This article was downloaded by: [Mr O LF Weyl] On: 30 December 2014, At: 07:56 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

African Journal of Aquatic Science Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/taas20 Species-specific impact of introduced largemouth bass Micropterus salmoides in the Groot Marico Freshwater Ecosystem Priority Area, South Africa PK Kimberga, DJ Woodfordbc, H Rouxd & OLF Weylbc a Hydrocynus Consulting, Kyalami, South Africa b South African Institute for Aquatic Biodiversity (SAIAB), Grahamstown, South Africa c Centre for Invasion Biology, SAIAB, Grahamstown, South Africa d Department of Rural, Environment and Agricultural Development, North West Provincial Government, Mahikeng, South Africa Published online: 02 Dec 2014. Click for updates

To cite this article: PK Kimberg, DJ Woodford, H Roux & OLF Weyl (2014) Species-specific impact of introduced largemouth bass Micropterus salmoides in the Groot Marico Freshwater Ecosystem Priority Area, South Africa, African Journal of Aquatic Science, 39:4, 451-458, DOI: 10.2989/16085914.2014.976169 To link to this article: http://dx.doi.org/10.2989/16085914.2014.976169

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions African Journal of Aquatic Science 2014, 39(4): 451–458 Copyright © NISC (Pty) Ltd Printed in South Africa — All rights reserved AFRICAN JOURNAL OF AQUATIC SCIENCE ISSN 1608-5914 EISSN 1727-9364 http://dx.doi.org/10.2989/16085914.2014.976169

Species-specific impact of introduced largemouth bass Micropterus salmoides in the Groot Marico Freshwater Ecosystem Priority Area, South Africa

PK Kimberg1, DJ Woodford2,3*, H Roux4 and OLF Weyl2,3

1 Hydrocynus Consulting, Kyalami, South Africa 2 South African Institute for Aquatic Biodiversity (SAIAB), Grahamstown, South Africa 3 Centre for Invasion Biology, SAIAB, Grahamstown, South Africa 4 Department of Rural, Environment and Agricultural Development, North West Provincial Government, Mahikeng, South Africa * Corresponding author, e-mail: [email protected]

Largemouth bass Micropterus salmoides are among the world’s 100 worst invaders and negatively affect aquatic biodiversity in many regions worldwide. In South Africa there is a paucity of empirical studies describing their impacts. The impact of M. salmoides on the fish community in the Groot Marico River catchment, an otherwise near-pristine river ecosystem and a freshwater ecosystem priority area, was assessed from surveys conducted in 2012. Fish presence and abundance were enumerated using multiple survey techniques, and their association with key habitat variables and the presence or absence of M. salmoides were assessed. A total of 14 native fish species were recorded, besides introduced M. salmoides, which occupied the majority of the mainstem and several tributaries downstream of barriers to upstream movement. Canonical correspondence analysis showed that only one native species, the Marico barb Barbus motebensis, had a negative spatial association with M. salmoides. Assessment of relative distributions showed this species to be excluded from M. salmoides-invaded river reaches, whereas the other native species were not visibly affected by the invader. This species-specificity of the impact of M. salmoides indicates that their impacts in South African streams may be dependent on predator-naiveté of prey.

Keywords: Barbus motebensis, barriers, conservation, environmental guild, invasion, predation

Introduction

Invasive alien species present one of the greatest case studies. The species has been linked to the extirpation challenges to conservation in freshwater ecosystems, which of native fishes in the Swartkops River in the Eastern Cape are more threatened than terrestrial or marine ecosystems (Ellender et al. 2011) and to changes in their abundance (Dudgeon et al. 2006). Of the many invasive species that and behaviour in the Driehoeks River in the Western Cape cause harm, introduced sport fishes have had perhaps the (Shelton et al. 2008). Furthermore, M. salmoides was found Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 most profound negative impacts on freshwater fish conser- to alter invertebrate community composition by removing vation (Cambray 2003). The North American largemouth large, conspicuous insect taxa in the Wit River, Eastern bass Micropterus salmoides (Lacepède, 1802), listed as Cape (Weyl et al. 2010). one of the 100 worst invasive alien species in the world Micropterus salmoides has been present in the Groot (Lowe et al. 2000), was first introduced into South Africa Marico River, which forms the south-western headwaters for angling purposes in 1928, and has subsequently spread of the Limpopo River catchment, for at least two decades throughout the county (de Moor and Bruton 1988; van (Skelton et al. 1994). This river originates in the dolomitic Rensberg et al. 2011). While primarily a still-water species aquifer plateau region of the North West province and flows that thrives in impoundments (Warren 2009), M. salmoides through a variety of geomorphological features to its conflu- can also survive in rivers and low-salinity reaches of ence with the Crocodile River (King 1951). The perennial estuaries (Wasserman et al. 2011; Glover et al. 2013). tributaries of the Groot Marico River emanate from dolomitic Introduced M. salmoides have been found to affect native eyes, formed at contact zones between igneous rocks and aquatic fauna negatively through predation in Canada the dolomites of the Transvaal Supergroup (Eriksson et al. (Trumpickas et al. 2011), Portugal (Godinho and Ferreira 2006). Perennial flow from these springs is an important 2000), Japan (Takamura 2007) and Zimbabwe (Gratwicke ecological attribute of the upper reaches in this semi-arid and Marshall 2001). In South Africa, there is anecdotal catchment, forming aquatic refugia for endemic biota within evidence that the species has negatively affected local this system. The upper reaches flow in deeply incised biodiversity in the Cape Floristic Region catchments of the gorges that are relatively unimpacted and sheltered from Western and Eastern Cape provinces (de Moor and Bruton human disturbances such as intensive agricultural activities 1988; Tweddle et al. 2009), though there are few published (Grobler et al. 2007).

African Journal of Aquatic Science is co-published by NISC (Pty) Ltd and Taylor & Francis 452 Kimberg, Woodford, Roux and Weyl

Micropterus salmoides’ invasion of the catchment was were placed parallel to the flow in areas with submerged or probably the result of multiple independent introductions, overhead cover in the evening and retrieved the following particularly in the Marico Bosveld Dam, the first major morning. All fish captured were identified to species, impoundment on the river. The catchment upstream of measured to the nearest mm, and returned to the water. this dam is of significant conservation importance (Skelton Habitat parameters were recorded at each sampling reach et al. 1994; Smith-Adao et al. 2006; Nel et al. 2011), and after the capture and release of fish. For each sampling was identified as a freshwater ecosystem priority area reach, length and width were measured, with a minimum (FEPA) because of the free-flowing nature of the Groot of four transects being used to assess the latter. On each Marico River, its ecological category of A or B, and as a width transect, five points were randomly selected, where fish sanctuary for the Marico barb Barbus motebensis the depth and substrate type (bedrock, cobble, gravel, Steindachner, 1894, listed as Vulnerable in the International sand, vegetation) were noted. The locality of each site was Union for Conservation of Nature (IUCN) Red List (Nel et al. recorded using a Garmin GPS receiver, and the altitude of 2011). The catchment has historically supported up to 18 each site was later confirmed using a digital elevation model fish species (Kleynhans et al. 2007). (DEM) generated from 5 m contours obtained from the The invasion of M. salmoides in the Groot Marico River surveyor general department of South Africa. catchment is clearly a cause for concern, and represents a good opportunity to assess the impact of this species on Data analysis a native fish community in a near-pristine South African Overall prevalence of fish species within the Groot Marico ecosystem. In addition, the Groot Marico represents a catchment was assessed using frequency of occurrence semi-arid catchment in a substantially different region of the (FO), i.e. the percentage of sites at which a species was country to those where previous studies on bass invasions recorded. Fish distributions were described by means of have been performed. multivariate procedures using CANOCO version 4.5 (ter The objectives of this study were to assess the spatial Braak and Šmilauer 2002). Presence/absence data were extent of the bass invasion across the catchment, and to used to assess species distributions, as potential variability determine whether the presence of M. salmoides has in species and size selectivity across the three sampling altered distributions of species within native fish community. methods prevented comparison of relative abundances. All analyses of presence/absence data were performed Methods under the assumption of equal detectability of each species across all habitats. Field data collection Canonical correspondence analysis (CCA) was used to A total of 94 sites in the upper Groot Marico catchment assess variables or environmental data that could drive the were sampled during surveys conducted in March and distributions of individual fish species across the catchment November 2012. Sites were located in the mainstem of (ter Braak and Šmilauer 2002). The environmental data the Groot Marico River and in its tributaries the Kaaloog se used included: elevation (m asl); site length (m); average Loop, Draaifonteinspruit, Van Straatensvlei, Ribbokfontein, width (m); average depth (cm); surface area of site (m2); Bokkraal, Rietspruit and Polkadraaispruit (Figure 1). invasion status (M. salmoides present/absent); and habitat Fish sampling was conducted primarily by means of type (pool, riffle or run). Micropterus salmoides was single-pass electrofishing, which was employed in shallow removed from the fish data in order to avoid autocorrela- riffle and run areas in all surveyed streams. Electrofishing tion with the invasion status parameter. Species with low was performed using a Smith Root LR24 (200 V, 0.1 ms occurrences were also removed from the fish data prior to

Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 pulse duration, 120 Hz frequency), and a Samus 725G the CCA assessment, as incidental records such as these backpack electrofisher (0.3 ms pulse duration, 90 Hz can skew the community responses to environmental frequency) by two teams working in parallel, each working variables (ter Braak and Šmilauer 2002). from opposite banks towards the centre of the stream. Fish In the event that the CCA identified significant negative were captured using dip nets and a 1 m wide push-seine associations between of the presence and absence of net placed downstream of the electrofisher anode. individual fish species and invasion state, the relationships Electrofishing was conducted over a set area to obtain of fish species densities to invasion state and any potentially densities per m2 sampled. confounding environmental factors were assessed, using To complement the electrofishing, snorkel surveys and electrofishing and snorkel survey data separately. For each fyke nets were used in pool habitats too deep to electro- of these sampling methods, an estimate of density based fish. Snorkel transects were conducted following the on the number of fish enumerated per unit area, were used. method described by Ellender et al. (2011), whereby the The effect of bass presence on a species’ density was number of fish are enumerated during two consecutive tested for significance with a non-parametric Mann–Whitney snorkel passes and averaged to give an estimate for the U-test using the statistical package Statistica 12. number of fish present in the pool. Where high turbidity hampered snorkel surveys, fyke nets were set overnight as Results an additional method to assess fish species presence within a reach. These nets comprised either two fyke traps with Fish diversity and distributions circular mouth diameter 55 cm joined by an 8 m guide wing, A total of 14 indigenous and two alien fish species were or a single trap with D-shaped mouth 40 cm wide with a 3 m collected during the March and November 2012 surveys guide wing attached to the centre of the mouth. These nets (Table 1). Alien fish species were recorded at 17 sites, African Journal of Aquatic Science 2014, 39(4): 451–458 453

Limpopo River system

AFRICA See enlarged area

South Africa

SOUTH AFRICA

km

Figure 1: Map of the Groot Marico River catchment showing locations of sampling sites invaded and not invaded by Micropterus salmoides, and barriers to upstream fish including small barriers (waterfalls and causeways) and major impoundments. The two small on-channel barriers represent known upper limits of M. salmoides within the tributaries of the Groot Marico River

Table 1: Indigenous and introduced fish species recorded at 94 sites in the upper reaches of the Groot Marico catchment in March and November 2012 using electrofishing (EF), snorkel surveys (SN) and fyke nets (FN). Frequency of occurrence (FO%)  percentage of sites at which a species occurred within the dataset

Species Common name FO% (n  94) Detection method Amphilius uranoscopus Stargazer catfish 37.2 EF, SN, FN

Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 Barbus brevipinnis Shortfin barb 1.1 EF Barbus motebensis Marico barb 57.4 EF, SN, FN Barbus paludinosus Straightfin barb 9.6 EF, SN, FN Chetia flaviventris Canary kurper 2.1 EF Chiloglanis pretoriae Shortspine suckermouth 24.5 EF, SN Clarias gariepinus Sharptooth catfish 7.4 EF, FN Cyprinus carpio (alien) Common carp 3.2 EF, SN Labeo molybdinus Leaden labeo 4.3 EF Labeoba rbus marequensis Largescale yellowfish 63.8 EF, SN Labeobarbus polylepis Smallscale yellowfish 9.6 EF, SN Mesobola brevianalis River sardine 1.1 EF Micropterus salmoides (alien) Largemouth bass 17 EF, SN Oreochromis mossambicus Mozambique tilapia 3.2 EF, SN Pseudocrenilabrus philander Southern mouthbrooder 13.8 EF, SN, FN Tilapia sparrmanii Banded tilapia 23.4 EF, SN

whilst 77 sites were uninvaded. All species were detected 63.8%; Table 1). Incidental species included the shortfin by electrofishing, with the majority of species also being barb Barbus brevipinnis, which was collected only at a recorded in snorkel surveys or fyke net catches (Table 1). single site in the upper reaches of Kaaloog se Loop, as well The Lowveld largescale yellowfish Labeobarbus as canary kurper Chetia flaviventris, Mozambique tilapia marequensis was the most widespread fish species (FO  Oreochromis mossambicus and river sardine Mesobola 454 Kimberg, Woodford, Roux and Weyl

brevianalis, which were sampled only from mainstem Table 2: Statistical significance of habitat parameters driving sites using electrofishing (Table 1). Micropterus salmoides the presence/absence of indigenous fish species used in CCA was collected at 17% of the sampling sites, and its range ordination included sites on the Groot Marico River mainstem, the Polkadraaispruit, Kaaloog se Loop and Draaifonteinspruit Parameter p-value (Figure 1). On the Draaifonteinspruit M. salmoides occupied Width 0.002 the lowest 1 km of stream up to a road causeway (Figure 1) Invaded 0.002 that appeared to have prevented it from moving further Pool/riffle/run 0.002 upstream into the catchment. Micropterus salmoides were, Elevation 0.098 however, also present upstream of several manmade Site length 0.218 barriers such as the Twyfelspoort Dam and a large weir Depth 0.516 on the Groot Marico River downstream of a waterfall that marked the upper limit of bass on the mainstem (Figure 1). These findings suggest historical stocking between the LMOL weir and the waterfall. Common carp Cyprinus carpio were Invaded recorded at three sites, and appeared to be present in low 0.6 Stream width abundances in the catchment (Table 1). MBRE 0.4 Interactions with habitat and invasion state LPOL CGAR The statistical significance of the various habitat param eters 0.2 TSPA Habitat PPHI CPRE driving the presence/absence of indigenous fish species BPAU LMAR was investigated (Table 2). Based on this assessment, 0.0 AURA average width, invasion status (M. salmoides present/ CCA 2 BMOT absent) and habitat type (pool/riffle/run) parameters were í0.2 found to be statistically significant (p < 0.05) in terms of the presence/absence of indigenous fish species, whereas site í0.4 length and average depth were not significant (p > 0.05) and were excluded from the CCA ordination. Although the í0.6 p-value of the elevation (m asl) parameter exceeded 0.05 Elevation (Table 2), this factor was included in the assessment for í0.6 í0.4 í0.20.00.20.40.6 the insight it provided into the distribution of B. motebensis CCA 1 in particular. Based on the CCA ordination, leaden labeo Labeo Figure 2: Canonical correspondence analysis plot of correlations molybdinus and M. brevianalis were negatively correlated between fish community structure and habitat type (pool/riffle/ with elevation and positively correlated with the presence run), stream width, elevation and invasion status (presence/ of M. salmoides (Figure 2). Both species were recorded absence of bass). AURA  Amphilius uranoscopus; BMOT  only in the lower reaches of the Groot Marico River. Three Barbus motebensis; BPAU  Barbus paludinosus; CGAR  species, lowveld smallscale yellowfish Labeobarbus Clarias gariepinus; CPRE  Chiloglanis pretoriae; LMAR  polylepis, straightfin barb Barbus paludinosus and banded Labeobarbus marequensis; LMOL  Labeo molybdinus; LPOL  tilapia Tilapia sparrmanii, were positively correlated with Labeobarbus polylepis; MBRE  Mesobola brevianalis; PPHI  Pseudocrenilabrus philander; TSPA Tilapia sparrmanii Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 increased stream width (Figure 2), with the latter species having a significant positive association with pool habitats (Figure 3; 2  9.78; df  1, p  0.002). The prevalences of shortspine suckermouth Chiloglanis pretoriae (2  salmoides invasion. Although M. salmoides is typically 23.07; df  1, p <0.0001) and stargazer catfish Amphilius associated with pool habitats, the species was equally uranoscopus (2  25.44; df  1, p <0.0001) were signifi- prevalent in pool and riffle habitats (Figure 3). The cantly positively associated with riffle habitats (Figure 3) maximum recorded length of M. salmoides in the riffles was and negatively correlated with stream width (Figure 2). 170 mm (mean 121 mm), whereas the maximum observed Labeobarbus marequensis and southern mouthbrooder length in pools was 400 mm (mean 192 mm). Pseudocrenilabrus philander showed no correlation with Barbus motebensis was the only species that was any environmental variables (Figure 2), reflecting their positively correlated with elevation and strongly negatively ubiquitous distribution across the catchment. Sharptooth correlated with the presence of M. salmoides (Figure 2). catfish Clarias gariepinus was primarily recorded in pool Barbus motebensis was recorded in all the tributaries habitats in the invaded section of the Groot Marico River, except in the Polkadraaispruit. The Polkadraaispruit differs although it was also recorded with B. motebensis in the from the other tributaries in that it has a lower gradient upper reaches of the Van Straatensvlei tributary. All the across its length and is invaded by M. salmoides along abovementioned species were negatively correlated with its entire length (Figure 1). Barbus motebensis was also the elevation habitat parameter, confirming that these recorded in the Groot Marico River upstream of the invaded species characterise the fish communities in the lower reaches. Barbus motebensis densities at both electro- reaches of the catchment, which coincides with the M. fished and snorkelled sites were higher at mid-to-high African Journal of Aquatic Science 2014, 39(4): 451–458 455

(a) Chiloglanis pretoriae (b) Amphilius uranoscopus because the water bodies that these structures impound Present were sites of M. salmoides introduction. Given the global 60 a Absent a trend of impoundments facilitating fish invasions in aquatic landscapes (Johnson et al. 2008), the overall lack of on-channel impoundments in the upper Groot Marico basin 40 may have contributed to its largely invasion-free status. b b The majority of fish species did not appear to be negatively 20 associated with M. salmoides presence, with all but one species showing stronger correlations with habitat variables than with invasion state. It is important to note that these (c) Tilapia sparrmanii (d) Barbus paludinosus findings are contingent on the assumption that detection probability was equal for all species across all environmental 60 a gradients. As detectability of species can co-vary with habitat gradients, which in turn can lead to positive or negative bias 40 in observed habitat associations (Gu and Swihart 2004), any b species–habitat association where detection probability has 20 not been directly assessed should be viewed with caution. This is especially true of detectability along the axis of elevation, as higher turbidity in the lower reaches may have (e) Barbus motebensis (f) Micropterus salmoides hampered the ability to detect benthic or cryptic species

NUMBER OF SURVEYED SITES during snorkel surveys. Nonetheless, the majority of species 60 displayed neutral or positive association with low elevations, suggesting a limited bias toward detection of headwater- 40 specialists in the smaller, clearer upper reaches of the Groot Marico catchment. The Groot Marico River fish community is relatively 20 diverse, with 14 indigenous species recorded during the March and November 2012 surveys, compared to that in other South African stream fish communities that have Pool Riffle Pool Riffle displayed negative effects of M. salmoides. For example, HABITAT the Swartkops River had only four native species (Ellender et al. 2011), while the Driehoeks River, where Shelton et Figure 3: Prevalence of (a) Chiloglanis pretoriae, (b) Amphilius al. (2008) conducted their research, had only three. It is, uranoscopus, (c) Tilapia sparrmanii, (d) Barbus paludinosus, however, unlikely that species richness alone is responsible (e) Barbus motebensis and (f) Micropterus salmoides in two major habitat classes (pool, riffle) in the Groot Marico catchment in for the lack of community-wide impacts observed here, as 2012. Differing letters indicate species occurrence is significantly there are several international examples of fish communi- dependent on habitat type (2 test, p < 0.05) ties with equal or higher diversity where M. salmoides has significantly altered community structure (Takamura 2007). Given the ambiguous role of diversity in the vulnera- elevation sites than at lower elevations (Figure 4). Although bility of fish communities to M. salmoides, it may be more

Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 they appeared to occur at naturally lower densities at low pertinent to examine whether or not the species comprising elevations, when only sites below the maximum upstream the community respond more strongly to environmental elevation of M. salmoides occurrence were assessed (the factors than to M. salmoides presence. The relationships of ‘bass zone’), bass-occupied sites had significantly lower B. these fishes to their physical environment can be described motebensis densities than bass-free electrofished sites (U  in terms of environmental guilds (Welcomme et al. 2006). 134; replicates  23,36; p < 0.0001), whilst the species was The Groot Marico River fish community appears to be absent from all snorkelled pools containing bass (Figure 4). divided primarily into pool and riffle specialists in the upper reaches, and lowland river specialists in the lower reaches. Discussion Chiloglanis pretoriae and A. uranoscopus are clearly riffle specialists, given their spatial correlation with this habitat The extent of the M. salmoides invasion appears to have type, and conform to the rhithronic (headwater) riffle guild been limited to the Groot Marico mainstem and certain (Welcomme et al. 2006). Due to these fishes’ close associ- low-elevation tributaries by key barriers to upstream ation with riffle habitat they are likely to be inconspicuous movement. These barriers include both natural features prey for large, piscivorous M. salmoides individuals. In such as a large waterfall on the Groot Marico River, as southern Africa, M. salmoides becomes primarily pisciv- well as artificial structures such as the Draaifonteinspruit orous after attaining 200 mm FL (Weyl and Hecht 1999), causeway. An interesting trend in the catchment is that dam and specimens greater than this size were recorded only in walls, such as Twyfelspoort Dam on the Polkadraaispruit pools during this study. River, and other smaller on-channel weirs generally do not The majority of other fishes in the community, including L. represent upstream limits to M. salmoides, most probably marequensis, L. polylepis, L. molybdinus, B. paludinosus, 456 Kimberg, Woodford, Roux and Weyl

(a) All snorkelled sites (c) All snorkelled sites below 1 325 m amsl 0.12 6 0.10 Bass zone 0.08 4 0.06

) ) 0.04 í 2 í 0.02

(b) All electrofished sites (d) All electrofished sites below 1 325 m amsl 1.0 DENSITY (fish m DENSITY (fish m 0.3 0.8 Bass zone a 0.6 0.2

0.4 0.1 0.2 b 1150 1200 1250 1300 1350 1400 1450 1500 Not invaded Invaded ELEVATION (m) BASS INVASION STATUS

Figure 4: Mean Barbus motebensis densities along the elevation gradient, showing the maximum altitude at which M. salmoides was recorded in the Groot Marico catchment in 2012, with downstream reaches designated as the ‘bass zone’ at (a) snorkelled sites and (b) electrofished sites. Barbus motebensis densities at bass-invaded and non-invaded sites within the bass zone (below maximum elevation of M. salmoides records) at (c) snorkelled sites and (d) electrofished sites, indicating near-complementary distributions with bass. Differing letters indicate significant differences in densities between invaded and non-invaded habitats (Mann–Whitney U-test, p < 0.05). Error bars denote SE

C. gariepinus and P. philander, appear to represent densities were highest at middle altitudes, rather than in the eupotamonic (main stream) guild species, which share an headwaters. At sites below the maximum elevation of M. affinity for larger, deeper, slower-flowing habitats with M. salmoides, B. motebensis was almost completely absent salmoides, which in its native range conforms to the plesio- when M. salmoides was present. These patterns strongly

Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 potamonic (floodplain and lentic river environments) guild, indicate a direct negative effect of M. salmoides on B. restricted to large, low-gradient river systems and lakes motebensis densities and distributions, most probably as (Warren 2009). All of these species may be better adapted a result of direct predation. Although variation in detection to co-existence with predatory fish such as C. gariepinus probability may have positively biased the density–elevation and C. flaviventris, than headwater species, which in South relationship observed in electrofishing and snorkel survey Africa often have not co-evolved with piscivorous fish, data, the strong negative association observed between making them naïve to novel predators like M. salmoides B. motebensis densities and bass is unlikely to have been (Skelton 2002). affected by such bias. Our findings concur with those of Gratwicke and Marshall Barbus motebensis inhabits slow-flowing sections (2001), who found M. salmoides to have limited impact and shallow pools of small streams (Engelbrecht and on fish assemblages in Zimbabwean lowland rivers, apart Bills 2007), making it a rhithronic pool specialist typically from that on certain small cyprinids (Barbus spp.), which associated with the upper foothills of mountain streams. were noted to also be vulnerable to predation from native It is usually found in association with banks and marginal piscivores such as tigerfish Hydrocynus vittatus. vegetation (Engelbrecht and Bills 2007) which is also the Barbus motebensis is unique within the fish community preferred habitat for M. salmoides (Savino and Stein 1989). of the Groot Marico River in that it appears to have been This potentially makes this species sensitive to predation severely negatively affected by the presence of M. from M. salmoides. Such habitat-linked sensitivity is shared salmoides. While the presence of both M. salmoides and B. with several other small South African cyprinid species motebensis correlated with site elevation, the electrofishing that are negatively affected by bass invasions, including and snorkel survey data indicated that B. motebensis the fiery redfin Pseudobarbus phlegethon (Woodford et African Journal of Aquatic Science 2014, 39(4): 451–458 457

al. 2005) and the Eastern Cape redfin Pseudobarbus afer References (Ellender et al. 2011). It is possible that all these species are more vulnerable to predation from large piscivorous Cambray J. 2003. Impact on indigenous species biodiversity bass by their inhabiting the same microhabitats within the caused by the globalisation of alien recreational freshwater river (slow-flowing deeper areas), as well as lacking the fisheries. Hydrobiologia 500: 217–230. behavioural adaptations to avoid these introduced predators de Moor I, Bruton M. 1988. Atlas of alien and translocated indigenous aquatic animals in southern Africa. South African (Skelton 2002). The behavioural naiveté of small cyprinids National Scientific Programmes Report 144. Pretoria: Council for like B. motebensis in South African headwater streams to Scientific and Industrial Research. pursuit predators like bass is a key aspect of their vulner- Dudgeon D, Arthington AH, Gessner MO, Kawabata Z-I, Knowler ability, and merits more detailed study. DJ, Lévêque C, Naiman RJ, Prieur-Richard A-H, Soto D, Barbus motebensis is endemic to the Limpopo River Stiassny MLJ, Sullivan C. 2006. Freshwater biodiversity: system and is listed as a vulnerable species, thus requiring importance, threats, status and conservation challenges. conservation management plans and actions to determine Biological Reviews 81: 163–82. and prevent further threats to the species (Engelbrecht Ellender BR, Weyl OLF, Swartz ER. 2011. Invasion of a headwater and Bills 2007). The conservation status of B. motebensis stream by non-native fishes in the Swartkops River system, within the Groot Marico River FEPA requires more careful South Africa. African Zoology 46: 39–46. Engelbrecht J, Bills R. 2007. Barbus motebensis. In: IUCN 2013. assessment before practical conservation actions can be IUCN Red List of threatened species. Version 2013.2. Available implemented. In particular, the genetic variability of popula- at www.iucnredlist.org [accessed 05 June 2014] . tions across the catchment needs to be assessed, to Eriksson PG, Altermann W, Hartzer FJ. 2006. The Transvaal establish whether or not M. salmoides threatens a geneti- Supergroup and its precursors. In: Johnson MR, Anhaeusser cally unique stock at any locality. Furthermore, the status CR, Thomas RJ (eds), The geology of South Africa. and security of barriers to the spread of M. salmoides needs Johannesburg: Geological Society of South Africa, and Pretoria: to be confirmed, and assessments made as to whether Council for Geoscience. pp 237–260. any reaches currently invaded by M. salmoides would be Godinho FN, Ferreira MT. 2000. Composition of endemic fish practical for the application of eradication efforts. Criteria for assemblages in relation to exotic species and river regulation in such efforts would include low local diversity of native fishes, a temperate stream. Biological Invasions 2: 231–244. Gratwicke B, Marshall BE. 2001. The relationship between the practicality of the reach for deploying piscicides, and an exotic predators Micropterus salmoides and Serranochromis upstream source of native fishes including B. motebensis that robustus and native stream fishes in Zimbabwe. Journal of Fish could recolonise the reach. All these criteria existed on the Biology 58: 68–75. Rondegat River in the Western Cape, where the successful Glover DC, DeVries DR, Wright, RA. 2013. Growth of largemouth eradication of smallmouth bass M. dolomieu resulted in bass in a dynamic estuarine environment: an evaluation of the significant habitat gains for native fishes (Weyl et al. 2014). relative effects of salinity, diet, and temperature. Canadian Only by prioritising reaches of high conservation value Journal of Fisheries and Aquatic Sciences 70: 485–501. and low logistical requirements, can one hope to bring Grobler DF, Roux H, Koni S. 2007. Habitat integrity assessment the limited capacity of the conservation structures in the report of selected rivers of the North West province. DACE North West province to bear in improving the outlook for Report ACE199/06. Department of Agriculture Conservation and Environment, Mafikeng, South Africa. B. motebensis in the Groot Marico catchment. The availa- Gu W, Swihart RK. 2004. Absent or undetected? Effects of bility of spatial fish population data is a prerequisite for any non-detection of species occurrence on wildlife-habitat models. freshwater conservation actions. This study has added Biological Conservation 116: 195–203. significantly to the spatial fish population data, and has Johnson PT, Olden JD, Vander Zanden MJ. 2008. Dam invaders:

Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014 verified the importance of the study area for freshwater impoundments facilitate biological invasions into freshwaters. conservation endeavours, specifically in the middle to upper Frontiers in Ecology and the Environment 6: 357–363. reaches. Data emanating from this study will enable the King LC. 1951. South African scenery: a textbook of geomorphology furtherance of current conservation initiatives in the Groot (2nd edn). Edinburgh: Oliver and Boyd. Marico River catchment. Kleynhans CJ, Louw MD, Moolman J. 2007. Reference frequency of occurrence of fish species in South Africa. WRC Report No. Acknowledgements — Funding for this research was provided by TT331/08. Pretoria: Water Research Commission. the Water Research Commission (project K5-2039), with logistical Lowe S, Browne M, Boudjelas S, De Poorter M. 2000. 100 of the support from the South African Institute for Aquatic Biodiversity world’s worst invasive alien species: a selection from the global (SAIAB), Golder Associates Ltd and Department of Rural, invasive species database. Auckland: Invasive Species Specialist Environment and Agricultural Development (North West Provincial Group [ISSG], and Gland: World Conservation Union [IUCN]. Government). We thank J du Plessis, Riverstill Guest Farm, for Nel JL, Driver A, Strydom WF, Maherry A, Petersen C, Hill L, accommodation and the many landowners of the Groot Marico Roux DJ, Nienaber S, van Deventer H, Swartz E, Smith-Adao who allowed access to their streams. We thank KA van der Walt, LB. 2011. Atlas of freshwater ecosystem priority areas in South Rhodes University, T Msezane, Pretoria University, and A Basson Africa: maps to support sustainable development of water and D Kindler, Golder Associates, for assistance in the field. resources. WRC Report No. TT500/11. Pretoria: Water Research B Ellender is thanked for statistical analysis support. All fieldwork Commission. was performed under SAIAB animal ethics guidelines (SAIAB Savino JF, Stein RA. 1989. Behavior of fish predators and their 2012/03), and research permits were provided by the North West prey: habitat choice between open water and dense vegetation. Sub-directorate of Biodiversity Management and Conservation Environmental Biology of Fishes 24: 287–293. (permit No. 042 NW-lI). We thank two anonymous reviewers for Shelton J, Day J, Griffiths C. 2008. Influence of largemouth bass, their comments, which greatly improved this paper. Micropterus salmoides, on abundance and habitat selection of 458 Kimberg, Woodford, Roux and Weyl

Cape galaxias, Galaxias zebratus, in a mountain stream in the and distribution of freshwater biodiversity in southern Africa. Cape Floristic Region, South Africa. African Journal of Aquatic Gland: IUCN and Grahamstown: South African Institute for Science 33: 201–210. Aquatic Biodiversity. pp 21–37. Skelton ER, Ribbink AJ, Twentyman-Jones V. 1994. The van Rensburg B, Weyl O, Davies S, Van Wilgen N, Spear D, conservation of dolomitic ecosystems in the western Transvaal, Chimimba C, Peacock D. 2011. Invasive vertebrates of South South Africa. Final Report to the Department of Environment Africa. In: Pimentel D (ed.), Biological invasions: economic and Affairs, South Africa. JLB Smith Institute of Ichthyology, environmental costs of alien plant, animal, and microbe species Grahamstown. (2nd edn). Boca Raton, Florida: CRC Press. pp 326–378. Skelton PH. 2002. An overview of the challenges of conserving Wasserman R, Strydom N, Weyl O. 2011. Diet of largemouth bass, freshwater fishes in South Africa. In: Collares-Pereira MJ, Coelho Micropterus salmoides (Centrarchidae), an invasive alien in the MM and Cowx IG (eds), Conservation of freshwater fishes: lower reaches of an Eastern Cape river, South Africa. African options for the future. London: Fishing News Books. pp 221–236. Zoology 46: 378–386. Smith-Adao LB, Nel JL, Roux DJ, Schonegevel L, Hardwick D, Warren ML. 2009. Centrarchid identification and natural history. In: Maree G, Hill L, Roux H, Kleynhans CJ, Moolman J, Thirion C, Cooke SJ, Philipp DP (eds), Centrarchid fishes: diversity, biology Todd C. 2006. A systematic conservation plan for the freshwater and conservation. West Sussex: Wiley-Blackwell. pp 375–533. biodiversity of the Crocodile (West) and Marico Water Management Welcomme RL, Winemiller KO, Cowx IG. 2006. Fish environmental Area. CSIR Report CSIR/NRE/ECO/ER/2006/0133/C. Pretoria: guilds as a tool for assessment of ecological condition of rivers. CSIR Natural Resources and the Environment. River Research and Applications 22: 377–396. Takamura K. 2007. Performance as a fish predator of largemouth Weyl OLF, Finlayson B, Impson ND, Woodford DJ, Steinkjer J. 2014. bass [Micropterus salmoides (Lacepède)] invading Japanese Threatened endemic fishes in South Africa’s Cape Floristic Region: a freshwaters: a review. Ecological Research 22: 940–946. new beginning for the Rondegat River. Fisheries 39: 270–279. ter Braak CJF, Šmilauer P. 2002. CANOCO reference manual Weyl OLF, Hecht T. 1999. A successful population of largemouth and Canodraw for Windows user’s guide: software for canonical bass, Micropterus salmoides, in a subtropical lake in Mozambique. community ordination (version 4.5). Ithaca, New York: Environmental Biology of Fishes 54: 53–66. Microcomputer Power. Weyl PS, de Moor FC, Hill MP, Weyl OLF. 2010. The effect Trumpickas J, Mandrak NE, Ricciardi A. 2011. Nearshore fish of largemouth bass Micropterus salmoides on aquatic assemblages associated with introduced predatory fishes in lakes. macro-invertebrate communities in the Wit River, Eastern Cape, Aquatic Conservation: Marine and Freshwater Ecosystems 21: South Africa. African Journal of Aquatic Science 35: 273–281. 338–347. Woodford DJ, Impson ND, Day JA, Bills IR. 2005. The predatory Tweddle D, Bills I, Swartz E, Coetzer W, Da Costa L, Engelbrecht impact of invasive alien smallmouth bass, Micropterus dolomieu J, Cambray J, Marshall B, Impson N, Skelton PH, Darwall W, (Teleostei: Centrarchidae), on indigenous fishes in a Cape Smith K. 2009. The status and distribution of freshwater fishes. Floristic Region mountain stream. African Journal of Aquatic In: Darwall W, Smith K, Tweddle D, Skelton PH (eds), The status Science 30: 167–173. Downloaded by [Mr O LF Weyl] at 07:56 30 December 2014

Manuscript received 21 July 2014, revised 19 September 2014, accepted 30 September 2014 Associate Editor: AJ Booth