Journal of the Royal Society of Western Australia, 93: 147-151, 2010 Invasive potential of a South-American fish species, Geophagus brasiliensis, in the Swan River, Western Australia: based on tolerance to instantaneous and gradual changes in salinity M de Graaf1,2 & T Coutts Western Australia Marine & Fisheries Research Laboratories, PO Box 20, North Beach, WA 6920 ' Current address: Wageningen Imares, PO Box 68, 1970 AB IJmuiden, the Netherlands 2 Corresponding author El [email protected] Manuscript received August 2009; accepted May 2010 Abstract The south-west of Western Australia is a biodiversity hotspot and has a high proportion of endemic freshwater fishes. None of the native fish species are primary piscivores and with the exception of the freshwater cobbler (Tandanus bostocki) all species are small (< 200 mm total length (TL)). The introduction of non-native freshwater fish species is considered one of the most damaging threats to this region's native fish diversity. Recently a new, large (maximum size 300 mm total length) feral fish species, the South American Pearl Cichlid (Geophagus brasiliensis), was found in Bennett Brook, a small tributary of the Swan River. The aim of this study was to determine the salinity tolerance of the feral Pearl Cichlid in order to predict it's invasive potential in watercourses of the Swan River catchment. Geophagus brasiliensis tolerated direct transfer from fresh water to 18-27 PPT with no mortality but more importantly, G. brasiliensis was able to resist gradual transference from freshwater to salt water (36 PPT) with very low mortalities. Therefore, ongoing control efforts are required in order to limit/prevent the invasion of this feral species into the Swan and other catchments in the southwest of Western Australia. Keywords: aquatic biodiversity, freshwater fish, introduced species Introduction Silver Perch, Bidyanus bidyanus; Eastern Mosquitofish, Gambusia holbrooki; One-spot Livebearer, Phalloceros Biodiversity 'hotspots' are areas of exceptional caudimaculatus; Goldfish Carassius auratus; Carp, Cyprinus concentrations of endemic species undergoing carpio; Rosy Barb, Puntius conchonius) are found in the exceptional loss of habitat (Myers et al. 2000). Of the 34 south-west of Western Australia (Morgan et al. 2004; currently recognized biodiversity hotspots in the world Maddern 2008; DL Morgan and SJ Beatty personal only one is situated in Australia; the south-west corner of communications). Human-mediated translocation [e.g. Western Australia. The south-west of Western Australia ongoing stocking of trout species for recreational fishing; has the highest proportion of endemic freshwater fishes biological control (Eastern Mosquitofish); aquaculture (80%) of all of the major Australian Drainage Divisions. escapees (Silver Perch); ornamental escapes (Goldfish, Ten species of native freshwater fish occur in south¬ Koi Carp, Rosy Barb)], either deliberate or accidental, is western Australia, eight of which are endemic to the the major vector of introduction (Morgan et al. 2004). region (Morgan et al. 1998; Allen et al. 2002). None of the native fish species are large piscivorous predators; only Recently a new feral species (ornamental escape), the the Freshwater Cobbler (Tandanus bostocki) typically South American Pearl Cichlid (Geophagus brasiliensis) attains a maximum size greater than 200 mm total length (Figure 1), was recorded in Bennett Brook (Figure 2), a (TL) (Morgan et al. 1998; Allen et al. 2002). tributary of the Swan River. The coastal drainages of Uruguay (Rio da Prata) and south-eastern Brazil Invasions of non-native species are increasingly (Amazon Basin) are the natural habitat of G. brasiliensis, a recognised as one of the most damaging threats to secondary-division freshwater fish (Axelrod & Schultz biodiversity, especially the introduction of freshwater 1955; Lowe-McConnell 1991). Primary-division fish species (Kolar & Lodge 2002). At present nine species freshwater families are strictly confined to freshwater, of non-native fish (Rainbow Trout, Oncorhynchus mykiss; whereas secondary-division freshwater families are Brown Trout, Salmo trutta; Redfin Perch, Perea fluviatilus; generally restricted to freshwater but may occasionally enter salt water. Geophagus brasiliensis is known to occur © Royal Society of Western Australia 2010 in brackish water (14 PPT; Mazzoni & Iglesias-Rios 2002) 147 Journal of the Royal Society of Western Australia, 93(3), September 2010 Figure 1. Pearl Cichlid, Geophagus brasiliensis, captured in Bennet Brook. Snfouni^"0" °f Be"net br°°k ^ ‘he Swan RiV6r CatChment- Red diamond (cid:8482)rkers indicate locations where G. brasiliensis has 148 de Graaf & Coutts: Invasive potential of Geophagus brasiliensis and it may be expected that this species would tolerate Progressive Salinity Increase Tolerance Test elevated levels of salinity. Fish were captured in the wild using seine nets and The Swan River at the confluence with Bennett Brook acclimatized to laboratory conditions in the 72 L tanks is brackish for most of the year (Figure 2). Therefore, this described above. After two weeks, 18 fish (avg. 92.6 mm species may be able to tolerate salinities that occur in the SL, 5.5 95% Cl) were individually placed in 7 L Swan River. The aim of this study was to determine the containers. Six containers were randomly assigned as salinity tolerance of this population of Pearl Cichlid in controls (0 PPT), while salinity in the remaining 12 order to predict its invasive potential in watercourses of containers was gradually increased from 0 to 36 PPT over the Swan River catchment. a nine-day period (4.5 PPT per day). After the initial nine day period, salinity levels were retained at 36 PPT for an additional 15 days. Water in all 18 containers was Methods replaced daily and fish were fed commercial pellets twice a week. The surviving fish in 36 PPT were all transferred to one large tank (0.5 x 0. 5 x 1.2 m) for a further 26 days Instantaneous Salinity Increase Tolerance Tests before the experiment was concluded. Experimental 72 L aquaria (0.3 x 0.4 x 0.6 m) were partitioned in five equal sections using perforated plastic sheets. Pearl cichlids were captured from Bennett Brook Results using seine nets and placed individually in a section within experimental aquaria. Each aquarium was aerated In both instantaneous salinity increase experiments and 25% of its volume replaced three times per week. (Figures 4A,B), no mortality was observed up to 27 PPT. Fish were fed commercial fish pellets twice a week. Fish A direct transfer to 36 PPT caused mortality of all fish were acclimatised for two weeks in the experimental within 24 hours at 14°C and within six days at 21 °C. In aquaria prior to being subjected to salinity tolerance contrast, low mortality (< 10%) among specimens of G. trials. brasiliensis was observed when salinity was gradually The first trial was conducted at a water temperature of increased over a nine day period (Figure 4C). The -15 °C (avg. 14.8 °C, 0.62 95% Cl) using 60 fish (avg. 67.6 majority of fish survived at 36 PPT for a period of 41 mm Standard Length [SL], 4.3 95% Cl) divided over 12 days. experimental tanks. Treatments for the first trial consisted of 0 (control), 4.5, 9, 18 and 36 PPT. Aquaria Discussion were randomly assigned to each treatment and mortality was recorded daily. Geophagus brasiliensis from Bennett Brook tolerated The second trial was conducted at a water direct transfer from fresh water to 18-27 PPT with no temperature of -21 °C (20.7 °C, 0.16 95% Cl) using 64 fish mortality (Figures 4A,B). Fish transfer directly to (86.5 mm SL, 2.7 95% Cl) divided over 13 experimental seawater (36 PPT) did not survive in both instantaneous aquaria. Treatments for the second trial consisted of 0 trials but when conducted at 15°C most fish died within (control), 4.5, 9, 18, 27 and 36 PPT. Aquaria were 24 hours while in water of 21°C several individuals randomly assigned to each treatment and mortality was survived for six days. Compared to temperate fish recorded daily. species the neotropical G. brasiliensis has a low thermal tolerance (Tantin & Petersen 1985) and at 15°C the fish most likely suffer from both salinity and temperature stress. More importantly, this study clearly demonstrated that G. brasiliensis from Bennet Brook is able to resist gradual transference from fresh water to sea water and survive in sea water for a long period of time, i.e. up to at least 40 days) with very low mortalities (Figure 4C). Based on the results of this study it is highly unlikely that without effective management action, the feral G. brasiliensis will be confined to Bennett Brook. Geophagus brasiliensis is likely to spread throughout the interconnected watercourses of the upper Swan catchment. It is, however, unclear whether G. brasiliensis would be able leave the Swan system and to migrate along the coast and invade other river systems in south¬ western WA. The spread of G. brasiliensis may pose a threat to aquatic biodiversity for several reasons. In the first place, the relatively large G. brasiliensis is territorial and aggressive towards conspecifics and other fish species especially during the breeding season. Secondly, Month G. brasiliensis is an omnivore (de Moraes et al. 2004; Figure 3. Monthly changes in salinity (PPT) and water Figure 5, de Graaf unpublished data) and will compete temperature (°C) of the Swan River near the mouth of Bennett with most native fish species for the same food resources Brook between 1995-2007 (data source Swan River Trust). (Morgan et al. 1998). Furthermore, introductions of G. 149 Journal of the Royal Society of Western Australia, 93(3), September 2010 A) Salinity Tolerance (Instantaneous Trial) G. bras Mens is at -14 °C B) Salinity Tolerance (Instantaneous Trial) G. brasiiiensis at ~21 °C 0 50 100 150 200 250 300 Time (hours) C) Salinity Tolerance (Progressive Trial) G. brasiiiensis at-21 °C 150 de Graaf & Coutts: Invasive potential of Geophagus brasiliensis S Gastropod 0 Odonata Larvae (cid:9633) Diptera larvae (cid:9633) Detritus (cid:9633) Fish (cid:9632) Insect 0-49 50- 99 100-149 150-199 (n=7) (n=12) (n=9) (n=2) Size Class (mm Standard Length) Figure 5. Ontogenetic changes in the diet of Geophagus brasiliensis at Bennet Brook (May 2006). brasiliensis have been reported in Taiwan, USA, Kolar C S & Lodge D M 2002 Ecological predictions and risk Philippines (www.fishbase.org) and the Tweed River, assessment for alien fishes in North America. Science 298: 1233-1236. New South Wales, Australia. Lowe-McConnell R H 1991 Ecological studies in tropical fish Future research should focus on a) the reproductive communities. Cambridge University Press, Cambridge. biology to determine size-at-maturity, breeding period Maddern M G 2008 Distribution and spread of the introduced and its potential ability to reproduce in the Swan River, One-spot Livebearer Phalloceros caudimaculatus (Pisces: and b) the diet of G. brasiliensis to determine the potential Peociliidae) in southwestern Australia. Journal of the Royal impact on native fish through competition and/or Society of Western Australia 91: 229-235. predation. The results of this preliminary study clearly Mazzoni R & Iglesias-Rios R 2002 Environmentally-related life suggest that ongoing control efforts are required in order history variations in Geophagus brasiliensis. Journal of fish biology 61: 1606-1618. to limit/prevent the invasion of the species into the Swan catchment. Morgan D L, Gill H S & Potter I C 1998 Distribution, identification and biology of freshwater fishes in south¬ western Australia. Records of the Western Australian Acknowledgements: We would like to thank Roy Melville-Smith, Stephen Museum Supplement No. 56. Beatty, David Morgan for valuable comments on earlier drafts of the Morgan D L, Hambleton S J, Gill H S & Beatty S J 2002 manuscript. The present study was financially supported by the Swan- Distribution, biology and likely impacts of the introduced Canning Innovative Research Program of the Swan River Trust and the redfin perch (Perea fluviatilis) (Percidae) in Western Australia. Department of Fisheries WA. Marine & Freshwater Research 53: 1211-1221. Morgan D L, Gill H S, Maddern M G & Beatty S J 2004 Distribution and impacts of introduced freshwater fishes in References Western Australia. New Zealand Journal of Marine & Freshwater Research 38: 511-523. Allen G R, Midgley S H & Allen M 2002 Field guide to the Myers N, Mittermeier R A, Miltermeier C G, da Fonseca GAB freshwater fishes of Australia. Quality Press, Perth. & Kent J (2000) Biodiversity hotspots for conservation Axelrod H R & Schultz L P 1955 Handbook of tropical aquarium priorities. Nature 403: 853-858. fishes. MacGraw-Hill Book, New York. de Moraes M F P G, de Freitas Barbola I & Duboc L F 2004 Feeding habits and morphometry of digestive tracts of Geophagus brasiliensis (Osteichthyes, Ciehlidae), in a lagoon of High Tibagi River, Parana State, Brazil. Publ. UEPG Ci. Biol. Saude, Ponta Grossa 10: 37-45. 151