Envirctnmental Biologt of Fishes 62: 291-313,2001 . Q 2001 Kltover Academic Publishers. Printed in the Netherlands.

Milyeringa veritas (), a remarkably versatile cave fish from the arid tropics of northwestern Australia

William F. Humphreys Terrestrial Invertebrate Zoology, WesternAustralian Museum, Francis Street,Perth, WesternAustralia 6000, Aus t ralia (e - mail : bill.humphrey s @mus eum.w a. g ov. au )

Received13 April 2000 Accepted 10 January 2001

Key words: stableisotope, allozymes, physico-chemical environment, hydrogen sulphide, anoxia, food, anchialine

Synopsis

The blind cave gudgeon Mi\:eringa veritas is restricted to groundwaters of Cape Range and Barrow Island, northwesternAustralia. It occursin freshwatercaves and in seawaterin anchialinesystems. It is associatedwith the only other stygobiticcave vertebrate in Australia,the blind caveeel, Ophisternoncandidum, the world's longest 'tethyan' , and a diversestygofauna comprising lineages with tracksand widely disjunctdistributions, often from North Atlantic caves.The cave gudgeon inhabits a karst wetland developedin Miocene limestonesin an arid area.There is an almostcomplete lack of informationon the basicbiology of this cavefish, despiteit being listedas threatenedunder the WesternAustralian Wildlife ConservationAct. Allozyme frequenciesand distributions indicate significant population sub-structuringon the CapeRange peninsula such that the populations are essentiallyisolated geneticallysuggesting that more than one biological species is present.Furlher, they suggestthat the vicariantevents may havebeen associated with a seriesof eustaticlow sealevels.Analysis of intestinalcontents indicates that they are opporlunistic feeders,preying on stygofaunaand accidentalstrapped in the water, at least at the sites sampled which wereopen to the surface,a conclusionsupported by the resultsof stableisotope ratio analysis.The gudgeons are found in freshwater caves and throughout deep anchialine systemsin which they occur in vertically stratified water columns in which there is a polymodal distribution of water chemistries(temperature, pH, salinity, dissolved oxygen,redox, dissolved inorganic nitrogen series, hydrogen sulphide).

Introduction the blind cave gudgeon, veritas Whit- Iey, 1945 (Perciformes: Eleotridae) and the blind The arid tropical northwest of Australia is notable for cave eel, Ophisternon candidum (Mees, 1962) the diversity ofthe subterraneanfauna that inhabits the (Synbranchiformes:Synbranchidae) (Figure 1) which karstic Cape Range peninsulaand Barrow Island. It occursympatrically over much andpossibly all of their includes a terrestrial fauna with wet forest affinities ranges.The latter speciesis consideredby Romero & (Humphreys1993a,b,2000a,b), and an aquaticfauna Vanselow (2000b) and it will not be discussedfur- with both limnic and anchialinecomponents. The latter ther here.Globally, there are f-ewoccurrences of sym- includes many disjunct Tethyan lineages (Humphreys patric speciesof cave fishes (Thinds & Proudlove 1993a,2000b, Jaume & Humphreys2001, Jaume et al. 1986).Proudlove' (personalcommunication) records 2001) whose congenersor sister taxa inhabit caves 13 instancesof two speciesand three instances of three on either side of the North Atlantic. in the Caribbean species occurring in sympatry. Kuhajda & Mayden region and the CanaryIslands (see below). lProudlove, G.S. 1997. A synopsis of the hypogean fishes of Included in this subterraneankarst wetland (ANCA the world. In: Proceedings of the I 2th Intemational Congress of 1996, Humphreys 2000b) are the only known sty- Speleology,La Chaaux de Fonds, Switzerland, August 1997, 3: gobiontic cave vertebrates in Australia. They are 351 -354. 298

(e.g.Wilkens 1988,Bergstrom 1991, review in Weber et al. 1998b),little is known of those inhabiting the Cape Range peninsula.This is partly becauseeven the general affinities of M. veritas are unknown, pre- cluding the interestingquestions that can be posed about the evolution of those cave fishes that are part of lineageswith extant epigeansister taxa. The identification of the phyletic affinities of M. veritas would be the most effectiveinformation to stimulate researchon the Australian cave fishes.Milyeringa ver- ilas is the only stygobitic speciesof eleotrid known in the world (Romero & Vanselow2000a) and occu- pies a wide range of physico-chemicalenvironments (Humphreys 1999b).While no researchhas focused on these Australian cave fishes, a variety of data has been collectedin associationwith other projects and a synthesis of this information is presentedhere to stimulate and facilitate researchon this versatile species.

History and affinities

Mi$,eringa veritas was described from subterranean watersof the CapeRange peninsula in the monotypic family (Gobioidea) (Whitley 19215). Mees (1962) recognized the eleotrid afflnities of Milyeringa, a tamily widespreadin tropical and sub- tropical shallow marine to fresh waters mainly in the Indo-Paciflc region (Nelson 1984). The mono- typic genusMilyeringa is endemicto the CapeRange peninsula and Barrow Island but its phylogenetic relationships within the Eleotridae have not been established.Milyeringa yeritas's closest affinity is possibly with the eleotrid Bafis, a mangrove dwelling genus of brackish and fresh waters of the Indo-AustralianArchipelago (i.e. Malay Peninsulato northern Australia including Indonesia, Philippines and most of Melanesia) (G.R. Allen personalcom- Figure l. a- Mill,eringateri.tusfrom KuburaWell (photo: Dou- munication).Mangroves are still presenton the Cape glasElford, WesternAustralian Museum):b -M. r,erltaslbraging Rage peninsula although they were formerly more over various sulphur bacteria in Bundera Sinkhole, an anchia- widespread(Morse I 993). line cave (photo: Stefan Eberhard, Western Australian Museum); On considerationofthe geologicaland tectonichis- c - Ophisternon ccutclidum liom Kubura Well (photo: Douglas Elford. Westem Australian Museum). tory of the area Humphreys & Adams (1991) sug- gested that a southern extension of the range of M. veritas was unlikely owing to the salinity of 'it (2001) report the first known occurrenceof sympatry the groundwater but that may be worth search- betweenNorth Americancave fish species. ing wider afield for this fauna ...' to the northeast While some cave fishes, parlicularly the Mexi- (seealso Humphreys 1993a,b). In 1996,elements of can characid As4,s1,ro,fasciatus, and the Amerrcan the tethyan crustaceanfauna were fbund as predicted Amblyopsidae, have been extensively researched where the Precambrian plateau (the Western Shield, 299

o12

a 1z o o21

tro 1

Fi gure 2. The distribution of blind cave fishes in Australia. Filled circles denole Ophiste rnon candidum. Milyeringu yeritus is present at all locations denoted by open or filied circles except location zl (Tantabiddi Well). Top ofthe map is north.

Beard 1998)approaches the shoreline to the southof Regional context Karatha. and in 1999M. veritas itself was collected from deepgroundwater on Barow Island (Humphreys Distribtttion 1999c). Barow Island lies on the shallowNorth West Shelf The known distribution of M. veritas is shown in (Figure 2) which, until the early Holocene,ca. 8000 Figure 2 and details of the sites and the history of years ago, would have beenconnected by land to the their discoveryare given in Humphreys(1999c). The Cape Range peninsulaowing to the generallylower species' range encompassesthe coastal plain from sea level over much the last million years (Chap- Exmouth to JavisWell (location 15), a distancealong pell & Thom 1977). Hence, for much of that time there the coast of about 136km. and one site on Barrow would havebeen a broad karstic plain, encompassing Islandwhich lies I 60 km to thenortheast of North West the North West Shelf from the tip of the peninsula Capeacross the shallowNorth West Shelf. to the FortescueRiver and including Barrow Island, There has been extensivesampling for stygofauna that would have allowed widespread dispersal of the along the coast of both the mainland and the Cape subter:raneanfauna (Humphreys1993b). Indeed, both Rangepeninsula, as well asfrom Banow Island2(Pesce terrestrial (Harvey & Humphreys 1995) and aquatic et al. 1996, figure 29, W.F. Humphreysunpublished (Adams & Humphreys1993, Humphreys 2000a) con- specific subterraneanfauna occur on Barrow Island 'Humphreys, WF. 1994. The subteranean fauna of the Cape and the northeastcoast of the Cape Rangepeninsula Range coastal plain, northwestem Australia. Repoil to the (Humphreys1993b). Australian Heritage Commission and the Westem Australian Her- itage Committee. 202 pp. 300

data) and the species appears to be endemic to the the freshwateraquifers through the surfacekarst which Cape Range and Barrow Island karst formed on anti- affectsthe elevationof the groundwatersurface for pro- clines of Tertiary limestones(Wyrwoll et al. 1993, longed periods.Temperatures are high (mean 27'C) McNamara & Kendrick 1994). However,M. veritas and the mean monthly maximum temperatureexceeds has rarely, and O. candidum has never been col- 35"C for four monthsof the year. lected using methods not involving direct sighting On the coastalplain the vegetationis naturally dom- prior to capture, hence the mainland sampling, which inated by tussock grasseswith eucalypt shrubs and was conductedin boresand pastoralwells using haul Flcus associatedwith caves but which occurs more nets and traps without visual contact, could have widely on the karst. Further details on the flora and missedthem. It is notable that other elementsof the vegetationassociations of the rangecan be found else- Cape Range and Barrow Island stygofauna do occur wherea(Beard 1975,Keighery & Gibson 1993). on the mainland in the calcrete and alluvial fans near the mouths of the Robe and Fortescue rivers Regionalgroundwater (Humphreys 1999a). One tethyan element, the ther- mosbaenaceanHalosbanerz, has even been recorded The general hydrogeological model for oceanic from an altitude of 300 m in the Robe Valley, the islands has been applied to the groundwater of approximate high sea level stand in the Late Eocene Cape Range peninsula (Allen 1993). Following the on the Yilgarn (G.W. Kendrick personal communica- Ghyben-Herzberg principle (Ford & Williams 1989), tion 1999). a freshwater lens overlies saltwater so that locally a wedge of saltwaterintrudes under the freshwater con- Geology and geomorphology tained in the limestone, and between which occurs a zone of mixing that broadenstowards the coast.In the The Cape Range peninsula comprises an anticline of Exmouth area, on the northeasternpart of the Cape marine limestonesof Late Eocene-Miocene agerising Range peninsula, the freshwater-saltwater transition, to an altitudeof 314 m. The peninsulais fringed by a at about5 km from the coast,is exceptionallyfar inland. coastalplain interruptedon the west coastby a series In the Exmouth area the inland limit to the salt-water of erosion terracescut into the Tertiary limestonesdur- interface appearsto be controlled by the presenceof ing sea level stadia and overlain in places by Qua- solution cavities and channels below the water table ternary carbonatesforming a karst pavement except (Maftin 1990).Such aquifers are vulnerable to saltwa- where they are overlain by Pleistoceneand Holocene ter upconing as a result of pumping water from the dunes(Wyrwoll et al. 1993).The westernshore is bor- thin freshwater lens, and to the inland migration of deredby the Ningaloo Reef, the longestfringing reef in seawater. Australia, and consideredan integral part of the Cape Rangekarst province3.Further information on the geol- ogy andhydrogeology may be soughtin Allen (1993). Habitat

Climate and vegetation General

The area is in the arid tropics with high evapora- Milyeringaveritas hasbeen recorded exclusively from tion (mean3219 mm) andreceives its rainfall (median inland karst systemsfrom within I 50 m of the coastand 280mm) asa resultof severalmeteorological processes up to 4.3 km inland. They are found in freshwatercaves (Beard 1975,Gentilli 1912,1979). The predominantly as well as in anchialineecosystems; the latter com- episodic heavy rainfall may occur through much of the prise fresh to brackish groundwatersoverlying seawa- year but with very low predictability (Humphreyset al. ter that hasno surfaceconnection to the seabut which is 1989)and it is responsiblefor the principalrecharge of affectedby marinetides (Sket 1981,1986, 1996,Stock

3Hamilton-Smith, aPayne, E., K. Kiernan & A. Spate. 1996. Karst A.L., P.J. Curry & G.F. Spencer. 1987. An inven- management considerations for the Cape Range karst province, tory and condition survey of rangelands in the Carnarvon Basin, Western Australia. W.A. Department of Environmental Protec- Westem Australia. Westem Australian Department of Agriculture tion, Perth. 80 pp. Technical Bulletin No. 73. 478 pp. 301

et al. 1986).Much of the habitatof M. veritasis known known from freshwatercaves (0.3 g l-r TDS: Poore& to be affectedby marine tides (Humphreys et al. 1999, Humphreys1992) but also inhabit the anchialinesys- figure 1) but the tidal influence is much more marked tem with salinities up to 34 g I I TDS (Yager & on the westcoast of the peninsula(W.F. Humphreys & Humphreys 1996, Humphreys 1999b) - they are R.D. Brooks unpublished). reported to be susceptiblein captivity to fungus infec- may be found swimming in tions (Saprolegnia) in freshwater (Young 1986). The sunlight at the surface in small karst windows in the fish occur in hyperoxic surface water to suboxic (see anchialine system and at depths of up to 33m in Sket 1996) watersbelow the pycnocline(Humphreys cavescontaining seawater (Humphreys 1999b). It has 1999b) and they have been seenforaging immediately beensampled from groundwatersthrough bore holes in abovesediments (Figure 1) coveredby black andwhite which the water table is up to 47 m below the ground sulphurbacteria. and remote from known openingsto the surface. At one location, Bundera Sinkhole,M. veritas inhab- its an anchialinecave with markedphysico-chemical and biological stratification(Figure 3). The fish occur P hys ic o -chemic al erwironment below multiple layers ofhydrogen sulphide and redox and oxygen minima associatedvariously with sulphur Humphreys& Adams ( 199 I ) andHumphreys2 reported bacteriaand a cascadeof dissolvedinorganic nitrogen the occurrence of M. veritas from a range of water series(Humphreys 1999b). While M. veritas inhabits chemistries(Table l) and subsequentlyfrom more waters with a remarkable range of physico-chemical salinewaters (Yager & Humphreys1996) with complex characteristics(Yager & Humphreys 1996), it is not physico-chemicalstratification (Humphreys 1999b, known how rapidly they move between these differ- Humphreys et al. 1999). M. veritas populations are ent water masses:the same individuals could move

Table I . Warcr analysis for the sites inhabited by Milyeringaveritas (.n : 20). The data for individual sites are contained in Humphreys2. All units are in mg 1-r unlessotherwise stated.

Parameter Mean SE Min Max CV pH 1.45 0.06 1.02 8.08 3 Turbidity NTU 3.83 1.00 0.6 16 116 Colour TCU 4.1 l.'7 034 162 ConductivitymS m ' 1295 301 35 4'700 106 TFS 8168 1930 250 26500 106 TFS-CO, 8013 1927 210 26500 107 Fe-unfiltered 0.20 0.08 0 1.6 185 Mn-unliltered 0.04 0.02 0 0.45 256 AI 0.12 0.03 0.02 0.6 119 Na 2352 58 948 7780 lL2 K 94.0 22.7 2 310 108 Ca 161 22.3 15 390 62 Mg 288 68.7 3.4 940 t07 Hardnessas CaCO3 1580 336 50 4840 95 Alkalinity 4.5 0.34 1.6 8.2 34 CI 4321 1088 54 15050 113 So+ 579 145 7 1950 I12 Si - asSiO2 13.8 0.9 2.8 23 29 Filterableorganic carbon 19.6 4.r r 74 93 Totalphosphorous 0.05 0.01 0.009 0.185 9l Freereactive phosphorous 0.026 0.007 0.004 0.13 121 Total Kjeldahl nitrogen 0.36 0.09 0.04 1.1 il3 Nitrite asnitrogen 0.02 0.009 0 0.19 202 Nitrate asnitrogen 3.2 1.54 0.02 3l.5 215 Ammonia asnitrogen 0.06 0.02 0 0.36 135 302

..t {

"=+

22 23 24 100 -50 0 150 Temperature RedoxmV

t . ...._-.._._r.l-< \.-. E \ \ \ I

26 Salinity

Fish Amphipods Remiped6 Shrimps Copepods Ostracods r.-. \ E E

(...... I t !l

O !o I ? c ta 40 DOmgFl Numbet of Milyeringa 4 6 Domgl-t

Figure 3. Vertical profile with depth in Bundera Sinkhole in 1997 where the grey bands denote the zones ofelevated hydrogen sulphide concentration (see Humphreys 1999b: Figure 2f): (a) temperature ('C), (b) salinity (g I I TDS). (c) dissolved oxygen (mg I I DO), (d) oxidation reduction (redox) potential (mV), (e) dissolved inorganic nitrogen species(DIN pM) N-NHa" N-NO2. N-NO3, (t) vertical distribution of sightings of the fauna in Bundera Sinkhole reported by divers [(a) the closed circles denote the vertical distribution of the taxon named at the head of the column, (b) the open circles denote the number of blind gudgeons, Mil.reringa velit,Is, counted at one occasion at each ofthree depths (secondbottom scale),(c) the curve denotesthe DO concentration(bottom scale)1.These data are illustrative only as no regular sampling has been conducted through time. Further details in Humphreys (1999b) and Humphreys et al. ( 1999). regularlyor rarely betweenthe water types,there may 1993a,2000a,Danielopol et al. 2000). Theseinclude be life interval movements or different stocks may the only member of the class Remipedia (Yager & inhabit the different water masses. Humphreys 1996) and the order Thermosbaenacea (Poore& Humphreys1992) known from the southern Associatedfauna hemispherewhich are congenericwith speciesoccur- ring in caveson either sideof the North Atlantic. Milyeringa veritas is part of a rich stygofaunalcom- In the general groundwater of the Cape Range munity that containsa number of ecogeographicrelicts peninsulaboth epigeanand stygal taxa cohabit with andphyletic relics with tethyanaffinities2 (Humphreys M.veritas.The stygalspecies i'1closbaenatulkiPoore & 303

Humphreys (Thermosbaenacea),Haptolana pholeta sp. nov. (Spionidae:Polychaeta), Halicyclops longi- Bruce & Humphreys (Isopoda: Cirolanidae) and furcatus Pesceet al. (Copepoda:Cyclopidae), ostra- Nedsia douglasi Barnard & Williams occur in cods,Iravadia sp.(Mollusca: Iravadiidae, Slack-Smith freshwater. Sryngiocaris lancifera Holthuis and 1993),water striders(Gerridae: Hemiptera) and Kief- S.stylifera Holthuis (Malacostraca:Atyidae) arefound ferulus intertinctusSkuse (Chironomidae, Diptera). A in freshwaterand anchialinesystems. Grandidierella fuller descriptionof the epigeanfauna associated with sp.nov., a non-stygalspecies, occurs in openanchialine thesekarst wetlandsis given elsewhere(Humphreys pools in which M. veritas may occasionallyfbrage. 2000a,b). A number of the sites with M. yeritas are also the Although no systematicsampling has been con- type localitiesfor other caverestricted species, some ducted. it is clear that both the distribution and of which are prey of the fish, suchas Thermosbaenacea abundanceof the fauna, including M. veritas, differs and atyid shrimps(Humphreys & Feinberg1995). betweenyears (Humphreys 1999b).While M. veritcts On Barrow Island M. veritas is known from a has been recordedthroughout the water column, the single bore that also contains calanoid, cyclopoid vertical distribution difl'ers markedly between years. and harpacticoid copepods including Inennipes In 1994,there was a large concentrationof the fish at humphreysiLee & Huys (Harpacticoida:Ameiridae), the extremedepth of the cave,while they were entirely H alosbaena tulki (Thermosbaenacea),the amphipods lacking from this areain 1997 when there were con- Nedsia sculptilis Bradbury & Williams (Melitidae) centrationsjust below the pycnocline(ca. 10m) and at and Bogidomma australis Bradbury & Williams ca. 20m, between the HlS layers (Figure 3f). This (Bogidiellidae),Haptolana pholeta (Cirolanidae)and changein location of the principle predator(Humphreys Stygiocaris sQlfe ra (Atyidae). & Feinberg1995). and of D. kornickerl,suggests that The vertical stratilication of the faunal associa- the ecosystemis variablethrough time. tions is best documented from Bundera Sinkhole Like M. veritas. other fishes with reduced or ves- because it is accessibleby divers to a depth of tigial eyes, which are sometimesfound in total dark- 33 m and M. veritas occurs through the full depth ness,are known to inhabit cavesexhibiting a density of the water column (Figure 31). Bundera Sinkhole interface(thermohalocline: sensu Humphreys 1999b). is the only deep anchialine system known in Aus- Speciesof the genus Lucifuga variously occur both tralia and the only known continentalanchialine sys- above (Cohen & McCosker 1998) and below (Diaz tem in the southern hemisphere (see Iliffe 2000, Perez 1987b) the density interface at depths of up to figure 1) as well as the only known site for the crus- 40m (Diaz Perez 1987b),in freshwater(Diaz Perez tacean class Remipedia in the southernhemisphere 1981a) or saltwater (34mg 1: Palmer 1985). They (Yager & Humphreys 1996). It is the most complex may alsooccur with relictualcrustaceans (Yager 198 1, habitat known to be inhabited by M. yeritas. Sev- Palmer 1985,Diaz Percz 1987a)such as Remipedia, eral taxa are known only from below the thermohalo- Thermosbaenaceaand Thaurnatocypris (Ostracoda: cline in the cave,namely Lasionectesexleyi Yager & Yager1981, DiazPerez 1987a). Humphreys (Remipedia), Danielopolino kornickeri Danielopol et al. 1999 (Ostracoda: Halocyprida: Biology Thaumatocyprididae), Liagoceradocus branchialis Bradbury & Williams (Crustacea: Hadziidae), Appearance and befuniour Stygiocefis st1'lifera, and the copepods Bunderin misophagnJaume & Humphreys, 2001 (Calanoida: Mil ,-eringctveritas is eyeless,depigmented and translu- Epacteriscida e'), S t 1n go c v-cI o p ia aus t r alis Jaumeet al., cent, and in yivo the brain is visible through its 2001 (Calanoida:Pseudocyclopiidae) and Speleophrict braincase,the scales are reduced on the body and bunderaeJaume et al.,2001 (Misophrioida: Speleophri- absentfrom the head,all indicativeof a high degreeof idae). troglomorphy. The general morphology of M. yeritas Above the thermohalocline in the photic zone is convergentwith other cave dwelling fish such as this anchialine system also contains a number of Amblyopsis rosae, the Ozark cave fish of Missouri non-stygalplants and (the algaeRhizoclonium (Bergstrom 1997). Although visible eyes are overtly tortuosum (Dillwyn) (Chlorophyta:Cladophoraceae.) lacking it is not known whether the eyes of M. veritas and Lamprothamniumpapulosum J. Groves (Charo- areentirely regressed as is suggestedby thebehavioural phyta: Characeae),and the invertebratesPrionospio observation that the fish in caves show no raoid 304 responseto suddenbright light; however,they some- Food and stable isotope ratio analysis times swim in sunlight in anchialine pools. The degreeof eye reduction in M. veritas does suggest Young (1986) reported that, in captivity, M. veritas ate that the lineage has been troglobitic for a long time mosquito larvae, live brine shrimps, and that eventu- since the degree of eye reduction may be consid- ally they would take frozen brine shrimps and would ered indicative of the time since troglobite evolu- take prey up to one-third of their body length. He tion (Wilkens et al. 1989). The latter correlatedthe observed that M. veritas often missed live prey on degreeof eye reduction in Lucifuga and Ogilbia with the surface film but rarely missed them when live the phylogenetic age of the lineages in the New prey were in the water column; they took live prey World tropicsas deducedfrom geologicaland eustatic by a suddengulp while nearly stationary and with the information. fins flared. When offered food they frequently ignored Milyeringa veritas often hangsmore or lessmotion- it and allowed tubificid worms to move over them less in the water or percheson rock ledgesfor long without overt response even though they eventually periods.When the water is disturbedthe fish may swrm ate them. Allen (1989) consideredM. veritas to feed directly to investigate the disturbance, gradually dis- opportunistically upon detritus, algae and whatever tance itself from the area, or swim to the bottom and organisms,including insects, that accidentallyfell into move behind obstacles,especially to move beneath the water. Examination of the intestinal contents of ledges.Despite the reportedsluggish nature of these existing collections showed that the diet of M. veritas fish (Knott 1993), they may exhibit strong escape included tenestrial (19Vo')and aquatic speciesand that behaviour from a net, are strong swimmers for short amongstthe aquaticspecies both stygobitic(107o) and distances,even leaping from the water and, if stranded, epigeanspecies were ingested(Humphreys & Feinberg can flip their bodies into the air. Young ( I 986) reported l 995). The predominantprey of terrestrialorigin con- 'aggression' that, in captivity,they displayed with the sumed by M. veritas differed between the east and 'typical gill flaring andbody waggleof a gudgeon'and west coast of the peninsula,reflecting the land use. that when familiar with their surroundings could be Exotic cockroaches(Periplaneta) and native isopods quite evasive. (Buddelundin)predominated in the diet of fish,respec- In shallow conduits of the anchialinesystem that tively, inhabiting the settled east coast and the largely drain at low tide, M. veritas may be seenclosely fol- unpopulatedwest coast (Humphreys & Feinberg1 995). lowing the advancing tidal front, apparently foraging Hence, the flsh appearto f'eedopportunistically on ter- over the newly submerged substrate,an obser"vation restrial invertebratesaccidentally introduced into the consistentwith the apparentopportunistic intake of aquaticsystem (Humphreys & Feinberg1995). Recent food revealedby analysisof the intestinal contents. observations (1999) of Buddelundia foraging under- Although M. veritas occur in particular parts of caves, water in Kubura Well meansthat their presencein water often in somenumbers (see below), they do not overtly is not necessarilyaccidental. M. veritctsalso includes group. They move widely though the water column in its diet aquatic larvae of aerial species (chirono- within a given cave but it is not known whether indi- mid and caddis larvae) as well as specialized mem- vidualscross the pycnocline. bersof the stygofauna(Stygiocaris sp.) (Humphreys & Feinberg1995). The bigger west coastfish were asso- ciated with bigger food items (isopods).However, as Numbers ter:restrial Crustacea generally have an energy den- sity only abott T}Vc that of terrestrial insects (from SeveralM. veritas are commonly seenin caves,but Cummins & Wuycheck l97l), then this apparent there are some siteswhere more than 100 may be seen advantagein food size may not confer an energetic (Humphreys1999b). Formal determination of thenum- advantage. ber of M. veritas at any location is limited to Kubura The gut contents examined by Humphreys & Well, actually an open cave that is connectedto an Feinberg(1995) were predominantlyfrom large indi- entirelyflooded chamber extending to a depthof 1l m. viduals all of which had been sampledfrom wells or The minimum size of the populationof M. veritas in opencaves where there was direct open access between KuburaWell was estimatedby dilution samplingto be the ground surface and the groundwater and hence 48 (95VoCI 45 and 52) and43 (957oCI 38 and 50) fish whereepigean prey specieswould bepresent. Through- in Juneand July 1993,respectively (Humphreysr.). out most (>99.99Vo\ of the inferred distribution of 305 this obligate groundwater community (stygofauna) sulphur bacteriavia stygal shrimps (Atyidae) scrap- there are no known openings to the surface. Exten- ing the biofilm in the anoxic/oxic ecotone in the sive samplingdown boreholeshas shown that stygo- anchialinesinkhole at Bundera(details in Humphreys fauna, including M. veritas, is presentin such areas, l 999b). from where there is a total absenceof epigeanspecies (Humphreys& Feinberg1995). Hence, in general,the Size,growth, breeding and condition fish are probably dependenton the stygofaunafor food rather than on epigeanspecies accidentally in the water. Milyeringa veritas ranged from 5 to 57 mm standard The fish share their habitat with a diverse stygo- length(SL) andthose collected from theeast coast were fauna, many of which were not representedin their bigger in every respectfrom those on the west coast gut contents(Humphreys & Feinberg1995), probably but the meristic variables do not differ between areas owing to the specimenshaving been sampledat sites (Humphreys1994, figure 13). For example,the mean open to the surfacewhere stygofaunawould havebeen SL of fish from thewest coast was 32.8mm (7.9 : SD; scarceand where prey of epigeanorigin was abundant. 125 : samplesize) compared with a mean elsewhere The smallesttaxa, such as copepodsand ostracods, of 24.6mm(10.4,23, F,r.r+a : 18.61,p < 0.001). may lie outside the size range of prey items, which Milyeringa veritas isolatedin a cavepool for 42 days for M. veritas is known to include the length range of grewin lengthat arateof 0.6mmpermonth (957olimits 2.8-14 mm, and for O. candidum3 mm (Halosbaena) 0.2-1.0mm) but during this time lost weight indicat- to an estimated 8 mm (Odonata). Melitid amphipods ing that their conditionsmay not havebeen optimal.2 are abundant and sympatric with M. veritas but they Thesesparse data on growth rate areconsistent with the were not identifled from the gut contentsdespite being apparentcohort size indicated from length frequency within the known size range of prey (Humphreys & distributions and suggest that the larger fish are 3* Feinberg1995). yearsold (Figure5). The prey identifiedis consistentwith the behaviour The only data availableon seasonalchange in condi- of the fish. M. veritas moves widely through the water tion in M. veritas are derived from museumspecimens. column, often hovering in mid- to surfacewaters where The condition index differs significantly through the prey with hygrophobic integuments,such as cock- year being higher in August and Septemberthan in roaches,would be encounteredwhen they fell into the other months (Table 2). There are inherent problems water. In contrast,O. candidum inhabits the surfaceof. with the use of such material,including preservation and burrows into. the flocculent faecal ooze character- artifactsand poor distribution in the samplesof the size- istic ofcrustacean-richstygal habitats, and all the prey classesand seasons. However, the results are ofinterest itemsidentified were bottom dwellersor, in the caseof assmall individuals (less than 18mm SL, cf. maximum the isopod,would sink to the bottom when it fell into 57 mm SL) havebeen observed in December-Marchin the water. severalyears (K. Cameronpersonal communication) To further examine the food of M. veritas stable suggestingthat there may be a distinct breedingsea- isotope ratio analysiswas conductedon elementsof son. The style of breedingin M. veritas is unknown. the stygofauna and surrounding habitats (methods in The smallestindividuals caught are4.5 mm SL. Humphreys 1999b). The locations of the 3Cl3 and Annual,even seasonal breeding cycles may be unex- 0N15 valuesand the trajectoryof the EC13and 0N15 pectedin tropical caveanimals. What could be the cue coordinatesbetween tropic levelsprovides information to seasonalbreeding in M. veritas suggestedby the 'cohorts' on the trophic level of the speciesas well as the ori- in Figure 5 and the apparentperiodic appear- gin of the food source for each trophic level (detailed ance of small individuals? The growth data and size in Humphreys1999b). class frequency analyseswere derived from M. yeritas The stable isotope signature(Figure 4) provides sampledfrom cavesand open pastoralwells and so strong evidence that M. veritas is a predator and they are not dependenton the stygal ecosystemas does not rely to any extent on detritus or algae as they gain most of their food directly from animals a food source as suggestedby Allen (1989). Fur- inhabiting terrestrial habitats (Humphreys & Feinberg thermore, the 013Cratios suggestthat they depend 1995). Such M. veritas populations,closely tied to on the energy derived from plants using the C4 surface openings, may be expected to respond dif- photosynthetic pathway, a mixture of C3 and C4 ferently to those populations inhabiting groundwaters plants, and from chemoautotrophicenergy fixation by more remote from surface openings and dependent 306

z o WC

-35 -30 -20 -15 a13C

Figure 4. Stable isotope analyses (013C and 015N) of Milyeringa veritcrs (dark shading: C2U-D and C28-S : deep and shallow samples,respectively. from Bundera Sinkhole, west coast of the Cape Range peninsula (Humphreys 1999b); WC : east coast of the Cape Range peninsula and some elementsof the associatedsystem. Light shading are plants with C3 and C4 photosyntheticpathways. Unshaded - Stlgiocaris stylifera from various sites (C28 : Bundera Sinkhole; WC : east coast of the Cape Range peninsula; 82. L8 and C65 : anode protection wells on Barrow Island; 81 : anchialine cave on Barrow Island). The location of Ihe M. verittts data appears to reflect the widely varying stable isotope signature of the largest prey, the atyid shrimp Srygiocaris. It has been argued from these and other data that in places the atyids depend on chemoautotrophic energy fixed by sulphur bacteria associated with the oxic/anoxic interface (Humphreys 1999b). Data for Stygiocaris and M. veritas fiom location WC are from Davies, PM. 1996. Stable isotope analyses of the food-rvebs associated with stygofauna of Exmouth borefield. A report to the Water Corporation Appendix I , 13 pp. In: Supplementary Investigations of the Effects of Public Water Supply Abstraction on the Stygofauna and Aquit'er of the Cape Range, Water Conroration. Perth.

= I (J

25 30 35 40 Standardlength (mm)

Figure 5. The pooled size class distribution of Mi$eringa veritas from Kubura Well in 1993. Arows denote the projected movement of 'cchort' the lower over successive years as estimated from the measured growth rate (see text) in Kubura Well. 301

Table 2. Condition index (standard length (mm)/ 3J Genetics weight(g)) tor Milyeringa veritas delermtned from museum soecimens. Seven populations of M. veritas, from throughout its Morrth Mean distributionon the Cape Rangepeninsula, were sub- jectedto a comprehensiveallozyme analysis at 43 loci May 38.2a 0.28 16 (Adams June 31.0a 0.36 19 & Humphreys1993), a synopsisof their find- July 38.3a 0.46 31 ings is presentedhere. Of the 43 loci scored,28% August 39.8b 0.49 at of loci were polymoryhic (P) and the averagelevel September 39.8b 0.81 l8 of heterozygosityper population (11) was 0.028 + October 38.l 0.60 5 0.013 (n : 7). Comparedwith marine fish, generally, The means differ significantly (Fs.',r 4.313, M. veritas exhibits low F1but high P: for 169 popu- 0.001; October samples excluded owing to small lations of marine fish, after removal of studiesinvolv- sample size). Rows with common letters are not signilicantly ing few individuals(<20), Planes(1998) found 11 : dift'erent by the Fisher PLSD test. The samples were fixed in 0.058+0.037and P : 21.37+14.097c. Planes (1998) formaldehyde and represent combined samples from several found no effect of taxonomic placement in marine years of collecting in the 1960s. fish on the magnitudeof H or P. Low valuesof F1 havesometimes been associated with smallpopulations on stygal prey species.Cave and well populations (Soul6 1976, Nevo et al. 1984), stableenvironments of M. veritas (and other stygofauna,such as Stygio- (Levins 1969, Van Valen 1965) or habitatspecialization caris spp.) may be expectedto respond closely to (Nevo 1978) associatedwith M. veritcts.However, as changesin food resourcesrather than to variation in no comparable allozyme data are available for other the stygal community itself. The episodicheavy rain- eleotrids,such comparisons are premature. fall associatedwith summer cyclones may serve to The averageNei's D for all populationswas 0.04, wash copious food into the caves, or else the humid and an averageof 2.37o, with a maximum of l7o caves may serve as a refuge for terrestrial inverte- of loci had fixed differences.The mean F51between bratesduring the very dry and hot (frequently > 40"C) populationswas 0.457 (957o CI : 0.211-0.599) summer period. Clearly, detailed work on the flsh is while the mean FIS betweenpopulations was 0.208 required. (95VoCI : 0.071-0.300).Whilst some fixed differ- enceswere encountered,the overall levels of genetic differentiation betweenpopulations and the geographic Morphometrics patternof alleledistribution were both consistentwith the presenceof a single biological species(Figure 7). The morphometric developmentof caveflsh haslargely However, analysis of the F-statistics demonstrates beenneglected other than the generalobservations that the presenceof significantpopulation substructuring they often have reducedbody size and larger head and in this species.The averagegene flow per genera- (Nn) lins (Poulson1964) andthat this may be a 'neotenic' tion can be determinedfrom the conditional trend (Weber et al. 1998a). However, no thorough averagefiequency, F51 (Slatkin 1981). In an island examination of the heterochrony(McNamara 1990) of population model Nra represents the average num- cave fish has been undertaken with respect of their ber of individuals exchangedbetween populations in lineages or consideredgenerally as a stygomorphy eachgeneration and populationsdo not divergeunless (cf. troglomoryhy: Christiansen 1962, Humphreys Nm < 1 (Wright 1931). Generally, if F51 2000c). (Nm < 1) considerabledivergence will occur between A synopsisof the morphologicaltrends in M. veritas populations,and with valuesof Fsr > 0.33(Nrr < 0.5) is presentedhere (Figure 6, Table 3) for comparison populations will be essentiallyunconnected geneti- with other speciesof cave fish. The general trends cally (Trexler 1988, Slatkin & Barton 1989).Hence, apparent in these data are that as the flsh grow they the population of M. veritas is not genetically uni- becomemore bulky, the tail becomesshofter, the body form and there is evidence of restricted gene flow on becomesdeeper and narrower and the buccal region the peninsula(Humphreys & Adams 1991,Adams & shofter.How this developmentpattern relatesto habits Humphreys1993), as is also the casein the sympatric is unknown. shrimp Stygiocaris stylifera. As some populations of 308

Figure 6. Morphometric measurements on Mill-eringa verltas (see Table 3) from specimens in the Western Australian Museum.

Table 3. Morphometric relationships: regressions of linear measurements and weight I of Mily-eringa,-erilz7s(N:148)summarizedinthematrixofslope(upperright)andrr(lower left) of regressions of log X (vertical) on log Y (horizontal) of variables Ll : standard length (size range examined was 5-57 mm); L2 : snout vent length: L3 : length 3 (see Figure 6); L4 : length 4; Wl : width 1; W2 : width 2;D1 : depth 1; D2 : depth 2. The slope is shown only when Ihe95Vo conlidence intervals for the slope do not include a slope of 1.0.

wr L1 L3 L4 D1 D2 Wl W2 Ll-L2 Wt 0.95 0.90 1.04 0.96 0.93 L1 0.98 0.94 1.06 1.05 1.08 L2 0.91 0.98 0.91 0.92 L3 0.96 0.97 0.95 1.09 l.l0 1.13 1.05 L4 0.97 0.95 0.95 0.91 0.94 0.91 0.91 0.87 Dt 0.98 0.93 0.92 0.92 0.89 0.91 0.91 0.87 D2 0.98 0.91 0.93 0.93 0.90 0.92 0.90 0.89 0.85 w1 0.94 0.87 0.87 0.86 0.84 0.89 0.88 0.90 0.85 w2 0.96 0.96 0.95 0.94 0.94 0.93 0.94 0.90 0.93 LI-L2 0.90 0.94 0.85 0.89 0.85 0.86 0.86 0.'79 0.93

M. veritas are genetically isolated from others, this populationson the west coast.The exceptionis the may suggestthat more than one biological speciesis geographicallyintermediate wellC-273 at thenorlhern present. end of the peninsula,which is genetically intermediate The genetic relationships amongst populations are betweenthe two groupings.This within-speciesgeo- displayedvisually in Figure 7. Two major clustersare graphic trend broadly mimics the distribution of the 'western' evident and these reflect a general versus two speciesof shrimp, which overlapin C-24, C-25, 'eastern' geographicsplit. PopulationsC-27 and C-361 and C-213 (Humphreys & Adams 1991. Adams & from the east coast of Cape Range peninsula are Humphreys 1993).The gudgeonsare the only group invariant at all loci examined, and display 5-iVa frxed examined in which the F-statistics suggestthere may differences when compared to all but one of the not be random mating occurring within subpopulations 309

:

zI3

{t

361 t3

149 'r I I ?'?4J

33qo odo

Figure 7. Dendrogram depicting the genetic relationship (Nei's D) amongst the seven populations of Milyeringa rerltas (data from Adams & Humphreys 1993) plotted alongside a map depicting the position of the populations sampled. The numbers on the dendrogram are the molecuiar clock ages (thousand years) derived from Nei's relationship (Nei 1987) as discussed in the text. The numbers alongside the dendrogram and on the map denote the site numbers as recorded in the karst index for Cape Range. CRNP denotes the two population, sampled that are included within Cape Range National Park. Top of the map is north. but it is not possibleto explorethe significanceof this These data are consistentwith the hypothesisthat the result any further without additional sampling. regional genetic structure between M. veritas popu- Many cave organisms appear to conform to the lations was determined by vicariance associatedwith molecular clock (Sbordoni et al. 1990) proposedby eustaticevents. Nei (1987)in which one unit of Nei's D is equivalent to 5 million yearsof divergencebetween populations. Using this relationship, the estimateddivergence times Conservation status between the M. veritas populations (Figure 7) are con- cordantwith major periodsof low sealevel, up to 130m The remote location of these cave fish and their below that now present.Three of the last four extreme subterraneanhabitat restricts observation in life and sealevel minima occurredat about95 000. 55 000 and hence very little is known about the biology of 20 000 yearsago (Wyrwoll et al. 1993),the estimated either sympatricspecies. Nonetheless, M. veritas and times of the three most recent divergencesbetween the O. candidum have, respectively, been classified as M. veritas populations.There was anothermajor sea rare and recommended for total protection, and vul- level minimum at ca. 315000 years ago (Hays et al. nerable(Michaelis 1985). Both fish species(as well 1976),coinciding with the estimatedtime of the ini- as some stygal crustaceanspecies) have been listed tial divergence between the M. veritas populations. as endanseredunder ScheduleOne of the Wildlife 310

Protection Act of Western Australia owing to their others, discussedby Boulton (2000) for subsurface small geographicdistribution, low populations and systems. vulnerability of their habitat. Bundera Sinkhole and CameronsCave, sites inhabited by M. veritas,are clas- sifiedas endangered communities by theWestern Aus- Prospects tralian wildlife authorities.The conservationstatus of these speciesis discussedby Romero & Vanselow Although information on M. veritas is sparse,and of (2000a,b). variable quality, it is clear that this species inhabits While there has beena significantextension of the waters with an extraordinary physico-chemical and range of M. yeritas to Barow Island (Humphreys biological variety. There are clear interuptions to the 1999c'),the inclusive known range of the two species gene flow within the Cape Range peninsula that may on the CapeRange peninsula has not increasedsince reflect earlier eustaticevents. 1991(Humphreys & Adams1991). The numberof sites Few sites have been examined in respect of fiom which M. veritashas been recorded has increased the allozyme variation in M. veritctsand a more com- steadilyover time, from two sitesin 1962to 25 sitesin plete sampling may provide a more complex picture 1999(Humphreys 1999c). These additional sites have of the variationbetween local populations,and of the markedly increasedthe known range of habitatsand vicariancehypothesized to be associatedwith eustatic water types inhabitedby the species.However, 19 of events. In addition, the non-random mating apparent thesesites are insecure or havebeen lost - six of these from the allozyme data suggeststhat detailedpopula- accesssites have been lost in the interveningperiod tion genetics using molecular markers would provide due to infilling, drying or siltation of the sites,and a insightsinto the populationprocesses of the cavefish. turther l0 sitesare close to planneddevelopments or Individual long term marking is neededto provide are rvithin an urban area, and three are within a mili- the basic population data required intrinsically and tary bombing rangereserve. Only six sitesare within lbr possiblemanagement needs, such as growth rate, Cape Range National Park where infrastructure devel- dispersal,dispersion, population size and structure, and opmentsare alsoplanned. M. yeritasis certainlymore breeding.This would alsofacilitate examination of the widespreadin thekarst than indicated by thepoint sam- physico-chemical environment variation for a given ples. Despitethis, the loss of accesssites is crucially population,which, with their physiologicaltolerance important as they provide the only windows into the or ability to conform, are at presentunknown. karst systemto enablethe biological study and mon- The degreeof ocular regressionis unknown, as is itoring of cave flsh populationsand their associated the degree and nature of the stygomorphies.Finally, ecosystems. the affinities of the specieswithin the Eleotridae are Owing to the genetic diversity and restricted gene unknown, the resolution of which would stimulate flow in M. veritas on the peninsula(Humptueys & researchon the Australian caveflshes. It is now known Adams 1991, Adams & Humphreys 1993),the pres- that the populations of M. veritas are of sufficient size ence of M. yeritas within Cape Range National Park to sustain a carefully planned major research effort. doesnot in itself providean adequaterepresentation of Despitethe generalremoteness of the area,the popula- the geneticdiversity of the species(Figure 7), nor does tions occur at sufficientlyaccessible locations to facil- it adequatelyrepresent their habitat diversity. itate sampling. The stygal community on the Cape Range penin- sulais unusualin havingtwo speciesof sympatriccave fish within a rich communityof stygalCrustacea, many Acknowledgements belongingto relict groups.As the flsh and crustaceans are part of the samefood web (Figure 4), the broad I acknowledgewith pleasurethe sustainedtechnical conservationimplications are clear-the stygofaunars supportof JulianneWaldock, the voluntaryassistance a functionalcommunity, changes to part of which may of Caroline Lawrence who measured the fish. and affectthe remainder. Such relatively simple ecosystems the cooperation over many years of Darren Brooks, 'top-down' arepotentially vulnerable to disruptionby Ken Cameron, Brian Vine and Keith Hallett. Graham effects (e.g. removal of M. veritas as the top preda- Proudloveprovided unpublishedmaterial and useful 'bottom-up' tor), and M. veritas susceptibleto effects references.Part ofthe work derivedfrom funding by the (e.g.changes to biolilms by nutrients),factors, amongst National Estate Grants Scheme. Australian National 311

Parks and Wildlife Service, Water Authority of West- Gentilli, J. 1972.Australian climatic pattems.Nelson. Melboume. ern Australia, the Water Corporation, West Australian 285 pp. Petroleum Pty. Ltd. and Ampolex Ltd. Gentilli, J. 1919. Epitropical westerly jet advective storms. Q. Geogr.J. 5: 1 20. Harvey. M.S. & W.F. Humphreys. 1995. Notes on the genus Draculoides Harvey (Schizomida: Hubbardiidae), with the References cited descriptionof a new troglobitic species.Rec. West.Austr. Mus. (Suppl.)52:183 189. Adams,M. & W.F.Humphreys. 1993. Patterns of geneticdiver- Hays, J.D., J. Imbrie & N.J. Shackleton. 1976. Variattons tn sity within selected subtenanean fauna of the Cape Range the Ealth's orbit: pacemaker of the ice ages. Science 194: peninsula, Western Australia: systematic and biogeographic tl2l-1t32. implications. Rec. West. Austr. Mus. (Suppl.) 45:145-164. Humphreys, W.F. 1993a. Stygofauna in semi-arid tropical Allen, A.D. 1993. Outline of the geology and hydrogeology of WesternAustralia: a Tethyan connection?M6m. Biospdol. 20: Cape Range, Carnarvon Basin, Westem Australia. Rec. West. lll 116. Austr. Mus. (Suppl.) 45: 25-38. Hurnphreys. W.F. 1993b. The significance of the subteranean Allen. G.R. 1989. Freshwaterfishes of Australia. T.F.H. Publica- fauna in biogeographical reconstruction: examples from Cape tions Inc., Neptune City. 240 pp. Range peninsula. Westem Australia. Rec. West. Austr. Mus. ANCA. 1996. A directory of important wetlands in Australia, (Suppl.)45: 165-192. 2nd edn. Australian Nature Conservation Agency, Canbena. Humphreys, W.F. 1999a. Relict stygofaunasliving in sea salt. 478 pp. karst and calcrete habitats in arid northwestem Austraha con- Beard, J.S. 1975.Pilbara: vegetationsurvey ofWestern Australia tain many ancient lineages. pp. 219-227 .1n: W. Ponder & D. l: 1 000 000 vegetationsedes. Explanatory notesto sheet5, the Lunney (ed.) The Other 99Vo,The Conservation and Biodi- vegetation of the Pilbara area. University of Western Australia versity of Invertebrates, Transactions of the Royal Zoological Press,Perth. 120 pp. Society of New South Wales, Mosman. history of the central Beard, J.S. I 998. Position and development Humphreys, W.F. 1999b. Physico-chemical profile and energy watershed of the Western Shield, Western Australia. J. Roy. fixation in Bundera Sinkhole. an anchialine remiped habitat in Soc.West. Austr. 8l: 157-164. north-westernAustralia. J. Roy. Soc. West. Austr. 82: 89-98. Bergstrom, D.E. 1997. The phylogeny and historical biogeog- Humphreys, W.F. 1999c. The distribution of the Australian cave raphy of Missouri's Anblyopsis rosae (Ozark cavefish) and fishes.Rec. West. Austr. Mus. 19: 469472. Typhlichthl'ssubterraneus (southern cavehsh). M.Sc. Thesis, Humphreys. W.F. 2000a. The hypogean tauna of the Cape Range University of Missouri-Columbia, Columbia. 62 pp. peninsula and Barrow Island. north-west Australia. pp. 581- Boulton. A.J. 2000. The subsurf'acemacrofauna. pp. 331-361. In: 601. In: H. Wilkens. D.C. Culver & W.F. Humphreys (ed.) J. Jones & P. Mulholland (eds), Streams and Ground Waters, Ecosystemsof the World, Vol. 30, SubtemaneanEcosystems, Academic Press, New York. Elsevier. Amsterdam. Chappel1,J. & B.G. Thom. 1977. Sea levels and coasts. pp. Humphreys, W.F. 2000b. Karst wetlands biodiversity and conti- 215-291.1n: J. Allen, J. Golson & R. Jones (ed.) Sunda and nuity through major climatic change - an example fiom arid Sakul: PrehistoricStudies in SoutheastAsia. Melanesia& Aus- tropical Westem Australia. pp. 227-258.1n; B. Gopal. W.J. tralia, Academic Press,London. Junk & J.A. Davis (ed.) Biodiversity in Wetlands: Assess- Christiansen,K. 1962. Proposition fbr the classificationof cave ment, Function and Conservation,Vol. l. Backhuys Publishers, animals. Spelunca2: 76-78. Leiden. Cohen, D.M. & J.E. McCosker. I998. A new speciesof bythitid glossary.pp.3 14.ht: fish, genus Lut'ifuga, from the Gallpagos Islands. Bull. Mar. Humphreys, W.F.2000c. Background and D.C. W.F. Humphreys (ed.) Ecosystems Sci.63: 179-187. H. Wilkens, Culver & Cummins, K.W. & J.C. Wuycheck. 1971. Caloric equivalentsfor of theWolld, Vol.30, SubterraneanEcosystems, Elsevier, Ams- investigationsin ecological energetics.Mitt. intemat. Verein. terdam. Limnol. 18: 1-158. Humphreys, W.F. & M. Adams. 1991. The subtenaneanaquatic Danielopol, D.L., A. Baltands & W.F. Humphreys. 2000. fauna of the Nofih West Cape peninsula, Western Australia. Daniekryolina kornickeri sp. n. (Ostracoda: Thaumatocypri- Rec.West. Austr. Mus. 15: 383 411. doidea) from a western Australian anchialine cave * morphol- Humphreys, W.F., M. Adams & B. Vine. 1989. The biology of ogy and evolution. Zool. Scripta 29: l-16. Schizomusvinei (Chelicerata:Schizomida) in the cavesofCape Diaz Perez.P.A., A.M. Lima & E.G. Machado. 1987a.Morpholo- Range,Western Australia. J. Zool. Lond. 211 111-20r. gia externa de ejemplares machos de Lucifuga siralle Nalbant, Humphreys. W.F. & M.N. Feinberg. 1995.Food of the blind cave l98l (Ophidilformes, Bythitidae). Revista Biol. (Havana) l: fishes of northwestern Australia. Rec. West. Austr. Mus. 17: 11-84. 29-33. Diaz Perez,PA., E. Nieto Misas & G. Abio Virsida. 1987b.Peces Humphreys. W.F., A. Poole, S.M. Eberhard & D. Warren. 1999. ciegosdel generoLucifuga (Ophidiiformes: Bythitidae) en dos Effects of research diving on the physico-chemical profile casimbascubanas. Rev. Invest. Mar. 8: 4l-r17. of Bundera Sinkhole, an anchialine remiped habitat at Cape Ford, D.C. & P.W. Williams. 1989. Karst geomorphology and Range, Westem Australia. J. Roy. Soc. West. Austr. 82: 99- hydrology. Unwin Hyman, London. 601 pp. 108. 312

Iliff'e, T.M. 2000. Anchialine cave ecology. pp 59-16. In: H. Planes, S. 1998. Genetic diversity and dispersal capabilities in Wilkens. D.C. Culver & W.F. Humphreys (ed.) Ecosystems marine fish. Evol. Biol. 30: 253-298. of the World, Vol. 30, Subtenanean Ecosystems, Elsevier, Poore, G.C.B. & WF. Humphreys. 1992. First record of Ther- Amsterdam. mosbaenacea (Crustacea) from the Southern Hemisphere: a Jaume,D., G.A. Boxshall & WF. Humphreys. 2001. New stygo- new species fiom a cave in tropical Westem Australia. Inverl. biont copepods(Calanoida, Misophrioida) from Bundera Sink- Taxon. 6: 1 19-725. hole, an anchialinecenote on north-westernAustralia. Zoo1.J. Poulson,T.L. 1964.Animals in aquaticenvironments: animals in Linn. Soc., Lond. (in press). caves.pp. 749-771.In; D.B. Dull (ed.) Handbook of Physiol- Jaume,D. & WF. Humphreys.200l. A new genusof epacteriscid ogy, American Physiological Society, Washington. calanoid copepod from an anchialine sinkhole in northwestern Romero, A. & PB.S. Vanselow. 2000a. Threatened flshes of Australia.J. Crust.Biol.21: 157-169. the world: Mill-eringa r.,erlrasWhitley, 1945 (Eleotridae). Env. Keighery, G. & N. Gibson. 1993.Biogeography and composrtion Biol. Fish. 57: 36. Romero, A. & PB.S. Vanselow. 2000b. of the flora of the Cape Range peninsula, Western Australia. Threatened fishes of the world: Ophisternon (Mees) (Synbranchidae). Rec. West. Austr. Mus (Suppl.) 45: 51-85. candidum Env. Biol. Fish. 58:214. Knott, B. 1993.Stygofauna from Cape Rangepeninsula, Western Sbordoni, V., A. Caccone, G. Allegrucci & D. Cesaroni. 1990. Australia: tethyan relicts. Rec. West. Austr. Mus. (Suppl.) 4-5: t09-121. Molecular island biogeography. Atti Convegn. Lincei 85: 55-83. Kuhajda, B.R. & R.L. Mayden. 2001. Status of the feder- Sket, B. i981. Fauna of anchialine (coastal)cave warers.irs ori- al ly endangered Alabama cavefi sh, Sp eo p Ia r rhinu s p o ul soni 1,^ gin and importance. pp.646-64'7 . In: B.F. Beck (ed.) Proceed- (Amblyopsidae), in Key Cave and the surrounding caves, ings of the Eighth Intemational Congress of Speleology, Bowl- Alabama. Env. Biol. Fish. 62: 215-222 (this volume). ing Green, Kentucky, Vol. 6, National Speleological Society, Levins, R. 1969. Evolution in changing environments.Princeton Huntsville. University Press,Princeton. 120 pp. Sket, B. 1986. Ecology ofthe mixohaline hypogean fauna along Martin, M.W. 1990. Exmouth town water supply investigation the Yugoslav coast.Stygologia 2: 311-338. report and recommendations for future work. Hydrogeology Sket, B. 1996. The ecology of anchihaline caves. Trends Ecol. Report No. 1990/36, Western Australian Geological Survey, Evol. ll:221-255. Perth. 1I pp. Slack-Smith, S.M. 1993. The non-marine molluscs of the Cape McNamara, K.J. (ed.), 1990. Evolutionary Trends. Belhaven Range peninsula, Western Australia. Rec. West. Austr. Mus. Press,London. 368 pp. (Suppl.)45:81-101. McNamara, K.J. & G.W. Kendrick. 1994. Cenozoic molluscs Slatkin, M. 1981. Estimating level of gene flow in natural popu- and echinoidsof Barrow Island. Westem Australia. Rec. West. lations. Genetics99: 323-335. Austr. Mus. (Suppl.) 51: 1-50. Slatkin, M. & N.H. Barton. 1989. A comparisonof three indirect Mees, G.F. 1962.The subteraneanfauna of Yardie Creek statron. methods for estimating average ievels of gene flow. Evolution North West Cape, Western Australia. J. Roy. Soc. West. Austr. 43:1349-1368. 45:24-32. Soul6, M. 1976. Allozyme variation: its determinants in space Michaelis, F.B. 1985.Threatened fi sh. A report on the threatened and time. pp. 60-71 .1n: F.J.Ayala (ed.) Molecular Evolution, fish of inland water of Australia. Australian National Parks and SinauerAssociates. Sunderland. 'anchia- Wildlif'e Service,Report Series(3): l-45. Stock,J.H., T.M. IlifTe& D. Williams. 1986.The concepr line' reconsidered.Stygologia Morse, K. I 993. Who can seethe sea?Prehistoric aboriginal occu- 2: 90-92. Thinds, G. & G. Proudlove. 1986. Pisces. pp.109-733. In: L. pation of the Cape Range peninsula. Rec. West. Austr. Mus. Botosaneanu(ed.) Stygofauna (Suppl.) 45:221-242. Mundi, A Faunistic, Distribu- tional, and Ecological Synthesis ofthe World Fauna Inhabiting Nei, M. 1987. Molecular Evolutionary Genetics. Columbia Subteranean Waters, E,.J.Brill/Dr. W. Backhuys, Leiden. University Press,New York. 512 pp. Trexler, J.C. 1988. Hierarchical organization of genetic variation Nelson, J.S. 1984. Fishes of the World, 2nd edn. Wiley- in the sailfish molly, Poecilia latipinna (Pisces: Poeciliidae). Interscience,New York. 532 pp. Evolution42: 1006-1017. Nevo. E. 1978. Genetic variation in natural populations:pattems Van Valen, L. 1965. Morphological variation and width of and theory. Theor. Pop. Biol. 13: l2L-111. ecological niche. Amer. Nat. 94: 377-390. Nevo, E., A. Beiles & R. Ben-Sholmo. 1984. The evolutionary Weber, A., G.S. Proudlove & T.T. Nalbanr. 1998a. Pisces significance genetic of diversity: ecological,demographic, and (Teleostei). A. Morphology, systematic diversity, distribu- Iif'ehistory conelates.pp. (ed.) l3-213. 1n: G.S. Mani Lecture tion, and ecology of stygobitic fishes. pp. 1179-1190. In: C. Notes in Biomathematics, Vol. 53, Springer-Verlag,Berlin. Juberthie & V. Decu (ed.) Encyclopaedia Biospeologica II, Palmer.J. I 985. The blue holes of the Bahamas.Jonarhon Caoe. Soci6t6 de Biosp6ologie, Moulis & Academie Roumaine, London. 185 pp. Bucarest. Pesce,G.L., P De Laurentiis & W.F. Humphreys. 1996.Copepods Weber, A., G. S. Proudlove, J. Parzetall, H. Wilkens & from ground waters of Westem Australia, Pnrt I, The gen- T.T. Nalbant. 1998b. Pisces (Teleostei). pp. 1171-1213. In: era Metac'yclops, M esoc 1-clop s, Mi crocyclop s and Apocyclop s C. Juberthie & V Decu (ed.) Encyclopaedia Biospeologica (CrustaceaCopepoda: Cyclopidae). Rec. West.Austr. Mus. l8: II, Soci6t6 de Biosp6ologie, Moulis & Academie Roumaine, 67-'76. Bucarest. 313

Whitley, G.P i945. New sharks and fishes from Western Range-Exmouth Gulf region. Rec. West. Austr. Mus. (Suppl.) Australia,Part 2. Austr.ZooI. 1l:35-3'7. 45: 1-23. Wilkens,H. 1988.Evolution and geneticsof epigeanand cave Yager,J. 1981.Remipedia, a new classof Crustaceafrom a manne Astyanax.fascialzrs(Characidae, Piscesae, Pisces): support for cave in the Bahamas.J. Crust. Biol. l: 328-333. theneutral mutation theory. Evol. Biol. 23: 271 367. Yager, J. & W.F. Humphreys. 1996. Lasionectes exleyi sp. nov., Wilkens,H., U. Strecker& J. Yager.1989. Eye reductionand the first remipede crustacean recorded trom Australia and the phylogeneticage in ophidiformcave lish. Z. zool. Syst.Evolut. Indian Ocean, with a key to the world species. Inverl. Taxon. -forsch.27:126-134. l0:171-187. Wright.S. 1931.Evolution in Mendelianpopulations. Genetics Young, M. 1986. Bringing the blind gudgeon into caprivity.Fish. 16:91-159. Sahul4: 145-148. Wyrwoll,K.H., G.W. Kendrick & J.A.Long. 1993.The geomor- phology and Late Cenozoicgeological evolution of the Cape