For further details and circumstances of the May 2012 field survey see BROOKS (2012). The southernmost location, Savate at the Cubango, was just sampled randomly en route. At all other sampling sites the sampling procedure was similar with slightly varying sampling effort depending on different sampling time available.

57 | P a g e

With few exceptions adult dragonflies were collected. A hand net was used to catch resting or flying individuals. With a close-focus binocular observations of adults were made if these were not possible to catch or if not necessary to catch. In most cases the individuals were released after determination. For determination a hand lens was used in the field, a microscope later in the lab. In cases of taxonomic uncertainty and in order to get a representative collection of voucher specimens at each sampling site a selection of individuals were collected. In that case the specimens were killed in ethanol, transformed into an acetone bath till the next morning, dried and labeled. The specimens were stored in labeled paper envelopes within air-tight plastic boxes (freezer boxes). In order to keep them absolutely dry the boxes also kept a small amount of silica gel which also prevents them from damage by harmful insects. Of a selection of specimens a tissue sample for further molecular genetic studies was taken. This was stored in 99% ethanol and labeled as the dry specimen.

A SRL camera with macro-lens and ring-flash was used to gather photographs of living dragonflies in the field. In some cases photographs of caught individuals in the hand were taken to keep information about living coloration that might change after death in collected specimens. In a few cases also exuviae of dragonflies were collected. Larvae gathered randomly by the fish expert group are not determined yet and not included in the present report. A GPS device was used to get accurate coordination of the sampling sites. A few chemical water parameters were measured using a handheld Hannah Instruments perimeter for pH and conductivity. The pH was parallel compared with testing strips. For identification purposes the preliminary key of CLAUSNITZER & DIJKSTRA (in prep.) was used together with papers of the original descriptions for species that not covered by this.

58 | P a g e

Table 2.1: All localities sampled for Odonates on SAREP field survey in May 2012. Numbers according to map in Fig. 2.4. Ordered roughly from south to north.

date(s) River Altitude No Locality Province Latitude Longitude Habitat sampled sub-system (m a.s.l.)

1 Savate, around 20 km 07.05.2012 Cuando Cubango 5 -16,727439 17,89966 1.136 permanent north of Savate at road to Cubango river and Caiundo mixed Acacia woodland 2 Cuebe River rapids, 30 km 09.05.2012 Cuando Cuebe -14,9395 17,7188 1.260 permanent south of Menongue Cubango river, rapids, rocks

3 Cuebe River, above 09.05.2012 Cuando Cuebe -14,93748 17,71959 1.260 permanent rapids, calm river section, Cubango sandy river 30 km south of Menongue 4 Cassanga, dambo and 09.05.2012 Cuando Cuebe -14,43934 17,67458 1.314 dambo and stream at road Caiundo- Cubango stream Menongue, 18 km south of Menongue 5 Missombo, dambo and 09.05.2012 Cuando Cuebe -14,74583 17,66826 1.349 dambo and stream at road Caiundo- Cubango stream Menongue, 10 km south of Menongue 6 Menongue, Kambumbe 08.05.- Cuando Cuebe -14,67609 17,73542 1.366 artificial dam Lodge, seepages and dam 11.05.2012 Cubango with rich just south of the Lodge vegetation and seepage 7 Soba Matias, Cuebe River 10.05.2012 Cuando Cuebe -14,43934 17,81484 1.378 permanent and floodplains 27 km Cubango river, large northeast of Menongue floodplains

8 Luassingua River, 55 km 17.05. + Cuando Longa -14,58995 18,17057 1.326 sandy east of Menongue, at 20.05.2012 Cubango meandering road Menongue-Cuito river Cuanavale 9 Longa River at road 20.05.2012 Cuando Longa -14,60633 18,46739 1.309 sandy river Menongue-Cuito Cubango Cuanavale

10 Cuito River at Cuito 17.05.- Cuando Cuito 2 -15,17159 19,19331 1.185 sandy river Cuanavale, at road bridge 18.05.2012 Cubango east of town

11 Cuito River at confluence 19.05.2012 Cuando Cuito 2 -15,140075 19,191429 1.186 sandy river with Cuanavale River Cubango and floodplains

12 Cuanavale River, 4 km 19.05.2012 Cuando Cuanavale -15,13569 19,19571 1.186 sandy river northeast of Cuito Cubango Cuanavale

13 Cuito River 9 km north of 19.05.2012 Cuando Cuito 1 -15,08723 19,14978 1.187 sandy river Cuito Cuanavale Cubango and floodplains

14 Cuele River, 35 km west 22.05.2012 Cuando Cuelei -14,70171 17,37832 1.402 sandy river of Menongue at road to Cubango Cuchi

15 Cuchi River at Cuchi town, 22.05.2012 Cuando Cuchi 2 -14,65186 16,60427 1.341 rocky river, around road bridges Cubango stony ground

59 | P a g e

date(s) River Altitude No Locality Province Latitude Longitude Habitat sampled sub-system (m asl)

16 Cuchi River, gorge with 21.05.2012 Cuando Cuchi 2 -14,58949 16,90687 1.364 rocky river, rapids and waterfall, Cubango waterfalls 7 km northeast of Cuchi

17 Satchijamba, Cuele River 16.05.2012 Bié Cuelei -13,71337 17,08962 1.513 river and at road Chitembo- dambo Menongue

18 Chingueia, dambo and 12.05.2012 Bié Cuelei -13,6937 17,06125 1.555 dambo and stream flowing to Cuele stream River

19 Cacuchi River at road 11.05.- Bié Cacuchi -13,59445 16,88045 1.500 permanent bridge Menongue- 12.05.2012 river with Chitembo, sandy river rapids with rocks and rapids 20 Cacuchi River north of 11.05.- Bié Cacuchi -13,592652 16,881654 1.500 dambo and Mumbué at road bridge 12.05.2012 floodplains Menongue-Chitembo, dambo and floodplain 21 Chitembo, stream south 16.05.2012 Bié Cacuchi -13,52797 16,75872 1.577 stream and of Chitembo, tributary to dambo Cacuchi River

22 Watana stream, rocky 13.05.2012 Bié Cacuchi -13,27965 16,74611 1.611 rocky stream stream north of Chitembo, tributary to Cacuchi River 23 Chiuano Micha, little 13.05.2012 Bié Cacuchi -12,87496 16,76735 1.688 dambo and stream and dambo, stream tributary to Cacuchi River

24 Cassoco, rocky stream 13.05.2012 Bié Cuchi 1 -12,78683 16,75737 1.699 rocky stream north of Cassoco, tributary to Cuchi River

25 Cuchi River at road bridge 13.05. + Bié Cuchi 1 -12,5307 16,69836 1.693 permanent Huambo-Kuito 16.05.2012 rocky river

26 Cutato River at road 16.05.2012 Bié Cutato -12,56787 16,49713 1.688 permanent bridge Huambo-Kuito muddy river

27 Chicala-Cholohanga, 15.05.2012 Huambo Cubango 1 -12,67115 16,11166 1.772 stream and Cubango River headwater waterfall south of town, stream with small waterfall

28 Chicala-Cholohanga, 15.05.2012 Huambo Cubango 1 -12,71841 16,07262 1.853 dambo and Cubango River source, stream south of town, springbrook on grassy plateau 29 Keve (Cuvo) River, east of 14.05.2012 Huambo outside -12,2031 15,59522 1.415 rocky river Alto Hama, rocky river Cubango with rapids with rapids Cuanza 30 Keve (Cuvo) River, east of 14.05.2012 Huambo outside -12,203073 15,59631 1.415 dambo Alto Hama, dambo Cubango nearby Cuanza

60 | P a g e

Figure 2.1: Map showing Odonata sampling points of the May 2012 survey (green dots). Numbers according to Table 2.1. Red line deliniate the active Okavango catchment.

61 | P a g e

2.4 Results

2.4.1 Overall Odonata species inventory Altogether 88 dragonfly species could be recorded on the survey in May 2012. Table 2.2 lists all recorded species and gives information about their Red List status according the IUCN Freshwater Assessments (DARWALL et al. 2011b) and if species were recorded for the first time in the region or country. 31 of the species are belonging to damselflies (Zygoptera) and 57 to the dragonflies (Anisoptera). As in the usual inventories in Africa the dominating families are the Coenagrionidae and Libellulidae, for the abundance of the different families see Fig. 2.2.

Lestidae 2 Chlorocyphidae

3 Platycnem-

ididae 2

Libellulidae 48 Coenagrionidae 24

Aeshnidae 2 Gomphidae 6

Corduliidae 1

Figure 2.2: Graph showing the distribution of the recorded Odonata families.

Altogether more than 600 specimens of Odonata were collected and DNA samples of all recorded species.

The survey led to the discovery of 4 undescribed species. One Chlorocypha sp. nov. near aphrodite is close to the tropical Chlorocypha aphrodite but differs in some characters. A Pseudagrion sp. nov. 'sarepi' is near the Zambian Pseudagrion fisheri but unique in morphological features and colour markings. Mesocnemis sp. nov. 'assesstrix' and Zygonyx sp. nov. 'denticulatus' were recorded before in neighbouring countries but are still undescribed.

20 of the species were recorded for the first time in Angola (incl. the undescribed four), 27 are new for the whole Okavango catchment and 54 species are new for the Angolan part of the Okavango catchment. This survey brings the total number of Odonata known from the Okavango catchment to 160 species (see Error! Reference source not found.).

The number of Odonata species known from Angola increases from 164 to 184, an updated preliminary checklist of the Odonata of Angola is provided in Error! Reference source not found..

62 | P a g e

Table 2.2: All Odonata species recorded on the SAREP field survey in May 2012 in Angola.

new for IUCN new for the Angolan new for Species Red List Okavango part of the Angola status catchment catchment

Zygoptera

Lestidae

Lestes pinheyi Fraser, 1955 LC X

Lestes plagiatus (Burmeister, 1839) LC X

Chlorocyphidae

Chlorocypha crocea Longfield, 1947 NT

Chlorocypha fabamacula Pinhey, 1961 LC

Chlorocypha sp. nov. nr. aphrodite - X X X

Platycnemididae

Mesocnemis singularis Karsch, 1891 LC

Mesocnemis sp. nov. 'assestrix' - X X X

Coenagrionidae

Africallagma subtile (Ris, 1921) LC X X

Agriocnemis angolensis Longfield, 1947 LC X

Agriocnemis bumhilli Kipping, Martens & Suhling 2012 - X X X

Agriocnemis exilis Selys, 1872 LC

Agriocnemis victoria Fraser, 1928 LC X

Azuragrion nigridorsum (Selys, 1876) LC X

Ceriagrion corallinum Campion, 1914 LC

Ceriagrion glabrum (Burmeister, 1839) LC

Ceriagrion suave Ris, 1921 LC

Ceriagrion whellani Longfield, 1952 LC X X X

Ischnura senegalensis (Rambur, 1842) LC

Pinheyagrion angolicum (Pinhey, 1966) LC X

Pseudagrion coeleste Longfield, 1947 LC X

Pseudagrion deningi Pinhey, 1961 LC X X

Pseudagrion estesi Pinhey, 1971 - X X

63 | P a g e

new for IUCN new for the Angolan new for Species Red List Okavango part of the Angola status catchment catchment

Pseudagrion greeni Pinhey, 1961 LC X X

Pseudagrion hamoni Fraser, 1955 LC X X X

Pseudagrion inconspicuum Ris, 1931 LC

Pseudagrion kersteni (Gerstäcker, 1869) LC X X

Pseudagrion rufostigma Longfield, 1947 LC X

Pseudagrion salisburyense Ris, 1921 LC

Pseudagrion sjoestedti Förster, 1906 LC X X

Pseudagrion sp. nov. 'sarepi' - X X X

Pseudagrion sublacteum (Karsch, 1893) LC

Anisoptera

Aeshnidae

Anax imperator Leach, 1815 LC X

Anax speratus Hagen, 1867 LC X X

Gomphidae

Crenigomphus kavangoensis Suhling & Marais, 2010 LC X X

Ictinogomphus dundoensis (Pinhey, 1961) LC X

Ictinogomphus ferox (Rambur, 1842) LC

Paragomphus genei (Selys, 1841) LC X X X

Phyllogomphus annulus Klots, 1944 LC X X

Phyllogomphus selysi Schoudeten, 1933 LC

Corduliidae

Phyllomacromia contumax Selys, 1879 LC X X

Libellulidae

Acisoma inflatum Selys, 1882 - X

Aethiothemis bequaerti Ris, 1919 LC X X

Aethiothemis solitaria, Martin, 1908 LC X

Aethriamanta rezia Kirby, 1889 LC X

Brachythemis lacustris (Kirby, 1889) LC

64 | P a g e

new for IUCN new for the Angolan new for Species Red List Okavango part of the Angola status catchment catchment

Brachythemis leucosticta (Burmeister, 1839) LC X

Crocothemis brevistigma Pinhey, 1961 LC X X X

Crocothemis divisa Baumann, 1898 LC X X

Crocothemis erythraea (Brullé, 1832) LC

Crocothemis sanguinolenta (Burmeister, 1839) LC

Diplacodes deminuta Lieftinck, 1969 LC X

Diplacodes lefebvrii (Rambur, 1842) LC

Diplacodes pumila Dijkstra, 2006 LC X X X

Hemistigma albipunctum (Rambur, 1842) LC X

Nesciothemis farinosa (Förster, 1898) LC

Nesciothemis fitzgeraldi Longfield, 1955 LC X X X

Orthetrum abbotti Calvert, 1892 LC

Orthetrum brachiale (Palisot de Beauvois, 1805) LC X

Orthetrum caffrum (Burmeister, 1839) LC

Orthetrum chrysostigma (Burmeister, 1839) LC

Orthetrum guineense Ris, 1910 LC

Orthetrum hintzi Schmidt, 1951 LC X X

Orthetrum icteromelas Ris, 1910 LC X

Orthetrum julia Kirby, 1900 LC

Orthetrum machadoi Longfield, 1955 LC

Orthetrum macrostigma Longfield, 1947 LC X X

Orthetrum monardi Schmidt, 1951 LC X X

Orthetrum trinacria (Selys, 1841) LC X

Palpopleura jucunda Rambur, 1842 LC X X

Palpopleura lucia (Drury, 1773) LC X

Palpopleura portia (Drury, 1773) LC X X

Pantala flavescens (Fabricius, 1798) LC

Tholymis tillarga (Fabricius, 1798) LC X X

65 | P a g e

new for IUCN new for the Angolan new for Species Red List Okavango part of the Angola status catchment catchment

Tramea basilaris (Palisot de Beauvois, 1805) LC

Trithemis aconita Lieftinck, 1969 LC X X

Trithemis annulata (Palisot de Beauvois, 1805) LC X

Trithemis anomala Pinhey, 1956 LC X X X

Trithemis arteriosa (Burmeister, 1839) LC

Trithemis dorsalis (Rambur, 1842) LC

Trithemis kirbyi Selys, 1891 LC

Trithemis leakeyi (Pinhey, 1956) LC X X X

Trithemis monardi Ris, 1931 LC X

Trithemis morrisoni/palustris Damm & Hadrys, 2009 LC X X X

Trithemis pluvialis Förster, 1906 LC

Trithemis stictica (Burmeister, 1839) LC

Zygonyx natalensis (Martin, 1900) LC

Zygonyx sp. nov. 'denticulatus' - X X X

Zygonyx torridus (Kirby, 1889) LC X

total: 88 species 27 54 20

Of the species recorded only one, Chlorocypha crocea is listed as Near Threatened.

Seven others were not assessed and evaluated during the last Freshwater Assessments for various reasons: Agriocnemis bumhilli was only described in 2012, and Pseudagrion estesi has been simply forgotten, both tend to qualify for Least Concern. The same counts for Acisoma inflatum which was known so far as Acisoma panorpoides, a Least Concern species. Latest taxonomic research showed that "A. panorpoides" on the African continent probably consist of at least two species and A. inflatum is the common and widespread one (K.-D.B. Dijkstra in litt.).

The remaining four species are those that were not known to science until now. Chlorocypha sp. nov. near aphrodite, Mesocnemis sp. nov. 'assesstrix', Pseudagrion sp. nov. 'sarepi' and Zygonyx sp. nov. 'denticulatus' are waiting for description and would qualify for the status Data Deficient as the knowledge about their distribution and possible threats are insufficient. What does not mean that the knowledge of the remaining bulk of species, listed as Least Concern is good and sufficient.

The list above and the species inventory shows some curiosities.

66 | P a g e

A first one is that the family of Calopterygidae is absent in the collection. At least Phaon iridipennis is normally a locally common species at rivers and streams. In the region also Umma electa occur and especially in the upper Cubango section the Angola endemic Umma femina should occur. All these were not seen. Three interesting species of Chlorocyphidae were found but the most common and widespread species Platycypha caligata was only seen at Popa Falls in Namibia at the middle section of the Okavango. From the Lestidae only two species were found of 7 species that are expected in the area. Even not the most common Lestes pallidus that inhabits temporary habitats came on record. Protoneuridae were completely missing. The genus Elattoneura is with two species, E. glauca and E. tropicalis present in the Biome but were not found. Mesocnemis singularis is normally the most common Platycnemididae in the neighbouring countries but on the survey only found with few individuals at one locality. Coenagrionidae are well presented with 24 species but also here whole genera are missing. The genus Aciagrion is missing completely, these species are settling temporary pools and swamps. Of the genera Africallagma only a single species has been recorded. Two species of Aeshnidae is a very low number as well. Gomphidae are well presented and in general this family is rare in collections. Within the Libellulidae the absence of Urothemis edwardsii is the most striking curiosity. It is one of the most common dragonflies in the Okavango Delta and also has not been recorded at localities that showed similarities to the Delta. The same counts for Trithemis hecate, also common in the Okavango Delta and elsewhere in the region. The only explanation of the absence of these species might be the late season of the survey. Especially many species of ephemeral freshwater habitats are underrepresented and in fact there were not much ephemeral pools or swamps seen on the tour.

The large success of the survey was the gathered knowledge in terms of biogeographical aspects and ecology. It was found that some of the species formerly only known from northern Zambia and Katanga are more widespread towards the west. For example Aethiothemis bequaerti, Crocothemis brevistigma, Nesciothemis fitzgeraldi, Orthetrum macrostigma, Trithemis anomala, Trithemis leakeyi or Pseudagrion greeni are formerly thought to be much more range restricted which does not hold anymore. About some rare species information about habitats could be obtained for the first time. About Agriocnemis bumhilli, Chlorocypha crocea or Pseudagrion estesi was not much known about habitat preferences until now. More detailed background about species will be given within the next chapter where the single sampling sites with their habitat characters as well as their Odonata species assemblage is in a closer focus.

67 | P a g e

3. REPTILES AND AMPHIBIANS

3.1 Introduction

The herpetofauna of Angola remains one of the most poorly documented on the African continent. The only synthesis dates from the 19th Century (Bocage 1895) and this is now very out-of-date. Due to the protracted civil war, modern biodiversity surveys in the country are just beginning, (e.g. Huntley 2009 & 2011) and large areas of the country have never been scientifically surveyed. Most previous reports on the country’s herpetofauna have been restricted to the western or southern regions, e.g. Boulenger (1905), Monard (1931, 1937), Schmidt (1933), Parker (1936), Mertens (1938), Bogert (1940), Hellmich (1957a,b), FitzSimons (1959), Branch & McCartney (1992) and Poynton & Haacke (1993). An exception to the regional neglect of the herpetofauna of northeast Angola is a series of reports in by Laurent (1950, 1954, & 1964), based on collections in the Museo du Dundo. No formal surveys had been undertaken to the Cubango-Okovango river catchments, except of some random collections reported by Bocage (1889), Monrad (1930, 1937), Laurent (1964) and Branch & McCartney (1992).

3.2 Study area

The geographical scope of the survey is defined by the watershed of the Cubango-Okavango basin. The area consists out of two main rivers systems: the Cubango and the Cuito rivers, both draining eastward into the Okavango Delta. The Cubango River and its tributaries lay to the west of the basin and are underlain by granite outcrops and characterised by a series of rapids and some waterfalls in places. The Cuito River and its tributaries to the eastern side of the basin are characterised by Kalahari sands and large slow flowing rivers which act as a ‘sponge’ slowly releasing water into the system. The eastern topography is very flat with isolated higher grounds to the north. The vegetation is open woodland savannah. The soil is extremely sandy, with no stones or rocky outcrops.

Figure 3.1. Sampling sites during May 2012 survey.

68 | P a g e

3.3 Methods

The current herpetological survey was undertaken by Werner Conradie, based at Port Elizabeth Museum (Bayworld), South Africa. Opportunistic specimens were also collected by other members of the survey, particularly during the fish surveys (Roger Bills and Paul Skelton). Due to the dry, relatively cool evening conditions and lack of reptile movement no traplines were employed and most collections were done by active searching. Amphibian call surveys were done using a Nagra ARES-ML digital recording device and a Sony F-V4T Microphone. The following literature was consulted to compile historical records for the study area: Monrad (1937), Laurent (1964), Poynton & Haacke (1993), and Branch & McCartney (1998). Collections of specimens have been deposited in Port Elizabeth Museum (South Africa) and South African Aquatic Biodiversity Institute (South Africa) and a representative collection will be returned to Angola.

3.4 Amphibian and reptile Species lists

The following are updated species lists for amphibians and reptiles based on historical records as well as data from the new survey within the boundaries of the Cubango-Okavango Angola river basin. The IUCN- Redlist Categories are: NE = Not evaluated, DD = Data deficient and LC = Least concern. Type of sample: V = Voucher, A = Auditory, T = Tadpole, O = Observation, L = Literature record.

AMPHIBIANS ARTHROLEPTIDAE Leptopelis anchietae Anchieta’s Tree Frog* T LC Monrad 1937 BREVICIPTIDAE Breviceps adspersus Common Rain Frog L LC Monrad 1937 BUFONIDAE Amietophrynus funereus Angola Toad L LC Monrad 1937 Amietophrynus gutturalis Guttural Toad* V LC Amietophrynus lemairii Lemaire’s Toad* V, T LC Amietophrynus maculatus Flat-backed Toad* V LC HYPEROLIIDAE Ahl 1931; Monrad 1937; Hyperolius angolensis complex Angolan Read Frog* V, A, T, L LC Poynton & Haacke 1993 Hyperolius benguellensis Benguella Reed Frog* V, A, T, L LC Monrad 1937 Hyperolius nasutus Sharp-nosed Reed Frog* V, A, T, L LC Monrad 1937 Hyperolius seabrai Quilombo Reed Frog L DD Monrad 1937 Hyperolius cinereus Ashy Reed Frog* V, A, T DD Kassina kuvangensis Kuvangu Kassina* T, L LC Monrad 1933, 1937 Kassina senegalensis Bubbling Kassina* T LC HEMISOTIDAE Hemisus marmoratus Mottled Shovel-nosed Frog L LC Monard 1937 PHRYNOBATRACHIDAE Phrynobatrachus mababiensis Dwarf Puddle Frog* V, T LC Phrynobatrachus natalensis Snoring Puddle Frog* V, T, L LC Monrad 1937 PTYCHADENIDAE Hildebrandtia ornatissima Angola Ornate Frog L DD Monrad 1937

69 | P a g e

Ptychadena bunoderma Rough Grass Frog L LC Monrad 1937 Ptychadena mascareniensis Mascarene grass Frog* V, L LC Monrad 1937 Ptychadena keilingi Dala Grass Frog L LC Monrad 1937 Ptychadena oxyrhynchus Sharp-nosed Grass Frog* V, A, T, L LC Monrad 1937 Ptychadena subpunctata Speckled-bellied Grass Frog* V, A, T LC Ptychadena uzungwensis Udzungwa Grass Frog L LC Laurent 1964 Ptychadena taenioscelis Dwarf Grass Frog* V, A, T LC PIPIDAE Xenopus petersi Peter’s Clawed Frog* V, T, L LC Monrad 1937 PYXICEPHALIDAE Amietia angolensis Common River Frog* V, A, T, L LC Monrad 1937 Tomopterna tubercolosa Beaded Sand Frog* V, T, L LC Monrad 1937 RANIDAE Hylarana darlingi Golden-backed Toad* T, L LC Monrad 1937 * sampled during this survey Total = 28 (20*)

REPTILES ORDER: SQUAMATA COLUBRIDAE Monard 1937; Branch & Crotaphopeltis hotamboeia White-lipped Herald Snake* V, L NE McCartney 1992 Dasypeltis scabra scabra Common Egg Eater L NE Monard 1937 Dipsadoboa shervei shervei Sherve’s Tree Snake L NE Loveridge 1932 Monard 1937; Bocage 1895; Branch & McCartney Dispholidus typus punctatus Boomslang L NE 1992 Limnophis bicolour Bicolored Swamp Snake L NE Monard 1937 Philothamnus semivariegatus Spotted Bush Snake L NE Monard 1937 Philothamnus heterolepidotus Slender Green Snake L NE Monard 1937 Philothamnus irregularis Northern Green Bush Snake L NE Monard 1937 Philothamnus ornatus Ornate Green Snake * V NE Thelotornis capensis oatsei Western Vine Snake L NE Monard 1937 ELAPIDAE Elapsoidea guntherii Gunther’s Garter Snake L NE Bocage 1895 Naja anchietae Anchieta’s Cobra L NE Monrad 1937 Naja melanoleuca Forest Cobra L NE Bocage 1895 Naja nigrocollis Black-necked Spitting Cobra L NE Monrad 1937 LAMPROPHIIDAE Aparallactus capensis Cape Centipede Eater L NE Branch & McCartney 1992 Boaedon lineatus Striped House Snake L NE Monard 1937 Monard 1937; Branch & Gonionotophis capensis Cape File Snake L NE McCartney 1992 Lycophidion capense capense Cape Wolf Snake L NE Bocage 1895 Lycophidion multimaculatum Spotted Wolf Snake L NE Branch & McCartney 1992 Prosymna ambigua ambigua East Africa Shovelsnout Snake L NE Monard 1937 Psammophis phillipsi Phillips’ Sand Snake L NE Branch & McCartney 1992 Psammophis mossambicus Olive Whip Snake * V NE Psammophis brevirostris Short-snouted Grass Snake L NE Monrad 1937

70 | P a g e

Psammophis sibilans L Monard 1937 Psammophylax tritaeniatus Striped Skaapsteker* V, L NE Monard 1937 Pseudaspis cana Mole Snake L NE Bocage 1895 Xenocalamus mechowii inornatus Elongate Quill-snouted Snake L NE Branch & McCartney 1992 TYPHLOPIDAE Megatyphlops anomalus Angola Giant Blind-snake L NE Monard 1937 LEPTOTYPHLOPIDAE Leptotyphlops scutifrons scutifrons Peter’s Thread Snake L NE Monard 1937 Leptotyphlops sp. Thread Snake* V NE PYTHONIDAE Python sebae African Rock Python* V, L CITES II Monard 1937 VIPERIDAE Monard 1937; Branch & Bitis arietans Puff Adder* V, L NE McCartney 1992 Bitis heraldica Angolan adder L NE Bocage 1889 Causus rhombeatus Common Night Adder L NE Monrad 1937 AMPHISBAENIDAE Dalophia angolensis Worm Lizard L NE Monard 1937 Monard 1937; Branch & Dalophia pistillum Worm Lizard L NE McCartney 1992 Kalahari Round-snouted Worm Zagaspis quadrifrons Lizard L NE Monard 1937 Angola Spade-snouted Worm Monopeltis anchietae Lizard L NE Monrad 1937 VARANIDAE Varanus albigularis Rock Monitor L CITES II Monrad 1937 Varanus niloticus Water Monitor Lizard* O, L CITES II Monard 1937 CHAMAELEONIDAE Chamaeleo dilepis Flapneck Chameleon L CITES II Monard 1937 AGAMIDAE Acanthocercus cyancephalus Western Tree Agama* V, L NE Monard 1937 Agama aculeata Ground Agama* V, L NE Monard 1937 SCINCIDAE Eumecia anchietae anchietae Western Serpentiform Skink L NE Monrad 1937 Monard 1937; Branch & Sepsina angolensis Angola Skink L NE McCartney 1992 Monard 1937; Laurent, Trachylepis angolensis Skink L NE 1964 Trachylepis ivensii Iven’s skink* V NE Trachylepis varia Variable Skink L NE Monard 1937 Trachylepis striata Striped Skink L NE Monard 1937 Monrad, 1937; Branch & Trachylepis wahlbergi Wahlberg’s Striped Skink* V, L NE McCartney 1992 GEKKONIDAE Pachydactulys scutatus Large-scaled gecko L NE Monard 1937 GERRHOSAURIDAE Gerrhosaurus nigrolineatus Black-lined Plated Lizard L NE Monard 1937 Gerrhosaurus bulsi Laurent’s Plated Lizard * V NE Tetradactylus ellenbergeri Ellen's Whip Lizard* V NE LACERTIDAE

71 | P a g e

Ichnotropis bivatatta Angola Rough-scaled Lizard L NE Monard 1937 Monard 1937; Branch & Ichnotropis capensis capensis Cape Rough-scales Lizard L NE McCartney 1992 Ichnotropis grandiceps Caprivi Rough-scaled Lizard * V NE CROCODYLIDAE African Slender-snouted DD, Mecistops cataphractus Crocodile L CITES I Monard 1937 LC, Monard 1937; Branch & Crocodylus niloticus Nile Crocodile L CITES II McCartney 1992 ORDER: TESTUDINES TESTUDINIDAE Pelusios subniger subniger East Africa Black Mud Terrapin L NE Monard 1937 Pelusios sinuatus Serrated Hinged Terrapin L NE Monard 1937 Kinixys belliana Bell’s Hinged-back Tortoise* V, L CITES II Monard 1937 * sampled during this survey Total = 62 (16*)

3.5 Results

A total of 90 species of amphibians and reptiles occur in the study area. During the May 2012 survey eight new amphibian and seven new reptile species records were added to the study area (see species lists Paragraph 4). Figure 1 indicates the gap in the herpetofauna of Angola that the current survey filled. Figure 3 and 4 represent a compilation of all the current records of reptiles and amphibians for southern Angola.

Figure 3.2. Map indicating the gap in the distribution of reptiles and amphibians in southern Angola filed by the current survey.

72 | P a g e

3.5.1 Cubango River Basin In total 81 species of amphibians and reptiles were associated with the Cubanago river basin catchments to the west of the study area.

Lizards (21 species): Acanthocerus cyanocephalus, Agama aculeata, Chamaeleo dilepis, Dalophia pistillum, Dalophia angolensis, Monopeltis anchietae, Zygaspis quadrifrons, Eumecia anchietae, Gerrhosaurus bulsi, Gerrhosaurus nigrolineatus, Ichnotrophis bivittata, Ichnotropis capensis, Pachydactylus scutatus, Sepsisna angolensis, Trachylepis angolensis, Trachylepis ivensii, Trachylepis striata, Trachylepis varia, Trachylepis wahlbergi, Varanus albigularis and Varanus niloticus.

Snakes (29 species): Bitis arietans, Bitis heraldica, Boaedon lineatus, Causus rhombeatus, Crotaphopeltis hotamboeia, Dasypeltis scabra, Dipsadoboa shervei, Dispholidus typus, Elapsoidea guentherii, Gonionotophis capensis, Leptotyphlops scutifrons, Leptoptyphlops sp., Limnophis bicolor, Lycophidion capense, Megatyphlops anomalus, Naja anchietae, Naja melanoleuca, Naja nigricollis, Phillothamnus irregularis, Philothamnus heterolepidotus, Philothamnus semivariegatus, Prosymna ambigua, Psammophis brevirostris, Psammophis mossambicus, Psammophis sibilans, Psammophylax tritaeniatus, Pseudaspis cana, Python sebae and Thelotornis capensis.

Crocodilians (2 species): Crocodylus niloticus & Mecistops cataphractus

Tortoise & Terrapins (3 species): Kinixys belliana, Pelusios sinuatus & Pelusios subniger subniger

Amphibians (26 species): Amietia angolensis, Amietophrynus funereus, Amietophrynus maculatus, Amietophrynus gutturalis, Amietophrynus lemairei, Breviceps adsperus, Hemisus marmoratus, Hildebrandtia ornatissima, Hylarana darlingi, Hyperolius angolensis (parrallelus complex), Hyperolius benguellensis (nasutus complex), Hyperolius cinereus, Hyperolius seabrai, Kassina kuvangensis, Kassina senegalensis, Leptopelis anchietae, Phrynobatrachus natalensis, Phrynobatrachus mababiensis, Ptychadena oxyrhynchus, Ptychadena bunoderma, Ptychadena keilingi, Ptychadena mascareniensis, Ptychadena taenioscelis, Ptychadena uzungwensis, Tomopterna tuberculosa & Xenopus petersii.

3.5.2 Cuito River Basin

In total 24 species of amphibians and reptiles were associated with the Cuito river basin catchments to the east of the study area.

Lizards (7 species): Dalophia pistillum, Gerrhosaurus nigrolineatus, Ichnotropis capensis capensis, Ichnotropis grandiceps, Sepsina angolensis, Trachylepis striata & Trachylepis wahlbergi.

Snakes (10 species): Aparallactus capensis, Bitis arietans, Crotaphopeltis hotamboeia, Disopholidus typus, Goniotophis capensis, Lycophidion multimaculatus, Philothamnus ornatus, Psammophis mossambicus, Psammophis phillipsi, Psammophylax tritaeniatus & Xenocalamus mechowii inornatus.

Crocodilians (1 species):

73 | P a g e

Cordylus niloticus

Amphibians (6 species): Hyperolius angolensis (parallelus complex), Hyperolius benguellensis (nasutus complex), Kassina senegalensis, Ptychadena mascareniensis, Ptychadena subpunctata & Xenopus petersii.

3.5.3 Important Discoveries

Although relatively few reptiles were collected, the survey revealed a number of exciting discoveries, including:  The first record for Angola of the Caprivi Rough-scaled Lizard (Ichnotropis grandiceps). Two sub adult (see Addendum B) was collected at the HALO operational camp on the outskirts of Cuito Cuanavalle. Species belonging to this genus have been reported to have annual live cylcles. It is unclear if this specific species does the same as both adults and juveniles have been collected together (Broadley 1967).  A rarely sampled aquatic skink (Trachylepis ivensii) was recorded. It is a new record for the Cubango-Okovango system (previously only known to occur in the Zambezi, Kwanza and Congo drainage systems), from the floodplains of the Cuebe River north of Menongue (see Addendum B). Further surveys of the floodplains will add considerately to the knowledge of this species.  The fourth record of Ellenberger’s Long-tailed Seps (Tetradactylus ellenbergeri) (see Addendum B) from Angola.  The following seven species are new for the study area (see Addendum C): Gerrhosaurus bulsi, Ichnotrophis grandiceps, Leptotyphlops sp., Psammophis mossambicus, Philothamnus ornatus, Tetradactylus ellenbergeri & Trachylepis ivensii.

Figure 3.3. All known reptile records for southern Angola.

74 | P a g e

Amphibian activities was reduced. Most species had already bred, although some individuals were still present. Winter breeding species were abundant. The following exciting discoveries were made:  Four new populations of the poorly known Ashy Reed Frog (Hyperolius cinereus) were made ((see Addendum A). The species was described in 1937 by Monard and was only known from two sites (Culaqumbe and Bimbi). Subsequently Laurent (1964) assigned a population of reed frogs in Dundo (Lunda Norte) to this species. In 2009 this species was re-discovered at Humpata (Huila), this representing the most southerly distribution for the species. The new records from the current survey expand the distribution 300 km east. They also represent the first records for the Cubango-Okovango river system. The new data will help better understanding of the species’ distribution and conservation status.  Near topotypic material was collected for the Kuvangu Kassina (Kassina kuvangensis; see Addendum A). The type locality is Kuvangu [= Cubango]. This sample will be valuable for ongoing taxonomical studies.  The following eight records are new for the study area (see Addendum C): Amietophrynus maculatus, Amietophrynus gutturalis, Amietophrynus lemairei, Hyperolius cinereus, Kassina senegalensis, Phrynobatrachus mababiensis, Ptychadena subpunctata & Ptychadena taenioscelis.

Figure 3.4. All known amphibian records for southern Angola.

75 | P a g e

Way forward

The Southern Africa Regional Environmental Program (SAREP) in association with MINAMB and INIP are planning to undertake a second joint survey in early November 2012. The timing of the survey will be better suited to the sampling of emergent summer, wet season taxa, such as terrestrial and aquatic invertebrates, reptiles and amphibians, as well as for long-distant migrants such as birds. It is planned to sample the same sites as the first survey, as well as some additional new sites. Replicated monitoring is essential to determine the impacts of various drivers of change within the system and the second survey will build on the results of this. New sites will be selected that conform with the stated selection criteria of the sites sampled during this survey. Of key importance will be the inclusion of sites within the lower reaches of the Cubango and Cuito systems (Fig. 34). The proposed survey will include specialists from the following taxa;  Fish  Amphibians  Reptiles  Birds  Molluscs  Mammals (small)  Aquatic micro & macro-invertebrates  Dragonflies  Ants  Herbaceous and flowering plants  Trees

SAREP and MINAMB also wish to undertake a biodiversity survey of the new national parks in the south east of Angola; Mavina and Luengue-Luiana covering 46,000km2 and 22,600km2 respectively and adjoining the Cuito and lower Kuvango Rivers. The survey would have a terrestrial focus, but also include aquatic aspects to sample the adjoining river systems.

76 | P a g e