Ppp 2004-wings-authigenic minerals in fossil bone

Authigenic minerals in fossil bones from the Mesozoic ofEngland: poor correlation with depositional environments

Oliver Wings �

Department of Earth Sciences, University of Bristol, Queen’s Road, Bristol BS8 1RJ, UK

Received 15 January 2003; received in revised form 3 September 2003; accepted 13 October 2003

Abstract

Petrographic microscopy, X-ray diffractometry, and scanning electron microscopy^energy dispersion spectrom-etry studies of cavity infills of bone samples from English Mesozoic vertebrate deposits (Isle of Wight, Swanage, LymeRegis, Aust Cliff, Westbury Garden Cliff, Tytherington) allow a better understanding of the distribution of authigenicminerals in bone voids and shed light on the importance of diagenetic processes on bone preservation. Surprisinglyfew minerals were identified as void fillers in all depositional environments, carbonates, sulphides, oxides, andsulphates being most abundant. Calcite is present in almost all samples. Pyrite is very common in voids as well asincorporated into bone francolite. In many samples several generations of infills are found, often consisting of layeredcements. The co-occurrence of the common minerals indicates anaerobic, slightly alkaline, and often sulphate-reducing environments in the bone voids. Later oxidation is common in iron minerals. Calcite, pyrite, and barite lackunambiguous environmental control, whereas sphalerite is possibly an indicator for marine deposits. Freshwaterdeposits show no particular mineral combination which would help in separating them from other environments. Insome instances, sediment infills and pyrite show geopetal structures.6 2003 Elsevier B.V. All rights reserved.

Keywords: diagenesis; cementation; authigenic minerals; bones; dinosaurs; Mesozoic; England

1. Introduction

The early diagenesis and mineralisation of boneis a crucial process in the fossilisation of verte-brates, but this diagenesis has not been ad-equately studied. Existing studies report mainlydiscrete aspects of bone diagenesis (e.g. Parkerand Toots, 1970; Pfretzschner, 2001) or give re-

sults from speci¢c sites without comparing thesewith other depositional settings (e.g. Barker et al.,1997; Holz and Schultz, 1998; Hubert et al.,1996). Furthermore, most research has been con-ducted on archaeological bones (e.g. Price et al.,1992) which typically are not fully mineralisedand which might never have entered the fossil rec-ord because of this and because of burial in anenvironment with low preservation potential (e.g.upland soil). This results in a major gap regardingour knowledge of the taphonomy and diagenesisof fossil vertebrates remains. Therefore, a compre-hensive study of cavity in¢lls of fossil bone sam-

0031-0182 / 03 / $ ^ see front matter 6 2003 Elsevier B.V. All rights reserved.doi:10.1016/S0031-0182(03)00709-0

* Present address: Institut fu«r Pala«ontologie, Universita«tBonn, NuMallee 8, D-53115 Bonn, Germany.Tel. : +49-228-734683; Fax: +49-228-733509.

E-mail address: [email protected] (O. Wings).

PALAEO 3243 27-1-04 Cyaan Magenta Geel Zwart

Palaeogeography, Palaeoclimatology, Palaeoecology 204 (2004) 15^32

www.elsevier.com/locate/palaeo

mailto:[email protected]

http://www.elsevier.com/locate/palaeo

ples from various English vertebrate deposits wascarried out to clarify the interactions betweenbone francolite, authigenic minerals, and waterin the enclosing sediment during diagenesis. Thus,for the ¢rst time, a direct comparison betweenfossil bones from several di¡erent depositionalenvironments is possible.

Beside common carbonates, sulphides, and sul-phates, the voids were a priori expected to be¢lled with a large variety of localised minerals,resulting from di¡erences in local depositional en-vironments (e.g. geochemical parameters, ion sup-ply in pore water, presence of bacteria in£uencingmineral precipitation, and decay of organic parts)during a very early diagenetic stage. Bacteria-trig-gered precipitations of rare minerals like kutno-horite (Clarke and Barker, 1993) were presumedto be relatively common in fossil bones. Becausethe bones are preserved as ‘apatite’, authigenicphosphate was expected in bone voids. Authigenicquartz and chlorite can be major void ¢llers infossil bone (Hubert et al., 1996). Gypsum andiron oxides should be frequent as weathering

products. More investigated minerals are listedin Section 3.

Such a variety of minerals might allow an addi-tional clue for taphonomic and depositional set-tings of vertebrate fossils. On the other hand, thebone cavities could also be ¢lled during later dia-genetic stages and therefore would not show areal environmental signature. This study aims toclarify the processes that lead to mineral precip-itation in bone and when these processes occur.Samples were obtained from a wide range of en-vironments, including caves, £oodplains, lakes,and oceans. To test the in£uence of late diagenesisoperating over long time periods, the sites alsocover a broad age range: Triassic, Jurassic, andCretaceous.

Furthermore, samples collected at Aust Cli¡provide the opportunity to examine di¡erencesin bone preservation and diagenetic cementationindependent from a number of external variablesbecause the samples come from the same stratum.This study also investigates the origin andtaphonomy of the non-marine bone material at

Fig. 1. Map of sampled English vertebrate fossil localities: (1) Isle of Wight; (2) Swanage; (3) Lyme Regis; (4) Aust Cli¡; (5)Westbury Garden Cli¡; (6) Tytherington.


O. Wings / Palaeogeography, Palaeoclimatology, Palaeoecology 204 (2004) 15^3216

Aust Cli¡. Lastly, the study discusses the suitabil-ity of techniques for fossil bone analysis.

2. Area descriptions

A variety of bone samples of di¡erent ages anddiagenetic histories were collected from a range ofwell-studied depositional environments. The sam-ples were taken from English vertebrate fossil oc-currences including the Wealden on the Isle ofWight (IW), the Lower Cretaceous bone-richlimestone at Swanage (SW), the Liassic shales atLyme Regis (LR), the Rhaetic bone beds at AustCli¡ (AC) and Westbury Garden Cli¡ (WB), andthe Late Triassic ¢ssure ¢llings at Tytherington(TY) (Fig. 1). These fossil sites were selected forthis study because of the large amount of back-ground research available (e.g. Barker et al.,1997; Benton and Spencer, 1995; Raiswell andBerner, 1985), the variety of depositional environ-ments, and the accessibility of samples. Despitethe prominence of some of the sites, this studyis concerned with standard types of fossilisationwhich are common in the fossil record. No specialmodes of preservation, such as soft tissue preser-vation, are known from most of the localities.

2.1. Isle of Wight

Samples were collected from the southwesterncoast of the Isle of Wight (50‡38P13QN,1‡25P21QW) between Hanover Point and BarnesHigh in the Wessex Formation (Wealden Group,Lower Cretaceous). The strata are mainly mud-stones, marls, and sandstones, deposited in acoastal £oodplain with meanders, lakes, andponds in a warm to hot climate (Martill andNaish, 2001). The fauna includes ¢sh, crocodiles,turtles, and dinosaurs (Insole et al., 1998). Mostbones collected from the beach exposure lackedrock matrix and are unassignable to speci¢cbeds. Some samples attached to sandstone matrixevidently are from the Sudmoor Point Sandstone.The majority of bones are considered dinosaurianbased on their large size. The most abundant di-nosaurs in the Isle of Wight strata are Iguanodonand Hypsilophodon, and probably they comprise

the majority of the bones collected (Benton andSpencer, 1995). Identi¢cation of the type of bone(e.g. tibia) was mostly impossible in the abradedand broken fragments, although some are clearlyvertebrae.

2.2. Swanage

This heavily weathered material was collectedabout 200 m westward of Peveril Point in Swan-age (50‡36P22QN, 1‡56P54QW). All of the section isin the Upper Purbeck Beds (Durlston Formation,Lower Cretaceous). The bones are in the BrokenShell Limestone, a massive shelly bioclastic lime-stone, which yields mostly ¢sh and turtle remains(House, 1993). However, a well-preserved verte-bral centrum of a crocodile or dinosaur was re-covered. The Purbeck Beds were deposited in alagoonal or lacustrine environment similar to theWessex Formation of the Wealden. Unabradedshells of Viviparus in the Broken Shell Limestoneindicate a fully freshwater origin for the collectedspecimens, because the fragile shells would havebeen broken during transport to the sea.

2.3. Lyme Regis

The bones were collected from the HettangianBlue Lias Beds (lower Liassic, Lower Jurassic) incli¡s along the coastline from Seven Rock Point(approximately 1 km west of Lyme Regis) andCharmouth (50‡42P51QN, 2‡57P19QW). The inter-bedded mudstones and laminated shales withlimestones were deposited in a shallow stagnantmarine basin (Melville and Freshney, 1982). Allbones are from shale, except one fragment of aplesiosaur (?) long bone found in a limestonebolder on the beach. Most samples are brokenichthyosaurian bones: phalanges, long bones,and ribs. Several fragments were identi¢ed as ple-siosaur ribs because of their size and shape.

2.4. Aust Cli¡ and Westbury Garden Cli¡

As two of the British classic localities of Rhae-tian bone beds (Westbury Formation, PenarthGroup, Upper Triassic), Aust Cli¡ (51‡36P26QN,2‡37P20QW) and Westbury Garden Cli¡


O. Wings / Palaeogeography, Palaeoclimatology, Palaeoecology 204 (2004) 15^32 17

(51‡48P38QN, 2‡24P27QW) represent marine highenergy environments, perhaps intertidal deposits(Antia, 1979). The Aust Cli¡ stratum is a con-glomerate of silty mudstone clasts with a largevariety of vertebrate remains (scales, teeth, bones)in a calcite cemented matrix of angular quartzsand. The Westbury Garden Cli¡ bone bed is athin sandstone layer, with abundant shell andvertebrate fragments. Authigenic euhedral pyriteis often present as the only cement. For a widerbackground regarding taphonomy, sedimentol-ogy, and genesis of these two bone beds, see True-man and Benton (1997) and Martill (1999).

The samples from Aust Cli¡ were collectedfrom one stratum. Unweathered rock sampleswere found to contain a mixture of completeand broken bones ranging from fresh and shinyto weathered pale-brown skeletal elements. How-ever, most of the bones are abraded and/or frac-tured and were therefore easily penetrated by min-eral-precipitating water. The bone fragmentschosen for this study belong to a diverse groupof vertebrates, including terrestrial and marinearchosaurs. However, the majority of the AustCli¡ samples are ¢sh bones.

2.5. Tytherington

The Mesozoic ¢ssure system within the Carbon-iferous limestone of Tytherington (51‡36P06QN,2‡29P18QW) was essentially formed by carbonatedissolution, producing caves and undergroundwatercourses, which were later ¢lled with terres-trial, herpetofauna-rich sediments under sub-

aqueous or sub-aerial conditions (Whiteside,1983). The sediment is a carbonate-cementedquartz-rich siltite or arenite, often dolomitised(Whiteside, 1983). Pocket-like structures with sev-eral bones were found in some rocks, but layers ofdense silt without bones are also present. Silt in-traclasts were commonly found near bones. The-codontosaurus long bone and vertebral fragmentswere used for this study.

3. Methods

To compensate for the variability among bonevoid in¢lls, approximately 300 nondescript bonesamples were cut with a rock saw and examinedmacroscopically. Typical, unusual, and preferablylarger bone pieces were selected for further exami-nation. To compare the cements from a wide va-riety of samples, 87 large thin sections (50U50mm and 75U50 mm) were made. Several thinsections also were made from di¡erent parts oflarger bones in order to identify and investigateany di¡erences in the mineralogy of cavity ¢llings.In thin sections with more than one bone, onlythe largest bone was counted for statistical pur-poses to avoid counting of unwanted regional dia-genetic changes. In every thin section, all mineralsin the bone voids were determined qualitatively(Table 1). In addition, a semi-quantitative analy-sis for the abundance (vol%) of pyrite in the thinsections was carried out visually (Table 2). Theabundance is estimated in 10% steps.

Images of thin sections were taken with a Ni-

Table 1Mineral abundances in bone and voids

Locality Isle of Wight Swanage Lyme Regis Aust Cli¡ Westbury Garden Tytherington All samples

Number of samples 21 14 22 17 3 10 87Calcite in voids 20 (95%) 14 (100%) 20 (91%) 17 (100%) 2 (67%) 10 (100%) 83 (95%)Pyrite in bone 14 (67%) 13 (93%) 18 (82%) 17 (100%) 3 (100%) 5 (50%) 70 (81%)Pyrite in voids 17 (81%) 10 (71%) 14 (64%) 12 (71%) 3 (100%) 6 (60%) 62 (71%)Sphalerite in voids 1 (5%) 2 (14%) 12 (55%) 8 (47%) 0 (0%) 0 (0%) 23 (26%)Barite in voids 11 (52%) 0 (0%) 7 (32%) 1 (6%) 1 (33%) 0 (0%) 20 (23%)Siderite in voids 9 (43%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 9 (10%)Sediment in voids 13 (62%) 10 (71%) 4 (18%) 17 (100%) 3 (100%) 9 (90%) 56 (64%)

The numbers show the quantity of samples containing the respective mineral. The percent values in brackets are based on the to-tal sample number of each locality.



kon FX-35A camera attached to a Nikon Opti-phot-Pol petrographic microscope. Slides werescanned with a Nikon Coolscan 3 scanner andscale bars and arrows were added in Adobe Pho-toshop. All thin sections were examined petro-graphically with re£ective and transmitted lightmicroscopy. Bone samples with representativemineral in¢lls were selected for X-ray di¡ractom-etry (XRD) analysis. Data collection was per-formed using a Siemens D5000 powder di¡rac-tometer (Bragg^Brentano geometry, Cu KK, 40kV, 35 mA) from 10‡ to 65‡ 2a using a step sizeof 0.02‡ and variable divergence slits. Sampleswere ground in an agate mortar and analysed us-ing a rotating standard £at-plate sample holder.Due to the small amount of powder, one sample(SW 3) was analysed using a re£ection-free siliconsample holder.

The di¡ractograms were compared with thePDF-computer database (ICDD) using the com-puter software EVA (Bruker-AXS). Characteristicdi¡ractograms of minerals which were likely to befound in the samples were selected, comparedwith each sample, and summarised in Table 3.The tested minerals were chosen from existingobservations and descriptions of other authors.The following minerals were searched for butnot found in the XRD samples: hydroxyapatiteCa5(PO4)3(OH), chlorapatite Ca5(PO4)3Cl, car-bonate hydroxyapatite (dahllite) Ca10(PO4)3(CO3)3(OH)2, calcium phosphate carbonate Ca10(PO4)6CO3, siderite FeCO3, chalcopyrite CuFeS2,vivianite Fe3(PO4)2W8H2O, iron sulphide FeS, goe-thite FeO(OH). Sediment grain analyses are notdiscussed here.

Major and minor elements were obtained byelectron probe microanalyses (EDS). The uncov-ered thin sections were coated with carbon andanalysed with a Stereoscan 250 Mk 3 scanningelectron microscope (SEM) with enabled back-scatter detection (BSE). The operation conditionsfor BSE were 15 keV and 1^5 nAmps; with varieddwell times. The results were compared with themicroscopic evidence. Additional examinations ofmineral habits in etched bone samples were per-formed.

Tab

le2

Abu

ndan

ceof

pyrite

asvo

idin¢lls

Abu

ndan

ceof

pyrite

invo

l%of

void

¢lls

Total

numberof

samples

Alllocalities(in%

ofall

samples

withpy

rite

invo

ids;n=62)

In%

oftotalsample

number;n=87

n=21

n=14

n=22

n=17

n=3

n=10

Isle

ofWight

Swan

age

Lym

eRegis

AustCli¡

WestburyGarden

Tythering

ton

Num

berof

samples

withpy

rite

invo

ids

610

105

69

636

(58%

)41%

204

42

313

(21%

)15%

302

13(5%)

4%40

12

3(5%)

4%50

11

2(3%)

2%60

22(3%)

2%70

11(2%)

1%80

11(2%)

1%90 100

11(2%)

1%



Table 3Results of the XRD analysis

mineral formula IW 2: C IW 5: S SW 3: C+S SW 13: C+S LR 4: S LR 17: C AC 2: C+S AC 3: S TY 1: C+S

calcite CaCO3 +, Mga ++, Mga ++, Mga ++, Mga ++, Mga ++, Mga ++ ++magnesium calcite (Ca,Mg)(CO3)2 ++quartz SiO2 o o + + ++£uorapatite Ca5(PO4)3F ++ ++ ++ ++ ++ ++ ++carbonate £uorapatite(francolite)

Ca10(PO4)5CO3F1:5(OH)0:5 ++ ++

witherite BaCO3 ?oligonite Fe(Mn,Zn)(CO3)2 ++kutnohorite, magnesian Ca(Mn,Mg)(CO3)2 omarcasite FeS2 ? o ? + ? o ?pyrite FeS2 ++ o + + + ++ ? osphalerite ZnS oiron oxide Fe2O3 (PDF 16-0653) ? + oiron oxide Fe2O3 (PDF 16-0895) + o ? olepidocrocite FeO(OH) ?iron oxide hydroxide FeOOH ogypsum CaSO4 W2H2O obarite BaSO4 ? o

Listed are the original sample numbers. Abbreviations: IW, Isle of Wight; SW, Swanage; LR, Lyme Regis; AC, Aust Cli¡; TY, Tytherington; C, compacta; S,spongiosa (cancellous bone).Symbols: ++, phase dominant; +, phase important; o, phase present; ?, phase probably present but due to peak superposition not clearly established.

a Position of the calcite [104]-peak indicates small amounts of Mg.

PALAEO

324327-1-04

Cyaan

Magenta

Geel

Zwart

O.Wings/P

alaeogeography,Palaeoclim

atology,Palaeoecology

204(2004)

15^3220

Fig. 2. (a) Plesiosaur (?) longbone fragment. The continuation of pyrite from void in¢lls into the sparite-¢lled crack suggests analmost simultaneous formation. The bone mineral is partly recrystallised (arrows). Crossed Nicols (XN), 1/4 V plate. (b) Isle ofWight bone fragment showing geopetal pyrite in¢ll. The pyrite accumulated on the bottom of the voids (left side), and a ¢nalcalcite cement ¢lled the voids. Plain Polarised Light (PPL). (c) Pyrite ¢lling in vascular canals of an Isle of Wight sample, show-ing the formerly interconnecting pathways between the bone voids and cracks probably caused by crystallisation pressure. Thewhite parts are resin. PPL. (d) Isle of Wight bone fragment showing stalactitic pyrite. The initial growing direction of all stalac-tites has been identical, the variant stalactite on the bone fragment in the lower centre (arrow) was broken o¡ during a later dia-genetic stage. PPL. (e) Heavily pyritised bone fragment from Westbury Garden Cli¡. Most voids are ¢lled with blocky sparite.Pyrite-cemented quartz grains are visible in the sediment beyond the bone surface to the right. PPL. (f) Aust Cli¡ bone with asparite-in¢lled crack. The iron minerals in the bone are pyrite on the right (opaque) and hematite (brown colour) on the left sideof the crack. PPL (for colour see online version).



4. Results and analyses

4.1. Apatite ^ Ca5(PO4)3(F, Cl, OH) andPhosphates ^ [PO33

4 ]

The XRD results (Table 3) indicate two clearlydi¡erent fossil bone minerals, the more common£uorapatite Ca5(PO4)3F in seven samples and car-bonate-£uorapatite (francolite) Ca10(PO4)5CO3

F1:5(OH)0:5 in two samples (AC 2 and LR 17).Recrystallisation in speci¢c areas of bone is fre-quently visible under polarised light (Fig. 2a).This recrystallisation is common in cortical bonefrom all sites and is often restricted to permeableparts like voids (osteons and lacunae) or cracks.Small amounts of bone francolite were found tohave been replaced by pyrite. However, replace-ment of phosphate by calcium carbonate in dis-solved bone as described by Holz and Schultz(1998) was not observed. No evidence of diage-netic phosphate in¢ll of voids has been observedin any of the samples studied.

4.2. Calcite ^ CaCO3 and Mg^Calcite ^(Ca, Mg)(CO3)2

Calcite was found in every XRD sample and in95% of all samples. Blocky sparite, drusy mosaic,bladed, and ¢brous crystals exist in the bonevoids in a range of sizes. Both, non-ferroan andferroan calcite are present.

The position of the calcite [104] peak indicatesa small amount of Ca2þ substitution by Mg2þ inmost of the XRD samples. The position of thispeak in sample IW-5 at 29.7‡ 2a for example in-dicates magnesium-rich calcite which may containaround 10 mol% MgCO3 (see Goldsmith et al.,1961). Small amounts of Mg are often integratedinto the calcite lattice.

4.3. Pyrite and marcasite ^ FeS2

FeS2 was found in every XRD sample as mar-casite or pyrite. Due to the overlap of peaks, itwas not possible to prove the presence of theseminerals in samples TY 1 and AC 2. Re£ectivelight microscopy showed that most FeS2 ispresent as pyrite rather than marcasite, making

pyrite the second most common mineral after cal-cite. It is common in voids of all sizes (Fig. 2a^e)as well as in the sediment as integral cement orisolated grains. Euhedral and subhedral crystalsof pyrite were found to encrust voids, cracks,and entire bone surfaces, but framboidal pyritewas found most commonly. Occurrences of ‘glob-ular’ pyrite, as described by Pfretzschner (2001),were also found within bone ‘apatite’. Pyrite andcalcite cements were frequently found in close as-sociation. For example, a bone fragment collectedat Lyme Regis exhibits a band of pyrite enclosedbetween two di¡erent blocky sparite generations(Fig. 3). A sample from Aust Cli¡, however, onlyhas the voids in the central area of the cortexin¢lled with pyrite. In several Isle of Wight sam-ples, at least three generations of in¢lls werefound: (1) an initial pyrite coating, (2) calcite pre-cipitation, and (3) mixture between pyrite andcalcite ¢lling in the remaining pore space.

Fig. 3. Polished section of a cancellous bone from LymeRegis with three generations of void in¢lls. The outer part is¢lled with a coarse sparite, followed by a band of pyrite(pale zone) and ¢nally an in¢ll of blocky calcite (light greyin centre, for colour see online version).



In Tytherington bones, Fe minerals are veryrare, although the surface of bone samples fre-quently exhibited a thin, brownish, iron-rich cov-er. Any pyrite found within bone samples waspresent as in¢lls in osteons and lacunae and iso-lated framboidal grains in poikilotopic calcite,while hematite was found more commonly invoids.

The amount of pyrite found in bone and invoids can be quite variable even at one location(Table 1). In material collected from Aust Cli¡,for instance, pyrite commonly appears in the bonefrancolite, but rarely occurs in the voids. Only inone Westbury Garden sample does pyrite occupyall the pore space (Table 2). Overall, in 79% ofsamples which were found to contain pyrite, themineral in¢lled less than 20% of the available voidspace. Samples with smaller void spaces werecompletely in¢lled with pyrite, whereas largervoids were only partially ¢lled and often exhibita pyrite fringe.

4.4. Hematite Fe2O3 and iron hydroxidesFeO(OH); FeOOH

Several iron oxides and hydroxides were identi-¢ed in the XRD samples (Table 3) except forsample IW 5 which contained none. One AustCli¡ sample shows the same amount of iron min-erals in a bone fragment divided by a crack, onone side still preserved as pyrite, but on the otherside oxidised to hematite (Fig. 2f).

4.5. Sphalerite ^ ZnS

Microscopic investigation revealed sphalerite in26% of the samples. ZnS is mostly found as euhe-dral crystals embedded in sediment or calcite. Asmall ichthyosaurian carpal collected from LymeRegis was found to be completely in¢lled by spha-lerite, while other bones from Lyme Regis havesphalerite present in voids, brecciated bone, andcracks (Figs. 4b and 5a). In Aust Cli¡ samples,sphalerite was frequently found in sediment sur-rounding bone samples and as isolated grains inthe voids.

Among the XRD samples, ZnS was only foundin AC 2. The N-spacings of sphalerite are nearly

coincident with those of iron sulphide (FeS) orpyrite (FeS2), which makes identi¢cation moredi⁄cult. In the case of AC 2, the relative intensityof the N-spacings were used to identify sphale-rite.

4.6. Barite BaSO4 and Witherite BaCO3

With XRD, barite was found in AC 2 and mayalso be present in LR 17, although the presence ofsmall amounts of barite are di⁄cult to determinedue to overlapping N-spacings of major phases.Subhedral clusters of prismatic crystals were com-monly found in thin sections. Other growth typesobserved were isolated blades, fascicular, and eu-hedal crystals. A few samples were observed tohave entire pore spaces in¢lled with barite, eitherin sparry (Fig. 4b) or fascicular form (Fig. 4d).Samples from Lyme Regis exhibited isolated bar-ite grains scattered in the voids.

XRD analysis of a Swanage sample indicatedminor amounts of witherite BaCO3.

4.7. Quartz and other minerals in sediments

In XRD results, no quartz was found in thesamples from Lyme Regis and the Isle of Wight.Quartz was one of the major phases in sample TY1, important in AC 2 and AC 3 and present inSW 3 and SW 13.

Nevertheless, there is no authigenic quartz inany of the examined bones. The quartz found inthe samples (mainly Tytherington, Aust Cli¡, andSwanage) results from sediments in¢lling voids(Fig. 4e). The sediment grains range in size frommud to sand and were found to be concentratedin the outermost voids of samples collected fromSwanage and Aust Cli¡. Isolated grains of glau-conite were identi¢ed in two larger cancellousbone fragments from Aust Cli¡. In Lyme Regisbones, sediment was found in the cracks, com-monly those near the sediment/bone boundary,but not in the voids.

4.8. Siderite FeCO3 and OligoniteFe(Mn,Zn)(CO3)2

Siderite is relatively abundant as isopachous



Fig. 4. (a) Di¡erent bone colours caused by iron oxides in an Aust Cli¡ bone fragment. These layers are growth marks whichwere enhanced during the recrystallisation of the bone francolite. PPL. (b) Plesiosaurus dolichodeirus vertebra, Lyme Regis. Can-cellous bone with isotropic sphalerite (black) in the lower centre of image. The voids in the upper left area are completely in¢lledwith barite (light grey). All other voids are ¢lled with coarse sparite. XN. (c) Void in¢lls in an Isle of Wight bone. The ¢rst smallisopachous fringe consists of siderite (arrow). A framboidal pyrite crystal served as nucleus for the crystallisation of bladed bariteand the remaining pore space is ¢lled in by blocky sparite. PPL. (d) Arch-shaped fascicular barite in an Isle of Wight sample.XN. (e) Thecodontosaurus long bone section from Tytherington. The bone is heavily cracked and the voids are sediment ¢lled(calcite-bound arenite). PPL. (f) Bone fragment from Swanage. The voids are ¢lled in by isolated framboidal pyrite clusters, fol-lowed by di¡erent generations of spar with changing iron content in the calcite. The void in the upper left corner is in¢lled withcarbonate-cemented silty sediment. XN, sensitive tint plate (for colour see online version).



cement in Isle of Wight material. Siderite is notidenti¢able in any of the samples from any othersite studied. Oligonite, which is a siderite (FeCO3)in which the Fe2þ is partly substituted by Mn2þ

and/or Zn2þ, was found with XRD to be a majorphase in sample IW 5. EDS analysis of thissample identi¢ed only Fe ions. However, the char-acteristic peak positions in the X-ray di¡racto-gram clearly show that the mineral is not puresiderite.

4.9. Gypsum CaSO4W2H2O

One sample from Aust Cli¡ (AC 3) was foundto contain CaSO4.2H2O in the XRD, identi¢ed byits characteristic [020] peak at 11.7‡ 2a. Micro-scopic investigation of the sample did not detectany gypsum.

4.10. Kutnohorite Ca(Mn,Mg)(CO3)2

A magnesium-rich kutnohorite was identi¢edwith XRD in sample SW 3, and microscopicinvestigation revealed in several thin sectionsfrom the Isle of Wight kutnohorite as an iso-pachous rim and as a non-layered inner-void in¢llprecipitated on siderite and pyrite-rich sedimentin¢ll.

5. Discussion and conclusions

5.1. General features of fossil bone

Di¡erent sections of the same bones werestudied on a plesiosaur rib from Lyme Regis (ap-proximately 20 cm in length) and a sauropod (?)bone fragment from the Isle of Wight (approxi-mately 15 cm in length). In both samples, uni-formity in the mineralogy of cavity in¢lls hasbeen observed. Unfortunately, veri¢cation of thehypothesis that longer bones are generally charac-terised by homogeneous mineral in¢lls was handi-capped by the limited sample size.

Cements such as calcite, siderite, and kutnohor-ite commonly show a layered appearance (Fig.4f), indicating microbial participation during theirformation (Clarke and Barker, 1993; Flu«gel,1982; Mortimer et al., 1997; Sagemann et al.,1999). Framboidal pyrite formation is also en-hanced by biologically mediated reactions (Can-¢eld and Raiswell, 1991).

5.2. Geopetal structures

In addition to common geopetal pyrite-richsediment in¢lls, pure pyrite also shows evidenceof gravitational orientation (Fig. 2b). The occur-

Fig. 5. (a) BSE-image of a Lyme Regis thin section showing four authigenic minerals in voids and brecciated bone. Note thesediment outlining euhedral sphalerite. (b) BSE-image of an Isle of Wight thin section showing close interactions between precipi-tated minerals. The minerals siderite (Fe), calcite (Ca), and kutnohorite (Mn, Ca) feature a gradual transition. Note that the kut-nohorite is a later in¢ll, predated by siderite. (c) BSE-image of an Isle of Wight thin section showing voids with several alternat-ing layers of calcite (dark grey) and pyrite (white).



rence of pyrite in distinct regions of bone voids isan indication of geopetal structures. The precipi-tation of pyrite can be initiated by bacteria whichconcentrate in low positions in the voids due togravity. This type of geopetal pyrite in¢lls was notreported by Hudson (1982), who described thegeopetal in¢lling of ammonite chambers with py-ritic sediment and pyrite stalactite. In this study,pyritic sediment similar to that found in ammo-nite chambers (Hudson, 1982) was found in bonesamples from all sites, whereas the generally rarestalactitic pyrite was only found in one samplefrom the Isle of Wight (Fig. 2d).

5.3. Cracking and brecciation

The sediment/bone interface seems to be easilypermeated by pore water. This might result fromcompaction-caused forces: sediment and bonehave di¡erent deformation characteristics andtherefore the interplay of compression and tensionoften causes a small gap on the bone surface. Inmost of the samples used in this study, this cavityis ¢lled with a small fringe of calcite cement andthe outermost region of any broken bone is some-times heavily pyritised which is not necessarily

due to the initial decay of soft tissue (Pfretzsch-ner, 2000). Pyrite is only enriched in the outer-most parts of the bone, indicating that this pyri-tisation happened later in the diagenetic process.The necessary supply of ion-carrying pore waterswas made possible by the gap between bone andsediment.

Almost every bone shows cracks (Fig. 2a). Inmany of the cancellous bone samples, crushed andbrecciated areas are visible (Fig. 5a). These zoneswere interpreted by Hubert et al. (1996) as evi-dence of trampling by large animals before thebones were buried. This explanation seems veryimplausible because such zones of crushing arealso commonly found in bones from fully marinesediments, like the Lyme Regis shales, and theyoften exhibit a di¡erent mineral in¢ll. Hence, adiagenetic origin due to overburden pressuremust be assumed for these structures. Pyrite isconcentrated in the crushed part of one LymeRegis sample which implies that precipitationtook place soon after the crushing occurred.

Small-scale cracks are much more likely to befound in certain areas of bone tissue. Dense boneparts are more a¡ected by the cracking than can-cellous areas. The structure of the cancellous bone

Fig. 6. Mineral occurrences in bone and voids. The graphs are created from data in Table 1. The values are based on the totalnumber of samples from each locality. Note the limited signi¢cance of some data due to the low number of samples from somelocalities (e.g. Westbury Garden Cli¡ with 3 samples).



is probably more £exible under initial burial pres-sure than compact bone, but tends to completelycollapse and brecciate under high compactionpressure.

5.4. Mineral occurrences

Precipitation and mineral replacement are twoof the complex diagenetic physicochemical pro-cesses which occur during in¢lling of openingsin bone (Downing and Park, 1998; Pate et al.,1989; Piepenbrink, 1989; Williams and Marlow,1987). This study revealed that relatively few min-erals are responsible for the in¢ll of bone voidsduring diagenesis. The most common in¢ll min-erals are carbonates (CaCO3, FeCO3, BaCO3),sulphides (FeS, ZnS), oxides (FeO, Fe2O3), andsulphates (BaSO4, CaSO4). The summarised para-genesis of the most important mineral in¢lls fromall localities can be seen in Fig. 7.

Calcite is present in almost all samples, indicat-ing its importance as a void in¢ller. The variouscarbonate cements observed in the material arebeyond the scope of this paper and are thereforediscussed in detail elsewhere (Wings, in prepara-tion). The most prevalent in¢ll observed was spar-ite, which appears to have been precipitated dur-ing the later stages of diagenesis, as suggested byFlu«gel (1982).

Pyrite is very abundant in the voids as well asin bone ‘apatite’. The reasons for the pyrite abun-

dance in fossil bone were discussed in detail byPfretzschner (2000, 2001), whereas the mecha-nisms of pyrite formation were summarised byCan¢eld and Raiswell (1991). The pyrite contentin the bone can vary very much in di¡erent bonesnext to each other in the sediment or even in thesame bone (Fig. 2e), implying di¡erent permeabil-ity pathways or a complex arrangement of micro-environments.

The described diagenetic order in several Isle ofWight samples is di¡erent from the very late pre-cipitation of pyrite described by Pfretzschner(2001). Samples used in this study demonstratethat both calcite and pyrite co-precipitated inter-mittently as a result of varying pore water chem-istry. An example of this pattern of mineralisationcan be seen in a crack in¢ll of a Lyme Regisplesiosaur bone (Fig. 2a). The integrated pyritein the sparite is a strong indicator for simulta-neous formation of both. Pyrite is also commonas crack in¢ll in Isle of Wight material, here sug-gesting a later diagenetic origin.

The higher pyrite content inside the voids ofSwanage material suggests that the decay of theorganic tissue was incomplete when the in¢ll be-gan ^ otherwise the same amount of pyrite wouldbe found in the surrounding sediment. However,the total amount of iron in body tissue is very low(about 4 g in an adult human; Schmidt-Nielsen,1997), making organic tissue a poor source of Fefor pyrite formation. The most important sourceof reactive iron are sedimentary iron minerals(Can¢eld, 1989; Can¢eld and Raiswell, 1991;Can¢eld et al., 1992). Yet, early pyrite formationis also controlled by the availability of sulphide(Can¢eld et al., 1992; Raiswell and Berner, 1985)and by the Eh/pH conditions caused by decay oforganic matter (Can¢eld and Raiswell, 1991;Krumbein and Garrels, 1952; Pfretzschner, 1998).

Interestingly, in Tertiary bones studied byParker (1966) and Parker and Toots (1970), nopyrite or any other mineral containing sulphurwas detected. This may be a result of late diage-netic oxidisation, as iron oxides were found(Parker, 1966; Parker and Toots, 1970). Limonite,hematite, and other iron oxides (Table 3) in thesamples are mainly weathering products of pyrite(Barker et al., 1997).

Fig. 7. Simpli¢ed paragenetic diagram for all localities.Listed are the most important mineral in¢lls in all bonevoids. Di¡erences between the localities are relatively insig-ni¢cant and are characterised mainly by the absence of someminerals (see Fig. 6).



The bands of darker coloured iron oxides inbone francolite (Fig. 4a) may have been causedby varying primary ion contents in the bone tis-sue, but more plausible is an early diagenetic pen-etration of iron-rich water into the collagen ma-terial along the borders of the recrystallisedcollophane/francolite areas (Pfretzschner, 2000).Therefore, the denser arrangement of francolitecrystals along growth marks caused a slowerspeed of the recrystallisation, and permitted thestorage of colouring ions due to a longer reactiontime. This distinctive layering, likely preservedduring recrystallisation of bone as ‘globular’ py-rite (Pfretzschner, 2001), is also visible inside thebone material. Some smaller cracks in other sam-ples were observed to be surrounded by FeO of alater diagenetic origin, indicating that iron miner-als were transported into the cracks and pene-trated the outer bone francolite.

The simultaneous precipitation of calcite andhematite (Holz and Schultz, 1998) is only possibleunder oxidising conditions (Krumbein and Gar-rels, 1952) and was not observed in this study.Far more common in the fossil record is thenearly concurrent precipitation of pyrite andcalcite (Krumbein and Garrels, 1952), and thelater oxidisation of pyrite to hematite. The con-current appearance of calcite and pyrite meansthat ions of both minerals were present in thepore £uids and precipitated interchangeably(Fig. 2a).

The isolated sphalerite grains in Aust Cli¡ sam-ples could also have been deposited and are notnecessarily authigenic. However, most sphaleritewas found to completely in¢ll the voids, indicat-ing precipitation in situ. The presence of sphaler-ite in cracks demonstrates that the mineralisationpostdated the main compaction phase, as reportedby Martill (1987). Overall, sphalerite is more com-mon than barite which was found in 23% of allsamples. Barite in the Isle of Wight material pre-dates most of the calcite cementation (Figs. 2dand 4c) and therefore cannot be considered alate-phase mineral in¢ll as reported by Martill(2001). The witherite founding in one samplemay have been formed by pseudomorph altera-tion of barite, which was not present in the sam-ple. Siderite always appears as a very early phase,

often forming the ¢rst isopachous cement in thevoids.

The sediment grains in the outermost bonevoids may imply an early in¢ll of broken bonewith embedding sediment and silicates. The distri-bution of quartz grains in Tytherington samples ismore uniform throughout the bone voids, perhapsindicating a longer transport and stronger disinte-gration of bone before sedimentation. The consis-tent grain distribution may also be a result ofstronger groundwater currents at Tytherington.Sediment in¢lling of the tiny surface cracks inLyme Regis bones may have occurred early inthe diagenesis when the sediment was still mobileenough to penetrate these spaces. The mineral in-¢lls, on the other hand, are the same in cracks andvoids, and likely ¢lled in the cavities later in thediagenesis.

The complete absence of authigenic quartz inthe bones indicates that soluble silica was not amajor component in the groundwater. This is con-sistent with the observation that the sediments atthe examined localities o¡er no exceptionally richsource for soluble silica, like silicic ash, which isperhaps the most common source of silica ce-ments in bones (personal observations in the Mor-rison Formation and the White River Group,USA, and in the Can‹ado¤n Calcareo Formation,Chubut province, Argentina).

Gypsum can be a sign of sample weathering(Barker et al., 1997). The Aust Cli¡ materialused in this study was very fresh and withoutany sign of recent weathering, which may sug-gest that any gypsum precipitation took placeduring a period of exposure prior to ¢nal bur-ial of the bone. On the other hand, as gypsumis not very resistant against weathering and ero-sion, the in¢ll precipitated probably quite recent-ly.

It is interesting that kutnohorite appears notonly as an isopachous rim, as reported by Clarkeand Barker (1993) and Barker et al. (1997), butalso as a non-layered inner-void in¢ll precipitatedon siderite and pyrite-rich sediment in¢ll. Thispermits speculation about the biological originof the kutnohorite, as suggested by Clarke andBarker (1993). A bacteria ¢lm in a later stage ofvoid in¢ll seems implausible because all nutrients



set free by the organic decay would likely havealready been consumed.

5.5. Geochemical conditions

The co-occurrence of calcite, pyrite, barite, andsiderite indicates a mineral stability ¢eld in theregion of pH7 to pH8, representing slightly alka-line conditions (Krumbein and Garrels, 1952).The redox potential of the burial environmentappears to have been slightly variable. While sid-erite precipitates at Eh=30.1, calcite and pyriteprecipitation is most common at Eh=30.3(Krumbein and Garrels, 1952). All minerals pre-sumably precipitated under anoxic conditions.Some minerals, especially the sulphides, under-went later oxidation, most commonly from pyriteto hematite. Unfortunately, the presence of spha-lerite and barite does not allow a tight constrain-ing of the geochemical conditions during theirformation. Both are common minerals with awide distribution (Klein and Hurlbut, 1993).

The question why so few minerals are respon-sible for bone void in¢lls in di¡erent environ-ments is perhaps answered by the chemical com-position of the pore waters. The constituent ionsof all minerals are common in groundwater aswell as marine water. As expected by the ion com-position of water, the most common ions (CO23

3 ,SO23

4 , Ca2þ, Fe3þ) formed the most abundantminerals (carbonates, sulphides, and sulphates)in the anaerobic, alkaline, and often sulphate-re-ducing environments of the bone voids. The ac-tual composition of water may easily vary on aseasonal basis (Hedges and Millard, 1995), caus-ing the precipitation of di¡erent mineral genera-tions (e.g. Fig. 5c) in a reasonably short time.However, the complete absence of some mineralgroups, like all Cu2þ-compounds and authigenicphosphate, is rather remarkable. Especially phos-phatisation of the bone voids should be expectedin phosphate-rich deposits with abundant organicmatter available and an oxygen-depleted environ-ment (Pre¤vot and Lucas, 1997).

5.6. Occurrence of phosphate

The presence of £uorapatite and francolite

shows a diagenetic change from original, unal-tered bone mineral, consisting of microcrystal-line, non-stoichiometric carbonate hydroxyapatiteCa5(PO4)3(OH) (Francillon-Vieillot et al., 1990;Hubert et al., 1996) to francolite. Phosphaticminerals have very rarely been interpreted authi-genic in bone (Macquaker, 1994) and these in-terpretations are argumentative. Martill (1999)found no authigenic phosphate in bone samplesfrom the same locality described by Macquaker(1994).

The absence of diagenetic phosphate in¢ll ofvoids is noteworthy. Phosphatic coprolites com-prise 20% of Aust Cli¡ sediments, indicating anample supply of phosphate. Phosphorites formunder the same Eh/pH conditions as calcite andpyrite (Krumbein and Garrels, 1952). These twominerals are most common as bone void in¢lls.However, the greatest di¡erence between experi-mental phosphatisation and the diagenetic condi-tions obtained for the bone samples is in the pHvalue. Authigenic phosphate forms during anoxicand acidic conditions (Kear et al., 1993) whereasthe void-¢lling minerals point to alkaline settings.Perhaps some kind of chemical rejection at thebone surface prevented phosphate ions from pre-cipitating in bone cavities. Special organic sub-stances which protected, for instance, aragoniticnacre of ammonites from solution (Keupp, 2000)could also have prohibited the phosphate precip-itation in bone voids. Yet, any speculation on thatsubject is idle without future experimental re-search.

5.7. Comparison to voids in carbonate fossils

Compared with chambered or segmented fossilswith a primary carbonate mineralogy (e.g. cepha-lopod shells) the bone void in¢lls show no greatdi¡erences in mineralogy. Calcite and pyrite arethe most abundant minerals in both fossil types.The presence of siderite, barite, and sphalerite isalso common in ammonites (Keupp, 2000). Phos-phatisation of organic tissue and the replacementof the shell with phosphate is common withammonites (Keupp, 2000; Weitschat, 1986), incontrast to fossil bone. In summary, the organichard part mineralogy (aragonite and calcite in



shells ; ‘apatite’ in bones), which do provide theporosity for authigenic mineral precipitation,is apparently irrelevant not only to pyrite forma-tion (Hudson, 1982), but also to most otherauthigenic minerals with the exception of phos-phate.

5.8. Relationship between minerals anddepositional environment

The main minerals found in all bone sampleslack apparent control by their environment offormation. Especially the distribution of calciteand pyrite does not follow a clear pattern. Never-theless, a few di¡erences in mineral distributionare obvious and permit limited insights (Fig. 6).Siderite is only present in Isle of Wight material,perhaps indicating a special microbial in£uence inthese bones (Mortimer et al., 1997). The karst siteTytherington is lacking sphalerite and barite, pos-sibly an indication for a restricted supply of thespeci¢c ions Zn2þ and Ba2þ. Tytherington bonealso shows less pyrite in bone and pores, resultingfrom the more aerated karst environment. Spha-lerite is abundant in two of the three marine de-posits. In addition, the absence of sphalerite in themarine Westbury Garden Cli¡ material could becaused by the insu⁄cient sample size. Therefore,sphalerite is a possible indicator for marine depos-its. In contrast, barite is present in marine andfreshwater deposits, but not consistently, thuslacking any environmental signal. The freshwaterdeposits of the Isle of Wight and Swanage showno unambiguous mineral combination whichcould help separate them from other environ-ments. Sedimentary grains are common in allsites except Lyme Regis. This is probably due todi¡erences in depositional energy levels in theaquatic environment. Abrasion opens up manynew pathways for sediment grains into thebone, whereas even ¢ne-grained sediments likethe Lyme Regis mudstones cannot penetratebones without su⁄cient connections to thebone cavities.

The greater variation in the mineralogy of Isleof Wight and Lyme Regis bone voids comparedwith the other sites can be explained by the great-er diversity of sediment types at these two local-

ities. No signi¢cant mineralogical variations werefound among samples from Aust Cli¡. Conse-quently, any distinction between bones of marineand non-marine animals was not possible on thisbasis.

5.9. Synopsis

In summary, authigenic minerals in the exam-ined bone voids show only little relationship tosedimentary settings. No unique mineral associa-tion was found in speci¢c sediments or rocks. Thedi¡erences in minerals in fossil bone voids repre-sent only very weak signals of the depositionalenvironment. Geopetal structures in bone voidsprovide useful information about taphonomicand diagenetic processes, because they revealtransport and reburial.

5.10. Suitability of techniques for fossil boneanalysis

Petrographic microscopy of thin sections is theeasiest and fastest method for identi¢cation ofminerals in voids. On one hand, bone thin sec-tions are often not ground down to exactly 30Wm because histologic features of fossil bonesare poorly visible at this standard thickness ofpetrographic thin sections. On the other hand,mineral optics demands this thickness for thinsection studies. Hence, petrographic microscopydata of bones may be inaccurate and should beinterpreted with care. For a general survey ofmineral contents, EDS/BSE are the best optionsand can also help in providing details of crystalgrowth (Fig. 5b,c). EDS is not very sensitive tosmall amounts of foreign ions in the crystallattice, as concentrations 6 2^5% are hardlydetectable. WDS (microprobe) is more sensi-tive but rather labour-intensive. A good alter-native to WDS can be XRD which is oftenaccurate in detection of small amounts of min-erals. Adversely, a peak overlap of certainminerals does often occur and prohibits un-ambiguous detection. Therefore, a combina-tion of methods, EDS/BSE with SEM, petro-graphic microscopy, and XRD gives very goodresults.



Acknowledgements

This research was conducted during my tenureas a Marie Curie Fellow at the University of Bris-tol. I would like to thank Mike Benton for super-vising my research. Discussions with David Mar-till and Martin Sander clari¢ed my view ondi⁄cult diagenetic features. Derek Briggs is ac-knowledged for a helpful discussion about theorigin and habits of pyrite. Robert van Geldernhelped with the XRD measurements and dis-cussed the XRD results with me. Stuart Kearnsprepared the SEM samples and gave an introduc-tion to the SEM at Bristol. I am indebted to anumber of people for providing samples for thisstudy (in alphabetic order) : Denver Fowler, MickGreene, Walter Joyce, David Martill, Octa¤vioMateus, Martin Sander, Gu«nther Viohl, andAdam Yates. I want to express my gratitude toClaudia Tro«gl, who helped preparing the ¢guresand to Patrice Hornibrook, Walter Joyce, JoeMacquaker, Martin Sander, and two anonymousreviewers for improvements to earlier drafts ofthe manuscript. The European Commission Im-proving Human Research Potential Programmefunded this research.

References

Antia, D.D.J., 1979. Bone-beds: A review of their classi¢ca-tion, occurrence, genesis, diagenesis, palaeoecology, weath-ering and microbiotas. Mercian Geol. 7, 93^174.

Barker, M.J., Clarke, J.B., Martill, D.M., 1997. Mesozoic rep-tile bones as diagenetic windows. Bull. Soc. Geol. Fr. 168,535^545.

Benton, M.J., Spencer, P.S., 1995. Fossil Reptiles of GreatBritain. Chapman and Hall, London, 386 pp.

Can¢eld, D.E., 1989. Reactive iron in marine sediments. Geo-chim. Cosmochim. Acta 53, 619^632.

Can¢eld, D.E., Raiswell, R., 1991. Pyrite formation and fossilpreservation. In: Briggs, D.E.G. (Ed.), Taphonomy. Ple-num, New York, pp. 337^387.

Can¢eld, D.E., Raiswell, R., Bottrell, S., 1992. The reactivityof sedimentary iron minerals toward sul¢de. Am. J. Sci. 292,659^683.

Clarke, J.B., Barker, M., 1993. Diagenesis in Iguanodon bonesfrom the Wealden Group, Isle of Wight, southern England.Kaupia 2, 57^65.

Downing, K.F., Park, L.E., 1998. Geochemistry and early dia-genesis of mammal-bearing concretions from the Sucker

Creek Formation (Miocene) of southeastern Oregon. Palaios13, 14^27.

Flu«gel, E., 1982. Microfacies Analysis of Limestones. Springer,Berlin, 633 pp.

Francillon-Vieillot, H., Bu¡re¤nil, V.d., Castanet, J., Ge¤raudie,J., Meunier, F.J., 1990. Microstructure and mineralizationof vertebrate skeletal tissues. In: Carter, J.G. (Ed.), SkeletalBiomineralization: Patterns, Processes and EvolutionaryTrends. Van Nostrand-Reinhold, New York, pp. 471^530.

Goldsmith, J.R., Graf, D.L., Heard, H.C., 1961. Lattice con-stants of the calcium^magnesium carbonates. Am. Mineral.46, 453^457.

Hedges, E.M., Millard, A.R., 1995. Bones and groundwater:Towards the modelling of diagenetic processes. J. Archaeol.Sci. 22, 155^164.

Holz, M., Schultz, C.L., 1998. Taphonomy of the south Brazil-ian Triassic herpetofauna; fossilization mode and implica-tions for morphological studies. Lethaia 31, 335^345.

House, M.R., 1993. Geology of the Dorset Coast. Geologists’Association Guide 22, Geologists’ Association, Oxford, 164pp.

Hubert, J.F., Panish, P.T., Chure, D.J., Prostak, K.S., 1996.Chemisty, microstructure, petrology, and diagenetic modelof Jurassic dinosaur bones, Dinosaur National Monument,Utah. J. Sediment. Res. 66, 531^547.

Hudson, J.D., 1982. Pyrite in ammonite-bearing shales fromthe Jurassic of England and Germany. Sedimentology 29,639^667.

Insole, A., Daley, B., Gale, A., 1998. The Isle of Wight. Geol-ogists’ Association Guides 60, Geologists’ Association, Lon-don, 132 pp.

Kear, A.J., Martill, D.M., Wilby, P.R., 1993. Phosphatizationof soft-tissue in experiments and fossils. J. Geol. Soc. Lond.150, 1035^1038.

Keupp, H., 2000. Ammoniten. Jan Thorbecke, Stuttgart, 165pp.

Klein, C., Hurlbut, C.S., 1993. Manual of Mineralogy (afterJames D. Dana). Wiley, New York, 681 pp.

Krumbein, W.C., Garrels, R.M., 1952. Origin and classi¢ca-tion of chemical sediments in terms of pH and oxidation^reduction potentials. J. Geol. 60, 1^33.

Macquaker, J.H.S., 1994. Palaeoenvironmental signi¢cance of‘bone-beds’ in organic-rich mudstone successions: An exam-ple from the Upper Triassic of south-west Britain. Zool.J. Linn. Soc. 112, 285^308.

Martill, D.M., 1987. A taphonomic and diagenetic case studyof a partially articulated ichthyosaur. Palaeontology 30,543^555.

Martill, D.M., 1999. Bone beds of the Westbury Formation.In: Martill, D.M. (Ed.), Fossils of the Rhaetian PenarthGroup. Palaeontological Association, London, pp. 49^64.

Martill, D.M., 2001. Taphonomy and Preservation. In: Naish,D. (Ed.), Dinosaurs of the Isle of Wight. PalaeontologicalAssociation, London, pp. 49^59.

Martill, D.M., Naish, D., 2001. The geology of the Isle ofWight. In: Naish, D. (Ed.), Dinosaurs of the Isle of Wight.Palaeontological Association, London, pp. 25^43.



Melville, R.V., Freshney, C.E., 1982. The Hampshire Basinand adjoining areas. British Regional Geology, Her Majes-ty’s Stationary O⁄ce, London, 146 pp.

Mortimer, R.J.G., Coleman, M.L., Rae, J.E., 1997. E¡ect ofbacteria on the elemental composition of early diageneticsiderite; implications for palaeoenvironmental interpreta-tions. Sedimentology 44, 759^765.

Parker, R.B., 1966. Electron microprobe analysis of fossilbones and teeth. Geol. Soc. Am. Spec. Pap. 101, 415^416.

Parker, R.B., Toots, H., 1970. Minor elements in fossil bone.Geol. Soc. Am. Bull. 81, 925^932.

Pate, F.D., Hutton, J.T., Norrish, K., 1989. Ionic exchangebetween soil solution and bone: Toward a predictive model.Appl. Geochem. 4, 303^316.

Pfretzschner, H.-U., 1998. Fru«hdiagenetische Prozesse bei derFossilisation von Knochen. Neues Jahrb. Geol. Pala«ontol.Abh. 210, 369^397.

Pfretzschner, H.-U., 2000. Pyrite formation in Pleistocenebones ^ a case of very early mineral formation during dia-genesis. Neues Jahrb. Geol. Pala«ontol. Abh. 217, 143^160.

Pfretzschner, H.-U., 2001. Pyrite in fossil bone. Neues Jahrb.Geol. Pala«ontol. Abh. 220, 1^23.

Piepenbrink, H., 1989. Examples of chemical changes duringfossilization. Appl. Geochem. 4, 273^280.

Pre¤vot, L., Lucas, J., 1997. Phosphate. In: Crowther, P.R.(Ed.), Palaeobiology: A Synthesis. Blackwell, Oxford/Bos-ton, pp. 256^257.

Price, T.D., Blitz, J., Burton, J., Ezzo, J.A., 1992. Diagenesisin prehistoric bone: Problems and solutions. J. Archaeol.Sci. 19, 513^529.

Raiswell, R., Berner, R.A., 1985. Pyrite formation in euxinicand semi-euxinic sediments. Am. J. Sci. 285, 710^724.

Sagemann, J., Bale, S.J., Briggs, D.E.G., Parkes, R.J., 1999.Controls on the formation of authigenic minerals in associ-ation with decaying organic matter: An experimental ap-proach. Geochim. Cosmochim. Acta 63, 1083^1095.

Schmidt-Nielsen, K., 1997. Animal physiology: Adaptationand Environment. Cambridge University Press, Cambridge,607 pp.

Trueman, C.N., Benton, M.J., 1997. A geochemical method totrace the taphonomic history of reworked bones in sedimen-tary settings. Geology 25, 263^266.

Weitschat, W., 1986. Phosphatisierte Ammonoideen aus derMittleren Trias von Central Spitzbergen. Mitt. Geol.-Palae-ontol. Inst. Univ. Hamburg 61, 249^279.

Whiteside, D.I., 1983. A Fissure Fauna from Avon. Ph.D.Thesis, University of Bristol, Bristol, 229 pp.

Williams, C.T., Marlow, C.A., 1987. Uranium and Thoriumdistributions in fossil bones from Olduvai Gorge, Tanzaniaand Kanam, Kenya. J. Archaeol. Sci. 14, 297^309.

Wings, O., in preparation. Microfacies methods applied tocarbonate cements in fossil bone ^ a new approach tobone diagenesis.



Education

Ppp 2004-wings-authigenic minerals in fossil bone