Survey of canine babesiosis in South Africa

A questionnaire, designed to obtain qualitative information on a number of variables concerning canine babesiosis (biliary fever) in South Africa, was sent to 510 veterinary practices in late 1993. Of the 157 practices that responded, all were presented with cases of babesiosis and most were situated in Gauteng, the Western Cape and KwaZulu-Natal. Apart from the Western Cape, a winter-rainfall region, the prevalence of babesiosis cases in dogs was highest in summer. Most of the respondent practices treated between 1,000 and 5,000 sick dogs that included 100 to 500 babesiosis cases each year. Respondents identified cerebral babesiosis, enterorrhagia, 'red' or haemoconcentrated babesiosis, acute renal failure and pulmonary babesiosis or 'shock lung', amongst others, as the most prevalent forms of complicated ('atypical') babesiosis. Diminazene, imidocarb and trypan blue were the most popular antibabesials. Trypan blue was most often used in shocked patients, whereas diminazene and imidocarb were preferred when there was a high parasitaemia in the absence of shock. At least 19 antibabesial treatment regimens were used in practices. These comprised the use of single doses of antibabesial drugs; split doses with repeat injections, and combined drug variations, some of which are undesirable due to possible sterilisation of Babesia infection or potential toxicity. Side-effects were most commonly associated with imidocarb use. Ninety-six percent of respondents used supportive treatment (e.g. corticosteroids, vitamins and 'liver support') in all cases of babesiosis. The use of blood transfusion as supportive treatment varied according to practice and severity of the case. Most practices never cross-matched blood to be transfused, and transfusion reactions were rare. Diminazene was most frequently incriminated in cases where drug 'resistance' or relapses occurred. Cerebral and 'red' cases resulted in high mortality. Treatment of babesiosis costs the dog-owning public in South Africa more than R20 million each year. Information on the distribution and possible complicating role of Ehrlichia canis was obtained. Development of a vaccine was the first research priority identified.     

Analysis of risk factors associated with recurrence of canine babesiosis caused by Babesia gibsoni

Canine babesiosis due to Babesia gibsoni (B. gibsoni) displays severe clinical manifestations. Recurrence of babesiosis after anti-babesial treatment is observable in over 10 % of the patients. The present study ascertains the risk factors and cumulative incidence of recurrence of canine babesiosis. For a sample of 145 dogs diagnosed with acute babesiosis, the following parameters were assessed over a period of 16 weeks: haematological parameters, status of anaemia, platelet count, total WBC count, haemoglobin concentration and RBC count, concurrent haemoparasitism, and secondary immune mediated haemolytic anaemia (IMHA). Patient demographics such as age, breed, sex were also recorded. The potential risk factors were statistically evaluated by the cumulative incidence function and the Kaplan-Meier method. The recurrent infections were observed in 11.8 % of the study sample. The following factors were found to associate with increased risk of recurrence: Rottweiler breed (CIR 21.8 % ± 6.9 %; p < 0.05), secondary IMHA (CIR 28.7 % ± 11.3 %; p < 0.05), RBC counts < 2 × 10⁶/μl on the day of diagnosis (CIR 16 % ± 4.6 %; p < 0.05), and persistent anaemia over 20 days post treatment (CIR 29.14 ± 7.9 %; p < 0.001). Dogs with concurrent haemoparasitic infections were predicted to have a fatal outcome in the survival analysis (disease related mortalities 25 % ± 13 %; p < 0.001). According to the findings, veterinarians need to pay attention to Rottweiler breed, dogs with secondary IMHA, concurrent haemoparasitism, low RBC counts on diagnosis and those with persistent anaemia to reduce the risk of relapse.     

Identification of three novel mycoplasma species from ostriches in South Africa

Mycoplasmas have been implicated in certain clinical syndromes in ostriches and are associated with upper respiratory tract infections. As these infections result in production losses, they are of considerable economic importance to the South African ostrich industry. Although poultry mycoplasmas have been shown to infect ostriches, the existence of unique ostrich-specific mycoplasmas has been suggested. In this study, mycoplasmas were isolated from ostriches in the Klein Karoo, Central Karoo and Garden Route areas of the Western and Northern Cape Provinces of South Africa and identified using 16S rRNA gene sequencing. These sequences indicated that ostriches in these areas carry three unique mycoplasmas and were not infected with chicken mycoplasmas. Phylogenetic analysis of the 16S rRNA sequences of the three isolated ostrich mycoplasmas showed them to be quite divergent and to fall into two distinct phylogenetic groupings. Unique sequences within the 16S rRNA gene of the ostrich mycoplasmas were subsequently used for the development of specific primers for the detection and diagnosis of mycoplasma infections in ostriches. Chickens kept in close proximity to infected ostriches were not infected with these ostrich mycoplasmas.     

Confirmation of occurrence of Babesia canis vogeli in domestic dogs in South Africa

The prevalence of Babesia infections in domestic dogs in South Africa was studied using reverse line blot hybridization and 18S sequence analysis. Babesia canis vogeli was confirmed for the first time in domestic dogs in South Africa. Out of a total of 297 blood samples collected from domestic dogs in Bloemfontein, East London, Johannesburg, Durban and from the Onderstepoort Veterinary Academic Hospital, 31 were positive for Babesia canis rossi, whereas B. c. vogeli was detected in 13 dogs. None of the dogs carried both parasites. The detection of B. c. vogeli has implications with regard to prevalence and varied clinical manifestation of canine babesiosis in South Africa.     

Retrospective study on the presence and pathogenicity of Dirofilaria repens in 5 dogs and 1 cat from Aosta Vally

Aosta Valley is the only region of Italy free of reported human cases of dirofilariosis due to Dirofilaria repens. Nonetheless, the retrospective analysis of five locally acquired D. repens infections in dogs, seems to indicate that an animal reservoir exist in the area. A feline case imported from Camargue, France, is also described. The most common dermatological signs were erythema, alopecia, crusting, papulae and nodules. All patients had lesional pruritus and dogs were affected by concurrent babesiosis. Eradication of the underlying condition, followed by a therapy with the specific adulticide and microfilaricide drugs led to a complete recovery of microfilaraemia. Increased pet travel and ecological changes are extending the ranges of vector borne, parasitic and zoonotic diseases. This fact pose a unique diagnostic challenge for the veterinarian, since the index of suspicion may be absent outside endemic areas.     

Anuran distribution,diversity and conservation in South Africa,Lesotho and Swaziland

The distributions of 96 anuran species in South Africa were mapped using published locality records, and the collections of the South African and Port Elizabeth museums. Centres of richness, endemism and Red Data Book (RDB) richness were identified, and compared with currently protected areas, using a geographical information system (GIS). Maputaland and the Kwazulu/Natal coast were identified as centres of species richness. Endemic ‘hotspots’ occur in the vicinity of Pietermaritzburg, Durban and Ihe fynbos region of the Western Cape, The Western Cape was also identified as a centre of RDB species richness. A large portion (95,8%) of South African anuran species are found in protected areas. The effective long-term viability of populations of many of these species within these areas, must be assessed. Currently four South African frog species are not found within protected areas. The Karoo biomes were identified as being under-represented, both in the species database and in the proportion of conserved areas. While a large proportion of the fynbos biome is contained in reserve areas, the low-lying fynbos habitats have almost disappeared as a result of urbanization and agriculture. Remnants of the sandy coastal fynbos and renosterveld veld types require urgent conservation. Several areas in the Western and Eastern Cape were identified as requiring additional research and conservation measures. A GIS proves a useful tool in the analysis of species distributions and the prioritization of areas and species for conservation. The importance of accurate collection data, for incorporation into species databases, and the regular publication of reserve species lists is emphasized.     

The South African form of severe and complicated canine babesiosis: clinical advances 1994-2004

Babesia (canis) rossi infection is common in dogs in South Africa, and frequently causes severe, life-threatening disease. Acidemia, persistent hyperlactatemia, hemoconcentration, elevated creatinine, cerebral babesiosis, pulmonary edema and pancreatitis are all associated with mortality rates above 30%, compared with overall mortality of 12% in admitted cases. Although half the admitted cases are severely anemic, hemoconcentration is associated with far higher mortality. Cerebral babesiosis is uncommon, but carries a poor prognosis. The pathological mechanism has been suggested to be endothelial cell damage and necrosis, followed by segmental microvascular necrosis with perivascular edema and hemorrhage. Renal involvement in babesiosis resembles the functional renal failure of sepsis. Hypotension is common, and other cardiovascular disturbances have been documented. Cerebellar ataxia, rhabdomyolysis and pancreatitis are recently identified complications. While the previous categorization into "severe" (life-threatening anemia) and "complicated" (complications not directly attributable to anemia) disease has proved useful, the distinction is artificial and probably unnecessary. An updated approach to classification is suggested, aimed at grouping animals by severity and prognosis, and using simple measures, such as clinical collapse and abnormal breathing, as much as possible. Although inflammatory mechanisms are undoubtedly important in the pathophysiology of babesiosis, there can be little doubt that tissue hypoxia plays a major role in the disease process.     

Mokola virus infection: description of recent South African cases and a review of the virus epidemiology

Five cases of Mokola virus, a lyssavirus related to rabies, are described. The cases occurred in cats from the East London, Pinetown and Pietermaritzburg areas of South Africa from February 1996 to February 1998. Each of the cats was suspected of being rabid and their brains were submitted for laboratory confirmation. Four of the cases were positive, but with atypical fluorescence, and 1 was negative. Mokola virus infection was identified by anti-lyssavirus nucleocapsid monoclonal antibody typing. As in rabies cases, the predominant clinical signs were of unusual behaviour. Aggression was present, but only during handling. Four of the 5 cats had been vaccinated for rabies, which is consistent with other studies that show that rabies vaccination does not appear to protect against Mokola virus. Since Mokola may be confused with rabies, the incidence of Mokola virus may be more common in Africa than is currently reported. As human infections may be fatal, the emergence of this virus is a potential threat to public health.     

Serological survey of Babesia bovis and Babesia bigemina in cattle in South Africa

A total of 719 serum samples collected from clinically healthy cattle from eight provinces located in different districts of South Africa were examined by the indirect enzyme-linked immunosorbent assay (ELISA) and the standard indirect fluorescent antibody test (IFAT) to determine the serological prevalence of Babesia bovis and Babesia bigemina. The results showed that 35.3% and 39.7% of cattle were positive for B. bovis and 30% and 36.5% were positive for B. bigemina antibodies on ELISA and IFAT, respectively. Mixed infections were detected in 18.2% and 26.3% of the samples using ELISA and IFAT, respectively. Consequently, the ELISAs with recombinant B. bovis spherical body protein-4 (BbSBP-4) and B. bigemina C-terminal rhoptry-associated protein-1 (BbigRAP-1/CT) were proven to be highly reliable in the serological diagnoses of bovine babesiosis in South African cattle, as evidenced by the significant concordance rates when the results were compared to those of IFAT. Moreover, the serological prevalence was significantly different among the tested provinces, in which the ranges exhibited between 15% and 73% for B. bovis infection and between 13% and 54% for B. bigemina infection. High sero-positive rates were present in Mpumalanga and KwaZulu-Natal provinces, while the lowest rate was in the North West province. Our data provide important information regarding the current seroprevalence of bovine babesiosis in South Africa, which might be beneficial in developing rational strategies for disease control and management.     

Acute pancreatitis: a newly recognised potential complication of canine babesiosis

This retrospective study describes 4 cases of canine babesiosis with histologically confirmed acute pancreatitis. In addition, 16 dogs with babesiosis are reported with serum amylase (>3500 U/l) and/or lipase (>650 U/l) activity elevations of a magnitude that would support a diagnosis of probable acute pancreatitis, although extra-pancreatic sources of the enzymes could not be excluded in these cases. Median time of pancreatitis diagnosis was 2.5 days post-admission, with primarily young (median age 3 years), sexually intact dogs affected. The development of pancreatitis was unrelated to the degree of anaemia at time of admission. In addition to pancreatitis, 80% of cases suffered from other babesial complications, namely icterus (13), acute respiratory distress syndrome (6), immune-mediated haemolytic anaemia (6), renal failure (3), haemoconcentration (2) and cerebral syndrome (2). Acute respiratory distress syndrome, renal failure and cerebral syndrome were associated with a poor prognosis, with 4 of the 5 dogs included in the overall 26% mortality rate having at least 1 of these complications. Haemolytic anaemia with ischaemia-reperfusion injury to the pancreas is proposed as a possible primary pathophysiological mechanism in babesial pancreatitis. Hypotensive shock, immune-mediated haemolytic anaemia, haemoconcentration and possibly altered lipid metabolism in babesiosis may also be involved. The previously postulated pro-inflammatory cytokine milieu of complicated babesiosis may underlie the progression, if not the primary initiation, of pancreatic pathology. Acute pancreatitis may represent the previously reported 'gut' form of babesiosis.     

Molecular characterisation of Babesia gibsoni infection from a pit-bull terrier pup recently imported into South Africa

Canine babesiosis caused by Babesia gibsoni was diagnosed in a 3-month-old Pit-bull pup during a routine clinical examination. Diagnosis was confirmed by way of smear examination, PCR, Reverse Line Blot (RLB) and sequence analysis which showed 100% homology with B. gibsoni (Japan AB118032) and Babesia sp. (Oklahoma) (AF205636). Haematology showed moderate anaemia and severe thrombocytopenia. Treatment was initiated with diminazene aceturate (Berenil RTU) followed by 2 doses of imidocarb diproprionate (Forray-65) 3 days and 14 days later, respectively. Babesia gibsoni DNA was still detectable 2 weeks post-treatment on the PCR/RLB test. A 10-day course of combination drug therapy using atovaquone and azithromycin was initiated. Blood samples taken on Day 1 and Day 40 after completion of treatment were negative for B. gibsoni DNA on PCR/RLB test. The implications of a possible introduction of B. gibsoni into South Africa are discussed.     

Nitric oxide metabolites in naturally occurring canine babesiosis.

Babesiosis, caused by the virulent haemoprotozoan parasite Babesia canis rossi, is an important disease of dogs in South Africa. The nitric oxide metabolites, nitrate and nitrite (collectively termed reactive nitrogen intermediates or RNIs) were measured in admission sera from dogs in a babesiosis-endemic area. Five groups were prospectively studied: mild uncomplicated (n=9), severe uncomplicated (severe anaemia) (n=10) and complicated babesiosis (n=11); and two groups of healthy aparasitaemic dogs: endemic controls from the study area (n=10) and experimental dogs kept in tick-free conditions (n=10). Four measures of RNI production were studied: (i) serum RNI; (ii) serum RNI/creatinine ratio; (iii) fractional clearance of RNI (FC(RNI)); (iv) fractional excretion of RNI (FE(RNI)). Marked elevations of serum RNI occurred in only two dogs, both in the severe uncomplicated group. The highest concentration (log value 5.29 micromol/l) was in a dog that died, but concentrations in the other four dogs that died were unremarkable (0, 0.34, 1.66 and 2.64 micromol/l). Age, appetite and free serum haemoglobin were significant covariates for measures of RNI production. There were no significant differences among the babesiosis groups for serum RNI. Adjustment for creatinine had minor effects on the results. All babesiosis groups had significantly higher serum RNI and RNI/creatinine than the tick-free control group, but did not differ from the endemic controls except for the severe uncomplicated group, which had higher RNI/creatinine. The complicated group had significantly lower FC(RNI) and FE(RNI) than all other groups, except for the tick-free control group, which had similar FE(RNI). The results indicate that, in an endemic area, measures of RNI production are unlikely to be useful indicators of severity or outcome in canine babesiosis.     

Furunculosis in rainbow trout (Salmo gairdneri) raised in sea water

Ulcerative skin lesions were encountered in rainbow trout raised in sea water by a commercial concern in the Western Cape, South Africa. Grossly, the lesions resembled furunculosis but, histopathologically, they differed from typical furunculosis in that bacterial colonies were rarely found in the organs, and also the kidneys and spleens were minimally involved. The causative organism was identified as an achromogenic Aeromonas salmonicida that shared characteristics with all 3 subspecies, salmonicida, masoucida and achromogenes. This is the first report of an outbreak of this disease in South Africa.     

Feline babesiosis: signalment, clinical pathology and concurrent infections

Fifty-six cats with naturally occurring Babesia felis infection were studied. No breed or sex predilection could be identified, but there was an apparent predilection for young adult cats less than 3 years of age. Macrocytic, hypochromic, regenerative anaemia was present in 57% of the cats and in-saline agglutination tests were positive in 16%. No characteristic changes were observed in total or differential leukocyte counts. Thrombocyte counts were variable and thrombocytopaenia was an inconsistent finding. Hepatic cytosol enzyme activity and total bilirubin concentrations were elevated in the majority of cats. Serum protein values were mostly normal, but increased values were occasionally observed and polyclonal gammopathies were observed in all cats with increased total globulin concentrations. No remarkable changes in renal parameters were observed. A variety of electrolyte abnormalities occurred in a number of cats, but no consistent pattern of change could be identified. A close correlation was evident between peripheral and central parasite counts. Concurrent infections with Haemobartonella felis, feline immunodeficiency virus and/or feline leukemia virus were identified in a number of cats.     

Patterns of distribution and conservation status of freshwater fishes in South Africa

The combined fish collection databases of the Albany Museum and the J.L.B. Smith Institute of Ichthyology are used to identify hotspots of endemism and threatened fish distributions in South Africa. Hotspots of fish species richness occur in the north-eastern lowveld sectors of South Africa and along the ecotone between the tropical/ subtropical and temperate faunal zones. Hotspots of endemic fish richness occur within both the tropical and temperate faunal regions, notably in the Olifants River system, Western Cape and in areas of high relief such as the Cape Fold Mountains, the Amatola-Wrnterberg (Eastern Cape), and the Drakensberg Escarpment {Kwazulu/ Natal-Eastern Transvaal). Threatened taxa are concentrated in the hotspots of endemic species richness which coincide largely with areas of major river conservation concern. There is limited scope for fish conservation within the ambit of formal (or informal) declared reserves, and the survival of the fauna depends on the success of river catchment conservation management. The value of museum collections in identifying areas of conservation concern for freshwater fishes is emphasized, and highlights the importance of well-preserved voucher specimens for biodiversity conservation.     

Seroprevalence of toxoplasmosis in sheep in South Africa

Serum samples from 600 sheep were collected from 5 different provinces randomly chosen in South Africa. Two sheep abattoirs (representing formal slaughter of sheep) and 1 rural location (representing informal slaughter of sheep) per province were also selected randomly. The serum samples were tested for anti-Toxoplasma gondii IgG antibodies using 2 different serological tests: an indirect fluorescent antibody (IFA) test and an enzyme-linked immunosorbent assay (ELISA) test available as a commercial kit. This study provides the first published data on seroprevalence of toxoplasmosis in sheep in South Africa, although positive titres have been found previously in wild felids, ferrets, chinchillas and a dog. Data on seroprevalence in sheep is considered important because consumption of mutton is universally considered to be a source of zoonotic transfer to humans. Seroprevalence in humans in South Africa was previously found to be 20 % and it is postulated that this may be linked to the informal slaughter and consumption of mutton. During this study, the overall national seroprevalence per province in sheep was found to be 5.6 % (IFA) and 4.3 % (ELISA), respectively. This is lower than in other countries, possibly because South Africa has an arid climate. Differences in seroprevalence in different areas studied suggested an association with the climate and a significant correlation (P > 0.05) was detected between the prevalence of T. gondii and the minimum average temperature. The seroprevalence was found to be significantly higher (P < 0.01) in sheep originating from commercial farms (7.9 %) than in rural sheep in the informal sector (3.4 %). Also, sheep managed extensively had a seroprevalence of 1.8 %, which was significantly lower (P < 0.05) than the seroprevalence in sheep under semi-intensive or intensive management systems (5.3%). An incidental finding of interest was the considerable movement of sheep to abattoirs and mutton after slaughter. The highest consumption of mutton was in the Western Cape Province (29.9%) while the highest concentration of sheep is found in the Eastern Cape Province (30.1%).     

Clinical manifestations of canine babesiosis in Hungary (63 cases)

Clinical observations of Babesia canis infection in 63 dogs during a 1-year period are summarised, demonstrating the pathogenicity of the Babesia strain endemic in Hungary. Most patients had babesiosis in the spring and autumn, correlating with the seasonal activity of ticks. Male animals appeared in higher numbers, probably due to an overrepresentation of outdoor dogs. Uncomplicated babesiosis was diagnosed in 32 cases. The disease affected dogs of any age in this study. Symptoms were similar to those published from other parts of the world: lethargy, fever, splenomegaly, pallor, icterus, haemoglobinuria and presence of ticks were the most common observations. Thrombocytopenia, lymphopenia and neutropenia were frequent haemogram changes. Imidocarb appeared to be highly effective in eliminating the Babesia infection. Thirty-one animals demonstrated babesiosis with complications. Most Rottweilers (7/9) developed complicated disease. Old age was a risk factor for multiple complications. Multiple organ manifestations had poor prognosis. Hepatopathy (44%), pancreatitis (33%), acute renal failure (ARF; 31%) and disseminated intravascular coagulation (DIC; 24%) were frequent complications, while immune-mediated haemolytic anaemia (IMHA; 10%), acute respiratory distress syndrome (ARDS; 6%) and cerebral babesiosis (3%) were rarely observed. There was a significant difference between the mean age of dogs having uncomplicated disease, babesiosis with a single complication and babesiosis with multiple complications (3.4, 4.8 and 8.6 years, respectively, p < 0.001). The recovery rate (78, 68 and 25%, respectively, p = 0.005) and mortality rate (3, 21 and 67%, respectively, p < 0.001) also tended to differ significantly in these groups. Systemic inflammatory response syndrome (SIRS) and DIC are two possible pathways leading to multiple organ dysfunction syndrome (MODS) in babesiosis. DIC was found to predict MODS more sensitively in this study than SIRS: there were 6 animals developing MODS out of 11 identified with DIC, while only 5 dogs developed MODS out of 22 having SIRS.     

Feline herpesvirus infection in a group of semi-captive cheetahs

Clinical disease caused by feline herpesvirus type-1 in wild felid species is similar to that in domestic cats. Herpesviruses are endemic in free-ranging lions in South Africa but actual clinical disease due to them has not been reported in free-ranging felids. The first reports of feline herpesvirus infection associated with clinical disease in wild felids came from Australia and the USA in 1970. Subsequent reports of clinical disease in cheetahs and other wild felid species were limited to captive animals. This report deals with clinical disease in a group of semi-captive cheetahs in which 18 animals were affected, and included 12 adult males, 4 adult females and 2 subadults. No mortalities occurred in this group, the most common clinical signs being sneezing, nasal discharge and loss of appetite.     

Gastroduodenal ulceration in cats: eight cases and a review of the literature

Gastroduodenal ulceration (GU) and blood loss was diagnosed in eight cats and compared with 25 previously reported cases of feline GU. Cats with GU presented in a critical condition. Clinical signs consistent with gastrointestinal bleeding were infrequently identified although anaemia was a common finding. Non-neoplastic causes of feline GU tended to have a shorter clinical course with ulcers confined to the stomach. Conversely, cats with tumour-associated GU usually had a more protracted clinical course, weight loss, and ulcers located in the stomach for gastric tumours and the duodenum for extra-intestinal tumours. In this series, definitive diagnosis was possible for cats with neoplasia (gastric tumours and gastrinoma), however, it was difficult to precisely identify the underlying aetiology in cats with non-neoplastic GU. Prompt stabilisation with a compatible blood transfusion, surgical debridement or resection, antibiotic and antiulcer therapy, and treatment of the underlying disease, if identified, was successful in the majority of cases. The prognosis for cats with appropriately managed GU depended on the underlying aetiology, but even cats with neoplasia could be successfully palliated for prolonged periods.     

Diagnosis of feline haemoplasma infection using a real-time PCR assay

Haemobartonella felis has been reclassified within the genus Mycoplasma as Mycoplasma haemofelis and 'Candidatus Mycoplasma haemominutum', collectively referred to as the feline haemoplasmas. A total of 78 cats from the Johannesburg area that had blood samples submitted to a private veterinary laboratory were tested using a real-time polymerase chain reaction (PCR) assay able to detect and distinguish the two feline haemoplasma (basonym Haemobartonella) species. All samples had been diagnosed with haemoplasma infection by cytological examination of blood smears. Statistical analysis was performed to evaluate associations between haemoplasma status, age, and haematological and biochemical parameters. On PCR assay 43 cats (55%) were haemoplasma negative, 25 (32.1%) positive for 'Candidatus Mycoplasma haemominutum', 5 (6.4%) positive for Mycoplasma haemofelis and 5 (6.4%) positive for both species. Significant inverse correlation was found between the amount of M. haemofelis DNA present in the blood and the haematocrit value. Cats that were positive for M. haemofelis showed macrocytic regenerative anaemia, monocytosis and thrombocytopaenia. This report documents the existence of both haemoplasma species in cats in South Africa.