EXCEPTIONAL PRESERVATION OF A NOVEL GILL
GRADE IN LARGE CRETACEOUS INOCERAMIDS:
SYSTEMATIC AND PALAEOBIOLOGICAL
IMPLICATIONS
by ROBIN I. KNIGHT1*, NOEL J. MORRIS2, JONATHAN A. TODD2,
LAUREN E. HOWARD3 and ALEXANDER D. BALL3
192 Vale Drive, Chatham, Kent ME5 9XA, UK; e-mail: [email protected] of Earth Sciences, Natural History Museum, Cromwell Road, London, SW7 5BD, UK; e-mails: [email protected], [email protected] Facilities Department, Imaging and Analysis Centre, Natural History Museum, Cromwell Road, London, SW7 5BD, UK; e-mails: [email protected],
[email protected]
*Corresponding author.
Typescript received 24 October 2012; accepted in revised form 24 March 2013
Abstract: Organised mineralised structures observed in
large inoceramids (valves on a metre scale) from the Late
Albian, Toolebuc Formation, Australia (Inoceramus suther-
landi McCoy, 1865), and the Santonian, Niobrara Formation,
USA (Platyceramus sp.), were investigated using variable
pressure scanning electron microscope (SEM) with energy-
dispersive X-ray spectroscopy (EDX), X-ray microcomputed
tomography (micro-CT) and X-ray diffraction (XRD) analy-
ses. These indicate that the structures comprised a phosphate
framework of aligned tubes and shallow troughs overlain per-
pendicularly by evenly spaced structures. In the Toolebuc
Inoceramus, these are U-shaped cross-structures, whilst in the
Niobrara Platyceramus, they comprise bundled fibre elements.
Comparison with modern bivalves indicates that the observed
phosphatised structures represent soft-tissue preservation of
the gills, as suggested in earlier publications. The tubes and
troughs are remnants of a filamental support framework com-
prising ordinary and primary filaments, whilst the U-shaped
cross-structures (I. sutherlandi) and fibrous bands (Platyce-
ramus) represent preserved longitudinal gill musculature.
Internal perforate and strand-like fabric observed on the inter-
nal surface of some Platyceramus tubular structures suggests
that the framework comprised collagen. The presence of
primary and ordinary filaments in numerous unusually large
plicae, in at least two lamellae, indicates that the gill structures
were heterorhabdic. Each plica has at least 40 ordinary
filaments, an exceptional number when compared with the
maximum of 20 present in modern heterorhabdic gills. The
absence of incontrovertible interfilament junctions makes it
difficult to say whether inoceramids were filibranch, pseudo-
lamellibranch or eulamellibranch. However, structures that are
best attributed to intraplical junctions between filaments sug-
gest the Inoceramidae had gills akin to those of pseudolamelli-
branch bivalves, although their unusually large number of
filaments per plica is more reminiscent of homorhabdic eu-
lamellibranch gills. The general form of the gill is similar to
that described in some other pteriomorphs, most specifically
Pteria. However, it has more complex junctions and intercon-
nections, although these are not as intricate or pervasive as
those observed in the pseudolamellibranch Ostreidae. The
connections and well-developed filament framework allowed
the gill to reach its unusually large size, supporting the large
size of these inoceramid species. The unusually large size of
the gill and its components indicate that the organism fed on
the larger suspended organic particles in the water column. It
would also have been capable of processing large volumes
of water quickly, leading to greater potential for food accumu-
lation and with likely implications for respiratory efficiency.
This may help explain the common association of inoceramids
with oxygen-deficient palaeoenvironments, particularly as
the general structure of the inoceramid gill is very different to
that observed in the commonest extant chemosymbiotic
bivalves.
Key words: inoceramid, gills, filament, musculature, fili-
branch, pseudolamellibranch, eulamellibranch.
SOFT-T I S SUE preservation does not commonly occur in
fossil Bivalvia. Those rare examples found in the fossil
record (Whyte 1992; Harper and Todd 1995; Wilby
and Whyte 1995; Klug et al. 2005; Dzik and Sulej
2007), as well as illuminating early diagenetic processes,
may also be of value for systematic and palaeobiological
analyses of bivalve families in the geological past. Such
soft-tissue information is of even greater importance in
© Crown copyright doi: 10.1111/pala.12046 37
[Palaeontology, Vol. 57, Part 1, 2014, pp. 37–54]
the case of extinct higher taxa such as the family Ino-
ceramidae.
The taxonomic affinities, evolutionary origins and life
habits of the Inoceramidae are still much debated. This
interest is largely driven by the importance of this family in
the Upper Cretaceous global oceans where taxa show par-
ticularly rapid evolution and wide distributions (Kauffman
1975; Dhondt 1992; Voigt 1995; Harries and Crampton
1998), making them of great value as biozonal and ancient
oceanographic environmental indices. Until now, most
information on the family’s soft-part anatomy has been
derived from attachment points of soft tissue to the shell
(muscle scars) (Koschelkina 1969; Kauffman and Powell
1977; Johnston and Collom 1998; Knight and Morris
2009). More intriguing, however, is a brief reference in the
literature to the preservation of gill structures in Platyce-
ramus platinus (Logan, 1898) from the Niobrara Formation
of Kansas (Western Interior Basin), USA (Stewart 1990).
If gill structures can be observed in the fossil Inoceram-
idae, then this will help in clarifying this extinct family’s
systematic affinities within the subclass Pteriomorpha,
whilst also giving an idea of feeding strategies and respi-
ratory characteristics. The latter is of particular interest
due to the unusual shell stable isotope signals that are
exhibited by some inoceramids, which have led to lengthy
consideration to the presence of possible chemosymbiont
associations in their gills (MacLeod and Hoppe 1992;
Grossman 1993). The prevalence of inoceramid preserva-
tion in organic-rich clays and oxygen-deficient environ-
ments has further added impetus to the postulated
chemosymbiotic gill association hypothesis (Sageman and
Bina 1997; Henderson 2004; Kauffman et al. 2007).
Stewart’s report led us to a detailed re-examination of a
specimen with soft-part preservation belonging to the
genus Platyceramus Seitz, 1967 (ex Heinz, 1932, nomen nu-
dum) that he donated to the collections at the NHM,
London, as well as an exquisite large specimen of Inoceram-
us sutherlandi McCoy, 1865 from the Late Albian, Toolebuc
Formation in Australia that exhibited similar soft-part pres-
ervation. This allowed us to examine soft-part structures
across two genera of large inoceramids and provide the first
detailed descriptions of the structure and diagenesis of a fossil
heterorhabdic gill (having two different forms of gill fila-
ment). Although the general features observed are reminiscent
of modern pseudolamellibranch gills, the overall structure is
different from anything else described in the literature.
MATERIAL EXAMINED
Platyceramus sp. (Natural History Museum, London:
NHMUK PI MB 1021) (Fig. 1C). A mineralised layer with
A
B
C
F IG . 1 . Inoceramid soft-part preservation. A–B, Inoceramus sutherlandi, Upper Albian, Toolebuc Formation, Australia. Part and
counterpart of specimen QMF34633 showing a structured mineralised layer adhering to the internal surface of the inner aragonite
layer. C, Platyceramus sp., Middle to Upper Santonian, Niobrara Formation, USA (NHMUK PI MB 1021). Structured mineralised layer
adhering to the internal surface of the shell fragment.
38 PALAEONTOLOGY , VOLUME 57
a well-defined structure that is directly adherent to the
internal surface of a fragment of external prismatic calcite
shell layer. Collected from the Smoky Hill Member of the
Niobrara Formation (Middle to Upper Santonian, Creta-
ceous; Scott and Cobban 1964) at Hills Bar, Kansas, USA
and donated by J. D. Stewart. This is a small (51 mm by
32 mm) fragment of an inoceramid that commonly grows
to an axial length of 1 m and in which some rare individu-
als can reach 2–3 m (Stewart 1990; Kauffman et al. 2007).
Inoceramus sutherlandi McCoy, 1865 (Queensland
Museum, Australia: QMF34633) (Fig. 1A–B). Large frag-
ments (roughly 186 mm by 92 mm) of inoceramid inner
aragonite nacre and outer shell prismatic calcite layers
with a mineralised layer adherent to the internal surface
that shows a well-defined structure (part and counter-
part). Collected from the Toolebuc Formation (Upper
Albian, Cretaceous; Henderson and Kennedy 2002) at
Siphon Paddock, Dunluce Street, near Hughendon, North
Queensland, Australia, by Drs A. Cook, M. Wade and E.
D. McKenzie. Intact I. sutherlandi specimens, including
the holotype, measured by Ludbrook (1966; holotype re-
figured pl. 8, fig. 1), indicate this species can also
approach an axial length of 1 m. The mineralised frame-
work studied is c. 210 9112 mm.
The small number of I. sutherlandi and Platyceramus
sp. specimens in the collections of the Natural History
Museum (NHM, London, UK) suggests that the two taxa,
like Inoceramus sensu stricto, are equivalved with the
umbones close to the anterior. The former species has a
slightly convex anterior margin and, unusually for the
family, has a very poorly developed lunule. Both inocer-
amid species have the multivincular hinge plate typical of
most other family members (Knight and Morris 2009).
ANALYSIS OF THE MINERALISEDSTRUCTURES
Specimens were examined using scanning electron micros-
copy (SEM; Philips XL30 FEG and LEO1455VP SEM) from
which images and elemental analyses were taken. Further
images for interpretation were derived from X-ray micro-
computed tomography, processed and viewed using the
SPIERS software package (http://spiers-software.org/), whilst
mineralogical analyses of the structures were undertaken
via X-ray diffraction (XRD) using the STOE powder dif-
fraction system. All imaging and analysis were carried out
in the Imaging and Analysis Centre at the NHM, London.
Preservation
Scanning electron microscopy elemental analyses
(LEO1455VP) of the structures observed in the Toolebuc
specimen indicate that they were preserved as calcium
phosphate (Fig. 2A). Associated cements and infills were
of calcite. XRD analyses of the structures in both speci-
mens indicate that the observed structures were preserved
in fluorapatite (Fig. 2B). The Australian specimen has a
better-defined crystalline structure than the American
one, which corroborates observations using SEM.
Gross morphology of structures
Hand specimen analyses using hand lens and light micro-
scopic examination of both specimens clearly show well-
organised and mineralised structures to be preserved. In
both cases, these take the form of an evenly spaced, lat-
tice-like framework of long rods, with evenly spaced
structures crossing them perpendicularly (Figs 3A–B, 7A).In the Toolebuc specimen, this appears to be one large
mass, whilst in Platyceramus, distinct separate packages
can be observed. SEM observations of the specimens indi-
cate that their rod structures are on the whole very simi-
lar, but there are key differences between specimens in
the structures preserved perpendicular to the rods.
Inoceramus sutherlandi. Diagenetic calcite cement
obscures the true nature of the structures preserved in the
Australian inoceramid gill material. Dissolution of these
cements in very weak acetic acid leaves the phosphate
framework intact, exhibiting extraordinarily well-pre-
served structures (nomenclature used to describe these
structures is illustrated in Fig. 4). The calcite cementation
has helped preserve the structures in an uncrushed state,
providing the best preservation of a fossil bivalve gill yet
described.
Robust rod/tube structures can be seen forming the
bulk of the framework, arranged 30–45 lm apart (rods
where the structures are totally infilled with cement, and
tubes where they are not) (Fig. 3B, D, F). U-shaped
cross-structures (spaced every 400–500 lm approxi-
mately) separate the rods and help demarcate layers, indi-
cating that they were stacked (Fig. 3A–B). The number of
layers within the stack is uncertain. The tubes were oval
in cross-section, with a long axis of around 45 lm and
short axis of 30 lm. Examination of the oval section of
the rods reveals that the majority of the tubes are hollow
and comprise two mineralised layers (Figs 2A, 3D–F).The oval section view of the tubes shows that the stacked
layers comprise folded units that lay on top of each other
(Fig. 3C–D). The tubes do not continue around the apex
of the fold, but are replaced by a shallow trough structure
running parallel to the long axes of the rods (Fig. 3E–F).Four distinct units can be seen where the folding occurs
in diametrically opposing directions, with paired units
present where the apical shallow trough structures are
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 39
A
B
F IG . 2 . Mineralogy of the inoceramid ultrastructural components. A, SEM elemental analyses of rod walls, infill and surrounding
cement in the Toolebuc inoceramid. B, XRD analyses of the mineralised organised structure in both Toolebuc and Niobrara specimens.
40 PALAEONTOLOGY , VOLUME 57
back-to-back (Fig. 3C). In places there appears to be a
join between the apical structures of each opposing unit
(Fig. 3E–F), but in general, they seem to be separate units
(Figs 3D, 6B).
The end of the fold unit opposite to that with the apical
shallow trough structure (‘distal’ with respect to the shal-
low troughs) was not commonly preserved intact. Instead,
it has become part of the structural melange of the long
section of stacked rods (Fig. 3A–B). In areas where there is
evidence that the folded units are joined distally, it is
unclear whether the rods return to enclose the apical struc-
tures or join them together (Fig. 5A). To clarify this aspect
of the morphology, we undertook micro-CT scans of a sur-
face lightly etched with dilute acetic acid. The resultant
images suggest that there are distal rod returns (Fig. 5B).
The joining of adjacent apical troughs via associated rod
rows (Fig. 5A, case c) seems most likely as there was a clear
divergence between these rod rows, suggesting that they did
not loop back distally so as to enclose the apical troughs
(Figs 5A, case b, 6B). In addition to the divergence evi-
dence, there are instances of poorly preserved distal rod
returns that suggest that the rod rows join so as to connect
adjacent shallow apical troughs (Fig. 6A).
The U-shaped cross-structures comprise a continuous
layer of fluorapatite, with a hollow centre (Figs 3B, 6C–F). They sit to one side of the rods, seemingly detached
from them (Fig. 6D–F), and are organised in juxtaposed
pairs to form long hollow structures. They can also be
observed on the flat surface parallel to the original shell,
although crushing and shear means the trough positions
are not as clear as in the sectional view (Figs 3B, 6E).
These juxtaposed U-shaped cross-structures form the cen-
tral separating layer of a two-rod layer composite. These
composite units were then organised back-to-back with
the rods sometimes interdigitated to apparently form one
layer, or alternatively were separated by calcite cement
(Fig. 3A–B, D). There are two pairs of small holes passing
through the whole trough-like structure (Fig. 6C–D, F).They are observed as paired perforations in the surface of
the outer cross-structure that were organised along the
juncture of the rod and trough mineralisation, with a cor-
responding set of perforations on the surface of the inner
cross-structure. The cross-sectional view suggests that
these holes were associated with projections from the rods
that pass thorough the structures that cross perpendicular
to them. The holes within the U-shaped tubes were fur-
ther aligned across the juxtaposed units suggesting the
rods were joined (Figs 3D, 6C, F).
In both structures, mineralisation of the outer wall is
fluorapatite, which has grown as crystallites perpendicular
to their edges (Figs 2A, 6G–H). Within the rods, a second
layer of euhedral radiating phosphate crystals can be
observed (Figs 2A, 6H). The surface of the internal phos-
phate layer appears to be textured, possibly moulding the
inner surface of the outer tube wall (Fig. 6G). When rods
are totally infilled, the occluding mineral is always calcite
(Fig. 2).
Platyceramus sp. The rod-like structures in the Platyce-
ramus specimen initially appear to be curved open struc-
tures (Fig. 7B–C), but closer inspection indicates that
they were originally hollow tubes. A lack of infilling
cement has led to the shell side of the rods breaking upon
compaction within the multilayered stack. This makes size
estimations difficult, although the crushed tube structures
were no more than 40 lm across.
In conjunction with the lack of infill, the tube walls of
this specimen are less crystalline and do not exhibit the
euhedral structures observed in the Toolebuc specimen
(Figs 2, 6G–H). Elongate indentations within a well-or-
ganised and strand-like fabric can be observed on their
inner surfaces (Fig. 7B). In places these can also be found
on the outer wall surface of the rods suggesting they may
have been perforate in nature (Fig. 7D).
The structures perpendicular to the rods in Platy-
ceramus differed markedly in form compared with those
in the Australian specimen. The general form of these
structures can be observed in hand specimen as 4 or 5
evenly spaced (approximately 600 lm apart) broad
strands crossing the rods (Fig. 7A, E). Closer scrutiny of
these structures using SEM indicates that they are thick
agglomerations of crystalline material seemingly organised
in discrete fibre bundles forming a rope-like morphology
(Fig. 7F). They do not interconnect the rods into a perfo-
rate framework, but rather lie across the framework as if
a separate entity (Fig. 7E). The fibre bundles lie perpen-
dicularly across packages of between 32 and 36 rods, with
each package being separated from each other by distinct
breaks (Fig. 7A).
PRESERVATION AND ANATOMICALINTERPRETATION OF THEMINERALISED STRUCTURES
Similar well-organised and regular structures observed in
the Toolebuc and Western Interior Seaway specimens
indicate unequivocally that biological structures are pre-
served in these inoceramids. Stewart (1990) believed that
these mineralised structures were remnants of the inocer-
amid bivalve gill structure.
Bivalve gills occur in four main forms: homorhabdic
filibranchs, heterorhabdic filibranchs, homorhabdic eu-
lamellibranchs and heterorhabdic pseudolamellibranchs
(Dufour and Beninger 2001). Heterorhabdic eulamelli-
branch gills are rare and only found in a few families
(Atkins 1937; Carter et al. 2012). Homorhabdic and
heterorhabdic refer to the gill having either only one type
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 41
of filament (homo) (Temkin 2006, fig. 13A) or two (het-
ero) (Beninger et al. 1988, fig. 2A–C; Cannuel and Benin-
ger 2006, fig. 6D–G; Temkin 2006, fig. 13C) (Fig. 8A–B),whilst the differentiation of filibranch, eulamellibranch
and pseudolamellibranch gill grades is linked to the nat-
ure of junctions between filaments, plicae and lamellae
(Mikkelsen and Bieler 2008; Carter et al. 2012). In fili-
branch gills, interfilament junctions are formed by inter-
twining cilia (e.g. as in the Pectinoidea (Beninger et al.
1988) and Mytilidae (Cannuel et al. 2009)), whereas in
pseudolamellibranch gills, they are formed by cilia and
organic tissue (Cannuel and Beninger 2006), and in eu-
lamellibranch gills, by organic tissue only (Gainey et al.
2003) (Fig. 8A, C). Interlamellar and intraplical junctions
are formed of organic tissue (Fig. 8A) in all the gill
grades, although the homorhabdic eulamellibranch lamel-
lae are conjoined by much larger interlamellar septa (Du-
four and Beninger 2001; Gainey et al. 2003, figs 1C, 6F).
These large interlamellar septa in the eulamellibranch gill
are associated with more elongated filaments (frontal to
abfrontal), another character that helps differentiate the
eulamellibranch gill from the pseudolamellibranch gill. In-
traplical junctions are most closely associated with the
increasing architectural complexity observed in the
pseudolamellibranch bivalves (Cannuel and Beninger
2006, fig. 6F–G), and are uncommon in the other grades.
Comparison with extant bivalve gills indicates that the
observed large-scale organised lattice structure we observe
must be the remnants of the gill structure in these speci-
mens (Beninger et al. 1988; Cannuel and Beninger 2006;
Temkin 2006). The structural preservation could be the
remnants of a gill support framework as seen in some
bivalve families (Beninger et al. 1988; Medler and Silver-
man 1997, 1998; Dufour and Beninger 2001; Cannuel and
Beninger 2006) and/or phosphatisation of soft tissue
(Briggs and Kear 1993; Briggs et al. 1993; Wilby and
Whyte 1995; Wilby and Briggs 1997).
There have been numerous references to solid frame-
work structures within all the gill grades observed in
extant bivalves. Yonge (1926) and Cox (1969) alluded to
‘chitinous’ fibrous structures occurring in some bivalve
gills, whilst slices through gills of Lucinidae and Placopec-
ten (Pectinidae) indicate collagen supports along the axes
of filaments (Le Pennec et al. 1988; Frenkiel et al. 1996).
Ridewood (1902) also indicated the presence of ‘chitin-
ous’ fibres helping support the bivalve gill, whilst high-
lighting the presence of discontinuous phosphatic ‘rods’
along the filament axes of the Unionidae. These ‘rods’ in
the Unionidae are also recorded by Whyte (1992) in both
modern and fossil specimens, along with similar struc-
tures in fossil Trigoniidae. Observations on the decay of
modern Unio gills have shown morphologically similar
structures to occur at about the same scale as those
observed in this study (Skawina 2010) that resulted from
gill frameworks remaining after soft-tissue breakdown.
The presence of collagen fibres supporting filaments
among a diverse range of taxa (Dreissena polymorpha
(Dreissenidae), Corbicula fluminea (Corbiculidae) and
Toxolasma texasensis (Unionidae)), together with cross-
strut continuations of the fabric joining the framework
together between ostia, has been described by Medler and
Silverman (1997, 1998).
Although the rod structures observed in Inoceramidae
from this study resemble gill support frameworks, they
differ in being hollow and phosphatic rather than solid
and chitinous (Figs 3A, C–G, 7A–C). This indicates that
the original framework is not preserved, but that a phos-
phogenic process has formed a facsimile of the original
structure. Work on the taphonomic processes within the
breakdown of the gill of Unio suggests that all original
soft tissue is lost (Skawina 2010), and that its breakdown
may lead to phosphogenesis around the gill support
framework. The breakdown of soft tissue within an
enclosed environment produces high concentrations of
phosphate ions, plus the right conditions to precipitate
calcium phosphate (Briggs and Kear 1993; Briggs et al.
1993; Wilby and Whyte 1995; Wilby and Briggs 1997).
Inoceramid gill preservation
We consider that the tubular rod structures represent the
preservation of the inoceramid gill support framework via
phosphate formation using the framework as a precipita-
tion surface. In both inoceramid specimens studied,
the phosphate completely ‘coated’ the framework, and the
later loss of the embedded framework lattice led to the
formation of tubes. In the case of the Toolebuc specimen,
F IG . 3 . Inoceramus sutherlandi McCoy 1865 (QMF34633), Upper Albian, Toolebuc Formation, Australia. Mineralised structures. A,
rod/tube and cross-strut framework visible at low magnifications (long-section view looking down on the length of the rods/tubes). B,
stacking of rods/tubes and well spaced U-shaped cross-structures. C, tube oval section view (looking down the end of the rods/tubes)
indicating that there are four distinct organised layers of rod/tube fold unit preserved (representing four different lamella). D, tube
oval section view of the framework structures with a closer look at the rod/tube fold unit (filaments within a plica) indicating that they
are joined up by U-shaped cross-structures (longitudinal musculature) (Note two sets of loops diametrically arranged representing dif-
ferent lamellae). E, the apex of the loops are defined by a shallow trough structure (primary filament) that appears to join similar api-
cal structures directly behind. F, two sets of joined rods/tubes and shallow apical trough structures backed against and apparently
joined to each other.
42 PALAEONTOLOGY , VOLUME 57
A B
C D
E
F
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 43
the tubes were then infilled via a second phase of phos-
phate and calcite formation (Fig. 2A). This probably led
to further crystallisation of the original soft-tissue miner-
alisation, with the original good preservation of detail
seen in Platyceramus being lost (Figs 6G–H, 7). The
fibrous, perforate fabric observed in the American speci-
mens could represent a mould of the original collagen
framework (Fig. 7B).
We believe that similar phosphate ‘coating’ of a decay-
resistant gill component led to the formation of the
U-shaped cross-structures in I. sutherlandi (Figs 3B–D,6B–F), with the fibrous cross-structure in Platyceramus
representing the exceptional preservation of this gill com-
ponent (Fig. 7E–F). These organised structures are closely
associated with the gill support framework, but do not
appear to be integral to it. The paired perforations occur-
ring through, and aligned across, the U-shaped cross-
structures suggest a gill component that was initially
decay resistant, but not a good substrate for phosphogen-
esis (Fig. 6C–D, F). After initial phosphate mineralisation
occurred on the larger U-shaped components, voids sub-
sequently formed due to later decomposition of the
enclosed gill component. The preservation of the original
gill support structure as voids suggests that phosphogene-
sis was an early, rapid diagenetic event. Later reminerali-
sation and calcite cement formation suggest that changing
conditions within the sediment favoured decomposition
of the initially decay-resistant components.
Lamellae, plicae and filaments
The general morphology and size of the tube structures
observed in Platyceramus and the Toolebuc material sug-
gests that they are a record of the filament support frame-
work in these Inoceramidae. The organised rod bundles
are of a similar scale to the filaments observed in the ho-
morhabdic eulamellibranch gills of lucinids (Frenkiel
et al. 1996). However, within these large lucinid filaments,
only one collagen support is observed (i.e. the arrangement
observed in modern bivalves (Yonge 1926; Cox 1969; Le
Pennec et al. 1988; Medler and Silverman 1997, 1998)),
rather than numerous smaller elements. It is also difficult
to envisage how rod interdigitation could occur if they
were originally separated by the flesh of the filament. This
would be more easily achieved if there were a space
between the rod layers on this unsupported side, allowing
the structures to collapse together upon death.
F IG . 4 . Nomenclature used to
describe the phosphatised ultrastruc-
tural components observed in Inoce-
ramus sutherlandi.
44 PALAEONTOLOGY , VOLUME 57
Therefore, we consider the apparent folded units most
likely to be plicae, with the loop returns joining adjacent
shallow apical troughs (Figs 3D, 6A). These plicae were
observed as distinct breaks between rod bundles in Platy-
ceramus (Figs 7A, 8A). Each rod represents an individual
filament, whilst the shallow troughs at the apex of the
Toolebuc specimen plicae were primary filaments (Figs 3D–F,
9A). This indicates that the whole structure was heterorhabdic
and similar to that in modern Pteria, Pecten and Crassostrea
(Beninger et al. 1988, fig. 2A–C; Cannuel and Beninger 2006,
fig. 6D–G; Temkin 2006, fig. 13C).
Comparison with heterorhabdic modern bivalves indi-
cates that there were an unusually large number of ordin-
ary filaments per plica in the Toolebuc specimen (at least
A
B
F IG . 5 . The ‘loop’ back conun-
drum relating to the ends of the rod/
tube rows opposite the shallow apical
trough (primary filament). A, are the
ends of the rod/U-shaped cross-
structure composite units open (a)?
Alternatively, are the rod/tube rows
joined up in some way? Either
enclosing the shallow apical trough
(b) or joining together the rod/U-
shaped composite (c)? B, micro-CT
scan slices at the etched surface of the
Toolebuc specimen showing loop
backs at the rod/tube strip extremi-
ties; their relationship to the shallow
apical trough is uncertain, although
Figure 6A suggests that option (c) is
most likely, as does the divergence of
the rod/tube rows in Figure 6B.
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 45
40 compared with a maximum of 20 in most extant
bivalve gills) (Fig. 3C–D). The Platyceramus specimen
from the USA would seem to preserve half a plica with at
least 18 filaments (i.e. half of the 36 closely packed rods,
with the other half representing structures from the
underlying plica).
In the Australian inoceramid gill, the presence of four
separately organised fold units indicates that at least two
lamellae (ascending and descending) of one demibranch
are preserved (Figs 3C, 8A). It may be that the paired units
represent two sets of lamella in one demibranch organised
within a W-configuration, or alternatively, there may be
two separate demibranchs with single lamella V-configura-
tions, as described by Atkins (1937) in Heteranomia (Pteri-
omorpha: Anomiidae). This is an important difference, but
it cannot be ascertained from this specimen.
Lamellae can commonly be seen to be organised rela-
tive to each other in the Toolebuc specimen, and may be
joined by interlamellar junctions at the primary filaments.
However, the mineralisation process has made it difficult
to confirm the presence of these junctions. The side-by-
side joining of apical troughs from different lamella
(Fig. 3E–F) would appear to be a mineralisation effect,
especially upon consideration of interlamellar junctions in
extant heterorhabdic bivalves in which the primary fila-
ments are arranged directly back-to-back (Temkin 2006,
fig. 13C). Although the distinct rod clusters observed in
Platyceramus do represent plicae, it is unclear whether
they are from the same or different lamellae.
Musculature and interfilament junctions
Understanding the origins of the U-shaped cross-structures
perpendicular to the filaments is slightly more problematic.
Their positioning relative to the tubes, as well as their unu-
sual large size, is not comparable with the cross-struts
known from modern bivalve gill support frameworks. This
suggests they may represent a gill component other than
the gill support framework cross-strut (whether filamen-
tous or tissue in nature), but closely associated with it.
In the Toolebuc Inoceramus, the position of the
U-shaped cross-structure relative to the primary filament
within each plica indicates that it was located abfrontally
(Figs 3C–E, 9A). It represents a gill component that was
initially more resistant to decomposition and that became
a substrate upon which phosphatisation occurred. This
component was later lost leaving the void in the U-
shaped structure (Fig. 6D–E). In addition, the presence of
paired perforations on the inner and outer surfaces of the
U-shaped cross-structure, together with associated projec-
tions from the rods, suggests connective tissue was pres-
ent between the adjacent filaments in a plica (Figs 3D,
6C–D, F, 9B). In Platyceramus, there were no U-shaped
cross-structures, but rather a wide, flat fibrous structure
(Fig. 7A, E). Its general size and shape is comparable with
the Toolebuc U-shaped cross-structure, as is the spacing
between neighbouring structures. This suggests these mor-
phologies may represent different preservational states of
the same gill component.
The fibre bundles that are seen crossing the separated
swaths of gill support framework in Platyceramus are
reminiscent of phosphatised musculature (Wilby and
Briggs 1997; Fig. 7F). The nature and organisation of
these structures across the framework suggests that they
represent longitudinal gill muscles that altered the size of
ostia within the gill, which in turn controlled water flow
through the structure (Medler and Silverman 1997, 1998;
Fig. 9A–B). A major difference between modern bivalve
longitudinal gill musculature, as described in the litera-
ture, and the structures observed in Platyceramus, is that
the latter were much larger and much more widely spaced
(see Medler and Silverman 1997, 1998, Gainey et al. 2003
and Gainey 2010 for comparison with modern bivalves).
The larger size, and wider spacing, of the longitudinal gill
musculature in the two inoceramid specimens may have
aided these particular species to attain their unusual size.
The close juxtaposition of the longitudinal muscle on
each side of the plicae appears to be unusual when
compared with modern bivalves (Fig. 9A). However, this
tissue association was observed throughout the Toolebuc
specimen, suggesting that it was the normal organisation
of the gill, at least proximal to the primary filament
(Fig. 3B–F). The postulated tissue joins between adjacent
filaments through the longitudinal musculature would
hold the whole framework together in this ‘closed’ state.
F IG . 6 . Inoceramus sutherlandi McCoy 1865, QMF34633, Upper Albian, Toolebuc Formation, Australia. Mineralised structures. A,
rod/tube organisation at the opposite end of the fold structures to the apical structure suggest that the rod/tube layers separated by
the U-shaped cross-structures are joined together so as to form continuous loop structures that leave the apical structures open
(Fig. 4A, case c). B, rod/tube loops diverging at the opposite apices to those with the shallow apical trough structures and leaving a
wide access to the loop-joining structures. C, holes through the U-shaped cross-structures (longitudinal musculature) associated with
rods/tubes (filaments) and aligning across the loop (plica). D, the U-shaped cross-structures are hollow and have paired perforations
passing through the entire structure. E, hollow, aligned U-shaped cross-structures forming tubular structures. F, U-shaped cross-struc-
tures forming a tube and with associated joining structures. G, rods with a distinct outer layer, and solid, textured inner core. H, euhe-
dral, tangential crystals of the rod walls with microcrystalline infill.
46 PALAEONTOLOGY , VOLUME 57
A B
C
D
E
F
G
H
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 47
The space generated between the filaments on either side
of one plica in the inoceramid is comparable with that
observed in modern bivalves (Cognie et al. 2003; Cannuel
and Beninger 2006). The main argument against these
structures being intraplical junctions is that, on occassion,
the primary filaments within a lamella are slightly
A B
C
D
E
F
F IG . 7 . Platyceramus sp., NHMUK PI MB 1021, Middle to Upper Santonian, Smoky Hill Member, Niobrara Formation, USA. Min-
eralised structures. A, framework of rods/tubes visible in discrete packages, with large disorganised bands crossing perpendicular to
their long axes. B, crushed rods/tubes with a perforate internal structure. C, crushed rods/tubes stacking on top of each other suggest-
ing several layers. D, the perforate internal structure of the rods/tubes. E, band that crosses perpendicular to the long axes of the rod/
tube and lies on top of the framework. F, close-up of the band indicating that it is a rope-like, organised fibre bundle suggesting it is
some form of preserved musculature.
48 PALAEONTOLOGY , VOLUME 57
staggered (Fig. 3D). Although it is hard to envisage tissue
joins in these regions of the framework, this misalignment
may be due to crushing and shear leading to breaking of
the joining tissue (Figs 3E–F, 6B).Interfilamental tissue junctions are features characteris-
tic of pseudolamellibranch and eulamellibranch gills
(Mikkelsen and Bieler 2008; Carter et al. 2012). However,
tissue intraplical structures mainly occur at junctions
between primary filaments in pseudolamellibranch bival-
ves (Cannuel and Beninger 2006). We consider that there
were intraplical structures in I. sutherlandi, and we
observed that there were also occasional narrow joins
between filaments that were in line with the longitudinal
muscle U-shaped cross-structures (Fig. 6C). However,
they cannot be convincingly interpreted as interfilamental
junctions, but rather could be an artefact resulting from
voids being filled by phosphatisation.
IMPLICATIONS FOR THESYSTEMATICS AND PALAEOBIOLOGYOF THE INOCERAMIDAE
The structural components and organisation of the bivalve
gill are taxonomically defining characteristics at family
level. Therefore, it would be reasonable to assume that at
least all Cretaceous species of Inoceramus sensu stricto and
Platyceramus had a similar general gill form to that
described for the two large species considered in this study.
The heterorhabdic organisation with 30–45 lm diameter
ordinary filaments would not be unexpected in other nor-
mal-sized inoceramid species, and is only slightly larger
than the dimensions observed in extant families (Cannuel
and Beninger 2006; Temkin 2006). However, the number
of filaments per plica in inoceramid species other than
those in Inoceramus sensu stricto and Platyceramus could
A
B
C
F IG . 8 . Simplified model of the
bivalve gill illustrating terminology
used in this study. A, illustrates the
large-scale characters relating to
demibranchs, lamellae and fila-
ments. B, illustrates the concept of
ordinary and primary filaments in
plicae, and what that means in
terms of whether a gill is homo- or
heterorhabdic. C, illustrates the dif-
ferent nature of joining structures in
filibranch, pseudolamellibranch and
eulamellibranch gills.
KN IGHT ET AL . : INOCERAMID GILL PRESERVAT ION 49
be expected to be far fewer. The unusually well-developed
longitudinal musculature and intraplical junctions between
all the filaments may also be a specific adaptation within
Inoceramus sensu stricto and Platyceramus that allowed
them to develop unusually large gills, and so to have the
potential to reach unusually large sizes.
Inoceramid systematics
The heterorhabdic gill form observed in these large inoc-
eramid species is typical of other extant pteriomorph
families such as the Pteriidae and Pectinidae (Beninger
et al. 1988; Temkin 2006). However, the gill of the extant
family Isognomonidae, with which the Inoceramidae are
often compared due to their shared multivincular hinge
plate, is homorhabdic and thus very different (Temkin
2006, fig. 13A).
Subtle differences between the form of the ligament pit
organisation of Cremnoceramus and Isognomon, along
with the pattern of pedal muscle attachment, led Johnston
and Collom (1998) to suggest removal of Inoceramidae
from the order Pteriomorpha and relate the family to the
subclass ‘Cryptodonta’. Temkin’s (2006) study of the
Pteriomorpha indicates that multiple ligament pits proba-
bly arose more than once within that order. It therefore
should not come as a surprise that the ligament pit orga-
nisation is not identical in the different clades. We also
A
B
F IG . 9 . Interpretation of the struc-
tures observed in Inoceramus suther-
landi and Platyceramus sp. A,
intraplical arrangement of ordinary
and primary filaments, with possible
interfilamental and intraplical junc-
tions interacting with the longitudi-
nal musculature. B, the intraplical
junctions passing through the
unusually well-developed longitudi-
nal musculature.
50 PALAEONTOLOGY , VOLUME 57
note that the pattern of muscle attachment, preserved in
many inoceramids, is similar to that observed in modern
Isognomon (Knight and Morris 2009, Fig. 4A), which we
would interpret as pedal byssal attachment scars (Stanley
1972).
Unfortunately, the only description of a ‘cryptodont’
gill is that of the extant protobranch Solemya (Reid 1980;
Zardus 2002), which we agree with Johnston and Collom
(1998), does not fit comfortably within the ‘Cryptodonta’.
On the basis of this investigation (and an earlier one on
inoceramid hinge plate origins (Knight and Morris
2009)), we would suggest that the Inoceramidae fit well
within the Pteriomorpha. Any relationship between the I-
noceramidae and ‘Cryptodonta’ (Cardiolidae and Slavi-
dae) is based upon superficial resemblances inferred from
an inadequate first-hand consideration of actual fossil
material (Kříž 2007).
The large inoceramids considered in this study,
although having a heterorhabdic gill, differ from the Pter-
iidae and Pectinidae in their increased complexity of join-
ing structures within plicae. This feature is reminiscent of
that within the pseudolamellibranch Ostreidae, although
it is less well developed than the interlocking structures
observed in the oyster gill (Cannuel and Beninger 2006).
The large inoceramid intraplical joins were probably an
adaptation to help maintain the integrity and solidity of
the large gill structure. However, this ability to develop
such structures suggests other inoceramid species may
also have rudimentary tissue-joining structures within
their gills. This means they cannot be considered fili-
branchs, but conversely cannot accurately be classified as
pseudolamellibranchs. The large number of ordinary fila-
ments in each plica observed in the inoceramid gill is not
present in modern heterorhabdic bivalves (Beninger et al.
1988, Fig. 2A–C; Cannuel and Beninger 2006, Fig. 6D–G;Temkin 2006, fig. 13C). Only the homorhabdic eulamelli-
branch bivalves have this number of ordinary filaments
per plica (Gainey et al. 2003, figs 1C, 6C, F), with even
the rare heterorhabdic eulamellibranchs such as the
anomalodesmatids having fewer than 20 (Carter et al.
2012, fig. 134). The gills of studied inoceramids, there-
fore, represent a grade form that as yet has not been
described. However, we will defer formal definition of a
new gill grade pending more complete morphological
information.
Inoceramid palaeobiology
The wide spacing of the musculature (which is usually
closely associated with interfilament junctions whether cil-
iary or tissue) in the two species of inoceramid studied
would suggest that the ostia in these extinct bivalves were
relatively large. The maximum spacing of the cross-struts
is around 100 lm in modern eulamellibranch bivalves
(Medler and Silverman 1998) and around 300 lm in
modern pseudolamellibranch bivalves (Gainey 2010),
compared with around 500 lm observed in I. sutherlandi
and Platyceramus sp. (Figs 3A–B, 7A). Ostia size would
have been able to be further altered by the longitudinal
musculature (Medler and Silverman 1997, 1998; Gainey
2010), but when open, the gills evidently could have rap-
idly filtered unusually large volumes of water. The well-
developed musculature would aid in pumping the water
quickly and efficiently through the gill system (Medler
and Silverman 1997, 1998). This would allow for
increased potential for suspension feeding within the
structure as recorded in living Acesta excavata and Pinct-
ada margaritifera (Pouvreau et al. 1999, 2000; J€arnegren
and Altin 2006). The large size of the gill and amounts of
water passing through it may have increased gaseous
exchange, but oxygen uptake itself reflects the metabolic
rate of the bivalve (Bayne and Newell 1983). The large gill
may help explain why the studied inoceramids, and possi-
bly other smaller species, seemed to have thrived in low-
oxygen environments (Sageman and Bina 1997; Hender-
son 2004; Kauffman et al. 2007).
Large gill structures are often associated with bivalves
that have chemosymbiotic associations within their gills
(Alyakrinskaya 2003; Taylor and Glover 2006). Con-
versely, there has not been any chemosymbiotic associa-
tions reported for large gilled bivalves such as Acesta
excavata, Crassostrea gigas and Pinctada margaritifera. The
family Lucinidae is typified by the presence of chemo-
symbionts within the gills (Taylor and Glover 2006). This
family is eulamellibranch and has long fleshy filaments
(Frenkiel et al. 1996, figs 2, 3) that can be as long as the
inoceramid plica. The chemosymbiotic bacteria are found
in the large area of fleshy filament directly behind the
frontal surface with collagen support structure (Distel and
Felbeck 1987, fig. 2B). This is not comparable with the
inoceramid gill observed in this study, which has closely
packed filaments within each plica, and so has little free
‘tissue’ area in which to house bacteria. The Inoceramidae
could have an extracellular chemosymbiotic association
on the filament abfrontal surfaces as in the homorhabdic
filibranch family Thyasiridae (Dufour 2005). However,
this would suggest poor development of the chemosymbi-
otic association as full development within the Thyasiri-
dae is associated with considerable extension of the
filament abfrontal surface. If this were the case, then the
inoceramid would not gain the full possible benefit from
the association. As such, we believe that the inoceramids
did not have chemosybiotic associations within their gills,
but relied on the unusual size and morphology of their
gills to survive in low-oxygen environments.
The formation of heterorhabdic plicae (Figs 3C–D; 6A–B) indicates that the gill could have filtered larger seston
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 51
particles (Ward et al. 1998; Cognie et al. 2003), the size
of food particle only being limited by the width of the
primary filament groove. Adults of the oyster Crassostrea
gigas have a primary filament width of around 35 lmand can filter large suspended algae (Cognie et al. 2003).
In comparison, the primary filament of I. sutherlandi is
around 100 lm. This, along with a slightly larger-than-
usual spacing between filaments, indicates that some
species could have been bacteriophages as suggested by
Henderson (2004), feeding on large bacterial fragments
broken from seafloor bacterial mats by strong currents.
CONCLUSIONS
This study has identified the exceptional preservation of
the gill support framework and associated musculature of
two Cretaceous giant inoceramid species. Preservation has
resulted from diagenetic phosphate ‘coating’ of the more
decay-resistant collagen musculature cores of filaments to
form rods and tubes (Figs 2A, 3, 7B–C), as well as pri-
mary filaments (Fig. 3D–F) and unusually large longitudi-
nal musculature of the gill (Figs 3B–D, 6C–F, 7A, E–F).The primary filaments are preserved as shallow apical
troughs that may interconnect across lamellae and indi-
cate the presence of unusually large plicae comprising in
excess of 40 filaments (Fig. 3C–D). The musculature
component is preserved as hollow U-shaped cross-
structures in I. sutherlandi, but is clearly observed as muscle
bands in Platyceramus (Fig. 7E–F).Intraplical connections are indicated by paired perfora-
tions through, and aligned across, U-shaped cross-
structures (Fig. 6C–D, F). The original joining structures
must have originally been resistant to decomposition, but
not a good substrate for phosphogenesis. After the initial
phosphate mineralisation they underwent decay, and were
lost, suggesting phosphogenesis to have occurred relatively
rapidly soon after death of the organism. Later recrystalli-
sation of phosphate components and the formation of
calcite cement indicate changes in the conditions within
the sediment.
The presence of primary filaments in a gill structure
that appears to have complex tissue intraplical, and possi-
bly interfilamental and interlamellar, connections would
suggest that these large inoceramids be loosely classified
as being heterorhabdic pseudolamellibranchs (Figs 3C–E,6C–D). However, they lack the more complex intercon-
nections observed in modern pseudolamellibranch bival-
ves (Cannuel and Beninger 2006, fig. 6D–G). The large
number of filaments per plica, presence of intraplical
junctions throughout each plica and well-developed longi-
tudinal musculature were extraordinary features that aided
the development of the gills’ large size, and are not
observed in modern heterorhabdic bivalves. We expect
other inoceramid species of Inoceramus sensu stricto and
Platyceramus to have had a similar general gill morphol-
ogy to that observed in the two large species in this study
(i.e. primary and ordinary filaments), indicating that they
were heterorhabdic. The presence of intraplical tissue-
joining structures suggests that the family tends towards
the pseudolamellibranch rather than filibranch grade. The
form of the gill organisation supports the view that
Inoceramidae do fall within the Pteriomorpha; the sug-
gestion that the musculature and ligament construction
suggest otherwise (Johnston and Collom 1998) is no
longer supported.
The general scale of the components of the gill in the
larger inoceramids studied were of a similar magnitude
to those observed in modern bivalves. However, relative
positioning of the musculature and rods indicates that
the ostia were larger than those observed in most mod-
ern bivalves. This along with the large size of the
primary filament suggests that larger food particles could
be filtered. The large size of the gill, and the arrange-
ment of its components, may also have implications
with regard to how these inoceramids survived in low-
oxygen environments (Sageman and Bina 1997; Hender-
son 2004), both in terms of optimising oxygen supply
and maximising food supply. The inoceramid gill does
not seem to have the same structures as the more com-
mon bivalves with chemosymbiotic gill associations. We
consider it more likely that the unusual size and struc-
ture of the gills allowed some taxa to thrive in oxygen-
deficient environments in the absence of chemosynthetic
gill bacteria.
Acknowledgements. We would like thank Dr John Taylor for his
expert advice relating to modern bivalve gills, Dr Russell
Garwood who ran our specimens through the micro-CT Scan
and then taught us how to process and view the results using
the SPIERS software package (http://spiers-software.org/), and
Dr Gordon Cressey who undertook the XRD analyses. We would
also like to acknowledge the contributions of the referees for this
manuscript, Dr James Crampton and Dr Ireneusz Walaszczyk,
whose suggestions and questions led to improvements in the
research and text. We are indebted to the Queensland Museum
for the loan of the exquisite specimen from the Toolebuc For-
mation (Australia) and to Mr J. D. Stewart for donating a frag-
ment of Platyceramus with adhering gill from the Niobrara
Formation (USA).
Editor. John Jagt
REFERENCES
ALYAKRINSKAYA, I. O. 2003. Tissue hemoglobins in Bival-
via (Mollusca). Biology Bulletin, 30, 617–626.ATKINS, D. 1937. On the ciliary mechanisms and interrela-
tionships of Lamellibranchs. Part III: Types of Lamellibranch
52 PALAEONTOLOGY , VOLUME 57
gills and their food currents. Quarterly Journal of Microscopical
Science, 2, 375–421.BAYNE, B. L. and NEWELL, R. C. 1983. Physiological ener-
getics of marine molluscs. 407–515. In SALEUDDIN,
A. S. M. and WILBUR, K. M. (eds). The Mollusca, Physio-
logy, Part 1. Academic Press, New York, 522 pp.
BENINGER, P. G., LE PENNEC, M. and SALA €UN, M.
1988. New observations of the gills of Placopecten magellanicus
(Mollusca: Bivalvia), and the implication for nutrition. I. General
anatomy and surface microanatomy.Marine Biology, 98, 61–70.BRIGGS, D. E. G. and KEAR, A. J. 1993. Fossilization of soft
tissues in the laboratory. Science, 259, 1439–1442.--MARTILL, D. M. and WILBY, P. R. 1993. Phos-
phatization of soft-tissue in experiments and fossils. Journal of
the Geological Society, 150, 1035–1038.CANNUEL, R. and BENINGER, P. G. 2006. Gill develop-
ment, functional and evolutionary implications in the Pacific
oyster Crassostrea gigas (Bivalvia: Ostreidae). Marine Biology,
149, 547–563.--MCCOMBIE, H. and BOUDRY, P. 2009. Gill
development and its functional and evolutionary implications
in the Blue Mussel Mytilus edulis (Bivalvia: Mytilidae). Biologi-
cal Bulletin, 217, 173–188.CARTER, J. G., HARRIES, P., MALCHUS, N., SAR-
TORI , A., ANDERSON, L., B IELER, R., BOGAN, A.,
COAN, E., COPE, J., CRAGG, S., GARCIA-MARCH,
J., HYLLEBERG, J., KELLEY, P., KLEEMAN, K.,
KRIZ, J., MCROBERTS, C., MIKKELSEN, P., POJE-
TA, J. Jr, TEMKIN, I., YANCEY, T. and ZIERITZ, A.
2012. Part N, Revised, Volume 1, Chapter 31; Illustrated glos-
sary of the Bivalvia. Paleontological Institute, University of
Kansas. Treatise Online, 48, 1–209.COGNIE, B., BARILL�E, L., MASS�E, G. and BENINGER,
P. G. 2003. Selection and processing of large suspended algae
in the oyster Crassostrea gigas. Marine Ecology Progress Series,
250, 145–152.COX, L. R. 1969. General features of Bivalvia, N3–N121. In
MOORE, R. C. (ed.). Treatise on Invertebrate Paleontology.
Part N. Mollusca, 6. Geological Society of America, Boulder,
Colorado and University of Kansas Press, Lawrence, Kansas,
489 pp.
DHONDT, A. V. 1992. Cretaceous inoceramid biogeography:
a review. Palaeogeography, Palaeoclimatology, Palaeoecology, 92,
217–232.DISTEL, D. L. and FELBECK, M. 1987. Endosymbiosis in
the lucinid clams Lucinoma aequizonata, Lucinoma annulata
and Lucina floridana: a re-examination of the functional mor-
phology of the gills as bacteria-bearing organs. Marine Biology,
96, 79–86.DUFOUR, S. C. 2005. Gill anatomy and the evolution of sym-
biosis in the bivalve family Thyasiridae. Biological Bulletin,
208, 200–212.-and BENINGER P. G. 2001. A functional interpretation
of cilia and mucocyte distributions on the abfrontal surface of
bivalve gills. Marine Biology, 138, 295–309.DZIK, J. and SULEJ , T. 2007. A review of the early Late
Triassic Krasiejow biota from Silesia, Poland. Palaeontologia
Polonica, 64, 3–27.
FRENKIEL, L., GROS, O. and MOU€EZA, M. 1996. Gill
structure in Lucina pectinata (Bivalvia: Lucinidae) with refer-
ence to hemoglobin in bivalves with symbiotic sulphur-oxidis-
ing bacteria. Marine Biology, 125, 511–524.GAINEY, L. F. Jr 2010. Seasonal potentiation of gill muscle
contraction in four species of bivalve molluscs. Journal of
Experimental Marine Biology and Ecology, 391, 43–49.-WALTON, J. C. and GREENBERG, M. J. 2003. Bran-
chial musculature of a venerid clam: pharmacology, distribu-
tion and innervation. Biological Bulletin, 204, 81–95.GROSSMAN, E. L. 1993. Evidence that inoceramid bivalves
were benthic and harbored chemosynthetic symbionts: com-
ment and reply. Geology, 21, 94–96.HARPER, E. M. and TODD, J. A. 1995. Preservation of the
adductor muscle of an Upper Jurassic oyster. Pal€aontologische
Zeitschrift, 69, 55–59.HARRIES, P. J. and CRAMPTON, J. S. 1998. The inocera-
mids. The American Paleontologist, 6 (4), 2–6.HEINZ, R. 1932. Aus der neuen systematic der Inoceramen
(Inoceramen XVII). Mitteilungen aus dem Mineralogisch –Geologischen Staatsinstitut in Hamburg, 15, 1–26.
HENDERSON, R. A. 2004. A Mid-Cretaceous association of
shell beds and organic rich shale: bivalve exploitation of
a nutrient-rich, anoxic sea-floor environment. Palaios, 19,
156–169.-and KENNEDY W. J. 2002. Occurrence of the ammonite
Goodhallites goodhalli (J. Sowerby) in Eromanga Basin,
Queensland; an index species for the late Albian (Cretaceous).
Alcheringa, 26, 233–247.J €ARNEGREN, J. and ALTIN, D. 2006. Filtration and respi-
ration of the deep living bivalve Acesta excavata (J. C. Fabri-
cius, 1779) (Bivalvia; Limidae). Journal of Experimental Marine
Biology and Ecology, 334, 122–129.JOHNSTON, P. A. and COLLOM, C. J. 1998. The bivalve
heresies – Inoceramidae are Cryptodonta, not Pteriomorphia.
347–360. In JOHNSTON, P. A. and HAGART, J. W. (eds).
Bivalves; an eon of evolution – palaeobiological studies honouring
Norman D. Newell. University of Calgary Press, Calgary, 461 pp.
KAUFFMAN, E. G. 1975. Dispersal and biostratigraphical
potential of Cretaceous benthonic Bivalvia in the Western
Interior. In CALDWELL, W. G. E. (ed.). The Cretaceous sys-
tem in the Western Interior of North America. Special Paper of
the Geological Association of Canada, 13, 163–194.-and POWELL J. D. 1977. Paleontology. In KAUFF-
MAN, E. G., HATTIN, D. E. and POWELL, J. D. (eds).
Stratigraphic, paleontologic and paleoenvironmental analysis of
the Upper Cretaceous rocks of the Cimarron County, Northwest-
ern Oklahoma. Memoirs of the Geological Society of America,
149, 47–114.-HARRIES , P. J., MEYER, C., VILLAMIL, T., AR-
ANGO, C. and JAECKS, G. 2007. Paleoecology of giant
Inoceramidae (Platyceramus) on a Santonian (Cretaceous)
seafloor in Colorado. Journal of Paleontology, 81, 64–81.KLUG, C., HAGDORN, H. and MONTENARI, M. 2005.
Phosphatized soft-tissue in Triassic bivalves. Palaeontology, 48
(4), 833–852.KNIGHT, R. I. and MORRIS , N. J. 2009. A reconsideration
of the origins of the ‘typical’ Cretaceous inoceramid calcitic
KNIGHT ET AL . : INOCERAMID GILL PRESERVAT ION 53
hinge plate in the light of new ultrastructural observations
from some Jurassic ‘inoceramids’. Palaeontology, 52, 963–989.KOSCHELKINA, Z. V. 1969. General characteristics and
morphology of the genus Retroceramus. In SHILO, N. A.
(ed.). Jurassic and Cretaceous inocerams of the northeastern
USSR. Trudy Severo-Vostochnogo Kompleksnogog Nauchno-
Issledovatel’skogo Instituta (SVKNII) [Magadan], 32, 5–13.[In Russian].
KŘ ÍŽ , J. 2007. Origin, evolution and classification of the new
superorder Nepiomorphia (Mollusca, Bivalvia, Lower Palaeo-
zoic). Palaeontology, 50, 1341–1365.LE PENNEC, M., BENINGER, P. G. and HERRY, A. 1988.
New observations of the gills of Placopecten magellanicus (Moll-
usca: Bivalvia), and the implication for nutrition. II. Internal
anatomy and microanatomy. Marine Biology, 98, 229–237.LOGAN, W. N. 1898. The invertebrates of the Benton, Niobr-
ara and Fort Pierre groups. University Geological Survey of
Kansas, Paleontology Pt.1. Upper Cretaceous, 4, 431–518.LUDBROOK, N. H. 1966. Cretaceous biostratigraphy of the
Great Artesian Basin in South Australia. Bulletin of the Geolog-
ical Survey of South Australia, 40, 1–223.MACLEOD, K. G. and HOPPE, K. A. 1992. Evidence that
inoceramid bivalves were benthic and harbored chemosyn-
thetic symbionts. Geology, 20, 117–120.MCCOY, F. 1865. Note on the Cretaceous deposits of Austra-
lia. Annals and Magazine of Natural History, 16, 333–334.MEDLER, S. and SILVERMAN, H. 1997. Functional organi-
zation of intrinsic gill muscles in zebra mussels, Dreissena
polymorpha (Mollusca: Bivalvia), and response to transmitters
in vitro. Invertebrate Biology, 116 (3), 200–212.--1998. Extracellular matrix and muscle fibers in the gills
of freshwater bivalves. Invertebrate Biology 117 (4), 288–298.MIKKELSEN, P. M. and BIELER, R. 2008. Seashells of
southern Florida: living marine molluscs of the Florida Keys and
adjacent regions, Bivalves. Princeton University Press, Prince-
ton, NJ, 503 pp.
POUVREAU, S., JONQUIERES , G. and BUESTEL, D.
1999. Filtration by the pearl oyster, Pinctada margaritifera,
under conditions of low seston load and small particle size in
tropical lagoon habitat. Aquaculture, 176, 295–314.-BODOY, A. and BUESTEL, D. 2000. In situ suspension
feeding of the pearl oyster Pinctada margaritifera: combined
effects of body size and weather-related seston composition.
Aquaculture, 181, 91–113.REID, R. G. B. 1980. Aspects of the biology of a gutless species
of Solemya (Bivalvia: Protobranchia). Canadian Journal of
Zoology, 58, 386–393.RIDEWOOD, W. G. 1902. On the structure of the gills of the
Lamellibranchia. Philosophical Transactions of the Royal Society
of London, B195, 147–284.
SAGEMAN, B. B. and BINA, C. R. 1997. Diversity and spe-
cies abundance patterns in Late Cenomanian blackshale bio-
facies, Western Interior, U.S. Palaios, 12, 449–466.SCOTT, G. R. and COBBAN, W. A. 1964. Stratigraphy of
the Niobrara Formation at Pueblo, Colorado. United States
Geological Survey Professional Paper, 454-L, L1–L30.SEITZ, O. 1967. Die Inoceramen des Santon und Unter-
Campan von Nordwest Deutschland, III Teil. Beihefte zum
Geologischen Jahrbuch, Heft 75, Hanover, Germany, 171 pp.
SKAWINA, A. 2010. Experimental decay of gills in freshwater
bivalves as a key to understanding their preservation in Upper
Triassic lacustrine deposits. Palaios, 25, 215–220.STANLEY, S. M. 1972. Functional morphology and evolution
of bysally attached bivalve mollusks. Journal of Paleontology,
46 (2), 165–212.STEWART, J. D. 1990. Preliminary account of halecostome-
inoceramid commensalism in the Upper Cretaceous of Kansas.
51–57. In BOUCOT, A. (ed.). Evolutionary Palaeobiology of
Behaviour and Coevolution. Elsevier, Amsterdam, 582 pp.
TAYLOR, J. D. and GLOVER, E. A. 2006. Lucinidae (Bival-
via) – the most diverse group of symbiotic molluscs. Zoologi-
cal Journal of the Linnean Society, 148, 421–438.TEMKIN, I. 2006. Morphological perspective on the classifica-
tion and evolution of Recent Pterioidea (Mollusca: Bivalvia).
Zoological Journal of the Linnean Society, 148, 253–312.VOIGT, S. 1995. Palaeobiogeography of early Late Cretaceous
inoceramids in the context of a new global palaeogeography.
Cretaceous Research, 16, 343–356.WARD, J. E., LEVINTON, J. S., SHUMWAY, S. E. and
CUCCI , T. 1998. Particle sorting in bivalves: in vivo determi-
nation of the pallial organs of selection. Marine Biology, 131,
283–292.WHYTE, M. A. 1992. Phosphate gill supports in living and
fossil bivalves. 427–431. In SUGA, S. and NAKAHARA, H.
(eds). Mechanisms and phylogeny of mineralisation in biology
systems. Springer, Tokyo, 517 pp.
WILBY, P. R. and BRIGGS, D. E. G. 1997. Taxonomic trends
in the resolution of detail preserved in fossil phosphatised soft
tissues. Geobios, 20, 493–502.-and WHYTE M. A. 1995. Phosphatized soft tissues in biv-
alves from the Portland Roach of Dorset (Upper Jurassic).
Geological Magazine, 132, 117–120.YONGE, C. M. 1926. Structure and physiology of the organs
of feeding and digestion in Ostrea edulis. Journal of the
Marine Biological Association of the United Kingdom, 14,
295–386.ZARDUS, J. D. 2002. Protobranch bivalves. Advances in Mar-
ine Biology, 42, 1–65.
54 PALAEONTOLOGY , VOLUME 57
