New insights into the ultrastructure, permeability, and integrity of conodont apatite determined by transmission electron microscopy

DOI 10.1111/j.1502-3931.2007.00024.x © 2007 The Authors, Journal compilation © 2007 The Lethaia Foundation LETHAIA Blackwell Publishing Ltd New insights into the ultrastructure, permeability, and integrity of conodont apatite determined by transmission electron microscopy JULIE A. TROTTER, JOHN D. FITZ GERALD, HARRI KOKKONEN AND CHRISTOPHER R. BARNES Trotter, J.A., Fitz Gerald, J.D., Kokkonen, H. & Barnes, C.R. 2007: New insights into the ultrastructure, permeability, and integrity of conodont apatite determined by trans- mission electron microscopy. Lethaia , Vol. 40, pp. 97–110. New crystalline structures have been observed in argon ion-milled conodont elements from a diverse suite of Ordovician taxa (‘ Cordylodus robustus ’, Drepanoistodus suberec- tus , Panderodus gracilis , Plectodina ? sp., Aphelognathus sp., Periodon aculeatus ), using transmission electron microscopy (TEM). Electron diffraction patterns of albid tissue reveal that the component crystals are extraordinarily large, in the order of hundred(s) of microns. These large albid crystals show typical cancellate porosity, although a dis- tinctly lamellar structure has also been observed within a large albid crystal positioned between hyaline lamellar and cancellate albid tissues. There is a distinct absence of ‘interlamellar space’ within all hyaline tissues examined, which are characterized by a polycrystalline matrix of micron-scale elongate crystals that are both strongly aligned and tightly bound within a broader lamellar structure. Optical opacity, caused by light scattering within large ( ≥ 0.5 µ m) pores, is also a feature of both albid and polycrystal- line lamellar crown tissues. Accordingly, conodont hard tissues are differentiated by crystal size and shape, as well as inter- and intracrystalline porosity. These new observations highlight the structural complexities of conodont histologies and the need for more comprehensive investigations particularly of transitional crown tissues, which are not well defined by terms typically used in the literature. Their his- tological structures are interpreted to be a product of in vivo crystallization and thus provide new insights into the relative porosity, permeability, and inherent integrity of the tissues as well as their growth relationships. Accordingly, these data not only have implications for earlier histological and palaeobiological interpretations of conodont hard tissues but are also fundamental in determining their chemical integrity, which is crucial for characterizing palaeoseawater composition and palaeoenvironmental change. The potential for conodont apatite to retain primary chemical information depends on crystal size and permeability, so the large albid crystal domains are consistent with parallel geochemical studies that suggest that cancellate albid crown is more resistant to diagenetic modification. � Apatite, conodont, histology, TEM, transmission electron microscopy, palaeoseawater chemistr y. Julie A. Trotter [[email protected]; [email protected]], Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia, and CSIRO Division of Petroleum Resources, North Ryde, NSW 2113, Australia; John D. Fitz Gerald [[email protected]] and Harri Kokkonen [[email protected]], Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia; Christopher R. Barnes [[email protected]], School of Earth and Ocean Sciences, University of Victoria, British Columbia, Canada V8W 2Y2; manuscript received on 21/10/2005; manuscript accepted on 27/3/2007. Conodont microfossils are composed of carbonate fluorapatite (Pietzner et al . 1968), typically range between 0.1 and 3.0 mm in length, and are believed to represent the cephalic feeding apparatuses of small, extinct, marine chordates. Each element occupies a specific position within a multi-element apparatus and, in contrast to vertebrate mineralized tissues (apart from tooth enamel), grew by the outer apposi- tion of successive apatite layers. The morphological diversity of conodont elements reflects their rapid evolution and, together with their ubiquity in Cambrian to Triassic marine sequences, signifies their importance as biostratigraphical indicators. Although there is consensus for a chordate affinity, debate ranges from a primitive relationship (e.g. Kemp & Nicoll 1995, 1996; Pridmore et al . 1997) to one within a vertebrate phylogeny (e.g. Sansom et al . 1992, 1994; Smith et al . 1996; Aldridge & Purnell 1996; Aldridge & Donoghue 1998). The vertebrate model has attracted most attention and has dominated current literature, with interpret- ations often based on the histological structure of the component mineralized tissues. Hyaline tissue has been defined as a radial crystallite enamel akin to that 98 Trotter et al. LETHAIA 40 (2007) formed in vertebrates (e.g. Dzik 1986; Burnett & Hall 1992; Sansom et al . 1992; Donoghue 1998, 2001; Donoghue & Chauffe 1998), albid crown tissue has been interpreted as dermal bone (Sansom et al . 1992, 1994; Smith et al . 1996), mesodentine (Donoghue 1998) or a tissue unique to conodonts (Donoghue 1998; Donoghue & Chauffe 1998) that is a derived form of enamel (Donoghue & Aldridge 2001), and basal body tissue has been represented as globular calcified cartilage (Sansom et al . 1992) or dentine (Dzik 1986; Sansom et al . 1994; Sansom 1996). Competing hypo- theses have promoted conodonts as having chaetognath origins (e.g. Szaniawski 1987) or cephalochordate relationships (Kemp & Nicoll 1996). The histologies of conodont mineralized tissues have long been studied, primarily within the context of taxonomy, palaeobiology, and phylogeny. Such studies have often utilized scanning electron microscopy (SEM) and form the basis of our current understanding of conodont ultrastructure. However, SEM sample preparation requires acid etching of sectioned (blocked) specimens, which potentially modifies existing struc- tures, especially those components more susceptible to dissolution, such as zones of greater porosity and reactivity. Furthermore, the resolution afforded by SEM techniques for examining mineralogical and crystalline structure is relatively limited. Transmission electron microscopy (TEM) has the potential to reveal nanometre-scale details of crystal geometry, morphology, alignment, and composition, yet to date there are only five such published studies of conodont apatite (Pietzner et al . 1968; Pierce & Langenheim 1969; Barnes et al . 1970; Szaniawski 1987; Kemp 2002). These earlier TEM studies employed either ultramicrotome (diamond blade) sectioning techniques that can create shear and ‘shatter’ artefacts, or peels of acid-etched conodonts, which reveal surface microstructures only. The present TEM study, how- ever, employed argon ion-milling techniques, which offer the benefit of thinning specimens to electron transparency without suffering the preparation artefacts mentioned above. The objective of this study was to examine conodont crown histologies from a diverse range of morphologies and taxa in order to better characterize their crystalline structure within the context of porosity and permeability. This also complemented parallel geochemical studies (Trotter & Eggins 2006) regarding the suitability of conodont apatite as a geochemical archive. The new crystalline structures observed provide greater insights into understanding conodont hard-tissue structure, which has broader implications for their mineralization, permeability, and growth relationships. Given that this study was mainly driven by interests in the physical and chemical integrity of conodont apatite, examination has focused on crown ultrastru- ctures including regions that are intermediate between hyaline and albid tissues. Basal tissues are reputably unsuitable for geochemical studies (Holmden et al. 1996; Trotter et al. 1999; Wenzel et al. 2000) and therefore are only noted briefly in this study. Although the ubiquity of conodont elements throughout marine sequences makes them particularly attractive for palaeoseawater studies, issues of post-depositional chemical overprinting of primary compositions remain a significant problem. The techniques employed herein are particularly well suited to help resolve these issues, as the physical properties, crystalline structure, and inherent permeability of the component conodont tissues dictate their susceptibility to diagenesis. Accordingly, the significant features of crown tissues observed during this histological study are interpreted and discussed within both a palaeontological and a geochemical context. Prior understanding of conodont ultrastructure The most commonly preserved components of conodonts are the mineralized crowns, which are composed of hyaline (translucent) and often albid (opaque white) tissues. Albid crown tissue may be densely to incipiently developed and variably distri- buted within cusps and denticles, or completely absent from conodont elements. Less frequently preserved is the basal body, which fills the basal cavity of the crown element and protrudes from its aboral margin. The present study uses the term ‘hyaline’ rather than ‘lamellar’, which is commonly used within the literature and usually considered synonymous with hyaline tissue, ‘albid’ is used rather than ‘white matter’, and ‘basal body’ is equivalent to other commonly used terms such as ‘basal filling’ or ‘basal tissue’. Earlier studies have typically described the ultra- structure of albid crown tissue as fine-grained (even though the crystals had not been resolved), lacking growth lamellae, and characterized by a cancellate porosity (Pietzner et al . 1968; Barnes et al . 1973a; Donoghue 1998) with variably (often irregularly) shaped pores ranging up to several microns in size. In contrast, hyaline crown has been considered to represent the most coarsely crystalline tissue that was mineralized in the form of growth layers (lamellae), with interlamellar spaces often described from SEM imaging but also recognized by TEM (Pietzner et al . 1968; Barnes et al . 1970, 1973a, b). The crystallites are reported to be elongate and small, typically in the order of several microns long but range from 10 nm to > 30 µ m, and are usually oriented parallel or oblique LETHAIA 40 (2007) Conodont ultrastructure, permeability & integrity using TEM 99 (Barnes et al . 1970, 1973b; Donoghue 1998; Kemp 2002), but in some cases perpendicular (Sansom 1996), to the growth axis. Basal bodies can be vari- ably mineralized, comprising a fine polycrystalline matrix of (sub)equant crystallites that are irregularly arranged within a broader lamellar structure, which is apparently confluent with adjacent hyaline crown tissue (Pietzner et al . 1968; Lindström & Ziegler 1971; Barnes et al . 1973b; Donoghue 1998). Donoghue (1998) provided an historical synopsis of conodont histology and growth structure. Of particular significance to the present study, Dono- ghue showed that some tissues appearing albid in reflected light have a lamellar structure when sectioned and interpreted this opacity to be an artefact of crys- tallite alignment and hypocalcification. Due to this association of optical opacity with lamellar tissues, Donoghue used the term ‘lamellar’ in preference to ‘hyaline’ in order to differentiate such tissues from cancellate albid crown, concluding that the latter can only be determined unequivocally by thin sectioning and SEM examination. However, the present study has highlighted ambiguities and impracticalities of some of these terms (see Discussion), and thus the necessity for further work to better characterize the range of complex structures within conodont tissues. Our observations have revealed new, key information about their crystalline structure, which clearly demon- strates that TEM analysis of ion-milled specimens is particularly well suited for examining and interpreting conodont hard tissues. Materials and methods Sampling Well-preserved, thermally unaltered single conodont elements with a colour alteration index (CAI) of 1 were selected for analysis. CAI is a measure of pyrolysis of the residual organics, an index of 1 indicating a burial temperature of < 50º to 80ºC (Epstein et al . 1977) eliminating the likelihood of alteration from metamorphism. However, this does not preclude the potential for chemical exchange from diagenetic pore fluids. Specimens with iron oxide precipitated on external surfaces and within apparent ‘interlamellar spaces’ were also examined to investigate crown permeability (Fig. 1). Conodont elements were made available from existing collections (C. R. Barnes) that had previously been extracted from the carbonate host-rock by con- ventional buffered acetic acid digestion and bromoform or tetrabromomethane heavy mineral separation tech- niques (Austin 1987). These samples derive from three Laurentian localities of: (1) shallow to deep subtidal facies from the Stony Mountain Formation, Williston Basin in Manitoba (Upper Ordovician, Amorphognathus ordovicicus Zone); (2) shallow subtidal slightly argil- laceous units from the Georgian Bay Formation, Manitoulin Island, southern Ontario (Upper Ordo- vician, A. ordovicicus Zone); and (3) mid-slope facies from the Cow Head Group, St. Paul’s Inlet, western Newfoundland (Middle Ordovician, Tripodus laevis Zone), which were also the focus of parallel geo- chemical studies. The nine conodont elements selected for TEM analysis represent a broad range of morphologies and phylogenetic groups, as well as different tissue types. ‘ Cordylodus robustus ’ Ethington & Clark s.f., (i.e. the form species that possibly represents a species of Plegagnathus ), Drepanoistodus suberectus (Branson & Mehl), Panderodus gracilis (Branson & Mehl), and Plectodina ? sp., represent three families (and orders), Drepanoistodontidae, Panderodontidae, and Plecto- dinidae, all of which have relatively early origins. Their histological structure is described in the text, with one representative of each taxon figured. These specimens are from the same host rock sample from Manitoba. Aphelognathus sp. and Periodon aculeatus Hadding, which represent two families (Plectodinidae and Periodontidae respectively) recovered from dif- ferent host rock samples, were also examined. However given that they displayed virtually identical hyaline and cancellate albid structures to that of the afore- mentioned taxa, are briefly mentioned as support Fig. 1. Iron oxide impurities precipitated within hyaline (lamel- lar) crown, particularly within the axial cavities, of Drepanoistodus suberectus (Branson & Mehl). Note the in vivo repair of cusp in specimen ‘A’, where lamellae are truncated and show subsequent regrowth. 100 Trotter et al. LETHAIA 40 (2007) material but not illustrated. These specimens are from southern Ontario and western Newfoundland, respectively. Techniques Specimens were embedded in small blocks of epoxy resin (Epofix) that, after setting, were manually ground flat using SiC grit to expose the conodont elements. A glass slide was attached to the ground surfaces with thermal-setting cement (Crystalbond). The back side of the blocks was then ground to produce thin-sections (~30 µ m) cutting through all specimens. Unfortu- nately, only longitudinal sections (parallel to the growth axis) were prepared in thin section as intense fracturing of hyaline tissue in the chosen taxa occurred when sectioned transversely, thus destroying the sample. Copper grids were glued to the surfaces of the sectioned specimens and the glass slide subsequently removed after heating the cement briefly at 140 ºC. Samples mounted on grids were ion-milled from both sides using 5 kV Argon ions at an incident angle of 15 degrees (Williams & Carter 1996, pp. 162–264 provide general information for this type of prepara- tion). By precise positioning of the copper grid on the sections and care during milling, suitably thin foils were produced from each specimen selected for TEM. Unlike microtomed sections, ion-milled foils suffer from major variations in thickness. Regions examinable using TEM are thus confined to margins of ion-milled perforations; in these margins foils are less than a few hundred nanometres thick. The milled specimen alongside each perforation is approximately wedge shaped with thickness increasing away from each hole; thick regions, though not electron trans- parent, are useful for examination by light microscopy (e.g. for detecting macropores). The thinned parts of the conodont elements remain supported by the relict copper grid, which is also partially eroded during the milling process. Aside from superficial ‘orange peel’ artefacts that correspond to thickness variations and are visible at lower magnification, ion-milled samples contain otherwise unmodified arrangements of crys- tals and defects across extensive regions around each ion-milled perforation. Carbon-coated specimens were introduced to a Philips CM300 Transmission Electron Microscope for imaging and diffraction using 300 keV electrons. Comparative examination by light microscopy (both before and after ion milling) pro- vided good location correlation of thinned regions examined in TEM. Qualitative chemical information (e.g. for identifying Fe oxides) was gathered from fluoresced X-rays using the EDAX ultra-thin-window, Phoenix, energy-dispersive spectrometer fitted to the TEM. Histological structure TEM images of ‘ C. robustus ’, D. suberectus , Pa. gracilis , and Plectodina ? sp. show consistent patterns across taxa of the general crystalline structure of each tissue. Their microstructural character is described below according to tissue type, but each figure is organized to show the range of histologies examined from an individual specimen, which also provides a conven- ient format to illustrate their structural and spatial relationships. Differences between the component tissues essen- tially pertain to crystal size, shape, and porosity. We noted that polycrystalline basal body tissues ( Pa. gracilis , Fig. 2B; Plectodina ? sp., Fig. 3C, D) show significant porosity at the crystal boundaries (i.e. intercrystalline porosity) with the isometric to elongate nanocrystals usually randomly arranged. Longitudinal growth lines visible at lower magnification by light microscopy ( Plectodina ? sp., Fig. 3B) continue across the basal body-hyaline crown boundary yet appear to diverge transversely near the aboral margin; crystal alignment within these zones also differs (Fig. 3C, D). Further characterization of basal tissues was not pursued. Extensive porosity is also developed in both albid and polycrystalline hyaline crown tissues, principally within the component crystals. The most striking new observation, however, is the extraordinarily large sin- gle crystals of cancellate albid crown, which contrasts markedly to the polycrystalline nature of basal and hyaline tissues. Regions intermediate between hyaline and cancellate albid tissues also show considerable structural complexity with seemingly ‘hybrid’ features. Considerable attention has been given to examining the ultrastructure of different crown tissues, which are described in detail within the subsections below. Hyaline crown tissues TEM examination of hyaline crown tissue in Pa. gracilis (Fig. 2C, D) and D. suberectus (Fig. 4E, F) revealed a polycrystalline matrix of small, elongate, microcrystals from ~100 nm wide to several microns in length. Crystals show strong shape and crystallographic alignment, the c-axis oriented within ~10º to the growth axis of the conodont element (Fig. 4A) as observed from crossed-polarizers with low power light microscopy. Crystals are usually densely packed and are commonly observed crosscutting the lamellar orientation (Fig. 4F). Grain boundaries can be intri- cately sutured (e.g. Fig. 2C) but feature little porosity. The only microstructures vaguely resembling ‘inter- lamellar spaces’ take the form of axial growth cavities within hyaline regions. LETHAIA 40 (2007) Conodont ultrastructure, permeability & integrity using TEM 101 Fig. 2. Light microscope and TEM images of Panderodus gracilis (Branson & Mehl). �A. Plane polarized light microscope image of thin section attached to copper grid showing ion-milled zones along base and near cusp tip of conodont, and locations of regions ‘B’, ‘C’, ‘D’, ‘E’ and ‘G’. �B. TEM image of region ‘B’ showing randomly arranged isometric nanocrystals of basal body tissue with significant intercrystalline porosity. �C. TEM image of region ‘C’ showing elongate microcrystals of hyaline tissue with ubiquitous intracrystalline subspherical nanopores. �D. TEM image of region ‘D’ showing subtle bands of alternating nanopore density in hyaline tissue, zones with a higher density of smaller pores indicated by arrows. �E. Plane polarized light microscope image of region ‘E’ showing albid and hyaline tissue transition: dashed line defines boundary of large single albid crystal (right) comprising longitudinally aligned pores (dark thin bands) that are confluent with lamellae (faint lines) of adjacent hyaline tissue (left); arrows mark an additional thin albid band within hyaline crown; note the 10 micron-scale ‘orange-peel’ pattern is an artefact from ion-milling that is most apparent at low power; location of region ‘F’ is also indicated. �F. TEM image of region ‘F’ showing discrete bands of intracrystalline pores within a large single albid crystal; note dislocations (fine black lines, e.g. arrow) within crystal. �G. Transmitted light microscope image of region ‘G’ showing ion-milled perforations (the few irregular holes > 5 µm) and numerous intracrystalline pores (< 5 µm) of cancellate albid cusp. �H. TEM image showing numerous subspherical pores (~20 nm) within albid crown of region ‘G’, and extinction contours reflecting a single crystal domain. 102 Trotter et al. LETHAIA 40 (2007) The dominant porosity represented in hyaline crown tissue is the ubiquitous and extensive development of intracrystalline nanopores. Pores are discrete with no apparent connectivity, are sub-spherical ranging up to tens of nanometres in size, and are a consistent feature evident when the crystals are first examined in all specimens. Subtle, alternating bands of smaller and more densely spaced pores are present in some parts of D. suberectus and Pa. gracilis (Fig. 2D, arrows). The orientations of these bands appear to be similar to the lamellae that were recognized during light microscope examination of these regions. Intercrystalline porosity is typically manifested as axial growth cavities (Figs 1 and 4A, D) but also occurs sporadically away from the growth axis. These larger pores appear unconnected and appear either empty or filled with secondary minerals, such as iron oxides/ hydroxides (Fig. 1). Essentially the same polycrystalline structure and intracrystalline porosity as described above has also been observed in Aphelognathus sp. and Pe. aculeatus . Albid crown tissues Crystal boundaries in albid crown were identified using electron diffraction patterns. Extraordinarily large crystals, as defined by continuous extinction contours and selected-area diffraction patterns, characterize cancellate albid crown tissue in ‘ C. robustus ’ (Fig. 5C, D) and Pa. gracilis . Crystal size is therefore distinctly different to the nano- to micron-sized crystals of hyaline crown and basal body tissues. The broken cusp of ‘ C. robustus ’ represents a single crystal (~180 µ m × ~70 µ m) of cancellate albid tissue (Fig. 5B), which possibly comprised most (if not all) of the original cusp. The intracrystalline pores are extremely Fig. 3. Light microscope and TEM images of Plectodina? sp. �A. Plane polarized light microscope image of thin section attached to copper grid showing ion-milled zone at base of conodont. �B. Plane polarized light microscope image of ion-milled zone of basal body tissue showing longitudinal growth lines (arrows) and locations of regions ‘C’ and ‘D’. �C. TEM image of region ‘C’ showing randomly arranged isometric nanocrystals of basal body tissue with significant intercrystalline porosity. �D. TEM image of region ‘D’ showing strong alignment of elongate nanocrystals with intercrystalline porosity. LETHAIA 40 (2007) Conodont ultrastructure, permeability & integrity using TEM 103 varied in shape and size, ranging from nano- to micron- scale. Pores are apparently unconnected and do not contain secondary precipitates within the pore spaces. Electron diffraction patterns along the albid cusp of Pa. gracilis also reveal large crystals. One crystal extends from the cusp tip with the c-axis sub-parallel to the growth axis of the cusp, displaying typical cancellate intracrystalline porosity (Fig. 2G) as well as numerous small pores (Fig. 2H). A second crystal at the base-cusp junction has its c-axis sub-parallel to the growth axis of the conodont base. It has a distinct longitudinal lamellar-type structure expressed as well-aligned elongate pores (Fig. 2F), which differs markedly from the cancellate porosity of the adjacent albid crystal. The gross morphology may have influ- enced the formation and orientation of these two crystals, as related to the change in curvature of the growth axis at the base–cusp junction. The crystal extending from the tip of the cusp (includes region ‘G’ of Fig. 2) is at least 250 µ m long. Dislocations and sub-grains have also been observed in albid crown tissue of Pa. gracilis. The density of Fig. 4. Light microscope and TEM images of Drepanoistodus suberectus (Branson & Mehl). �A. Plane polarized light microscope image of thin section prior to ion-milling, showing axial growth cavities and darker bands (optical opacity) near the cusp tip and along the growth axis; black lines represent optical extinction position when examined under crossed-polarizers, which show a change of inflexion at the growth axis. �B. Plane polarized light microscope image of thin section attached to copper grid showing ion-milled zones and location of regions ‘C’, ‘D’ and ‘F’; scale as in ‘A’. �C. Plane polarized light microscope image of region ‘C’ showing lamellar structure of hyaline tissue with lamellae parallel to growth axis (arrow). �D. Plane polarized light microscope image of region ‘D’ showing ‘albid’ axial zone of high porosity and axial cavities; note that the specimen was returned to the ion-thinner hence area D represents an enlarged ion- thinned perforation relative to that shown in images ‘B’ and ‘C’. �E. TEM image of margin of region ‘E’ showing well aligned elongate microcrystals and intracrystalline subspherical nanopores of hyaline tissue; lamellae are indistinct and ‘interlamellar spaces’ absent; growth axis denoted by arrow. �F. TEM image of region ‘F’ showing two crystal alignments (arrows) dominated by the orientation oblique to the growth axis, with some parallel to the growth axis. Note that lamellae are sub-parallel to growth axis (C), whereas crystal alignment is oblique (A, E). 104 Trotter et al. LETHAIA 40 (2007) dislocations (Fig. 2F, arrow) within these albid crys- tals may vary considerably, from low-density isolated dislocations to a moderately dense distribution. Similar dislocations as well as the large cancellate albid crystal structure described above were also recognized in Pe. aculeatus, although some sub-grains of ~1 µm in diameter were surrounded by dislocation walls that have a slight (< 1°) difference in orientation to the enclosing crystal. Albid–hyaline transition Regions intermediate between large albid crystal domains and polycrystalline hyaline crown examined in ‘C. robustus’ and Pa. gracilis have revealed ‘hybrid’ features. The albid single crystal adjacent to hyaline crown in Pa. gracilis displays a lamellar-type struc- ture of alternating dark and light bands. The dark bands represent pores that are confluent with the lamellae (faint lines) of the adjacent polycrystalline hyaline tissue (Fig. 2E), creating an impression of lamellar structure within albid crown that is definitely within a large single crystal. Significantly, the transition from this lamellar albid single crystal to polycrystal- line hyaline tissue is very subtle here and complex, in that no clear tissue boundary is discernable. In fact, an additional thin band of polycrystalline albid tissue (Fig. 2E arrows) appears to lie within the margins of Fig. 5. Light microscope and TEM images of ‘Cordylodus robustus’ Ethington & Clark s.f. �A. Plane polarized light microscope image of thin section prior to ion-milling showing albid tissue (dark patch) of remnant cusp. �B. Plane polarized light microscope image of cusp showing cancellate microporous albid tissue (dark zone left of dashed line) with irregular cuspate pores up to 5 µm diameter, hyaline tissue with faint lamellar structure (clear areas), and patchy albidity (dark streaks right of dashed line) within zone intermediate between hyaline and cancellate crown; line defines the boundary of single crystal cancellate albid crown; electron-transparent regions ‘C’ and ‘E’ are indicated. �C, D. respective extinction contour and spot electron diffraction patterns of region ‘C’ define its single crystal structure; note in ‘C’ that the large holes at the top are ion-milled perforations and not pores. �E. TEM images of region ‘E’ showing elongate intercrystalline micropores within a polycrystalline matrix otherwise identical to hyaline tissue. LETHAIA 40 (2007) Conodont ultrastructure, permeability & integrity using TEM 105 surrounding hyaline crown, with this band compris- ing crystals much larger than those in the adjacent hyaline tissue. Tissue intermediate between polycrystalline hya- line crown and the cancellate albid single crystal of ‘C. robustus’ is polycrystalline but hosts numerous large elongate intercrystalline pores (Fig. 5E). This high intercrystalline porosity observed by light microscopy is manifested as sporadic opaque streaks that interfinger with the adjacent crown tissues. The transition between albid and hyaline tissue is again irregular and indistinct, although translucent tissue appears to abut cancellate albid tissue in a few small patches. Discussion Our new observations are compared below with exist- ing descriptions from the literature, and are discussed in the context of mineralization, diagenesis, and growth relationships. The major differences in some funda- mental properties, such as crystal size and porosity, shown by this study have broad implications crossing different geoscience disciplines. Not only do they provide new insights into the histological structure and mineralization of conodont tissues, but are also indicative of the inherent integrity of the component tissues and consequently their potential as geochemical archives. Firstly, however, some discussion is required regarding the practicalities and usefulness of some commonly used terms, as well as some technical aspects relevant to interpreting TEM images. The new observations from this study highlight potential ambiguities of terms commonly used within the literature to differentiate conodont crown tissues. The white opacity characteristic of cancellate albid tissue is due to light scattering by large pores (≥ 0.5 µm), whereas nanopores of hyaline tissue are too small to scatter visible light, resulting in a translucent appear- ance. Accordingly, ‘lamellar’ and ‘hyaline’ have usually been considered synonymous. However, opacity is often observed in hyaline (polycrystalline) lamellar tissue, commonly manifested as large pores sporadically dis- tributed along crystal boundaries (i.e. intercrystalline porosity; Fig. 5B, E). Donoghue (1998) also recognized that opacity was not exclusive to cancellate albid crown tissues, revealing its occurrence within polycrystalline lamellar crown from thin section and SEM examina- tion of Cordylodus angulatus Pander and Ligonodina sp., although he attributed this feature to changes in crystallite alignment. Our observation of lamellae within single large albid crystals of Pa. gracilis (Fig. 2F) also shows that the term ‘lamellar’ cannot be restricted to polycrystalline hyaline (translucent) crown tissues. The terms ‘albid’ (and white matter) and ‘lamellar’ are therefore unsuitable sole descriptors for discrim- inating crown tissues. Given that lamellar structures may feature in both albid and hyaline tissues and that higher power examination is required to further resolve and chara- cterize the crystalline structure, a practical first-order distinction of crown tissues is required. The common terms ‘albid’ (opaque white) and ‘hyaline’ (translucent), though only diagnostic of relative porosity (pore size), are convenient for light microscopy of complete speci- mens as used by the majority of conodont workers. Future high resolution work, however, is clearly needed to identify a more complete range of crown tissues that represent a greater array of taxonomic groups. The challenge is therefore to create a nomenclature that, at some level, suitably encompasses both high resolution histology as well as low power reflected light microscopy of complete specimens. Other important considerations when interpreting observed microstructures are the potential artefacts created by the techniques employed, particularly those used during sample preparation. Some appar- ent contradictions and potential controversy between observations from the present study and earlier TEM studies (Pietzner et al. 1968; Kemp 2002) almost cer- tainly relate to technical issues. These are addressed separately below, prior to the primary discussion on mineralization, growth, and diagenetic processes, in order to first justify some fundamental interpretations of specific microstructural features. Histological artefacts The present study has examined sections of conodont elements that were Ar-ion-milled, and produced what we suggest are the first published images of undisrupted crystalline structure of basal body, hyaline, and albid crown tissues. In contrast, specimen preparation techniques used in earlier studies, being microtome sections of block-mounted conodont elements (Pietzner et al. 1968; Szaniawski 1987; Kemp 2002) and examination of peels and acid-etched surfaces of block-mounted specimens (Pierce & Langenheim 1969; Barnes et al. 1970), most likely modified the tissues to some extent and complicated characterization of their crystalline structure. While excellent images of superficially similar apatitic materials (e.g. lungfish dentine in Kemp & Barry 2006) are produced in carefully microtomed sections, we suspect that this is restricted to fine grain sizes (e.g. figs 5–6 in Kemp & Barry 2006). As far as we can ascertain, the images of ion-milled regions of conodont albid tissues shown in the present study are the first that illustrate larger apatite crystals. 106 Trotter et al. LETHAIA 40 (2007) Porosity has been well recognized in cancellate albid tissue by earlier SEM and TEM studies, however we have seen that extensive intracrystalline porosity is also common in hyaline crown tissues. The tiny sub- spherical pores reported in this study are similar to those recognized earlier by Barnes et al. (1970) in the neurodont hyaline species Polycaulodus bidentatus (Branson & Mehl), and more recently observed in Cordylodus and Panderodus by Kemp (2002) who noted that they ‘resemble artefacts that can be related to beam damage in enamel crystals of mammalian enamel viewed in the electron microscope (Warshaw- sky 1989)’ (Kemp 2002, p. 33). This requires some clarification; on the one hand, Warshawsky (1989) described nano-sized beam-damage features that resulted from electron irradiation and which were observed to develop further during prolonged beam irradiation. Kemp (2002), on the other hand, in char- acterizing conodont apatite using TEM, reports that ‘vacuities are evident when specimens are first exam- ined in the electron beam and do not develop further while the specimen is being examined’ (Kemp 2002, p. 33). Such features therefore are not due to electron beam damage. Similarly, to the best of our knowledge, all of the pores shown in our micrographs were evident when the sections were first examined, do not change during routine observation at moderate magnifica- tions, and hence are not electron irradiation artefacts. Nevertheless, care should always be taken in TEM imaging since nanoscale damage centres are easily generated in apatite crystals when examined by high electron doses, especially if those crystals have pre- existing defect structures (Warshawsky 1989) that possibly relate to calcium deficiency (Nelson & Barry 1989). Kemp (2002, p. 33) also stated that the ‘vacuities’ she observed ‘are a result of diagenesis of these specimens, of damage affecting the crystals of hydroxyapatite during its preservational history’. Her conclusion relies on the lack of pores in conodont tissue imaged by Pietzner et al. (1968). However, close examination of micrographs in Pietzner et al. (1968) indicates that pores of sub-50 nm diameter would be very difficult to detect in this case (e.g. plate 21 in Pietzner et al. 1968), since some ultra-microtomed sections from this first TEM study of conodont elements were badly shattered. Therefore, we argue that the sharply defined pore structures like those we imaged here in phase contrast, also seen by Barnes et al. (1970) and Kemp (2002), are significant microstructures likely related to in vivo conodont apatite crystallization. Further work on other specimens is required to confirm our conjecture. Of further note is the apparent contradiction between the large crystal domains within albid tissue observed in our study, and earlier findings of a polycrystalline structure in albid tissue from Hindeodella (plate 21, figs 1–3 in Pietzner et al. 1968). We conclude that the true crystalline structure of Hindeodella was disrupted by the intense shatter and shear artefacts created during microtome sample preparation that produced an arced electron diffraction pattern (plate 19, fig. 6 in Pietzner et al. 1968), both misinterpreted as chara- cterizing fine-grained and polycrystalline tissue. Mineralization of crown tissues The fine crystals of conodont hyaline tissue with their elongate form and intricate serrations suggest that they are well preserved and potentially primary, or at least mirror the original structure. Kemp (2002), however, argued that nanopores in hyaline tissue resulted from diagenesis, but we refute that argument (see Histological artefacts section above). Of particu- lar significance are the extraordinarily large crystals (100’s µm) of cancellate (i.e. coarsely porous) albid crown observed in ‘C. robustus’ (Fig. 5B), Pa. gracilis (Fig. 2G, H), and Pe. aculeatus, which contradict ear- lier findings (Pietzner et al. 1968, and see Histological artefacts section above) and widespread assumptions that cancellate albid crown is finely crystalline. The large crystal size is a feature that is likely common to many taxa, as preliminary electron back-scattered diffraction (EBSD) analysis of six additional specimens (previously blocked and polished for probe analysis) of Plectodina? sp., Plegagnathus sp., Pristognathus bighornensis, and indeterminate fragments, has also revealed large crystal domains in cancellate albid crown tissues. Such large crystal sizes are interpreted to have formed in vivo rather than from post-mortem recry- stallization processes, and dislocations noted within some albid regions are considered to be defects grown at the time of mineralization rather than from sub- sequent deformation. It is also well known that albid crown tissue is the least soluble conodont histology, which, together with its large crystal size and lower concentrations of trace and rare earth elements (REE) in particular (Trotter & Eggins 2006), suggests that secondary alteration has been minimal (see Permeability section below). Nevertheless, an alternative interpretation of the large single crystal character and trace element chemistry of albid crown tissues might invoke diage- netic recrystallization, involving removal of REE that had been previously incorporated during early dia- genesis. In this scenario, pores within cancellate albid crown might incorporate REE rejected from the recrystallizing albid tissue; however, these pores are typically devoid of impurities as evidenced by TEM examination and by the very low REE contents of LETHAIA 40 (2007) Conodont ultrastructure, permeability & integrity using TEM 107 albid regions. Furthermore, the exclusion of REE from a growing albid crystal might concentrate such elements at and ahead of the recrystallization front, manifested as REE-enriched zones in the vicinity of the albid–hyaline junction. No such enrichments have been observed using laser ablation inductively cou- pled plasma mass spectrometry (LA-ICPMS) at the transition from hyaline to albid regions in blocked and longitudinally sectioned conodont elements, but rather a sharp decrease in REE concentrations occurs across the boundary (Trotter & Eggins 2006). Fur- thermore, it is difficult to conceive that albid regions would consistently undergo extensive diagenetic recrystallization while adjacent permeable hyaline tissue escaped with comparatively little change, retain- ing a fine, well-preserved, microcrystalline structure. This would require the primary albid tissue to be considerably less stable during diagenesis than adjacent hyaline crown. It is therefore concluded from both TEM and geochemical evidence that the large crystal domains of cancellate albid crown are most likely produced in vivo. This may have occurred during a primary or secondary metabolic process, but given both the ubiquitous intracrystalline porosity of the various crown tissues and the structures observed within regions intermediate between hyaline and cancellate albid tissues (i.e. hybrid features), a secondary in vivo process may be more likely. Implications for tissue growth and conodont phylogeny Based on limited microstructural data, some earlier workers favoured cancellate albid tissue to be derived from a precursor material, thus representing a secondary in vivo mineralization event (Lindström 1964; Lind- ström & Ziegler 1971; Lindström et al. 1972; Barnes et al. 1973a, b). Barnes and co-workers further sug- gested the existence of a ‘transitional’ tissue type due to the absence of a clear boundary between cancellate albid and hyaline lamellar crown when viewed at high power by SEM, and the apparent convergence of their microstructures (Barnes et al. 1973a, b; Barnes & Slack 1975). More recently, Donoghue (1998) identified extremely sharp and step-like boundaries between hyaline and cancellate crown by oil immer- sion light and laser confocal microscopy, which were interpreted as simultaneous mineralization thus refuting the existence of transitional histologies. However, the clarity of any boundary is subject to the scale at which it is examined, as lower power analysis can potentially indicate a distinct boundary that is not structurally defined at higher resolution, as has been observed herein and previously by Barnes et al. (1973a). The extensive intracrystalline porosity of hyaline and cancellate albid crown tissues also suggests an intrinsic histological relationship. Following earlier models of in vivo resorption of existing hyaline lamellar crown, the coalescence of existing intracrystalline nanopores could produce a highly porous replace- ment tissue with large irregular pores, similar to that characterizing cancellate albid crown. The complex and indistinct boundaries between hyaline, cancellate albid, and intermediate crown tissues observed in this study in ‘C. robustus’ and Pa. gracilis in particular, suggest potential gradations that could be consistent with either a simultaneous or derivative growth relationship. Although the lamellar part of the single albid crystal in Pa. gracilis (Fig. 2E, F) represents intracrystalline porosity, it clearly resembles and is likely related to the elongate intercrystalline pores of polycrystalline tissue observed in ‘C. robustus’ (Fig. 5E). In any case, the confluent lamellar structure of hyaline and adjacent single crystal albid crown in Pa. gracilis is at least strongly indicative of a primary (i.e. in vivo) origin rather than post-mortem recystallization. A fundamental feature of polycrystalline hyaline crown tissue is its inherent lamellar structure, which was first recognized by Pander (1856) from light microscope examination. These lamellae have been the basis for interpreting the mode of growth and mineralization of conodont hard tissues. However, the continuous ‘interlamellar spaces’ described in many SEM studies were absent in all polycrystalline crown tissues examined herein by TEM, even though lamellae can be observed in these regions by light microscopy. Lamellar structures, as represented by continuous longitudinal white (Fig. 1A) and dark (Fig. 4C) lines in light microscope examination, were specifically targeted in this study, yet examination from the outer margin through to the growth axis showed no microstructural features corresponding to the lamellar spacing. All hyaline tissues examined are therefore characterized by dense crystal aggregates lacking continuous interlamellar spaces, the crystals often transecting the lamellar orientation (Fig. 4F), and occasional intercrystalline pores appearing spo- radic and unconnected. The nature of the lamellae in hyaline polycrystalline tissues is therefore somewhat enigmatic, but could be related to the subtle bands of different refractive index due to intracrystalline nan- opores of alternating size and density, which have been observed in both D. suberectus and Pa. gracilis (Fig. 2D). It should be considered that zones of higher porosity can be readily and preferentially etched during sample preparation for SEM imaging, hence may create the impression of continuous voids. Such spaces would represent those often described as interlamellar spaces, 108 Trotter et al. LETHAIA 40 (2007) which are most likely the accentuated expression of lamellae margins modified by acid etching that would be otherwise compact in pristine material. The regions intermediate between hyaline and cancellate albid crown tissues examined in Pa. gracilis may also provide insights into possible pore connectivity within high porosity zones (Fig. 2E, F). The apparent margins of lamellae (faint lines) within hyaline tissue appear confluent with the adjacent high porosity zones (dark bands) within the lamellar albid single crystal (Fig. 2E). The pores are clearly discrete, but where the ion- milled specimen becomes thicker (in the right side of Fig. 2E), continuous lines are seen in the albid regions (Fig. 2F) suggesting potential connectivity. The com- plete three-dimensional lamellar structure of hyaline tissue, however, remains unclear and certainly war- rants further study given that it is a fundamental aspect of conodont growth, but also has implications for assessing the permeability of hyaline crown. The new histological observations from albid tissue also provide some insights into better constraining the phylogenetic affinities of conodonts, although in- ferences regarding this complex and contentious issue are not within the scope of this study. Nonetheless, it is now clear that historical models incorporating the assumption that albid crown tissue is finely crystal- line are invalid (e.g. Sansom et al. 1992, 1994; Smith et al. 1996; Donoghue 1998; Donoghue & Aldridge 2001), as fine crystalline tissues such as dermal bone, dentine, and even those of enamel, clearly have no equivalence to the enormous crystal sizes of cancel- late albid tissues revealed by this study. Although enameloids may develop large crystals, most forms are composed of elongate fibre-like crystals and dif- ferences in composition (e.g. higher carbonate content) further provide inconsistencies with conodont albid tissue. Such inconsistencies may support earlier sug- gestions that albid tissue is unique to conodonts, although further high resolution histological and geochemical analyses are clearly needed to better characterize its development within various taxonomic groups, to better elucidate potential relationships to known tissue structures. Permeability and implications for geochemical studies The extreme differences in crystal sizes of conodont crown tissues together with their contrasting trace element compositions (Trotter & Eggins 2006) is parti- cularly important to conodont geochemical studies in the context of palaeoseawater reconstructions. In porous tissue, pore connectivity and crystal size are key controls of permeability, which has significant implications for its inherent susceptibility to diagenetic alteration. Tissues with high pore connectivity will have greater hydraulic conductivity, thus enhancing pore-fluid migration, and large crystal surface areas will promote the diffusion and adsorption of second- ary ions onto apatite crystallites. Although extensive porosity has been recognized in all crown tissues, the assessment of pore connectivity and consequent tissue permeability is complex. The small subspherical intracrystalline nanopores of hyaline tissue clearly cannot enhance permeability at low temperature, similarly for the apparently unconnected large pores of cancellate albid crown. The intricate grain-boundary shapes and low grain- boundary porosity in hyaline tissue further suggest low permeability. However, the precipitation of impurities (e.g. Fe oxide) on both the outer surfaces and within hyaline tissues along crystal boundaries and in axial cavities (Fig. 1) reflects a diagenetic origin, presum- ably by interaction with external fluids or perhaps even bacterially induced, and thus some degree of connectivity between the intercrystalline pores. The clear relationship of hyaline lamellae with those of the adjacent single albid crystal (Fig. 2E) in Pa. gracilis and the lines of pores (Fig. 2F) in particular, suggest potential connectivity within the lateral longitudinal plane. Furthermore, the apparent confluent growth relationship of hyaline crown with more permeable basal body tissue may provide migration pathways for secondary fluids, potentially enhancing uptake by lamellar crown tissues. It is also noted that albid crown in general is well recognized as being significantly more resistant to acid etching than hyaline tissue (e.g. Stauffer & Plummer 1932; Lindström 1964; Barnes et al. 1973a), which may suggest comparatively lower permeability, as well as particular compositional differences such as lower carbonate content, which has been suggested earlier (Pietzner et al. 1968) and recently determined by Fourier Transform Infra-Red Spectroscopy of conodont tissues (Trotter & Eggins 2006). The extraordinarily large crystals (100’s µm) and apparent lack of pore connectivity of cancellate single crystal albid tissue would inhibit chemical exchange with impurities from diagenetic pore fluids. Conversely, the numer- ous small crystals of hyaline crown would provide a significantly greater surface area with many sites available for chemical exchange, assuming sufficient pore connectivity and thus accessibility to these surfaces. Distinct permeability differentials between various conodont tissues have recently been inferred from high-resolution in situ geochemical analysis using LA-ICPMS (Trotter & Eggins 2006). Continuous depth profiles through albid crown tissue showed REE compositions to be typically U-shaped with very low to LETHAIA 40 (2007) Conodont ultrastructure, permeability & integrity using TEM 109 undetectable interior concentrations, which contrasts markedly with the enriched and essentially equili- brated (uniform) profiles in hyaline crown and basal body tissues, the latter similar to ichthyoliths and phosphatic brachiopods. Given that significant REE uptake by biogenic apatites is a post-mortem pheno- menon, variations in concentrations can potentially reflect the relative permeabilities of these tissues. The strong linear correlation of their REE compositions defines albid crown as the lowermost concentration ‘end-member’ implying minimal post-mortem uptake, whereas compositions of hyaline crown are interme- diate between the high concentration ‘end-members’ of conodont basal tissue, ichthyolith, and brachiopod apatites (Trotter & Eggins 2006). These systematic compositional differences are supported by the new histological observations described herein, and col- lectively suggest that cancellate albid tissue has restricted permeability, thereby offering greater potential for retaining primary geochemical signatures. Conclusions This study highlights the benefits of ion-milled speci- men preparation for TEM imaging and electron dif- fraction to better characterize conodont tissues, and thus shows that it is particularly well suited for exam- ining small fragile materials. The resultant data have enhanced our understanding of conodont apatite ultrastructure by revealing new, key, crystalline properties that cannot be determined within the limits of conventional light microscopy, SEM, and TEM analysis of ultramicrotomed sections. Cancellate albid crown comprises large (100’s µm) apatite crystals, much larger than those in hyaline lamellar crown tissue, hence is not a fine-grained aggregate as previously believed. These extraordinarily large crystals likely reflect an in vivo crystallization event, and served as a less-permeable barrier to fluid exchange during diagenesis. Opacity may be variable, caused by intracrystalline porosity within single cry- stal albid tissue from either a cancellate or a lamellar structure, or by large sporadic intercrystalline pores within a polycrystalline matrix otherwise identical to hyaline crown tissue. Hyaline crown comprises elongate and well-aligned crystals containing numerous nanopores, however, the nature of the lamellar structure is unclear at micron- and nano-scale, and continuous interlamellar spaces often reported in SEM studies were not observed by TEM. The extensive intracrystalline porosity recognized in all crown tissues examined, and the confluent lamellar structure between hyaline and single crystal albid tissue of Pa. gracilis, suggest a possible derivative growth relationship between hya- line and albid tissues. These observations of porosity and crystal size challenge some of the orthodox views on conodont histology and have wide ranging implications that cross various geoscience disciplines, although further work is clearly required to encapsulate a broader range of taxonomic groups. In addition to providing new insights into conodont ultrastructure and possible growth relationships between the component tissues, which clearly relates to broader conodont phyloge- netic research, these data are also directly relevant to geochemical studies. In particular, inferences regard- ing the relative permeability of conodont tissues help constrain their suitability as geochemical tracers for palaeoseawater applications, and highlight the import- ant relationship between sample microstructure and diagenesis. In this context, albid crown most likely represents the least permeable conodont tissue and thereby offers the greatest potential for retaining primary geochemical signatures. Acknowledgements. – The authors thank S.M. Eggins and M.T. McCulloch for informative discussion and editorial com- ments on the manuscript, U. Faul for providing EBSD analyses, and R.S. Nicoll for samples analyzed by EBSD. Some wide-ranging comments from J. Barry, P. Donoghue, A. Kemp, M. Lindström, and I. Sansom, improved the manuscript. Hedy Dechert is acknow- ledged for her English translation of Pietzner et al. (1968) that was kindly provided by Gilbert Klapper. An Australian Postgrad- uate Award funded this research, and a student award granted by the Paleontological Society and Mid-American Paleontological Society is also gratefully acknowledged. References Aldridge, R.J. & Purnell, M.A. 1996: The conodont controversies. Trends in Ecology & Evolution 11, 463–468. Aldridge, R.J. & Donoghue, P.C.J. 1998: Conodonts: a sister group to hagfish? In Jørgensen, J.M.J., Lomholt, J.P., Weber, R.E. & Malte, H. (eds): The Biology of Hagfishes, 15–31. Chapman & Hall, London. Austin, R.L. 1987: Conodonts: Investigative Techniques and Applications, 422 pp. British Micropaleontological Society, Ellis Horwood, Chichester, UK. Barnes, C.R. & Slack, D.J. 1975: Conodont ultrastructure: the subfamily Acanthodontinae. Royal Ontario Museum, Life Sciences Contribution 106, 1–21. Barnes C.R., Sass, D.B. & Monroe, E.A. 1970: Preliminary studies of the ultrastructure of selected Ordovician conodonts. Royal Ontario Museum, Life Sciences Contribution 76, 1–24. Barnes, C.R., Sass, D.B. & Monroe, E.A. 1973a: Ultrastructure of some Ordovician conodonts. Geological Society of America Special Paper 141, 1–30. Barnes, C.R., Sass, D.B. & Poplawski, M.L.S. 1973b: Conodont ultrastructure: the family panderodontidae. Royal Ontario Museum, Life Science Contribution 90, 1–36. Burnett, R.D. & Hall, J.C. 1992: Significance of ultrastructural features in etched conodonts. Journal of Paleontology 66, 266– 276. Donoghue, P.C.J. 1998: Growth and patterning in the conodont skeleton. Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences 353, 633–666. 110 Trotter et al. LETHAIA 40 (2007) Donoghue, P.C.J. 2001: Microstructural variation in conodont enamel is a function of adaptation. Proceedings of the Royal Society of London B, Series B, Biological Sciences 268, 1691–1698. Donoghue, P.C.J. & Chauffe, K.M. 1998: Conchodontus, Mitrel- lataxis and Fungulodus: conodonts, fish or both? Lethaia 31, 283–292. Donoghue, P.C.J. & Aldridge, R.J. 2001: Origin of a mineralized skeleton. In Ahlberg, P.E. (ed.): Major Events in Early Verte- brate Evolution, 85–105. Taylor & Francis, London. Dzik, J. 1986: Chordate affinities of the conodonts. In Hoffman, A. & Nitecki, M.H. (eds): Problematic Fossil Taxa. Oxford Monographs on Geology and Geophysics 5, 240–254. Oxford University Press, New York. Epstein, A.G., Epstein, J.B. & Harris, L.D. 1977: Conodont color alteration – an index to organic metamorphism. US Geological Survey Professional Paper 995, 1–27. Holmden, C., Creaser, R.A., Muehlenbachs, K., Bergstrom, S.M. & Leslie, S.A. 1996: Isotopic and elemental systematics of Sr and Nd in 454 Ma biogenic apatities: implications for paleoseawater studies. Earth and Planetary Science Letters 142, 425–437. Kemp, A. 2002: Hyaline tissue of thermally unaltered conodont elements and the enamel of vertebrates. Alcheringa 26, 23–36. Kemp, A. & Barry, J.C. 2006: Prismatic dentine in the Australian lungfish, Neoceratodus forsteri (Osteichthyes: Dipnoi). Tissue and Cell 38, 127–140. Kemp, A. & Nicoll, R.S. 1995: Protochordate affinities of cono- donts. Courier Forschungsinstitut Senckenberg 182, 235–245. Kemp, A. & Nicoll, R.S. 1996: A histochemical analysis of biological residues in conodont elements. Modern Geology 20, 287–302. Lindström, M. 1964: Conodonts, 196 pp. Elsevier, Amsterdam. Lindström, M. & Ziegler, W. 1971: Feinstruckturelle untersuc- hungen an conodonten. 1. Die überfamilie Panderodontacea. Geologica et Palaeontologica 5, 9–33. Lindström, M., McTavish, R.A. & Ziegler, W. 1972: Feinstruck- turelle untersuchungen an conodonten. 2. Einge Prioniodontidae aus dem Ordovicium Australiens. Geologica et Palaeontologica 6, 33–43. Nelson, D.G.A. & Barry, J.C. 1989: Electron microscopy of non-stoichiometric apatite crystals. The Anatomical Record 224, 265–276. Pander, C.H. 1856: Monographie der Fossil Fische des Silurischen Systems der Russisch-Baltischen Gouvernements. 91 pp. Akad- emie der Wissenschaften, St. Petersburg, Russia. Pierce, R.W. & Langenheim, R.L. Jr 1969: Ultrastructure in Palmatolepis sp. and Polygnathus sp. Geological Society of America Bulletin 80, 1397–1400. Pietzner, H., Vahl, J., Werner, H. & Ziegler, W. 1968: Zur chemis- chen zusammensetzung und mikromorphologie der conodonten. Palaeontographica Abhandlungen A 128, 115–152. Pridmore, P.A., Barwick, R.E. & Nicoll, R.S. 1997: Soft anatomy and affinities of conodonts. Lethaia 29, 317–328. Sansom, I.J. 1996: Pseudooneotodus: a histological study of an Ordovician to Devonian vertebrate lineage. Zoological Journal of the Linnean Society 118, 47–57. Sansom, I.J., Smith, M.P., Armstrong, H.A. & Smith, M.M. 1992: Presence of the earliest vertebrate hard tissues in conodonts. Science 256, 1308–1311. Sansom, I.J., Smith, M.P. & Smith, M.M. 1994: Dentine in conodonts. Nature 368, 591. Smith, M.M., Sansom, I.J. & Smith, M.P. 1996: ‘Teeth’ before armour: the earliest vertebrate mineralized tissues. Modern Geology 20, 303–319. Stauffer, C. & Plummer, H.J. 1932: Texas Pennsylvanian cono- donts and their stratigraphic relations. University of Texas Bulletin 3201, 13–50. Szaniawski, H. 1987: Preliminary structural comparisons of protoconodont, paraconodont, and euconodont elements. In Aldridge, R.J. (ed.): Palaeobiology of Conodonts, 35–47. British Micropaleontological Society, Ellis Horwood, Chichester, UK. Trotter, J.A. & Eggins, S.M. 2006: Chemical systematics of cono- dont apatite determined by laser-ablation ICPMS. Chemical Geology 233, 196–216. Trotter, J.A., Korsch, M.J., Nicoll, R.S. & Whitford, D.J. 1999: Sr isotopic variation in single conodont elements: implications for defining the Sr seawater curve. In Serpagli, E. (ed.): Studies on Conodonts. Proceedings of the Seventh European Conodont Symposium 1998. Bolletino della Società Paleontologica Italiana 37, 507–514. Warshawsky, H. 1989: Organization of crystals in enamel. The Anatomical Record 224, 242–262. Wenzel, B., Lécuyer, C. & Joachimski, M.M. 2000: Comparing oxygen isotope records of Silurian calcite and phosphate δ18O compositions of brachiopods and conodonts. Geochimica et Cosmochimica Acta 64, 1859–1872. Williams, D.B. & Carter, C.B. 1996: Transmission Electron Micro- scopy: A Textbook for Materials Science, 730 pp. Plenum Press, New York. /ColorImageDict > /JPEG2000ColorACSImageDict > /JPEG2000ColorImageDict > /AntiAliasGrayImages false /DownsampleGrayImages true /GrayImageDownsampleType /Bicubic /GrayImageResolution 120 /GrayImageDepth -1 /GrayImageDownsampleThreshold 1.00000 /EncodeGrayImages true /GrayImageFilter /DCTEncode /AutoFilterGrayImages true /GrayImageAutoFilterStrategy /JPEG /GrayACSImageDict > /GrayImageDict > /JPEG2000GrayACSImageDict > /JPEG2000GrayImageDict > /AntiAliasMonoImages false /DownsampleMonoImages true /MonoImageDownsampleType /Bicubic /MonoImageResolution 300 /MonoImageDepth -1 /MonoImageDownsampleThreshold 1.00000 /EncodeMonoImages true /MonoImageFilter /FlateEncode /MonoImageDict > /AllowPSXObjects false /PDFX1aCheck false /PDFX3Check false /PDFXCompliantPDFOnly false /PDFXNoTrimBoxError true /PDFXTrimBoxToMediaBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXSetBleedBoxToMediaBox true /PDFXBleedBoxToTrimBoxOffset [ 0.00000 0.00000 0.00000 0.00000 ] /PDFXOutputIntentProfile (None) /PDFXOutputCondition () /PDFXRegistryName (http://www.color.org) /PDFXTrapped /Unknown /Description > >> setdistillerparams > setpagedevice