Bivalve taphonomy in tropical mixed siliciclastic ...geosci.uchicago.edu/pdfs/kidwell/Best-Kidwell-2000-I.pdf · ical siliciclastic sediments. Focusing on unambiguous damage to bivalve

q 2000 The Paleontological Society. All rights reserved. 0094-8373/00/2601-0006/$1.00

Paleobiology, 26(1), 2000, pp. 80–102

Bivalve taphonomy in tropical mixed siliciclastic-carbonatesettings. I. Environmental variation in shell condition

Mairi M. R. Best and Susan M. Kidwell

Abstract.—Contrary to the geological stereotype of pure-carbonate reef platforms, approximately50% of shallow shelf area in the Tropics is accumulating siliciclastic and mixed siliciclastic-car-bonate sediments. Taphonomic characterization of these settings is thus essential for assessing var-iation among major facies types within the Tropics, as well as for eventual comparison with higher-latitude settings. Our grab samples and dredge samples of bivalve death assemblages from ninestations in five subtidal habitats in a large marine embayment of Caribbean Panama (Bocas delToro) provide the first actualistic information on the taphonomic condition of shells in Recent trop-ical siliciclastic sediments. Focusing on unambiguous damage to bivalve shell interiors, we foundthat the quality of shell preservation in fine-grained siliciclastics is superb: commonly K10% ofspecimens are affected by encrustation, boring, edge-rounding, and fine-scale surface alterationvia dissolution, microbioerosion, and maceration. Pure-carbonate and mixed siliciclastic-carbonateenvironments containing hard substrata (patch reefs, Halimeda gravelly sand, mud among patchreefs) contain higher numbers of more severely damaged shells (generally .25%) and also higherdiversities of fossilizable encrusters and borers. Disarticulation and fragmentation are pervasiveacross all environments and are probably related to predation rather than to postmortem processes.As in other shallow subtidal study areas, the taxonomic compositions of death assemblages havenot been homogenized by postmortem transport but show high spatial fidelity to the distributionof living species. Assemblages from the five sedimentary environments have distinct taphonomicsignatures, but the strongest differences are between the two fine-grained, exclusively soft-sedi-ment siliciclastic environments on the one hand and the three environments containing hard sub-strata on the other. Experimental tests for rates and agents of damage, still in progress, indicatethat the most critical environmental variables are exhumation cycles and burial rate. Bivalve deathassemblages from Bocas del Toro demonstrate that damage levels in tropical fine-grained silici-clastic environments are much lower than in closely associated reefs and algal sands, and suggesta less filtered record of biological information.

Mairi M. R. Best and Susan M. Kidwell. Department of the Geophysical Sciences, University of Chicago,Chicago, Illinois 60637. E-mail: [email protected] and [email protected]

Accepted: 29 June 1999

Introduction

One of the major unknowns in marine ta-phonomy is the effect of climate and latitude onpatterns of skeletal preservation. The physicalcondition of death assemblages has been stud-ied extensively in temperate latitudes, and anal-ogous information for tropical carbonates hasgrown rapidly, especially in the last decade(mollusk studies by Ekdale 1977; Tudhope andScoffin 1984; Miller 1988; Tudhope 1989; Par-sons 1989, 1993; Miller et al. 1992; Dent 1995,1996; Perry 1996, 1998; echinoderms by Nebel-sick 1992, 1995; Llewellyn and Messing 1993;Nebelsick et al. 1997; reef corals by Pandolfi andMinchin 1995; Greenstein and Moffat 1996;Greenstein and Pandolfi 1997; Pandolfi andGreenstein 1997a,b). However, although carbon-ate platforms and other sites of predominantlycarbonate accumulation constitute the geologi-

cal stereotype for low-latitude settings, they oc-cupy only about half of shallow tropical shelfarea, for example occupying ;45% in AmericanTropics (Best 1998a). Siliciclastic and mixed sil-iciclastic-carbonate sediments characterize theremaining shelf area. Tropical siliciclastics maydiffer from temperate-latitude analogues in to-tal input and seasonality of organic matter; ratesand seasonality in bioturbation and, thus, sed-iment aeration; types and quantities of dis-solved ions (tropical subaerial weathering); andthe composition, rate, and timing of skeletal car-bonate input (e.g., possible latitudinal differenc-es in shell robustness, body size, productivity).However, actualistic taphonomic data are lack-ing and even sedimentological information islimited for such environments, despite their ob-vious importance today and in the geologicpast.

81TROPICAL TAPHOFACIES

FIGURE 1. Bocas del Toro embayment of Caribbean Panama, showing location of sampling stations and regionalgeomorphology. Polygons enclose sampling area of multiple grabs and dredges per station.

There are many reasons to suspect funda-mental differences in the taphonomy of mac-roinvertebrate hard parts within tropical sili-ciclastic and carbonate sediments. Siliciclasticsediments are typically more abrasive, havemore turbid overlying waters, and, owing tothe presence of iron and other metals, mayhave less-saturated pore-water chemistriesthan in pure carbonate sediments (e.g., Aller1982; Walter and Burton 1990; Green et al.1996; but see Aller et al. 1996 for evidence ofoversaturation due to iron). Siliciclastic sedi-ments also typically support different benthiccommunities of shell producers and modifi-ers. The net effect of these physical, chemical,and biological factors on shell preservation intropical siliciclastics is unknown. Some ofthese environmental attributes might conveyadvantages to shell preservation in siliciclas-tics (e.g., less light penetration of water col-umn to support boring algae), whereas others

are expected to be disadvantageous. For ex-ample, oxidation of pyrite, which lowers pore-water pH and promotes shell-carbonate dis-solution in temperate latitudes (Aller 1982),could operate at even higher rates in the Trop-ics.

To examine these issues, we are investigat-ing siliciclastic and carbonate subtidal sea-floors along the Caribbean coast of Panamathrough taphonomic evaluation of death as-semblages, experimentation with arrays oftethered shells, sediment analysis, and pore-water geochemistry (Best and Kidwell 1995,1996; Best 1996a,b, 1998c; Ku and Walter 1998;Best et al. 1999). Here, we report results on bi-valve death assemblages from the Bocas delToro area of Panama (98N), a region of exten-sive, primarily siliciclastic sedimentation in arelatively deep (37 m), geomorphically com-plex embayment (Fig. 1). Taphonomic analysisfocused on bivalves because of their abun-

82 MAIRI M. R. BEST AND SUSAN M. KIDWELL

dance in both reef and nonreef tropical envi-ronments. The Caribbean coast of Panamaprovides sedimentary and climatic conditionstypical of the American Neotropics, and itsmodern fauna and fossil record are of partic-ular interest to studies of marine biodiversityand its response to Neogene uplift of the Pan-ama Isthmus (Jackson et al. 1997).

To identify postmortem damage unambig-uously, our analysis focuses exclusively ondamage to shell interiors. Most previous ta-phofacies studies have summed damage toshell interiors and exteriors, some of whichmay be acquired during the life of the shellproducer.

Geologic Setting

The Bocas del Toro embayment lies on theCaribbean coast of Panama near the Costa Ri-can border (Fig. 1) and is situated within thelate Cenozoic back-arc basin of the andesiticvolcanic arc that forms much of the Panaman-ian Isthmus. Outcrops of Neogene marine sil-iciclastic sediments, derived from the arc andnow uplifted along a network of small faults(Coates et al. 1992; Seyfried et al. 1994), forma series of mangrove-fringed peninsulas andislands within the embayment. These faultsare probably also a major control on local ba-thymetry (Greb et al. 1996). Windward cliffedshorelines are local sources of siliciclastic sed-iment, supplementing that derived from themainland watershed, which reaches elevationsof .2200 m. Along the coast immediately eastand west of the Bocas embayment, the openCaribbean reef front is attached to the main-land and occupies the inner edge of a sea-ward-dipping shelf. Across the broad mouthof the Bocas embayment, the reef front is at-tached to bedrock islands in the west (IslasColon and Bastimentos) and, in the east, con-tinues as a series of carbonate sand keys onsubmerged horsts (Cayos Zapatillas) (Greb etal. 1996). A deep natural channel (Canal delTigre) cuts across the reef front between Cay-os Zapatillas and the mainland Penınsula Val-iente, which defines the eastern edge of theBocas embayment.

Greb et al. (1996) surveyed the sedimentol-ogy and macrobenthic ecology of the shallow-water (,12 m) pure-carbonate environments

that fringe islands in the western sub-basin(Bahıa Almirante; Fig. 1), including data onthe Cayos Zapatillas reef platform. Sedimen-tologic and ecological information for the east-ern siliciclastic-dominated portion of Bocasdel Toro is limited to ecological surveys of fo-raminiferans and corals (Havach and Collins1997; Guzman and Guevara 1998a,b).

Hydrographic information for this part ofthe Caribbean coast is summarized by Greb etal. (1996). The prevailing regime is microtidal(0.3 m normal, 0.5 m spring tide), and the arealies outside Caribbean hurricane tracks. Al-though most of the Bocas embayment is pro-tected by the outer line of bedrock islands andcarbonate keys, wind-driven swells from thenortheast typically sweep directly into Canaldel Tigre, breaking on the north coast of CayoAgua and the tip of Penınsula Valiente (NEprevailing wind; storms from northwest dur-ing rainy season [Sadler et al. 1987]). Theseswells, combined with tidal exchange of alarge volume of water in Laguna de Chiriquı(max. depth 37 m), create substantial currentsthrough Canal del Tigre, especially at its nar-rowest segment along the eastern edge ofCayo Agua. A longshore current also sweepsthe Bocas del Toro coastline from the north-west. Whereas a part of this current is appar-ently deflected into Bahıa Almirante in thenorthwest (max. depth 24 m) (Greb et al.1996), it is probably less important in thesoutheastern portion of the embayment whereour work has concentrated.

Precipitation is strongly seasonal, rangingfrom virtually no rain in February and Marchto as much as 566 mm in November; the an-nual average is 3305 mm (Greb et al. 1996).Runoff during the fall rainy season, particu-larly from the Rios Guarumo, Guariviara, andCricamola, delivers a large but as yet unquan-tified volume of fine siliciclastic sedimentsfrom the mainland. Sediment plumes off theserivers can extend .30 km and persist for sev-eral days after major rains (Best and Kidwellpersonal observations 1994, 1995, 1997). Visi-bility measured with a Secchi disk during adry week in November 1994 was quite high(9–16 m; our data), with highest values on thewindward side of Cayo Agua where oceanicwaters enter the embayment via Canal del Ti-


gre. Water temperatures are also seasonal,with the mean shifting from 28.08C during theMay–December wet season to 26.68C duringthe dry season off the Bocas coast (Sadler et al.1987). In November 1994, we measured tem-peratures of 28.38C throughout the water col-umn in stations adjacent to Canal del Tigre(12–29 meters) and found warmer surface wa-ter in protected backwater stations (30.08Csurface, 0.6–1.1 degrees higher than bottomwaters in 12–20 m depths). The exceptionswere an extreme backwater station (station 12in Fig. 1) where the entire water column hadbeen heated to 308C, and immediately off RioUyama (station 8) where the temperaturestratification was inverted, probably becauseof freshwater outflow. These measurementsare consistent with a larger hydrographic sur-vey conducted in 1997 by L. D’Croz (unpub-lished.).

Methods

The eastern embayment and adjacent shelfof Bocas del Toro were sampled from the RVUracca (Smithsonian Tropical Research Insti-tute) in November 1994 to explore sedimenttypes and molluscan assemblages at a rangeof depths and oceanographic facings. Sam-pling focused on areas .10 m both to take ad-vantage of the vessel and to examine environ-ments with stratigraphically significant rec-ords. Several subareas within the embaymentwere reoccupied for supplementary environ-mental information in June 1997, and at thattime we also sampled two cross-shelf tran-sects of the (siliciclastic) open shelf. This paperfocuses on taphonomic data from the nine sta-tions established within the embayment in1994 (nine nonconsecutively numbered poly-gons in Fig. 1).

We collected at least three dredge samplesand three Petersen grab samples at each sta-tion. Each was subsampled for sediment anal-ysis, and the remainder was wet-sievedthrough a 2-mm mesh. Live specimens wereremoved and preserved in alcohol. Death as-semblages were later wet-sieved with freshwater through 8-mm and 2-mm mesh. Afterair-drying, we determined the total volume of2–8-mm and .8-mm grains and the percent-age of bivalves, by volume and weight, within

the .8-mm fraction. Grain size (sedigraphand sieving), percent nitrogen, organic car-bon, and carbonate (total weight percent ni-trogen and carbon by elemental analysis, car-bonate by digestion, following Verardo et al.1990) were determined for bulk sedimentsamples by the Department of Geology,Northern Illinois University. For patch reefs,we tallied information from both loose bivalveshells within sediment and those attached toor wedged within dead coral rubble, whichdominated these remotely collected samples;this contrasts with previous workers (Parsons1993; Dent 1995), who sampled using SCUBAand examined shells exclusively from pocketsof sediment within reef frameworks.

In preliminary subsamples, we found thatsome kinds of taphonomic damage in deathassemblages were sensitive to shell size-class,as recognized previously in temperate deathassemblages (e.g., Boekschoten 1966; Bosence1979; but contra Davies et al. 1989; Staff andPowell 1990a,b), and thus we targeted thecoarsest fraction (.8 mm) where levels ofdamage were thought to be greatest andwhere species identification would be moststraightforward. The vast majority of theshells fell within a range of 8–20 mm. To fa-cilitate organization of the database, and to al-low information on specimens to be verified oraugmented at a later date, we gave each bi-valve shell or shell fragment a unique identi-fying number.

Interior surfaces of all bivalve specimenswere examined by 103 microscope, and re-sults for six major categories of damage are re-ported here: disarticulation, fragmentation,edge-rounding, and, based exclusively onshell interior surfaces, encrustation, nonpred-atory boring, and fine-scale alteration of sur-face texture. For each category of damage,both the specific type and extent of damageper specimen were tallied (Table 1), followinga protocol similar to that of other workers(e.g., Davies et al. 1989; Kowalewski et al.1995). Results are presented as a damage pro-file (bar graph), summarizing the percentageof shells that display fragmentation (any de-gree), rounding of the commissure, presenceof encrusters or borers, and relatively severe


TA

BL

E1.

Dat

aty

pes

per

shel

l.

Eco

log

ical

info

rmat

ion

Gen

us/

spec

ies

Min

erol

og

yR

obu

stn

ess

Lif

ep

osi

tion

Att

ach

men

tFe

edin

gst

rate

gy

Acc

ord

ing

toa

taxo

nli

stB

ased

onty

pe

spec

imen

sar

agon

ite

calc

ite

nac

reou

sar

agon

ite

thin

(,0.

5m

m)

thic

k( .

0.5

mm

)in

fau

nal

sem

i-in

fau

nal

epif

aun

al

free

-liv

ing

byss

ate

cem

ente

rb

orer

filt

erfe

eder

mix

edfe

eder

chem

osy

nth

etic

dep

osit

feed

erTa

ph

onom

icin

form

atio

nF

rag

men

tati

onan

dd

isar

ticu

lati

onE

ncr

ust

atio

nB

orin

gE

dg

e-ro

un

din

gF

ine-

scal

esu

rfac

eal

tera

tion

Val

ves

arti

cula

ted

Sin

gle

wh

ole

valv

eL

arg

efr

agm

ent

( .20

%va

lve)

Smal

lfr

agm

ent

( ,20

%va

lve)

no

encr

ust

atio

nfi

ne

wor

mtu

bes

serp

uli

d/

spir

orbi

dag

glu

tin

ated

wor

mtu

be

bar

nac

lesh

eet

bryo

zoan

net

bryo

zoan

no

bor

ing

spon

ge

wor

mbi

valv

ero

otet

chin

gg

razi

ng

bryo

zoan

un

mod

ified

edg

esch

ipp

eded

ges

rou

nd

eded

ges

thin

ned

edg

es

pri

stin

ed

ull

chal

kyp

itte

dch

alky

and

pit

ted

erod

ed

biva

lve

verm

etid

spon

ge

cora

lH

omot

rem

ad

isk

fora

mag

glu

tin

ated

fora

mca

lcar

eou

sal

gae

brac

hio

pod


fine-scale damage (either chalky and pitted, oreroded).

In our protocol, fine-scale alteration refersto various degrees of degradation of originalluster; it may be due to any combination of mi-croboring and other microbioerosion, partialdissolution of mineral crystallites, macerationof shell organic matrix, and physical abrasion,which require a scanning electron microscope(SEM) to distinguish (see Cutler 1995; Bestand Kidwell 1996; Best 1998c; Best et al. 1999).This suite of damage types is grouped underthe general category of ‘‘dissolution’’ or ‘‘cor-rosion’’ in many other studies. Because visiblechalkiness can be produced by mantle anaero-biosis during the life of a bivalve (Crenshaw1980), we tallied fine-scale alteration usingonly the portion of the shell interior outsidethe pallial line, an area not prone to the arti-fact of in vivo dissolution and characterized ina given species by a single microstructuraltype. This latter point is important, becausestyles and rates of fine-scale alteration are al-most certainly a function of shell microstruc-ture (i.e., the size, shape, mineralogy, andpacking of mineral crystallites in organic ma-trix) (see Glover and Kidwell 1993) rather thanof environmental conditions alone. Our sub-division of shell interiors into microstructur-ally and physiologically distinct anatomicalsectors for taphonomic characterization con-trasts with division into geometric sectors re-lated to life position (e.g., Davies et al. 1989),which is most relevant for tallying damage toshell exteriors.

Sedimentary Environments

Our benthic surveys in the eastern two-thirds of the Bocas embayment indicate a mo-saic of largely siliciclastic sediments withpoint sources of carbonate production (Table2, Fig. 2). The geomorphic and hydrographiccomplexity of the embayment creates a mix-ture of low- and high-energy shallow-waterenvironments. Windward shorelines of bed-rock islands (including the western edges ofPenınsula Valiente) are characterized by pock-et beaches of lithic sand interspersed with oc-casional small patch reefs and isolated corals.These foreshore and shoreface sands gradeinto muddy sand and sandy mud at 20–30 m

depth. The largest and most extensive patchreefs rise up from shallow muddy seafloors onthe mangrove-rimmed lee sides of bedrock is-lands and in the muddy leeside of Cayos Za-patillas. These reefs include both live and rel-ict material, are 5–10 m diameter, and have ;5m relief. They are the primary sources of car-bonate sediment, along with small meadowsof occasionally coral-rich Halimeda gravellysand (e.g., rimming the constricted portion ofCanal del Tigre east of Cayo Agua) and car-bonate sand and seagrass beds (leeside ofCayos Zapatillos). The embayment is domi-nated by green, gray, and brown clayey silici-clastic muds, which surround patch reefs andalso form a low-relief plain in the innermostembayment (Laguna de Chiriquı). Muds in-crease in carbonate-mud content toward theopen sea (to ;50%), and contain only a fewpercent skeletal debris .2 mm.

Five distinct sedimentary environmentscharacterize waters .10 m (Table 2, Fig. 2).Three of these include hard substrata (patchreef, Halimeda gravelly sand, mud amongpatch reefs), and the other two present exclu-sively soft sediments (sandy mud, mud).Hard-substrata environments do not show asimple correlation with water energy—thelargest patch reefs are in relatively protectedlee areas, whereas Halimeda meadows are incurrent-swept areas—but all are in the shal-lowest-water portions of the embayment andare not found along the mainland shore. En-vironments are arranged in order of decreas-ing carbonate content and grain size in Table2 and other figures.

Abundance and Taxonomic Composition ofBivalve Death Assemblages

Skeletal material .2 mm constitutes only afew percent by volume of sediments in the em-bayment (Table 2). However, virtually all ofthis material is taxonomically identifiable, andin siliciclastic sediments, bivalve shells andshell fragments constitute the majority. Con-sequently, relatively small (8–10 liters) sam-ples of sediment yielded reasonable numbersof specimens .8 mm for taphonomic evalua-tion (2–15 specimens per liter of sediment)(Table 3). This abundance matches or exceedsthat in other subtidal tropical sediments. For


TA

BL

E2.

Sed

imen

tary

env

iron

men

tch

arac

teri

zati

on.

En

vir

onm

ent

Wat

erd

epth

Sed

imen

tco

lor

and

gra

insi

zeSe

dim

ent

com

po

siti

onC

har

acte

riza

tion

ofb

enth

os

Pat

chre

ef(s

tati

ons

1an

d4)

14–1

6m

(to

top

sof

reef

s)sm

all

cora

l-d

omin

ated

reef

s,5–

10m

dia

met

er,w

ith

5m

reli

efab

ove

surr

oun

din

gm

ud

dy

sea-

floo

r

carb

onat

eh

ard

gro

un

dm

ixtu

reof

live

and

dea

dco

rals

,pre

dom

i-n

antl

yA

gari

cia

(wit

hS

ider

astr

ea,M

onta

s-tr

ea,

Pori

tes)

and

oth

erla

rgel

yep

ifau

nal

taxa

(rh

yn

con

elli

dbr

ach

iop

ods,

bryo

zoan

s,ca

lcar

eou

sal

gae

,ver

met

ids,

rop

ysp

ong

es,

Hom

otre

ma,

Spo

ndyl

us,

and

Cha

ma)

Hal

imed

ag

rave

lly

san

d(s

tati

on10

)12

–13

mta

nfi

ne–

med

ium

–coa

rse

carb

on-

ate

san

d(m

ostl

yH

alim

eda

de-

bris

,som

eco

rall

ine

alg

aean

dm

ollu

skfr

agm

ents

),tr

ace

mu

dan

dsp

ong

esp

icu

les;

smal

lp

atch

reef

sof

cora

lsan

dre

dco

rall

ine

alg

ae

no

quan

tita

tive

dat

afo

rlo

ose

sed

imen

t,bu

tk

50%

carb

onat

e;sk

el-

etal

deb

ris

.2

mm

53.

5%of

bulk

sam

ple

byvo

lum

e

abu

nd

ant

clu

mp

sof

live

Hal

imed

a,fl

esh

yg

reen

alg

ae,a

nd

bot

hcr

ust

ose

and

ram

ose

red

alg

aew

ith

abu

nd

ant

san

dto

gra

vel-

size

dca

lcar

eou

sd

ebri

s,in

clu

din

gsh

ells

ofst

rom

bid

san

dve

ner

ids;

her

mit

crab

s;ar

eain

clu

des

smal

lre

efs

ofp

rim

aril

yp

late

yco

rals

,esp

ecia

lly

Aga

rici

a,cr

ust

ose

red

al-

gae

,an

dot

her

atta

ched

epif

aun

a(S

pond

y-lu

s,B

arba

tia,

and

Ano

mia

biva

lves

,bry

ozo-

ans,

rhy

nco

nel

lid

brac

hio

pod

s)M

ud

amon

gp

atch

reef

s(s

tati

ons

1,5,

and

8)14

–20

mli

ght–

med

ium

gre

ento

gra

yis

h-

brow

nm

ud

,4–6

%.

63m

(mos

tly

shel

lg

rave

l).S

amp

les

vary

inp

roxi

mit

yto

pat

chre

efs

and

inth

eir

con

ten

tof

coar

sesk

elet

ald

ebri

s

mu

dsi

ze-f

ract

ion

con

-ta

ins

18–5

2%ca

rbon

-at

e;sk

elet

ald

ebri

s.

2m

m5

3–6%

ofbu

lksa

mp

leby

volu

me

sulf

uro

us

odor

(sta

tion

5),

1–2%

org

anic

car-

bon

,0.1

7–0.

27%

nit

ro-

gen

spar

seto

abu

nd

ant

mol

lusc

ansh

ell

mat

eria

l,m

ang

rove

deb

ris;

live

luci

nid

and

ven

erid

biva

lves

,hol

oth

uri

ans,

pri

apu

lid

s;am

oun

tof

reef

-der

ived

cora

ld

ebri

sva

ries

wit

hp

roxi

mit

yto

pat

chre

efs

(see

pat

ch-r

eef

fa-

cies

des

crip

tion

abov

e)

San

dy

mu

d(s

tati

ons

7an

d11

)24

–29

mg

reen

ish

-gra

ysa

nd

ym

ud

,10–

20%

.63

m(m

ostl

yd

ark

fin

e–m

edi-

um

mafi

can

dli

thic

gra

ins)

mu

dsi

ze-f

ract

ion

con

-ta

ins

19–3

5%ca

rbon

-at

e;sk

elet

ald

ebri

s.

2m

m5

0.3–

1.0%

ofbu

lksa

mp

leby

volu

me

sulf

uro

us

odor

(sta

tion

11),

0.8–

3.6%

org

anic

carb

on,0

.10–

0.35

%n

i-tr

ogen

coar

ser-

gra

ined

site

(sta

tion

7)h

asab

un

dan

tla

rge

mil

ioli

dfo

ram

s,h

erm

itcr

abs,

brac

h-

yu

rid

s,st

omat

opod

s,li

vesa

nd

dol

lars

,m

ang

rove

deb

ris,

and

mol

lusk

shel

ls;fi

n-

er-g

rain

edsi

te(s

tati

on11

)in

clu

des

dea

dsp

atan

goid

s,st

rom

bid

san

dot

her

gas

tro-

pod

s,m

ang

rove

deb

ris,

brac

hy

uri

ds,

sto-

mat

opod

s,bi

valv

esh

ells

Mu

d(s

tati

ons

12an

d13

)13

mg

reen

ish

-gra

yto

brow

nso

ftcl

ay-

eym

ud

mu

dsi

ze-f

ract

ion

con

-ta

ins

20–3

0%ca

rbon

-at

e;sk

elet

ald

ebri

s.

2m

m5

0.4–

1.0%

ofbu

lksa

mp

leby

volu

me

sulf

uro

us

odor

(sta

tion

13),

2.8–

5.6%

org

anic

carb

on,u

pto

0.35

%n

i-tr

ogen

live

card

iid

,tel

lin

id,c

orbu

lid

,ven

erid

,an

dp

ecti

nid

biva

lves

and

shri

mp

;mu

cus

mat

sw

ith

shel

ld

ebri

s,m

ang

rove

deb

ris,

spa-

tan

goid

deb

ris,

brac

hy

uri

ds,

mol

lusc

ansh

ells

;in

sam

ple

sfr

omsi

tes

nea

rer

tom

ang

rove

shor

elin

e(s

tati

on13

),al

sod

ead

Spo

ndyl

us,

Cha

ma,

ram

ose

cora

ls,r

egu

lar

ech

inoi

ds


FIGURE 2. Distribution of sedimentary environments in Bocas del Toro study area showing dominance of silici-clastic sediments with point sources of carbonate. Subtidal areas in white are outside study; no data available.

TABLE 3. Percent abundance of bivalve life styles; d 5 dredge sample, g 5 grab sample.Deposit feeders: several species of NuculanaChemosynthetic feeders: lucinids Diplodonta, Codakia, Lucina, and DivaricellaMixed feeders: many species and several genera of TellinaceaSuspension feeders (infaunal): venerids, cardiids, corbulids, Glycymeris, Tagelus, Cardiomya, and PandoraFree-living epifauna: Anadara brasiliana and all pectinids except for one byssate ChlamysByssate epifauna: Lima, Chlamys, arcids, anadarids, Barbatia, Pteria, Malvimalleus, Pinna, Brachiodontes, and Gastro-

chaenaCementers and borers: several chamid genera, Plicatula, Spondylus, Ostrea, Spengleria, and Lithophaga

Environment

Samplesize(n)

No. oftaxa

Infauna

Depositfeeders

Chemo-syntheticfeeders

Mixedfeeders

Suspensionfeeders

Epifauna

Free-livingepifauna

Byssateepifauna

and nestlersCementersand borers

Patch reef1-d3.0-01-d3.1-04-d19-0

Halimeda gravelly sand10-g44-0

483849

101

81114

26

———

—

———

—

—86

13

—3

22

20

25

14

3

312624

16

675833

4510-d59-0

Mud among patch reefs1-d3.3-08-d36-45-d23-4

94

343100111

27

472727

—

——4

3

118

5

22

212235

43

384141

2

672

11

824

18

2610

7Sandy mud

7-d35-1.17-d25-1.211-g49-07-g31-011-g47-0

7378981744

121410

410

3733—71—

513——

333366—61

1113242436

1118

662

—————

31

———

Mud12-d63-213-d72-112-g61-0

109103147

91313

———

—15

1

266232

732065

—3—

——1

1—1


FIGURE 3. Cumulative sampling curve of taxa vs. in-dividuals for samples pooled by environment. Eachcurve is the average of 200 bootstraps, with one stan-dard deviation error indicated.

example, in the San Blas Archipelago of Pan-ama, one liter of sediment in both carbonateand siliciclastic environments yields 1–18identifiable bivalve specimens .8-mm (Best1996a,b). Shell densities in pure-carbonatesediments were 0–30 specimens/liter in Ma-dang, Papua New Guinea (.8 mm, bivalvesonly) (Best and Pandolfi 1992 and unpub-lished data), 1–25/liter in the Virgin Islands(.5 mm, combined bivalves and gastropods)(Parsons 1993), and an average of 38/liter inthe Florida Keys (.2 mm, combined bivalvesand gastropods) (Dent 1995).

The total number of bivalve species persample ranged from 4 (n 5 17 specimens) to47 (n 5 343) (Table 3, summarizing all pro-cessed samples). Pooled total diversity per en-vironment ranged from 18 (mud) to 66 (mudamong patch reefs) (Fig. 3). Cumulative sam-pling curves, pooled by environment, indicatethat species richness per environment beginsto level off (using maximum convexity oncurve; see Heck et al. 1975; Sanders 1968) at;8 for muds (n ; 30 individuals), 14 for sandymuds (n ; 60 individuals), 19 for patch reefs(n ; 70 individuals), 29 for Halimeda gravellysands (n ; 90 individuals), and 36 for mudsamong patch reefs (n ; 100 individuals) (Fig.3). These data suggest that bivalve diversitiesin soft siliciclastic seafloors (mud, sandy mud)are significantly lower than those in carbon-ate-rich and hard substrata (Halimeda gravellysand, patch reef). Moreover, a reasonable di-versity of bivalve taxa is captured in quitemodest samples. These estimates are almost

certainly conservative because the analysiswas limited to the coarse fraction (.8 mm),the regional fauna is underdescribed, andsome fragments were difficult to assign taxo-nomically to a species. Lumping of species ismost significant among pectinids, oysters,and small infaunal species with few distin-guishing features, but the large majority ofspecimens in samples could be assigned atleast to the genus level, and virtually all couldbe assigned to ecological groups (Table 3).

These taxonomic richnesses are similar tosingle samples from other tropical subtidalenvironments, even those including finer sizefractions of the death assemblage and report-ed as total richness rather than from a cumu-lative sampling curve. In San Blas Archipela-go, total per-sample species richness rangedfrom 6 to 19 in siliciclastics (sample sizes of 40to 89 individuals) and from 9 to 16 in carbon-ates (sample sizes of 33 to 44; Best 1996a,b,1998c, and unpublished). In pure-carbonateenvironments, pooled bivalve richness rangedfrom 15 to 28 in Papua New Guinea (samplesizes of 73 to 158; Best and Pandolfi 1992, un-published data), and per-sample total mollus-can richness ranged from 8 to 39 in the VirginIslands (sample sizes of 45 to 116; Parsons1993) and from 17 to 75 in Yucatan (samplesizes of 39 to 3634, 3-mm mesh; Ekdale 1972,1977).

Spatial Fidelity of Death Assemblages

In the study area, live mollusks (both bi-valves and gastropods) were extremely rare atall stations, and so an evaluation of spatial fi-delity via a live-dead comparison (e.g., John-son 1965; Cadee 1968; Warme 1971; Miller1988; Feige and Fursich 1991) was not feasible:in washing 72 samples from the Bocas embay-ment and adjacent open shelf totaling ;2600l sediment, we encountered only a few dozenlive individuals .2 mm. Such low numbersare not unusual for single censuses of livemollusks in live-dead studies; many replicatesamples in space and over time are generallynecessary to build a rigorous database of mol-luscan diversity from living material alone(see Kidwell and Bosence 1991). Death assem-blages can also be compared with other sourc-es of information on the habitat preferences of


FIGURE 4. Taphonomic signatures of bivalve death as-semblages based on percent frequency of shells with en-crusters, macroscopic borers (nonpredatory), chalkyand pitted or eroded shell outside the pallial line (‘‘sur-face degraded’’), rounded shell commissures (‘‘roundedges’’), and fragmentation. A, Values per sample.Curve indicates water depth in meters. B, Values forsamples pooled by environment; error bars are 95% con-fidence intervals.

local fauna for the purpose of a live-dead com-parison (e.g., Henderson and Frey 1986; Da-vies et al. 1989; Staff and Powell 1990a), butthe Bocas fauna is insufficiently known to dothis at the species level.

However, the spatial fidelity of death assem-blages can also be assessed by evaluating tax-onomic and ecologic compositions in light ofthe environment (e.g., Zuschin and Hoheneg-ger 1998). In Bocas, the proportion of ecolog-ical groups in the bivalve death assemblagesis significantly different at the 0.05 level foreach of the five environments (Kolmogorov-Smirnov test of frequency; Sokal and Rohlf1995) (Table 3), indicating that postmortemtransport has not been so severe as to homog-enize surviving death assemblages. Moreover,the relative proportions of these ecologicalgroups are consistent with the physical char-acteristics of the collection station (Tables 2and 3), suggesting the preservation of a strongsignal from the original environmental distri-bution of these taxa during life. For example,the proportion of specimens from epifaunalbivalves requiring hard substrata (cementers,borers, byssate or nestling bivalves) is greatestin patch reefs, moderate in Halimeda gravellysands (where there are both small patch reefsand live Halimeda for attachment), highly var-iable but generally lower in muds amongpatch reefs (depending on reef proximity),and virtually nil in sandy muds and muds. In-faunal bivalves are present in all but one sam-ple (1-d3.0, patch reef), but numbers of indi-viduals and diversity of infaunal feedingtypes are greatest in organic-rich sandy mudsand muds.

Facies-Level Differences in PostmortemDamage

Disarticulation.—To account for possibledisarticulation during collection, processing,and shipping, matching valves were countedas articulated, which, if anything, would pro-duce an overestimate of articulation. Nonethe-less, articulated valves are generally rare. Insamples of soft substrata (sandy mud, mud),only 1–12% of specimens are articulated andthese are limited to species with taxodontdentition (most commonly Nuculana). In set-tings that include hard substrata (patch reef,

Halimeda gravelly sand, mud among patchreefs), articulation is more common (e.g., 20–50% per sample from patch reefs), but is lim-ited to specimens that are physically wedgedwithin interstices of coral rubble (thus makingthe separation of valves difficult) and to spe-cies that have very thick ligaments or denti-tions resistant to disarticulation (chamids,spondylids).

Fragmentation.—Levels of fragmentation areuniformly high across all environments (39–53% of bivalve specimens per sample; Fig. 4).The only exception is patch reefs (5–22% persample), where low levels are attributable tothe high abundance of cementing species (Ta-ble 3), which are represented primarily by in-tact valves still tightly attached to pieces ofcoral rubble.

Because shells can easily be broken duringcollection and processing, fragmentation (likedisarticulation) is a problematic taphonomicvariable. However, initial inspection of sam-ples on deck and during gentle washing in-dicated that high levels of fragmentation were


FIGURE 5. Percent frequency of fragmentation in deathassemblages is not correlated with the frequency of thinshells or with environment.

FIGURE 6. A, Percent frequency of shells that are en-crusted and relative abundance of encrusters. B, Percentfrequency of shells that are macroscopically bored andrelative abundance of borers. Both graphs are pooled byenvironment with error bars showing 95% confidenceintervals.

not artifacts of sampling gear—grab anddredge samples had similar appearances—orof handling. Thus, although some additionalfragmentation may have occurred duringshipping, the qualitative pattern is probablyrobust (as argued by others for other study ar-eas, e.g., Davies et al. 1989).

Fragmentation is not significantly correlat-ed with water depth (12–29 m; Fig. 4A), rela-tive water energy (inferred from sedimentgrain size and setting), or shell thickness (Fig.5). Thus, although physical agents may con-tribute to fragmentation, they are not the pri-mary or sole agents determining environmen-tal differences. Predators and other organismsare the most likely agents of fragmentation inthe low-energy muds, a conclusion supportedby distinctive break patterns on shells (narrowinvaginated breaks and chipped edges createdby crabs and whelks; see Schafer 1972; Cateand Evans 1994) and consistent with the ob-served abundance of shell-crushing rays andtheir feeding pits (Best personal observationon SCUBA).

Encrustation.—The percentage of bivalveshell interiors that are encrusted decreaseswith decreasing availability of hard substrata,from ;45% in patch reefs and Halimeda grav-elly sands (no significant difference betweenthese two environments, which both have ex-tensive hard-substrata) to zero in muds (Fig.4).

Rank orders of the different encrusting taxaare quite similar among environments, withcalcareous tube-forming worms most abun-dant in all (serpulids and spirorbids), but sig-

nificant quantitative differences do exist (Fig.6A). Patch reefs and Halimeda gravelly sandshave the highest diversities: calcareous tube-forming worms and bryozoans are most com-mon, followed by calcareous algae; foraminif-era, bivalves (chamids, ostreids), corals, andvermetids occur with lower frequency. Inmuds among patch reefs, worms overwhelm-ingly dominate other encrusting taxa; smallnumbers of bryozoans, calcareous algae, fo-raminifera, bivalves (chamids, ostreids), andcorals are mostly associated with shells thatwere probably derived from nearby patchreefs (i.e., attached epifaunal bivalves). In allthree of these hard-substrata environments,monospecific encrustations usually consist ofrunner bryozoans, suggesting that they areearly colonists. In contrast, worms are virtu-ally the only encrusters in bivalve death as-semblages from exclusively soft substrata;barnacles, brachiopods, agglutinated wormtubes, and sponges (the latter two having poorpreservation potential) occur only in traceamounts.

Within the study area, there is no correla-


FIGURE 7. Association of macroscopic boring and en-crustation on individual shells. Boring rarely occurswithout encrustation, suggesting that encrusters are thefirst colonists of shell interiors.

tion between the level of encrustation and wa-ter depth (Fig. 4A), perhaps because the rangewas relatively narrow (12–29 m) and becausesome shallow-water muds have both low en-ergy and relatively turbid overlying waters.The clearest environmental correlate of en-crustation is the presence of hard substrata onthe seafloor. The extent of hard substrata with-in an environment is difficult to quantifywithout detailed mapping using SCUBA, butnonetheless has a strong, essentially dichoto-mous effect: environments that include hardsubstrata all have assemblages with at least20% encrustation levels, whereas exclusivelysoft substrata are characterized by encrusta-tion levels of only a few percent, even whentested at the individual sample level (Fig. 4A).Water turbidity may be an additional factor, aslight-dependent encrusters (calcareous algae,Foraminifera, corals) and those most sensitiveto sediment smothering (bryozoans) occurprimarily in nonmuddy stations.

Nonpredatory Boring.—Boring of shell inte-riors (examined at 103) also covaries with thepresence of hard substrata, ranging from 26%in patch reefs to zero in muds (Fig. 4B). Levelsof boring are significantly higher in patchreefs than in the other two hard-substrate-bearing environments (Halimeda gravellysand, mud among patch reefs), which are notsignificantly different from each other. Thetwo soft-sediment environments (sandy mud,mud) have significantly lower boring levelsbut are not significantly different from eachother.

Clionid sponges are the most frequent bor-ers in all environments, and environmentswith hard substrata have the highest diversi-ties of borers (Fig. 6B). Clionid sponges arefollowed or matched in frequency by spionidworms, followed by lithophagid bivalves andbryozoans. (Microborers are considered un-der fine-scale surface alteration.)

As with encrustation, there is no correlationbetween the frequency of bored shells and wa-ter depth within the study area (Fig. 4A), andthe clearest environmental links are with thepresence of hard substrata. All environmentsthat include hard substrata have bivalve deathassemblages with levels of boring 10% orhigher, whereas environments that are exclu-

sively soft have levels of only a few percent,even when tested at the level of individualsamples. An additional environmental factoris indicated at station 8 (mud among patchreefs, sample 8-d36; Fig. 4A), where an anom-alously high incidence of worm borings coin-cides with elevated nutrients from Rio Uyama(confirmed by L. D’Croz personal communi-cation 1997). The fact that such overwhelming-ly high levels of boring worms were not seenelsewhere suggests that, away from the im-mediate influence of a river, nutrient levelsvary little throughout the lagoon (for an ex-ample of this in oligotrophic reefs see Kiene1997).

Except at nutrient-rich station 8 (describedabove), boring levels are generally lower thanencrustation levels (compare within Fig. 4B).However, the two types of damage are closelycorrelated, and in fact individual bivalve in-teriors are rarely bored unless they are alsoencrusted (Fig. 7). This suggests that encrus-ters are generally the first colonists of deadshells in our area, at least in stations of com-parable nutrient regimes, and thus that shellsthat are merely encrusted might have been ex-posed at the sediment/water interface forshorter periods of time than those that areboth encrusted and bored. We are presentlytesting this pattern in a series of experimentalarrays in the Bocas embayment and in the SanBlas Archipelago, Panama (Best 1998c).

Edge-Rounding.—Shells with rounded com-missures (not rounding of the broken edges ofshell fragments) are most frequent in environ-ments with hard substrata, ranging from 32%


in Halimeda gravelly sand to 5% in mud (Fig.4). There is no significant difference in levelsamong the hard-substrata environments, andonly a marginal difference between the twoexclusively soft-sediment environments.

Edge-rounding does not correlate with ei-ther water depth (Fig. 4A) or shell thickness.SEM examination (in progress) suggests algalor other micro-bioerosion and grazing of shelledges, perhaps by fish or echinoids, judgingfrom ‘‘nibble’’ marks (Best and Kidwell 1996).Edge-rounding thus appears to be largely bi-ological in origin.

Fine-Scale Alteration.—Shells whose interiorsurfaces outside the pallial line have lost theiroriginal luster and present a chalky, pitted, oreroded appearance (103 examination) aremost common in environments with hardsubstrata (grand means 44–52%, reaching 71%in one sample) and lowest in exclusively soft-sediment environments (grand means 7–9%)(Fig. 4).

Although many causes are possible, initialSEM examination suggests that chalky andpitted textures were generated both by algaland fungal microboring and by microbialmaceration of organic matrix, indicated byloosened crystallites, rather than by chemicaldissolution (Best and Kidwell 1996; Best et al.1999). Alteration can be quite high even inmuddy, organic-rich sediments (e.g., ;20% insome individual samples from sandy muds;Fig. 4A), consistent with microbial attack,which is not impeded by shallow burial ofshells (see Simon and Poulicek 1990).

Total Damage Profile.—As expected fromtrends in individual variables, the primarydistinction among overall damage profiles(Fig. 4B) is between environments that includehard substrata (patch reef, Halimeda gravellysand, mud among patch reefs) and those com-posed exclusively of soft sediments (sandymud, mud). With the exception of fragmen-tation, which is high in all environments ex-cept patch reefs, death assemblages from ex-clusively soft substrata are virtually pristine,with significantly lower levels of all kinds oftaphonomic alteration (,10% specimens al-tered, and commonly K10% within individ-ual samples). In contrast, environments con-taining hard substrata show high frequencies

of damage (.10% specimens altered, and gen-erally .25%), and higher diversities of infest-ing taxa. Replicate samples from individualenvironments (Fig. 4A) yield consistent levelsof damage for exclusively soft-sediment en-vironments and heterogeneous results fromhard-substrata environments. This is not un-expected given that the soft-sediment envi-ronments are far more homogeneous physi-cally than the other environments. This dif-ference in heterogeneity of taphonomic sig-nature is being explored using rarefactiontechniques (Best 1998b, unpublished).

Pair-wise comparison of environmentswithin each of these two sets of environmentsreveals second-order differences in damageprofiles (Fig. 4B). For example, muds havemarginally less boring, encrustation, andedge-rounding than do sandy muds. Halimedagravelly sand assemblages have marginallymore edge-rounding and fine-scale surfacedamage, and less boring, than do assemblagesfrom patch reefs, and mud among patch reefsdiffers from patch reefs only in having fewerencrusted shells. While these differences aresufficient to change the rank ordering of var-iables among these environments, the signifi-cances of these differences are marginal, andit is unwise to overinterpret them in light ofthe heterogeneity of the individual samples.The fundamental difference is between hard-and soft-substrata environments.

Discussion

Differences in Methods.—Quantitative com-parison of trends in Bocas death assemblageswith other studies is not straightforward.Studies diverge considerably in examining allmollusks, bivalves only, all species, or only afew target taxa, and also diverge in the sizefraction, set of taphonomic variables, andmethod of scoring damage per specimen (Ta-ble 4). The robustness of taphonomic signa-tures to such differences in approach has nev-er been fully tested, although individual stud-ies have evaluated some aspects (e.g., effect ofsize fraction and of whole versus fragmentedshells [Davies et al. 1989; Staff and Powell1990b]; target taxa versus pooled data [Dent1995, 1996]).

Our method diverges most strongly from


most previous studies by focusing exclusivelyon shell interiors, so that postmortem damagecould be quantified unambiguously. Thisshould not affect the comparability of our re-sults for some variables, namely disarticula-tion, fragmentation, and edge-rounding. Forother variables, we expect our observed fre-quencies will be equivalent to or lower thanthe frequencies we would have obtained bysumming damage to both surfaces, makingthe Bocas dataset a conservative test of trendsin postmortem damage. Other studies areconsistent with this expectation. For example,considering only infaunal bivalves, whichshould be least sensitive to this methodologi-cal difference, Kowalewski et al. (1995) foundno differences in damage to shell interiors andexteriors in Cholla Bay and San Felipe inter-tidal assemblages (Gulf of California), but tal-lied a larger number of pristine interior sur-faces than exterior surfaces in North Sea in-tertidal assemblages. In Texas subtidal assem-blages dominated by infauna, Staff and Powell(1990b) found no significant interior-exteriordifferences on the inner shelf, but Davies et al.(1989) did find greater abrasion and ‘‘disso-lution’’ (5 fine-scale alteration of diverse ori-gins) on exterior shell surfaces in a nearby mi-crotidal inlet. This differential damage mayindicate selective attack of shell exteriors bytaphonomic processes, but it may also haveoccurred during life, especially among infau-nal species with little or no periostracum. Thisdoes not mean that tallying information fromshell exteriors is not useful for facies discrim-ination, only that it is not unambiguouslytaphonomic.

Because it would have doubled the numberof observations (already 41,931 for the 1553specimens examined), we did not gather in-formation on damage to exterior surfaces inBocas assemblages, and thus do not knowhow large an artifact this common methodmight introduce, particularly in epifauna-richassemblages (see concern of Parsons and Brett[1991] for reef-bearing study areas; see Bestand Kidwell this issue). However, because wefound highest damage in hard-substrata en-vironments, just as Parsons (1993) did in hernortheastern Caribbean study area using ex-terior as well as interior damage (see discus-

sion below), we suspect that combining exte-rior and interior shell damage may have onlya quantitative rather than qualitative effect onperceived taphonomic trends.

Spatial Fidelity.—Differences among envi-ronments in the taxonomic composition ofdeath assemblages, and the ecological consis-tency of assemblages with host sediments,suggest high spatial fidelity of molluscan re-mains at the facies-level within the Bocas delToro embayment: postmortem transport isrelatively minor both in distance and in quan-tities of shell moved out of life habitats. Thisresult is consistent with the pattern emergingfrom other enclosed and open-marine envi-ronments, both siliciclastic and carbonate, andappears to be robust to the method of analysisand even to the perils of mobile shell-ingest-ing predators and intermittent high energy(e.g., Cadee 1968, 1984; Warme et al. 1976; Bos-ence 1979; Fursich and Flessa 1987; Miller etal. 1992; Cate and Evans 1994; Flessa 1998; An-derson et al. 1998; see review by Kidwell andBosence [1991]).

High spatial fidelity is also implied by thesignificantly different damage profiles seenfor Bocas death assemblages from differentenvironments (see reasoning of Parsons andBrett [1991]). Habitat-specific shell damagehas been used elsewhere to recognize proba-ble transport directions of sediments (Pilkeyand Curran 1986) and the likely exotic originsof selected species in death assemblages (Da-vies et al. 1989), including hurricane-sweptback-reef settings (Miller et al. 1992 referenceto unpublished work; Perry 1996). In Austra-lia’s cyclone-affected Great Barrier Reef, Tud-hope (1989) found only a few meters’ trans-port of skeletal debris from patch reefs intoadjacent soft sediments. We observe the sameminimal off-reef transport around patch reefsin Bocas del Toro and San Blas Archipelago,Panama: shells from patch reefs are ecologi-cally and taphonomically distinctive, andtheir abundance in surrounding muds dimin-ishes to trace quantities within a few metersof the reef face (Best and Kidwell unpub-lished).

Biogenic Disarticulation and Fragmentation.—The high frequency of disarticulation andfragmentation in samples from quiet-water


soft sediments at Bocas suggests that thisdamage is largely biogenic rather than phys-ical in origin. This and distinctive break pat-terns also suggest that this damage is gener-ated largely by predators, rather than accruedduring postmortem time-averaging on theseafloor. Similarly high values, attributed topredators and in many instances contrary togradients in physical energy, have been re-ported from other low-energy settings (Cadee1994; see review by Cate and Evans 1994). Fei-ge and Fursich (1991) reasoned that fragmen-tation was largely biogenic in their intertidalstudy area because it decreased across a gra-dient of increasing physical energy and de-creasing sedimentation rate.

We attribute low fragmentation and disar-ticulation in Bocas patch reefs to the high pro-portion of species that are cemented to or oth-erwise firmly lodged in reef rubble in that lo-cal bivalve community. In contrast, Parsons(1993; Parsons and Brett 1991) found that frag-mentation levels were greatest in reef (andbeach) environments sampled in the Virgin Is-lands. We surmise that this difference resultsfrom her sampling areas of sand within reefframeworks, whereas we sampled the reefframework itself, and thus the damage pro-files are from two different subhabitats.

Genesis and Significance of Other Damage.—All other kinds of taphonomic damage in theBocas embayment—encrustation, postmortemboring, edge-rounding, and fine-scale surfacedegradation—are strictly postmortem in ori-gin and are greatest in assemblages frompatch reefs, Halimeda gravelly sands, andmuds that receive skeletal debris from imme-diately adjacent reefs (Fig. 4B). High frequen-cies of such damage may correlate with hardsubstrata for many reasons, which need not bemutually exclusive. These include (1) intrinsicecological factors, such as a larger proportionof epifaunal species, which, unlike infaunalspecies, necessarily experience a period ofpostmortem exposure on the seafloor beforeburial (see Best and Kidwell this issue), and(2) a host of extrinsic, covarying environmen-tal factors.

The most likely environmental factor isgreater exposure of death assemblages onpatch reefs and coral-algal meadow seafloors,

due both to exhumation cycles and to slowernet sediment accumulation rates. Water clarityprobably also contributes to higher rates of en-crustation in these settings. In contrast, wesuspect that sedimentation (and thereforeburial) rates in muddy environments are quitehigh in absolute terms, judging from burial ofexperimental arrays (e.g., up to 8 cm of sedi-ment accumulated in one year on experimen-tal shells deployed at the surface). Even in theabsence of high sedimentation, shell exposureat the sediment-water interface would not befavored, because these muddy siliciclastic-richsediments readily resuspend when touchedby divers (Best personal observation on SCU-BA, in both Bocas and San Blas habitats). Highorganic contents (Table 2) also indicate thatthese muds have quite ‘‘fluffy’’ sediment-wa-ter interfaces unfavorable to many infesters,even under quiet water conditions. These ben-thic conditions would pertain even when Sec-chi-disk visibility in overlying waters is high,as they were during our fieldwork.

Elevated levels of fine-scale alteration inhard-substrata environments may reflect sev-eral different factors, operating both at theseafloor and during shallow burial: (1) greaterexposure to light, fostering algal microboringand grazing-type bioerosion; (2) intermittent-ly higher energy, providing more opportuni-ties for physical abrasion; (3) stronger irriga-tion of pore waters in the coarser-grained sed-iments of these stations, fostering higher ratesof sulfate reduction, greater reoxidation of re-duced sulfide, and thus greater potential forcarbonate dissolution in addition to any acid-ification from aerobic decomposition (see Wal-ter and Burton 1990); (4) microbial attack oforganic matrix, which may proceed in all set-tings (Simon and Poulicek 1990; Glover andKidwell 1993), but which might be elevated incoarser-grained stations because of greaterpore-water supply of oxidants (see Allison1990; Walter and Burton 1990). SEM exami-nation of shells from analogous environmentsin the Caribbean San Blas Archipelago of Pan-ama (in progress) indicates that all of these ex-cept abrasion are probably factors in the fine-scale alteration associated with hard-substratain the Bocas embayment. In muddy siliciclas-tic sediments, all of these factors would be re-


duced. Preliminary results indicate that thehigh iron content of siliciclastics in our SanBlas study area fosters iron reduction andoversaturated pore waters (cf. Aller et al. 1996;Ku and Walter 1998, unpublished; Best et al.1999), and this would cause an even greatertaphonomic contrast between assemblagesfrom hard-carbonate and soft-siliciclastic en-vironments such as sampled in Bocas.

Finally, although edge-rounding has beenused as a proxy for combined physical abra-sion and chemical solution by many workers(Cutler 1987; Davies et al. 1989), based in parton tumbling experiments by Chave (1964), itis probably largely biological in origin in Bo-cas assemblages and dependent upon seafloorexposure. SEM examination suggests algal orother micro-bioerosion, exacerbated by graz-ing of shell edges (Best and Kidwell 1996).Moreover, in contrast to previous studies,edge-rounding in Bocas assemblages is not re-stricted to settings of prolonged or high-en-ergy reworking or to thick shells (thought tobetter withstand such treatment).

The types of damage that are most commonin hard-substrata death assemblages in Bocasare thus linked to seafloor exposure, either in-termittent or continuous, and to the opennessof sediments to solute exchange, rather than tohigh energy per se. Death assemblages inmuds among patch reefs include skeletal de-bris that carries the taphonomic signature ofthese hard-substrate habitats. In contrast,postmortem damage to death assemblagesfrom exclusively soft-sediment environmentsis virtually nil (Fig. 4B) and is primarily a mi-crobial signature of sediment burial. The lowfrequency of encrustation and boring in theseenvironments appears to reflect the limitedexposure of shells above the sediment-waterinterface: infesting taxa occur in the samerank order as in hard-substrata death assem-blages, simply at much lower frequencies, in-dicating a shared larval species pool (Fig. 6).The few shells that are exposed in soft sedi-ments may in fact be more intensely colonizedbecause of substratum limitation and lowergrazing intensity (e.g., Jackson 1977; support-ed by experimental results of Best and Kid-well 1996; Best 1998c).

Broader Comparisons.—Working in pure-car-

bonate habitats of northeastern Caribbean,Parsons (1993; see Table 4 for methods) alsofound generally higher levels of damage inreef and high-energy sand environments thanin bioturbated sands and muds, with the low-est damage levels in mud (Dent [1995] indi-cates that this mud includes considerable sili-ciclastic component). Encrustation and clionidsponge borings (the only endobiont tallied)were consistently high in all reef habitats andin rippled sands of the adjacent shallow shelf(1.5–22 m; 13–67% of shells, versus ,2% ofshells in muds and bioturbated sands ,5.5 m;compare with our per-sample data in Fig. 4A).Fine-scale alteration (excluding rhizome etch-ings and polishing from abrasion) was alsohigh in reefs and in sandy grass beds andbeach sands (;40–95% of shells damaged), incontrast to shells in bioturbated sands andmuds (0–8% of shells damaged, except for onestation; compare with our Fig. 4). All of theseenvironments would have experienced simi-larly high levels of water clarity, and thus dif-ferences in damage levels were attributed todifferences in water energy, frequency of shellexhumation, and sediment accumulationrates, particularly where sedimentary veneerson hard substrata were especially thin. On thecarbonate platform of the Florida Keys, Dent(1995) also found the highest damage levels inreef-framework and back-reef rubble habitats,and attributed this to more frequent physicalreworking and exposure of shells. His singlemud station (pure carbonate) was in a channelof presumed low net sedimentation, whichwas invoked to explain why shells exhibitedmore damage than found by Parsons (1993) inher more onshore, sediment-trapping mudhabitat.

Workers in extratropical siliciclastic sedi-ments have also underscored the importanceof shell exhumation cycles and of low net sed-imentation. High damage levels have beendocumented in mollusk shells from storm-re-worked clastic-bypassed tidal inlets (Staff andPowell 1990a); from beaches, channels, andwave-influenced portions of macrotidal flats(Meldahl and Flessa 1990; Feige and Fursich1991); from clastic-starved nearshore areas be-tween active coastal alluvial fans (Meldahl etal. 1997); from palimpsest offshore sands


TA

BL

E4.

Com

par

ison

ofre

sult

sfr

omqu

anti

tati

veta

ph

ofac

ies

stu

die

sin

mod

ern

env

iron

men

ts,

ord

ered

lati

tud

inal

ly.

Un

less

not

ed,a

lld

amag

eb

ased

onp

oole

dob

-se

rvat

ion

son

shel

lin

teri

oran

dex

teri

orsu

rfac

es.n

.s. 5

not

spec

ified

.Cod

esfo

rty

pes

ofd

amag

eev

alu

ated

:Abr

5ab

rasi

on,B

I5

biot

icin

tera

ctio

ns

(su

mof

bioe

rosi

on,

encr

ust

atio

n,p

red

atio

n,e

tc.)

,Bio

5bi

oero

sion

,Bor

5b

orin

g(m

acro

scop

ic),

C5

crac

kin

g,C

L5

colo

rlo

ss,C

orr

5co

rros

ion

,Dis

art5

dis

arti

cula

tion

,Dis

s5

dis

solu

tion

,E

nc

5en

cru

stat

ion

,ER

5ed

ge-

rou

nd

ing

,Exf

5ex

foli

atio

n,F

rag

5fr

agm

enta

tion

,FSA

5fi

ne

scal

eal

terc

atio

n(o

fex

t5

exte

rior

&in

t5

inte

rior

surf

aces

)L

EO

5lo

ssof

exte

rior

orn

amen

tati

on,

MB

5m

icro

bor

ing

,P

pt

5p

reci

pit

atio

n,

R/

L5

rati

ori

ght/

left

valv

es,

Rad

5ra

du

lar

mar

ks,

RE

5ro

otet

chin

gs,

SF5

size

freq

uen

cy,T

M5

thin

ned

mar

gin

s

Stu

dy

area

(au

thor

)

En

vir

onm

ents

sam

ple

d(d

epth

)

Size

frac

tion

( .X

mm

)Ta

xaT

yp

esof

dam

age

Met

ho

dof

scor

ing

dam

age

Key

resu

lts

Sili

cicl

asti

cE

nv

iron

men

ts

Skag

erra

kco

ast,

Nor

way

(Cu

tler

and

Fle

ssa

1995

)B

each

3in

fau

nal

and

epif

au-

nal

biva

lves

Bio

,Cor

r(d

isso

lu-

tion

and

mac

era-

tion

),P

pt

pre

sen

ce/

abse

nce

ofea

chty

pe

ofd

amag

eon

each

shel

l(S

EM

)

maj

orit

yof

shel

lssh

owm

icro

bor

ing

,hal

fsh

owsi

gn

sof

dis

solu

-ti

onor

mac

erat

ion

,an

d10

%sh

owp

reci

pi-

tati

onN

orth

Sea

coas

t,G

erm

any

(Kow

alew

ski

etal

.19

95)

low

-en

erg

yti

dal

flat

san

dsu

bti

dal

chan

nel

sn

.s.

infa

un

albi

valv

eC

er-

asto

derm

aed

ule

Abr

,Bio

,En

c,F

rag

,F

SA(i

nt

and

ext)

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

gen

eral

lym

oder

ate

abra

-si

on,v

ery

litt

lefr

ag-

men

tati

on,a

nd

hig

hly

vari

able

surf

ace

alte

r-at

ion

Cap

eC

od,M

assa

chu

sett

s(M

eld

ahl

and

Fle

ssa

1990

)sa

ltm

arsh

,in

ner

flat

s,m

idd

lefl

ats,

oute

rfl

ats,

bea

ch,s

ub

tid

al(0

–2.5

m),

sub

tid

al(2

.5–1

0m

)

6in

fau

nal

biva

lve

Mer

-ce

nar

iam

erce

nar

iaA

br,B

io,C

orr,

En

c,F

rag

pre

sen

ce/

abse

nce

ofea

chty

pe

ofd

amag

eon

each

shel

l

mar

shan

dsu

bti

dal

san

ds

gen

eral

lyd

iffe

r-en

td

amag

efr

omin

-te

rtid

alfl

ats,

but

no

sim

ple

tren

ds

Sap

elo

Isla

nd

,G

eorg

ia(F

rey

and

How

ard

1986

)of

fsh

ore

pal

imp

sest

san

d(1

2–15

m)

2.5

enti

rem

ollu

scan

dea

thas

sem

blag

eB

or,C

L,E

nc,

FSA

pre

sen

ce/

abse

nce

ofea

chty

pe

ofd

amag

eon

each

shel

l

hig

hes

td

amag

ein

site

sw

ith

slow

est

sed

imen

-ta

tion

rate

s

San

Lu

isP

ass,

Texa

s(D

avie

set

al.1

989)

mic

roti

dal

inle

t(s

torm

-d

omin

ated

tid

ald

el-

ta)

24

sub

sets

ofm

ollu

s-ca

nd

eath

asse

m-

blag

e,se

gre

gat

edby

life

hab

itat

s(i

n-

ner

shel

f,b

each

/in

-le

t,b

ay,e

ury

top

ic)

Abr

,BI,

Cor

r 1M

B,

Dis

art,

ER

,Fra

g,

SF

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

dam

age

lin

ked

toli

feh

abit

ats

ofsp

ecie

s,ac

-cr

ued

bef

ore

tran

spor

tto

tid

ald

elta

;cor

ro-

sion

/m

icro

bor

ing

hig

hes

tin

eury

top

icsp

ecie

s,ed

ge-

rou

nd

-in

gan

dab

rasi

onh

igh

-es

tin

bay

spec

ies

Texa

sco

ast

(Sta

ffan

dP

owel

l19

90a)

mic

roti

dal

inle

t,in

ner

shel

f(1

5–22

m)

4en

tire

mol

lusc

and

eath

asse

mbl

age

Abr

,BI,

Cor

r 1M

B,

Dis

art,

ER

,Fra

g,

R/

L,S

F

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

dam

age

dec

reas

esof

f-sh

ore,

exce

pt

corr

o-si

on;d

amag

efo

cuse

don

shel

lfr

agm

ents


TA

BL

E4.

Con

tin

ued

.

Stu

dy

area

(au

thor

)

En

vir

onm

ents

sam

ple

d(d

epth

)

Size

frac

tion

( .X

mm

)Ta

xaT

yp

esof

dam

age

Met

ho

dof

Scor

ing

Dam

age

Key

resu

lts

Ch

olla

Bay

,Gu

lfof

Cal

ifor

nia

,So

nor

a(F

eig

ean

dF

urs

ich

1991

)

salt

mar

shch

ann

el,i

n-

ner

flat

,mid

dle

flat

,m

idd

lech

ann

el,o

ute

rfl

at,

oute

rch

ann

el

3en

tire

mol

lusc

and

eath

asse

mbl

age

and

targ

etg

ener

aP

roto

thac

a,C

hion

e,D

osin

ia,T

agel

us,

Cer

ithi

um

,Cer

ithi

-de

a,an

dE

ncop

e

Bor

,CL

,Cor

r,D

i-sa

rt,E

nc,

Fra

g,

LE

O,R

/L

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

hig

hes

tov

eral

ld

amag

ein

chan

nel

s;m

ost

dam

age

incr

ease

sse

a-w

ard

acro

ssfl

ats

asen

erg

yan

dbi

ogen

icre

wor

kin

gin

crea

sean

das

sed

imen

tati

onra

ted

ecre

ases

Ch

olla

Bay

,Gu

lfof

Cal

ifor

nia

,So

nor

a(C

utl

er19

95)

salt

mar

sh,m

arsh

chan

nel

,bea

ch,s

pit

wit

hd

un

es,i

nn

erfl

at,

mid

dle

flat

,ou

ter

flat

,ro

cky

flat

,su

bti

-d

alro

cky

san

d(1

–5m

)

6in

fau

nal

biva

lves

Chi

-on

eflu

ctif

raga

and

Chi

one

cali

forn

ien

sis

Abr

,Cor

r,M

B,R

ad,

RE

pre

sen

ce/

abse

nce

ofea

chty

pe

ofd

amag

eon

each

shel

l

hig

hes

tov

eral

ld

amag

ein

spit

;mic

rob

orin

gd

omin

ates

all

exce

pt

bea

ch,w

her

eab

rasi

ond

omin

ates

;oth

erd

am-

age

typ

esh

ave

lim

ited

dis

trib

uti

ons

San

Feli

pe,

Gu

lfof

Cal

ifor

nia

,B

aja

Cal

ifor

nia

(Kow

ales

ki

etal

.199

4)

mod

ern

chen

ier,

sub

-m

oder

nch

enie

r,su

b-

foss

ilch

enie

r

12.5

infa

un

albi

valv

eM

u-

lin

iaco

lora

doen

sis

Bio

,C,E

nc,

ER

,Exf

,F

rag

,FSA

(ext

.vs.

int.

)

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

dam

age

incr

ease

sw

ith

geo

log

icag

ean

dis

gre

ates

tfo

rsh

ells

sam

ple

dfr

omch

enie

rsu

rfac

eB

ahıa

Con

cep

cion

,Gu

lfof

Cal

ifor

nia

,Baj

aC

alif

orn

iaSu

r(M

eld

ahl

etal

.199

7)

inte

rtid

alm

ang

rove

,n

ears

hor

e(1

–5m

),n

ears

hor

e(5

–20

m),

offs

hor

em

ud

( .20

m),

nea

rsh

ore

bio-

clas

tic

san

d

15in

fau

nal

biva

lve

spe-

cies

Bor

(pre

dat

ory

1p

ostm

orte

m,i

ncl

.M

B),

CL

,FSA

(in

t),L

EO

,TM

each

shel

lin

teri

orsc

ored

acco

rdin

gto

over

all

dam

age

pro

file

(4g

rad

esof

dam

age)

hig

hes

tov

eral

ld

amag

ein

shal

low

est

wat

er;

amon

gn

ears

hor

esi

tes,

hig

hes

td

amag

ein

area

sof

slow

est

sed

-im

enta

tion

Pu

re-c

arb

onat

een

vir

onm

ents

Flo

rid

aK

eys

(Den

t19

95,

1996

)on

shor

e-of

fsho

retr

an-

sect

ofFl

orid

ash

elf:

gras

sbe

ds,

sand

s,re

efh

ard

grou

nds 1

rubb

le(1

–7m

),p

lus

2sa

ndan

dh

ard

grou

ndfo

re-

reef

site

s(2

1–33

m)

215

mos

tab

un

dan

tm

ollu

scan

spec

ies

inst

ud

yar

ea(o

f38

6to

tal)

Abr

,Bor

(sp

ong

eon

ly),

CL

,En

c,E

R,F

rag

,FSA

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

dam

age

onre

efis

gre

at-

erth

anin

all

oth

eren

-v

iron

men

tsw

ith

the

exce

pti

onof

frag

men

-ta

tion

and

colo

rlo

ss


TA

BL

E4.

Con

tin

ued

.

Stu

dy

area

(au

thor

)

En

vir

onm

ents

sam

ple

d(d

epth

)

Size

frac

tion

( .X

mm

)Ta

xaT

yp

esof

dam

age

Met

ho

dof

Scor

ing

Dam

age

Key

resu

lts

St.

Cro

ix,

Vir

gin

Isla

nd

s,an

dM

ona

Isla

nd

,P

uer

toR

ico

(Par

son

s19

89,

1992

,P

ar-

son

san

dB

rett

1991

)

bea

ch,m

ud

,bio

turb

at-

edsa

nd

,gra

ssb

ed,

san

dp

ocke

tsw

ith

inp

atch

reef

,bac

kre

ef,

fore

reef

,alg

alsa

nd

onop

ensh

elf

har

d-

gro

un

d

5en

tire

mol

lusc

and

eath

asse

mbl

age

Dis

art,

Fra

g,E

nc,

Bio

,RE

,ER

,Abr

,C

L

deg

ree

ofea

chty

pe

ofd

amag

eon

each

shel

l

hig

hes

tov

eral

ld

amag

ein

reef

san

din

bea

chsa

nd

,lea

std

amag

ein

mu

d;r

eef

faci

esn

otd

isti

ng

uis

hab

lefr

omea

chot

her

,bu

td

isti

nct

from

all

oth

erfa

cies

Bot

hsi

lici

clas

tic

and

carb

onat

een

vir

onm

ents

Boc

asd

elTo

ro,

Car

ibb

ean

Pan

ama

(th

isp

aper

)sa

nd

ym

ud

(sil

ici.

),m

ud

(sil

ici.

),m

ud

amon

gp

atch

reef

s,p

atch

reef

,Hal

imed

ag

rave

lly

san

d

8al

lbi

valv

esp

ecie

sD

isar

t,F

rag

,En

c,B

or,E

R,F

SA(i

nt.

)d

egre

eof

each

typ

eof

dam

age

onea

chsh

ell

inte

rior

hig

hes

td

amag

ein

env

i-ro

nm

ents

wit

hh

ard

sub

stra

ta

(Frey and Howard 1986); and, within chenierplains, from relatively old shell ridges thathave suffered greatest meteoric exposure (Ko-walewski et al. 1994) (Table 4). In the most rel-evant subset of subtidal muddy sediments,damage levels (based on both shell exteriorand interior surfaces) are only slightly higherthan those we observed in tropical siliciclasticmuds of the Bocas embayment. For example,between fair-weather and storm wave base onthe Texas shelf, 91–98% of bivalve shells 4–12mm were disarticulated, 16% were fragment-ed, 16% had significant fine-scale alteration(‘‘major dissolution’’), and 8–13% were en-crusted or bored (latter figure for pooled bi-valves and gastropods) (Staff and Powell1990b; compare with our Fig. 4B). In offshoresiliciclastic muds of Bahıa Concepcıon, pres-ervation of bivalve death assemblages is con-sistently ‘‘good’’ to pristine, with .50% ofspecimens in ‘‘excellent’’ condition in all sam-ples (color retained, no postmortem infesta-tion, no edge alteration, may have some lossof luster) (Meldahl et al. 1997) .

The excellent condition of shells from sili-ciclastic muds in Bocas del Toro is not uniquewithin the Tropics, as we are finding similarlywell-preserved material from subtidal silici-clastic muds in the San Blas Archipelago ineastern Panama (Best 1996b, 1998b,c). In theabsence of pure-carbonate muds within the Bo-cas study area, it is not possible to isolate theeffects of sediment composition on the onehand from sediment grain size and its physi-cal environmental correlates on the other.Nonetheless, the differences in damage profilewould be difficult to explain by sedimentcomposition alone, so we feel confident thatsediment composition is probably of only sec-ondary importance among the set of environ-ments examined in Bocas. The effect of sedi-ment composition can be tested more straight-forwardly in the San Blas area, where both sil-iciclastic and pure-carbonate muds arepresent. In our preliminary data from SanBlas, bivalve shell condition is poorer in pure-carbonate muds, primarily because of greaterfine-scale surface alteration (Best et al. 1999and unpublished). Such differences in sedi-ment composition may in fact explain in partwhy Dent (1995) found that shells from his


pure-carbonate mud in Florida were in poorercondition than shells from Parsons’s (1993) ap-parently impure muds in the Virgin Islands.

Conclusions

These results from Bocas del Toro constitutethe first taphonomic characterization of mol-luscan death assemblages in Recent tropicalsiliciclastic sediments and can be comparedwith assemblages from pure-carbonate hardsubstrata in the same back-arc embayment,broadening the actualistic database for strati-graphically significant shallow-marine envi-ronments. To the advantage of taphofaciesanalysis in analogous fossil records, bivalveshells are relatively abundant, death assem-blages have not been homogenized by post-mortem transport, and each environmentyields a distinctive damage profile. Between-sample variation is high within the physicallyheterogeneous hard-substrata environments,however, requiring greater pooling of data.The most distinct differences are betweenhard-substrata environments (patch reef, Hal-imeda gravelly sand, mud among patch reefs)and exclusively fine-grained, soft-sedimentenvironments (sandy mud, mud), with the lat-ter characterized by superb shell preservation.Because we only tallied damage to shell inte-riors, these differences are unambiguouslytaphonomic (postmortem) in origin.

Tests to determine rates and specific agentsof damage are still in progress. However, theprimary environmental (extrinsic) controls ondamage are probably differences in exhuma-tion cycles and burial rate, given the depen-dence of the most common kinds of shell dam-age on seafloor exposure. In hard-substrataenvironments, residence times at or very nearthe seafloor probably depend upon the fre-quency of physical reworking (relatively highin Halimeda gravelly sands but less so in lee-positioned patch reefs) and the thickness ofthe sedimentary veneer (i.e., sediment avail-able to bury shells; not great in either environ-ment). In muddy substrata, we suspect dam-age levels are kept low by high rates of netsediment accumulation, low frequency ofphysical reworking of shells, and easily (e.g.,tidally) resuspended sediments at the seafloorthat would rapidly coat and effectively bury

shells upon death (for rare epifauna) or ex-humation (including exhumation by burrow-ers). The only significantly non-zero tapho-nomic damage to shells in these exclusivelysoft-sediment environments, aside from peri-mortem disarticulation and fragmentation, isfine-scale alteration of shell interiors and com-missural rounding. This alteration is probablylargely microbial in origin and can proceedduring shallow burial within anaerobic sedi-ments.

Regardless of cause, the excellent conditionof shells in assemblages from muddy tropicalsiliciclastic sediments suggests a less filteredrecord of original biological input than in as-semblages from closely associated reefs and al-gal meadows. Low damage levels suggest thattaphonomic processes have operated at verylow intensity, thereby causing little differentialdestruction of species (i.e., little bias), and/orthat sedimentation rates are very high, result-ing in little time-averaging. Such optimistic ex-trapolations require testing before applicationto the fossil record, but preliminary resultsfrom experimental shells suggest that rates ofdamage for buried shells in all environmentsare quite low, with at most a few percentweight loss in a 27-month period and all of thisachieved within the first 12 months (Best andKidwell 1996; Best 1998c; Best et al. 1999). Alltypes of damage show higher levels in hard-substrata assemblages, but these also includetypes that accrue only during exposure. Al-though key rates are yet to be quantified, givensuch significant differences in shell condition,the compositions of assemblages from thesecontrasting sets of facies (or alternating shaleand limestone beds) are not likely to be isota-phonomic, that is providing the same qualitiesof paleobiologic data. However, within each ofthe two sets of environments tested here, as-semblages are quite similar in condition, andthus might be accepted as taphonomicallyequivalent sources of paleontologic informa-tion.

Acknowledgments

We are grateful to J. B. C. Jackson and A. G.Coates of the Panama Paleontology Project,Smithsonian Tropical Research Institute, forencouraging and facilitating this project; L. S.


Collins for the loan of sampling gear; and D.West and the crew of the RV Uracca for theircheerful cooperation. We thank A. I. Miller, G.Cadee, and D. Jablonski for helpful reviews ofthe original manuscript. Analysis supportedby National Science Foundation grant EAR-9628345.

Literature Cited

Aller, R. C. 1982. Carbonate dissolution in nearshore terriginousmuds: the role of physical and biological reworking. Journalof Geology 90:98–154.

Aller, R. C., N. E. Blair, Xia Q., and P. D. Rude. 1996. Reminer-alization rates, recycling, and storage of carbon in Amazonshelf sediments. Continental Shelf Research 6:753–786.

Allison, P. A. 1990. Variation in rates of decay and disarticula-tion of Echinodermata: implications for the application of ac-tualistic data. Palaios 5:432–440.

Anderson, L. C., R. A. McBride, M. J. Taylor, and M. R. Byrnes.1998. Late Holocene record of community replacement pre-served in time-averaged molluscan assemblages, Louisianachenier plain. Palaios 13:488–499.

Best, M. M. R. 1996a. Preservation of calcium carbonate skele-tons in tropical shelf environments of the San Blas Archipel-ago, Caribbean Panama. Eighth international coral reef sym-posium (Panama), Abstracts of papers, p. 18.

. 1996b. Actualistic bivalve taphonomy in carbonate andsiliciclastic tropical marine shelf environments of the San BlasArchipelago, Caribbean Panama. Geological Society of Amer-ica Abstracts with Programs 28:364.

. 1998a. Distribution and nature of siliciclastic and car-bonate sediments on the tropical American shelves: signifi-cance for carbonate burial. Geological Association of CanadaAnnual Meeting Abstracts 23:17.

. 1998b. Taxonomic and taphonomic rarefaction: a metricfor assessing sampling and divergence of taphonomic signa-ture. Canadian Paleontology Conference Program and Ab-stracts 8:8.

. 1998c. Experimental and death assemblage bivalve ta-phonomy in tropical carbonate and siliciclastic environments,San Blas Archipelago, Caribbean Panama. Geological Societyof America Abstracts with Programs 30:383.

Best, M. M. R., and S. M. Kidwell. 1995. Actualistic bivalve ta-phonomy in tropical marine siliciclastics: preliminary resultsfrom natural death assemblages and experimental arrays inCaribbean Panama. Geological Society of America Abstractswith Programs 27:134.

. 1996. Bivalve shell taphonomy in tropical siliciclasticmarine environments: preliminary experimental results.Sixth North American paleontology convention, Abstracts ofpapers. Paleontological Society Special Publication 8:34.

. 1999. Bivalve taphonomy in tropical mixed siliciclastic-carbonate settings. II. Effect of bivalve life habits and shelltypes. Paleobiology 26:103–115.

Best, M. M. R., and J. M. Pandolfi. 1992. Molluscan taphonomyin Madang Lagoon, Papua New Guinea. Proceedings of theSeventh International Coral Reef Symposium, Guam 1:437.

Best, M. M. R., S. M. Kidwell, T. C. W. Ku, and L. M. Walter. 1999.Bivalve taphonomy and porewater geochemistry in tropicalcarbonate and siliciclastic marine environments: implicationsfor the preservation of carbonate. Geological Association ofCanada Annual Meeting Abstracts, Vol. 24.

Boekschoten, G. T. 1966. Shell borings of sessile epibiontic or-ganisms as paleoecological guides (with examples from the

Dutch coast). Palaeogeography, Palaeoclimatology, Palaeo-ecology 2:333–379.

Bosence, D. W. J. 1979. Live and dead faunas from coralline algalgravels, County Galway. Palaeontology 22:449–478.

Cadee, G. H. 1968. Molluscan biocoenoses and thanatocoenosesin the Ria de Arosa, Galicia, Spain. Zoologische Verhandelin-gen (Rijksmuseum Natuurlijke Historie Leiden) 95.

. 1984. Macrobenthos and macrobenthic remains on theOyster Ground, North Sea. Netherlands Journal Sea Rsearch18:160–178.

. 1994. Eider, shelduck, and other predators, the mainproducers of shell fragments in the Wadden Sea—paleoeco-logical implications. Palaeontology 37:181–202.

Cate, A. S., and I. Evans. 1994. Taphonomic significance of thebiomechanical fragmentation of live molluscan shell materialby a bottom-feeding fish (Pogonias cromis) in Texas coastalbays. Palaios 9:254–274.

Chave, K. E. 1964. Skeletal durability and preservation. Pp. 377–387 in J. Imbrie and N. D. Newell, eds. Approaches to paleo-ecology. Wiley, New York.

Coates, A. G., J. B.C. Jackson, L. S. Collins, T. M. Cronin, H. J.Dowsett, L. M. Bybell, P. Jung, and J. A. Obando. 1992. Clo-sure of the Isthmus of Panama: the near-shore marine recordof Costa Rica and western Panama. Geological Society ofAmerica Bulletin 104:814–828.

Crenshaw, M. A. 1980. Mechanisms of shell formation and dis-solution. Pp. 115–132 in D. A. Rhoads and R. A. Lutz. Skeletalgrowth of aquatic organisms: biological records of environ-mental change. Plenum, New York.

Cutler, A. H. 1987. Surface textures of shells as taphonomic in-dicators. In K. W. Flessa, ed. Paleoecology and taphonomy ofrecent to Pleistocene intertidal deposits, Gulf of California.Paleontological Society Special Publication 2:164–176.

. 1995. Taphonomic implications of shell surface texturesin Bahıa la Choya, northerm Gulf of California. Palaeogeog-raphy, Palaeoclimatology, Palaeoecology 114:219–240.

Cutler, A. H., and K. W. Flessa. 1995. Bioerosion, dissolution andprecipitation as taphonomic agents at high and low latitudes.Senckenbergiana maritima 25:115–121.

Davies, D. J., E. N. Powell, and R. J. Stanton. 1989. Taphonomicsignature as a function of environmental process: shells andshell beds in a hurricane-influenced inlet on the Texas coast.Palaeogeography, Palaeoclimatology, Palaeoecology 72:317–356.

Dent, S. R. 1995. A taphofacies model of the recent South Floridacontinental shelf: a new perspective for a classic, exposed car-bonate environment. Ph. D. dissertation. University of Cincin-nati, Cincinnati, Ohio.

. 1996. Recent molluscan taphofacies of south Florida: acomparison of basic analytical methods. Geological Society ofAmerica Abstracts with Programs 28:A364.

Ekdale, A. A. 1972. Ecology and paleoecology of marine inver-tebrate communities in calcareous substrates, northeast Quin-tana Roo, Mexico. M.A. thesis. Rice University, Houston.

. 1977. Quantitative paleoecological aspects of modernmarine mollusk distribution, northeast Yucatan coast, Mexi-co. American Association of Petroleum Geologists Studies inGeology 4:195–207.

Feige, A., and F. T. Fursich. 1991. Taphonomy of the Recent mol-luscs of Bahıa la Choya (Gulf of California, Sonora, Mexico).Zitteliana 18:89–113.

Flessa, K. W. 1998. Well-traveled cockles: shell transport duringthe Holocene transgression of the southern North Sea. Geol-ogy 26:187–190.

Frey, R. W., and J. D. Howard. 1986. Taphonomic characteristicsof offshore mollusk shells, Sapelo Island, Georgia. TulaneStudies in Geology and Paleontology 19:51–61.

Fursich, F. T., and K. W. Flessa. 1987. Taphonomy of tidal flat


molluscs in the northern Gulf of California: paleoenviron-mental analysis despite the perils of preservation. Palaios 2:543–559.

Glover, C. P., and S. M. Kidwell. 1993. Influence of organic ma-trix on the post-mortem destruction of molluscan shells. Jour-nal of Geology 101:729–747.

Greb, L., B. Saric, H. Seyfried, T. Broszonn, S. Brauch, G. Gugau,C. Wiltschko, and R. Leinfelder. 1996. Okologie und sedimen-tologie eines rezenten rampensystems an der Karibikkustevon Panama. Institut fur Geologie und Palaontologie, Uni-versitat Stuttgart. Profil 10.

Green, M. A., R. C. Aller, and J. Y. Aller. 1996. Benthic biogeo-chemical changes associated with calcite undersaturation inLong Island Sound sediments. In Geology of Long Island andmetropolitan New York. Department of Earth and Space Sci-ences, SUNY Stony Brook, Stony Brook, N.Y.

Greenstein, B. J., and H. A. Moffat. 1996. Comparative taphon-omy of modern and Pleistocene corals, San Salvador, Baha-mas. Palaios 11:57–63.

Greenstein, B. J., and J. M. Pandolfi. 1997. Preservation of com-munity structure in modern reef coral life and death assem-blages of the Florida Keys: implications for the Quaternaryfossil record of coral reefs. Bulletin of Marine Science 61:431–452.

Guzman, H. M., and C. A. Guevara. 1998a. Arrecifes coralinosde Bocas del Toro, Panama. I. Distribucion, estructura y es-tado de conservacion de los arrecifes continentales de la La-guna de Chiriquı y la Bahıa Almirante. Revista de BiologıaTropical 46:601–623.

. 1998b. Arrecifes coralinos de Bocas del Toro, Panama.II. Distribucion, estructura y estado de conservacion de losarrecifes de las Islas Bastimentos, Solarte, Carenero y Colon.Revista de Biologıa Tropical 46:893–916.

Havach, S. M., and L. S. Collins. 1997. The distribution of recentbenthic foraminifera across habitats of Bocas del Toro, Carib-bean Panama. Journal of Foraminiferal Research 27:232–249.

Heck, K. L., Jr., G. Van Belle, and D. Simberloff. 1975. Explicitcalculation of the rarefaction diversity measurement and thedetermination of sufficient sample size. Ecology 56:1459–1461.

Henderson, S. W., and R. W. Frey. 1986. Taphonomic redistri-bution of mollusk shells in a tidal inlet channel, Sapelo Island,Georgia. Palaios 1:3–16.

Jackson, J. B. C. 1977. Habitat area, colonization, and develop-ment of epibenthic community structure. Pp. 349–358 in B. F.Keegan, P. O. Ceidigh, and P. J. S. Boaden, eds. Biology of ben-thic organisms. Pergamon, Oxford.

Jackson, J. B. C., A. F. Budd, and A. G. Coates, eds. 1997. Evo-lution and environment in tropical America. University ofChicago Press, Chicago.

Johnson, R. G. 1965. Pelecypod death assemblages in TomalesBay, California. Journal of Paleontology 39:80–85.

Kidwell, S. M., and D. W. J. Bosence. 1991. Taphonomy and time-averaging of marine shelly faunas. In P. A. Allison and D. E.G. Briggs, eds. Taphonomy: releasing the data locked in thefossil record. Topics in Geobiology 9:115–209. Plenum, NewYork.

Kiene, W. 1997. Enriched nutrients and their impact on bioero-sion: Results for ENCORE. Proceedings of the Eighth Inter-national Coral Reef Symposium, Panama 1:897–902.

Kowalewski, M., K. W. Flessa, and J. A. Aggen. 1994. Taphofa-cies anaysis of Recent shelly cheniers (beach ridges), north-eastern Baja California, Mexico. Facies 41:209–242.

Kowalewski, M., K. W. Flessa, and D. P. Hallman. 1995. Ternarytaphograms: triangular diagrams applied to taphonomicanalysis. Palaios 10:478–483.

Ku, T. C. W., and L. M. Walter. 1998. Pore water geochemistry ofshallow marine carbonates vs. tropical siliciclastics: carbon,

sulfur and iron systematics and implications for carbonatepreservation. Geological Society of America Abstracts withPrograms 30:374.

Llewellyn, G., and C. G. Messing. 1993. Compositional andtaphonomic variations in modern crinoid-rich sedimentsfrom the deep-water margin of a carbonate bank. Palaios 8:554–573.

Meldahl, K. H., and K. W. Flessa. 1990. Taphonomic pathwaysand comparative biofacies and taphofacies in a Recent inter-tidal/shallow shelf environment. Lethaia 23:43–60.

Meldahl, K. H., K. W. Flessa, and A. H. Cutler. 1997a. Time-av-eraging and postmortem skeletal survival in benthic fossil as-semblages: quantitative comparisons among Holocene envi-ronments. Paleobiology 23:207–229.

Meldahl, K. H., O. G. Yajimovich, C. D.Empedocles, C.S.Gustafson, M. M. Hidalgo, and T. W. Reardon. 1997. Holo-cene sediments and molluscan faunas of Bahıa Concepcion: amodern analog to Neogene rift basins of the Gulf of Califor-nia. Geological Society of America Special Paper 318:39–56.

Miller, A. I. 1988. Spatial resolution in subfossil molluscan re-mains: implications for paleobiological analyses. Paleobiolo-gy 14:91–103.

Miller, A. I., G. Llewellyn, K. M. Parsons, H. Cummins, M. R.Boardman, B. J. Greenstein, and D. K. Jacobs. 1992. Effect ofHurricane Hugo on molluscan skeletal distributions, Salt Riv-er Bay, St. Croix, U.S. Virgin Islands. Geology 20:23–26.

Nebelsick, J. H. 1992. Echinoid distribution by fragment iden-tification in the northern Bay of Safaga, Red Sea, Egypt. Pa-laios 7:316–328.

. 1995. Comparative taphonomy of Clypeasteroids. Eclo-gae Geologica Helvetica 88:685–693.

Nebelsick, J. H., B. Schmid, and M. Stachowitsch. 1997. The en-crustation of fossil and recent sea-urchin tests: ecological andtaphonomic significance. Lethaia 30:271–284.

Pandolfi, J. M. and B. J. Greenstein. 1997a. Taphonomic alter-ation of reef corals: effects of reef environment and coralgrowth form. I. The Great Barrier Reef. Palaios 12:27–42.

. 1997b. Preservation of community structure in death as-semblages of deep-water Caribbean reef corals. Limnologyand Oceanography 42:1505–1516.

Pandolfi, J. M., and P. R. Minchin. 1995. A comparison of taxo-nomic composition and diversity between reef coral life anddeath assemblages in Madang Lagoon, Papua New Guinea.Paleogeography, Paleoclimatology, Paleoecology 119:321–341.

Parsons, K. M. 1989. Taphonomy as an indicator of environ-ment: Smuggler’s Cove, St. Croix, U.S.V.I. In D. K. Hubbard,ed. Terrestrial and marine ecology of St. Croix, U.S. Virgin Is-lands. West Indies Laboratory Special Publication 8:135–143.

. 1993. Taphonomic attributes of mollusks as predictorsof environment of deposition in modern carbonate systems:northeastern Caribbean. Ph.D. disseration. University ofRochester, Rochester, N.Y.

Parsons, K. M., and C. E. Brett. 1991. Taphonomic processes andbiases in modern marine environments: an actualistic per-spective on fossil assemblage preservation. Pp. 22–65 in S. K.Donovan, ed. The processes of fossilization. Columbia Uni-versity Press, New York.

Perry, C. T. 1996. The rapid response of reef sediments to chang-es in community composition: implications for time averag-ing and sediment accumulation. Journal of Sedimentary Re-search 66:459–467.

. 1998. Grain susceptibility to the effects of microboring:implications for the preservation of skeletal carbonates. Sed-imentology 45:39–51.

Pilkey, O. H., and H. A. Curran. 1986. Molluscan shell trans-port—you ain’t seen nothin’ yet. Palaios 1:197.

Sadler, J. C., M. A. Lander, A. M. Hori, and L. K. Oda. 1987.


Tropical marine climatic atlas, Vol. 1. Indian Ocean and At-lantic Ocean. University of Hawaii, Honolulu.

Sanders, H. L. 1968. Marine benthic diversity: a comparativestudy. American Naturalist 102:243–282.

Schafer, W. 1972. Ecology and palaeoecology of marine envi-ronments. University of Chicago Press, Chicago.

Seyfried, H., and W. Hellmann, eds. 1994. Geology of an evolv-ing island arc: the isthmus of southern Nicaragua, Costa Rica,and western Panama. Institut fur Geologie und Palaontologie,Universitat Stuttgart. Profil 7.

Simon, A., and M. Poulicek. 1990, Biodegradation anaerobiquedes structures squelettiques en milieu marin. I. Approchemorphologique. Cahiers de Biologie Marine 31:95–105.

Sokal, R. R., and F. J. Rohlf. 1995. Biometry, 3d ed. W. H. Free-man, New York.

Staff, G. M, and E. N. Powell. 1990a. Taphonomic signature andthe imprint of taphonomic history: discriminating betweentaphofacies of the inner continental shelf and a microtidal in-let. Paleontological Society Special Publication 5:370–390.

. 1990b. Local variability of taphonomic attributes in aparautochthonous assemblage: can taphonomic signature dis-tinguish a heterogeneous environment? Journal of Paleontol-ogy 64:648–658.

Tudhope, A. W. 1989. Shallowing-upwards sedimentation in acoral reef lagoon, Great Barrier Reef of Australia. Journal ofSedimentary Petrology 59:1036–1051.

Tudhope, A. W., and T. P. Scoffin. 1984. The effects of Callianassabioturbation on the preservation of carbonate grains in DaviesReef lagoon, Great Barrier Reef, Australia. Journal of Sedi-mentary Petrology 54:1091–1096.

Verardo, D., P. Froelich, and A. McIntyre. 1990. Determinationof organic carbon and nitrogen in marine sediments using theCarlo-Erba NA-1500 Analyzer. Deep-sea Research 37:157–165.

Walter, L. M., and E. A. Burton. 1990. Dissolution of recent plat-form carbonate sediments in marine pore fluids. AmericanJournal of Science 290:601–643.

Warme, J. E. 1971. Paleoecological aspects of a modern coastallagoon. University of California Publications in GeologicalSciences 87:1–110.

Warme, J. E., A. A. Ekdale, S. F. Ekdale, and C. H. Peterson. 1976.Raw material of the fossil record. Pp. 143–69 in R. W. Scott andR. R. West, eds. Structure and classification of paleocommun-ities. Dowden, Hutchinson, and Ross, Stroudsburg, Penn.

Zuschin, M., and J. Hohenegger. 1998. Subtropical coral-reef as-sociated sedimentary facies characterized by molluscs (north-ern Bay of Safaga, Red Sea, Egypt). Facies 38:229–254.

Documents

Bivalve taphonomy in tropical mixed siliciclastic ...geosci.uchicago.edu/pdfs/kidwell/Best-Kidwell-2000-I.pdf · ical siliciclastic sediments. Focusing on unambiguous damage to bivalve