Ecological Parameters of Selected Helminth

ECOLOGICAL PARAMETERS OF SELECTED HELMINTH SPECIES IN LABEOBARBUS AENEUS AND LABEOBARBUS KIMBERLEYENSIS IN THE VAAL DAM, AND AN EVALUATION OF THEIR INFLUENCE ON INDICATORS OF ENVIRONMENTAL HEALTH

ALESSANDRA BERTASSO

SHORT DISSERTATION

submitted in partial fulfilment of the requirements for the degree

MASTER OF SCIENCE

ENVIRONMENTAL MANAGEMENT

in the

FACULTY OF SCIENCE

at the

RAND AFRIKAANS UNIVERSITY

SUPERVISOR: PROF. A. AVENANT-OLDEWAGE MAY 2004

ACKNOWLEDGEMENTS

My sincere thanks and gratitude go to:

∗ My partner in life and love, Dwayne, for all his support, love and continual motivation.

∗ My supervisor, Professor A. Avenant-Oldewage, for her support and patience throughout my study.

∗ Rand Water for making water quality data available.

∗ Rand Afrikaans University for financial support.

∗ Zoology Department at RAU for use of facilities.

∗ The staff and students at the Zoology Department for assistance in field work.

∗ Groot Eiland staff for their help and use of facilities.

∗ S.N. Mashego for his assistance in identifying the tapeworms collected.

∗ Riette Eiselen from RAU Statkon for statistical analyses of my data.

∗ My colleague and friend, Bronwyn Gernet for assisting with the maps.

∗ My Parents, for giving me the opportunity to further my studies and for their unconditional love and support.

ABSTRACT

Surveys conducted by the Rand Afrikaans University fish parasitology group have shown unexpectedly high numbers of helminth parasites (endoparasites) in yellowfish species in the Vaal Dam. The high number of helminth parasites can be attributed to a cestode species (Bothriocephalus spp.) which has been introduced with cyprinid fish into South Africa from Asia. It was expected that this opportunistic introduced species, in its high numbers, may affect the accuracy of the fish health assessment index (HAI). The main purpose of this research project was to determine the infection (in terms of prevalence, mean intensity and abundance) of the helminth species found in yellowfish in the Vaal Dam and determine whether it plays a role in calculating the health status of the fish as expressed in the Health assessment Index (HAI) and thus have an impact on the value obtained for the quality of the environment they live in.

Seasonal surveys were conducted at the Vaal Dam between April 2000 and January 2001. Twenty smallmouth yellowfish (Labeobarbus aeneus) and 20 largemouth yellowfish (Labeobarbus kimberleyensis) were collected with the aid of gill nets and used as indicator fish species. The two yellowfish species sampled were identified based by the distinctive size of their mouths. Fish were examined for external parasites after which they were weighed and measured. Blood was drawn and slime smears were made. Blood and slime smears were examined with a compound microscope for the presence of parasites. Fish were killed, dissected and then examined as prescribed in the fish HAI. Internal parasite numbers were recorded. Helminth parasites were extracted from the intestines of infected fish and examined. The cestodes were identified as either Bothriocephalus acheilognathi Yamaguti, 1934 based on the heart-shaped scolex and presence of bothria or “other cestode spp.”. The majority (99.8%) of the cestodes found for both yellowfish species were identified as B. acheilognathi. The prevalence, mean intensity and abundance of B. acheilognathi (Asian tapeworm) in both yellowfish species were calculated. Ecological parameters including species specificity, seasonality, gender specificity and relationships between fish size and the Asian tapeworm prevalence were also included. From the endo- and ectoparasite data collected, infestation statistics and ecological parameters (species specificity, seasonality and gender specificity) were calculated. Two HAI values were calculated for each fish species; one including all parasites found and the second excluding the number of B. acheilognathi found. To verify the results of the HAI, water quality, made available by Rand Water, was included in the study.

The Vaal Dam, located in the upper reaches of the Vaal River, is believed to have reasonably good water quality. This theory was supported in this study; water quality in the Vaal Dam during the four surveys was relatively good with high turbidity, high dissolved oxygen and low TDS characteristics. In this study B. acheilognathi preferred Lb. kimberleyensis over Lb. aeneus although a low infection rate was observed in smallmouth yellowfish. Furthermore the infection (in terms of prevalence, abundance and mean intensity) in largemouth yellowfish was markedly high. Seasonal patterns observed in the Asian tapeworm’s infection of smallmouth yellowfish were attributed to breeding and subsequent

feeding patterns of this fish species with relatively high infections recorded in winter and spring. For Lb. kimberleyensis no explanation could be given regarding the seasonal patterns observed for the mean intensity and abundance of B. acheilognathi. The maximum and minimum mean intensity and abundance values in largemouth yellowfish were recorded in autumn and spring respectively. In addition the prevalence of B. acheilognathi was high throughout the four surveys. Regression analysis indicated that the presence/absence of B. acheilognathi most successfully predicted (80.6%) the fish as belonging to the correct fish species. In terms of the HAI, higher values were calculated for Lb. kimberleyensis than Lb. aeneus when including the number of Asian tapeworms in the HAI calculations. When excluding the number of Asian tapeworms from the HAI calculations, slightly higher but similar values were obtained for both fish species. Seasonal changes in water quality in the Vaal Dam were not reflected in the Asian tapeworm’s infection or the HAI. Therefore B. acheilognathi and its high infection in Lb. kimberleyensis, directly influences the high health index value obtained for this fish species. However it does not reflect a poorer water quality as would be expected when observing higher HAI values. Statistical analyses indicate that if the HAI excluding the number of B. acheilognathi is a good indication of fish health then the HAI including all parasites would also be a good indication of fish health except the HAI values would be 1.03 times higher.

OPSOMMING

Opnames wat uitgevoer is deur die visparasitologiese groep by die Randse Afrikaanse Universiteit het onverwagte hoë getalle helmint parasiete (endoparasiete) in geelvisspesies in die Vaaldam gevind. Die hoë getalle helminte kan toegeskryf word aan die voorkoms van ‘n lintwurm (Bothriocephalus sp) oorspronklik ingevoer saam met verteenwoordigers van die Cyprinidae uit Asië. Daar was verwag dat die hoë getalle van hierdie oppertunistiese ingevoerde spesie en die akkuraatheid van die Visgesondheidsindeks sal beïnvloed. Die hoofdoel van die navorsingsprojek was om die infeksie (in terme van persentasiebesmetting, gemiddelde besmettingsintensiteit en besmettingsmoontlikheid) van die helmintspesies in geelvisse in die Vaaldam te bepaal en vas te stel of dit ‘n effek het op die berekening van die gesondheidstatus van die visse soos uitgedruk in die Visgesondheidsindeks (VGI) en dus ‘n rol speel by waarde wat verkry word vir die omgewing waarin hulle leef.

Seisoenlikse opnames is uitgevoer in die Vaaldam tussen April 2000 en Januarie 2001. Twintig kleinbekgeelvisse (Labeobarbus aeneus) en 20 grootbekgeelvisse (Labeobarbus kimberleyensis) is versamel met behulp van kieuenette en aangewend as indikatorspesies. Die twee geelvisspesies is geïdentifiseer op grond van die grootte van hulle bekke. Visse is deurgesoek vir die teenwoordigheid van uitwendige parasiete en daarna geweeg en gemeet. Bloed is onttrek en ‘n slymsmere is gemaak. Bloed en slymsmere is met behulp van ‘n saamgestelde mikroskoop vir die teenwoordigheid van parasiete ondersoek. Vis is gedood, gedissekteer en ondersoek soos voorgeskryf in die VGI. Inwendige parasiete se getalle is aangeteken. Helmintparasiete is verwyder uit die intestinum van besmette visseen bestudeer. Die cestode is geïdentifiseer as òf Bothriocephalus acheilognathi Yamaguti, 1934 gebasser op die hartvormige skoleks en aanwesigheid van bothria òf “’n ander cestood spesie”. Die meerderheid (99.8%) van die cestode wat in beide geelvisspesies voorgekom het, is geïdentifiseer as B. acheilognathi. Die persentasiebesmetting, gemiddelde besmettingsintensiteit en besmettingsmoontlikheid van B. acheilognathi (Asiese lintwurm) in beide geelvisspesies is bereken. Ekologiese parameters soos spesiespesifisiteit, seisoenaliteit, geslagsvoorkeur en die verband tussen visgrootte en die voorkoms van die lintwurm is ingesluit. Uit die endo- en ektoparasietdata wat versamel is, is infestasiestatistiek bereken. Twee waardes is bereken vir die VGI; een waar alle parasiete se voorkoms ingesluit is en een waar die getalle verkry vir B. acheilognathi uitgesluit is. Ten einde die waardes verkry vir die VGI te verifieer is waterkwaliteitsdata, wat van Randwater verkry is, ingesluit in die studie.

Daar word geglo dat die Vaaldam wat in die bolope van die Vaalrivier geleë is oor redelike goeie waterkwaliteit beskik. Hierdie teorie word ondersteun in die studie; waterkwaliteit in die Vaaldam gedurende die vier opnames was relatief goed met hoë waardes vir turbiditeit, hoë opgeloste suurstof en lae totale opgeloste soliede partikels. In die studie het B. acheilognathi Lb. kimberleyensis verkies bo Lb. aeneus en ‘n lae infeksiekoers is waargeneem by kleinbekgeelvis. Voorts was die besmetting (in terme van persentasiebesmetting, gemiddelde besmettingsintensiteit en besmettingsmoontlikheid) merkbaar hoër in grootbekgeelvisse. Seisoenlikse patrone waargeneem in die Asiese lintwurm se

besmetting van kleinbekgeelvisse is toegeskryf aan broei- en gevolglike voedingsgewoontes van die visspesie met redelike hoë infeksies in die winter en lente. Vir Lb. kimberleyensis kon geen verklaring gebied word vir die seisoenlikse variasie in intensiteit en besmettingsmoontlikheid met B. acheilognathi. Die maksimum en minimum gemiddelde besmettingsmoontlikheid en besmettingsintensiteitswaardes vir grootbekgeelvis is onderskeidelik in die herfs en lente aangeteken. Voorts was die persentsiebesmetting gedurende al vier opnames hoog. Regressie analises het aangetoon dat die aan/afwesigehid van B. acheilognathi suksesvol kan voorspel (80.6%) tot watter spesie ‘n bepaalde vis behoort het. In terme van die VGI is hoër waardes bereken vir Lb. kimberleyensis as vir Lb. aeneus wanneer die getal Asiese lintwurms ingesluit is. Wanneer die getal Asiese lintwurms uitgesluit word, word effe hoër maar soortgelyke waardes vir beide spesies verkry. Seisoenlikse veranderinge in watergehalte in die Vaaldam was nie gereflekteer in die Asiese lintwurm se voorkoms of die VGI nie. Gevolglik het die hoë voorkoms van B. acheilognathi ‘n direkte rol gespeel in die verkryging van die hoë waardes vir die visgesondheidsindeks van die visspesie. Tog word swakker waterkwaliteit nie gereflekteer wanneer hoër VGI waardes waargeneem word nie, soos verwag is. Statistiese analieses dui aan dat indien die visgesondheidsindeks met B. acheilognathi uitgesluit ‘n goeie aanduider is van die visgesondheid die VGI met alle parasiete ingesluit ook ‘n goeie aanduider van visgesondheid is maar dat die waardes 1.03 keer hoër is.

DECLARATION

I declare that this report is my own original work. It is being submitted in partial fulfilment of the requirements for the degree of Master of Science in the Faculty of Science of the Rand Afrikaans University, Johannesburg, South Africa. It has not been submitted before for any other degree or examination in any other university.

______Alessandra Bertasso

TABLE OF CONTENTS

1. INTRODUCTION...... 1

1.1. INTRODUCTION TO THE HEALTH ASSESSMENT INDEX...... 1

1.2. INTRODUCTION TO THE CURRENT STUDY...... 2

1.3. STRUCTURE OF THIS REPORT ...... 3 2. THE VAAL DAM AND WATER QUALITY ...... 4

2.1. INTRODUCTION ...... 4

2.1.1. THE VAAL RIVER SYSTEM ...... 4

2.1.2. THE STUDY SITE ...... 5

2.1.3. WATER QUALITY AND THE SOUTH AFRICAN WATER QUALITY GUIDELINES ...... 8

2.1.4. FACTORS POTENTIALLY INFLUENCING THE QUALITY OF WATER IN THE VAAL DAM ...... 9

2.2. MATERIALS AND METHODS ...... 13

2.3. RESULTS ...... 13

2.3.1. SURFACE WATER VARIABLES...... 13

2.3.2. MACRO WATER ANALYSIS ...... 14

2.3.3. TRACE METAL ANALYSIS ...... 16

2.4. DISCUSSION...... 17

2.4.1. SURFACE WATER VARIABLES...... 18

2.4.2. MACRO WATER ANALYSIS ...... 21

2.4.3. TRACE METAL ANALYSIS ...... 26

2.5. SUMMARY AND CONCLUSION ...... 30 3. ASPECTS OF THE ECOLOGY OF THE ASIAN TAPEWORM, BOTHRIOCEPHALUS ACHEILOGNATHI YAMAGUTI, 1934 FOUND IN YELLOWFISH IN THE VAAL DAM ...... 33

3.1. INTRODUCTION ...... 33

3.2. MATERIALS AND METHODS ...... 34

3.2.1. THE STUDY LOCATION...... 34

3.2.2. COLLECTION OF FISH...... 35

3.2.3. EXTRACTION AND COLLECTION OF CESTODES ...... 35

3.2.4. IDENTIFICATION OF CESTODES ...... 35

3.3. RESULTS ...... 37

3.3.1. IDENTIFICATION OF CESTODES ...... 37

3.3.2. PARASITE NUMBERS ...... 40

3.3.3. INFECTION STATISTICS OF Bothriocephalus acheilognathi...... 40

3.3.4. ECOLOGICAL PARAMETERS ...... 42

3.4. DISCUSSION...... 43

3.5. SUMMARY AND CONCLUSION ...... 47

4. THE HEALTH ASSESSMENT INDEX...... 48

4.1. INTRODUCTION ...... 48

4.2. MATERIAL AND METHODS ...... 50

4.2.1. THE STUDY LOCATION...... 50

4.2.2. FISH HEALTH ASSESSMENT INDEX (HAI) AND ASSOCIATED PARASITE INDEX ...... 50

4.2.2.1. FIELD WORK...... 50

4.2.2.2. LABORATORY PROCEDURES...... 52

4.2.2.3. CALCULATIONS ...... 53

4.3. RESULTS ...... 55

4.3.1. PARASITES ...... 55

4.3.1.1. PARASITES COLLECTED ...... 55

4.3.1.2. INFESTATION STATISTICS FOR ECTO- AND ENDOPARASITES...... 56

4.3.2. HEALTH ASSESSMENT INDEX AND ASSOCIATED PARASITE INDEX ...... 59

4.3.2.1. HEALTH ASSESSMENT INDEX ...... 59

4.3.2.2. LOGISTIC REGRESSION ...... 62

4.3.3. CONDITION FACTOR ...... 63

4.3.4. ECOLOGICAL PARAMETERS ...... 64

4.4. DISCUSSION...... 66

4.5. SUMMARY AND CONCLUSION ...... 74 5. GENERAL SUMMARY AND CONCLUSION...... 75 6. REFERENCES...... 80

Introduction

1. INTRODUCTION CHAPTER 1 INTRODUCTION

“If you violate nature’s laws you are your own prosecuting attorney, judge, jury and hangman.” - Luther Burbank

In South Africa, the quality of water resources was primarily determined by carrying out chemical analysis of water samples and measuring physical variables (Roux, Van Vliet and Van Veelen 1993). Chemical analysis has the disadvantage that it is only accurate at the time that the sample was taken (Abel 1989). With the above in mind, alternative methods of determining the quality of water sources are continuously being investigated. One of the alternative methods is biological monitoring. Biological monitoring is an integral part of monitoring programmes (Roux et al. 1993). Biological responses have proven to be effective in determining the effect of changing environmental conditions and allow effective protection of aquatic resources. This technique is already being used on a continuous basis in North America and the British Isles to determine the effect of pollution on the environment (Goede and Barton 1990; Klemm, Stober and Lazorchak 1992; Adams, Brown and Goede 1993). In South Africa the fish Health Assessment Index (HAI) and associated Parasite Index is one of the many biomonitoring tools that can be used to constructively contribute to environmental management in South Africa (Avenant-Oldewage and Swanepoel 1993).

1.1. INTRODUCTION TO THE HEALTH ASSESSMENT INDEX Since 1994, studies have been conducted in the Olifants River to determine the effectiveness of a Health Assessment Index developed in the USA as a bio-monitoring tool under South African conditions. Results obtained by Marx (1996) showed that fish populations with higher HAI values are found in poor quality water, while better quality water harboured healthier fish populations with lower HAI values. Further studies conducted by Watson (2000), in the Olifants River revealed that the incorporation of endoparasites in the fish HAI provides a more accurate indication of water quality. High numbers of endoparasites and hence a higher HAI value reflects a highly polluted area while a high number of ectoparasites and hence a lower HAI reflects a better quality of water and healthier fish populations. This model gives the Olifants River an accurate forecast of approximately 90% in terms of water quality (Avenant-Oldewage 2001). Lyons, Wang and Simonson (1996) made the statement that a biomonitoring index yielding reliable results in one system would not necessarily perform the same in a different system. The HAI was therefore tested by Crafford (Crafford 2000) over a two-year period, from 1999 to 2000, in the Vaal River system and more specifically the Vaal Dam and Vaal Barrage. The Vaal Dam has always been considered as a control point (non-polluted) in research work whereas the Vaal Barrage has been considered as the polluted area. Results obtained by Crafford (2000) substantiated the theory behind the HAI. He found that HAI values were higher for the Vaal Barrage and lower for the Vaal Dam thereby distinguishing successfully between the two localities based on water quality. The number of endoparasites was higher at the Barrage and

Introduction the number of ectoparasites was higher at the Vaal Dam thereby supporting the theory proposed by Watson (2000). In addition, the HAI has been used by Robinson, Hines, Sorensen and Bryan (1998) to assess the effects of parasite infestation on the health of two endangered desert fish.

Recent field trips by the Rand Afrikaans University fish parasitology group have shown unexpectedly high numbers of helminth parasites (endoparasites) in fishes at the Vaal Dam, which shown in previous studies should have low or less endoparasites than fishes in the Barrage. It is thought that the high number of helminth parasites (sometimes as many as 450) can be attributed to a cestode species, namely Bothriocephalus spp. which has been introduced into South Africa (Van As, Schoonbee and Brandt 1981; Mashego 1982) from Asia. It is also thought that this introduced species may affect the accuracy of the health assessment index (HAI). This high infection of cestodes however has no apparent effect on the health of the fish as is evident from the condition factors and HAI values.

1.2. INTRODUCTION TO THE CURRENT STUDY The main purpose of this research project was to determine the infection (in terms of prevalence, mean intensity and abundance) of the helminth species found in yellowfish in the Vaal Dam and determine whether it plays a role in calculating the health status of the fish as expressed in the Health assessment Index (HAI) and thus have an impact on the value obtained for the quality of the environment they live in. This study had the following aims: ∗ to determine whether the change in water quality during the various seasons is reflected in the Health Assessment Index whereby the parasite index is included; ∗ to identify the tapeworms to species level (Bothriocephalus acheilognathi); ∗ to determine whether the health/condition of the fish is related to water quality; and ∗ to determine if the exclusion of the introduced tapeworm has an effect on or plays a role in the fish Health Assessment Index.

The study location was the Vaal Dam. The smallmouth yellowfish (Labeobarbus aeneus Burchell 1822) (renamed from Barbus aeneus) and largemouth yellowfish (Labeobarbus kimberleyensis Gilchrist and Thompson 1913) (renamed from Barbus kimberleyensis) were used in this study as indicator fish species. Both these species are important South African angling fish (Skelton 1993; Skelton 2001). Lb. kimberleyensis is becoming scarce and is being artificially cultured and restocked (Skelton 1993; Skelton 2001). The fish Health Assessment Index as described by Avenant-Oldewage et al. (1995) and printed (with the inclusion of additional variables) in Marx (1996) [based on the necropsy-based system developed by Goede and Barton (1990) and refined by Adams et al. (1993)] was employed with the addition of a colour chart developed by Watson (2001). The Inverted Parasite Index assessed by Crafford (2000) was utilized. To verify the results of the HAI, water quality supplied by Rand Water at the Vaal Dam was included in the study. Helminth parasites were extracted from infected fish and examined. The helminths were identified as B. acheilognathi Yamaguti, 1934. Statistical methods were used to determine the prevalence, mean intensity and abundance of

Introduction

B. acheilognathi (Asian tapeworm).

1.3. STRUCTURE OF THIS REPORT The results and findings of this study are submitted as outlined below. ∗ Chapter 2 describes the Vaal Dam, the catchment in which it is located, the water quality and the factors influencing its water quality. This chapter contains the water quality data. ∗ Chapter 3 focuses on aspects of the ecology of the cestode endoparasite B. acheilognathi Yamaguti, 1934 present in large numbers during the surveys. ∗ Chapter 4 discusses the HAI and associated inverted PI in determining the health of the fish studied. ∗ Chapter 5 provides a short general summary of the results obtained and conclusions drawn. ∗ Chapter 6 contains references used in this dissertation. References referred to in separate chapters are included in this chapter to avoid repetition.

Water quality

2. THE VAAL DAM AND WATER QUALITY CHAPTER 2 THE VAAL DAM AND WATER QUALITY

“Our liquid planet glows like a soft blue sapphire in the hard-edged darkness of space. There is nothing else like it in the solar system. It is because of water.” – John Todd

2.1. INTRODUCTION Without water ‘earth would have no oceans, no life as it exists and no people’ (Miller 1999). Water quality as defined by Department of Water Affairs and Forestry (1996a) describes the ‘physical, chemical, biological and aesthetic properties of water which determine its fitness for a variety of uses and for protecting the health and integrity of aquatic ecosystems’. This chapter describes the study site in terms of its geographical location including the river system and catchment in which it is located, its water quality and factors potentially influencing its environmental well-being.

2.1.1. THE VAAL RIVER SYSTEM The Vaal River rises on the western slopes of the Drakensburg escarpment (Braune and Rogers 1987) in the vicinity of Lake Chrissie near Breyten (Department of Water Affairs and Forestry 1993) in Mpumalanga. From Lake Chrissie the river flows in a west-south-west direction across the interior plateau to join the Orange River near Douglas in the Northern Cape Province (Braune and Rogers 1987). The Orange River ultimately flows into the Atlantic Ocean at Alexander Bay (Department of Water Affairs and Forestry 1993). The Vaal River runs for a distance of approximately 900 km (Braune and Rogers 1987) and traverses ten categories of veld types as described by Acocks (1988). The change in veld types results in diversity of landscapes with various scenic characteristics, fauna and flora life.

The major tributaries of the Vaal River drain the Drakensberg in the east, the Witwatersrand in the north and the Maluti mountains in the south. Major dams on the Vaal River include the Grootdraai, Vaal Dam, Vaal Barrage, Bloemhof, Vaalharts and Douglas Weir (Braune and Rogers 1987; Department of Water Affairs and Forestry 2000).

The Vaal River has a catchment area of approximately 192 000 km2 (19.2 million hectares) (Braune and Rogers 1987) and is the main source of water for the central industrial, mining and metropolitan regions in South Africa (Kasrils 2000). It supplies water to the industrial powerhouse of the whole country (Department of Water Affairs and Forestry 2000), including Gauteng, parts of Mpumalanga, Northern Province, North West Province, Orange Free State and Northern Cape Province (Kasrils 2000). More than 50% of South Africa’s gross geographical product (GGP) and more than 80% of the country’s electricity is generated by the area supplied by the Vaal River system (Department of Water Affairs and Forestry 2000). This area also includes some of the largest gold, platinum and coal mines in the world (Department of Water Affairs and Forestry 2000).

Water quality

As explained by Braune and Rogers (1987), the Vaal River catchment can be divided into four zones based on water quality (Figure 2.1). The four zones, in order of decreasing water quality, are the Vaal Dam, the Barrage, Bloemhof and Douglas Weir catchment. The water quality along the Vaal River is in line with the general water uses along the river, ranging from the highest quality requirements for power stations in the Vaal Dam catchment to comparatively lower quality for agriculture in the lower reaches of the system (Braune and Rogers 1987). The Vaal Dam is discussed in further detail below.

Figure 2.1: Water quality zones of the Vaal River catchment (adopted from Braune and Rogers 1987)

2.1.2. THE STUDY SITE The sampling sites used in the current study are located in the Vaal Dam, near to RAU Island (Groot Eiland) (S 26o 52.249′, E 28o 10.249′). The largest portion (60%) of the Vaal Dam catchment is located in the Free State Province with the remainder being located in the Gauteng and Mpumalanga Provinces (Department of Water Affairs and Forestry 2000) (Figure 2.2). The location of the Vaal Dam on the Vaal River is illustrated on Figure 2.3. Key features of the Vaal Dam and the Vaal Dam catchment are outlined in the text box below.

The Vaal Dam is situated at an altitude of between 1 200 and 1 450 m. Topography surrounding the dam slopes gently towards the dam. The dam is located in the Highveld Climatic Zone as defined by Schulze (1965). The rainfall is mostly due to showers and thunderstorms and falls mainly in summer

Water quality from October to March with the maximum fall THE VAAL DAM occurring in January. On the whole winds are light Surface area - Approximately 32 107 hectares (321 km2) except for short periods during thunderstorms Capacity - Approximately 2 536 million m3 (Schulze 1965). Mean annual runoff to dam - ± 8% of annual rainfall Rainfall - average annual of ± 700 mm The Vaal Dam drainage area is underlain by rocks Evaporation - annual potential of ± 1 500 mm predominantly of the Karoo Supergroup which when Rivers feeding the dam - Vaal, Liebenbergvlei, eroded liberate mainly illitic clay minerals (Birch Wilge and Nuwejaarspruit Rivers Catchment in which dam is located - Vaal Dam 1983). As highlighted by Birch (1983), the Vaal catchment which straddles two quaternary sub Dam also contains sediments enriched in kaolinite catchments C12L and C83M and smectite. VAAL DAM CATCHMENT Catchment area - Approximately 38 500 km2 (approximated in area by a 261km by 287km The Vaal Dam is one of the major dams in the Vaal rectangle) River catchment (Braune and Rogers 1987). The Water Management Area - Upper Vaal WMA dam has the largest surface area and second Sources: Department of Water Affairs and Forestry 2000; Water Research Commission largest dam capacity in the Vaal River catchment 1994; Department of Water Affairs and Forestry making it the fourth largest storage reservoir in 1991a; Braune and Rogers 1987. South Africa (Department of Water Affairs and Forestry 1993). As highlighted in the text box, evaporation from the Vaal Dam greatly exceeds the average rainfall it receives.

The Vaal Dam acts as the central structure in the Vaal River system and Vaal River water supply system (Department of Water Affairs and Forestry 1993). The Vaal Dam is the primary supplier of potable water to the heartland of the South African economy (Department of Water Affairs and Forestry 1991a). More than 25% of the country’s people (i.e. more than 10 million people) are dependent on water stored in the Vaal Dam. The Vaal Dam catchment was proclaimed a government water control area in terms of the old Water Act (54 of 1956) in 1970 in order to protect the available sources against over-exploitation (Department of Water Affairs and Forestry 1993).

Control over the dam and surface water rests with Department of Water Affairs and Forestry and the distribution of water is undertaken mainly by the Rand Water Board within its area of supply (Office of the Prime Minister 1982; Department of Water Affairs and Forestry 1993). The Department of Water Affairs and Forestry and the Rand Water Board are responsible for routine monitoring of water quality. Each organisation monitors specific points within the Vaal Dam.

2.1.3. WATER QUALITY AND THE SOUTH AFRICAN WATER QUALITY GUIDELINES In 1993 the Department of Water Affairs and Forestry published the first edition of the South African Water Quality Guidelines. These water quality guidelines are scientific and technical information provided for a particular water quality constituent in the form of numerical data and/or narrative descriptions of its effects on the fitness of water for a particular use or on the health of aquatic ecosystems (Hohls, Silberbauer, Kuhn, Kempster and Van Ginkel 2002). There are seven guidelines

Water quality each for a specific water use, namely domestic, recreational, industrial, irrigated agriculture, livestock watering (agricultural), aquaculture (agricultural) and aquatic ecosystems.

For the purposes of this study, the water quality data given below is compared to the target water quality range (TWQR) as outlined in the water quality guidelines for domestic use and aquatic ecosystems. The TWQR for a particular constituent and water use is defined as the ‘range of concentrations or levels at which the presence of the constituent would have no adverse or anticipated effects on the fitness of the water assuming long term continuous use and for safeguarding the health of aquatic ecosystems’ (Department of Water Affairs and Forestry 1996a). These two user groups, along with irrigated agriculture use, generally have the most stringent requirements for water quality (Hohls et al. 2002).

2.1.4. FACTORS POTENTIALLY INFLUENCING THE QUALITY OF WATER IN THE VAAL DAM Water is an excellent solvent and transport medium for particulates and therefore tends to become polluted by both natural processes and man-induced processes and waste (Hohls et al. 2002). Natural processes influencing water quality includes topography (influences climatic conditions and surface run-off), climate (rainfall volumes and distribution over space and time), geology (determines riverbed characteristics and water chemistry), soils (determines natural nutrient and salt loads) and vegetation (influences stream flow and surface runoff from water banks) (Avenant-Oldewage and Eddy 2003).

Goudie and Viles (1997) explained that the human activities conducted on land adjacent to rivers and water bodies dominate the quality of natural river systems. Land cover and land use is known to affect water quality (Hohls et al. 2002). Land cover in the Vaal Dam catchment as shown on the 1:250 000 CSIR National Landcover map of South Africa comprises mainly grassland (59.6%) and cultivated land (26.5%) with a small percentage comprising built-up areas (urban, commercial, industrial) (0.65%), mines (and quarries) (0.1%), forests (0.34%), natural vegetation (0.67%), wetlands and waterbodies (1.04%), degraded land (0.35%) and bare rock and soil (0.14%). Apart from the natural vegetation and wetland and waterbodies which would have little pollution effect (as explained by Hohls et al. 2002), factors caused by the land uses mentioned above are included in the sections below.

About 60% of South Africa’s mines and industries occur in the broader Vaal River catchment area (Department of Water Affairs and Forestry 2003). A report written by Braune and Rogers (1987) specifically dealing with the Vaal River catchment and its problems highlighted the factors listed below as potentially affecting water quality in the Vaal Dam catchment. ∗ Gold mining The majority of South Africa’s gold mines lie within the Barrage, Bloemhof Dam and Vaal Dam catchments (Braune and Rogers 1987). However, only a few are found in the Vaal Dam catchment. Gold mine effluent could affect water quality in the Vaal Dam (Crafford 2000). The main causes of water pollution from gold mining operations is acid mine drainage, old mine

Water quality

dumps, washing processes and the use of explosives. The resulting impacts include low pH values (2 to 4.4), high salt content (TDS ranging from 1 500 to 4 000 mg/ℓ), dissolved iron and sulphuric acid (Braune and Rogers 1987). ∗ Coal mining (underground and opencast) Most of South Africa’s coal occurs in the northern and eastern parts of the country (Gauteng, eastern Orange Free State and northern Natal) (Department of Mines 1980). There are 19 coalfields identified in South Africa (Department of Minerals and Energy Affairs 1989). In the Vaal River catchment, coal mining is limited to the Vaal Dam and Barrage catchments and consists mainly of underground mining with some opencast mining (Braune and Rogers 1987). Existing coalfields in the Vaal Dam catchment include the Highveld and Mpumalanga coalfields (Department of Minerals and Energy Affairs 1989). In relation to these, there are various collieries in the Vaal Dam catchment (Department of Minerals and Energy Affairs 1990). Some collieries produce export coal (for example Ermelo), some are linked to power stations (for example New Denmark colliery and Tutuka power station and Springfield colliery and Grootvlei power station) and other collieries are linked to synthetic fuel plants (for example Secunda colliery and Sasol II and Sasol III plants) (Figure 2.3). There are two potential coal reserves near the Vaal Dam; the one runs in a north easterly to south westerly direction under the Vaal Dam while the second is located below the Wilge River as it feeds into the Vaal Dam (Office of the Prime Minister 1982). The potential impacts of coal reserves and coal mining operations on water quality in the Vaal Dam catchment include increased acidity, sulphates and TDS, lowered oxygen content and destruction of living organisms. ∗ Industrial development Department of Water Affairs and Forestry (1993) is concerned about the increase in pollution in the Vaal River in general due to diffuse sources from industry. The major wet industrial operations in South Africa include power generation, textile manufacturing, paper and pulp production, iron and steel, synthetic fuels and abattoirs (Avenant-Oldewage and Eddy 2003). Of these the most significant in the Vaal Dam catchment include power generation and fuel production (Braune and Rogers 1987). As mentioned above, the Sasol II and Sasol III synthetic fuel plants are found in the Vaal Dam catchment (Figure 2.3). The main water-related problems are as a result of atmospheric pollution, oil spills, highly alkaline ash water and effluent pollution (sodium, fluorides, non-biodegradable organic compounds, phosphorus, nitrates, ammonia). Atmospheric pollution in one of the long term pollution threats to the Vaal Dam catchment (Braune and Rogers 1987). Spills and effluent pollution to the environment would have serious immediate effects on the local stream however these problems should not occur if they are well managed on site (Braune and Rogers 1987). There is a possibility that spills from Sasol’s operations can have an impact on water quality in the Vaal Dam. Although industries co-operate well in the prevention of pollution, the assimilation capacity of the Vaal River for certain water quality variables is exceeded and it is foreseen that industries in these areas will in future have to comply with even stricter point specific effluent standards (Department of Water Affairs and Forestry 1993). However the diffuse sources still need to be determined.

Water quality

∗ Urban (including squatting) development Urban sources which impact on water quality include, amongst others, soil erosion, litter, decay of vegetation, application of fertilizers and pesticides to gardens, car washes, swimming pool backwashing and animal waste (Avenant-Oldewage and Eddy 2003). The main water-related impacts associated with urban developments are as a result of activities leading to eutrophication, solid waste pollution, hydrological effects and organic contamination from domestic and industrial effluents discharges (Braune and Rogers 1987). Solid wastes from municipal areas and informal settlements can lead to the occurrence of pesticides, heavy metals (lead associated with use of petrol and zinc associated with galvanized steel roofs and gutters), salts and polychlorinated hydrocarbons in water (Braune and Rogers, 1987; Avenant-Oldewage and Eddy 2003). Should this water find its way into the Vaal Dam it could cause pollution. Many of the dams in South Africa suffer from eutrophication (a process of nutrient enrichment). The Vaal River catchment in general suffers from eutrophication as a result of the high levels in phosphates in urban effluent (O’Keefe et al. 1992). Although phosphates can occur naturally, the increase in human activities is causing eutrophication to occur more rapidly (O’Keefe et al. 1992). As can be seen in Figure 2.3, there are various towns located along the tributaries feeding into the Vaal Dam. In the immediate vicinity of the Vaal Dam, there are two townships Deneysville and Oranjeville (Office of the Prime Minister 1982) and a third informal township known as Metsimaholo. The locations of these are shown on Figure 2.2. The exact location and extent of informal settlements in the vicinity of the dam is unknown however it is expected that these would negatively influence the quality of water in the Vaal dam. In addition, there are various holiday homes and resorts located on the banks of the Vaal Dam (author’s personal observation). Construction of these homes and resorts result in land disturbance which causes erosion, and the associated release of plant nutrients, metals and organic compounds (Avenant-Oldewage and Eddy 2003). Domestic sewage water from Deneysville and Oranjeville is treated in septic tanks or discharged into stone quarries or other suitable sites. The Office of the Prime Minister (1982) highlights that adequate measures are in place to prevent pollution of the Vaal Dam. A purification works (small package plant) is located at Vaal Marina on the Vaal Dam (Figure 2.2). Effluent from these works is used for irrigation purposes (Office of the Prime Minister 1982). ∗ Agriculture The broader Vaal River catchment is responsible for approximately 42% of South Africa’s agricultural production (Braune and Rogers 1987) with agricultural activities such as intensive cultivation, mixed farming and stock farming covering the whole catchment (Slabber 1980 cited by Braune and Rogers 1987). The population of the Vaal Dam area consisted mainly of an agricultural community (Office of the Prime Minister 1982). The main irrigated crops grown within the Vaal Dam catchment include maize, vegetables, flowers, grazing and grass (Department of Water Affairs and Forestry/Rand Water 1996). Agricultural practices which could impact on water quality include the application of fertilizers and pesticides, livestock rearing and feedlots (Avenant- Oldewage and Eddy 2003). The impact on water quality in the Vaal Dam could include increased salinity, decreased turbidity, nutrient pollution (from fertilisers and animal food) and pesticide

Water quality

pollution. One of the long term pollution threats to the Vaal Dam catchment, as described by Braune and Rogers (1987), is diffuse agricultural sources. ∗ Recreational use Although the Vaal Dam was primarily built to store and supply water, at present due to the lack of large stretches of water in the PWV Complex and the increasing demand for open air recreational facilities, the Vaal Dam and its riparian areas are being used for recreational activities (Office of the Prime Minister 1982). Recreational activities include, amongst others, developments (homes and resorts) in or close to the riparian zone; consumptive activities such as fishing, swimming, canoeing, water skiing, sailing and powerboat racing; and non consumptive activities such as bird watching, wildlife watching, camping, weekend accommodation and picnicking (Department of Water Affairs and Forestry/Rand Water 1996). The major concentration of recreational activities (consumptive and non-consumptive) occurs between the Vaal Dam and Bloemhof (Braune and Rogers 1987) (Figure 2.1). Although the area immediately surrounding the Vaal Dam (approximately 500 m from the waters edge), except Orangeville and Deneysville, is zoned as recreation or tourist attractions, nature areas or open spaces (Office of the Prime Minister 1982) several developments (as can be seen by the small black squares on Figure 2.2) occur within the riparian zone of the Vaal Dam. There are several holiday homes and resorts located on the banks on the Vaal Dam (author’s personal observations). Problems caused by recreational use of the Vaal Dam could include fuel and oil spills, bank erosion and re-suspension of sediment from power boating, modification of the riparian zone and introduction or spread of alien plants (Braune and Rogers 1987). ∗ Lesotho Highlands Water project (LHWP) The LHWP is a bi-national water transfer scheme that imports water from the upper reaches of the Orange River in Lesotho (the Senqu River) (Department of Water Affairs and Forestry 1993) to the Vaal Basin in South Africa (Lesotho Highlands Water Project 1995). The main purpose of the project is to capture the Lesotho high summer rainfalls through dams and to transfer this water to parts of South Africa (Lesotho Highlands Water Project 1995; Horizons of Winter 1997). The project involves 5 major dams, 200 kilometres of tunnels and a hydro electricity station in Lesotho (Horizons of Winter 1997). The ongoing phases of the project will import increasing amounts of water. Phase 1 of the proposed four phase scheme was completed this year (2004) resulting in a total transfer rate of 30 m3 of water per second (Department of Water Affairs and Forestry 1993). It is proposed that by the year 2020, the LHWP will be able to deliver approximately 70 m3/s of water from the highlands of Lesotho to the Vaal River system (Lesotho Highlands Water Project 1995). It is expected that the less turbid nature of the high mountain waters would decrease the turbidity of water in the Vaal Dam catchment. This in turn would increase algal production resulting in eutrophication (Braune and Rogers 1987). Braune and Rogers (1987) do however indicate that the later phases of the LHWP will yield more turbid water.

It must be noted that further research is needed to determine all the factors that could affect water quality in the Vaal Dam (Department of Water Affairs and Forestry 1993).

Water quality

2.2. MATERIALS AND METHODS For the duration of the study water quality data was made available from the Rand Water Board. The Rand Water Board sampling point used was C-VD1I (Rand Water Board Inlet at the Vaal Dam wall) as shown on Figure 2.2. Although this water sampling point is approximately 6 km from the fish sampling sites, water quality in the Vaal Dam is fairly consistent throughout (Van Wyk 2001, personal communication1). The data was collected during routine monitoring activities made by the Rand Water Board during April 2000, June 2000, October 2000 and January 2001. These months correlate with the months in which fish were sampled at the Vaal Dam (information on fish sampling is given in Chapter 4). Standard techniques were used by the Rand Water Board to analyse the water samples. For some of the months (April and October 2000), various parameters were not measured. The reason could be that sampling and/or measuring instruments were either not in working order or being serviced.

2.3. RESULTS The results are presented in terms of surface water variables (pH, temperature, light penetration and turbidity, conductivity and dissolved oxygen), macro determinants (calcium, chlorine, fluoride, calcium carbonate, potassium, magnesium, sodium, ammonium, nitrate, phosphate, sulphates, silicon and total dissolved solids) and trace metals (aluminium, cadmium, chromium, copper, iron, manganese, nickel, lead and zinc).

2.3.1. SURFACE WATER VARIABLES Seasonal changes in surface water quality variables (pH, temperature, light penetration, conductivity and dissolved oxygen) are illustrated in Figure 2.4 and the results presented in this section.

32 30 28 26 24 22 20 pH (pH units) 18 16 Temperature (Degrees Celsius) 14 Electrical conductivity (mS/m)

Legend units 12 Dissolved Oxygen (mg/l O2) 10 8 Secchi disc readings (cm) 6 4 2 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surveys

Figure 2.4: Graph depicting surface water quality variables recorded seasonally at the Vaal Dam

1 F. van Wyk 2001, personal communication. Rand Water, Head: Catchment Management. Tel: (011) 682-0480.

Water quality pH: Although pH values were fairly constant, a slight variation in seasonal values is exhibited. The pH values ranged from 7.34 in winter to 8.26 in summer. The data shows a seasonal decrease in pH values from autumn to winter followed by an increase in pH in spring and a further increase in summer.

Temperature: As for pH, temperature values showed a slight variation between seasons even though no value was record in autumn. The minimum surface water temperature was recorded in winter (19.9oC) and the maximum in summer (25oC). Temperature increased seasonally in the following order: winter spring summer.

Light penetration/turbidity: Very little variation was observed in the Secchi disc readings (light penetration). The minimum and maximum readings were 28 cm (spring and summer) and 30 cm (winter), respectively. There was no reading recorded in autumn. Turbidity values (not shown in Figure 2.4), on the other hand exhibited a greater variation with a maximum of 100 NTU (nephelometric turbidity unit) recorded in autumn and a minimum of 50 NTU recorded in summer. Values decreased in the following order: autumn winter spring summer. It would be expected that an increase in Secchi disc readings would correlate with a decrease in turbidity; however this is not the case in this study. Secchi disc readings and turbidity values both increased throughout the seasons although Secchi disc readings were fairly constant.

Electrical conductivity: Electrical conductivity values varied little between autumn, winter and spring after which a significant increase in conductivity was observed in summer. The electrical conductivity values ranged from 23 mS/m (summer) to between 16 mS/m (winter) and 17 mS/m (autumn and spring).

Dissolved oxygen: The only variable which exhibited a noticeable variation (as can be seen in Figure 2.4) between the seasons is dissolved oxygen ranging from 6.3 mg/ℓ in summer to 12.5 mg/ℓ in winter. However, there were no readings recorded in autumn and spring.

2.3.2. MACRO WATER ANALYSIS Macro determinants (expressed in mg/ℓ) for the Vaal Dam are given in Table 2.1 and the results presented in this section. In Table 2.1, values exceeding the domestic use TWQR are highlighted in grey; values exceeding the TWQR for aquatic ecosystems are highlighted in bold.

Calcium (Ca): Calcium concentrations varied considerably from 10 mg/ℓ in autumn (April 2000) and 16 mg/ℓ in summer to 44 mg/ℓ in winter. No reading was recorded in spring.

Fluoride (F): Fluoride concentrations remained fairly constant during all seasons with values ranging from 0.16 mg/ℓ (autumn and winter) to 0.17 mg/ℓ (summer). However, no reading was recorded in spring.

Water quality

Table 2.1: Concentrations (in mg/ℓ) of water macro determinants analysed at the Vaal Dam # Season Month Ca Cl F CaCO3 K Mg Na Autumn April (2000) 10.00 <10.00 0.16 50.50 3.40 6.20 9.80 Winter June (2000) 44.00 <10.00 0.16 132.52 4.30 5.50 10.00 Spring October (2000) * <10.00 * * * * * Summer January (2001) 16.00 <10.00 0.17 79.48 2.20 9.60 4.60 TWQR Aquatic ecosystems NA ≤0.0002 ≤0.75 NA NA NA NA Domestic use 0-32 0-100 0-1 50-100 0-50 0-30 0-100

Season Month NH4 NO3 PO4 SO4 Si TDS Autumn April (2000) <0.05 0.52 0.12 15.00 1.00 * Winter June (2000) <0.05 0.60 <0.05 <10.00 0.64 130.00 Spring October (2000) 0.05 0.45 <0.05 20.00 * * Summer January (2001) <0.05 0.11 <0.05 20.00 4.70 145.00 TWQR Aquatic ecosystems ≤0.007 NA ∆ NA NA ∆ Domestic use 0-1.0 0-6 NA 0-200 NA 0-450 NOTES: * No reading available. # Calcium carbonate (total hardness) values calculated from calcium and magnesium concentrations as follows: 2.497 x [mg Ca/ℓ] + 4.118 x [mg Mg/ℓ] (Department of Water Affairs and Forestry 1996a). ∆ Concentrations should not be changed by more than 15% from the normal cycles of the water body under unimpacted conditions at any time of the year and amplitude and frequency of natural cycles should not be changed. NA No guideline available. TWQR Target Water Quality Range given in mg/ℓ (South African Water Quality Guidelines, Department of Water Affairs and Forestry 1996c). The domestic use TWQR is based on water quality fit for human consumption.

Chloride (Cl): Recorded chloride concentrations were less than 10 mg/ℓ during all four surveys.

Calcium carbonate (CaCO3): Calcium carbonate (total hardness) was calculated using the calcium and magnesium concentrations (see Table 2.1) recorded by Rand Water during the four surveys as follows [formula given by Department of Water Affairs and Forestry (1996a)]:

[mg CaCO3/ℓ] = 2.497 x [mg Ca/ℓ] + 4.118 x [mg Mg/ℓ] Concentrations varied considerably from 50.5 mg/ℓ in autumn (minimum) to 132.5 mg/ℓ in winter

(maximum). A CaCO3 concentration could not be calculated for spring as there was no calcium or magnesium concentration recorded in spring in this study. The following seasonal trend was observed: concentrations increased from autumn to winter and then decreased in summer.

Potassium (K): Potassium concentrations showed little variation ranging from a minimum of 2.2 mg/ℓ in summer to a maximum of 4.3 mg/ℓ in winter. A concentration of 3.4 mg/ℓ was recorded for autumn. No reading was recorded for spring.

Magnesium (Mg): Recorded magnesium concentrations varied with the lowest concentration (5.5 mg/ℓ) recorded in winter, followed by a concentration of 6.2 mg/ℓ recorded in autumn and the highest concentration (9.6 mg/ℓ) recorded in summer. No reading was recorded for spring.

Water quality

Sodium (Na): Sodium exhibited similar concentrations in autumn (9.8 mg/ℓ) and winter (10 mg/ℓ) with a considerable decrease in concentration in summer (4.6 mg/ℓ). No reading was available in spring.

Ammonium (NH4): During all four surveys, recorded ammonium concentrations were 0.05 mg/ℓ (spring) or less than 0.05 mg/ℓ (autumn, winter summer).

Nitrate (NO3): Except for summer where a lower nitrate concentration of 0.11 mg/ℓ was recorded, concentrations for the remaining three surveys varied between 0.45 mg/ℓ (spring) and 0.60 mg/ℓ (winter).

Phosphate (PO4): Phosphate concentrations remained constant (less than 0.05 mg/ℓ) during winter, spring and summer. In autumn however a considerably higher concentration of 0.12 mg/ℓ was recorded.

Sulphate (SO4): Sulphate concentrations varied slightly from less than 10 mg/ℓ in winter to between 15 mg/ℓ in autumn and 20 mg/ℓ in spring and summer.

Silicon (Si): The highest silicon concentration (4.7 mg/ℓ) was recorded in summer. Concentrations in autumn and winter were 1 mg/ℓ and 0.6 mg/ℓ, respectively. No reading was recorded in spring.

Total Dissolved Solids (TDS): TDS concentrations showed little variation ranging from 130 mg/ℓ in winter to 145 mg/ℓ in summer. No readings were recorded for the autumn and spring surveys.

2.3.3. TRACE METAL ANALYSIS Trace metal concentrations (expressed in mg/ℓ) for the Vaal Dam are given in Table 2.2 and the results presented in this section. In Table 2.2, values exceeding the domestic use TWQR are highlighted in grey; values exceeding the TWQR for aquatic ecosystems are highlighted in bold.

Aluminium (Al): Aluminium concentrations varied from less than 0.1 mg/ℓ and 0.29 mg/ℓ in winter and summer respectively to 1.9 mg/ℓ in autumn. No reading was recorded for spring.

Cadmium (Cd): Cadmium concentrations remained constant for three of the four surveys (autumn, winter and summer) at a concentration less than 0.05 mg/ℓ. No reading was recorded in spring.

Chromium (Cr): Recorded chromium concentrations were 0.05 mg/ℓ (winter) or less than 0.05 mg/ℓ (autumn and summer). No reading was recorded for spring.

Copper (Cu): Concentrations for copper remained constant during the autumn, winter and summer surveys at a value less than 0.10 mg/ℓ. No readings were recorded for the spring survey.

Water quality

Table 2.2: Concentrations (in mg/ℓ) of trace metals analysed at the Vaal Dam Season Month Al Cd Cr Cu Fe Autumn April (2000) 1.90 <0.05 <0.05 <0.10 1.10 Winter June (2000) <0.10 <0.05 0.05 <0.10 0.14 Spring October (2000) * * * * * Summer January (2001) 0.29 <0.05 <0.05 <0.10 0.11 TWQR Aquatic ecosystems ≤0.005 ≤0.00015** <0.007 <0.0003** ∞ Domestic use 0-0.15 0-0.005 0-0.05 0-1 0-0.1

Season Month Mn Ni Pb Zn Autumn April (2000) <0.10 <0.10 <0.10 <0.10 Winter June (2000) <0.10 <0.10 <0.10 <0.10 Spring October (2000) * * <0.10 * Summer January (2001) <0.10 <0.10 <0.10 <0.10 TWQR Aquatic ecosystems ≤0.18 NA ≤0.0002 ≤0.002 Domestic use 0-0.05 NA 0-0.01 0-3 NOTES: * No reading available. ** Where there is a series of TWQR depending either on the pH or hardness of the water, the most conservative value is included in the table. ∞ Concentrations should not be allowed to vary by more than 10% of the background concentration for a particular site or case, at a specific time. NA No guideline available. TWQR Target Water Quality Range given in mg/ℓ (South African Water Quality Guidelines, Department of Water Affairs and Forestry 1996c). The domestic use TWQR is based on water quality fit for human consumption.

Iron (Fe): Similar iron concentrations were recorded in summer (0.11 mg/ℓ) and winter (0.14 mg/ℓ) with an increased concentration recorded in autumn (1.1 mg/ℓ). No reading was recorded for spring.

Manganese (Mn): Except for spring (where no readings were recorded), manganese cconcentrations remained constant during all surveys at a concentration less than 0.10 mg/ℓ.

Nickel (Ni): Nickel concentrations exhibited no variation. Concentrations remained constant for the autumn, winter and summer surveys at a value less than 0.10 mg/ℓ. No readings were recorded during the spring survey.

Lead (Pb): Lead concentrations remained constant (less than 0.01 mg/ℓ) during all four surveys.

Zinc (Zn): Concentrations for zinc remained constant for the autumn, winter and summer surveys at a value less than 0.10 mg/ℓ. No readings were recorded for the spring survey.

2.4. DISCUSSION Factors that aggravate water pollution can include the physical and chemical properties of the water and pollutant. Certain substances occur naturally in an ecosystem and are harmless to living organisms (Mason 1991). Many water quality problems are as a result of geological characteristics of the source area and therefore a decrease in the quality of water of a specific water body or system cannot be attributed to a specific point source (discharges from industry, discharges from sewage treatment works, agriculture) or non-point source (atmospheric fallout, stormwater runoff, groundwater

Water quality influences, disturbance of sediments, agricultural activities, mining activities) (Department of Water Affairs and Forestry 1991b).

Environmental factors can influence the toxicity of these substances. The toxic effects of pollutants vary with the quality of the water (Mason 1991). Temperature, pH and hardness are the main factors affecting toxicity. These factors in terms of this study are discussed below in Section 2.4.1 (temperature and pH) and Section 2.4.2 (hardness). However it must be noted that a single factor never works alone; there is always a combination of factors operating in an ecosystem (Mason 1991).

2.4.1. SURFACE WATER VARIABLES The pH value is a measure of the hydrogen ion activity in a water sample (Department of Water Affairs and Forestry 1996b). It affects the taste and corrosivity of the water (Department of Water Affairs and Forestry 2002). Most freshwaters, in South Africa, are relatively well buffered and are more or less neutral with pH values ranging from 6 to 8 (Department of Water Affairs and Forestry 1996b). In this study, pH values were within the above-mentioned range except in summer where a slightly higher pH value was observed. However, the Department of Water Affairs and Forestry’s guideline document for domestic use (Department of Water Affairs and Forestry 1996a) mentions that the pH of most raw waters lies between 6.5 and 8.5 due to the geology and geochemistry of underlying rocks and soils. If this pH range is taken into account, then pH values for this study are still within Department of Water Affairs and Forestry’s acceptable range. Van Veelen et al. (1990) classify a pH range of 6.6 to 8.5 as “suitable for most uses”. It is important to note however that the toxicity of a substance is dependent on a specific pH and that a change in pH can affect the chemical speciation of various metals making them more toxic. The ionic form of a metal is generally more toxic (Mason 1991). The pH of surface water may vary as a result of discharges of effluent (industrial, municipal), runoff, acidic rainfall and microbial activity (Department of Water Affairs and Forestry 1996a). In this study, pH values varied slightly between seasons although remaining fairly constant (Figure 2.4). A difference of 0.92 pH units was measured between the maximum and minimum values recorded. The seasonal variation showed the following trend: a decrease from autumn to winter followed by an increase in spring and a further increase in summer.

Temperature is defined by Department of Water Affairs and Forestry (1996b) as “the condition of a body that determines the transfer of heat to, or from, other bodies”. The temperature of inland waters in South Africa generally ranges from 5 to 30oC (Department of Water Affairs and Forestry 1996b). In this study temperatures varied from 19 to 25oC (Figure 2.4). Temperature influences both the organisms (change in metabolic activity and behaviour) and the chemical and physical state of the pollutant (Mason 1991). In general, toxicity increases with temperature (Mason 1991) however there are always exceptions to the rule. Temperature could vary as a result of heated industrial effluent discharges, heated return flows of irrigation water, removal of riparian vegetation cover (increasing the amount of solar radiation reaching the water) and inter-basin water transfers (Department of Water Affairs and Forestry 1996b). All organisms associated with freshwater, excluding birds and mammals,

Water quality are poikilothermic. In other words, they are unable to control their body temperatures and are therefore highly dependent on ambient water temperature and very susceptible to changes (Department of Water Affairs and Forestry 1996b). Temperature plays an important role in the life cycles of many organisms acting as a cue for the timing of migration, spawning and emergence (Department of Water Affairs and Forestry 1996b). In this study, surface water temperatures varied slightly between seasons with a difference of 6oC between winter (lowest temperature recorded) and summer (highest temperature recorded). As was expected, temperatures increased seasonally in the following order: winter, spring, summer. No value was recorded in autumn however it is expected that the temperature in autumn would be slightly higher than winter but lower than in summer.

Turbidity is the measure of light scattering ability of water and is indicative of the concentration of suspended matter in water (Department of Water Affairs and Forestry 1996a). Turbidity refers to how clear the water is (Water on the Web 2003). Turbidity can either be measured using a nephelometer (ideal) or Secchi disc which measures visibility (light penetration - the depth to which one can see into the water). In this study both visibility and turbidity were measured by Rand Water. Visibility (water clarity) in the Vaal Dam was fairly high and stable during the four surveys (Figure 2.4). Similar Secchi readings to those recorded in this study were obtained in a separate study conducted by Crafford (2000) in the Vaal Dam from January 1999 to February 2000 showing very little variation in light penetration in the Vaal Dam.

When reviewing the turbidity of water in the Vaal Dam, it would be expected that turbidity values would be fairly low (correlating with the high Secchi readings) due to the transfer of water from the less turbid Lesotho Highlands (Braune and Rogers 1987; Graham 2004, personal communication2) however this is not seen in this study. Turbidity values were extremely high and varied greatly between seasons. All seasons had a turbidity value significantly higher than the TWQR for domestic use. These high turbidity results correlate with a study conducted by Groenewald (2000) during 1997 and 1998 in the Vaal Dam indicating little change in turbidity over four years. The effects of such high turbidity values include severe aesthetic effects such as appearance, taste and odour; and significant effects on the microbiological quality of the water (Department of Water Affairs and Forestry 1996a). For fish living in this system, the high turbidity could lead to poor visibility and reduced feeding rates (Water on the Web 2003). High turbidity values observed in this study could explain the reason why excessive algal blooms are not seen in the Vaal Dam when compared to other water bodies. Turbidity also affects the risk of infectious disease transmission (Department of Water Affairs and Forestry 2002). Soil particles, discharge of sewage and other wastes can contribute significantly to turbidity (Department of Water Affairs and Forestry 1996a). In the Vaal Dam, the high turbidity is related to the types of clay found in the dam (Birch 1983). Seasonal variations in turbidity values were as follows: values decreased from autumn to winter to spring to summer. It would be expected that an increase in Secchi disc readings would correlate with a decrease in turbidity. However in this study, Secchi disc readings and turbidity values had the same seasonal decrease although Secchi disc readings were fairly constant and no

2 Dr M Graham 2004, personal communication. Umgeni Water, Pietermaritzberg. Tel: (033) 341-1111.

Water quality reading was recorded in autumn. The reason for the same seasonal trend is unknown.

Electrical conductivity is the measure of the ability of water to conduct electrical current (Department of Water Affairs and Forestry 1996b) and estimates the amount of total dissolved solids or dissolved ions in the water (Water on the web 2003). The ability to conduct electrical current is as a result of the presence of the following ions in water, which carry an electrical charge: carbonate, bicarbonate, chloride, sulphate, nitrate, sodium, potassium, calcium and magnesium (Department of Water Affairs and Forestry 1996b). As will be seen in the section discussed below (Section 2.4.2), all these ions are present in the Vaal Dam water. However the electrical conductivity in the Vaal Dam for all four surveys was fairly low when compared to the TWQR for domestic use of 0 to 70 mS/m. Electrical conductivity values for all four surveys were below 24 mS/m and therefore within the TWQR (Figure 2.4). There is no TWQR for aquatic ecosystems. Electrical conductivity, along with total dissolved solids (discussed in Section 2.4.2 below), serves as a general indicator of change in water quality and effects the taste and “freshness” of the water (Department of Water Affairs and Forestry 2002).

Dissolved oxygen is one of the best indicators of the health of a water ecosystem (Cleveland 1998). Dissolved oxygen is needed for the respiration of all aerobic organisms and therefore is critical for the survival and functioning of aquatic biota (Department of Water Affairs and Forestry 1996b). Low dissolved oxygen levels combined with high water temperatures can compound stress effects on aquatic organisms (Department of Water Affairs and Forestry 1996b). A decrease in the dissolved oxygen levels is usually an indication of an influx of some type of organic pollutant (Cleveland 1998). The amount of suspended material in the water affects the saturation concentration of dissolved oxygen either chemically or physically. Factors affecting saturation solubility (non-linearly) include temperature, salinity, atmospheric pressure and other site-specific chemical and physical factors. Typical saturation at sea level and at TDS value below 3 000 mg/ℓ is 9.09 mg/ℓ at 20oC (Department of Water Affairs and Forestry 1996b). As explained by Department of Water Affairs and Forestry (1997) the atmospheric pressure on the Highveld is 20% lower than at sea level. Therefore the saturation concentration for identical waters will differ by the same percentage meaning that in the Highveld typical saturation is expected to be 7.272 mg/ℓ at 20oC (20% lower than sea level). In this study, values varied from 6.3 mg/ℓ at 25oC to 12.5 mg/ℓ at 19.9oC (Figure 2.4). The dissolved oxygen concentration in this study for water with a temperature of 20oC is almost double the expected concentration. Oxygen concentrations above saturation can cause gas bubble disease (Department of Water Affairs and Forestry 1996b). Increases in saturation correlate with a decrease in temperature and similarly a decrease in saturation will correlate with in an increase in temperature (Department of Water Affairs and Forestry 1996b). This is observed in this study where a lower dissolved oxygen concentration was recorded in summer and a higher concentration in winter. Seasonal variations also arise from changes in biological productivity.

Water quality

2.4.2. MACRO WATER ANALYSIS This section discusses the macro determinant concentrations presented in Section 2.3.2 in terms of the South African target water quality ranges (TWQR) as defined by the Department of Water Affairs and Forestry (1996a, 1996b and 1996c).

Calcium (Ca) occurs naturally in most waters and is the main constituent of water hardness. It is an essential element for all living organisms and an important constituent of bony skeletons in mammals. The solubility of calcium is governed by the carbonate/bicarbonate equilibrium and is therefore strongly influenced by temperature and pH. Calcium itself influences the integrity of cell membranes and therefore strongly influences the absorption and toxicity of heavy metals (Department of Water Affairs and Forestry 1996a). Department of Water Affairs and Forestry (1996a) defines a typical concentration of calcium in freshwater as 15 mg/ℓ. Calcium concentrations recorded in this study were slightly lower (10 mg/ℓ) than the typical concentration in autumn and slightly higher (16 mg/ℓ) in summer but still within the TWQR for domestic use (Table 2.1). The calcium concentration in winter exceeded the TWQR for domestic use considerably (Table 2.1) however the elevated concentration has no health effects. The elevated concentration in winter could have an affect on the toxicity of heavy metals. There is no calcium TWQR available for aquatic ecosystems.

Elemental chlorine (Cl2) is too reactive to persist in the aquatic ecosystem (Department of Water Affairs and Forestry 1996b). It occurs in the aquatic environment as chlorides (an anion of chlorine), free forms of chlorine or combined available chlorine (chloramines). Chloramines may be less toxic in water but are more persistent (Department of Water Affairs and Forestry 1996b). The toxicity of chlorine to fish is increased by a reduction in dissolved oxygen concentration and is also affected by organic carbon and ammonia. Chloride on the other hand is a common constituent in water, is highly soluble, and once in solution tends to accumulate. Small amounts of chlorides are required for normal cell functions in plant and animal life (Water Watch 2003). Concentrations of chloride in freshwater typically ranges from a few to several hundred mg/ℓ (Department of Water Affairs and Forestry 1996a). Recorded concentrations for all four surveys were less than the detection concentration (limit) of 10 mg Cl/ℓ (Table 2.1). The detection limit is within the TWQR for domestic use but higher than the TWQR for aquatic ecosystems and the chronic and acute effect value for aquatic ecosystems. Therefore it is unknown whether the TWQR and/or chronic and/or acute effect values for aquatic ecosystems is exceeded.

Fluoride, a halogen (the most electronegative), readily forms complexes with many metals. In the elemental form fluorine dissolves in water to form hydrofluoric acid. Fluoride reacts readily with various ions to form insoluble complexes (calcium and phosphates) or complexes not easily absorbed by aquatic biota (magnesium and aluminium). A small amount of fluoride is necessary for the hardening of dental enamel and to increase the resistance to attack by bacterial acids (Department of Water Affairs and Forestry 1996a). Typically the concentration of fluoride in unpolluted surface water is approximately 0.1 mg/ℓ (Department of Water Affairs and Forestry 1996a). Fluoride concentrations

Water quality in this study were fairly constant and slightly higher than the typical concentration but well within the TWQR for both domestic use and aquatic ecosystems (Table 2.1).

Water hardness was originally described as the soap destroying power of water caused by the presence of calcium and magnesium salts. The current definition of total hardness as defined by Department of Water Affairs and Forestry (1996a) is the sum of calcium and magnesium concentrations expressed as mg/ℓ of calcium carbonate (CaCO3). The natural hardness of water is influenced by the geology of the catchment and the presence of soluble calcium and magnesium minerals. Other metals that occasional contribute to water hardness include strontium, iron, aluminium, zinc and manganese. In general, pollutants tend to be more toxic in softer water

(Mason 1991). The hardness of surface water rarely exceeds 100 mg CaCO3/ℓ and should be limited to between 50 and 100 mg CaCO3/ℓ. Calcium carbonate concentrations in the Vaal Dam varied from

50.5 to 132.52 mg CaCO3/ℓ (Table 2.1). The water can therefore be described as moderately soft in autumn and summer to slightly hard in winter. The winter concentration is above the recommended hardness of surface water. The reason for this could be the elevated calcium concentration recorded in winter (see discussion on calcium).

Potassium is an alkali metal which reacts violently with water to form potassium ions. It always occurs in water in association with anions such as chloride, but also occurs with sulphate, bicarbonate or nitrate (Department of Water Affairs and Forestry 1996a). Potassium is an essential dietary element and is the main intracellular cation in living organisms. Typically the concentration of potassium in freshwater is within the range of 2 to 5 mg/ℓ. The concentrations in this study at the Vaal Dam were within this range and therefore within the TWQR for domestic use (Table 2.1). There is no potassium TWQR available for aquatic ecosystems.

Magnesium is an alkaline earth metal, common in water, and reacts with oxygen and water to form magnesium oxide and magnesium hydroxide respectively. The solubility of magnesium in water is governed by the carbonate/bicarbonate equilibrium and therefore the pH. Magnesium, along with calcium is responsible for the hardness of water. Magnesium is an essential nutritional element and is a basic essential element for plants (Department of Water Affairs and Forestry 1996a). In freshwater, the typical concentration of magnesium ranges from 4 to 10 mg/ℓ. In this study recorded concentrations were within this range and therefore within the TWQR for domestic use (Table 2.1). There is no magnesium TWQR available for aquatic ecosystems.

Sodium, like potassium, is an alkali metal which reacts with water to form highly soluble sodium ions. It usually occurs as sodium chloride (table salt), but sometimes occurs as sodium, sulphate, bicarbonate or nitrate. Sodium is an essential dietary element and is present in all food in varying degrees. The levels of sodium are usually low in high rainfall areas and high in arid areas with low rainfall (Department of Water Affairs and Forestry 1996a). The Vaal Dam region is a fairly high rainfall area (see textbox on page 7) and therefore it is expected that sodium levels will be low. This is the

Water quality case in this study (Table 2.1). Sodium concentrations are significantly below the TWQR of 0 to 100 mg/ℓ for domestic use. There is no sodium TWQR available for aquatic ecosystems.

Ammonia can occur either in an ionised or un-ionised state. Toxicity is primarily attributed to the un- ionised ammonia form (NH3) as opposed to the ionised ammonium form (NH4) (The Krib 2002). However, ammonia can readily take up an additional hydrogen ion to form ammonium (Department of

Water Affairs and Forestry 1996b). Tests for ammonia usually measure total ammonia (NH3 + NH4) (The Krib 2002). The most significant factors that affect the proportion and toxicity of ammonia in aquatic ecosystems are water temperature and pH. The South African water quality guidelines for aquatic ecosystems (Department of Water Affairs and Forestry 1996b) provide a table to determine the percentage contribution of NH3 to total ammonia. Using the pH values and corresponding temperature values in this study (Figure 2.4), the contribution (as a percentage) of NH3 to total ammonia was extrapolated and the results given in Table 2.3. There were no temperature values recorded in autumn and therefore the contribution of NH3 could not be calculated.

Table 2.3: Table depicting the contribution of un-ionised ammonia (NH3) (as a percentage) to total ammonia recorded in the Vaal Dam during the four surveys

Survey Approximate NH3 Approximate NH3 contribution (%) concentration (mg N/ℓ) Autumn (April 2000) * * Winter (June 2000) 1.2 <0.0006 Spring (October 2000) 3.8 0.0019 Summer (January 2001) 15 <0.0075 Note: Values were extrapolated from a table relating temperature and pH to the un-ionised ammonia contribution published by Department of Water Affairs and Forestry (1996b). * No extrapolation or calculation can be made as no temperature values were recorded for this survey.

The aquatic ecosystems TWQR for un-ionised ammonia is less than or equal to 0.007 mg N/ℓ and the chronic effect value (CEV) is 0.015 mg N/ℓ. As can be seen in Table 2.3, in spring and winter, the calculated concentration of NH3 was below the TWQR and the chronic effect values. In summer the calculated concentration is slightly higher than the TWQR but below the chronic effect value. In addition, ammonia is more toxic under alkaline (higher pH values) than neutral conditions but has a very low toxicity under acidic conditions (lower pH values) (Department of Water Affairs and Forestry 1996a). In this study, pH values are fairly neutral (minimum pH of 7.34) to slightly alkaline (maximum pH of 8.3) (Section 2.3.1).

Nitrate (NO3) is the end product and nitrite (NO2) is the inorganic intermediate of the oxidation of ammonia and organic nitrogen. Nitrate is more stable than nitrite and is more abundant in aquatic ecosystems (Department of Water Affairs and Forestry 1996b). A significant source of nitrates in natural waters is as a result of the oxidation of vegetable and animal debris and animal and human excrement. Concentrations of nitrate in water are typically less than 22 mg/ℓ nitrate (Department of Water Affairs and Forestry 1996b). In this study, nitrate concentrations were well below the typical concentration and within the TWQR for domestic use (Table 2.1). No TWQR for aquatic ecosystems

Water quality is available. In terms of inorganic nitrogen it has stimulatory effects on aquatic plant growth and algae (Department of Water Affairs and Forestry 1996a). In South Africa, inorganic nitrogen in unimpacted aerobic surface water is usually below 0.5 mg N/ℓ but may increase to between 5 and 10 mg N/ℓ in highly enriched waters. In this study, spring and summer concentrations were below 0.5 mg N/ℓ which in terms of nitrogen concentration is indicative of oligotrophic conditions. These systems are usually low in productivity with rapid nutrient recycling resulting in no nuisance growth of aquatic plants or blue-green algal blooms. In autumn and winter, concentrations were between 0.5 and 0.6 mg N/ℓ respectively which in terms of nitrogen concentrations is indicative of mesotrophic conditions. These systems are often productive with nuisance growth of aquatic plants and algal blooms; however the algal blooms are seldom toxic (Department of Water Affairs and Forestry 1996b).

Phosphorous is a key element necessary for growth of plants and animals (Water Watch 2003). Phosphorous can occur in numerous organic and inorganic forms but elemental phosphorous does not occur in the natural environment (Department of Water Affairs and Forestry 1996b). Phosphates -2 (formed from this element) and more specifically orthophosphates (H2PO4 and HPO4 ) (produced by natural processes and found in sewage) are the only forms of soluble inorganic phosphorous directly used by aquatic biota (Water Watch 2003). Phosphorous is an essential macronutrient, which has a major role in building nucleic acids and in the storage and use of energy in cells. Phosphorous is also thought to be the main nutrient controlling the degree of eutrophication in aquatic ecosystems (Department of Water Affairs and Forestry 1996b). Phosphate will stimulate growth of plankton and aquatic plants. Excessive growth leads to eutrophication (Water Watch 2003). In autumn, the concentration was considerably high at 0.12 mg/ℓ (Table 2.1). In terms of phosphate concentrations, the Vaal Dam in autumn can therefore be described as exhibiting eutrophic conditions (highly productive system, with nuisance growth of aquatic plants and blooms of blue-green algae often including species toxic to man, livestock and wildlife). Phosphate concentrations for winter, spring and summer were less than the detection limit of 0.05 mg/ℓ (Table 2.1). It is therefore possible that the type of conditions exhibited during those surveys could range from eutrophic to oligotrophic (see nitrate discussion above). However turbidity plays an essential role in algal blooms due to its influence on light penetration. In this study turbidity values were extremely high (Section 2.4.1).

Sulphate is a common constituent of water arising from the dissolution of mineral sulphates in soil and rock. It forms salts with various cations such as potassium, sodium, calcium, magnesium, barium, lead and ammonium. Sulphates are discharged from acid mine wastes and many other industrial processes (Department of Water Affairs and Forestry 1996a). Sulphates when added to water tend to accumulate to progressively increasing concentrations. Atmospheric sulphur dioxide from the combustion of fossil fuels can give rise to sulphuric acid in rainwater (acid rain) resulting in the return of sulphate to surface waters. The typical concentration of sulphate in surface water is 5 mg/ℓ. In this study, sulphate concentrations were considerably higher, two to four times higher, than the typical concentration but still within the TWQR of 0 to 200 mg/ℓ for domestic use (Table 2.1). The elevated concentrations above the typical concentration could be as a result of the combustion of fossil fuels

Water quality

(there are several roads running along the banks of the Vaal dam). There is no TWQR available for aquatic ecosystems.

Silicon is not discussed in any of Department of Water Affairs and Forestry’s South African water quality guidelines. Since it is unlikely to affect the result of this study it is not discussed further.

Total dissolved solids, as mentioned previously under electrical conductivity (Section 2.4.2) serve as a general indicator of change in water quality (Department of Water Affairs and Forestry 2002). Total dissolved solids (TDS) is a measure of the amount of various inorganic salts dissolved in water. Virtually all natural waters contain varying concentrations of TDS as a result of the dissolution of minerals in rocks, soils and decomposing plant material. TDS is normally dependent on the characteristics of the geological formations the water was or is in contact with. According to Walmsley et al. (1998) the Vaal River has major problems with total dissolved solids (TDS). However this problem is not seen in the current study at the Vaal Dam. For those surveys where TDS concentrations were recorded, the values were well within the TWQR for domestic use (Table 2.1).

Due to the fact that only two readings were recorded for TDS concentrations in the Vaal Dam during this study, the following information has been included. TDS concentrations are directly proportional to the electrical conductivity (EC) of water (Wetzel 1975). Conductivity is often used as an estimate for TDS. According to Department of Water Affairs and Forestry (1996a), EC is usually a factor (ranging between 5.5 and 7.5 for most natural waters) lower than the TDS concentration. The exact value of the conversion factor is dependent on the ionic composition of water especially pH and bicarbonate concentration. The average conversion factor commonly used and given by the Department of Water Affairs and Forestry (1996a and 1996b) for most South African inland waters is 6.5. When applying this average conversion factor to EC values recorded in this study, the resulting approximation for TDS concentrations still fall well within the TWQR for domestic use of 0 to 450 mg/ℓ and correlate to a certain extent with the measured TDS values (Table 2.4). It is important to note that the rate of change of TDS concentration and the duration of change is more important than absolute changes in TDS concentrations. Therefore annual, seasonal mean or mean values should be used when comparing concentrations with the criteria for TDS (Department of Water Affairs and Forestry 1996a).

Table 2.4: Estimate of total dissolved solids concentration (in mg/ℓ) using electrical conductivity values recorded for this study in the Vaal Dam Survey Measured TDS Measured EC Calculated TDS concentration concentration value using average conversion factor (mg/ℓ) (mS/m) of 6.5 (mg/ℓ) Autumn (April 2000) * 17 110.5 Winter (June 2000) 130 16 104 Spring (October 2000) * 17 110.5 Summer (January 2001) 145 23 149.5 Note: TDS concentrations were calculated using the average conversion factor of 6.5 given by Department of Water Affairs and Forestry (1996a and 1996b) as follows: TDS (mg/ℓ) = EC (mS/m at 25 oC) x 6.5 * No reading available.

Water quality

2.4.3. TRACE METAL ANALYSIS The transformation of compounds released into a water source may render them more toxic. The toxicity of a chemical is directly related to its bioavailabilty and the chemical and physical form in which an organism encounters the compound is determined by its bioavailabilty (Wade 1999; personal communication3). Metals enter into a number of reactions with environmental components resulting in complexation, precipitation and sorption. These reactions affect their mobility and bioavailabilty and therefore their toxicity. However, sediments can act as a sink causing concentrations to be higher in sediments than water (Roebuck 1992; Gouws and Coetzee 1997).

This section discusses the metal concentrations presented in Section 2.3.3 in terms of the South African target water quality ranges (TWQR) as defined by the Department of Water Affairs and Forestry (1996a, 1996b and 1996c).

Aluminium is the third most abundant metal (8.1%) in the earth’s crust and is described as a non- critical element (Department of Water Affairs and Forestry 1996b). However there is growing concern over the effects of elevated aluminium concentrations in the environment specifically as a result of acid mine drainage and acid precipitation (Department of Water Affairs and Forestry 1996b). Under acidic (pH<6.0) or alkaline (pH>8.0) conditions elevated concentrations of aluminium may be mobilised to the aquatic environment (Department of Water Affairs and Forestry 199b). However at neutral pH values (between 6 and 8) aluminium is insoluble and occurs as a hydrated Al(III) cation. In this study pH values ranged from 7.34 to 8.3 (Section 2.3.1). The aluminium TWQR for aquatic systems and domestic use is less than, or equal to, 0.005 mg/ℓ and 0 to 0.15 mg/ℓ respectively. The limit for detectable concentrations in water for analyses in this study is 0.1 mg Al/ℓ. For the winter survey (June 2000) aluminium concentrations were below detection and therefore within the TWQR for domestic use (Table 2.2). For the remaining surveys (autumn and summer) aluminium concentrations were above the TWQR for both aquatic ecosystems and domestic use (Table 2.2). The main effects of aluminium concentrations above the TWQR in domestic water are aesthetic, relating to discolouration in the presence of iron and manganese (Department of Water Affairs and Forestry 1996a). In aquatic ecosystems, elevated concentrations of bio-available aluminium in water are toxic to a wide variety of organisms (Department of Water Affairs and Forestry 1996b). The toxic effects are however dependant on the species and life stage of the organism, the concentration of calcium in the water and the pH (Department of Water Affairs and Forestry 1996b). Calcium concentrations for this study were fairly low (within the TWQR for domestic use) except in winter (Section 2.3.2). The pH values in this study were fairly neutral (pH of 7.34) to slightly alkaline (pH of 8.3) (Section 2.3.1). Increased concentrations may be the result of effluent from metal construction. Aluminium is however also used as a flocculant in water treatment processes which may result in increased concentrations of aluminium in the final water.

Cadmium is a metal element which is highly toxic to marine and freshwater aquatic life. It is found

3 P Wade 1999, personal communication. CSIR, Environmentek, Tel: (012) 841-2911.

Water quality naturally in the earth’s crust in association with zinc, lead and copper sulphide ore bodies. Large quantities enter the global environment annually as a result of natural weathering processes (Department of Water Affairs and Forestry 1996b). Trace concentrations of cadmium are found in fresh waters mostly as a result of industrial activity (Department of Water Affairs and Forestry 1996b). The presence of cadmium in the aquatic environment and in drinking water is of concern because it bioaccumulates (Department of Water Affairs and Forestry 1996a) and is highly toxic to all higher organisms (Department of Water Affairs and Forestry 2002; Kempster et al. 1980). Cadmium has a low solubility under neutral or alkaline pH conditions and is highly soluble under acidic conditions. As mentioned above, pH values in this study are fairly neutral to alkaline (7.34 to 8.26). The TWQR value for cadmium is less than, or equal to, 0.00015 mg/ℓ for aquatic ecosystems and 0 to 0.005 mg/ℓ for domestic use. Cadmium concentrations were below the detectable concentration limit of less than 0.05 mg/ℓ for all surveys (Table 2.2). This detection limit is higher than both TWQR and therefore it is not known whether the TWQR for aquatic ecosystems and/or domestic use is exceeded.

Chromium is a relatively scarce metal and therefore its occurrence and concentration in aquatic ecosystems is usually very low (Department of Water Affairs and Forestry 1996b). Chromium occurs in various forms. The most important of these forms is chromium (VI) salts which are highly soluble at all pH values (Department of Water Affairs and Forestry 1996b). Chromium (II) and (III) forms are less toxic and therefore less hazardous and are not usually found at near-neutral pH (Department of Water Affairs and Forestry 1996a) and therefore are not expected to occur in the Vaal Dam during this study. The equilibrium between chromium (VI) and (III) is strongly influenced by redox potential and pH and the toxicity of both forms is affected by pH and water hardness. As water hardness and pH increase acute toxicity decreases (Department of Water Affairs and Forestry 1996b). The pH values remained fairly constant during all four surveys (Figure 2.4). Water hardness was fairly constant except in winter where an increase in hardness was observed (Table 2.1). This increase in water hardness correlates with a slightly higher chromium concentration recorded in winter. During three of the four surveys, chromium concentrations were either equal to the detection concentration of 0.05 mg/ℓ (winter) or less (autumn and summer) and therefore within the TWQR for domestic use (Table 2.2). No values were recorded in spring. The TWQR for aquatic ecosystems is lower than the detection concentration and therefore it is not known whether the guideline in terms of aquatic ecosystems is exceeded.

Copper is an essential trace element to plants, animals and humans (Department of Water Affairs and Forestry 1996a). It is a common metallic element in rocks and minerals of the earth’s crust and is commonly found as an impurity in mineral ores (Department of Water Affairs and Forestry 1996b). A significant source of copper in domestic water arises from the dissolution of copper from plumbing systems in areas with soft or low pH waters. Natural sources of copper in the aquatic environment occur due to weathering processes or dissolution of copper minerals and native copper. The copper TWQR for domestic use is between 0 and 1 mg/ℓ. The detectable limit for copper concentrations in water in this study is less than 0.1 mg/ℓ. Copper concentrations were below the detectable limit for all surveys and therefore fall within the TWQR for domestic use (Table 2.2). For aquatic ecosystems the

Water quality

TWQR value is less than 0.0003 mg/ℓ. Even though copper concentrations recorded during this study were below the detection limit, it is not known whether concentrations exceed the TWQR for aquatic ecosystems as the detection limit is above the TWQR. Copper in the aquatic environment may result from sewage treatment plant effluents; runoff and groundwater contamination in the use of copper in treating soils; liquid effluents and atmospheric fallout from industrial sources such as mining, smelting and refining industries, coal-burning and iron- and steel- producing industries (Department of Water Affairs and Forestry 1996b). Even at low concentrations, copper in water is toxic (Department of Water Affairs and Forestry 1996b). The toxicity of copper is however dependent on local water quality conditions such as water hardness, dissolved oxygen, pH, the presence chelating agents (amino acids, humic acids, suspended solids), when present in combination with other metals and in the presence of sulphate, calcium and magnesium (Department of Water Affairs and Forestry 1996b). The status of dissolved oxygen and pH in the Vaal Dam during this study is outlined in Sections 2.4.1. The presence of sulphate, calcium, magnesium and the status of water hardness are discussed in Section 2.4.2. The main constituents which exhibited higher concentrations than normal included dissolved oxygen and calcium.

Lead is principally released into the aquatic environment through the weathering of sulphide ores (Department of Water Affairs and Forestry 1996b). Lead also has many industrial applications; it is used in batteries, in domestic water distribution pipes, in paints and as an organic lead compound which could lead to sources of lead pollution in water supplies (Department of Water Affairs and Forestry 1996a). The lead TWQR for aquatic ecosystems and domestic use is less than, or equal to, 0.0002 mg/ℓ and 0 to 0.01 mg/ℓ respectively. The detectable concentration for lead in this study is less than 0.1 mg/ℓ. Although during all surveys concentrations were below the detectable limit (Table 2.2), it is not known whether the TWQR for aquatic ecosystems and/or domestic use is exceeded as the detection limit is above both these values.

Iron is the fourth most abundant element in the earth’s crust. Natural waters contain varying amounts of iron. It is a trace element required by both plants and animals as a vital oxygen transport mechanism in blood (Water Watch 2003). The toxicity of iron depends on whether it is in the ferrous or ferric state and in suspension or solution. However, iron has a limited toxicity and bio-availability and is therefore classified as a non-critical element. Typically, in unpolluted surface water dissolved iron concentrations range from 0.001 to 0.5 mg/ℓ (Department of Water Affairs and Forestry 1996a). During all surveys in which iron concentrations were recorded, the concentrations were within this range but above the TWQR for domestic use of 0 to 0.1 mg/ℓ (Table 2.2). In winter and summer, concentrations were only slightly above the TQWR at 0.11 and 0.14 mg/ℓ respectively while in autumn concentrations were considerably higher than the TWQR at 1.1 mg/ℓ. The effects of the elevated levels of iron are very slight aesthetic effects but no health effects (Department of Water Affairs and Forestry 1996a). The concentration of dissolved iron is dependent on pH, redox potential, turbidity, suspended matter, aluminium concentration and the occurrence of several heavy metals, notably manganese.

Water quality

Manganese is the eighth most abundant metal in nature and occurs in a number of ores. It is an essential micronutrient for plants and animals however in high concentrations it is toxic (Department of Water Affairs and Forestry 1996b). In aquatic ecosystems, manganese occurs in various salts and minerals and not as a metal. Natural sources of manganese include soils, sediments and metamorphic and sedimentary rocks; non-natural sources include industrial discharges and acid mine drainage. The concentration of dissolved manganese is influenced by changes in redox potential, dissolved oxygen, pH, turbidity, organic matter and concentration of aluminum (Department of Water Affairs and Forestry 1996a; Department of Water Affairs and Forestry 1996b). The average typical concentration of manganese in freshwater is 0.008 mg/ℓ with a range of 0.00002 to 0.13 mg/ℓ (Department of Water Affairs and Forestry 1996a). During all surveys in which manganese was measured, concentrations were below detection (Table 2.2). As a result, manganese concentrations fall within the typical concentration range and below the TWQR for aquatic ecosystems. However, for domestic use the TWQR is higher than the detection limit. Therefore it is not known whether the TWQR for domestic use is exceeded.

Nickel is not discussed in any of Department of Water Affairs and Forestry’s South African water quality guidelines and therefore no TWQR exist for nickel. Nickel concentrations for all surveys where nickel was measured were less than the detection concentration of 0.1 mg/ℓ (Table 2.2). It is unlikely to affect the results of this study and is not discussed further.

Zinc is a metallic element which is essential for the nutrition of plants and animals. Humans have a high tolerance for zinc while fish are highly susceptible to poisoning (Department of Water Affairs and Forestry 1996a). Zinc (relatively low in toxicity) strongly interacts with cadmium (highly toxic) and copper resulting in the following effects: zinc reduces the toxicity of cadmium and copper increases the toxicity of zinc in soft waters (Department of Water Affairs and Forestry 1996b). In addition, the pH of the water determines the concentration of soluble zinc (Department of Water Affairs and Forestry 1996a; Department of Water Affairs and Forestry 1996b). More acidic waters (low pH) give rise to relatively high concentrations of zinc in solution. In this study, pH values were fairly neutral to slightly alkaline (Section 2.4.1). Dissolved oxygen concentrations also play a role in zinc toxicity; lower oxygen concentrations increase zinc toxicity. Dissolved oxygen in this study was considerably higher than the typical saturation levels (Section 2.4.1). Zinc can enter the aquatic environment through natural processes such as weathering and erosion and through industrial activity (Department of Water Affairs and Forestry 1996b). Zinc may affect the taste of water making it bitter (Department of Water Affairs and Forestry 2002). The concentration of zinc in water is typically as low as 0.015 mg/ℓ. The zinc TWQR for aquatic ecosystems and domestic use are less than, or equal to, 0.002 mg/ℓ and 0 to 3 mg/ℓ respectively. In this study, zinc concentrations were below the detection limit of 0.1 mg/ℓ and therefore within the TWQR for domestic use (Table 2.2). As the detection concentration is greater than the TWQR for aquatic ecosystems, it is unknown whether zinc concentrations exceed the guideline.

Water quality

2.5. SUMMARY AND CONCLUSION Summary With regards to surface water variables such as pH, temperature and electrical conductivity, values recorded during this study were all within the recommended and available Department of Water Affairs and Forestry guideline values. The pH of the water varied from fairly neutral in winter to slightly alkaline in summer. Temperature values ranged from 19.9oC in winter to 25oC in summer. Electrical conductivity values were fairly constant with a slight increase in summer. Electrical conductivity serves as a general indicator of change in water quality (Department of Water Affairs and Forestry 2002).

On the other hand dissolved oxygen and turbidity values were noticeable higher when compared to Department of Water Affairs and Forestry’s guideline values. Dissolved oxygen values were almost double the typical concentration expected for inland (Highveld) waters and turbidity values were between 50 and 100 times higher than the guideline values for domestic use. Dissolved oxygen is an essential element needed for the survival and functioning of aquatic biota but in high concentrations can cause gas bubble disease in fish (Department of Water Affairs and Forestry 1996b). Dissolved oxygen is one of the best indicators of the health of a water system (Cleveland 1998). The cause of the elevated dissolved oxygen concentrations is not known. The extremely high turbidity values could be as a result of the turbid nature of the rivers feeding into the Vaal Dam as mentioned by Braune and Rogers (1987), clay mineralogy of the dam (Birch 1983), disturbance of soil particles, discharge of sewage and/or other wastes into the water system at the time of sampling. The highest dissolved oxygen and turbidity values were recorded in winter and autumn respectively. The lowest values for both variables were recorded in summer. Turbidity values decreased seasonally from autumn to winter to spring to summer.

In terms of water hardness (calcium carbonate), water in the Vaal Dam was moderately soft in autumn and summer and slightly hard in winter. The increase in water hardness in winter can be attributed to the elevated calcium concentrations recorded in winter. Although the increased calcium concentration in winter has no health effects it could have an effect on the toxicity of metals (Department of Water Affairs and Forestry 1996a). With respect to the following macro determinants: fluoride, chloride, potassium, magnesium, sodium, silicon, sulphate and total dissolved solids (TDS), it is unlikely that the low concentrations recorded will affect the quality of water in the Vaal Dam and this study.

In terms of inorganic nitrogen (linked to nitrates), the conditions in autumn and winter due to slightly elevated concentrations of nitrogen can be described as mesotrophic (often productive with nuisance growth of aquatic plants and algal blooms however algal blooms are seldom toxic). In terms of phosphate concentrations, conditions in the Vaal Dam in autumn can be described as eutrophic (highly productive system, with nuisance growth of aquatic plants and blooms of blue-green algae often including species toxic to man, livestock and wildlife).

Water quality

Trace metal concentrations are relatively low for all measured metals except iron, chromium and aluminium. Iron concentrations, although they exceeded the domestic use guideline values were still within the typical concentrations found in unpolluted surface waters and therefore is not thought to be a problem in this study. With regards to chromium, only one survey (winter) definitely exceeds the guideline value for aquatic ecosystems. It is possible that the slightly harder water in winter is responsible for this elevated chromium concentration. Aluminium, on the other hand exceeds both the domestic use and aquatic ecosystem guideline values in autumn and summer. These elevated levels of aluminium could result from the use of aluminium in water treatment processes although it is not known whether this type of water treatment is used in the vicinity of the Vaal Dam.

For many of the metals, such as cadmium, copper, manganese, nickel, lead and zinc, the concentrations measured were less than the detectable concentration limit for the particular metal analysis in this study and in some cases this limit was greater than the Department of Water Affairs and Forestry’s domestic and aquatic ecosystem guideline values. Therefore it is difficult to say whether these metals are present in concentrations greater or less than the guideline values. It is however not expected that these trace metals will affect the quality of water in the Vaal Dam in this study due to the low concentration of industrial and mining activities in the Vaal Dam catchment. In addition, sediments have the ability to act as a sink for trace metals causing concentrations to be higher in sediments than in water (Roebuck 1992; Gouws and Coetzee 1997). This was observed in a recent study conducted at the Vaal Dam where metal concentrations were relatively low in the water and higher in sediments (Crafford 2000).

Conclusion In general the best quality waters along the Vaal River are found in the Vaal Dam catchment (Braune and Rogers 1987). Braune and Rogers (1987) mentioned that rivers of the Vaal Dam catchment are characterized by turbid, bicarbonate water with low total dissolved solids (TDS). These variables could influence the quality of water in the Vaal Dam. In this study, high turbidity values and low TDS concentrations were recorded in the Vaal Dam. In addition, noticeably high levels of dissolved oxygen were also recorded.

Eutrophication, already a problem in the Vaal River (Braune and Rogers 1987), is in general as a result of high levels of phosphates in urban effluent (O’Keefe et al. 1992). The signs of eutrophication can be seen in this study in the slightly elevated nitrate concentrations recorded in autumn and winter and elevated phosphate concentrations recorded in winter indicating a mesotrophic to eutrophic condition in the Vaal Dam during autumn and winter. However due to the high turbidity in the Vaal dam, the algal blooms one would expect from the high eutrophication (due to phosphates and nitrates) was not observed.

Organic contaminants (micro-pollutants) from both industrial and agricultural sources have been detected in water, plants and fish but not in high concentrations (Braune and Rogers 1987). This

Water quality information correlates with the results obtained in this study indicating that there has been little change in water quality in the Vaal Dam over thirteen years. Low concentrations of all trace metals except aluminium and iron were recorded in water samples from the Vaal Dam. In 1997, Gouws and Coetzee conducted a study on heavy metals in sediments in the Vaal Dam. It was concluded that the extractable metal content of the sediments was low which is indicative of a relatively unpolluted system.

The seasonal variations observed in water quality in the Vaal Dam will be used to determine correlations, if any, between water quality, parasite infestation (discussed in the following chapter) and the fish health assessment index (discussed in Chapter 4).

Asian tapeworm

3. ASPECTS OF THE ECOLOGY OF THE ASIAN TAPEWORM, BOTHRIOCEPHALUS ACHEILOGNATHI YAMAGUTI, 1934 FOUND IN YELLOWFISH IN THE VAAL DAM CHAPTER 3 ASPECTS OF THE ECOLOGY OF THE ASIAN TAPEWORM, BOTHRIOCEPHALUS ACHEILOGNATHI YAMAGUTI, 1934 FOUND IN YELLOWFISH IN THE VAAL DAM

3.1. INTRODUCTION This chapter focuses on the occurrence of the genus Bothriocephalus Rudolphi, 1808 (Cestoda: Pseudophyllidea) and more specifically Bothriocephalus acheilognathi Yamaguti, 1934 found in Labeobarbus aeneus (previously named Barbus aeneus) Burchell, 1822 (smallmouth yellowfish) and Labeobarbus kimberleyensis (previously named Barbus kimberleyensis) Gilchrist & Thompson, 1913 (largemouth yellowfish) collected during four seasonal surveys at the Vaal Dam. B. acheilognathi, the Asian tapeworm belongs to the Order Pseudophyllidea Carus, 1863 and the family Bothriocephalidae Blanchard, 1849 (Khalil and Polling 1997).

The Asian tapeworm is a key parasite of carps and other cyprinids (Tindall 1989). It was originally a parasite of the Chinese grass carp (Ctenopharyngodon idella Valenciennes, 1844) and the silver carp (Hypothalmichthys molitrix Valenciennes, 1844) in the southern parts of China (Boomker, Huchzermeyer and Naudé 1980). A brisk trade in grass carp for food, sport and weed control has resulted in the rapid spread of this tapeworm to other countries by means of infected fish (Pool 1987) and has adapted itself successfully to the common carp (Cyprinus carpio Linnaeus, 1758) (Boomker et al. 1980). According to Scholz (1999) the spreading of parasites such as B. acheilognathi has been facilitated through insufficient veterinary control during the import of fish. The Chinese grass carp and the silver carp are currently distributed virtually throughout the world (Mashego 1982). B. acheilognathi is presently found in most countries in Europe, Asia, United States of America, New Zealand and South Africa (Boomker et al. 1980).

In South Africa, bothriocephalid parasites have been found in various dams and freshwater systems in Gauteng (formerly the Transvaal) (Hänert 1984). Brandt, Van As, Schoonbee and Hamilton-Atwell (1981) established that B. acheilognathi was introduced into South Africa in 1975 together with fingerlings of the grass carp from West Germany. The parasite was first recorded in South Africa (as B. gowkongensis Yeh, 1955) from Lb. kimberleyensis in 1978 (Brandt et al. 1981) and has subsequently been recorded from C. carpio, Lb. mattozzi Guimaraes, 1884, Lb. trimaculatus Peters, 1952 (Brandt et al. 1981; Van As, Schoonbee and Brandt 1981) and Oreochromis mossambicus Peters, 1852 (Paperna 1996). Localities in South Africa where B. acheilognathi has been found include the Komatipoort area (Eastern Transvaal) (Boomker et al. 1980), Marble Hall (Lowveld) (Brandt et al. 1981), Boskop Dam (Mooi River) (Van As et al. 1981), Hartebeespoort Dam, Piet Gouws Dam (Lebowa) (Mashego 1982), Olifants River (Lebowa) (Mashego 1982), Glen Alpine Dam (Lebowa)

Asian tapeworm

(Mashego 1982) and the Vaal Dam (Mashego 1982). The tapeworm’s presence in most of the localities mentioned above can be attributed to the supply of common carp fry to commercial farmers (Boomker et al. 1980). However, according to Mashego (1982) its presence in the Vaal Dam cannot be accounted for. Brandt et al. (1981) mentioned the possibility that this tapeworm was imported into South Africa with the common carp (C. carpio) as early as 1859 or with the Dinkelsbühl Aischgrund variety of the common carp in 1952.

B. acheilognathi was originally described as three different species, described from wild fish in Japan (as B. acheilognathi Yamaguti, 1934 and B. opsariichthydis Yamaguti, 1934) and from grass carp (C. idella) from South China (as B. gowkongensis Yeh, 1955) (Paperna 1996). These three species were later recognised as synonyms (Körting 1975; Molnár 1977) with B. acheilognathi taking priority. Pool (1988) believes that the three species, B. acheilognathi, B. kivuensis and B. aegyptiacus are synonymous, with B. acheilognathi having priority. Various authorities (as cited by Pool and Chubb 1985 and Pool 1988) note that B. opsariichthydis Yamaguti, 1934, B. fluviatilis Yamaguti, 1952, B. gowkongensis Yeh, 1955, B. phoxini Molnár, 1968 and Schyzocotyle fluviatilis Akhmerov, 1960 are synonyms of B. acheilognathi Yamaguti, 1934. Pool and Chubb in 1985 concluded that there is only one Bothriocephalus species parasitizing cyprinid fish and the continued use of the name B. acheilognathi should be made. This is the approach that has been adopted in this study.

For the purposes of this study, cestodes were identified as either B. acheilognathi or ‘other cestode spp.’. The reason for this, as mentioned previously, is that the occurrence of B. acheilognathi in South Africa is as a result of introduced fish species from Asia (originally) of which this tapeworm is a parasite and the occurrence of the large numbers of this particular cestode could affect the accuracy of the Health Assessment Index values obtained. Although helminth parasites are rarely implicated as causes of disease (Tindall 1989), B. acheilognathi is probably the most pathogenic of the Bothriocephalus genus (Boomker et al. 1980). Heavy infections in fish can cause the abdomen to swell and the intestinal lumen to become blocked by the parasite. Such fish become sluggish, emaciated and stop feeding. The fish develop haemorrhagic enteritis with destruction of the intestinal epithelium (Tindall 1989). The Health Assessment Index in relation to the presence of B. acheilognathi and its affect on fish health is discussed in the following chapters.

The purpose of this chapter is to confirm the identification of the tapeworms collected during the four surveys and provide a brief description of the parasite’s infection, seasonality, gender specificity and species (host) specificity. A comparison between the parasite’s infection in Lb. aeneus and its infestation in Lb. kimberleyensis is also included.

3.2. MATERIALS AND METHODS

3.2.1. THE STUDY LOCATION Four surveys were conducted in the Vaal Dam at RAU Island (Groot Eiland) (see Chapter 2 for more detail on the study site), one per season namely in April 2000 (early autumn), June 2000 (winter), October 2000 (late spring) and January 2001 (summer). The fish species collected during the four

Asian tapeworm surveys were Lb. aeneus (smallmouth yellowfish) and Lb. kimberleyensis (largemouth yellowfish). Fish were identified based on the size of the mouth (Skelton 1993) (see Chapter 4 for more detail).

3.2.2. COLLECTION OF FISH During each survey 20 largemouth yellowfish and 20 smallmouth yellowfish were collected, killed and dissected in a field laboratory (see Chapter 4 for more detail on the process followed).

3.2.3. EXTRACTION AND COLLECTION OF CESTODES The extraction and collection of cestodes was done in the field. Once the fish have been dissected, the intestines were removed and placed in petri dishes with saline solution for examination. The following method as described by Khalil (1991) was used for processing the Platyhelminth parasites found during the surveys. Parasites were collected as soon as possible after the death of the fish to prevent any deterioration. The intestines were pulled open carefully using two Dumont No. 7 sharp tweezers to ensure that the cestodes were kept intact. Intestines were not cut open as this could result in the inaccurate identification and counting of the specimens. Each cestode was carefully and slowly dislodged from the intestine wall ensuring that the cestode was collected intact (the scolex, mature and gravid proglottids). Specimens were transferred to a clean sampling bottle containing saline solution using brushes or plastic pipettes. Some specimens were placed on ice or in a refrigerator and left overnight to relax before being transferred to 70% ethanol for preservation. Other specimens (in sampling bottle containing saline solution) were shaken vigorously for a few minutes to dislodge debris and induce muscle fatigue, which deters strong contraction and relaxes the scolex. While swirling the sampling bottle, an equal amount (equal to the amount of saline solution already present in the sampling bottle) of hot Alcohol-formaldehyde-acetic acid (AFA) solution (50 mℓ 95% ethyl alcohol, 6 mℓ 40% formalin, 4 mℓ glacial acetic acid and 40 mℓ distilled water) was added to kill and fix the specimens. Khalil (1991) described AFA as a good general fixative with rapid penetrative action and therefore it was used in this study. Specimens were transferred to 70% alcohol where they were stored for later staining and identification.

3.2.4. IDENTIFICATION OF CESTODES The following procedures as outlined below were followed in the laboratory.

Staining: An initial staining test phase was conducted using Borax Carmine (Grenacher) stain (Pantin, 1964). The recipe for the stain and the staining procedure followed is given in the text box below. Specimens were mounted with a cover slip using commercially prepared Entellan (Merck Cat No. 1.07961.0500) as a mounting agent. This test phase showed that specimens that were stored on ice or in a refrigerator overnight during the field work could not be identified using the internal organs as a result of deterioration of these organs. The field work was modified (specimens were shaken vigorously to induce muscle fatigue instead of being placed on ice or in the fridge) to prevent further specimens from deteriorating.

Asian tapeworm

Following the test phase, specimens were stained for photography purposes. It is important to note that due to the size of the cestodes, specimens were stained individually.

BORAX CARMINE (GRENACHER) STAIN (Pantin, 1964) Staining solution: Concentrated solution of borax carmine (add carmine powder to 4% borax solution and boil for ½ hour). Dilute solution with equal volume of 70% alcohol. Allow to stand and then filter. Staining procedure (short way): Transfer specimen from 70% alcohol to the borax carmine. Stain until the specimen is thoroughly penetrated (10 minutes). Thoroughly differentiate specimen in acid alcohol (4 drops strong hydrochloric acid (HCl) to 100mℓ of 70% alcohol) until specimen assumes bright transparent appearance. Dehydrate in 90% alcohol (10 minutes). Note: Times may vary Pass through two sets of absolute alcohol (5 to 30 minutes each). depending on the size Clear in xylene and mount. of specimens

Identification of cestodes: Following the test phase of the staining process, it was evident that identification of the genus and species could take place using a light microscope as the scolex of many of the cestodes collected was characteristic of the species. Following this, microscopic analysis and micrographs were compared to available literature for diagnosis of the specimens.

Statistical analyses: All specimens were counted and the data included in the calculations of the Health Assessment Index (see Chapter 4) and used for statistical analyses of the cestode parasites. Statistical analyses were conducted by Rand Afrikaans University Statistical Consultation Services. Prevalence, abundance and mean intensity of B. acheilognathi was calculated per season for each fish species. Infection statistics were calculated by making use of the following definitions set by Margolis, Esch, Holmes, Kurtis and Schad (1982) and Bush, Lafferty, Lotz and Shostak (1997): ∗ prevalence (as a percentage) = number of infested individuals of a host species ÷ by the total number of hosts; ∗ mean intensity = total number of a particular parasite species ÷ by the number of infested hosts; ∗ abundance (relative density) = total number of a particular parasite species ÷ by the total number of hosts in a sample.

Secondly, data was analysed to determine the seasonality and species and gender specificity of B. acheilognathi. This included determining whether the presence or absence of B. acheilognathi is dependent on fish gender and species (using Pearson Chi-square test) and determining whether there are significant differences between fish gender (using T-tests) and seasons (using ANOVA and then Scheffe or Dunnet T3). The infection in the two fish species was compared to determine whether significant differences exist (using T-test). In addition, the infection of B. acheilognathi (intensity) in each fish species was compared to the size (fork length) of the fish sampled.

Photography: Digital micrographs of stained specimens were taken using a Zeiss Axioplan 2 Imaging microscope. The images were processed with AxioVision 3.1 software.

Asian tapeworm

3.3. RESULTS

3.3.1. IDENTIFICATION OF CESTODES According to Mashego (1982), the classification of the bothriocephalid worms is based primarily on the shape of the scolex. Pool concluded from his study on B. acheilognathi in 1984 that the identification of adults should be based on the heart-shaped scolex and prominent square apical disc. Identification of cestodes in this study was therefore based on these characteristics of this particular species. As outlined in Section 3.1, species of the genus Bothriocephalus which have a heart-shaped scolex have been synonymised with B. acheilognathi and the name B. acheilognathi takes priority. Tapeworms were therefore identified as either B. acheilognathi (S Mashego 2000, personal communication4) or not. The reason for distinguishing between B. acheilognathi and other cestodes, as outlined in the introduction to this chapter (Section 3.1), is that the occurrence of B. acheilognathi in South Africa is as a result of an introduced fish species from Asia of which B. acheilognathi is a cestode parasite and the occurrence of the large numbers of this particular cestode could affect the accuracy of the Health Assessment Index values obtained (Chapter 4).

Specimens found in the current study were compared to sketches provided by various authorities as well as the diagnosis of B. acheilognathi (as B. gowkongensis) by Yeh (1955), as cited in Paperna (1996). When reviewing available sketches it is the author’s opinion that the scolex of specimens from the current study compared fairly well with many of the Bothriocephalus species that have a heart- shaped scolex. This is visible when comparing Figure 3.1 C and D. The diagnosis as given in Paperna (1996) and corresponding micrographs of specimens collected in this study [in brackets] are detailed below. Where possible, sketches of specimens from available literature have been included in Figure 3.1 and Figure 3.2 for comparison. ∗ Eggs are operculated and premature when laid [Figure 3.1A]. ∗ Worms are variable in size and number in segments [Figure 3.1B]. ∗ Scolex is heart-shaped, laterally flat usually with a distinct terminal disk and deep lateral grooves (bothria) [Figure 3.1C and E]. ∗ Mature segments are broader than long. Gravid segments are longer than broad. ∗ Each segment contains 50 to 90 testes. ∗ The cirrus sac is round and genital atrium is situated in the median line of the dorsal surface of the segment [Figure 3.2A]. ∗ The ovary is comprised of two lateral lobes connected by an isthmus. ∗ Vitellaria, approximately 200, are scattered laterally [Figure 3.2C].

According to the above mentioned diagnosis, mature and gravid segments vary in breadth and length; however in this study this was not the case. Most segments were broader than longer (Figure 3.2 A and B). When labelling the reproductive organs of collected specimens (Figure 3.2 B), the author disagrees with the labels proposed by Yamaguti (1934) and seen in Figure 3.2 D3 and D4. Instead, the author believes (after reviewing available literature) that the labelling of the reproductive organs should be as indicated in Figure 3.2 B and used by Mashego (1982) (Figure 3.2 D5).

4 S Mashego 2000, personal communication. Vice Rector, University of the North West. Tel: 083 627 1181.

Asian tapeworm

A B SPECIMEN 2

SPECIMEN 1

EGG D

C BOTHRIA

4 3

Figure 3.1: Micrographs of Bothriocephalus acheilognathi collected during the four surveys at the Vaal Dam and sketches taken from available literature. a) Operculated egg b) Variable size of specimens collected during this study c) & e) Heart-shaped scolex and presence of bothria in collected specimen d) Sketches of scolices taken from available literature 1) B. kivuensis from Mashego (1982) 2) B. opsariichthydis from Yamaguti (1934) 3) & 4) B. gowkongensis from Yeh (1955)

Asian tapeworm

TESTES E A 1 GENITAL ATRIUM

CIRRUS SAC

UTERUS

OVARY

3 2

B COILS OF VAS DEFERENS

CIRRUS SAC UTERINE OVARY COILS 4

TESTES TESTES

vdcs uc

C 5

t ov vit VITELLARIA Abbreviations in sketches: cs - cirrus sac; ut - uterus; ov - ovary; vit - vitellaria; t – testes; uc - uterine coils; vd - vas deferens

Figure 3.2: Micrographs of Bothriocephalus acheilognathi collected during the four surveys at the Vaal Dam and sketches taken from available literature. a) Mature proglottid with reproductive organs b) Young adult proglottid c) Vitellaria scattered d) Sketches of specimens taken from available literature. 1) & 2) B. gowkongensis from Molnár & Murai (1973) 3) B. acheilognathi from Yamaguti (1934) 4) B. opsariichthydis from Yamaguti (1934) 5) B. kivuensis from Mashego (1982)

Asian tapeworm

3.3.2. PARASITE NUMBERS Tapeworms in this study were grouped as either B. acheilognathi or ‘other cestode spp.’. The number of B. acheilognathi and other cestode spp. collected during the four surveys is tabulated in Table 3.1. Of the 160 fish sampled, only 19 out of 80 (23.8%) Lb. aeneus harboured B. acheilognathi while none (0%) harboured other cestode spp. and 68 out of 80 (85%) Lb. kimberleyensis harboured B. acheilognathi while only six (6) (7.5%) harboured other cestode spp.

Table 3.1: Table summarizing number of cestodes collected from Labeobarbus aeneus and Labeobarbus kimberleyensis at the Vaal Dam during the four surveys Labeobarbus aeneus (n = 80) Labeobarbus kimberleyensis (n = 80) Survey B. acheilognathi Other cestode B. acheilognathi Other cestode spp. spp. Autumn (April 2000) 4 0 3417 4 Winter (June2000) 298 0 1651 0 Spring (October 2000) 256 0 1120 1 Summer (January 2001) 24 0 2040 3 Total 582 0 8216 8 Total cestodes 582 8228

3.3.3. INFECTION STATISTICS OF Bothriocephalus acheilognathi The percentage of hosts (prevalence) infected with B. acheilognathi, the abundance (relative density) of B. acheilognathi and the intensity (mean intensity) of the Asian tapeworm infection in both Lb. aeneus and Lb. kimberleyensis is illustrated graphically in Figure 3.3, Figure 3.5 and Figure 3.4 respectively. A statistical comparison (T-test) of the two fish species in terms of B. acheilognathi prevalence, abundance and mean intensity is also included to determine whether or not there are significant differences between fish species.

Prevalence: The prevalence of B. acheilognathi in Lb. kimberleyensis is fairly constant over all seasons ranging from 80% to 90% whereas the prevalence in Lb. aeneus is fairly constant (10 to 15%) but much lower throughout autumn, spring and summer with a considerable increase in prevalence (55%) in winter (Figure 3.3). When comparing fish species, the prevalence of B. acheilognathi in Lb. kimberleyensis is considerably higher. Statistical analyses indicate that, there is a significant difference (T-test, p = 0.001) between the prevalence of B. acheilognathi in the two fish species sampled.

100 90 80 Labeobarbus aeneus 70 60 Labeobarbus 50 kimberleyensis 40 Labeobarbus aeneus 30 trendline Prevalence (%) 20 Labeobarbus 10 kimberleyensis trendline 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Survey Figure 3.3: Graph depicting the prevalence of Bothriocephalus acheilognathi in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

Asian tapeworm

Abundance (relative density): During all surveys, the abundance of the Asian tapeworm was considerably higher in Lb. kimberleyensis (Figure 3.4). In this fish species, abundance values ranged from 55.0 (spring) to 170.9 (autumn) and in Lb. aeneus, values ranged from 0.2 (autumn) to 14.9 (winter). The following seasonal trend was observed for largemouth yellowfish: values decreased from autumn to winter then again in spring followed by an increase in summer. For smallmouth yellowfish the opposite trend was observed. Values increased from autumn to winter followed by a decrease in spring and a further decrease in summer. Statistical analyses indicate that, there is a significant difference (T-test, p = 0.011) between the relative densities of B. acheilognathi in the two fish species sampled.

180 160 140 Labeobarbus aeneus 120 Labeobarbus 100 kimberleyensis 80 Labeobarbus aeneus

Abundance 60 trendline 40 Labeobarbus 20 kimberleyensis trendline 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Survey Figure 3.4: Graph depicting the abundance (relative density) of Bothriocephalus acheilognathi in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

Mean intensity: Except in spring, the infection was considerably more intense in Lb. kimberleyensis (Figure 3.5). The highest value recorded for Lb. kimberleyensis was 213.6 in autumn and the lowest value was 68.8 recorded in spring. For the remaining surveys mean intensity decreased to 102.2 and 85.8 in summer and winter respectively. As for Lb. kimberleyensis, mean intensities in Lb. aeneus differed considerably from 2.00 in autumn to 78.7 in spring. During winter and summer a mean intensity was recorded of 27.1 and 4.3 respectively. Statistical analyses indicate that, there could be a significant difference between the mean intensities of B. acheilognathi in the two fish species (T-test, p value was slightly above 0.05 at 0.053) however either due to the sample size being too small or the variance being too big this cannot be said for certain.

250

200 Labeobarbus aeneus Labeobarbus 150 kimberleyensis 100 Labeobarbus aeneus trendline Mean Intensity Mean 50 Labeobarbus kimberleyensis trendline 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Survey Figure 3.5: Graph depicting the mean intensity of Bothriocephalus acheilognathi in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

Asian tapeworm

3.3.4. ECOLOGICAL PARAMETERS Gender specificity (data pooled according to fish gender): In both fish species, there were no significant differences (T test, p values were greater than 0.05) in the average number of B. acheilognathi found in males and females even though the average number in males and females differed. In Lb. aeneus the average number of B. acheilognathi in male and female fish was 12.85 and 2.70 respectively. In Lb. kimberleyensis the average number was 91.43 and 107.42 respectively.

In addition, the presence or absence of B. acheilognathi is not dependent on the gender of fish species (Pearson Chi-square test; p values for both fish species were greater than 0.05). A similar number of male (7) and female (12) Lb. aeneus and male (37) and female (31) Lb. kimberleyensis were found to have B. acheilognathi. In Lb. aeneus the prevalence of B. acheilognathi in males and females was 0.21 and 0.26 respectively. Similarly in Lb. kimberleyensis the prevalence of B. acheilognathi was 0.84 in males and 0.86 in females.

Seasonality (data pooled according to seasons): In Lb. aeneus, the highest number of B. acheilognathi were observed during the winter survey while in Lb. kimberleyensis the highest number was observed during the spring survey (Table 3.1). From the statistical analyses (Pearson Chi-square test) performed on the data, the presence/absence of B. acheilognathi in Lb. aeneus is dependent on the season (p = 0.002) with the highest number of infected fish (11) caught in winter and lowest (2) in autumn. This is not the case for Lb. kimberleyensis. Similar numbers of Lb. kimberleyensis (between 16 and 18) were infected with B. acheilognathi during all four surveys.

However, when conducting the ANOVA test to determine if there are significant differences between seasons, the results showed that there are significant (p = 0.003) seasonal differences based on the number of B. acheilognathi found in Lb. kimberleyensis. Post Hoc tests were then undertaken to distinguish which seasons differed significantly. The results of the statistical analyses (Dunnet T3 test) showed that autumn and spring differed significantly with a p value of 0.011. No significant seasonal differences were observed in the number of B. acheilognathi found in Lb. aeneus.

Species specificity (data pooled according to fish species): When comparing fish species (Table 3.1), Lb. kimberleyensis harboured a higher number of B. acheilognathi than Lb. aeneus. The number of B. acheilognathi found in Lb. kimberleyensis totalled 8216 while Lb. aeneus only had a total of 582 B. acheilognathi in the same number of hosts (80). In addition, a higher number of Lb. kimberleyensis (68) were infected with B. acheilognathi than Lb. aeneus (19). From the statistical analyses (Pearson Chi-square) performed, when data is pooled according to fish species, the presence/absence of B. acheilognathi is highly dependent on fish species with p values of 0.000.

Size specificity: As can be seen in Figure 3.6 below, no correlations were observed between the sizes (fork lengths) of yellowfish (both smallmouth and largemouth) sampled and the number of Asian tapeworms recorded.

Asian tapeworm

A B 450 450 p-value = -0.089 p-value = 0.174 400 400 350 350

300 300

250 250

200 200

150 150 Bothriocephalus acheilognathi Bothriocephalus acheilognathi Bothriocephalus 100 100

50 50 Number of of Number 0 Number of 0 300 320 340 360 380 400 420 440 460 280 320 360 400 440 480 520 560 Length of fish (mm) Length of fish (mm)

Labeobarbus aeneus Labeobarbus kimberleyensis Figure 3.6: Graphs depicting no correlation between the size (fork length) of Labeobarbus aeneus (A) and Labeobarbus kimberleyensis (B) sampled during the four surveys and the number of Bothriocephalus acheilognathi. The statistically calculated p-value is included.

3.4. DISCUSSION This study focused on the Asian tapeworm, B. acheilognathi introduced into South Africa via imported infected fish (Cyprinidae). Exceptionally high numbers of this parasite were found during this study. On the other hand, low numbers (an insignificant amount) of other cestode spp. were found in the yellowfish sampled. For the purposes of this study the following four questions relating to B. acheilognathi are answered in the discussion. 1. What is the infection of the tapeworm in yellowfish collected during this study and are there noticeable seasonal differences in the infection in the two yellowfish species? 2. Is the tapeworm specific to the fish species it infects in terms of the two fish species collected? 3. Is the tapeworm specific to the gender of the fish species it infects? 4. Is there a correlation between the size of the fish infected and the intensity of the infection?

As highlighted by Khan and Thulin (1991), parasites are a natural part of the aquatic community and their distribution and abundance are potentially either directly or indirectly affected by a number of biotic and abiotic factors.

Infection The infection of B. acheilognathi (in terms of prevalence, abundance and mean intensity) in Lb. kimberleyensis (largemouth yellowfish) was greater than that observed in Lb. aeneus (smallmouth yellowfish). Noticeably higher numbers of this tapeworm were observed in largemouth yellowfish and a noticeably higher number of this fish species was infected. Similar prevalence and mean intensity values to those recorded in this study were recorded for Lb. aeneus and Lb. kimberleyensis in a separate study conducted by Nickanor, Reynecke, Avenant-Oldewage and Mashego (2002) in the Vaal Dam during 2001.

Asian tapeworm

The infection statistics indicate that both prevalence and abundance of B. acheilognathi differ significantly (statistically), between the fish species with Lb. kimberleyensis exhibiting higher values. The mean intensity of B. acheilognathi is higher in Lb. kimberleyensis for most seasons, except in spring when the opposite is true. In spring a small number of Lb. aeneus were infected with a relatively high number of tapeworms thereby increasing the mean intensity. It is possible that the mean intensities of the fish species differ significantly but because either the sample size was too small or the variance was too big to distinguish between fish species this has not been proven in this study.

It would be expected that the high numbers (and subsequently high prevalence, abundance and mean intensity) of B. acheilognathi found in Lb. kimberleyensis in this study are linked to the life cycle of the tapeworm as the transmission of parasite to host is via an intermediate host eaten by the fish (Paperna 1996). Körting (1975) mentioned that the intermediate host of the Asian tapeworm for carp is a crustacean. He also mentioned that a number of crustaceans can act as intermediate hosts. When looking at the food preference of largemouth yellowfish, it initially feeds on insects and crustaceans and once the yellowfish reaches a fork length of more than 300 mm it feeds on other fish (Skelton 2001). Although the majority of infected largemouth yellowfish collected during the surveys varied in fork length between 360 and 420 mm it is possible that infected crustaceans are eaten occasionally resulting in an infection in Lb. kimberleyensis. However it is not expected that the infection would be as heavy as that seen in this study due to the fish’s preference for larger food particles. When looking at the infection of B. acheilognathi’ (in terms of prevalence, mean intensity and abundance) in relation to fish size (fork length), no correlations were recorded in Lb. kimberleyensis (Figure 3.6). It is important to note that Lb. kimberleyensis was the first recorded host of B. acheilognathi (as B. gowkongensis) in South Africa in 1978 (Brandt et al. 1981). The type of copepod acting as an intermediate host for B. acheilognathi in this fish species should be determined. This opportunistic tapeworm has already adapted successfully to the common carp, Cyprinus carpio (Körting 1974; Boomker et al. 1980) in South African waters. This carp species feeds on a range of plant and animal matter (Skelton 2001) and more specifically carp fry, which tend to be more heavily infected with the tapeworm, feed on zooplankton (Boomker et al. 1980).

Other authorities such as Marcogliese and Esch (1989) and Williams and Jones (1994), mention that metacestodes use planktonic or benthic copepods as intermediate hosts. If this is the case then Lb. aeneus would be the preferred host as this fish species is broadly omnivorous with zooplankton, benthic invertebrates, vegetation, algae and detritus forming the major food of the species (Dörgeloh 1985; Skelton 2001). Although Lb. aeneus is not the preferred host (Nickanor et al. 2002 and this study), it was still infected (although low) with B. acheilognathi. Poulin (1998) mentions that parasites that enter their host through ingestion, such as the Asian tapeworm, cannot prevent non-host species from eating the infected intermediate hosts. This method of transmission (ingestion) allows more host species to be infected. Another reason both fish species are infected with this tapeworm could be that the intermediate copepod host can vary considerably. Various genera of copepods have been found

Asian tapeworm to be compatible intermediate hosts (Williams and Jones 1994; Paperna 1996) and in this case it may be a larger crustacean such as a crab which may act as paratenic hosts and would explain the enigma behind the higher infection observed in largemouth yellowfish.

In a study conducted by De Leon, Garcia-Prieto, Leon-Regagnon and Choudhury (2000) in Mexico it was found that helminth communities (in general) were generally more abundant in carnivorous fish species than herbivores and detritivores. This matter needs to be researched further in order to gain a better understanding of the Asian tapeworm’s infection.

Seasonal trends In Lb. aeneus the prevalence of B. acheilognathi was fairly constant except in winter when a noticeably higher numbers of smallmouth yellowfish were infected; prevalence values in winter were approximately four to five times higher than in the remaining seasons (Figure 3.3). This could be due to a change in feeding regime; however as will be discussed in the following chapter Lb. aeneus fed well during all four surveys. Körting (1974) indicated that early spring, when plankton grows, is likely to be a significant season in terms of seasonal incidence and infective period. However, this is not seen in Lb. aeneus in this study. In spring the prevalence was fairly low. The reason could be that more infected food (copepods) was available in that particular winter than is normally the case. When reviewing temperature data obtained for a separate study in the Vaal Dam (Crafford 2000), values recorded during the winter of this study were higher than the year before (1999). In winter visibility (light penetration) in the Vaal Dam was at its maximum although for the remaining three surveys similar (slightly lower) values were exhibited. It could be assumed that the increased prevalence in winter is an exception and caused by an external unknown factor. Statistical analyses indicate that, as a result of this high prevalence in winter, the presence of B. acheilognathi in Lb. aeneus is dependent on the season. In Lb. kimberleyensis, no trend was observed for prevalence of the Asian tapeworm throughout the four seasons (Figure 3.3); a similar (very high) number of fish were infected throughout the year.

Abundance values, in Lb. aeneus, exhibited a pattern similar to the prevalence values observed in the same fish species (Figure 3.4). Values increased considerably from autumn to winter followed by a decline in spring and a further decline in summer. It is assumed that this seasonal trend is related to the breeding and subsequent feeding patterns of Lb. aeneus. Feeding habits of the host account for a large percentage of the variation in the total number of parasites per host species (Williams and Jones 1994). The number of parasites in a host would depend on how much the host eats and whether the food is infected. In winter, the fish eat enough food to sustain them through the breeding season which lasts from spring through to late summer (Skelton 2001) explaining the decrease in abundance recorded in spring and summer. As the breeding season ends fish start eating again resulting in the increase in abundance from autumn to winter. In Lb. kimberleyensis, the seasonal trend in abundance could not be attributed to the fish’s breeding patterns. Largemouth yellowfish breed in mid to late summer (Skelton 2001) which would mean that abundance values should be higher in winter and

Asian tapeworm spring however the opposite was observed (Figure 3.4). A sharp decrease in abundance values was recorded from autumn to winter after which abundance values remained relatively constant decreasing slightly in spring but increasing again in summer. This seasonal trend could be related to changes in the amount of food and subsequently infected food available. In autumn the high abundance relative to the remaining three seasons was due to the considerably higher number of B. acheilognathi found in sampled Lb. kimberleyensis.

A similar pattern to that recorded for the abundance of B. acheilognathi in Lb. kimberleyensis was recorded for the mean intensity (Figure 3.5). This is due to the fact that most of the fish sampled during the four surveys were infected with B. acheilognathi. The mean intensity in Lb. aeneus also exhibited a similar pattern to that of the abundance in Lb. aeneus except the mean intensity peaked in spring rather than winter (Figure 3.5). Mean intensities in the smallmouth yellowfish increased considerably from autumn to winter followed by a significant increase (three times that of winter) in spring after which the intensity decreased in summer to an intensity similar to that recorded in autumn. The mean intensity recorded in Lb. aeneus was even higher than that recorded in Lb. kimberleyensis. Opposite seasonal trends were recorded for the mean intensity of the two fish species.

Statistical analysis indicates that the presence of B. acheilognathi in smallmouth yellowfish is dependent on the season with the highest number of fish infected in winter. However no statistically meaningful differences were observed in the number of B. acheilognathi found during each season. The opposite is true for largemouth yellowfish sampled in this study. Statistical analysis indicates that the presence of B. acheilognathi is not dependent on the season, even though there were significant statistical differences between the number of B. acheilognathi recorded in autumn and that recorded in spring.

Fish gender and species specificity Statistical analysis indicates that the presence of B. acheilognathi is highly dependent on the species of fish with Lb. kimberleyensis, as mentioned earlier, having the highest infection between the two yellowfish species. Poulin (1998) noted that high host specificity could be an artefact of inadequate sampling however in this study 20 fish per species per season were collected cancelling out this possibility. In addition, various studies conducted in the Vaal and Olifants River systems, in which cestode endoparasites have been incorporated, have indicated either an absence of cestodes or low cestode infections in a range of fish species namely Clarias gariepinus (Marx 1996; Crafford 2000; Watson 2001), Labeobarbus marequensis (Watson 2001), Oreochromis mossambicus (Watson 2001), Labeo capensis and Labeo umbratus (Groenewald 2000).

On the other hand, when pooling data according to the sex of the fish, the tapeworms exhibited no preference for male or female fish. A similar number of male and female fish were infected in Lb. kimberleyensis. Even though there were noticeably lower numbers of infected female smallmouth yellowfish when compared to males there was no dependency (statistically) on fish gender.

Asian tapeworm

Size specificity In both yellowish species sampled no correlations were observed between fish size and Asian tapeworm infection (Figure 3.6). The statistical p-values were closer to 0 than 1.

3.5. SUMMARY AND CONCLUSION Summary In this study the majority of the tapeworms were identified as B. acheilognathi based on the heart- shaped scolex and presence of bothria. This was achieved after comparing collected specimens with available literature and sketches. B. acheilognathi in this study was species (host) specific with a considerably higher infection (in terms of prevalence, abundance and mean intensity) recorded in Lb. kimberleyensis. The reason for this has still not been determined. The Asian tapeworm in this study was not fish gender specific.

Seasonal trends were observed for prevalence, abundance and mean intensity of the tapeworm in Lb. aeneus although statistical analyses indicate that no significant differences exist between seasons. Seasonal trends in smallmouth yellowfish were attributed to breeding and subsequent feeding patterns of the fish.

In Lb. kimberleyensis, no seasonal trend was recorded for prevalence. The abundance and mean intensities of B. acheilognathi however varied seasonally with the highest value recorded in autumn and the lowest value recorded in spring. Statistical analyses indicate that there is a significant difference between the presence of B. acheilognathi in these two seasons. The reason for the seasonal trends observed in largemouth yellowfish could not be explained. Further research on largemouth yellowfish concentrating on factors such as post-spawning migrations of the host, schooling behaviour, age of host, reproductive behaviour, host feeding behaviour, host hormone levels/state of maturity, immunological response of host, availability of infected intermediate hosts as food, site of infection, negative interaction between parasites (as outlined by Williams and Jones 1994) should be conducted to provide more explanations on the seasonal variations observed in B. acheilognathi’s infection in this fish species.

The outcome of this section of this study will now be integrated into the health assessment index (HAI) performed on the fish sampled and the effect of this high infection in Lb. kimberleyensis on the HAI values obtained will be discussed in the following chapters.

Health Assessment Index

4. THE HEALTH ASSESSMENT INDEX CHAPTER 4 THE HEALTH ASSESSMENT INDEX

Biological systems are dynamic entities, ever changing and shifting to maintain the equilibrium once thought stagnant - Anonymous

4.1. INTRODUCTION This chapter focuses on the health of the fish sampled. Although there are many different ways of assessing fish health, this study concentrated on the bio-monitoring tool known as the Fish Health Assessment Index (HAI) and associated (inverted) Parasite Index (PI). Bio-monitoring or biological monitoring came about based on the assumption that the measurement of the condition or health of biota could be used to assess the health of an ecosystem (Herricks and Cairns 1982). Aquatic biomonitoring makes use of aquatic organisms in order to reflect the quality of the associated water (Roux 1993). As mentioned by Roux, Van Vliet and Van Veelen (1993), biological communities are generally sensitive to low levels of disturbances in a wide variety of environmental factors therefore acting as good indicators of water quality and general ecosystem health. Adams, Brown and Goede (1993) added that in aquatic ecosystems fish, that control a high position in the food chain, are good representative indicators of overall system health.

The Fish HAI, which uses fish as indicator organisms, was developed to complement the SASS 2 (South African Scoring System, Version 2) (Avenant-Oldewage, Oldewage and Van Vuren 1995). In 1990 Goede and Barton developed a field necropsy autopsy-based system. The system was developed to meet the needs for a rapid, inexpensive and easily used method for biologists to detect changes in the health of fish populations, early enough to take corrective action (Goede and Barton 1990). The system was designed in association with the Fish Condition Profile (HCP) which involved a series of simple, ordered observations and measurements of external characteristics, simple blood parameters and external organs of a sample of twenty fish (Novotny and Beeman 1990). This system was later refined by Adams et al. (1993), in the USA, who proposed a quantitative index that allowed for statistical comparisons of fish health among data sets such as sites, species and years. This index was originally developed for monitoring organic pollution. It provides a health profile of fish based on percentages and degrees of anomalies in tissues and organs of sampled fish as a result of environmental stressors. This refinement, made by Adams et al. (1993) led to the development of the Fish Health Assessment Index (HAI). Adams et al. (1993) found the HAI to be a “simple and inexpensive means of rapidly assessing general fish health in field situations”. In 1993, Avenant-Oldewage and Swanepoel suggested the use of fish health studies in South Africa. A user manual was developed by Avenant-Oldewage et al. (1995) in response to a request by the Institute of Water Quality Studies to test the HAI in South Africa. In 1996, a study conducted by Marx, involved testing of the HAI for the first time in South Africa in the Olifants River catchment in the Kruger National Park. To date, the HAI has been successfully applied on the Olifants River system (Avenant-

Health Assessment Index

Oldewage et al. 1995, Marx 1996, Robinson 1996, Luus-Powell 1997, Watson 2001) and Vaal River system (Crafford 2000) using various fish species as indicator organisms and variations of the Parasite Index (discussed below). In 2001, Watson made an addition to the South African user manual and included a colour chart which would limit the subjective nature of colour assessments made during the evaluation of the liver, bile colour and spleen.

The fish HAI, developed by Adams et al. (1993) made use of the following components: post mortem, blood and parasites. Parasites were seen as an indication of disease - an indication of bad fish condition - and therefore only their presence or absence was recorded in the health assessment. In 1994, Avenant-Oldewage suggested the use of fish endo- and ectoparasite community composition as an indicator of environmental health. Parasite data, as highlighted by Crafford (2000) can be used to a great effect in environmental management. The results of a study conducted by Marcogliese and Cone in 1997 supported the hypothesis that parasite communities are good indicators of environmental stress and biodiversity. Robinson, Hines, Sorensen and Bryan (1998) used the HAI to assess the effects of parasite infestation on the health of two endangered desert fish. During the study conducted by Marx (1996), the interrelationship between fish health and parasitism was investigated to determine whether parasites should be incorporated into the South African HAI or used as a separate entity in association with the HAI. The study went one step further than Adams et al. (1993) in terms of parasites and looked at two parasite variables, namely endoparasites and ectoparasites. The presence or absence of these variables as well as a refinement of their scores were assessed in the HAI by Marx (1996). She suggested that a low level parasitic survey distinguishing between ecto- and endoparasites should be used as a separate entity to accompany and enhance the results of the HAI. Crafford in 2000 assessed the use of four parasite indices, namely the original parasite index by Adams et al. (1993) (distinguished between the presence and absence of parasites), inserted parasite index by Marx (1996) (distinguished between the presence and absence of ecto- and endoparasites), refined parasite index by Marx (1996) (distinguished between the number of ecto- and endoparasites) and the inverted parasite index. The inverted parasite index is based on the premise that ectoparasites are more directly exposed to the effects of poor water quality than endoparasites (Crafford 2000) therefore a smaller number of ectoparasites and larger number of endoparasites would be found at a more polluted locality and vice versa at a less polluted locality. His study concluded that the inverted parasite index more accurately reflected the HAI values and therefore this parasite index was adopted in the current study. Watson (2001) stated that by using the separate evaluation of ecto- and endoparasites, the HAI could be seen as a biomonitoring technique combined with a Parasite Index component. Parasites could be used as possible bio-indicators of water quality rather than simple indicators of poor fish condition (Watson 2001).

In terms of indicator fish species, this study focused on two yellowfish species, namely Labeobarbus aeneus (smallmouth yellowfish) Burchell, 1822 and Labeobarbus kimberleyensis (largemouth yellowfish) Thompson & Gilchrist, 1913. Both these fish species are important South African angling

Health Assessment Index species (Skelton 1993, Skelton 2001). In addition, field trips to the Vaal Dam have shown that largemouth yellowfish seem to be most affected by endoparasites of the fish sampled.

The purpose of this chapter is to determine first the health of the fish sampled by using the HAI and inverted PI as well as the condition factor, second if there are seasonal variations in fish health and third whether there are differences between the fish species sampled. The reason for comparing fish species in this study, is that in the previous chapter it was concluded that the Asian tapeworm, Bothriocephalus acheilognathi, ‘prefers’ to infest Lb. kimberleyensis when given the ‘option’ between the two fish species sampled in this study with the infection (in terms of prevalence, abundance and mean intensity) in Lb. kimberleyensis being considerably higher. By comparing the health of the fish sampled it will allow the author to determine whether there are correlations, if any, with the high infection of the Asian tapeworm both seasonally and between species. Although HAI values for a locality should only be compared with different seasons at the same locality and with the same sampling species (Groenewald 2000), it is necessary to compare the HAI values for the two yellowfish species sampled in this study in order to achieve the objectives of this study.

4.2. MATERIAL AND METHODS

4.2.1. THE STUDY LOCATION Four surveys were conducted in the Vaal Dam at RAU Island (Groot Eiland) (see Chapter 1 for more detail on the study site), one per season namely in April 2000 (early autumn), June 2000 (winter), October 2000 (late spring) and January 2001 (summer). The fish species collected during the four surveys were Lb. aeneus (smallmouth yellowfish) and Lb. kimberleyensis (largemouth yellowfish). Fish were identified based on the size of the mouth; in Lb. aeneus the snout length is less than the orbit to preopercular groove while in Lb. kimberleyensis the snout length is equal to or less than the eye to the preopercular groove (Skelton 1993).

4.2.2. FISH HEALTH ASSESSMENT INDEX (HAI) AND ASSOCIATED PARASITE INDEX The Fish Health Assessment Index as described by Avenant-Oldewage et al. (1995) and printed (with the inclusion of additional variables) in Marx (1996) [based on the necropsy-based system developed by Goede and Barton (1990) and refined by Adams et al. (1993)] was employed with the addition of the colour chart developed by Watson (2001). The Inverted Parasite Index assessed by Crafford (2000) was utilized.

4.2.2.1. Field work A field laboratory was set up for each survey. During each survey 20 largemouth yellowfish and 20 smallmouth yellowfish were collected using gill nets. Each gill net had four sections with varying stretched mesh sizes of 90, 110 and 130 mm respectively. Nets were thrown out in the late afternoon and left overnight. Gill nets were collected in the early morning. As the fish were removed from the gill nets they were checked for mobile external parasites (body surface, fins, gill cavity and buccal cavity). When external parasites were found, the number of parasites were recorded and placed in

Health Assessment Index small sampling bottles containing dam water, the fish were marked using a tagging gun and plastic tags and the tag number recorded. The fish were then placed in a circulating boat live tank and once on shore transferred to a holding tank through which dam water was circulated using a hosepipe and pump. The fish were weighed (in grams) and measured (in millimeters). Both the fork and total length were recorded.

Fish were placed on a dissecting board where blood samples were drawn from the caudal aorta situated below the lateral line. Blood samples were collected using a “Vacutainer” system. This system includes a collar, needle and untreated “Vacutainer” blood collection tubes. Prior to each survey, the “Vacutainer” tubes were filled with 0.1mℓ of Heparin (an anticoagulant) using a needle and syringe. The tubes were inverted and swirled a few times to ensure they were well rinsed with Heparin. Previous studies have shown that use of the “Vacutainer” system for collecting blood decreases the chances of haemolysis (rupturing of red blood cells) occurring. It is important to remember not to remove the lid of the “Vacutainer” tube as this would defeat the purpose of using a vacuum based system to draw up the blood. Blood smears were made for each fish. This involved placing a drop of blood on a clean microslide and using a second microslide to make the blood smear. This procedure ensures even consistency of the blood on the slide and prevents blood parasite aggregation on the edges of the slide.

Each microslide was allowed to dry and then preserved by dipping it into methanol. Once the slides had air-dried they were stored in a safe place and taken back to the laboratory for further analysis. Whole blood samples were stored on ice prior to being centrifuged. First, the haematocrit was determined using a Heraeus-christ microhaematocrit centrifuge. Blood was drawn into a capillary tube, sealed at one end with citroseal clay and centrifuged for three minutes at 15 000 revolutions per minute. The haematocrit and leukocrit was determined by expressing the amount of red and white blood cells (in millimeters using a ruler) respectively as a percentage of the total measurement. Second, the remaining whole blood samples were centrifuged in a Heraeus-christ centrifuge for 10 minutes at 3 000 revolutions per minute. The plasma was removed using a plastic Pasteur pipette and placed in clean untreated “Vacutainer” tubes. These tubes were then placed and stored on ice for plasma protein analysis in the laboratory. Protein analysis should be carried out as soon as possible to prevent any protein denaturation. In the instance that protein analysis cannot be carried out immediately, the plasma should be frozen.

Directly after drawing blood, a slime smear was made and examined under a light microscope for external parasites. Slides were then discarded.

Fish were killed prior to dissection by severing the spinal cord behind the head. Fish were examined as outlined previously using the manual compiled by Avenant-Oldewage et al. (1995) and printed (with the inclusion of additional variables) in Marx (1996) with the addition of the colour chart developed by Watson (2001). The colour chart was used to limit the subjective nature of the colour assessments

Health Assessment Index made for the liver, bile and spleen. External organs and tissue (eyes, skin, fins, opercules and gills) were examined for damage and cysts. Any abnormalities were recorded. The gills were dissected out, placed in petri dishes containing dam water and examined for parasites and cysts with the aid of a dissection microscope. The parasites were recorded and placed in sampling bottles containing dam water. Parasites were then preserved using 70% ethanol. The fish were subsequently dissected by making an insertion from the anus towards the head and the internal organs (mesenteric fat, hindgut, kidney, liver, bile and spleen) examined. Any abnormalities were recorded. The sex of the fish was also noted. Data sheets on which data was recorded were adopted from Watson (2001). Parasites found in the visceral (body) cavity were recorded and placed in sampling bottles containing normal saline solution (8.5 mg NaCl in 1 000 mℓ distilled water).

The intestines were removed and placed in petri dishes with saline solution for further examination. The intestines were pulled open carefully using two Dumont No. 7 sharp tweezers. All endoparasites were recorded and placed in sampling bottles containing normal saline solution (8.5 mg NaCl in 1 000 mℓ distilled water). Nematodes found in the intestine were processed according to the methods suggested by Khalil (1991). Nematodes were washed in normal saline to remove mucus and fixed by placing the specimens in Glacial Acetic Acid (GAA). GAA causes the nematodes to die, uncoil and stretch. Specimens were transferred to 70% alcohol for storage. The extraction and collection of cestodes found in the intestine is described in detail in Chapter 3 and more specifically Section 3.2. Fish were then discarded.

4.2.2.2. Laboratory procedures Total blood plasma protein concentrations were determined using a Boehringer Mannheim Total Protein test kit (MPR 3 Total protein 124281) and Roche Total Protein test kit (Total protein 1553836). The Roche test kit was used for a single survey, namely the January 2001 survey, as the Boehringer Mannheim test kit was no longer being manufactured. The test procedure carried out was as described in the enclosed instruction pamphlets.

Blood slides were stained using a commercially prepared Giemsa solution (Saarchem Serial No. 2645020) manufactured by Saarchem (Pty) Ltd. Slides were placed in the solution for 10 minutes and rinsed well in distilled water after which they were left to air dry. Each slide was mounted with a cover slip using commercially prepared Entellan (Merck Cat No. 1.07961.0500) as a mounting agent. Blood counts were performed using a light microscope (100 x objective i.e. 1 000 x magnification if use is made of a 10 x eyepiece). Five fields were randomly chosen on each slide. The number of red and white blood cells was recorded for each field and the percentage of white blood cells was determined. In addition, each slide was scanned (20 x objective) for blood parasites and parasite numbers noted.

Health Assessment Index

4.2.2.3. Calculations Health Assessment Index (HAI) and associated Parasite Index The numerical values used to classify recorded CALCULATING THE HEALTH abnormalities were adopted as demonstrated in ASSESSMENT INDEX (Adams et al. 1993) the user manual developed by Avenant- 1. Assign characters/values to internal and external variables from the necropsy based system. Oldewage et al. (1995) and printed (with the 2. Substitute assigned characters/values with inclusion of additional variables) in Marx (1996). corresponding numerical values. Slight modifications discussed by Watson (2001) 3. Calculate the HAI for each fish by summing the numerical values for all the variables. were adhered to. 4. Calculate the HAI for a sample population by summing all individual HAI values and dividing by the total number of fish examined for that sample The Inverted Parasite Index assessed by Crafford (mean). (2000) was used when assigning numerical 5. Standard deviation (SD) for each sample was calculated as follows: values to the number of ecto- and endoparasites N 2 observed. The Parasite Index is based on the SD = Σ (Vi – X) i=1 premise that ectoparasites are more directly N - 1 exposed to the effects of poor water quality than Where: N = number of fish per site X = average index for each site endoparasites (Crafford 2000). The Inverted Vi = index value for fish i Parasite Index results in the larger numbers of 6. Coefficient of variation (CV) was calculated as ectoparasites (indicative of better water quality) follows: having a lower score thereby correlating with the CV % = 100 x SD/X HAI value. Conversely large numbers of endoparasites have a higher score. The numerical values from the Inverted Parasite Index were included in the HAI calculations.

A summary of the method [as developed by Adams et al. (1993) and printed in Marx (1996)] used in this study to calculate the HAI value is given in the text box above. Two HAI values were calculated for each fish, one was a straightforward calculation including all parasites found and the other excluded the number of Asian tapeworms (B. acheilognathi Yamaguti, 1934) found in each fish. The reason for this was to determine whether or not this particular cestode species, which has been introduced into South Africa from Asia (see Chapter 3 for more detail), has an effect on the results of the Health Assessment Index.

Condition Factor The Condition Factor (CF) of a fish is determined using the weight (in grams) and length (total length in millimeters) of the fish. The Condition Factor was calculated using the following formula described by Carlander (1969) and suggested by Klemm, Stober and Lazorchak (1992):

CF = weight x 105 / length3

The minimum, maximum and mean condition factor, mean length, mean weight, standard deviation (SD) and standard error (SE) were statistically determined for each fish species per season to compare condition factors between the various seasons and fish species.

Health Assessment Index

Statistical analyses Certain data collected during the health assessment of the fish as well as the health assessment index values calculated and the condition factors calculated were analysed statistically. The analyses were conducted by Rand Afrikaans University Statistical Consultation Services. First, the abundance, prevalence and mean intensity of ecto- and endoparasites was calculated per season for each fish species. Infestation statistics were calculated by making use of the following definitions set by Margolis et al. (1982): ∗ prevalence (as a percentage) = number of infested individuals of a host species ÷ by the total number of hosts; ∗ mean intensity = total number of a particular parasite species ÷ by the number of infested hosts; ∗ abundance (relative density) = total number of a particular parasite species ÷ by the total number of hosts in a sample.

Secondly, data was analysed to determine if there are any significant differences between various variables. These analyses (as summarized in Table 4.1) were carried out for each fish species individually and then a comparison between the two fish species was conducted. The variables used in the analyses were as follows: ∗ number of Bothriocephalus acheilognathi; ∗ weight (in grams); ∗ number of other cestode spp.; ∗ condition factor; ∗ number of ectoparasites; ∗ Health Assessment Index (HAI) including all ∗ number of endoparasites; parasites; and ∗ length of fish (total in millimeters); ∗ Health Assessment Index (HAI) excluding the ∗ length of fish (fork in millimeters); number of B. acheilognathi.

Table 4.1: STATISTICAL ANALYSES PERFORMED ON DATA COLLECTED What was asked Statistical test performed For Labeobarbus aeneus Differences between gender groups T-test Differences between seasons ANOVA Differences between the absence/presence of certain variables* Pearson Chi-Square test For Labeobarbus kimberleyensis Differences between gender groups T-test Differences between seasons ANOVA Differences between the absence/presence of certain variables* Pearson Chi-Square test Comparison of the two species In terms of absence/presence of certain variables* Pearson Chi-Square test In terms of abundance, mean intensity and prevalence T-tests

Accuracy of predicting correct fish species using HAI variables Logistic regression Note: * - Variables include B. acheilognathi, other cestode spp., ectoparasites, endoparasites.

Water quality variables were not included as only one locality (Vaal Dam) was investigated and the water quality was constant for both fish species. Water quality at the Vaal Dam during the four surveys is discussed in detail in Chapter 2.

Health Assessment Index

4.3. RESULTS In total 160 fish were collected from the Vaal Dam; 20 fish per species [Lb. aeneus (smallmouth yellowfish) and Lb. kimberleyensis (largemouth yellowfish)] per season (autumn, winter, spring, summer). For Lb. aeneus, the number of males and females were 33 (41.3%) and 47 (58.7%), respectively. For Lb. kimberleyensis, 44 (55%) males and 36 (45%) females were collected.

4.3.1. PARASITES 4.3.1.1. Parasites collected Parasites were collected from both Lb. aeneus and Lb. kimberleyensis during all four surveys. Diplozoidae specimens were collected from the gills. Argulus japonicus was collected from the fins and skins. No parasites were observed from the slime smears during the four surveys. Nematodes were collected from both the visceral cavity (larval nematodes) and intestine (adult). No blood parasites were observed during the surveys. Cestodes were collected from the intestines. A large numbers of immature cestodes were noted during the surveys. Unidentified trematode cysts were found on the skin, fins, gill filaments and gill arches. The number of ecto- and endoparasites recorded during each survey is given in Table 4.2 and Table 4.3 respectively.

Table 4.2: Table summarizing the number of ectoparasites collected from Labeobarbus aeneus and Labeobarbus kimberleyensis at the Vaal Dam during the four surveys Labeobarbus aeneus (n = 80) Survey Diplozoidae Argulus japonicus Total ectoparasites Autumn (April 2000) 27 6 33 Winter (June 2000) 50 5 55 Spring (October 2000) 23 0 23 Summer (January 2001) 18 0 18 Labeobarbus kimberleyensis (n = 80) Survey Diplozoidae Argulus japonicus Total ectoparasites Autumn (April 2000) 18 3 21 Winter (June 2000) 12 3 15 Spring (October 2000) 7 0 7 Summer (January 2001) 1 0 1

Table 4.3: Table summarizing the number of endoparasites collected from Labeobarbus aeneus and Labeobarbus kimberleyensis at the Vaal Dam during the four surveys Labeobarbus aeneus (n = 80) Survey B. acheilognathi Other cestode Nematodes Trematode Total spp. cysts endoparasites Autumn (April 2000) 4 0 2 55 62 Winter (June 2000) 298 0 0 50 352 Spring (October 2000) 256 0 6 40 312 Summer (January 2001) 24 0 0 5 30 Labeobarbus kimberleyensis (n = 80) Survey B. acheilognathi Other cestode Nematodes Trematode Total spp. cysts endoparasites Autumn (April 2000) 3417 4 4 0 3425 Winter (June 2000) 1651 0 2 260 1916 Spring (October 2000) 1120 1 5 10 1138 Summer (January 2001) 2040 3 0 0 2043

Health Assessment Index

4.3.1.2. Infestation statistics for ecto- and endoparasites Parasites were grouped as ectoparasites or endoparasites. Of the 160 fish sampled, 48 out of 80 smallmouth yellowfish (60%) harboured ectoparasites while 36 (45%) harboured endoparasites and 28 out of 80 largemouth yellowfish (35%) harboured ectoparasites while 71 (88.8%) harboured endoparasites. As can be seen from these numbers, a larger percentage of Lb. aeneus was infested with ectoparasites than endoparasites while the opposite was the case for Lb. kimberleyensis. This was also clear when viewing the graphs below.

The total numbers of ecto- and endoparasites were used to determine the percentage of hosts infected (prevalence), the intensity of the infection (mean intensity) and the relative density (abundance) of parasites per season for each fish species. Infestation statistics of ecto- and endoparasites are illustrated graphically in Figure 4.1 to Figure 4.6. A statistical comparison (T-test) of fish species (Lb. aeneus and Lb. kimberleyensis) in terms of prevalence, mean intensity and abundance is also included to determine whether or not there are significant differences.

Prevalence: With regards to the percentage of hosts infested (prevalence) with ectoparasites, seasonal trends were seen in both Lb. aeneus and Lb. kimberleyensis (Figure 4.1). Prevalence of ectoparasites in Lb. aeneus increased from autumn to winter followed by a decrease in spring and a further decrease in summer. Prevalence ranged from 40% (summer) to 85% (winter). An almost similar pattern was seen in the prevalence of ectoparasites in Lb. kimberleyensis, except instead of an increase in winter prevalence decreased gradually from autumn (55%) through to summer (5%). A considerable difference was seen between the minimum and maximum values calculated for both fish species. Although a greater percentage of Lb. aeneus were infested with ectoparasites than Lb. kimberleyensis in all four surveys, statistical analyses indicate that the difference between the two fish species was not statistically meaningful.

For endoparasites, Lb. aeneus and Lb. kimberleyensis exhibited completely different seasonal trends to those observed for ectoparasites as well as completely different seasonal trends when comparing fish species (Figure 4.2). The percentage of Lb. aeneus infected with endoparasites increased considerably from autumn (20%) to winter (70%), then decreased in spring (65%) and decreased steeply in summer (25%). The highest and lowest values calculated for this fish species differed considerably. For Lb. kimberleyensis, a slight (almost constant) seasonal pattern was seen with prevalence increasing from autumn (85%) to winter (95%) then decreasing in spring (85%) and increasing again in summer (90%). The maximum values for both fish species were recorded in winter. The minimum values for Lb. aeneus and Lb. kimberleyensis were recorded in autumn and autumn/spring respectively. The prevalence of endoparasites during all four surveys was however higher in Lb. kimberleyensis. Statistical analyses indicate that there was a significant difference (T- test, p = 0.042) between the two fish species. This is indicated clearly in Figure 4.2.

Health Assessment Index

90 80 70 Labeobarbus aeneus 60 Labeobarbus kimberleyensis 50 Labeobarbus aeneus trendline 40 Labeobarbus kimberleyensis 30 trendline

Prevalence (%) 20 10 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surveys

Figure 4.1: Graph depicting the prevalence of ectoparasites in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

100 90 80 Labeobarbus aeneus 70 Labeobarbus kimberleyensis 60 50 Labeobarbus aeneus trendline 40 Labeobarbus kimberleyensis 30 trendline Prevalence (%) 20 10 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surveys

Figure 4.2: Graph depicting the prevalence of endoparasites in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

Mean intensity: The same seasonal trends as were observed for ectoparasite prevalence (Figure 4.1) in each fish species were seen in the mean intensities of ectoparasites (Figure 4.3). Mean intensities in Lb. aeneus were: 2.54 (autumn), 3.24 (winter), 2.30 (spring) and 2.25 (summer). For Lb. kimberleyensis mean intensities differed seasonally as follows: 1.91 (autumn), 1.50 (winter), 1.17 (spring) and 1.00 (summer). Once again, the maximum mean intensity in Lb. aeneus was recorded in winter, the maximum mean intensity in Lb. kimberleyensis was recorded in autumn and the minimum value for both fish species was recorded in summer. As can be seen from the mean intensities listed above, the ectoparasite infestation was more intense in Lb. aeneus during all four surveys. Statistical analyses indicate that, there was a significant difference (T-test, p = 0.008) between the mean intensity of ectoparasites in Lb. aeneus and Lb. kimberleyensis as indicated in Figure 4.3.

Seasonal variations in the mean intensity of endoparasites are shown in Figure 4.4. As can be seen from this figure, in Lb. kimberleyensis a completely different pattern to that observed in the prevalence of endoparasites was seen. Mean intensities decreased considerably from autumn (201.47) to winter (87.16), then decreased again in spring (66.35) followed by an increase in summer (113.50). When looking at the full cycle, this shows that from summer to autumn the mean intensity of endoparasites in largemouth yellowfish almost doubles. In Lb. aeneus the same pattern as was seen in the prevalence

Health Assessment Index of endoparasites was observed for the mean intensity. Autumn (1.75) and summer (5.00) exhibited markedly lower values than winter (21.57) and spring (20.92). For both fish species, there was a considerable difference between the highest and lowest values calculated. The endoparasite infection was considerably more severe in Lb. kimberleyensis throughout the study. Statistical analyses indicate that, there was a significant difference (T-test, p = 0.013) between the mean intensities of endoparasites in the two fish species. This is clearly observed in Figure 4.4.

3.5 3.0 Labeobarbus aeneus 2.5 Labeobarbus kimberleyensis 2.0 Labeobarbus aeneus trendline 1.5 Labeobarbus kimberleyensis trendline

Mean Intensity 1.0 0.5 0.0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surevys

Figure 4.3: Graph depicting the mean intensity of ectoparasites in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

250

200 Labeobarbus aeneus

150 Labeobarbus kimberleyensis Labeobarbus aeneus trendline 100 Labeobarbus kimberleyensis

Mean Intensity trendline 50

0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surveys

Figure 4.4: Graph depicting the mean intensity of endoparasites in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

Abundance (relative density): The seasonal trend observed in the abundance (relative density) of ectoparasites (Figure 4.5) was similar to that already observed in the prevalence and mean intensity (Figure 4.1 and Figure 4.3 respectively) of ectoparasites. Abundance values for Lb. aeneus were: 1.65 (autumn), 2.75 (winter), 1.15 (spring) and 0.9 (summer). Values for Lb. kimberleyensis were: 1.05 (autumn), 0.75 (winter), 0.35 (spring) and 0.05 (summer). The maximum values for Lb. aeneus and Lb. kimberleyensis were recorded in winter and autumn respectively, while the minimum value for both fish species was recorded in summer. The abundance of ectoparasites during all four surveys was greater in Lb. aeneus. Statistical analyses indicate however that there was no significant difference between the abundance of ectoparasites in the two fish species.

The same trend was observed in the abundance (relative density) of endoparasites (Figure 4.6) in each fish species as was seen in the mean intensity of endoparasites in each fish species

Health Assessment Index

(Figure 4.4). In Lb. aeneus, mean intensities in autumn (0.35) and summer (1.25) exhibited considerably lower values than winter (15.1) and spring (13.6). The highest and lowest value recorded for Lb. aeneus was in winter and autumn respectively. In Lb. kimberleyensis, abundance values decreased considerably from autumn (171.25) to winter (82.8), then decreased again in spring (56.4) followed by an increase in summer (102.15). When looking at the full cycle again, this shows that from summer to autumn the mean intensity of endoparasites in largemouth yellowfish almost doubles. The highest and lowest value recorded in Lb. kimberleyensis was in autumn and spring respectively. Statistical analyses indicate that, there was a significant difference (T-test, p = 0.009) between the abundance of endoparasites in Lb. aeneus and Lb. kimberleyensis. This is clearly demonstrated in Figure 4.6.

3.0

2.5 Labeobarbus aeneus 2.0 Labeobarbus kimberleyensis

1.5 Labeobarbus aeneus trendline Labeobarbus kimberleyensis

Abundance 1.0 trendline 0.5

0.0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surveys

Figure 4.5: Graph depicting the abundance of ectoparasites in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

180 160 Labeobarbus aeneus 140 120 Labeobarbus kimberleyensis 100 Labeobarbus aeneus trendline 80 Labeobarbus kimberleyensis

Abundance 60 trendline 40 20 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Surveys Figure 4.6: Graph depicting the abundance of endoparasites in Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

4.3.2. HEALTH ASSESSMENT INDEX AND ASSOCIATED PARASITE INDEX 4.3.2.1. Health Assessment Index The frequency of abnormal conditions recorded during all four surveys while conducting the health assessment on the two fish species collected is outlined in Table 4.4. The numbers given in the table refer to the number of fish observed with the abnormality.

Health Assessment Index

Table 4.4: Table depicting the frequency of abnormal categorical HAI variables observed for each survey per fish species Autumn Winter Spring Summer Variables (April 2000) (June 2000) (October 2000) (January 2001) Lb.a Lb.k Lb.a Lb.k Lb.a Lb.k Lb.a Lb.k Eyes 0 0 0 0 0 0 0 0 Skin 0 0 0 0 0 0 0 0 Fins 1 1 1 3 0 0 0 0 Opercules 0 0 0 1 0 0 0 0 Gills 4 7 9 11 9 9 12 10 Liver 20 20 20 18 20 20 20 20 Spleen 0 0 0 0 0 0 0 0 Hindgut 0 1 1 6 0 2 11 11 Kidney 0 0 0 1 0 0 2 0 Blood (haematocrit) 6 5 6 6 5 8 9 11 Leukocrit 3 0 4 0 0 1 4 2 % White blood cells 3 0 4 0 0 1 4 2 Abbreviations: Lb.a – Labeobarbus aeneus Lb.k – Labeobarbus kimberleyensis

Following the assessment done in the field, two Health Assessment Index (HAI) values were calculated for each fish. The first HAI value was a straightforward calculation including all parasites collected and the second HAI value was calculated excluding the number of B. acheilognathi (Asian tapeworm) found in each fish. Table 4.5 below outlines both HAI calculations including the standard deviation (SD) and coefficient of variations (CV) (as a percentage). The coefficient of variation indicates the level of stress experienced by a fish population (Adams et al. 1993). The average HAI value for each season and fish species is highlighted in grey.

Table 4.5: Table showing the HAI calculations for Labeobarbus aeneus and Labeobarbus kimberleyensis during four surveys at the Vaal Dam Labeobarbus aeneus Labeobarbus kimberleyensis Survey HAI with all HAI without HAI with all HAI without parasites B. acheilognathi parasites B. acheilognathi Autumn Mean 100 100 125.5 104 April SD 25.13 25.13 26.25 19.84 2000 CV% 25.13 25.13 20.92 19.08 Winter Mean 113 110 133.5 114.5 June SD 25.15 24.28 21.83 21.14 2000 CV% 22.26 22.07 16.35 18.46 Spring Mean 105.5 103.5 124.5 111.5 October SD 17.91 14.61 28.74 25.4 2000 CV% 16.98 14.12 23.08 22.78 Summer Mean 125 124.5 141 124.5 January SD 39.54 39.53 38.10 30.86 2001 CV% 31.63 31.75 27.02 24.79

In Lb. aeneus, both HAI values (including all parasites and excluding the number of Asian tapeworms) increased from autumn (100 and 100) to winter (113 and 110), decreasing in spring (105.5 and 103.5) and increasing again in summer (125 and 124.5) with the highest value recorded in summer and the lowest in autumn. As can be seen from Table 4.5, for all four surveys except autumn HAI values including all parasites were slightly higher than the HAI values excluding the number of B. acheilognathi.

Health Assessment Index

The standard deviations calculated for the HAI including all parasites ranged from 17.91 in spring to 39.54 in summer. Coefficient of variations (as a percentage) were >15% for all seasons with a maximum of 31.63% recorded in summer and a minimum of 16.98% recorded in spring. Standard deviations calculated for the HAI excluding the number of B. acheilognathi showed similar variations with a standard deviation of 14.61 in spring and 39.53 in summer. The coefficient of variations for the surveys were >15%, except in spring where the coefficient of variation was <15%.

In Lb. kimberleyensis, as for Lb. aeneus, both HAI values (including all parasites and excluding the number of B. acheilognathi) increased from autumn (125.5 and 104) to winter (133.5 and 114.5), decreasing in spring (124.5 and 111.5) and increasing again in summer (141 and 124.5). The highest HAI value was recorded in summer and the lowest in spring (HAI including all parasites) and autumn (HAI excluding number of B. acheilognathi). The HAI including all parasites was higher during all four surveys than the HAI excluding the number of B. acheilognathi (Table 4.5).

The standard deviations calculated for the HAI including all parasites ranged from 21.83 in winter through 26.25 (autumn) and 28.74 (spring) to 38.1 in summer. Coefficient of variations (as a percentage) were >15% for all seasons with a maximum of 27.02% (summer) and a minimum of 16.35% (winter). Standard deviations calculated for the HAI excluding the number of B. acheilognathi varied from 19.84 in autumn to 30.86 in summer. For winter and spring the standard deviations were 21.14 and 25.4 respectively. The coefficient of variations for all surveys were above 15% with the maximum being 24.79% (summer) and the minimum being 18.46% (winter).

The seasonal variations between the two calculated Health Assessment Index values for both fish species is seen more clearly in the figures below. Both figures show a seasonal pattern (red for Lb. aeneus and green for Lb. kimberleyensis) in HAI values. Figure 4.7 illustrates the straightforward calculation including all parasites while Figure 4.8 illustrates the HAI value excluding the number of B. acheilognathi.

140 120 Labeobarbus aeneus 100 Labeobarbus kimberleyensis 80 Labeobarbus aeneus trendline 60 Labeobarbus kimberleyensis HAI value HAI 40 trendline 20 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Survey

Figure 4.7: Graph illustrating the Health Assessment Index (calculated including the number of Bothriocephalus acheilognathi) for Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

Health Assessment Index

120 100 Labeobarbus aeneus 80 Labeobarbus kimberleyensis 60 Labeobarbus aeneus trendline

HAI value HAI Labeobarbus kimberleyensis 40 trendline 20 0 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Survey

Figure 4.8: Graph illustrating the Health Assessment Index (calculated excluding the number of Bothriocephalus acheilognathi) for Labeobarbus aeneus and Labeobarbus kimberleyensis during the four surveys

4.3.2.2. Logistic regression The stepwise logistic regression model, determines with what probability a fish can be attributed to the correct species, using HAI variables. The variables used in the analyses included the presence/absence of B. acheilognathi, presence/absence of other cestode spp., presence/ absence of ectoparasites, presence/absence of endoparasites, total length of fish, fork length of fish, weight, condition factor, HAI including all parasites, HAI excluding the number of B. acheilognathi.

Table 4.6 summarises the model and delineates the variables that would most accurately predict from which fish species (Lb. aeneus or Lb. kimberleyensis) a randomly chosen fish was collected. These variables included the absence/ presence of B. acheilognathi, the absence/ presence of ectoparasites and the fork length of the fish. On its own, the absence/presence of B. acheilognathi would accurately predict 80.6% of the fish as belonging to the correct fish species. With the addition of the absence/ presence of ectoparasites, the predictive ability stays the same at 80.6% but with the addition of the fish’s fork length, the predictive ability increases to 83.1%. The use of the fish fork length as a predictive variable in this analysis will be discussed below in the discussion.

Table 4.6: Summary of the stepwise logistic regression model illustrating the three HAI variables most accurately predicting the species of the fish Model Percentage Step Chi- Degree of correctly Variable Significance square freedom classified 1 65.238 1 0.000 80.6% Presence/absence of B. acheilognathi 2 79.415 1 0.001 80.6% Presence/absence of B. acheilognathi Presence/absence of ectoparasites 3 87.263 1 0.011 83.1% Presence/absence of B. acheilognathi Presence/absence of ectoparasites Fork length

Health Assessment Index

4.3.3. CONDITION FACTOR The minimum, maximum and mean condition factor, mean mass (in grams) and mean length (in millimeters) for both yellowfish species is given in Table 4.7. The average condition factor for each fish species and per season is highlighted in grey.

Table 4.7: Table showing the condition factor (CF) calculated for each fish species during the four surveys Survey Condition variable Labeobarbus aeneus Labeobarbus kimberleyensis Autumn Min. CF 0.839 0.544 April 2000 Max. CF 1.296 1.069 Mean CF 0.996 0.903 SD 0.118 0.134 SE 2.64 x 10-2 3.00 x 10-2 Mean mass (g) 744.00 825.50 Mean length (mm) 368.30 392.20 Winter Min. CF 0.350 0.425 June 2000 Max. CF 1.065 1.040 Mean CF 0.851 0.818 SD 0.182 0.157 SE 4.06 x 10-2 3.52 x 10-2 Mean mass (g) 666.00 593.75 Mean length (mm) 373.25 364.50 Spring Min. CF 0.581 0.706 October 2000 Max. CF 1.108 1.092 Mean CF 0.988 0.932 SD 0.123 9.40 x 10-2 SE 2.74 x 10-2 2.10 x 10-2 Mean mass (g) 814.00 924.25 Mean length (mm) 383.35 405.70 Summer Min. CF 0.806 0.793 January 2001 Max. CF 2.039 1.026 Mean CF 1.020 0.918 SD 0.254 6.68 x 10-2 SE 5.69 x 10-2 1.49 x 10-2 Mean mass (g) 702.50 970.00 Mean length (mm) 370.90 414.90 Abbreviations: CF - Condition Factor Min - Minimum Max - Maximum SD - Standard deviation SE - Standard error

The condition factor of a fish is classified as ideal when the factor value is 1.00 (Doyon, Downing and Magnin 1988). However this is dependent on the fish species sampled as weight and length relationships differ between fish species and condition factors may vary within fish species depending on the geographic location (Doyon, Downing and Magnin 1988). In autumn, the average condition factor for Lb. aeneus was considerably higher than in Lb. kimberleyensis. In winter, although the average condition factors for both fish species dropped, Lb. aeneus still exhibited a slightly higher condition factor. In spring, Lb. aeneus was once again in better health than Lb. kimberleyensis. In summer, a considerable difference in the average condition factor was seen between the two fish species. As can be seen in Figure 4.9 below, Lb. aeneus, as a general trend had a higher condition factor than Lb. kimberleyensis. Statistical analyses indicate that the average condition factor between fish species differed significantly (p = 0.05, Pearson Chi-square).

Health Assessment Index

The seasonal variations in the condition of the fish are clearly observed in Figure 4.9. Lb. aeneus and Lb. kimberleyensis demonstrate the same seasonal variations except in summer where the condition factor increases for Lb. aeneus and decreases for Lb. kimberleyensis.

1.00

0.80 Labeobarbus aeneus

0.60 Labeobarbus kimberleyensis Labeobarbus aeneus trendline 0.40 Labeobarbus kimberleyensis Condition Factor 0.20 trendline

0.00 April 2000 June 2000 October 2000 January 2001 Autumn Winter Spring Summer Survey

Figure 4.9: Graph illustrating seasonal changes in the condition factor of Labeobarbus aeneus and Labeobarbus kimberleyensis sampled at the Vaal Dam

4.3.4. ECOLOGICAL PARAMETERS The variables used in the statistical analyses, as outlined in Section 4.2 (Table 4.1), included the number of B. acheilognathi, the number of other cestode spp., the number of ectoparasites, the number of endoparasites, total length of fish, fork length of fish, weight, condition factor, HAI including all parasites and HAI excluding the number of B. acheilognathi found. The presence/absence of certain variables, namely B. acheilognathi, other cestode spp., ectoparasites and endoparasites were also included in certain analyses as indicated in Table 4.1. It is important to note that no ‘other cestode spp.’ were found in Lb. aeneus during the four surveys (Table 4.3) and therefore for certain statistical analyses this variable could not be used. The statistics regarding B. acheilognathi have been discussed in detail in Chapter 3. Only certain observations have been repeated here to aid in discussing the health assessment index.

Gender specificity (data pooled according to fish gender): When the data for Lb. aeneus is pooled according to gender, the only significant difference (T-test, p = 0.021) observed between males and females was in the average weight of the fish. The average weight for males and females was 662.58 and 780.11 grams respectively. The averages of other variables such as number of B. acheilognathi and the number of endoparasites differed between males and females however the differences were not significant (T-test, p values were greater than 0.05). Although the presence of ectoparasites differed between males (16) and females (32), the presence of ectoparasites was not dependent of fish gender (Pearson Chi-square, p > 0.05).

For Lb. kimberleyensis, the only significant difference (T-test, p = 0.033) observed between males and females was in the average number of ectoparasites found, which was 0.32 and 0.83 respectively. The averages of other variables such as number of B. acheilognathi, number of other cestode spp., number of endoparasites, total length of fish, fork length of fish and weight differed between males and

Health Assessment Index females however the differences were not significant (T-test, p values were greater than 0.05). Even though a significant difference was seen between the average number of ectoparasites found in males and females, the presence/absence of ectoparasites is not dependent on the gender of Lb. kimberleyensis (Pearson Chi-square, p > 0.05).

Seasonality (data pooled according to seasons): For Lb. aeneus the highest number of both ectoparasites and endoparasites were recorded in winter (June 2000) while for Lb. kimberleyensis the highest numbers of both ectoparasites and endoparasites were recorded in autumn (April 2000) (Table 4.2 and Table 4.3). From the statistical analyses, the presence/absence of ectoparasites and presence/absence of endoparasites in Lb. aeneus are dependent on the season (Pearson Chi- square, p = 0.022 and p = 0.001 respectively). The highest number of Lb. aeneus infested with ectoparasites was 17 (winter) and the lowest was 8 (summer). The highest and lowest number of endoparasites found in Lb. aeneus was 14 (winter) and 4 (autumn) respectively. In Lb. kimberleyensis, only the presence/absence of ectoparasites is dependent on the season (Pearson Chi-square, p = 0.003). The maximum and minimum number of ectoparasites found in Lb. kimberleyensis was 11 (autumn) and 1 (summer) respectively. Similar numbers of Lb. kimberleyensis (between 17 and 19) were infected with endoparasites during all four seasons.

The ANOVA test was conducted to determine if there are significant differences in variables between seasons. For Lb. aeneus, variables which exhibited significant seasonal variations included the number of ectoparasites (p = 0.009), condition factor (p = 0.016), HAI including all parasites (p = 0.037) and HAI excluding the number of B. acheilognathi (p = 0.031). For Lb. kimberleyensis there were significant seasonal differences between number of ectoparasites (p = 0.006), number of endoparasites (p = 0.005), total length of fish (p = 0.009), fork length of fish (p = 0.002), weight (p = 0.004) and condition factor (p = 0.015).

Post Hoc tests (multiple comparison tests), namely the Scheffe test and Dunnet T3 test, were then used to determine which seasons differed significantly. Seasons differing significantly and the correlating HAI variable for are presented in Table 4.8.

Table 4.8: Table outlining the significantly different seasons and correlating HAI variables for Labeobarbus aeneus and Labeobarbus kimberleyensis Labeobarbus aeneus HAI variable Seasons differing significantly P-value Statistical test Number of ectoparasites Winter Summer p = 0.018 Scheffe Winter Spring p = 0.055 Scheffe Condition factor Winter Summer p = 0.036 Scheffe HAI including all parasites Autumn Summer p = 0.055 Scheffe HAI excluding the number of Autumn Summer p = 0.104 Dunnet T3 B. acheilognathi* Note: Months highlighted in grey may be significant because p values are close to 0.05 but this is not certain. * - Sample size too small and/or variance to big to be able to distinguish between seasons. The closest seasons are given in the table.

Health Assessment Index

Table 4.8 continued Labeobarbus kimberleyensis HAI variable Seasons differing significantly P-value Statistical test Number of ectoparasites Autumn Summer p = 0.033 Dunnet T3 Winter Summer p = 0.039 Dunnet T3 Number of endoparasites Autumn Spring p = 0.012 Dunnet T3 Total length of fish Winter Summer p = 0.014 Scheffe Fork length of fish Winter Spring p = 0.031 Scheffe Winter Summer p = 0.005 Scheffe Weight Winter Spring p = 0.033 Scheffe Winter Summer p = 0.011 Scheffe Condition factor Winter Spring p = 0.033 Scheffe

Species specificity (data pooled according to fish species): When data was pooled to determine differences between fish species, the presence/absence of the following HAI variables were used: B. acheilognathi, other cestode spp., ectoparasites and endoparasites. Statistical analyses (Pearson Chi-square) show that all the above-mentioned variables are highly dependent on fish species with p values of 0 (presence/absence of B. acheilognathi and presence/absence of endoparasites), 0.028 (presence/absence of other cestode spp.) and 0.003 (presence/absence of ectoparasites). When comparing the two fish species: ∗ ectoparasites were present in 60% of Lb. aeneus and 35% of Lb. kimberleyensis; ∗ endoparasites were present in 45% of Lb. aeneus and 88.8% of Lb. kimberleyensis; ∗ B. acheilognathi were present in 23.8% of Lb. aeneus and 85% of Lb. kimberleyensis; and ∗ other cestode spp. were present in 0% of Lb. aeneus and 0.1% of Lb. kimberleyensis.

4.4. DISCUSSION This discussion focuses on the observations made while conducting the Health Assessment Index (HAI) and associated inverted Parasite Index (PI) on the two yellowfish species sampled during this study. Attention is given to each individual fish species in terms of seasonal patterns observed in parasite numbers (in terms of endo- and ectoparasites), parasite infestation (in terms of endo- and ectoparasites), the HAI (including all parasites versus excluding the number of Bothriocephalus acheilognathi) and condition factor.

The discussion has been presented in columns to assist with easy comparison between the two yellowfish species sampled. The reason for comparing fish species, as discussed in the introduction to this chapter is to enable the author to determine whether there are correlations, if any, between the high infection of B. acheilognathi (discussed in Chapter 3) and the health of the infected fish.

In addition a general comparison between the two fish species sampled with regards to parasite numbers (in terms of endo- and ectoparasites), parasite infestation (in terms of endo- and ectoparasites), the HAI (including all parasites versus excluding the number of B. acheilognathi), condition factor and species specificity will also be discussed.

Health Assessment Index

Labeobarbus kimberleyensis – the largemouth yellowfish Labeobarbus aeneus – the smallmouth yellowfish Parasite numbers Lb. kimberleyensis, as concluded in the previous chapter, had a high infection A fairly high number of endoparasites were recorded in Lb. aeneus of which of the introduced Asian tapeworm, B. acheilognathi. As the majority of most belonged to the tapeworm species B. acheilognathi. However the high endoparasites found in this study belonged to this tapeworm species it makes numbers were only observed in winter and spring as discussed below. The sense that high numbers of endoparasites, as a group, would also be found in number of endoparasites in Lb. aeneus was higher than the number of this fish species. This confirms assumptions made during previous field trips ectoparasites observed due to the presence of B. acheilognathi in Lb. aeneus to the Vaal Dam that largemouth yellowfish are affected by high numbers of (Chapter 3). However, if the number of B. acheilognathi were excluded from endoparasites. The number of endoparasites in Lb. kimberleyensis was the calculations a slightly higher but almost similar number of endoparasites considerably higher (due to the number of Asian tapeworms) than the number would have been recorded in comparison to ectoparasites. of ectoparasites observed. However, if one were to exclude the number of B. acheilognathi found in Lb. kimberleyensis during the study the same would Although relatively high numbers of endoparasites were observed in still be true based on the number of trematode cysts. Lb. aeneus during winter and spring when compared to autumn and summer statistical analyses indicate that there were no significant differences between Statistical analyses indicate that both the number of endoparasites and the seasons. On the other hand, for ectoparasites, statistical analyses indicate number of ectoparasites differed significantly between seasons. The that significant differences exist between seasons with the highest and lowest maximum and minimum number of endoparasites was recorded in autumn and numbers recorded in winter and summer respectively. In addition, it is spring respectively. For ectoparasites seasons which differed significantly (in possible that there is a significant difference between winter and spring terms of numbers) were autumn and summer in addition to winter and although this is not certain (either the sample size was too small or the summer. variance was too large to distinguish between seasons).

Statistical analyses indicate that significant differences exist in the average In terms of endo- and ectoparasite numbers, no significant differences were number of ectoparasites found in male and female largemouth yellowfish observed between male and female smallmouth yellowfish and therefore their however the presence of ectoparasites was not dependent on fish gender. presence is not dependent on fish gender. A study conducted by Le Roux (2001) on Lb. aeneus in the Vaal Dam supports the results of this study that ectoparasites are not fish gender specific.

Health Assessment Index

Labeobarbus kimberleyensis – the largemouth yellowfish Labeobarbus aeneus – the smallmouth yellowfish Prevalence, abundance and mean intensity of endo- and ectoparasites The prevalence of endoparasites in Lb. kimberleyensis remained fairly high Statistical analyses indicate that the presence/absence of endoparasites in and almost constant throughout the four surveys (Figure 4.2). It is possible Lb. aeneus is dependent on the season. This is observed clearly when that a slight seasonal pattern was exhibited but this is not clearly reviewing seasonal differences in endoparasite prevalence. The prevalence, demonstrated. The high prevalence of endoparasites can be attributed to the abundance and mean intensity of endoparasites in Lb. aeneus exhibited the fact that most largemouth yellowfish sampled were infected with same seasonal variation with notable increases and decreases observed endoparasites. This corresponds with the high prevalence of B. acheilognathi during all four surveys (Figure 4.2). Subsequently, winter and spring had reported in the previous chapter. The mean intensity and abundance of similar values and autumn and summer had similar values with values in endoparasites exhibited similar trends to each other (Figure 4.4 and Figure 4.6 winter and spring being considerably higher than autumn and summer. The respectively). High values were recorded in autumn followed by a decrease in maximum and minimum values were recorded in winter and autumn winter after which the seasonal variations were minimal. respectively.

The prevalence of ectoparasites in Lb. kimberleyensis exhibited a gradual Similarly the presence/absence of ectoparasites in Lb. aeneus is dependent seasonal decrease from autumn (highest value recorded) through to spring on season. The infestation statistics for ectoparasites (Figure 4.1, Figure 4.3 (lowest value recorded) (Figure 4.1) resulting in the presence/absence of and Figure 4.5) demonstrate a similar pattern in all three statistics with the ectoparasites being statistically dependent on season. The mean intensity highest and lowest values recorded in winter and summer respectively. The (Figure 4.3) and abundance (Figure 4.5) of ectoparasites exhibited a similar seasonal variation observed in the mean intensity however was more gradual. seasonal trend as observed in the prevalence of ectoparasites although more Apart from the extremely high endoparasite infection recorded in winter and gradual. spring, the remaining surveys exhibited similar values to those observed in the ectoparasite infestation. The prevalence, abundance and mean intensity of endoparasites in Lb. kimberleyensis is far greater than the ectoparasite infestation. For Lb. aeneus the prevalence of ectoparasites was in general greater than the endoparasite prevalence. In contrast, the general abundance and mean intensity of endoparasites was higher than ectoparasites due to the high infestation in winter and spring.

Health Assessment Index

Labeobarbus kimberleyensis – the largemouth yellowfish Labeobarbus aeneus – the smallmouth yellowfish Health Assessment Index Fish abnormalities Most variables used in the HAI are given an index rating of 0 for normal and 30 an indication of a severe deviation from the norm.

When assessing these variables several abnormal conditions were seen in When assessing these variables for Lb. aeneus, a number of abnormal Lb. kimberleyensis. The following observations were made (the percentages conditions were seen. The following observations were made (the given below refer to the number of fish in the sample (n = 20 per season) percentages given below refer to the number of fish in the sample (n = 20 per showing the abnormal condition): season) showing the abnormal condition): ∗ no abnormal conditions were recorded in variables such as eyes, skin and ∗ no abnormal conditions were recorded in variables such as eyes, skin, spleen; opercules and spleen; ∗ in some seasons only a few fish had abnormal fins [in autumn (5%) and ∗ in some seasons only a few fish had abnormal fins [in autumn (5%) and winter (15%)], opercules [in winter (5%)], kidney [in winter (5%)], leukocrit winter (5%)], kidney [in summer (10%)], leukocrit and percentage white and percentage white blood cells [in spring (5%) and summer (10%)]; blood cells [in autumn (15%), winter (20%) and summer (20%)]; ∗ abnormal gills were seen during all four surveys ranging from 35% in autumn ∗ abnormal gills were seen during all four surveys ranging from 20% in autumn to 55% in winter; to 60% in summer; ∗ 5% of the fish sampled in autumn, 10% of the fish sampled in spring, 30% of ∗ abnormal haematocrit (blood) values were recorded during all four surveys the fish sampled in winter and more than 50% of the fish sampled in summer ranging from 25% in spring to 45% in summer; suffered from inflamed hindguts; ∗ 5% of the fish sampled in winter and 55% of the fish sampled in summer ∗ abnormal haematocrit (blood) values were recorded during all four surveys suffered from inflamed hindguts; ranging from 25% in autumn to 55% in summer; ∗ all the fish sampled exhibited abnormal conditions in the liver. ∗ almost all the fish sampled (97.5%) exhibited abnormal conditions in the liver.

There are two HAI variables which are not included in the calculations but do give a significant amount of information regarding the status of the fish sampled, these include the colour of bile and mesenteric fat and are discussed below.

Health Assessment Index

Labeobarbus kimberleyensis – the largemouth yellowfish Labeobarbus aeneus – the smallmouth yellowfish Colour of bile According to Love (1980) the colour of the bile is a good short term indication of the nutritional status of the fish and the fish’s feeding activity. Indications of no recent feeding could be due to a change in feeding behaviour or a lack of available food (Goede and Barton 1990).

When assessing the colour of the bile, most largemouth yellowfish (85 - 90%) The colour of the bile in all Lb. aeneus sampled in winter, spring and summer sampled in autumn, summer and spring were feeding well (eaten in the last indicated that fish were feeding well (eaten in the last couple of days/hours) couple of days/hours) (bile was yellow or straw colour and the bladder was full while in autumn only 70% of the fish were feeding well with the remaining 30% to empty) with the remaining fish having fed in the last week while in winter all having fed in the last week (bile was light green to grass). No indication of of the largemouth yellowfish sampled were feeding well. No indications of starvation due to a change in feeding behaviour or lack of food was observed. change in feeding behaviour or lack of food were observed. As a result the This explains the high endoparasite infection observed in winter and spring endoparasite infection as discussed above was high. although one would then expect the same in summer however this is not the case.

Mesenteric fat Mesenteric fat (level of fat) on the other hand reflects the intensity of feeding and energy deposition over the long term (Goede and Barton 1990). It also gives an idea of the stress experienced by fish although it is not directly related to stress (Goede and Barton 1990). The fat index as pointed out by Goede and Barton (1990) can vary with season, sex and aquatic systems. All fish in this study were collected from one locality therefore changes in fat levels cannot be attributed to the aquatic system. Newsome and Leduc (1975) have reported that the autumn condition of females shows a decrease in mesenteric fat as they normally use fat in the autumn and winter for ovary development.

During this study, largemouth yellowfish were generally fatter (had more The mesenteric fat deposition in Lb. aeneus during the study was generally mesenteric fat) in spring, followed by winter and summer and leaner in autumn higher in summer and spring with lower fat levels observed in autumn and the although the difference in seasons was minimal. It is interesting to note that in least in winter. This pattern correlates with suggestions made by Newsome general more male fish were caught in spring and summer and more female and Leduc (1975) above regarding ovary development. Once again more fish were caught in autumn and winter therefore supporting suggestions made male fish were caught in spring and summer and more female fish were by Newsome and Leduc (1975) above however this does not explain the caught in autumn and winter further supporting the suggestion made by relatively high levels of fat observed in winter. Newsome and Leduc (1975).

Health Assessment Index

Labeobarbus kimberleyensis – the largemouth yellowfish Labeobarbus aeneus – the smallmouth yellowfish Index values With an index rating of 0 for normal and 30 an indication of a severe deviation from the norm, the higher the HAI calculation for the population being assessed the poorer the health profile of the fish in that aquatic system (Adams et al. 1993). The coefficient of variation (CV) calculated for each index value, as highlighted by Adams et al. (1993) indicates the level of stress experienced by a fish population.

The two index values calculated for Lb. kimberleyensis increased and For both index values calculated for Lb. aeneus, a slight variation between decreased slightly between seasons with the highest values recorded in seasons was observed with a maximum value recorded in summer (indicating summer (both indices) (indicating a poorer health profile) and the lowest in a poorer health profile) and a minimum value recorded in autumn (indicating a spring (HAI including all parasites) and autumn (HAI excluding the number of better health profile) (Figure 4.7 and Figure 4.8). Statistical analyses however B. acheilognathi) (indicating a better health profile) (Figure 4.7 and Figure 4.8). indicate that there is a significant difference between the HAI (excluding the As a result no significant differences (statistically) were observed between number of B. acheilognathi) calculated in autumn and that calculated in seasons. When comparing the two index values the HAI calculated using all summer. It is also possible that there is a significant difference in HAI values parasites was slightly higher than the HAI excluding the number of (including all parasites) recorded in autumn and summer although this is not B. acheilognathi during all four surveys. certain either due to the sample size being too small or the variance to large to distinguish accurately between seasons. The same seasonal pattern was In this study, the CV calculated for both HAI indices was the highest in observed in both indices. When comparing the two HAI indices calculated, summer and the lowest in winter. The following seasonal pattern was similar values (almost the same) were recorded for both indices with the HAI observed: CV values decreased from autumn to winter followed by an increase including all parasites, exhibiting fraction higher values. in spring and a further increase in summer. When comparing the CV for the two HAI indices, the CV for the HAI including all parasites was slightly higher The coefficient of variation (CV), for both HAI indices was the highest in than the second index (HAI excluding the number of B. acheilognathi) in summer followed by autumn, then winter, then spring (lowest CV value). autumn, spring and summer but in winter was slightly lower than the second When comparing the CV for the two HAI indices, the CV for the HAI including index value. all parasites was the same as the second index (HAI excluding the number of B. acheilognathi) in autumn, higher than the second index in winter and spring but lower than the second index in summer.

Health Assessment Index

Labeobarbus kimberleyensis – the largemouth yellowfish Labeobarbus aeneus – the smallmouth yellowfish Condition factor The condition factor is based on the relationship between the fish’s weight and length (Carlander 1969). There are a number of factors, as highlighted by Bolger and Connoly (1989), that influence fish weight such as changes due to breeding activity, food availability and metabolic rate (due to temperature changes). Environmental stress and parasitic infestation as highlighted by Crafford (2000) could influence feeding time. In addition feeding time is reduced in stressed fish (Heath 1987).

The condition factor calculated for Lb. kimberleyensis showed a slight variation The condition factor for Lb. aeneus varied slightly between seasons with the between seasons (Figure 4.9) decreasing from autumn to winter then following trend observed: values decreased from autumn to winter, then increasing in spring and decreasing again in summer. However statistical increased in spring followed by a further increase in summer. Statistical analyses indicate that there is a significant difference between the condition analyses indicate that there is a significant difference between the condition factor recorded in winter (minimum) and that recorded in spring (maximum). factor recorded in winter (minimum) and that recorded in summer (maximum). Seasonal differences in largemouth yellowfish conditions are expected due to Seasonal differences in smallmouth yellowfish conditions are expected due to the spawning activity in mid to late summer (Skelton 2001) and lowered the spawning activity in spring to mid summer (Skelton 2001) and subsequent metabolism and therefore feeding in winter. For Lb. kimberleyensis the feeding. However, as outlined previously most fish were feeding well highest and lowest factor values were recorded in spring and winter throughout the year. Statistical analyses indicate that the average weight of respectively. Statistical analyses indicate that there were significant seasonal Lb. aeneus differed significantly between male and female fish which could differences between the total length of fish, fork length of fish and weight of contribute to the seasonal patterns observed as more male fish were sampled fish recorded in winter, summer and spring which could account for the in spring and summer and more female fish in autumn and winter. The variation in this fish species’ condition factor. The calculated condition factors calculated condition factors suggest that smallmouth yellowfish sampled suggest that largemouth yellowfish sampled during this study were in fairly during this study were in fairly good condition. This is supported by good condition. This is supported by other observations made during the observations discussed previously namely the colour of the bile (most fish assessment of the fish’s health namely colour of the bile (most fish were were feeding well) and mesenteric fat levels (most fish had more than 50% feeding well) and mesenteric fat levels (most fish had more than 50% body body fat). fat).

Health Assessment Index

∗ Comparison between the two yellowfish species sampled Parasite numbers and infestation When comparing fish species, Lb. kimberleyensis hosted a considerably higher number of endoparasites compared to Lb. aeneus and Lb. aeneus on the other hand hosted a larger number of ectoparasites than Lb. kimberleyensis. In addition more largemouth yellowfish were infected with endoparasites and more smallmouth yellowfish were infested with ectoparasites. As a result the endoparasite infection (in terms of prevalence, mean intensity and abundance) was greater in Lb. kimberleyensis while the ectoparasite infestation (in terms of prevalence, mean intensity and abundance) was greater in Lb. aeneus during all four surveys. Statistical analyses indicate that significant differences exist between the mean intensities of ectoparasites in the two yellowfish sampled. Similarly significant differences were recorded between endoparasite prevalence in the two yellowfish species sampled as well as in endoparasite mean intensities and endoparasite abundance. It would be expected that both fish species would exhibit the same seasonal trend when looking at ectoparasite infestation as both species and therefore their ectoparasites are exposed to the same quality of water. However this is not the case in this study. Seasonal trends in ectoparasite infestation varied greatly between fish species. In terms of seasonal variations in endoparasite infections, reverse trends in terms of mean intensities and abundance between the two yellowfish species were observed and completely different seasonal trends in prevalence were observed.

Health Assessment Index For both fish species, a similar amount of fish exhibited abnormalities per season and a similar amount of abnormalities were observed throughout the study. In terms of abnormalities, for both fish species, the only variable which exhibited the abnormality in almost all the fish collected (158 out of 160) was the liver. Bile colour and mesenteric fat levels recorded for both yellowfish species throughout the study indicate that the fish sampled are feeding well and are generally healthy with largemouth yellowfish fish ranking slightly higher than smallmouth yellowfish in terms of bile colour and mesenteric fat levels. However in terms of index values, Lb. kimberleyensis exhibited a higher index value than Lb. aeneus when calculating the HAI using all parasites indicating that Lb. aeneus had a healthier profile than Lb. kimberleyensis. When excluding the number of B. acheilognathi from the HAI calculation, largemouth yellowfish had a slightly higher but almost similar index value to Lb. aeneus. When comparing index values, the HAI including all parasites was greater than the HAI excluding the number of B. acheilognathi. In general, the higher index values (including all parasites) for Lb. kimberleyensis seem to agree with the slightly better feeding pattern (as observed while in the bile colour) observed in this fish species.

Condition factor When comparing fish species, Lb. aeneus had a slightly higher condition factor value than Lb. kimberleyensis during all four surveys. This correlates with the HAI values calculated for each yellowfish species as discussed above.

Health Assessment Index

Species specificity in terms of endo- and ectoparasites The presence/absence of ectoparasites, endoparasites, B. acheilognathi and other cestode spp. are all dependent on the species of fish. A greater number of Lb. kimberleyensis were infected with B. acheilognathi and other cestode spp. and therefore endoparasites while a larger number of Lb. aeneus were infested with ectoparasites.

Logistic regression As mentioned under the results of this chapter, logistic regression analyses determines with what probability a fish could be attributed to the correct fish species using HAI variables. The variables found to most accurately predict the species of a randomly chosen fish included the presence/absence of B. acheilognathi, presence/absence of ectoparasites and the fork length of the fish. On its own, the presence/absence of B. acheilognathi would accurately predict 80.6% of the fish as belonging to the correct fish species. With the addition of the presence/ absence of ectoparasites, the predictive ability stays the same at 80.6% but with the addition of the fish’s fork length, the predictive ability increases to 83.1%. However it must be noted that fish fork length in this case cannot be effectively used as a variable to distinguish between the two fish species sampled, as Lb. kimberleyensis as a species is generally larger and therefore longer than Lb. aeneus. In addition, the size of the mesh gill nets used in the study may contribute to the size of fish sampled.

4.5. SUMMARY AND CONCLUSION In summary, Lb. kimberleyensis exhibited a higher endoparasite infection while Lb. aeneus exhibited a higher ectoparasite infestation. The high endoparasite infection in the largemouth yellowfish was due to the high numbers of B. acheilognathi recorded in this fish species (Chapter 3). In terms of HAI variables which give a significant amount of information regarding the general health of fish (bile colour and mesenteric fat), Lb. kimberleyensis exhibited a fraction healthier profile. However when comparing the index values calculated and the condition factors for both fish species Lb. aeneus exhibited a healthier profile. It can therefore be assumed that the high endoparasite infection in Lb. kimberleyensis is causing the increase in HAI values. It must be noted though that the differences observed between Lb. kimberleyensis and Lb. aeneus in terms of bile colour, mesenteric fat, HAI values and condition factors are slight and in some cases almost negligible.

The conclusions drawn from this chapter will now be discussed in relation to the water quality recorded in the Vaal Dam (Chapter 2) and the high B. acheilognathi infection observed during the study (Chapter 3).

Summary and Conclusion

5. GENERAL SUMMARY AND CONCLUSION CHAPTER 5 GENERAL SUMMARY AND CONCLUSION

The helminth parasites observed in both Lb. kimberleyensis (largemouth yellowfish) and Lb. aeneus (smallmouth yellowfish) were identified as Bothriocephalus acheilognathi, the Asian tapeworm, introduced to South Africa via imported cyprinid fish. The reason for identifying the tapeworms was to determine whether the high numbers of endoparasites observed in yellowfish in the Vaal Dam belonged to the introduced Asian species. The aim of this chapter is to identify correlations, if any, between water quality in the Vaal Dam (discussed in Chapter 2), the health of the yellowfish sampled (discussed in Chapter 4), and the high infection rate by the Asian tapeworm (discussed in Chapter 3). To achieve this, the aims of this research project (as outlined in the introduction) in terms of a conclusion are discussed below.

Is the health/condition of the fish related to water quality? In order to answer this question, conclusions related to the quality of water observed in the Vaal Dam during this study (Chapter 2) and the HAI results obtained (Chapter 4) must be highlighted. The Vaal Dam falls into a water quality zone that is defined by Braune and Rogers (1987) as good quality water (Zone 1) with the main pollutants coming from mining and fuel processing plants. This view is supported by the current study. Water quality in the Vaal Dam during the four surveys was relatively good with high turbidity, high dissolved oxygen and low TDS characteristics. Surface water temperatures varied slightly, as would be expected, from 19.9oC in winter to 25oC in summer. The pH of water varied from fairly neutral in winter to slightly alkaline in summer. In terms of hardness (calcium carbonate), water in the Vaal Dam was moderately soft in autumn and summer and slightly harder in winter. Of the twelve (excluding calcium carbonate) macro determinants analysed only calcium (in winter) exhibited above guideline-value concentrations correlating with the slightly harder water recorded. The slightly elevated nitrate (in autumn and winter) and phosphate (in autumn) concentrations recorded could lead to eutrophication in the Vaal Dam but due to the high turbidity, algal blooms are negligible. With reference to trace metals (nine analysed in this study by Rand Water), only slightly elevated levels of aluminium (in autumn and summer), chromium (in winter) and iron (in autumn, winter and summer) were recorded during the current study.

In terms of the health of the fish sampled, both yellowfish species seemed to be fairly healthy in terms of their feeding behaviour/patterns (no indications of lack of food or change in feeding behaviour were observed in the bile colour), mesenteric fat levels and condition factors although several anomalies were observed in the fins, opercules, white blood cells, gills, hindguts, haematocrit, kidney and liver with the liver exhibiting the greatest frequency of abnormality in almost all the fish (both smallmouth and largemouth yellowfish) sampled throughout the study. For Lb. aeneus condition factors varied from 0.851 (in winter) to 1.020 (in summer) while for Lb. kimberleyensis condition factors varied from 0.818 (in winter) to 0.932 (in spring).

Summary and Conclusion

A problem encountered with the HAI, as pointed out by Groenewald (2000), is that there is no scale or range of minimum or maximum HAI values recorded in literature. In addition no other studies on yellowfish in terms of the HAI have been conducted in the Vaal Dam therefore no baseline data is available for comparison. One can only speculate that both fish species were in a fairly healthy condition with Lb. aeneus exhibiting a slightly healthier profile than Lb. kimberleyensis.

To answer the question, to a certain extent the fairly good health of the fish can be related to the fairly good quality of water found in the Vaal Dam during the study. However both fish species were exposed to the same quality water and therefore it is not expected that the slightly higher HAI values observed in largemouth yellowfish would be attributed to water quality.

Is the change in water quality during the various seasons reflected in the Health Assessment Index (HAI) whereby the Parasite Index (PI) is included? In previous studies, as outlined in the introduction to Chapter 4, the HAI and associated PI have been used to distinguish between localities in terms of water quality however in this study only one locality, namely the Vaal Dam was studied over a one year period incorporating all four seasons. Groenewald (2000) suggested that the HAI should be used to compare between different seasons at the same locality and with the same sampling species.

To determine whether any correlations exist between water quality and the HAI values obtained a summary of the seasonal variations observed in water quality variables and the HAI values (both indices) calculated for both yellowfish species are outlined below. For many of the water quality variables, readings were not recorded in spring and in some cases autumn by Rand Water (equipment was being serviced or was out of commission) therefore complete seasonal patterns could not be observed. For these variables (marked with an *) only the maximum and/or minimum values are outlined below. For many of the trace metals, except aluminium, iron and chromium, the measured concentrations were below the detectable concentration limit for the particular metal analysis and therefore no seasonal patterns could be observed. For aluminium, iron and chromium (marked with an *) only the maximum and/or minimum values are outlined below.

Seasonal variations observed in water quality ∗ Surface water temperatures*, pH readings and electrical conductivity increased from winter through spring to summer followed by a decrease in autumn and then a decrease in winter. The minimum and maximum values for all three variables were recorded in winter and summer respectively. Even though no temperature reading was recorded in autumn the above seasonal trend is expected. ∗ Visibility* was the highest in winter and the lowest in spring and summer. ∗ Turbidity decreased from autumn (maximum) through winter and spring to summer (minimum) and then increased in autumn. ∗ Water hardness* and calcium* concentrations were the highest (water was slightly hard) in winter and the lowest (water was moderately soft) in autumn.

Summary and Conclusion

∗ Dissolved oxygen*, potassium*, sodium* and nitrate* concentrations were the highest in winter and the lowest in summer. ∗ Phosphates* were the highest in autumn. ∗ Ammonium* concentrations were the highest in spring. ∗ TDS*, magnesium* and silicon* concentrations were the highest in summer and the lowest in winter. ∗ Fluoride* concentrations were the highest in summer and the lowest in autumn and winter. ∗ Sulphate* concentrations were the highest in spring and summer and the lowest in winter. ∗ Aluminium* and iron* concentrations were the highest in autumn and lowest in winter (for aluminium) and summer (for iron). ∗ Chromium* concentrations were the highest in winter.

Seasonal variations observed in calculated HAI values ∗ HAI values (both indices) for both yellowfish species exhibited the following seasonal trend: values increased from autumn to winter, then decreased in spring followed by an increase in summer. From summer values decreased in autumn. Maximum and minimum values for Lb. aeneus were recorded in summer and autumn respectively. For Lb. kimberleyensis, highest value was recorded in summer while the lowest value was recorded in spring (HAI including all parasites) and autumn (HAI excluding the number of B. acheilognathi).

To answer the question, seasonal changes in water quality in the Vaal Dam over the four surveys is not reflected in the HAI values obtained. It can be noted though that the highest HAI value for both indices in both fish species was calculated in summer which correlates with the highest temperature, pH, electrical conductivity, TDS, fluoride, magnesium, sulphate and silicon concentrations as well as the lowest turbidity, visibility, dissolved oxygen, potassium, sodium, nitrate and iron concentrations recorded during this study. The minimum HAI value recorded in autumn for Lb. aeneus (both indices) and Lb. kimberleyensis (HAI excluding the number of B. acheilognathi) correlates with the highest turbidity, phosphate, aluminium and iron concentrations as well as slightly softer water and the lowest fluoride concentration recorded during this study. The minimum HAI value recorded in spring for Lb. kimberleyensis (HAI including all parasites) correlates with lowest visibility as well as the highest ammonium and sulphate concentrations recorded in this study.

Does the introduced tapeworm, B. acheilognathi have an effect on or play a role in the fish Health Assessment Index? Before one can discuss the role B. acheilognathi played in the fish HAI it is important to outline the main conclusions drawn regarding its infection. B. acheilognathi, as highlighted by Körting (1974), is not restricted to a specific locality for its survival as was observed in this study. From the results obtained B. acheilognathi was host specific preferring Lb. kimberleyensis over Lb. aeneus although a low infection rate was observed in smallmouth yellowfish. In contrast the infection (in terms of prevalence, abundance and mean intensity) in largemouth yellowfish was markedly high. Seasonal patterns observed in the Asian tapeworm’s infection of smallmouth yellowfish were attributed to

Summary and Conclusion

breeding and subsequent feeding patterns of this fish species with relatively high infections recorded in winter and spring. For Lb. kimberleyensis an explanation could not be given regarding the seasonal patterns observed for the mean intensity and abundance of B. acheilognathi. The maximum and minimum mean intensity and abundance values in largemouth yellowfish were recorded in autumn and spring respectively. In addition the prevalence of B. acheilognathi in Lb. kimberleyensis was high throughout the four surveys. Due to this high infection, the stepwise logistic regression model, which determines with what probability a fish can be attributed to the correct species, using HAI variables showed that on its own the presence/absence of B. acheilognathi would accurately predict 80.6% of the fish as belonging to the correct fish species.

When calculating the HAI, two index values were calculated namely one including all parasites observed during the surveys and the other excluding the number of B. acheilognathi found. The reason for doing this was to determine whether the high infection of B. acheilognathi as discussed above affects the HAI values obtained. When comparing the index values calculated for Lb. aeneus the only difference observed between the two indices (HAI including all parasites and the HAI excluding the number of B. acheilognathi) was that during winter and spring slightly higher values were calculated for the HAI including all parasites. In summer even though a slightly higher value was calculated for the HAI including all parasites, the difference in index values was negligible. For Lb. kimberleyensis on the other hand the two indices varied considerably with the HAI including all parasites exhibiting higher values throughout the study than the HAI excluding number of B. acheilognathi. This suggests that the HAI correlates seasonally with the Asian tapeworm infection observed in both fish species. When comparing the two index values calculated for both yellowfish species, statistical analysis indicates that a linear correlation exists between the two indices calculated as demonstrated below in Figure 5.1.

300

200

Labeobarbus aeneus HAI VALUE INCLUDING ALL PARASITES Labeobarbus kimberleyensis

100

0 100 200 300 HAI VALUE EXCLUDING THE NUMBER OF BOTHRIOCEPHALUS ACHEILOGNATHI

Figure 5.1: Graph depicting the relationship between the HAI calculated including all parasites and the HAI calculated excluding the number of Bothriocephalus acheilognathi

Summary and Conclusion

To conclude, as outlined in the introduction, the main purpose of this research project was to determine whether it plays a role in calculating the health status of the fish as expressed in the HAI and thus have an impact on the value obtained for the quality of the environment they live in. The infection of B. acheilognathi was considerably higher in Lb. kimberleyensis than Lb. aeneus thereby resulting in higher HAI values calculated for Lb. kimberleyensis when including the number of Asian tapeworms in the HAI calculations. When excluding the number of Asian tapeworms from the HAI calculations, slightly higher but almost the same values were obtained for both fish species. Therefore B. acheilognathi and its high infection in Lb. kimberleyensis, directly influences the high health index value obtained for this fish species. However it does not reflect a poorer water quality as would be expected when observing higher HAI values. Statistical analyses indicate that if the HAI excluding the number of B. acheilognathi is a good indication of fish health then the HAI including all parasites would also be a good indication of fish health except the HAI values would be 1.03 times higher.

References

6. REFERENCES CHAPTER 6 REFERENCES

ABEL, P.D. 1989. Pollutant toxicity to aquatic animals – Methods of study and their applications. Reviews on Environment Health, 8(1-4):119-55.

ADAMS, S.M. 1990. Status and use of biological indicators for evaluating the effects of stress on fish, in Biological indicators of stress in fish, edited by S.M. Adams. American Fish Society Symposium, 8:1-8.

ADAMS, S.M., BROWN, A.M. & GOEDE, R.W. 1993. A quantitative health assessment index for rapid evaluation of fish condition in the field. Transactions of the American Fisheries Society, 122: 63-73.

AVENANT-OLDEWAGE, A. & EDDY, L.J. 2003. Sewage wastewater management in South Africa. Final report by RAU to Rand Water. Report No.: 2003/02/28. BIOM.GEN. (H4). Hydrobiology Section, Analytical Services, Rand Water, Vereeniging, South Africa.

AVENANT-OLDEWAGE, A. 2001. Parasites as biomonitoring tool for detection of habitat degradation. Journal of the South African Veterinary Association, 72(3):179-180.

AVENANT-OLDEWAGE, A., OLDEWAGE, W.H. & VAN VUREN, J.H.J. 1995. Development of a fish health and condition assessment procedure. Final report to the Institute of Water Quality Studies, Pretoria, South Africa. 107pp

BIRCH, G.F. 1983. The clay mineralogy of sediments in the Vaal Dam and in the rivers supplying and draining the dam upstream of the barrage. Water SA, 9(3):117-123.

BOLGER, T. & CONNOLY, P.L. 1989. The selection of suitable indices for the measurement and analysis of fish condition. Journal of Fish Biology, 34:171-182.

BOOMKER, J., HUCHZERMEYER, F.W. & NAUDÉ, T.W. 1980. Bothriocephalus in the common carp in the Eastern Transvaal. Journal of the South African Veterinary Association, 51(4):263- 264.

BRANDT, F. DE W., VAN AS, J.G., SCHOONBEE, H.J & HAMILTON-ATTWELL, V.L. 1981. The occurrence and treatment of Bothriocephalus in the common carp, Cyprinus carpio in fish ponds with notes on its presence in the largemouth yellowfish Barbus kimberleyensis from the Vaal Dam, Transvaal. Water SA, 7:35-42.

BRAUNE, E., & ROGERS, K.H. 1987. The Vaal River catchment: Problems and research needs. Pretoria: Foundation for Research Development.

References

BRAY, R.A., JONES, A. & ANDERSEN, K.I. 1994. Keys to the cestode parasites of vertebrates, edited by L.F. Khalil, A. Jones & R.A. Bray. UK: CAB International.

BUSH, A.O., LAFFERTY, K.D., LOTZ, J.M. & SHOSTAK, A.W. 1997. Parasitology meets ecology on its own terms: Margolis et al. revisited. Journal of Parasitology, 83(4):575-583.

CARLANDER, K.D. 1969. Handbook of freshwater fishery biology. Iowa State: University Press.

CLEVELAND, J. 1998. Science Junction. http://www.ncsu.edu/sciencejunction/depot/experiments/water/lessons/do/index.html

CRAFFORD, D. 2000. Application of a Fish Health Assessment Index and associated parasite index on Clarius gariepinus (sharptooth catfish) in the Vaal River System, with reference to heavy metals. Unpublished M.Sc. dissertation, Rand Afrikaans University.

DE LEON, G.P.P., GARCIA-PRIETO, L., LEON-REGAGNON, V. & CHOUDHURY, A. 2000. Helminth communities of native and introduced fishes in Lake Patzcuaro, Michoacan, Mexico. Journal of Fish Biology, 57:303-325.

DEPARTMENT OF MINERALS & ENERGY AFFAIRS. 1989. Operating and developing coal mines in the Republic of South Africa. Braamfontein: Minerals Bureau.

DEPARTMENT OF MINERALS & ENERGY AFFAIRS. 1990. Operating and developing coal mines in the Republic of South Africa. Braamfontein: Minerals Bureau.

DEPARTMENT OF MINES. 1980. Coal – A commodity review with special reference to South Africa. Internal Report No. 63. Braamfontein: Minerals Bureau.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1991a. Vaal Dam pamphlet. Department of Water Affairs and Forestry, Communication Services, Republic of South Africa.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1991b. Water quality management policies and strategies in the R.S.A. Pretoria: Department of Water Affairs and Forestry.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1993. Vaal River: Life-blood of a nation. Pretoria: Department of Water Affairs and Forestry.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1996a. South African Water Quality Guidelines: Volume 1: Domestic Use. (2nd ed) Pretoria: Department of Water Affairs and Forestry.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1996b. South African Water Quality Guidelines: Volume 7: Aquatic Ecosystems. (1st ed) Pretoria: Department of Water Affairs and Forestry.

References

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1996c. South African Water Quality Guidelines: Volume 8: Field Guide. (1st ed) Pretoria: Department of Water Affairs and Forestry.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1997. Water quality sampling manual for aquatic ecosystems. Pretoria: Department of Water Affairs and Forestry.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 1998: National Water Act (Act 36 of 1998). Pretoria: Government printer.

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 2000. The Orange River Basin. http://www.dwaf.gov.za/orange/vaaldam.htm

DEPARTMENT OF WATER AFFAIRS & FORESTRY. 2002. Quality of domestic water supplies. Volume 2: Sampling guide. (1st ed). Pretoria: Department of Water Affairs and Forestry.

DEPARTMENT OF WATER AFFAIRS & FORESTRY/RAND WATER. 1996. Vaal River Barrage catchment, Water Quality Management Plan. Volume 1: Phase 1: Scoping study. Pretoria: Department of Water Affairs and Forestry/Rand Water. pp 23.

DEPARTMENT OF WATER AFFAIRS. 1985. Draft report on the management of South African water resources. Pretoria: Department of Water Affairs.

DÖRGELOH, W. 1985. Food selection and competition for food among three fish species, Salmo giardneri, Barbus aeneus and Clarias gariepinus. South African Journal of Science, 81:693.

DOYON, J-F., DOWNING, J.A. & MAGNIN, E. 1988. Variation in the condition of northern pike, Esox lucius. Canadian Journal of Fisheries and Aquatic Science, 45:479-483.

FAIRBANKS, D.H.K., THOMPSON, M.W., VINK, D.E., NEWBY, T.S., VAN DEN BERG, H.M. & EVERARD, D.A. 2000. The South African landcover characteristics database: a synopsis of the landscape. South African Journal of Science, 96:69 - 82.

GOEDE R.W. & BARTON, B.A. 1990. Organismic indices and an autopsy-based assessment as indicators of health and condition of fish. In Biological indicators of stress in fish, edited by S.M. Adams. American Fish Society Symposium, 8:93-108.

GOUDIE, A. & VILES, H. 1997: The Earth Transformed. Oxford: Blackwell.

GOUWS, K. & COETZEE, P.P. 1997. Determination and partitioning of heavy metals in sediments of the Vaal Dam system by sequential extraction. Water SA, 23:217-226

GRANATH, W.O. (Jr) & ESCH, G.W. 1983. Temperature and other factors that regulate the composition and infrapopulation densities of Bothriocephalus acheilognathi (Cestoda) in Gambusia affinis (Pisces). Journal of Parasitology, 69:1116-1124.

References

GROENEWALD, M. 2000. Bioaccumulation of metals and the general health of fish from the Vaal dam and Vaal River Barrage. Unpublished M.Sc. dissertation, Rand Afrikaans University.

HÄNERT, E.T. 1984. ŉ Morfologiese studie van die vislintwurm Bothriocephalus (Cestoda: Pseudophyllidea). Unpublished M.Sc. thesis, Rand Afrikaans University.

HEATH, A.G. 1987. Water pollution and fish physiology. Florida: CRC Press Inc.

HERRICKS, E.E. & CAIRNS, J. (Jr). 1982. Biological monitoring: Part III – receiving system methodology based on community structure. Water Research, 16:141-153.

HOHLS, B.C., SILBERBAUER, M.J., KUHN, A.L., KEMPSTER, P.L. & VAN GINKEL, C.E. 2002. National water resource quality status report: Inorganic chemical water quality of surface water resources in SA – The Big Picture. Institute of Water Quality Studies, Department of Water Affairs and Forestry, Pretoria, South Africa (Report No. N/0000/REQ0801).

HORIZONS OF WINTER. 1997. What is the Lesotho Highlands Water Project? http://www.geocities.com/wim_klunne/lesotho/

HUGO, M.L., VILJOEN, A.T. & MEEUWIS, J.M. 1997. The ecology of natural resource management. Pretoria: Kagiso.

KASRILS, R. 2000. Media statement by the Minister of Water Affairs and Forestry. Department of Water Affairs and Forestry.

KEMPSTER, P.L., HATTINGH, W. A. J. & VAN VLIET, H. R. 1980. Summarised water quality criteria. Pretoria: Department of Water Affairs (Technical report No TR108).

KHALIL, L. 1991. Techniques for identification and investigative helminthology. St Albans: UK: International Institute of Parasitology.

KHALIL, L.F & POLLING, L. 1997. Check list of the helminth parasites of African freshwater fishes. University of the North, Republic of South Africa.

KHAN, R.A. & THULIN, J. 1991. Influence of pollution on parasites of aquatic animals. Advances in Parasitology, 30:201-238.

KLEMM, D.J., STOBER, Q.J. & LAZORCHAK, J.M. 1992. Fish field and laboratory methods for evaluating the biological integrity of surface water. Environment Monitoring Systems Laboratory, Office of Research and Development, United States of America (Report No. EPA/600/R-92/111), p239-88.

KÖRTING, W. 1974. Bothriocephalosis of the carp. Veterinary Medical Review, 2(74):165-171.

References

LE ROUX, L.E. 2001. Aspects of the morphology and the ecology of a Paradiplozoon species from Barbus aeneus in the Vaal Dam, South Africa. Unpublished M.Sc. Rand Afrikaans University.

LEIGHTON, F.A. 1995. Pathogens and disease, in Handbook of eco-toxicology edited by D.J. Hoffman, B.A. Rattner, G.A. Burton (Jr). & J. Cairns (Jr). United States of America: Florida: CRC Press: Lewis publishers, 509-518.

LESOTHO HIGHLANDS WATER PROJECT. 1995. Overview of the Lesotho Highlands Water Project. New Civil Engineer, September: 20-29.

LOVE, R.M. 1980. The chemical biology of fishes, Volume 2: Advances 1968-1977. London: Academic Press.

LUUS-POWELL, W.J. 1997. Evaluation of a health assessment index with reference to bioaccumulation of metals in Labeo species and aspects of the morphology of Chonopeltis victori. Unpublished M.Sc. dissertation, Rand Afrikaans University.

LYONS, J., WANG, L. & SIMONSON, T.D. 1996. Development and validation of an index of biotic integrity for coldwater streams in Wisconsin. North American Journal of Fisheries Management, 16(2):241-256.

MAKEIG, K.S. 1991. Regulatory mandates for controls on groundwater monitoring, in Practical handbook of groundwater monitoring edited by D.M. Nielsen. Chelsea: Lewis.

MARCOGLIESE, D.J. & CONE, D.K. 1997. Parasite communities as indicators of ecosystem stress edited by I. Paperna. Parasitologia, 39(3):227-232.

MARCOGLIESE, D.J. & ESCH, G.W. 1989. Experimental and natural infection of planktonic and benthic copepods by the Asian tapeworm, Bothriocephalus acheilognathi. Proceedings of the Helminthology Society of Washington, 56:151-155.

MARGOLIS, L., ESCH, G.W., HOLMES, J.C., KURTIS, A.M. & SCHAD, G.A. 1982. The use of ecological terms in parasitology (Report of an ad hoc committee of the American Society of Parasitologists). Journal of Parasitology, 68(1):131-133.

MARTIN, E.A. 1976. A dictionary of life sciences. London: Macmillan.

MARX, H.G, DYSON, L.L. & VAN HEERDEN, J. 2003. Verification of rainfall forecasts for the Vaal Dam catchment for the summer rainfall seasons of 1994 to 1998. Water SA, 29(2):195-200.

MARX, H.M. 1996. Evaluation of a Health Assessment Index with reference to metal bioaccumulation in Clarias gariepinus and aspects of the biology of the parasite Lamproglena clariae. Unpublished M.Sc. thesis, Rand Afrikaans University.

References

MASHEGO, S.N. 1982. A seasonal investigation of the helminth parasites of Barbus species in water bodies in Lebowa and Venda, South Africa. Ph.D. thesis, University of the North, Sovenga, South Africa.

MASON, C.F. 1991. Biology of freshwater pollution. (2nd ed) England: Longman.

MILLER, G.T. 1999. Environmental science: Working with the earth. (7th ed) California: Wadsworth.

NICKANOR, N., REYNECKE, D.P., AVENANT-OLDEWAGE, A. & MASHEGO, S.N. 2002. A comparative study of stomach and intestine contents in Barbus aeneus and Barbus kimberleyensis in the Vaal Dam to clarify variance in tapeworm infestation. Journal of South African Veterinary Association, 73(3):142-159.

NOVOTNY, J.F. & BEEMAN, J.W. 1990. Use of a fish condition profile assessing the health and condition of juvenile Chinook salmon. The Progressive Fish Culturist, 52:967-973.

O’KEEFE, J.H., UYS, M. & BRUTON, M.N. 1992. Freshwater systems, in Environmental management in South Africa edited by R.F. Fuggle, & M.A. Rabie. Pretoria: Juta: 277-315.

OFFICE OF THE PRIME MINISTER. 1982. Vaal River complex: guide plan. Pretoria: Government Printer.

OVERSTREET, R.M. 1982. Abiotic factors affecting marine parasitism in Parasites – their world and ours. Volume II. edited by D.F. Mettrick and S.S. Desser. Plenary and discipline lectures and abstracts, 5th International Congress of Parasitology, Toronto, Canada 7-14 August 1982, under the auspices of the World Federation of Parasitologists. Toronto, Canada, pp36-39.

PAPERNA, I. 1996. Parasites, infections and diseases of fish in Africa – An update. (CIFA Technical Paper, no.31). Rome: FAO.

PARK, C. 1997. The Environment: principles and applications. London: Routledge.

PIETERSE, B. 1997. SASAQS Newsletter Spring 1997. North-West regional report back. http://www.dwaf.gov.za/iwqs/sasaqs/nl/1997/nw.htm

POOL, D. 1984. A scanning electron microscope study of the life cycle of Bothriocephalus acheilognathi Yamaguti, 1934. Journal of Fish Biology, 25(3):361-364.

POOL, D.W. & CHUBB, J.C. 1985. A critical scanning electron microscope study of the scolex of Bothriocephalus acheilognathi Yamaguti 1934, with a review of the taxonomic history of the genus Bothriocephalus parasitizing cyprinid fish. Systematic Parasitology, 7:199-211.

POOL, D.W. 1987. A note on the synonymy of Bothriocephalus acheilognathi Yamaguti 1934, B. aegyptiacus Ryšavý and Moravec, 1975 and B. kivuensis Baer and Fain, 1958. Parasitology Research, 73:146-150.

References

POOL, D.W. 1988. An experimental study of the biology of Bothriocephalus acheilognathi Yamaguti 1934 (Cestoda: Pseadophyllidea). Abstract of Thesis, University of Liverpool.

POULIN, R. 1998. Evolutionary ecology of parasites: From individuals to communities. London: Chapman and Hall.

PRIEMER, J. 1987. On the life cycle of Proteocephalus exigus (Cestoda) from Salmo giardneri (Pisces). Helminthologia, 24:75-85.

ROBINSON, A.T, HINES, P.P, SORENSEN, J.A. & BRYAN, S.D 1998. Parasites and fish health in a desert stream, and management implications for two endangered fishes. North American Journal of Fisheries Management, 18(3):599-608.

ROBINSON, J. 1996. Evaluation of a health assessment index with reference to bioaccumulation of metals in Oreochromis mossambicus (Peters, 1852) and aspects of the morphology of Lernaea cyprinacea, Linnaeus, 1758. Unpublished M.Sc. dissertation, Rand Afrikaans University.

ROEBUCK, B.D. 1992. Heavy metals, in A primer of environmental toxicology, edited by R.P. Smith. Philadelphia: Lea and Febiger.

ROUX, D.J. 1993. Role of biological monitoring in water quality assessment and a case study on the Crocodile River, Eastern Transvaal. Unpublished M.Sc. thesis, Rand Afrikaans University.

ROUX, D.J., VAN VLIET, H.R. & VAN VEELEN, M. 1993. Towards integrated water quality monitoring: assessment of ecosystem health. Water SA, 19(4):275-280.

SCHOLZ, T. 1999. Parasites in cultured and feral fish. Veterinary Parasitology, 84:317-335.

SCHULZE, B.R. 1965. Climate of South Africa, Part 8, General Survey. Weather Bureau: Department of Environment Affairs.

SKELTON, P. 1993. A complete guide to the freshwater fishes of Southern Africa. Halfway House: Southern Book Publishers.

SKELTON, P. 2001. A complete guide to the freshwater fishes of Southern Africa (2nd ed). Halfway House: Southern Book Publishers.

THE KRIB. 2002. Water chemistry. www.thekrib.com

TINDALL, B. 1989. Fish pathology. (2nd ed), edited by R.J. Roberts. Great Britain: Mackays of Letchworth.

References

VAN AS, J.G., SCHOONBEE, H.J. & BRANDT, F. DE W. 1981. Further records of the occurrence of Bothriocephalus (Cestoda: Pseudophyllidea) in the Transvaal. South African Journal of Science, 77(8):343.

WATER ON THE WEB. 2003. Monitoring Minnesota Lakes on the internet and training water science technicians for the future. http://wow.nrri.umn.edu

WATER RESEARCH COMMISSION. 1994. Surface water resources of South Africa 1990. Volume II: Drainage Region C: Vaal. WRC Report No. 298/2.2.94.

WATER WATCH. 2003. http://www.state.ky.us/nrepc/water/wwhomepg.htm

WATSON, R. 2001. The evaluation of a Fish Health Assessment Index as a biomonitoring tool for heavy metal contamination in the Olifants River catchment area. Unpublished Ph.D. thesis, Rand Afrikaans University.

WETZEL, R.G. 1975. Limnology. W.B. Saunders Company.

WILLIAMS, H. & JONES, A. 1994. Parasitic worms of fish. London: Taylor and Francis.