Abstract: A dense assemblage of fossil isopod crustaceans (Brunnaega tomhurleyi Wilson, sp. nov.) from the Lower Cretaceous (Albian) Toolebuc Formation of Queensland, Australia, has been found within the carcass of a large actinopterygian fish, Pachyrhizodus marathonensis (Etheridge). Preservation of fine anatomical details supports referral to the genus Brunnaega Polz, which is herein reassigned to the family Cirolanidae. Furthermore, placement of this taxon within the cirolanid subfamily Conilerinae Kensley and Schotte is significant because the group includes modern species that are well known as voracious scavengers. This isopod–fish association represents the oldest unequivocal evidence of scavenging by Mesozoic cymothoidean isopods on a large vertebrate carcass.
I sopods are an important component of marine ecosystems and, despite their small size (compared to larger ‘keystone’ invertebrates like decapods), have a significant impact on energy transfer via their preference for scavenging lifestyles (Shafir and Field 1980). In many habitats, the tropics for example, isopods can comprise a substantial part of the scavenging guild (Keable 1995) and are frequently known to consume vertebrate carcasses or even attack injured fish (Stepien and Brusca 1985; Brusca et al. 1995). Scavenging in crustaceans probably dates back to at least the Early Cambrian (520 Ma) with the appearance of stem-group lineages such as the Bradoriida (Wilkinson et al. 2007; Hou et al. 2010). The antiquity of scavenging habits among isopods, however, is not completely certain. Isopod taxa, such as PalaegaWoodward, 1870 (related to modern BathynomusMilne-Edwards, 1879), that today use scavenging as their primary feeding mode have been present in the oceans since the Mesozoic (e.g. Hessler 1969; Wieder and Feldmann 1989, 1992). Nevertheless, direct evidence of this dietary strategy is rare and so far limited to isolated finds occurring in association with a water bug from the Upper Jurassic (Tithonian) Solnhofen Lagerstätte (Polz 2004) and a teuthoid squid from the Upper Jurassic (Kimmeridgian) Nusplingen Lithographic Limestone of Germany (Polz et al. 2006); inferences of scavenging habits have also been put forward based on specimen numbers and stratigraphical affinities (Feldmann et al. 1998), or morphological comparisons (e.g. Bowman 1971).
Here, we report the discovery of a densely packed assemblage of more than 130 individual isopods infesting the carcass of a large actinopterygian fish, Pachyrhizodus marathonensis (Etheridge, 1905), from the Lower Cretaceous (Albian) Toolebuc Formation of Queensland, Australia. The specimens are referred to a cymothoidean genus, BrunnaegaPolz, 2005, and the species B. tomhurleyi Wilson sp. nov., which is described herein. Because these specimens provide previously unknown anatomical information, a revised generic diagnosis is provided for Brunnaega, and the genus is reassigned to the family Cirolanidae. This finding marks the first incontrovertible evidence of scavenging by Mesozoic cymothoidean isopods on a vertebrate carcass.
Numerous fossil isopods (South Australian Museum registration numbers (SAM) P45635–P45654) were found eroding out of the limestone matrix mostly filling the opercular cavity in the skull of an actinopterygian fish (Pachyrhizodus marathonensis; SAM P41209; Text-fig. 1). The fish specimen was collected from the Toolebuc Formation at Canary Station near Boulia in central-western Queensland, Australia (Text-fig. 2). Senior et al. (1975, 1978) and Moore et al. (1986) have provided detailed accounts of the lithology, stratigraphical relationships and distribution (including geological and locality maps) for this rock unit. In brief, the Toolebuc Formation comprises a thin series (<65 m) of calcareous and carbonaceous mudstones with abundant coquinite (composed mainly of Inoceramus) and bituminous shale. Stratigraphically, it constitutes a medial subdivision within the Rolling Downs Group (Eromanga Basin) and has a conformable lower boundary with the Wallumbilla Formation and an upper conformity with the Allaru Mudstone (Moore et al. 1986). Age determinations for the Toolebuc Formation based on dinoflagellates and palynomorphs indicate a latest middle to late Albian range: Pseudoceratium ludbrookiae dinoflagellate zone/upper Coptospora paradoxa spore-pollen zone (Moore et al. 1986; McMinn and Burger 1986).
The Toolebuc Formation outcrops around Boulia incorporate around 25–36 m (Boulia A type section; Senior et al. 1975) of carbonaceous mudstone with fossil-rich concretionary limestone; this is often exposed and broken up during exfoliation of the laminated parent rock. Invertebrate (ammonites, bivalves; Day 1969) and vertebrate macrofossils (marine reptiles, fish; Kear 2003, 2007; Kear and Lee 2006) are common and typically manifest complete/articulated specimens consistent with accumulation under well-oxygenated shallow water conditions. However, the presence of dark organic-rich shale layers indicates stratification within the water column and dysaerobia at depth (Moore et al. 1986). Depositional interpretations of the Toolebuc Formation suggest that it was laid down slowly in a restricted marine environment (Morgan 1980; McMinn and Burger 1986). This was part of a maximum transgressive episode extending south-west into the pre-Great Artesian Basin complex from the southern Tethys Ocean (Morgan 1980).
Materials and methods
The isopod fossils have been partially replaced by hard crystalline calcite, such that the interior parts of the limbs are well preserved. Conversely, the exterior cuticle is phosphatized and could thus be etched by acid. Further mechanical removal of the soft limestone matrix revealed additional structures and permitted a composite reconstruction of the whole animal based on multiple individuals.
Because the density of individuals within the mass was so high and many specimens were overlapping, a coordinate system was created (Text-fig. 1) to locate isopod fossils on the rock slab (Table 1). Positions were designated by x, y coordinates oriented to the approximate centre of the specimen. Only those fossils that showed useful features were recorded and described (see Supporting Information). Body length measurements represent the observed value, or estimates where the fossil was incomplete (denoted by ‘+’). Some specimens were prepared by chipping away the surrounding matrix, which exposed uneroded anatomy.
Table 1. Specimens employed for the description of Brunnaega tomhurleyi sp. nov.
SAM Reg. number
More information provided in Supporting Information. In the length column (+) or (++) indicates different degrees of missing specimen so that the reported length is a minimum.
After preparation, individual fossils were photographed from several angles and focal depths. The fossils were photographed either as single digital images or were composited to merge different focal levels. These images were used as templates in a vector image editor (http://inkscape.org), and the outlines of the fossils were drawn; the reconstruction was created in the same way. Character states were recorded in a taxonomic database using the DELTA system (Dallwitz 1980; Dallwitz et al. 2000a, b), with definitions derived from the Cirolanidae character set developed by S. J. Keable (Keable 2001, 2006) but modified to provide consistent nomenclature (Wilson 2009) and to accommodate specific information from the fossils. A natural language description, generated using the DELTA tool CONFOR, was edited for clarity (e.g. merging repetitive text). Diagnoses were created from lists of similarities and differences between the two species of Brunnaega using the DELTA tool INTKEY. The composite reconstruction of the new species was based on photographs compiled from specimen x152 y068 (SAM P45651, holotype) for the anterior part of the body with overall approximate shape also drawn from x071 y097 (SAM P45638). The antennulae and antennae were derived from x132 y068 (SAM P45646), lateral view from x081 y071 (SAM P45640) and legs from the intact appendages on x146 y075 (SAM P45648) and x164 y051 (SAM P45653). An explicit taxonomic authority (International Code of Zoological Nomenclature (ICZN) recommendation 50A) is given for new species name, Brunnaega tomhurleyi Wilson, sp. nov. Institutional abbreviation: SAM, South Australian Museum.
Diagnosis. Body size small, adult length less than 20 mm. Cuticular surfaces smooth, unornamented. Head without ornamentation, furrows or ridges absent, dorsal surface smooth, not set deeply into pereonite 1. Eyes present, well developed, visible in dorsal and ventral view, moderate in size, ovate in lateral view, length less than 2.0 height, partially overlapped by pereonite 1. Pereonites without tubercles. Pereonite 1 longest, less than 1.5 length of pereonite 2, pereonites 2–7 lengths subequal. Pleonite 1 posterolateral margins projecting ventrally. Pleonites 2–5 posterolateral margins not projecting equally, dorsal posterolateral margin subequal with ventral posterolateral margin, distally acute; pleonite 3 posterolateral margins projecting, extending posterior to posterodorsal margin of pleonite 5, acute; pleonite 4 posterolateral margins distinctly projecting posterior to pleonite 3, ventral margin separated from pleonite 3, posterodorsal margin sinuate, convex proximal to meeting ventral margin at apex, apex forming narrow acute point, extending distinctly posterior to posterodorsal margin of pleonite 5; pleonite 5 posterolateral margin encompassed by pleonite 4. Pleotelson broad (basal width greater than length), smooth, without ornamentation and sculpting, anterodorsal uropodal sutures on anterolateral margins convex; posterolateral margins without spines, sinuate, apex not projecting, lateral margins tapering smoothly to posterior point. Pereopods I–III dactylus weakly curved, not strongly curved and hook-like, propodus longer than dactylus. Pereopods anterior coxae shallow – not as high as long, posterior coxae deep – higher than long; central ridges (furrows) present on all pereopodal coxae, moderately developed, anterior oblique ridge on pereopods II–VII; coxae II–VII articulations with body distinct; coxae I–II posterior margins straight or slightly convex, other coxae projecting beyond posterior margin of associated pereonite; angular apices absent on all coxae; coxae II with anteroventral corner rounded. Uropods extending near posterior extent of pleotelson; protopod medial margin strongly projecting, protopod distolateral angle rounded; endopod triangular, medial margin rounded, apex entire, without notch; endopod lateral margin convex, without pit or distinct excision.
Remarks. The original diagnosis for Brunnaega was necessarily short owing to missing anatomy on the type species B. roeperiPolz, 2005. That diagnosis included features that do not distinguish this genus from other similar isopods (e.g. ‘uropods biramous’), which are assumed implicitly in the revised diagnosis above. Although we do not have evidence for many features included in the diagnosis for the type species, the shared anatomical similarity between it and the new species described below supports extension of the generic diagnosis. The diverse family Cirolanidae has 61 extant genera, so a detailed comparison with these taxa is not possible owing to the limited anatomical data from the fossils. Brunnaega can nevertheless be recognized among most cirolanid taxa, fossil or extant, by its smooth dorsal surfaces, elongate distally rounded coxae, elongate uropods with subtriangular endopods and a pleotelson lacking carinae or marginal denticles. Its classification is further discussed below. The specimens assigned by Hansen and Hansen (2010) to AegaLeach, 1815, from the Miocene of Denmark also may belong to Brunnaega, which would further extend the stratigraphic range of this genus. These latter fossils, however, are too fragmentary to be certain.
Derivation of name. This species of Brunnaega is named in honour of Mr Tom Hurley, an enthusiastic private collector of Cretaceous fossils in the Boulia district of Queensland.
Holotype. Here designated, one of four specimens in a group, SAM P45651, 8.5 mm (+) (Text-fig. 4).
Material Examined. See Table 1 and Supporting Information for details on surveyed specimens.
Diagnosis. Fossils belonging to Brunnaega. Body length 8–10 mm. Pereonites without transverse ridges (carinae). Pereonite 1 laterally with 2 longitudinal lateral carinae. Pleonites ornamentation absent, all pleonites smooth, without tubercles.
Overall adult body form of small size, 8–10 mm long (n = 8), medium width, length approximately 2.5 greatest width, cuticular surfaces smooth and polished.
Head. Without ornamentation, furrows or ridges absent, surface smooth, not set deeply into pereonite 1, length less than width. Frontal margin angled ventrally, extending to frontal lamina; anterior margin smoothly rounded in dorsal view, distinct but not overriding antennulae; apex overlapping frontal lamina. Cephalic ridges absent, submarginal furrow present, submarginal cephalic furrow well developed, extending along entire length of anterior margin to eyes; dorsal interocular furrow absent. Eye ommatidia arranged in rows (visible on holotype, x152 y068, SAM P45651), 7–8 horizontally, 4–6 vertically, ommatidia not divided by unfaceted band.
Frontal lamina, clypeus and labrum. Frontal lamina present, broad, length approximately 2.5 basal width, forming angle of c. 90 degrees with ventral surface of head; ventral surface flat, not projecting, not sculpted, not expanded, in 1 plane (not ‘stepped’); lateral margins straight and parallel (x151 y076, SAM P45649); apex not projecting, not visible in dorsal view, anterior margin rounded. Clypeus subtriangular, width greater than length, anterior half of ventral surface lying in same plane as frontal lamina, but posterior half lying in different plane and forming projection, not sculpted. Labrum flat, narrower than clypeus, smooth, without lamina projection.
Pereonites. Dorsal surfaces smooth, without ornamentation other than lateral ridges, ornamentation absent (x152 y068, SAM P45651–2; x081 y071, SAM P45640). All pereonites smooth, transverse ridges (carinae) absent, tubercles absent. Pereonite 1 medial length greater than other pereonites length.
Pleonites. All equally visible along dorsal margin, pleonite 5 present (x152 y068, SAM P45651–2; x081 y071, SAM P45640). Ornamentation absent, all pleonites smooth, tubercles absent. Pleonites 2–4 posterolateral margins projecting posteriorly, lengths not equal. Pleonite 2 ventral posterolateral margin acute, formed into short process. Pleonite 3 posterolateral margins extending posterior to posterodorsal margin of pleonite 5, distally acute. Pleonite 4 posterolateral margins extending posterior to pleonite 3, posterodorsal margin sinuate, convex proximal to meeting ventral margin at apex, extending posterior to posterodorsal margin of pleonite 5. Pleonite 5 lateral margin encompassed by pleonite 4.
Pleotelson. Broad, length 0.94 basal width (x152 y068, SAM P45652, specimen on right), smooth, ornamentation and sculpting absent, anterodorsal depression present, anterodorsal uropodal sutures on anterolateral margins present. Posterolateral margins without spines, sinuate, apex not projecting, lateral margins tapering smoothly to point.
Antennula. Short, not reaching pereonite 1 (x132 y068, SAM P45646), length less than antenna. Articles 1–2 flexibly articulated, weakly geniculate (anterodistal angle of article 1 rounded; x128 y035, SAM P45644; x132 y068, SAM P45646). Article 1 length greater than width and greater than article 2. Article 2 width subequal to length. Article 3 short, subequal to articles 1 and 2 but shorter than combined lengths of articles 1–2; length greater than width. Article 4 present and small. Flagellum shorter than podomeres, not formed into callynophore, of 10 articles, not axially compressed (lengths of most greater than half width); first flagellar article (article 5) length not much greater than width, longest, subsequent articles decreasing in length distally.
Antenna. Short, reaching between anterior of head and posterior margin of pereonite 1. Basal podomeres composed of five articles, articles 1–2 subequal, 3–5 longer than 1–2, 3–4 together subequal and shorter than article 5, article 2 longer than article 3, article 4 length subequal to article 3, article 5 subequal in length to article 4. Flagellum with 15 articles (at least; see x132 y068, SAM 45646).
Mouthparts. Mandible gracile, not massive with well-developed molar process (x124 y074, SAM P45643; other structures not visible); palp article 1 subequal to article 3, article 2 length c. 2.0 article 3 length, article 3 broader distally than proximally. Maxillula lateral lobe relatively slender. Maxilliped palp article 3 length less than width, distal margin width greater than proximal margin of article 4; article 4 length subequal to width, distal margin width greater than proximal margin of article 5; article 5 rounded, length greater than width. Endite attenuated, extending to first palp article.
Pereopodal coxae. Anterior coxae shallow, not as high as long; posterior coxae deep, higher than long (x081 y071, SAM P45640); central ridges (furrows) present on all coxae, moderately developed, anterior oblique ridge present; coxa-body, articulation complete on pereopods II–VII; I–II posterior margins straight or slightly convex, apex not projecting, II with anteroventral corner rounded.
Pereopods I–VII. Bases I–VII median longitudinal surface with broadly angular ridge (x146 y075, SAM P45648; x164 y051, SAM P45653). Pereopod I merus anterodistal angle short, extending distally just beyond posterodistal extent of carpus, posterior margin sinuate; carpus short, less than half length of propodus; dactylus long, length between 0.5–1 propodus length, slender. Propodus V–VII slender, subequal to that of pereopods II and III (assumed using relative size of propodus on x146 y075, SAM P45648). Pereopods V–VII without flattened articles, propodus long, but pereopod V length less than 2.0 pereopod VII length. Pereopod VII basis narrow (greatest width <0.48 length), medial longitudinal ridge positioned midway between anterior and posterior margins, anterior margin convex, posterior margin convex and rounded, more than one-third of margin straight; merus length greater than width; propodus longer than carpus.
Uropods. Extending to tip of pleotelson (x152 y068, SAM P45651; specimen on right). Endopod subtriangular with convex margins; apex entire, without notch. Exopod lanceolate, medial and lateral margins convex; apex acute, entire, without notch.
Occurrence and stratigraphical range. Toolebuc Formation near Boulia in central-western Queensland, Australia (23°16.248′S 140°30.574′E; Text-fig. 2); latest middle to late Albian range, c. 106–112 Ma.
Remarks. Brunnaega tomhurleyi Wilson sp. nov. is distinguished from the type and only other species B. roeperiPolz, 2005, by smaller body size 8–10 mm (versus > 14 mm). Pereonites lack dorsal transverse ridges or other ornamentation, and pereonite 1 bears two longitudinal lateral ridges (only one is present in B. roeperi). All pleonites are smooth dorsally, unlike B. roeperi in which pleonites have fine denticles on the posterior margin.
The specimens preserved with the fish skull (see Supporting Information: descriptions of individual specimens) are classifiable within the Cirolanidae because they possess a 5-articled maxillipedal palp, and a mandible with both a molar process and incisor process that broadens distally. These features disallow their classification in one of the cymothoidean free-swimming families (e.g. Aegidae; Corallanidae; Barybrotidae; Tridentellidae), which otherwise have variously reduced mouthparts. Although the fossil isopod literature contains previous reports of Aegidae (Polz 2005; Hansen and Hansen 2010), these are based on general body structure rather than specific character states that support indisputable attribution to one of the modern families. Hansen and Hansen (2010) recently presented a multivariate analysis of two isopod fossils from the Gram Formation (Denmark) with a broad range of modern species using a 12 character data set. They concluded that their fossils belonged to the genus Aega Leach, family Aegidae, even though they had no preserved limbs. Referral of these fossils to the Cirolanidae may be more feasible. Other non-Palaega cymothoidean taxa have been placed in the Cirolanidae (Wieder and Feldmann 1992; Hiller, 1999). We argue that Brunnaega Polz also belongs in the Cirolanidae because B. tomhurleyi has mostly ‘ambulatory’ anterior pereopods I–III with short weakly curved dactyli and a well-developed mandibular molar, while taxa in the Aegidae have strongly curved dactyli of raptorial, grasping pereopods I–III and a substantially reduced mandibular molar.
The specimens at hand do not belong in PalaegaWoodward, 1870, or BathynomusMilne-Edwards, 1879 (the latter synonymized into Palaega by Wieder and Feldmann (1989) but subsequently rejected by ICZN (1992) to maintain nomenclatural stability), because they manifest a dorsally smooth, uncarinate pleotelson that lacks marginal teeth or denticles. Debodea mellitaHiller, 1999, which was tentatively placed in the Cirolanidae, is not a cymothoidean owing to the anterior placement of its antennae and eyes on the head, and its narrow body form with small coxae. Debodea also has a habitus more similar to Phreatoicidea, but the original description of this taxon lacks sufficient evidence to make a clear determination. The morphologically diverse genus Cirolana, with more than 128 modern species, differs from Brunnaega in its broad frontal lamina and robust setae on the margins of the uropods. Species of Cirolana also tend to have broader bodies and shorter uropods. With explicit reference to the fossil forms, C. enigmaWieder and Feldmann, 1992, from the Upper Cretaceous (Campanian) of South Dakota, USA, contrasts with Brunnaega in its deeper body with limited expression of coxal articulations. Cirolana garassinoiFeldmann, 2009, from the Upper Cretaceous (Cenomanian) of Lebanon conversely has a much wider and flatter body, broad distally rounded uropods and a distal indentation in the pleotelson. Several other fossil isopods have been attributed to Cirolana (Feldmann et al. 2008; de Angeli and Rossi 2006), but they cannot be compared directly with Brunnaega.
The affinity of our fossils appears to lie within the ‘Conilera’ group (Bruce 1986), now referred to the subfamily Conilerinae Kensley and Schotte, 1989 (reviewed inRiseman and Brusca 2002). Distinctive conilerine attributes (sensuBruce 1986; Kensley and Schotte 1989) include: antennal articles 4 and 5 longest and approximately subequal; frontal lamina not strongly projecting or elongate; clypeus posteriorly flattened, frontal lamina flat narrow, pereopods I–III with ischium and merus anterodistally produced; pleonite 3, although not overlapping pleonite 4, is projected posteriorly; and pleonite 5 overlapped by pleonite 4. Conilerinae comprises small- to medium-sized cirolanids with generalized features that may not be well defined phylogenetically (Brusca et al. 1995; S. J. Keable, 2010 pers. comm.). Using the key of extant ‘Conilera’ group taxa developed by Wetzer et al. (1987), with the assumption that the antennular article 4 is tiny and therefore not resolved in our specimens, one arrives at NatatolanaBruce, 1981 (recently summarized inKeable 2006). Our fossils, however, do not belong in this genus because the posterior pereopods are narrow and lack expanded bases. We therefore suggest that species from the Cretaceous are unlikely to conform to the same genus-level clades as extant taxa.
Of the known Mesozoic forms, BrunnaegaPolz, 2005, appears to be the most similar to the specimens from Boulia, even though this genus was originally assigned to the Aegidae. Unfortunately, material of B.roeperiPolz, 2005, lacks a well-preserved head, so more detailed comparisons are not possible. The published diagnosis of Brunnaega comprises generalized features (paraphrased from Polz 2005, p. 74): body oval, dorsally convex, more than twice as long as broad; head with smoothly convex anterior margin; pereonites 1–7 free, coxae II–VII with distinct articulation to pereonites, with posterolateral tips; pleonites 1–5 free, marginally narrower than pereonites, dorsomedial lengths subequal, pleurae with posterolateral tips; pleotelson subtriangular, narrower than pleonites, posterior margin smooth; uropods biramous, lateral, endopod extending to apex of pleotelson. Although Polz (2005, fig. 2) reconstructs the pleonites as being approximately similar, his photographs (Polz 2005, plates 3–4) show that pleonite 4 laterally overlaps pleonite 5, similar to our specimens. Brunnaega roeperi also has pleonite 3 projecting substantially more than other pleonites, which features in the ‘Conilera’ group (sensuBruce 1986). This configuration of the pleonites is common within Cirolanidae, especially the form of pleonite 5, so these similarities may be uninformative. In addition, the large coxae of B. roeperi have a somewhat different shape anteriorly, and lateral ridges on the anterior coxae are distinctive. Despite these limitations, both species share comparable size and articulations of the coxae, the presence of lateral coxal ridges, and posterior coxae with approximately similar outlines. The pleotelson is also smooth and triangular and extends near to the posterior extent of the uropods (less so in B. roeperi). Because of these shared traits, we have assigned our specimens to a new species of BrunnaegaPolz, 2005, B. tomhurleyi. We also propose a transfer of this genus to Cirolanidae. This conservative placement extends the stratigraphical range of Brunnaega by approximately 50 myr from the upper Kimmeridgian (Upper Jurassic) to upper Albian (Lower Cretaceous).
Over 130 individuals of the cirolanid Brunnaega tomhurleyi Wilson, sp. nov. form dense clusters on the fish remains and are primarily concentrated around the gill (underneath the gill plate) and anterior body cavity where substantial amounts of soft respiratory and visceral tissue would have been situated.
Modern cirolanid isopods are typically voracious scavengers and micropredators (particularly among the Conilerinae) and can be abundant in shallow marine communities where they constitute one of the most important recyclers of organic matter (Stepien and Brusca 1985; Bruce 1986; Keable 1995; Wong and Moore 1995; Barradas-Ortiz et al. 2003; Castro et al. 2005). Cirolanids, together with other closely related cymothoideans (e.g. Aegidae, Cymothoidae), are also commonly associated with actinopterygian fishes, either on carcasses or injured individuals where they attach themselves to branchial and buccal surfaces as well as flesh and internal organs (Stepien and Brusca 1985; Bunkley-Williams and Williams 1998; Bunkley-Williams et al. 2006). In the light of this recognized lifestyle, the isopod–fish carcass association from the Toolebuc Formation most probably represents an example of scavenging. This inference is further supported by the primary feeding mode in modern cirolanids and other related taxa, which use the massive molar and incisor processes on their mandibles to process edible tissues. The high prevalence of scavenging in the ‘Conilera’ group (e.g. Brusca et al. 1995; Keable 1995, 2006; Wong and Moore 1995) is also coherent with this assessment.
The Toolebuc Formation depositional setting is consistent with a scavenging habit for Brunnaega tomhurleyi. The predominance of dark organic-rich shales indicates a stratified water column with fluctuating dysoxic to anoxic bottom waters (Morgan 1980; McMinn and Burger 1986; Moore et al. 1986; Henderson 2004); similar marine conditions were widespread during the mid-Cretaceous (Erbacher et al. 2001; Wilson and Norris 2001). This palaeoenvironment is reflected in the monospecific benthic communities, which tend to be dominated by inoceramid bivalves, a group interpreted as being tolerant of oxygen-poor habitats (Henderson 2004; see also Kauffman et al. 2007 for a discussion of inoceramid palaeoecology). Brunnaega tomhurleyi therefore might have been a specialized low-oxygen sea floor scavenger, comparable to extant cirolanids. For example, Natatolana borealis is known to flourish in dysoxic and anoxic concentrations (de Zwaan and Skjoldal 1979; Taylor and Moore 1995). Moreover, B. tomhurleyi could have had the ability to withstand long periods (weeks to months) of starvation, because of the scarcity of benthos, and respond rapidly to sporadic food falls such as vertebrate carcasses (as observed in N. borealis; Wong and Moore 1995; Kaïm-Malka 1997; Castro et al. 2005).
Based on fossil associations (e.g. Polz 2004; Polz et al. 2006) and phylogenetic affinities with modern detritivorous lineages (e.g. Polz 2005), scavenging lifestyles in isopods would appear to extend as far back as the Late Jurassic. Nevertheless, no direct link has hitherto been found with vertebrates. Feldmann et al. (1998) suggested that Archaeoniscus aranguthyorum from the Lower Cretaceous (Albian) Tlayúa Formation in Puebla, Mexico, might have been either an ectoparasite or scavenger because it occurred alongside numerous fishes within the assemblage. The spectacular discovery of a direct cirolanid–actinopterygian association in the Lower Cretaceous (Albian) Toolebuc Formation of Australia is therefore highly significant because it constitutes the oldest unequivocal example of isopod scavenging on a vertebrate carcass in the fossil record and confirms that such habits might have been widespread throughout the Mesozoic.
Acknowledgements. We are grateful for a Cirolanidae DELTA character list provided by S. J. Keable, and also for discussions with him on relationships of cirolanid taxa. S. Lindsay in the Australian Museum Imaging Lab assisted with the photography of the specimens. M. A. Binnie assisted with the registration of the types at the South Australian Museum and suggested corrections to the Supporting Information. B. P. Kear acknowledges the financial support of the Australian Research Council. We thank S. J. Keable, R. Feldmann and an anonymous referee for helpful comments on the manuscript.
Author responsibilities: G. D. F. Wilson prepared and illustrated the fossils, entered the data into the database, generated and edited the description, and determined the taxonomic classification of the fossils. J. R. Paterson and B. P. Kear interpreted the stratigraphy and palaeoecology.