the role of fecundity and reproductive effort in defining

Biol. Rev. (2013), 88, pp. 745–766. 745doi: 10.1111/brv.12028

The role of fecundity and reproductive effortin defining life-history strategies of NorthAmerican freshwater mussels

Wendell R. Haag∗

US Forest Service, Center for Bottomland Hardwoods Research, Forest Hydrology Laboratory, Oxford, Mississippi, 38655, U.S.A.

ABSTRACT

Selection is expected to optimize reproductive investment resulting in characteristic trade-offs among traits such asbrood size, offspring size, somatic maintenance, and lifespan; relative patterns of energy allocation to these functionsare important in defining life-history strategies. Freshwater mussels are a diverse and imperiled component of aquaticecosystems, but little is known about their life-history strategies, particularly patterns of fecundity and reproductiveeffort. Because mussels have an unusual life cycle in which larvae (glochidia) are obligate parasites on fishes, differencesin host relationships are expected to influence patterns of reproductive output among species. I investigated fecundityand reproductive effort (RE) and their relationships to other life-history traits for a taxonomically broad cross sectionof North American mussel diversity. Annual fecundity of North American mussel species spans nearly four ordersof magnitude, ranging from < 2000 to 10 million, but most species have considerably lower fecundity than previousgeneralizations, which portrayed the group as having uniformly high fecundity (e.g. > 200000). Estimates of RE alsowere highly variable, ranging among species from 0.06 to 25.4%. Median fecundity and RE differed among phylogeneticgroups, but patterns for these two traits differed in several ways. For example, the tribe Anodontini had relatively lowmedian fecundity but had the highest RE of any group. Within and among species, body size was a strong predictor offecundity and explained a high percentage of variation in fecundity among species. Fecundity showed little relationshipto other life-history traits including glochidial size, lifespan, brooding strategies, or host strategies. The only apparenttrade-off evident among these traits was the extraordinarily high fecundity of Leptodea, Margaritifera, and Truncilla, whichmay come at a cost of greatly reduced glochidial size; there was no relationship between fecundity and glochidial sizefor the remaining 61 species in the dataset. In contrast to fecundity, RE showed evidence of a strong trade-off withlifespan, which was negatively related to RE. The raw number of glochidia produced may be determined primarily byphysical and energetic constraints rather than selection for optimal output based on differences in host strategies or othertraits. By integrating traits such as body size, glochidial size, and fecundity, RE appears more useful in defining mussellife-history strategies. Combined with trade-offs between other traits such as growth, lifespan, and age at maturity,differences in RE among species depict a broad continuum of divergent strategies ranging from strongly r-selectedspecies (e.g. tribe Anodontini and some Lampsilini) to K -selected species (e.g. tribes Pleurobemini and Quadrulini;family Margaritiferidae). Future studies of reproductive effort in an environmental and life-history context will be usefulfor understanding the explosive radiation of this group of animals in North America and will aid in the development ofeffective conservation strategies.

Key words: life history, trade-off, fecundity, reproduction, glochidia, unionidae, margaritiferidae, unionoida.

CONTENTS

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746II. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747

(1) The data set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747(2) Life-history variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753(3) Data analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753

* Author for correspondence (Tel: 662-234-2744; E-mail: [email protected]).

Biological Reviews 88 (2013) 745–766 Published 2013. This article is a U.S. Government work and is in the public domain in the USA.

746 Wendell R. Haag

III. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754(1) Overall patterns of fecundity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 754(2) Overall patterns of reproductive effort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 755(3) Relationships of fecundity and life-history variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 755(4) Relationships of reproductive effort and life-history variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 759(5) Models with multiple life-history variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 759

IV. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 760V. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 763

VI. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764VII. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764

I. INTRODUCTION

Fecundity is an important life-history trait and is considereda measure of an organism’s fitness (Charlesworth, 1994).When resources are limiting, selection is expected to optimizereproductive investment resulting in characteristic trade-offsamong brood size, offspring size, and reproductive lifespan.For example, high fecundity is possible often because ofreduced parental care or production of smaller offspringrequiring lower energetic investment (e.g. Winemiller &Rose, 1992). High investment in reproduction may comeat a cost to somatic maintenance, resulting in decreasedlifespan for highly fecund organisms. However, lifetimereproductive effort is predicted to be invariant of lifespanbecause of the longer reproductive life of species that investless in reproduction per unit time (White & Seymour, 2004;Charnov, Warne & Moses, 2007).

Freshwater mussels (order Unionoida) are a diverseand ecologically important group (Vaughn & Hakenkamp,2001; Gutierrez et al., 2003; Vaughn, Nichols & Spooner,2008). Mussels have an unusual life history in whichlarvae (glochidia) of most species are obligate parasiteson fishes. After hatching, glochidia are brooded by thefemale mussel, but brooding time varies among speciesgenerally corresponding to either a short-term or long-termbrooding strategy. Short-term brooders brood glochidia forapproximately 2–6 weeks before release usually in springor summer. Long-term brooders brood glochidia for up toeight months, usually from late summer or autumn, untilrelease the following spring or summer (e.g. Weaver, Pardue& Neves, 1991; Bruenderman & Neves, 1993; Garner,Haggerty & Modlin, 1999). The portion of the gills usedfor brooding also varies among species ranging from the useof all four gills to only a portion of the outer gills. Afterrelease from the female, glochidia must encounter a suitablehost species within a few days. Host use varies among speciesfrom generalists that are able to parasitize nearly any fishspecies, to specialists that can use as hosts only one or a fewclosely related fishes (Haag, 2012). Host use of specialistsencompasses a taxonomically and ecologically wide range offishes from drift-feeding minnows to large top predators, andone species is a specialist on an aquatic salamander (Necturusmaculosus). Accordingly, a wide array of strategies to transmitglochidia to hosts has evolved ranging from broadcast offree glochidia to complex strategies that target specific host

species (Barnhart, Haag & Roston, 2008; Haag, 2012). Inpart because of this complex life cycle, mussels are now oneof the most imperiled groups of organisms on Earth (Regnier,Fontaine & Bouchet, 2009).

As a group, freshwater mussels have been portrayedtraditionally as having uniformly high annual fecundity(> 200 000) to compensate for the low chances of glochidiaencountering hosts (McMahon & Bogan, 2001). However,fecundity is known for few species and relationships offecundity to other life-history traits have received littleattention. Recent studies show a much wider range infecundity including species that produce < 10000 glochidiaannually (Haag & Staton, 2003; Pilarczyk et al., 2005; Joneset al., 2010), and energetic investment in glochidia is lowin some species (Bauer, 1998). Bauer (1994) showed aninverse relationship between glochidial size and fecundityas predicted by life-history theory, but his analysis includedonly seven species. In addition to predictions of generallife-history theory, life-history traits specific to mussels alsomay be expected to influence fecundity. For example, if hostinfection strategies such as broadcasting are less efficient thanother, more specialized, strategies, fecundity of broadcastersshould be higher. Brooding glochidia may incur a cost tofemales because the presence of glochidia in the gills canreduce feeding and respiratory efficiency (Richard, Dietz &Silverman, 1991; Tankersley & Dimock 1993; Tankersley,1996), and the duration of the brooding period and portionof the gill used for brooding are therefore expected to beinversely related to fecundity. The wide range of variationamong species in fecundity, host strategies, and other traitssuch as growth and lifespan (Haag & Rypel, 2011) coupledwith the great taxonomic diversity of this group suggests thatmussels use a broad array of life-history strategies. Becauseof the central role of fecundity and reproductive effort in lifehistory, patterns of variation in these traits among species andtheir relationships to other life-history traits are importantfor understanding mussel ecology and developing effectiveconservation strategies for these imperiled animals.

I examined fecundity and reproductive effort in NorthAmerican freshwater mussels using original data and litera-ture reports. First, I examined patterns of variation at severallevels including within populations, among species, andamong phylogenetic groups. Second, I report relationships offecundity and reproductive effort with other life-history vari-ables including body size, glochidial size, lifespan, brooding


Fecundity and reproductive effort in mussels 747

traits, host use, and host-infection strategy. I then used mul-tiple regression to evaluate the relative importance of thesetraits in explaining variation in fecundity among species andto show patterns of covariation among traits. Finally, I dis-cuss the significance of fecundity and reproductive effort inlife-history evolution of freshwater mussels.

II. METHODS

(1) The data set

I compiled fecundity data for 72 mussel taxa (71 speciesplus 1 subspecies) from the literature, from original datagenerated by this study, and from unpublished data providedby colleagues (Table 1). These species represent a broadcross section of North American mussel diversity, containingmultiple representatives of all major phylogenetic lineages(sensu Graf, 2002; Campbell et al., 2005), all recognizedhost-attraction strategies (Barnhart et al., 2008), and an arrayof life-history strategies (see Haag, 2012). Several species arerepresented by estimates from more than one population,but for most, sample sizes were too small to examinestatistically differences in fecundity among populations, andno species are represented by estimates from across theirgeographic range.

I generated original fecundity estimates from preservedgravid female mussels following the methods of Haag &Staton (2003). I made estimates only from females whosegill marsupia appeared fully charged, in order to maximizethe likelihood that estimates represented the full reproductivecomplement of the individual and were not biased by femaleshaving previously released a portion of their brood. For eachspecimen, I measured total shell length (nearest 0.1 mm)with dial calipers. For 33 species, I also measured shellvolume (nearest 1.0 ml) to provide a surrogate measure ofbody mass. I measured shell volume by filling each shellvalve with water and recording the volume of water usinga graduated cylinder. I estimated fecundity by dissectinggravid gills, suspending glochidia in a known volume ofwater, and extrapolating total fecundity from the meannumber of glochidia in three replicate aliquots. I countedonly the number of developing eggs or glochidia in a sampleso that fecundity estimates reflected the number of viableoffspring in a brood and were not inflated by the presenceof structural or other non-developing eggs (see Barnhartet al., 2008). Most mussel species appear to produce onlya single brood per year, but multiple clutches have beenreported for a few species, primarily short-term brooders.However, most of these reports are from subtropical regionssouth of about 31◦N latitude, and the evidence for multiplebrood production is equivocal in most cases (Haag, 2012).Because all of the populations sampled in this study arefrom temperate latitudes, I considered estimates to representannual fecundity for an individual.

In addition to fecundity, I estimated reproductiveeffort (RE) for nine species (Table 2). Reproductiveeffort is a measure of the relative amount of energy

allocated to offspring production and can be calculatedas: RE = (offspring mass)/(offspring mass + adult bodymass + mass of organic component of the shell) (Bauer, 1998;Gosling, 2003). This estimate assumes one brood per yearand that the energetic cost of producing a mass of adult tissueis equivalent to the energetic cost of producing the same massof offspring tissue (Charnov et al., 2007). All individuals in thisanalysis were brooding fully mature glochidia, and I flushedglochidia from gravid gills into pre-weighed pans. Glochidia,soft tissues (including flushed gills), and shells were then driedat 60◦C for 48 h and weighed separately. I did not estimatedirectly the mass of the organic component of the shell. Shellsof adult mussels typically contain about 2.3–3.8% organicmatter (Cameron, Cameron & Paterson, 1979), and I usedthe midpoint of this range (3.1%) to estimate organic shellmass for all species. I also compiled other reported values ofRE from the literature. One disadvantage of this method isthat inorganic shell mass is included in estimates of glochidialmass but not in estimates of adult tissue mass, and resultantestimates of RE are likely inflated. Therefore, I also ashedadult soft tissue and glochidia at 500◦C in a muffle furnace,and then subtracted ash mass from dry mass to obtain ash-free dry mass, an estimate of the organic fraction of samples.I then calculated RE using estimates of organic mass.

Measurements of tissue and shell mass necessary forcalculating RE were unavailable for other species in thedataset. I used length versus body mass and length versus shellmass relationships to estimate mass (dry mass only) for anadditional 10 species (13 populations) with fecundity data(Table 2). In all cases, length-mass relationships explaineda high percentage of variation in body and shell mass(r2 = 0.85–0.98; see also Benke et al., 1999) and werederived from the same population as fecundity estimates.I estimated total glochidial volume for these species as(glochidia size)3 × fecundity (see Charnov et al., 2007),and converted these values to mass according to therelationship between volume and mass derived from thenine species with direct measurements of glochidial mass(log glochidial mass = 0.985(log glochidial volume) − 4.067;r2 = 0.817, P < 0.0001). Adult length and fecundity usedto predict mass were mean values for each population (seeTable 1), and glochidial size was determined from sourceslisted subsequently. In Fusconaia cerina and Pleurobema decisum,approximately 50% of eggs typically are unfertilized andprovide structure to the conglutinate (Haag & Staton, 2003;Barnhart et al., 2008; see Section II.2). These unfertilizedeggs were not included in fecundity estimates, but becausethey represent an energetic cost, I multiplied estimates ofglochidial mass by two for these species. I then used theseestimates of adult body mass, shell mass, and glochidial massto calculate RE for these 13 populations. I tested the accuracyof this method by estimating RE for eight of the species withdirect mass measurements. For all species except Leptodea

fragilis, direct and estimated values of RE were very similarand were strongly correlated (r = 0.941 without L. fragilis,r = 0.712 with L. fragilis; P < 0.0001 for both correlations;see Table 2).


748 Wendell R. HaagT

able

1.Fe

cund

ityda

tafo

rN

orth

Am

eric

anfr

eshw

ater

mus

sels

Spec

ies

Mea

nle

ngth

(mm

)M

ean

fecu

ndity

Fecu

ndity

rang

eN

Len

gth-

fecu

ndity

rela

tions

hip

Site

Sour

ce

Fam

ily

Mar

gari

tife

rid

aeC

umbe

rlan

dia

mon

odon

ta15

1.2

4967

500

1930

000

–95

7000

08

—G

asco

nade

and

Mer

amec

Riv

ers,

MO

Bai

rd(2

000)

Mar

gari

tife

ram

arga

ritife

ra77

.944

2168

810

3060

0–

1685

1000

8—

Riv

erD

ee,U

KY

oung

&W

illia

ms

(198

4)F

amil

yU

nio

nid

aeT

ribe

Am

blem

ini

Am

blem

apl

icat

a78

.532

5709

7605

0–

6813

0013

y=

0.60

0x3.

0275

,r2

=0.

879*

**L

ittle

Tal

laha

tchi

eR

iver

,M

SH

aag

&St

aton

(200

3)

Am

blem

apl

icat

a86

.922

9738

4050

0–

4653

0012

y=

0.00

9x3.

825 ,

r2=

0.83

9***

Sips

eyR

iver

,AL

Haa

g&

Stat

on(2

003)

Tri

beA

nodo

ntin

iA

lasm

idon

tahe

tero

don

29.0

2661

2067

–29

973

——

R.M

air

(unp

ublis

hed

data

)A

lasm

idon

tam

argi

nata

55.0

4729

842

559

–54

391

3—

—R

.Mai

r(u

npub

lishe

dda

ta)

Ala

smid

onta

viri

dis

35.6

6914

4432

–84

763

——

R.M

air

(unp

ublis

hed

data

)A

rcid

ens

conf

rago

sus

144.

010

5075

0—

1—

St.F

ranc

isR

iver

,AR

Thi

sst

udy

Las

mig

ona

com

plan

ata

158.

514

4950

010

4800

0–

1930

000

4—

St.F

ranc

isR

iver

,AR

Thi

sst

udy

Las

mig

ona

hols

toni

a38

.628

83—

5—

—R

.Mai

r(u

npub

lishe

dda

ta)

Peg

ias

fabu

la1

25.0

1950

1500

–24

00—

——

J.L

ayze

r(u

npub

lishe

dda

ta)

Pyg

anod

onca

tara

cta2

54.3

†53

811

——

y=

912.

011x

2.41

0 ,r2

=0.

397*

*L

ayto

n’s

Lak

e,N

SPa

ters

on&

Cam

eron

(198

5)

Pyg

anod

onca

tara

cta2

60.7

†17

482

——

y=

1.47

9x5.

200 ,

r2=

0.46

2**

Mor

ice

Lak

e,N

BPa

ters

on&

Cam

eron

(198

5)

Pyg

anod

ongr

andi

s10

7.6

4123

0069

300

–72

4050

9y=

0.00

2x4.

092

r2=

0.86

32**

*Fa

rmpo

nd,M

ST

his

stud

y

Pyg

anod

ongr

andi

s12

7.4

4896

0017

4000

–88

8000

10y=

0.22

0x2.

990

r2=

0.42

3**

Farm

pond

,MS

Thi

sst

udy

Pyg

anod

ongr

andi

s73

.558

453

2000

0–

2150

00†

63y=

0.00

006x

4.84

2 ††,

r2=

0.88

0**

Lon

gan

dN

arro

wL

akes

,A

BH

anso

n,M

acka

y&

Prep

as(1

989)

Sim

pson

aias

ambi

gua

42.3

5225

0—

1—

Wis

cons

inR

iver

,WI

Thi

sst

udy

Str

ophi

tus

subv

exus

55.6

2321

420

500

–25

927

2—

Bru

shy

Cre

ek,A

LT

his

stud

yU

tter

back

iaim

beci

llis

65.7

5026

513

500

–85

500

10y=

0.87

6x2.

600 ,

r2=

0.85

7***

Farm

pond

,MS

Thi

sst

udy

Utter

back

iaim

beci

llis

79.5

7779

725

500

–14

6500

8y=

1.42

9x2.

460 ,

r2=

0.27

1**

Pusk

usL

ake,

MS

Thi

sst

udy

Tri

beL

amps

ilini

Act

inon

aias

liga

men

tina

3—

—80

616

–15

6122

464

—G

reen

Riv

er,K

YM

oles

&L

ayze

r(2

008)

Act

inon

aias

liga

men

tina

128.

2†65

2843

4586

91–

8149

0417

y=

1.08

0x2.

743 ††

,r2

=0.

580*

**G

reen

Riv

er,K

YM

oles

&L

ayze

r(2

008)

Act

inon

aias

liga

men

tina

118.

9†81

6063

3109

48—

1170

959

22y=

4.44

0x2.

537 ††

,r2

=0.

640*

**G

reen

Riv

er,K

YM

oles

&L

ayze

r(2

008)

Cyp

roge

nia

aber

ti53

.013

2363

—7

—V

erdi

gris

Riv

er,K

SE

cker

t(20

03)

Cyp

roge

nia

aber

ti58

.063

182

—4

—St

.Fra

ncis

Riv

er,M

OE

cker

t(20

03)



Tab

le1.

(Con

t.)

Spec

ies

Mea

nle

ngth

(mm

)M

ean

fecu

ndity

Fecu

ndity

rang

eN

Len

gth-

fecu

ndity

rela

tions

hip

Site

Sour

ce

Cyp

roge

nia

aber

ti44

.069

634

—7

—O

uach

itaR

iver

,AR

Eck

ert(

2003

)C

ypro

geni

ast

egar

ia41

.943

487

2239

3–

6354

44

—C

linch

Riv

er,T

NJo

nes

&N

eves

(200

2)D

rom

usdr

omas

57.4

1083

81.6

4826

7–

2530

507

nsC

linch

Riv

er,T

NJo

nes

etal

.(20

04)

Ellip

sari

aline

olat

a56

.097

833

1400

0–

1635

0018

y=

0.07

1x3.

482,

r2=

0.81

0***

Ten

ness

eeR

iver

,AL

Thi

sst

udy

Epi

obla

sma

ahls

tedt

i45

.716

843

7066

–31

946

4—

Duc

kR

iver

,TN

J.W

.Jon

es(u

npub

lishe

dda

ta)

Epi

obla

sma

caps

aefo

rmis

441

.413

589

1049

4–

1687

69

y=

16.7

2x1.

798,

r2=

0.61

5*C

linch

Riv

er,T

NJ.

W.J

ones

(unp

ublis

hed

data

)

Epi

obla

sma

triq

uetr

a35

.315

511

1344

0–

2100

04

—B

ourb

euse

Riv

er,M

OB

arnh

art(

2002

)E

piob

lasm

aflo

rent

ina

aure

ola

29.6

8068

5818

–12

558

7ns

Clin

chR

iver

,VA

J.W

.Jon

es(u

npub

lishe

dda

ta)

Epi

obla

sma

flore

ntin

aw

alke

ri42

.05

7213

1828

–12

822

4—

Big

Sout

hFo

rkC

umbe

rlan

dR

iver

,KY

J.W

.Jon

es(u

npub

lishe

dda

ta)

Gle

bula

rotu

ndat

a510

6.4

5310

00—

5—

Bay

ouPe

yron

net,

LA

Park

er,H

ackn

ey&

Vid

rine

(198

4)H

amio

taau

stra

lis6

62.0

9180

0—

1—

Nat

ural

Bri

dge

Cre

ek,A

LB

lalo

ck-H

erod

,Her

od&

Will

iam

s(2

002)

Lam

psilis

orna

ta74

.228

1776

4862

5–

7396

0029

y=

0.83

7x2.

955

r2=

0.62

0***

Sips

eyR

iver

,AL

Haa

g&

Stat

on(2

003)

Lam

psilis

ovat

a13

6.3

7172

50—

1—

Ten

ness

eeR

iver

,AL

Thi

sst

udy

Lam

psilis

rafin

esqu

eana

113.

012

7400

011

7300

0–

1466

000

4—

Spri

ngR

iver

,KS

Bar

nhar

t(20

03)

Lam

psilis

siliqu

oide

a510

4.8

5330

00—

9—

Eas

tFor

kL

ittle

Mia

miR

iver

,O

HPe

rles

,Chr

istia

n&

Ber

g(2

003)

Lam

psilis

siliqu

oide

a68

.392

667

7400

0–

1250

003

——

Cok

eret

al.(

1921

)L

amps

ilis

stra

min

ea7

66.8

2684

6417

8000

–37

0000

7y=

1.03

4x2.

973,

r2=

0.81

2**

Sips

eyR

iver

,AL

Thi

sst

udy

Lam

psilis

tere

s96

.779

4667

6050

00–

9990

003

—L

ittle

Tal

laha

tchi

eR

iver

,MS

Thi

sst

udy

Lam

psilis

tere

s12

9.0

4631

633

1434

000

–84

3450

06

—St

.Fra

ncis

Riv

er,A

RT

his

stud

yL

amps

ilis

tere

s84

.438

6250

2940

00–

4785

002

—Si

psey

Riv

er,A

LT

his

stud

yA

llL

amps

ilis

tere

s11

2.1

2813

300

—10

y=

0.00

0010

x5.49

6

r2=

0.88

8***

—T

his

stud

y

Lem

iox

rim

osus

36.6

5306

4132

–64

792

—C

linch

Riv

er,T

NJo

nes

etal

.(20

10)

Lem

iox

rim

osus

36.6

2036

070

66–

5870

06

—D

uck

Riv

er,T

NJo

nes

etal

.(20

10)

Lep

tode

afr

agilis

79.9

9586

987

1571

250

–23

4015

0015

y=

5.86

4x3.

217,

r2=

0.77

0***

St.F

ranc

isR

iver

,AR

Thi

sst

udy

Lep

tode

afr

agilis

99.7

1073

5000

1070

7500

–10

7625

002

—L

ittle

Tal

laha

tchi

eR

iver

,MS

Thi

sst

udy

Lep

tode

afr

agilis

43.7

6882

50—

1—

Sips

eyR

iver

,AL

Thi

sst

udy

Lep

tode

afr

agilis

66.5

9080

00—

1—

Tow

nC

reek

,MS

Thi

sst

udy

Lep

tode

afr

agilis

74.0

2225

000

—1

—C

oker

etal

.(19

21)

Lep

tode

ale

ptod

on52

.210

7766

741

9000

–20

0000

03

—G

asco

nade

and

Mer

amec

Riv

ers,

MO

Bar

nhar

t(20

01,2

002,

2003

)

Lig

umia

rect

a15

1.6

1076

833

8955

00–

1226

500

3—

Ten

ness

eeR

iver

,AL

Thi

sst

udy


750 Wendell R. HaagT

able

1.(C

ont.)

Spec

ies

Mea

nle

ngth

(mm

)M

ean

fecu

ndity

Fecu

ndity

rang

eN

Len

gth-

fecu

ndity

rela

tions

hip

Site

Sour

ce

Lig

umia

rect

a11

9.6

6042

6928

3500

–10

9696

49

y=

0.00

6x3.

833 ,

r2=

0.72

7**

St.C

roix

Riv

er,M

NT

his

stud

y

Lig

umia

subr

ostr

ata

60.0

7482

135

500

–13

7500

7y=

11.3

77x2.

133

r2=

0.84

4**

Farm

pond

,MS

Thi

sst

udy

Med

ioni

dus

acut

issi

mus

21.7

1625

814

50–

3537

515

y=

0.00

3x4.

937

r2=

0.90

0***

Sips

eyR

iver

,AL

Thi

sst

udy

Med

ioni

dus

conr

adic

us8

33.2

1640

011

400

–21

400

7y=

6.91

9x2.

215 ,

r2=

0.77

9**

Clin

chR

iver

,TN

Thi

sst

udy

Obl

iqua

ria

refle

xa48

.840

975

447

–13

5750

21sq

rt(y

)=5.

238x

−53

.191

,r2

=0.

679*

**L

ittle

Tal

laha

tchi

eR

iver

,MS

Haa

g&

Stat

on(2

003)

Obl

iqua

ria

refle

xa40

.825

767

3250

–82

500

12sq

rt(y

)=5.

238x

−53

.191

,r2

=0.

679*

**Si

psey

Riv

er,A

LH

aag

&St

aton

(200

3)

Obo

vari

aun

icol

or9

32.5

5639

024

200

–10

5925

31y=

572.

160x

1.31

0

r2=

0.40

8***

Sips

eyR

iver

,AL

Thi

sst

udy

Pot

amilus

ohie

nsis

68.7

1062

768

—1

—St

.Fra

ncis

Riv

er,A

RT

his

stud

yP

otam

ilus

purp

urat

us10

4.5

4174

0775

000

–68

8750

10y=

0.00

2x4.

134

r2=

0.89

1***

St.F

ranc

isR

iver

,AR

Thi

sst

udy

Pot

amilus

purp

urat

us11

3.3

4255

0027

1000

–58

0000

2—

Litt

leT

alla

hatc

hie

Riv

er,M

ST

his

stud

yP

tych

obra

nchu

ssu

bten

tum

576

.124

7000

4330

0–

5020

0029

—C

linch

Riv

er,T

NV

.Men

gel&

J.L

ayze

r(u

npub

lishe

dda

ta)

Tox

olas

ma

texa

sens

is27

.620

089

8214

–31

964

2—

Qua

rter

liah

Cre

ek,M

ST

his

stud

yT

oxol

asm

ate

xase

nsis

35.4

3350

018

250

–48

750

2—

St.F

ranc

isR

iver

,AR

Thi

sst

udy

Tru

ncilla

dona

cifo

rmis

20.8

3519

00—

1—

St.F

ranc

isR

iver

,AR

Thi

sst

udy

Tru

ncilla

trun

cata

45.3

8838

8928

6000

–18

5100

05

—St

.Cro

ixR

iver

,MN

Thi

sst

udy

Ven

usta

conc

hael

lips

ifor

mis

62.4

1387

0012

3500

−161

500

5—

Pom

me

deT

erre

Riv

er,M

OT

his

stud

yV

enus

taco

ncha

plea

sii

37.3

4694

7—

1—

Jam

esR

iver

,MO

Bar

nhar

t(20

01)

Villo

sair

is41

.757

667

4850

0–

6775

03

—L

ittle

Riv

er,T

NT

his

stud

yV

illo

salien

osa

63.5

5255

00—

1—

Cat

alpa

Cre

ek,M

ST

his

stud

yV

illo

salien

osa

43.5

7302

532

600

–10

7100

4—

Sips

eyR

iver

,AL

Thi

sst

udy

Villo

sane

bulo

sa46

.514

0000

3975

0–

2402

502

—H

olly

Cre

ek,G

AT

his

stud

yV

illo

sava

nuxe

men

sis

42.7

5866

745

400

–68

000

3—

Hol

lyC

reek

,GA

Thi

sst

udy

Villo

sava

nuxe

men

sis

39.2

2627

517

875

–33

750

3—

Spri

ngC

reek

,TN

Thi

sst

udy

allV

illo

sa—

——

—y=

0.01

0x4.

169

r2=

0.79

4***

—T

his

stud

y

Tri

bePl

euro

bem

ini

Ellip

tio

arca

69.9

1362

2719

300

–20

6875

35y=

−100

.930

x2+

1709

3x–

5654

29,

r2=

0.64

5***

Sips

eyR

iver

,AL

Haa

g&

Stat

on(2

003)

Ellip

tio

com

plan

ata2

58.0

†16

668

1878

–68

655

23y=

12.6

18x4.

088 ††

,r2

=0.

563*

**M

oric

eL

ake,

NB

Pate

rson

(198

5)

Ellip

tio

cras

side

ns11

4.7

6512

50—

1—

Sips

eyR

iver

,AL

Thi

sst

udy

Ellip

tio

dila

tata

84.1

8108

948

393

–11

2589

5—

St.C

roix

Riv

er,M

NT

his

stud

y



Tab

le1.

(Con

t.)

Spec

ies

Mea

nle

ngth

(mm

)M

ean

fecu

ndity

Fecu

ndity

rang

eN

Len

gth-

fecu

ndity

rela

tions

hip

Site

Sour

ce

Fus

cona

iabu

rkei

44.1

6058

3880

–10

395

4—

Eig

htm

ileC

reek

,FL

Pila

rczy

ket

al.(

2005

)F

usco

naia

ceri

na46

.023

890

8750

–55

422

36y=

3.97

3x2.

273

r2=

0.47

6*Si

psey

Riv

er,A

LH

aag

&St

aton

(200

3)

Fus

cona

iacu

neol

us5

63.0

1130

00—

1—

Clin

chR

iver

,VA

Bru

ende

rman

&N

eves

(199

3)F

usco

naia

ozar

kens

is68

.915

840

8582

–29

470

8ns

Spri

ngR

iver

,MO

M.C

.Bar

nhar

t(un

publ

ishe

dda

ta)

Ple

urob

ema

deci

sum

46.7

2943

367

21–

7358

116

y=

0.65

3x2.

788,

r2=

0.76

54**

*Si

psey

Riv

er,A

LH

aag

&St

aton

(200

3)

Ple

urob

ema

deci

sum

47.9

4088

727

860

–69

553

10y=

30.3

13x1.

863,

r2=

0.42

9*Si

psey

Riv

er,A

LH

aag

&St

aton

(200

3)

Ple

urob

ema

collin

a49

.912

423

—8

—Jo

hns

and

Dic

ksC

reek

s,V

AH

ove

&N

eves

(199

4)P

leur

obem

ast

rode

anum

1043

.118

413

1032

0–

3973

39

nsE

ight

mile

Cre

ek,F

LPi

larc

zyk

etal

.(20

05)

Tri

beQ

uadr

ulin

iM

egal

onai

asne

rvos

a11

7.1

6074

5536

1000

–81

5250

11y=

156.

170x

1.73

2

r2=

0.79

9**

*T

enne

ssee

Riv

er,A

LT

his

stud

y

Qua

drul

aas

pera

ta40

.496

4725

0–

2825

074

sqrt

(y)=

3.79

6x−

55.0

40,

r2=

0.64

6***

Sips

eyR

iver

,AL

Haa

g&

Stat

on(2

003)

Qua

drul

acy

lind

rica

91.5

1150

00—

5—

Clin

chR

iver

,VA

Yea

ger

&N

eves

(198

6)Q

uadr

ula

pust

ulos

a53

.028

369

49–

5062

530

sqrt

(y)=

−0.0

99x2

+13

.758

x−

281.

810

r2=

0.64

6***

Litt

leT

alla

hatc

hie

Riv

er,M

SH

aag

&St

aton

(200

3)

Qua

drul

apu

stul

osa

59.9

3687

5—

1—

St.C

roix

Riv

er,M

NT

his

stud

yQ

uadr

ula

rum

phia

na11

50.7

3075

025

00–

6825

010

y=

(8×

10−1

6 )x11

.371

,r2

=0.

857*

*Si

psey

Riv

er,A

LT

his

stud

y

Uni

omer

uste

tral

asm

us55

.586

0654

20–

1072

54

—T

hom

pson

Cre

ek,M

ST

his

stud

yU

nion

idae

,inc

erta

ese

dis

Ple

ctom

erus

dom

beya

nus

100.

455

3500

1635

00–

9435

002

—L

eafR

iver

,MS

Thi

sst

udy

Fecu

ndity

refe

rshe

reto

asi

ngle

clut

chan

din

clud

esfe

rtile

eggs

only

;fo

rso

me

spec

ies,

repo

rted

estim

ates

have

been

adju

sted

toac

coun

tfo

rth

ech

arac

teri

stic

pres

ence

ofa

larg

enu

mbe

rof

unfe

rtili

zed

eggs

(e.g

.Fus

cona

iasp

p.,P

leur

obem

asp

p.;s

eeH

aag

&St

aton

,200

3).M

ultip

leva

lues

for

asp

ecie

sre

pres

entd

iffer

entp

opul

atio

ns.F

orle

ngth

(x)v

ersu

sfe

cund

ity(y

)re

latio

nshi

ps,l

engt

his

inm

mun

less

note

dot

herw

ise.

—=

notr

epor

ted

orno

tava

ilabl

e.*P

<0.

05;*

*P<

0.01

;***

P<

0.00

1;ns

=no

tsig

nific

ant.

†Val

ues

estim

ated

from

grap

hica

lpr

esen

tatio

nof

data

.††E

quat

ion

give

nor

igin

ally

inlin

ear

form

usin

glo

g-tr

ansf

orm

edda

ta.

1 Len

gth

not

avai

labl

e;st

ated

leng

this

aty

pica

lval

uefo

rth

esp

ecie

s;2 le

ngth

-fec

undi

tyre

latio

nshi

pba

sed

onle

ngth

ince

ntim

eter

s;3 ra

nge

acro

ssth

ree

popu

latio

ns;

4 excl

udes

one

outli

er:7

780

(42.

0m

m);

5 mea

nle

ngth

not

give

n,va

lue

ism

idpo

int

ofre

port

edle

ngth

rang

e;6 fe

cund

ityre

port

edas

anes

timat

edra

nge

(612

00–

1224

00),

valu

eis

mid

poin

tof

rang

e;7 ex

clud

eson

eou

tlier

:203

000

(73.

3m

m);

8 estim

ates

excl

ude

one

outli

er:6

200

(34.

6m

m);

9 estim

ates

excl

ude

two

outli

ers:

1970

00(3

9.9

mm

leng

th)a

nd17

5000

(32.

3m

m);

10ex

clud

estw

oou

tlier

s:86

0(4

1.1

mm

leng

th)a

nd33

0(4

6.2

mm

);11

excl

udes

two

outli

ers:

6000

(53.

2m

m)a

nd10

750

(60.

0m

m).


752 Wendell R. Haag

Tab

le2.

Rep

rodu

ctiv

eef

fort

(RE

)in

fres

hwat

erm

usse

ls

Spec

ies

Site

N

Obs

erve

dR

E(%

)dr

ym

ass

Est

imat

edR

E(%

)dr

ym

ass

Obs

erve

dR

E(%

)as

h-fr

ee

Ash

(%)

adul

t(w

/osh

ell)

Ash

(%)

gloc

hidi

a(w

/she

ll)

Fam

ilyM

arga

ritif

erid

aeM

arga

ritife

ram

arga

ritife

ra1

Cen

tral

Eur

ope

—0.

8–

5.3

——

——

Fam

ilyU

nion

idae

Tri

beA

mbl

emin

iA

mbl

ema

plic

ata

Litt

leT

alla

hatc

hie

Riv

er,M

S—

—3.

8—

——

Am

blem

apl

icat

aSi

psey

Riv

er,A

L—

—1.

6—

——

Tri

beA

nodo

ntin

iA

nodo

nta

anat

ina2

Swed

en—

15–

45—

——

—P

ygan

odon

cata

ract

a3M

oric

eL

ake,

NB

——

10.8

——

—P

ygan

odon

cata

ract

a3L

ayto

n’s

Lak

e,N

B—

—19

.9—

——

Pyg

anod

ongr

andi

sH

atch

ery

pond

,AL

3525

.4(1

.8)

——

——

Utter

back

iaim

beci

llis

Pusk

usL

ake,

MS

814

.3(1

.6)

12.0

——

—

Tri

beL

amps

ilini

Lam

psilis

orna

taSi

psey

Riv

er,A

L4

7.3

(1.0

)6.

4—

——

Lam

psilis

stra

min

eaSi

psey

Riv

er,A

L6

8.5

(0.5

)8.

23.

4(0

.6)

20.0

(1.2

)66

.5(6

.4)

Lam

psilis

tere

sSi

psey

Riv

er,A

L2

8.9

(0.7

)8.

75.

1(0

.6)

17.5

(1.3

)53

.1(1

.7)

Lep

tode

afr

agilis

Sips

eyR

iver

,AL

225

.4(8

.7)

8.5

——

—M

edio

nidu

sac

utis

sim

usSi

psey

Riv

er,A

L7

15.6

(0.9

)17

.68.

0(0

.6)

18.4

(1.4

)58

.9(4

.1)

Obl

iqua

ria

refle

xaL

ittle

Tal

laha

tchi

eR

iver

,MS

——

1.7

——

—O

bliq

uari

are

flexa

Sips

eyR

iver

,AL

——

1.8

——

—O

bova

ria

unic

olor

Sips

eyR

iver

,AL

245.

5(0

.3)

5.0

2.0

(0.2

)22

.7(1

.0)

67.7

(3.2

)T

runc

illa

dona

cifo

rmis

Litt

leT

alla

hatc

hie

Riv

er,M

S—

—7.

5—

——

Villo

salien

osa

Sips

eyR

iver

,AL

49.

3(0

.3)

10.8

3.8

(1.0

)32

.7(1

4.8)

63.2

(9.1

)

Tri

bePl

euro

bem

ini

Ellip

tio

arca

Sips

eyR

iver

,AL

——

6.2

——

—F

usco

naia

ceri

naSi

psey

Riv

er,A

L—

—0.

9—

——

Ple

urob

ema

deci

sum

Sips

eyR

iver

,AL

——

3.4

——

—

Tri

beQ

uadr

ulin

iQ

uadr

ula

aspe

rata

Sips

eyR

iver

,AL

——

1.0

——

—Q

uadr

ula

pust

ulos

aL

ittle

Tal

laha

tchi

eR

iver

,MS

——

1.1

——

—Q

uadr

ula

rum

phia

naSi

psey

Riv

er,A

L—

—0.

1—

——

Obs

erve

dva

lues

ofR

Ear

efr

omth

isst

udy

exce

ptw

here

othe

rwis

eno

ted.

Sam

ple

size

sin

this

anal

ysis

diffe

rfr

omth

ose

inT

able

1be

caus

em

ass

was

notd

eter

min

edfo

ral

lind

ivid

uals

with

fecu

ndity

estim

ates

.Est

imat

edR

Ew

asba

sed

onle

ngth

vers

usdr

ym

ass

rela

tions

hips

(W.R

.Haa

g,un

publ

ishe

dda

ta,e

xcep

twhe

reot

herw

ise

note

d)us

ing

estim

ates

ofm

ean

adul

tle

ngth

and

fecu

ndity

for

each

popu

latio

n(T

able

1;se

ete

xt).

Val

uesi

npa

rent

hese

sare

stan

dard

erro

rs.—

deno

test

hatv

alue

seith

erw

ere

notd

eter

min

ed,n

otre

port

edin

the

sour

ce,o

rar

eno

tapp

licab

lebe

caus

ees

timat

edR

Eis

base

don

mea

nva

lues

ofm

ass

and

fecu

ndity

for

the

popu

latio

n.O

ther

sour

ces:

1 Bau

er(1

998)

;2 Hau

kioj

a&

Hak

ala

(197

8),c

ited

inB

auer

(199

8);3 C

amer

onet

al.(

1979

;len

gth

–dr

ym

ass

rela

tions

hips

).



(2) Life-history variables

I examined variation in fecundity among mussel species withregard to seven life-history variables: body size, offspring(glochidia) size, lifespan, brooding period, area of the gillused for brooding, host-infection strategy, and host use.When possible, I used estimates of these variables specific tothe same population from which fecundity estimates weremade, but estimates from other populations were used insome cases. Sources and methods for determining life-historyvariables are described below.

Body size: determined as the mean shell length (mm) ofindividuals from which fecundity estimates were made asreported in Table 1. I also used shell volume when availableas a surrogate measure of body mass.

Glochidia size: defined as the mean of glochidial shell lengthand height (mm) following Barnhart et al. (2008). Data onglochidia size were obtained from Hoggarth (1999), Kennedy& Haag (2005), and Barnhart et al. (2008), and were availablefor 67 species in the dataset.

Lifespan: determined as the maximum age (years) reportedfor a species. Information was obtained primarily from Haag& Rypel (2011), but also van der Schalie & van der Schalie(1963), Harmon & Joy (1990), Barnhart (2001), Rogers,Watson & Neves (2001), Fobian (2007), Jones et al. (2010)and Jones & Neves (2011). If estimates of lifespan wereavailable for more than one population of a species, I usedthe maximum reported value. Estimates of lifespan wereavailable for 51 species in the dataset.

Brooding period: classified as short-term or long-termbrooders according to information summarized in Williams,Bogan & Garner (2008). Brooding period was available forall species.

Gill brooding area: determined as the proportion of the gillsused for brooding glochidia. For example, species that broodglochidia throughout the length of all four gills were assigneda value of 1.00, and species that brood only in the posteriorhalf of the outer gills were assigned a value of 0.25. Cyprogeniaspp. and Obliquaria reflexa brood only in a few water tubes inthe outer gills and were assigned a value of 0.17. Broodingarea was determined mostly from information summarizedin Williams et al. (2008) and was available for all species.Because gill area had only four values (0.17, 0.25, 0.50, and1.00), I treated this as a nominal variable in data analysis.

Host infection strategy: determined as one of five strategiesused by mussels to transmit glochidia to hosts, includingbroadcasting, mucus webs, mantle lures, conglutinates, andmantle magazines (see Barnhart et al., 2008). Broadcastingincludes species that release free glochidia, which encounterhosts apparently by chance. Release of glochidia in mucuswebs is similar to broadcasting, but glochidia are bound ina mucus matrix that entangles potential host fishes. Mantlelures are modified portions of the female mantle that attractpotential fish hosts to the mussel by mimicking prey items offishes; glochidia are released when attacking fishes rupturethe gravid gill which is also displayed as part of the lure.Conglutinates are discrete packets of glochidia that mimicprey items of fishes; after release from the female mussel,

conglutinates are ingested by fishes facilitating transmittalof glochidia. The mantle magazine strategy involves short-term storage of glochidia within a chamber associated withthe excurrent aperture; glochidia are discharged rapidlyfrom the chamber apparently upon tactile, photosensory,or chemosensory stimulation by a host fish. Host infectionstrategies were determined for 58 species based on Barnhartet al. (2008). Host infection strategies are well-described formost species in the data set except for a few species classifiedas broadcasters by default because of the apparent absenceof other adaptations for host infection (e.g. Leptodea, Potamilus,Truncilla; see Section IV).

Host use: determined as the primary hosts for a musselspecies (see Haag & Warren, 1997; O’Brien & Williams,2002). To facilitate broad comparisons in host use, I catego-rized host use into guilds according to the highest taxonomiclevel that encompassed the range of fishes or amphibians usedby a particular mussel species. Host guilds included fresh-water drum (Aplodinotus grunniens), black basses (Micropterusspp.), sunfishes (Lepomis spp.), darters (Percidae), sculpins(Cottus spp.), catfishes (Ictaluridae), minnows (Cyprinidae),salmonids (Salmonidae), gar (Lepisosteidae), skipjack herring(Alosa chrysochloris), mudpuppy (Necturus maculosus), sauger andwalleye (Sander spp.), and generalists that do not specializeon a particular host group. I obtained host information froma large number of studies summarized in the Mussel/hostdatabase, The Ohio State University Museum of BiologicalDiversity, Division of Mollusks, http://www.biosci.ohio-state.edu/˜molluscs/OSUM2/index.htm.

(3) Data analysis

For original data and datasets provided by colleagues, Itested for length-fecundity relationships for populations withsample sizes exceeding seven individuals. When samplesizes were sufficient, I generated separate length-fecundityrelationships for species represented in the dataset by morethan one population. Because mussel fecundity typically isrelated to length by a power function (e.g. Paterson, 1985;Haag & Staton 2003; Moles & Layzer, 2008), I fitted powerfunctions for all species, but also evaluated alternative modelsby examining residuals and coefficients of determination. Iomitted a small number of outlier observations for somespecies if these values deviated widely from length–fecundityrelationships. In most cases, outliers were exceptionally smallnumbers suggesting that these females had already released aportion of their brood; all omitted outliers are reported in thefootnotes to Table 1. For species with shell volume data andsufficient sample sizes, I also tested for relationships betweenfecundity and volume. Although the dataset included two ormore populations for some species, small sample sizes did notallow me to test for differences in length-specific fecundityamong any populations.

I examined bivariate relationships between fecundity andlife-history variables using linear regression and analysis ofvariance with all continuous variables log10-transformed.For these analyses, I used mean fecundity, shell length, andshell volume for species represented in the dataset by more


754 Wendell R. Haag

than one estimate. Because shell length was strongly relatedto fecundity, I accounted for the influence of length onrelationships with other life-history variables either by usinglength-standardized fecundity (obtained as the residuals offecundity regressed on length) or with analysis of covariance.I also examined relationships between reproductive effort(RE) and life-history variables (shell length, glochidial size,lifespan, brooding period, and fecundity). For these analyses,life-history variables were specific to the population forwhich RE was estimated except for M. margaritifera; forthis species, I used data from other central Europeanpopulations (see Table 2). I did not examine relationshipsbetween RE and host strategies because of the smallernumber of species with estimates of RE. In addition toestimates of annual fecundity presented in Table 1, I wasinterested in how lifetime fecundity was related to life-history variables, especially lifespan. In organisms withindeterminate growth, lifetime fecundity can be calculatedby summing mean fecundity of each age class. However,because age-fecundity relationships were available for fewspecies in the dataset, I estimated lifetime fecundity asthe product of mean fecundity and lifespan, assuming onebrood per year. I evaluated the accuracy of this approachin three species for which complete age-specific fecundityinformation was available (Haag & Staton, 2003). Estimateswere similar for both methods for all three species (Elliptio arca:summed age-specific fecundity = 4077374, mean fecundityx lifespan = 4086810; Lampsilis ornata: 6220106, 5071968;Quadrula pustulosa: 1433387, 1361712).

I examined the relationship between annual fecundityand suites of life-history variables simultaneously usingmultiple regression. I used this approach to provide asimple comparison of the relative amount of variation infecundity explained by various factors and not to develop apredictive model for estimating fecundity. Further, to avoidproblems with stepwise procedures, I analyzed only the fullmodels including all effects (e.g. Whittingham et al., 2006).In this analysis, I used as independent variables shell length,glochidial size, lifespan, brooding period, and host-infectionstrategy. I did not include host use in this analysis because it ishighly associated with host-infection strategy; host-attractionstrategies typically target specific fish feeding or behaviouralguilds (Barnhart et al., 2008; Haag, 2012). The categoricalvariables brooding period and host-infection strategy alsocovary to a considerable extent. For example, all specieswith mantle lures are long-term brooders, but long-termbrooders are also present among other infection strategies.In addition, brooding was weakly correlated with glochidialsize (r = 0.306, P < 0.0121; on average, larger glochidiain long-term brooders) and lifespan (r = 0.654, P < 0.0001;on average, shorter lifespan in long-term brooders), butnot with shell length; despite these correlations, glochidialsize and lifespan overlapped widely among broodingperiods. There was a weak relationship between shelllength and lifespan (r = 0.281, P < 0.0464), but the othercontinuous independent variables were not correlated witheach other.

0.00

0.05

0.10

0.15

0.20

0.25

0.30

2500

075

000

1250

0017

5000

2250

0027

5000

3250

0037

5000

4250

004 7

5000

5250

0057

5000

6250

0067

5000

7250

0077

5000

8250

008 7

5000

9250

009 7

5000

>100

0000

Pro

port

ion

of s

peci

es

Fecundity (bin size = 25,000)

Fig. 1. Distribution of average fecundity values for NorthAmerican freshwater mussel species (N = 71 species plus onesubspecies).

III. RESULTS

(1) Overall patterns of fecundity

Fecundity varied widely among mussel species and spannednearly 4 orders of magnitude (Table 1). Mean population-level species fecundities ranged from 1950 (Pegias fabula) to10735000 (Leptodea fragilis). Fecundity was even more variableamong individuals and ranged from 49 for a small specimenof Quadrula pustulosa to 23401500 for a large adult Leptodeafragilis. Fecundity values showed an apparently multimodaldistribution (Fig. 1). Species with relatively low fecundity(< 25000) composed the largest percentage (30%), and 65%of species had fecundity < 150000; 14% of species had veryhigh fecundity (> 1000000), and about 20% had fecunditybetween 250000–950000.

Among populations of a species, mean fecundity was gen-erally similar and well within the same order of magnitude.In many cases where populations appeared to vary morewidely, these differences were due to samples containingindividuals that differed widely in length (e.g. Lampsilis teres,Ligumia recta, Leptodea fragilis). For example, mean fecundity ofPyganodon grandis from two Mississippi populations was morethan seven times higher than a population from Alberta,but mean length of these populations also differed widely,and the predicted fecundity of a 90 mm individual wassimilar in all three populations (Mississippi, 198514, 153323;Alberta, 174018; see Hanson et al., 1989). However, somepopulations did appear to show substantial differences inlength-specific fecundity. Predicted fecundity of a 55 mmPyganodon cataracta from Layton’s Lake, Nova Scotia (55498),was more than five times higher than a similarly sizedindividual from nearby Morice Lake, New Brunswick(10468), and the slopes of these relationships also differedwidely (Table 1; see Paterson & Cameron, 1985). Slopesof length-fecundity relationships were similar between twopopulations of Actinonaias ligamentina in the Green River,Kentucky, but length-specific fecundity was over 1.5 timeshigher in one population (see Moles & Layzer, 2008).

Within a population, fecundity varied among individualsby 0–2 orders of magnitude and this variation was strongly



related to female size (Table 1). Shell length was asignificant predictor of fecundity for all populations withdatasets of seven or more individuals with the exceptionof Epioblasma florentina aureola (P < 0.4373, N = 7), Fusconaia

ozarkensis (P < 0.4468, N = 8), and Pleurobema strodeanum

(P < 0.0532, N = 9). Length explained a high percentageof variation in fecundity in most populations, even thosewith small sample sizes (r2 = 0.40–0.90; Table 1). Length-fecundity relationships for most species were best describedby power functions. Values of the allometric constant b

ranged from 1.3 to 5.5 and averaged 3.2 (median value3.0). The value of b for Quadrula rumphiana departed widelyfrom this range (11.4). Length-fecundity relationships forfive species were best described by equations of other formsincluding second-order polynomials (Elliptio arca) and square-root functions (Obliquaria reflexa, Quadrula asperata, and Q.

pustulosa) (see Haag & Staton, 2003). The length-fecundityrelationship for Obovaria unicolor was described equally wellby a power function (r2 = 0.408, see Table 1) and a linearfunction (r2 = 0.441, P < 0.001; fecundity = 2491.6[length]− 24031); however, the 95% confidence interval for b in thepower function contained 1, providing little support for acurvilinear relationship in this species.

Shell volume and body mass were generally equivalentpredictors of fecundity compared with length (Table 3).For 8 of 11 populations, volume explained less variationin fecundity than length, but r2 values were similar.For nine populations, volume-fecundity relationships werebest described by power functions, which explained morevariation in fecundity than linear functions. However, inmost cases, the 95% confidence interval for b contained 1,providing little support for a curvilinear relationship betweenvolume and fecundity. Female body mass also was a nearlyequivalent predictor of fecundity compared with length. Therelationship between mass and fecundity was best describedby a linear function for two species and by a power functionfor two others; however, in the latter cases, 95% confidenceintervals for b contained 1.

Fecundity differed among phylogenetic groups, but washighly variable within most groups (Fig. 2). The Margar-itiferidae had by far the greatest median fecundity of anygroup, and the two species in the dataset had similarly highfecundity (Cumberlandia monodonta and Margaritifera margari-

tifera). The Amblemini and Lampsilini had moderately highmedian fecundity, but fecundity of the Lampsilini spannedan enormous range. The Lampsilini included the highestobserved fecundity of any species (Leptodea fragilis) and severalother species with fecundity > 500000 (Actinonaias ligamentina,

Glebula rotundata, Lampsilis spp., Leptodea leptodon, Ligumia recta,

Potamilus ohiensis, and Truncilla truncata), as well as specieswith fecundities < 20000 (Epioblasma spp., Lemiox rimosus,and Medionidus spp.). The Anodontini, Pleurobemini, andQuadrulini all had similar, low median fecundities, but eachgroup had conspicuous outliers with high fecundity (Anodon-tini: Arcidens confragosus, Lasmigona complanata, Pyganodon grandis;Pleurobemini: Elliptio crassidens; Quadrulini: Megalonaias

nervosa). Species with very low fecundity (< 10000) occurred

in all phylogenetic groups except the Margaritiferidae andAmblemini (Anodontini: Pegias fabula, Alasmidonta heterodon,A. viridis, Lasmigona holstonia; Lampsilini: Epioblasma florentinaaureola, E. florentina walkeri; Pleurobemini: Fusconaia burkei,Quadrulini: Uniomerus tetralasmus, Quadrula asperata).

Fecundity standardized by length showed a similar patternamong phylogenetic groups (Fig. 2). Length-standardizedfecundity was highest on average for the Margaritiferidae,intermediate for the Amblemini and Lampsilini, and lowestfor the Anodontini, Pleurobemini, and Quadrulini. However,length-specific fecundity was less variable than total fecundityin the Anodontini, Pleurobemini, and Quadrulini and wasuniformly low compared to other groups (studentized resid-uals for most species < 0). Similar to total fecundity, length-specific fecundity was highly variable within the Lampsiliniand included the highest values for any species (Truncillaspp., Leptodea spp.), as well as other species with high valuesdespite their low-to-moderate total fecundity (e.g. Lampsilisspp., Medionidus acutissimus, Obovaria unicolor, Villosa spp.).

(2) Overall patterns of reproductive effort

Direct estimates of reproductive effort (RE) based on drymass ranged among nine species from 5.5 to 25.4%(Table 2). Estimates of RE based on ash-free dry masswere roughly half of those based on dry mass. This wasdue to a much higher percentage of inorganic material(ash) in glochidia compared with adult soft tissues. Onlythree species had sufficient sample sizes to test relationshipsbetween size and RE within a population. For Obovariaunicolor, dry mass RE was negatively related to length butexplained little of the variation in RE (r2 = 0.328, P < 0.004;log length = −19.482[arcsine(RE/100)] + 42.438). Ash-freeRE of O. unicolor and dry mass RE of Pyganodon grandis andUtterbackia imbecillis were not related to length (P < 0.210).Estimated RE ranged from 0.06 to 19.9%. Estimates of REdiffered substantially between populations of Amblema plicataand Pyganodon cataracta but were similar between populationsof Obliquaria reflexa.

Patterns of variation in RE among phylogenetic groupsdeparted in several ways from those for fecundity (Fig. 2).Despite their high fecundity, RE of Margaritifera margaritifera(Margaritiferidae) and Amblemini was low. Conversely,although median fecundity of Anodontini was relatively low,their reproductive effort was the highest of any group. Similarto fecundity, median RE of Lampsilini was moderately highbut was highly variable among species, and RE was uniformlylow for the Pleurobemini and Quadrulini.

(3) Relationships of fecundity and life-historyvariables

Across species, shell length was a strong, significant predictorof fecundity and explained just over half of the variationamong species [log fecundity = 2.808(log length) 0.055;P < 0.0001, r2 = 0.513]. This relationship was influencedby five outliers representing species with very high fecundity(Leptodea fragilis, L. leptodon, Margaritifera margaritifera, Truncilla


756 Wendell R. Haag

Tab

le3.

Com

pari

son

ofsh

elll

engt

h,vo

lum

e,an

dbo

dym

ass

for

expl

aini

ngva

riat

ion

infe

cund

ityam

ong

indi

vidu

als

infr

eshw

ater

mus

selp

opul

atio

ns

Spec

ies

Popu

latio

nV

aria

ble

r2a

bN

Am

blem

apl

icat

aL

ittle

Tal

laha

tchi

eR

iver

,MS

Len

gth

0.82

80.

006

3.92

110

Vol

ume

0.80

517

701.

246†

10A

.pl

icat

aSi

psey

Riv

er,A

LL

engt

h0.

879

0.60

03.

027

13V

olum

e0.

669

9323

0.92

1†13

Ellip

sari

aline

olat

aT

enne

ssee

Riv

er,A

LL

engt

h0.

810

0.07

13.

482

18V

olum

e0.

762

3385

1.17

0†18

Lam

psilis

stra

min

eaSi

psey

Riv

er,A

LL

engt

h0.

825

0.26

83.

300

6V

olum

e*0.

701

8780

261

326

Mas

s*0.

752

4961

472

0841

6L

epto

dea

frag

ilis

St.F

ranc

esR

iver

,AR

Len

gth

0.77

05.

864

3.21

715

Vol

ume

0.69

821

2253

0.98

4†15

Med

ioni

dus

acut

issi

mus

Sips

eyR

iver

,AL

Len

gth

0.93

90.

037

4.09

58

Vol

ume

0.86

619

201

0.89

2†8

Mas

s0.

947

4732

651.

125†

8M

egal

onai

asne

rvos

aT

enne

ssee

Riv

er,A

LL

engt

h0.

799

156.

170

1.73

211

Vol

ume

0.73

191

810.

888†

11O

bova

ria

unic

olor

Sips

eyR

iver

,AL

Len

gth

0.40

857

2.16

01.

310†

24V

olum

e*0.

469

2840

963

2624

Mas

s*0.

3873

2847

073

026

24P

ygan

odon

gran

dis

Farm

pond

,MS

Len

gth

0.42

30.

220

2.99

010

Vol

ume

0.43

331

090.

952†

10P

otam

ilus

purp

urat

usSt

.Fra

ncis

Riv

er,A

RL

engt

h0.

891

0.00

24.

134

10V

olum

e0.

874

441

1.46

010

Utter

back

iaim

beci

llis

Pusk

usL

ake,

MS

Len

gth

0.27

11.

429

2.46

08

Vol

ume

0.38

241

900.

824†

7M

ass

0.24

769

744

0.94

4†8

aan

db

are

fitte

dco

nsta

ntso

fthe

allo

met

ric

equa

tion:

fecu

ndity

=a(

leng

th,v

olum

e,or

mas

s)b,e

xcep

tfor

rela

tions

hips

deno

ted

by*,

whi

char

elin

ear:

(fecu

ndity

=a+

b(le

ngth

,vol

ume,

orm

ass).

All

rela

tions

hips

are

sign

ifica

ntat

P<

0.00

1ex

cept

for

P.

gran

dis,

volu

me

(P<

0.03

).In

som

eca

ses,

cons

tant

sfo

rle

ngth

rela

tions

hips

diffe

rfr

omth

ose

pres

ente

din

Tab

le1

beca

use

mas

san

dsh

ellv

olum

ew

asno

tdet

erm

ined

for

alli

ndiv

idua

lsw

ithfe

cund

ityes

timat

es.†

deno

tes

the

95%

confi

denc

ein

terv

alfo

rth

eco

nsta

ntco

ntai

ns1.



35647

108382

277742

23890 31686

0

100000

200000

300000

400000

4694594

500000

2500000

4500000

6500000

8500000 (A)T

otal

fecu

ndity

-2

-1

0

1

2

3

4

Leng

th-s

tand

ardi

zed

fecu

ndity

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Rep

rodu

ctiv

e ef

fort

(%

)

Margaritiferidae

Amblemini

Lampsilini

Anodontini

Pleurobemini

Quadrulini

(B)

(C)

Fig. 2. Phylogenetic patterns of fecundity and reproductiveeffort for North American freshwater mussel species. (A) Totalannual fecundity (N = 70 species plus 1 subspecies, excludingPlectomerus dombeyanus, which is of uncertain phylogenetic affinity,see Campbell et al., 2005). Numbers and horizontal barsare median fecundity for each group; note change in scalebetween 400000 and 500000. (B) Length-standardized fecundityobtained as the residuals of fecundity regressed on length (bothvariables log-transformed). Horizontal bars are median residualsfor each group. (C) Reproductive effort for 20 species. Horizontalbars are median values.

donaciformis, and T. truncata; studentized residuals ≥ 2.0). Afteromitting these species from the analysis, the proportionof variation in fecundity explained by length increaseddramatically (Fig. 3). Other species with high length-specificfecundities but not identified as significant outliers wereCumberlandia monodonta, Lampsilis teres, Medionidus acutissimus,Obovaria unicolor, Potamilus ohiensis, and Villosa nebulosa. Shellvolume was nearly equivalent to length for predictingfecundity. For 33 species with volume data (excluding

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

1.2 1.4 1.6 1.8 2.0 2.2

Log

fecu

ndity

Log shell length (mm)

Fig. 3. Relationship between fecundity and shell length forNorth American freshwater mussels (N = 71 species plus onesubspecies). Open circles were identified as significant outliersbased on very high length-specific fecundity (see text). Theplotted regression line is for the dataset excluding theseoutliers [log fecundity = 3.146(log length) − 0.672; P < 0.0001,r2 = 0.751].

the outliers L. fragilis and Truncilla spp.), length explained69.8% of the variation in fecundity [log fecundity = 2.739(logshell volume) + 0.109; P < 0.0001], and volume explained68.1% [log fecundity = 0.958(log shell volume) + 3.766;P < 0.0001]. Among species, shell volume was closely relatedto length [log shell volume = 2.716(log shell length) − 3.551;r2 = 0.940, P < 0.0001].

Glochidial size was significantly and negatively relatedto fecundity, but glochidial size explained little of thevariation in fecundity among species [Fig. 4; log fecun-dity = −1.6984(log glochidial size) + 8.9575; P < 0.0008,r2 = 0.163, N = 67]. This relationship was driven by a groupof six species with small glochidia and very high fecundity;this group included the five high-fecundity species identifiedpreviously as outliers and also Cumberlandia monodonta.Another species with small glochidia, Quadrula rumphiana, hadlow fecundity. There was a significant interaction between

3.0

4.0

5.0

6.0

7.0

8.0

1.7 1.9 2.1 2.3 2.5 2.7

Log

fecu

ndity

Log glochidial size (µm)

Cumberlandia monodonta,Margaritifera margaritifera, Leptodea spp., Truncilla spp.

Quadrularumphiana

Fig. 4. Relationship between offspring (glochidia) size andfecundity in North American freshwater mussels. Points withinthe ellipse are six species with small glochidia and high fecundity(see text).


758 Wendell R. Haag

3.0

4.0

5.0

6.0

7.0

0.5 1.0 1.5 2.0

Log

annu

al fe

cund

ity

Log lifespan (years)

y = 1.193x + 4.863r 2 = 0.167, P < 0.003

3.0

4.0

5.0

6.0

7.0

8.0

9.0

0.5 1.0 1.5 2.0

Log

lifet

ime

fecu

ndity


y = 1.032x + 1.098r 2 = 0.882, P < 0.001

4.0

5.0

6.0

7.0

8.0

9.0

3.0 4.0 5.0 6.0 7.0

Log

lifet

ime

fecu

ndity

Log annual fecundity

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

0.5 1.0 1.5 2.0

Leng

th-s

tand

ardi

zed

annu

alfe

cund

ity


Leptodea spp., Truncilla spp.

Margaritiferamargaritifera

(A) (B)

(C) (D)

Fig. 5. (A–D) Relationships between lifespan and fecundity in North American freshwater mussels.

glochidial size and adult shell length (ANOVA, F = 5.99,P < 0.0178), showing that, across all species, the relationshipbetween fecundity and length was confounded by variation inglochidial size. After omitting from the analysis the six specieswith small glochidia and high fecundity, glochidial size wasnot related to fecundity (Fig. 4; P < 0.1995, r2 = 0.033,N = 61) and there was no interaction between glochidial sizeand length (F = 0.16, P < 0.6907). Further, there was norelationship between glochidial size and length-standardizedfecundity (ANCOVA, F = 2.03, P < 0.1604).

Fecundity showed little relationship with lifespan (Fig. 5).Annual fecundity was not related to lifespan, either for allspecies (P < 0.6189, N = 51), or for the dataset excludingspecies with high fecundity and small glochidia (P < 0.1274,N = 46). Similarly, there was no relationship between length-standardized fecundity and lifespan (P < 0.1538, N = 51).There was an apparent trend of higher length-specificfecundity in short-lived species, driven by Leptodea spp. andTruncilla spp., but long-lived Margaritifera margaritifera alsohad high length-specific fecundity. Lifetime fecundity waspositively related to lifespan, but lifespan explained littleof the variation in fecundity; omitting species with highfecundity and small glochidia improved the explanatorypower of this model by only a small margin (r2 = 0.2014,P < 0.004). However, this relationship was apparently anartifact of the weak, but positive relationship between sizeand lifespan (see Section II.3 and Haag & Rypel, 2011). Afterremoving the effect of size on lifetime fecundity, lifetimefecundity was not related to lifespan (ANCOVA, F = 0.88,

P < 0.3537, N = 51). Rather, lifetime fecundity was simplya strong function of annual fecundity (Fig. 5).

Fecundity was weakly related to brooding period. Totalfecundity did not differ between long-term and short-term brooders (ANOVA, F = 0.15, P < 0.7009, N = 71).Length-standardized fecundity differed significantly betweenbrooding strategies but explained very little variation infecundity (ANCOVA, F = 5.30, P < 0.0244, r2 = 0.071,N = 71; no interaction between log length and brooding,F = 2.41, P < 0.1258). Mean length-standardized fecunditywas higher for long-term brooders but confidence intervalsoverlapped [least-square mean (95% confidence interval):long-term brooders = 120216 (80716–179045); short-termbrooders = 56453 (33742–94450)], and in general, datascatter for the two strategies was similar (Fig. 6A). Removingspecies identified previously as outliers had no substantialeffect on relationships between fecundity and broodingperiod.

There was little relationship between fecundity and thearea of the gills used for brooding. Total fecundity wasnot related to gill brooding area (ANOVA, F = 0.05,P < 0.8217, N = 71). Length-standardized fecundity differedsignificantly among brooding area classes (ANOVA,F = 7.26, P < 0.0003; Fig. 6B). Fecundity was significantlyhigher for species that brood glochidia only in the posteriorhalf of the outer gills (i.e. class 0.25, including most speciesin the Lampsilini) than species that brood glochidia in theentire outer gills (i.e. 0.50; Anodontini, Elliptio spp., Pleurobemaspp.), but did not differ among the other brooding groups.



3.0

4.0

5.0

6.0

7.0

1.2 1.4 1.6 1.8 2.0 2.2

Log

fecu

ndity

Log length (mm)

Long-term broodersShort-term brooders

-2.0

-1.0

0.0

1.0

2.0

3.0

4.0

0.00 0.25 0.50 0.75 1.00

Leng

th-s

tand

ardi

zed

fecu

ndity

Proportion of gills used for brooding

1,21

21,2

(A) (B)

Fig. 6. (A,B) Relationship between brooding traits and fecundity in North American freshwater mussels. In (B), horizontal barsare mean length-standardized fecundity for each brooding area class; classes with the same number are not significantly different(P > 0.05).

Despite these significant differences, total and length-specificfecundity overlapped greatly among brooding classes (Fig. 6).

Length-standardized fecundity (residuals of log fecundityregressed on log length) was significantly different amonghost-infection strategies (ANOVA, F = 23.86, P < 0.0001,N = 58). Length-standardized fecundity was higher forbroadcasters than for all other infection strategies, but didnot differ significantly among any other strategies (Fig. 7).Among putative broadcasters, only Potamilus purpuratus hadlength-standardized fecundity that was similar to species withother infection strategies. Even though length-standardizedfecundity did not differ among the remaining strategies,residuals were centered nearly on zero only for species thatdisplayed mantle lures. For conglutinates, mucus webs, andmantle magazines, most residuals were < 0, indicating a weaktrend for lower length-specific fecundity in these strategies.Within the mantle lure strategy, length-standardizedfecundities were lowest in Epioblasma, which uses a host-trapping infection strategy in conjunction with lure display(see Section IV). Overall, the lowest observed values of length-standardized fecundity were for three species that produceconglutinates (Fusconaia burkei, F. ozarkensis, and Pleurobemacollina), and two species that apparently release glochidia inmucus webs (Alasmidonta heterodon, Elliptio complanata).

Patterns of length-standardized fecundity and hostuse were similar to those for infection strategy anddiffered significantly among groups (ANOVA, F = 10.10,P < 0.0001, N = 59). Length-standardized fecundity wassignificantly higher for species that use freshwater drumthan for all other host groups; among other groups fecunditydiffered only between species that use bass and those thatuse minnows (Fig. 7). In addition, length-specific fecunditywas high for each of the single species in the dataset that usegar (Lampsilis teres) and salmonids (Margaritifera margaritifera).Residuals were centered nearly on 0 for bass, sunfish,darter, sculpin, skipjack herring, mudpuppy, and sauger,suggesting that mussels using these hosts have similar length-specific fecundities. Most residuals for minnows, catfishes,and generalists were < 0, suggesting a weak trend of lower

length-specific fecundity for mussel species using these hostgroups. Despite these patterns, fecundity overlapped greatlyfor all host strategies.

(4) Relationships of reproductive effort andlife-history variables

Reproductive effort was not related to length, fecundity, orglochidial size, either for the combined dataset of observedand estimated values (P < 0.138–0.891, r2 = 0.001–0.107),or for observed values only (P < 0.850–0.960,r2 = 0.001–0.005). For observed and estimated REcombined, RE was strongly negatively related to lifespan(Fig. 8). For observed values only, lifespan explained aneven higher percentage of the variation in RE (P < 0.0001,r2 = 0.861). Reproductive effort also differed significantlyamong brooding strategies and was higher in long-termbrooders (ANOVA: F = 40.84, 1 d.f., P < 0.0001; Fig. 8).

(5) Models with multiple life-history variables

The multiple regression model including length, glochidialsize, lifespan, brooding period, and host infection strategysupported results of bivariate relationships and the primaryimportance of size in predicting fecundity. For the full dataset,size was clearly the most important variable determiningfecundity, and this model explained a large percentage of thevariation in fecundity (Table 4). However, glochidial size andhost-infection strategy were also significant factors althoughtheir sums of squares were much smaller than for length.Lifespan and brooding period were not significant factors. Inthe dataset excluding the five species identified as significantoutliers (Leptodea spp., Margaritifera margaritifera, and Truncillaspp.; see Section III.3), length was the only significant factorbut infection strategy was marginally significant; this modelexplained a similar proportion of the variation in fecundityas the full dataset (Table 4). For the 40 species in this dataset,length alone explained 73% of the variation in fecundity.This suggests that knowledge of infection strategy could


760 Wendell R. Haag

-2

-1

0

1

2

3

4

Broadcaster1

Mantle lure2Conglutinate2

Mucus web2Mantle magazine2

-2

-1

0

1

2

3

4

drum

1

saug

er

catfi

sh2,

3

gene

ralis

t2,3

dart

er2,

3

sunf

ish2,

3

bass

2

gar

salm

onid

scul

pin2,

3

min

now

3

mud

pupp

y

skip

jack

her

ring

Leng

th-s

tand

ardi

zed

fecu

ndity

(A)

(B)

Fig. 7. Patterns of fecundity among host strategies of NorthAmerican freshwater mussels. (A) Shows variation in fecundityamong host-infection strategies and (B) shows variation amonghost-use strategies. For both (A) and (B), length-standardizedfecundity was obtained as the residuals of fecundity regressedon length (both log-transformed). Horizontal bars are meanresiduals for each group; groups with the same number are notsignificantly different (P > 0.05); on the (B), host groups withoutnumbers were not included in the analysis.

provide at best only a small improvement in prediction offecundity compared to length alone.

IV. DISCUSSION

Like many organisms, fecundity of mussels is stronglyrelated to body size, both within and among species. Withina population, a high proportion of variation in fecundityamong individuals is accounted for by shell length. Becausefecundity increases as a power function of length in mostspecies, large individuals are expected to contribute adisproportionately high fraction of reproductive output ina population. Age is typically a poorer predictor of musselfecundity (Haag & Staton, 2003), suggesting that fecundityis determined primarily by energetic constraints related tobody size. Although shell volume and mass were roughlyequivalent to length as predictors of fecundity, femalebody mass or condition may explain some of the variationin fecundity unaccounted for by length. In Margaritiferamargaritifera, heavier females produced more glochidia per

unit length, and there was an apparent body mass thresholdbelow which other females did not reproduce (Bauer, 1998).

The extent to which fecundity varies among populationsand the significance of this variation remains poorly known.In Elliptio arca, Fusconaia cerina, and Quadrula asperata, meanfecundity and length-specific fecundity did not differ amongpopulations in a river or among nearby rivers (see Haag &Staton, 2003). In other cases, mean fecundity differed widelybetween populations, but length-fecundity relationships weresimilar, and differences in fecundity were simply a result ofpopulation differences in adult size (e.g. Obliquaria reflexa,Pyganodon grandis). Growth rate and body size can beinfluenced to a large extent by environmental factors (seeHaag & Rypel, 2011), and concomitant differences infecundity are an expected consequence of this plasticity.More interesting are cases in which length-specific fecunditydiffers among populations, indicating different levels ofenergetic allocation to reproduction. Such differences havebeen documented among populations in different waterbodies (Amblema plicata, Pyganodon cataracta), and even amongpopulations in a river (Actinonaias ligamentina, Pleurobema

decisum; see Paterson & Cameron, 1985; Haag & Staton,2003; Moles & Layzer, 2008). The potential adaptivesignificance of these differences is unknown, but somealso may be simple responses to variation in body size.For example, even though length-specific fecundity differedwidely among populations of P. cataracta, these populationsalso differed in length-specific body mass, and fecundity didnot differ when expressed as a function of mass (Paterson& Cameron, 1985). Despite these examples, no studies haveexamined patterns of fecundity and reproductive allocationat large scales, such as across the geographic range of aspecies or across broad environmental gradients.

The great variation in fecundity among mussel speciessuggests that this may be an important trait for describingdivergent life-history strategies within the group. However,similar to within- and among-population variation, variationin fecundity among species was explained primarily by bodysize and there was little evidence of trade-offs betweenfecundity and other life-history traits. For most species,there was no evidence of a trade-off between fecundity andoffspring size, and little evidence of a trade-off betweenfecundity and lifespan. The degree of parental care investedin offspring was opposite to that predicted by life-historytheory: species that brood glochidia for extended periods(long-term brooders) had on average significantly higherfecundity than short-term brooders.

The strongest evidence for a fecundity trade-off are thediminutive glochidia of Margaritifera margaritifera, Leptodea spp.,and Truncilla spp., which are less than one third the size ofmost other species. Even though they have extraordinarilyhigh fecundity, RE of these species is lower than, or compa-rable to, many other species. This suggests that reduced ener-getic investment in individual glochidia allows maximizationof total glochidial output as predicted by life-history the-ory. A cost of small glochidia may be the requirement for anextended period of glochidial development on hosts (Howard



0

5

10

15

20

25

30

0 10 20 30 40 50 60 70 80

Re

prod

uctiv

e e

ffort

(%)

Lifespan (years)

observed RE

predicted RE

1.9

13.8

0

5

10

15

20

25

30(A) (B)

short-termbrooders

long-termbrooders

Fig. 8. Relationship between reproductive effort (RE), lifespan, and brooding period. (A) Regression equation is arcsinRE =−16.564(log lifespan) + 35.254; P < 0.0001, r2 = 0.632 (observed and predicted RE combined), but data are plotteduntransformed. (B) Horizontal bars indicate mean values back-transformed from arcsin-transformed data.

Table 4. Results of multiple regression models for predicting fecundity of freshwater mussels based on life-history variables

Variable d.f. Partial sums of squares F P < r2

Full dataset (N = 46) 8 24.470 30.29 0.0001 0.871Length 1 11.010 109.04 0.0001 —Glochidial size 1 1.575 15.60 0.0003 —Lifespan 1 0.003 0.03 0.8646 —Host-infection strategy 4 1.101 2.73 0.0444 —Brooding period 1 0.081 0.80 0.3766 —

Dataset without outliers (N = 40) 8 15.697 21.98 0.0001 0.850Length 1 9.731 109.00 0.0001 —Glochidial size 1 0.213 2.39 0.1326 —Lifespan 1 0.010 0.12 0.7355 —Host-infection strategy 4 0.884 2.47 0.0648 —Brooding period 1 0.001 0.01 0.9096 —

& Anson, 1922; Young & Williams, 1984; Steingraeber et al.,2007; Barnhart et al., 2008), during which the chances ofglochidial or host death are increased. Despite the costs ofsmall glochidia, the host-infection strategy of these speciesmay place them under strong selective pressure to maximizefecundity. Margaritifera margaritifera broadcasts free glochidiainto the water column where host infection apparently islargely a matter of chance and may be strongly dependenton high glochidial abundance in the drift (Murphy, 1942). InM. margaritifera, small glochidia are further thought to allowhigh fecundity despite the low productivity of streams inwhich it typically occurs (Bauer, 1998). Host-infection strate-gies of Leptodea spp. and Truncilla spp. are unknown, and theforces that may have selected for small glochidia and highfecundity in these species are unclear. Gravid female Trun-cilla spp. show subtle mantle movements that are absent inmales (B. Sietman, personal communication), but otherwise,both genera lack conspicuous mantle lures, conglutinates,or other obvious host attraction adaptations, suggesting thatthey also may broadcast free glochidia. Alternatively, itis proposed that their host, the molluscivorous freshwaterdrum, becomes infected with glochidia by feeding on gravidfemale mussels (Coker et al., 1921; Howard & Anson, 1922;Barnhart et al., 2008). This host-infection mode has not beendocumented directly, but small glochidia may make possible

the unusually high glochidial infestations typically seen ondrum (e.g. Surber, 1915; Coker et al., 1921; Weiss & Layzer,1995) therefore maximizing the benefits of female sacrifice.Regardless, the occurrence of this trait in two disparate mus-sel lineages (Margaritiferidae and Lampsilini) suggests that itis an adaptive trade-off rather than a phylogenetic artifact.

Diminutive glochidia also occur in the Quadrula quadrulagroup (including Q. rumphiana) and in Quadrula verrucosa(Barnhart et al., 2008). These species attract and infect hostswith mantle magazines, but this trait is shared with othermembers of the Quadrulini that do not have diminutiveglochidia (Cyclonaias tuberculata, Quadrula metanevra group, Q.pustulosa group; Barnhart et al., 2008; Sietman, Davis &Hove, 2012). Unlike other species with diminutive glochidia,fecundity is low in Q. rumphiana and is comparable to Q.asperata and Q. pustulosa, and much lower than Q. cylindrica, allof which have glochidia that are nearly three times larger thanQ. rumphiana. Consequently, the presence of small glochidia inthe Quadrulini clearly does not represent a trade-off allowinghigher fecundity, and the potential adaptive significance ofthis trait is unexplained.

There were no strong differences in fecundity amongother host-infection strategies suggesting that differencesin efficiency among these strategies have not been largeenough to exert substantial selective pressure on fecundity.


762 Wendell R. Haag

Nevertheless, differences among or within some strategiescould reflect the relative efficiency of host infection. Forexample, the lower length-specific fecundity of Epioblasma

spp. compared to other species with mantle lures may bedue to the unusual host trapping behaviour exhibited bythese species. Females trap hosts (darters) between their shellvalves and pump glochidia over the fish’s gills (Barnhart et al.,2008); this behaviour may be highly efficient in transferringglochidia to hosts thus relaxing selection for high fecundity.For mantle-lure strategies that do not include host trapping,it is more difficult to propose that their generally highlength-specific fecundity reflects lower infection efficiencycompared with conglutinate, mucus web, and mantlemagazine strategies that are typically associated with lowlength-standardized fecundity. First, pelagic conglutinatesare composed of a large number of unfertilized eggs thatprovide structure to these conglutinates (Barnhart et al.,2008). These structural eggs represent an energetic cost tofemales and necessarily lower fecundity dramatically in thesespecies. Second, most species that produce mucus webs arehost generalists. Rather than reflecting the efficiency of theweb strategy itself in infecting fishes, lower fecundity maybe possible in these species because of a greater chance ofinfecting hosts brought about by a broadening of the hostresource. Finally, there is likely a phylogenetic componentto some of these patterns. For example, all species that usemantle magazines to infect hosts are in the Quadrulini, andlower fecundity for this strategy may be simply a shared,inherited trait within this lineage.

Other shared traits within lineages may be responsible fordifferences in fecundity among species or tribes. Broodingperiod and the area of the gill used for brooding have a strongphylogenetic component and may not be easily modified byselection. Brooding glochidia in the gills likely decreasesfeeding and respiratory efficiency. In the Anodontini andLampsilini, glochidia are brooded for an extended period,and gravid gills are greatly distended to several timestheir normal thickness leaving little or no space for watercirculation. This problem is dealt with in the Anodontiniby the development of accessory gill water tubes in gravidfemales (Richard et al., 1991; Tankersley & Dimock, 1993;Tankersley, 1996) and in the Lampsilini by using only asmall portion of the total gill area. In species that use theentire length of the gill for brooding and lack accessory watertubes (Margaritiferidae, Pleurobemini, Quadrulini), gills areusually incompletely filled with glochidia and are not greatlydistended. This condition has been interpreted as a result ofpoor fertilization success or female energetic status (Haggertyet al., 1995; Bauer, 1998). However, incompletely chargedgills seems to be a characteristic of these species (Coker et al.,1921; Haag & Staton, 2003) and, along with a short broodingperiod, may be necessary to maintain filtering and respiratoryefficiency. Such constraints could partially explain the lowfecundity of most Pleurobemini and Quadrulini.

In contrast to fecundity, patterns of variation in repro-ductive effort among species showed evidence of a strongtrade-off with lifespan as predicted by life-history theory. A

central tenet of life-history evolution is that, when resourcesare limiting, organisms with longer life spans shoulddevote less energy to reproduction per unit time comparedwith short-lived organisms (Williams, 1966; Stearns,1992; Charlesworth, 1994). High and early investment inreproduction can decrease lifespan because of diversion ofresources from somatic maintenance, and also because ofadult mortality associated with reproductive behaviours.In addition to differences among species, this trade-off alsoappears to be manifested among populations of a species.Reproductive effort of Amblema plicata from a long-livedpopulation in the Sipsey River (54 years) was about half thatof a short-lived population in the Little Tallahatchie River(18 years; see Haag & Rypel, 2011). Among 13 Finnish pop-ulations of Anodonta piscinalis, there was a negative correlationbetween RE and lifespan in one year when food resourcesappeared to be limiting, but this trade-off was not apparentwhen resources were more abundant (Haukioja & Hakala,1978). One unusual feature of RE in mussels is the higherreproductive effort of long-term brooders, which presumablyinvest more energy in parental care than short-term broodersand would be expected accordingly to devote less energy tooffspring production. Factoring in potential costs of broodingmay further widen differences in RE among species.

Trade-offs among life-history traits are depicted by theconcepts of the fast-slow continuum and r- and K - selection(Pianka, 1970; Promislow & Harvey, 1990; Bielby et al.,2007). In a fast life history (generally corresponding tor-selected species), high or variable extrinsic mortality isthought to favour fast growth and early and high investmentin reproduction, which results in decreased lifespan. Lowor constant mortality and increased competition selects forthe opposite suite of traits represented by a slow life history(K -selected species) (Arendt, 1997; Cichon, 1997).

In addition to the apparent trade-off between reproductiveeffort and lifespan, relationships among other mussel life-history traits are emerging that suggest a broad range oflife-history strategies in the group. Lifespan varies widelyamong mussel species, ranging from < 5 to > 100 years, andis strongly negatively related to growth rate such that speciesthat invest heavily in growth have predictably short lives(Haag & Rypel, 2011). Similarly, age at maturity rangesfrom < 1 to 20 years and also is negatively related togrowth rate but positively related to lifespan (Haag, 2012).Together, these relationships depict a continuum of life-history diversity with at least two endpoints corresponding tor-selected species with fast growth, short lifespan, high annualreproductive investment, and early maturity, and K -selectedspecies with the opposite suite of traits. Among species inthe present study, r-strategists include the lampsiline Leptodea

fragilis, and the anodontines Anodonta anatina, Pyganodon spp.,and Utterbackia imbecillis. Species at the K -selected end ofthe continuum include the Pleurobemini (Fusconaia spp.,Pleurobema spp.), Quadrulini, slow-growing populations ofAmblema plicata, and potentially lampsilines such as Obliquaria

reflexa and Obovaria unicolor. Despite its high fecundity,Margaritifera margaritifera has low reproductive effort, late age



at maturity, and a long lifespan concordant with a K -selected life-history strategy (Bauer, 1998). Other species inthis study fall at intermediate positions along the continuum,and for some, such as Amblema plicata, life-history traitsmay be plastic resulting in variable expression of life-historystrategy among populations in different environments (Haag,2012). Particularly interesting is the very low reproductiveeffort of Quadrula rumphiana, which was at least an order ofmagnitude lower than all other species. This species has onlya moderate life span (∼28 years), but its low reproductiveeffort and diminutive glochidia suggest that it represents aunique life-history strategy.

V. CONCLUSIONS

(1) Fecundity of North American mussel species spansnearly four orders of magnitude, but most species haveconsiderably lower fecundity than previous characterizationsfor the group (e.g. > 200000; McMahon & Bogan, 2001).Nevertheless, the primary role of body size in predictingfecundity suggests that glochidial production is mainlydetermined by physical and energetic constraints ratherthan trade-offs among life-history traits. For species thatbroadcast free glochidia, extraordinarily high fecundity atthe cost of reduced glochidial size may be selected for tomaximize chances of host infection in the drift. However,for most other species with more elaborate host-infectionstrategies, differences in fecundity among species may havelittle adaptive significance. The high fecundity seen in somemussel species has been explained as an adaptation associatedwith their parasitic life cycle (McMahon & Bogan, 2001).Parasites were long considered to have higher fecundity thanfree-living organisms to compensate for the high mortalitythat typically occurs during transmission to hosts (Poulin,1995; Whittington, 1997). However, recent studies haveshown that, similar to mussels, fecundity of many parasitesis primarily a function of body size, and length-specificfecundity of parasitic organisms is not different from free-living relatives (e.g. Trouve & Morand, 1998). Glochidiaexperience extremely high mortality (� 99%) betweenrelease from the female and recruitment to the benthicjuvenile stage, and recruitment can be highly variable amongyears probably according to environmental factors (Jansen,Bauer & Zahner-Meike, 2001; Haag, 2012). Accordingly,variation in fecundity has little influence on populationgrowth compared to adult survival (Haag, 2012). Despitethe traditional focus on fecundity as a measure of fitness,changes in adult survivorship are increasingly viewed ashaving a much greater effect on population growth in awide range of organisms (Crone, 2001). For mussels, as inother organisms, differences among habitats in the relativeimportance of environmental stochasticity and its effects onadult survivorship may be primary drivers of diversificationin life-history strategies, which in turn, influence speciesresponses to human impacts (Haag, 2012).

(2) Unlike fecundity, the relative allocation of energy toreproduction, as measured by reproductive effort, appearsto be an important variable for understanding mussel life-history evolution. The negative relationship between REand lifespan, as well as apparent trade-offs among otherlife-history traits, indicates a range of divergent life-historystrategies with endpoints similar to those described for manyother organisms (Haag, 2012; see also Winemiller & Rose,1992; Grime, 2001).

(3) Despite its strong relationship to other life-historyvariables, RE of freshwater mussels is generally low(mean = 7.2% in this study for dry-mass estimates) comparedwith marine bivalves, in which RE commonly exceeds40% (Dame, 1996). Further, estimates of RE that do notinclude the inorganic fraction of glochidial mass are abouthalf of estimates based on total dry mass, indicating evenlower investment in reproduction. This could suggest thatproduction of glochidia incurs comparatively little cost tofemale mussels, and consequently, patterns of reproductiveallocation may have had little role in shaping life histories (seeTuomi, Hakala & Haukioja, 1983). This idea is supportedby the observation that transplanting freshwater musselsamong sites with apparently varying productivity resultedin large changes in growth but only slight or no changesin reproductive output (Jokela & Mutikainen, 1995). Analternative explanation for this result is that allocationto reproduction is a high-priority heritable trait that isinfluenced by environmental variation to a lesser extentthan phenotypically plastic traits such as growth (Glazier,2002). Additional studies of RE in mussels, particularlythose addressing the influence of environmental and bioticfactors and accordant variation in RE among populations,will be valuable in elucidating mechanisms of life-historydiversification.

(4) Measures of RE based on biomass can be a usefulindex of relative differences among species or populations,but this method has several shortcomings (Roff, 1992). First,it does not account for potential differences in caloric contentof reproductive versus somatic tissues. Second, it is not basedon actual production and cannot account for differences inreproductive allocation over the life of an individual. In mostbivalves, allocation of production to gonads increases withage and often approaches 100% in older individuals withslow growth (Thompson, 1984; Nakaoka, 1994; Gosling,2003). In addition, differences in age at maturity amongspecies will influence schedules of reproductive allocationand their demographic consequences.

(5) Several issues specific to mussels also complicateaccurate estimation of RE. First, in the method used by Bauer(1998), measurements of adult female body mass includedthe mass of the soft tissue and the organic fraction of the shellonly, but glochidial mass included these components as wellas the inorganic fraction of glochidial shells. In my dataset,the mean inorganic fraction of glochidial mass across specieswas 62% – nearly three times higher than for adult tissuemass (mean = 22%) – resulting in estimates of RE based onash-free dry mass that were about half of those based on total


764 Wendell R. Haag

dry mass. However, the inorganic fraction of glochidia maynot be completely cost-free for female mussels. For example,maternal carbon and calcium is transferred to developingglochidia (Wood, 1974; Silverman, Kays & Dietz, 1987), butthe magnitude of these and other contributions is unknown.Second, other aspects of reproduction likely incur as yetunspecified costs. Unfertilized structural eggs, membranesand pigments associated with complex conglutinates suchas those in Ptychobranchus spp., copious mucus produced inassociation with glochidia in many species, and reductionsin filtering efficiency associated with brooding all have anenergetic cost that must be accounted for. Production anddisplay of complex mantle lures in the Lampsilini doubtlessrequires considerable energy, and display of these lures mayexpose females to higher predation risk (Jones & Neves,2011). Third, reproductive periodicity is poorly known formany species and it is possible that some, particularly short-term brooders, produce multiple broods in a year (e.g. Price& Eads, 2011; see Haag, 2012). Frequent multiple-broodproduction would substantially alter the low annual fecundityand RE apparent for the Pleurobemini and Quadruliniunder the assumption of single broods. Finally, populationdynamics of most mussel species, particularly short-livedspecies, are poorly known. Short-lived species are expectedto have higher rates of instantaneous mortality, which isexpected to influence reproductive allocation (Haukioja &Hakala, 1978; Heino & Kaitala, 1997, 1999; Haag, 2012).

(6) The apparent breadth of mussel life histories contrastswith traditional depictions of the group as uniformlylong lived and with typically high fecundity. Effectiveconservation efforts will need to consider fundamentalecological differences among species because strategiesdesigned for long-lived, slow-growing species at the K -selected end of the life-history continuum may be ineffectivefor species with r-selected traits. For example, efforts toreestablish extirpated populations of short-lived species withhigh RE may require very large initial introductions tooffset the high natural mortality typically associated with thislife history. Short-lived species with high growth rates andhigh RE also may be dependent on high food availabilitymaking them more susceptible to food competition in densemussel assemblages. Understanding the evolution of mussellife histories and the factors that shape mussel assemblages isultimately dependent on specifying the selective forces thatregulate populations, particularly the roles of mortality anddensity dependence.

VI. ACKNOWLEDGEMENTS

I thank the following people for their various contributions tothis study. Steve Ahlstedt, Chris Barnhart, Ryan Evans, SteveFraley, Jeff Garner, Mark Hove, John Harris, Paul Johnson,Larry Shaffer, Jim Stoeckel, and Tom Watters providedor helped collect specimens. Chris Barnhart, Al Christian,Rachel Mair, Michael Stewart, Jess Jones, and Jim Layzerkindly provided unpublished data on fecundity of several

species. Leann Staton, Mickey Bland, Amy Commens-Carson, Angela Greer, Karen Higgins, Liz McQuire, andAnthony Rietl helped in the field and conducted much of thelaboratory work. This study was supported by the NationalFish and Wildlife Foundation, Weyerhaeuser Company, andthe Southern Research Station, US Forest Service.

VII. REFERENCES

Arendt, J. D. (1997). Adaptive intrinsic growth rates: an integration across taxa.Quaterly Review of Biology 72, 149–177.

Baird, M. S. (2000). Life history of the spectaclecase, Cumberlandia monodonta Say,

1829 (Bivalvia, Unionoidea, Margaritiferidae). M.S. Thesis: Southwest Missouri StateUniversity, Springfield.

Barnhart, M. C. (2001). Fish hosts and culture of mussel species of specialconcern. Unpublished report. Report to US Fish and Wildlife Service andMissouri Department of Conservation, Columbia. Available at http://biology.missouristate.edu/faculty_pages/Barnhart%20pubs/barnhart_pubs.htm.

Barnhart, M. C. (2002). Propagation and culture of mussel species of specialconcern. Unpublished report. Report to US Fish and Wildlife Service andMissouri Department of Conservation, Columbia. Available at http://biology.missouristate.edu/faculty_pages/Barnhart%20pubs/barnhart_pubs.htm.

Barnhart, M. C. (2003). Culture and restoration of mussel species ofconcern. Unpublished report. Report to US Fish and Wildlife Service andMissouri Department of Conservation, Columbia. Available at http://biology.missouristate.edu/faculty_pages/Barnhart%20pubs/barnhart_pubs.htm.

Barnhart, M. C., Haag, W. R. & Roston, W. N. (2008). Adaptations to hostinfection and larval parasitism in Unionoida. Journal of the North American Benthological

Society 27, 370–394.Bauer, G. (1994). The adaptive value of offspring size among freshwater mussels

(Bivalvia; Unionoidea). Journal of Animal Ecology 63, 933–944.Bauer, G. (1998). Allocation policy of female freshwater pearl mussels. Oecologia (Berlin)

117, 90–94.Benke, A. C., Huryn, A. D., Smock, L. A. & Wallace, J. B. (1999). Length-mass

relationships for freshwater macroinvertebrates in North America with particularreference to the southeastern United States. Journal of the North American Benthological

Society 18, 308–343.Bielby, J., Mace, G. M., Bininda-Emonds, O. R., Cardillo, M., Gittleman, J.

L., Jones, K. E., Orme, C. D. & Purvis, A. (2007). The fast-slow continuum inmammalian life history: an emperical reevaluation. American Naturalist 169, 748–757.

Blalock-Herod, H. N., Herod, J. & Williams, J. D. (2002). Evaluation ofconservation status, distribution, and reproductive characteristics of an endemicGulf Coast freshwater mussel, Lampsilis australis (Bivalvia: Unionidae). Biodiversity and

Conservation 11, 1877–1887.Bruenderman, S. A. & Neves, R. J. (1993). Life history of the endangered fine-rayed

pigtoe, Fusconaia cuneolus (Bivalvia: Unionidae) in the Clinch River, Virginia. American

Malacological Bulletin 10, 83–91.Cameron, C. J., Cameron, I. F. & Paterson, G. C. (1979). Contribution of organic

matter to biomass estimates of unionid bivalves. Canadian Journal of Zoology 57,1666–1669.

Campbell, D. C., Serb, J. M., Buhay, J. E., Roe, K. J., Minton, R. L.& Lydeard, C. (2005). Phylogeny of North American amblemines (Bivalvia,Unionoida): prodigious polyphyly proves pervasive across genera. Invertebrate Biology

124, 131–164.Charlesworth, B. (1994). Evolution in Age-Structured Populations. Cambridge University

Press, Cambridge.Charnov, E. L., Warne, R. & Moses, M. (2007). Lifetime reproductive effort.

American Naturalist 170, E129–E142.Cichon, M. (1997). Evolution of longevity through optimal resource allocation.

Philosophical Transactions of the Royal Society of London B 264, 1383–1388.Coker, R. E., Shira, A. F., Clark, H. W. & Howard, A. D. (1921). Natural history

and propagation of fresh-water mussels. Bulletin of the Bureau of Fisheries 37, 79–181.Crone, E. E. (2001). Is survivorship a better fitness surrogate than fecundity? Evolution

55, 2611–2614.Dame, R. F. (1996). Ecology of Marine Bivalves: An Ecosystem Approach. CRC Press, Boca

Raton.Eckert, N. L. (2003). Reproductive biology and host requirement differences among isolated

populations of Cyprogenia aberti (Conrad, 1850). MS Thesis: Missouri State University,Springfield.

Fobian, T. (2007). Reproductive biology of the rabbitsfoot mussel ( Quadrula cylindrica)

(Say, 1817) in the upper Arkansas River system. MS Thesis: Missouri State University,Springfield.



Garner, J. T., Haggerty, T. M. & Modlin, R. F. (1999). Reproductive cycleof Quadrula metanevra (Bivalvia: Unionidae) in the Pickwick Dam tailwater of theTennessee River. American Midland Naturalist 141, 277–283.

Glazier, D. S. (2002). Resource-allocation rules and the heritability of traits. Evolution

56, 1696–1700.Gosling, E. (2003). Bivalve Mollusks: Biology, Ecology, and Culture. Blackwell, Oxford.Graf, D. L. (2002). Molecular phylogenetic analysis of two problematic freshwater

mussel genera (Unio and Gonidea) and a re-evaluation of the classification of NearcticUnionidae (Bivalvia: Palaeoheterodonta: Unionoida). Journal of Molluscan Studies 68,65–71.

Grime, J. P. (2001). Plant Strategies, Vegetation Processes, and Ecosystem Properties. John Wiley& Sons, Chichester.

Gutierrez, J. L., Jones, C. G., Strayer, D. L. & Iribarne, O. O. (2003). Mollusksas ecosystem engineers: the role of shell production in aquatic habitats. Oikos 101,79–90.

Haag, W. R. (2012). North American Freshwater Mussels: Natural History, Ecology, and

Conservation. Cambridge University Press, Cambridge.Haag, W. R. & Rypel, A. L. (2011). Growth and longevity in freshwater mussels:

evolutionary and conservation implications. Biological Reviews 86, 225–247.Haag, W. R. & Staton, J. L. (2003). Variation in fecundity and other reproductive

traits in freshwater mussels. Freshwater Biology 48, 2118–2130.Haag, W. R. & Warren, M. L. (1997). Host fishes and reproductive biology of 6

freshwater mussel species from the Mobile Basin, USA. Journal of the North American

Benthological Society 16, 576–585.Haggerty, T. M., Garner, J. T., Patterson, G. H. & Jones, L. C. (1995). A

quantitative assesment of the reproductive biology of Cyclonaias tuberculata (Bivalvia:Unionidae). Canadian Journal of Zoology 73, 83–88.

Hanson, J. M., MacKay, W. C. & Prepas, E. E. (1989). Effect of size-selectivepredation by muskrats (Ondatra zibethicus) on a population of unionid clams (Anodontagrandis simpsoniana). Journal of Animal Ecology 58, 15–28.

Harmon, J. L. & Joy, J. E. (1990). Growth rates of the freshwater mussel, Anodonta

imbecillis Say 1829, in five West Virginia Wildlife Station ponds. American Midland

Naturalist 124, 372–378.Haukioja, E. & Hakala, T. (1978). Measuring growth from shell rings in populations

of Anodonta piscinalis (Pelecypoda, Unionidae). Annales Zoologici Fennici 15, 60–65.Heino, M. & Kaitala, V. (1997). Should ecological factors affect the evolution of age

at maturity in freshwater clams? Evolutionary Ecology 11, 67–81.Heino, M. & Kaitala, V. (1999). Evolution of resource allocation between growth

and reproduction in animals with indeterminate growth. Journal of Evolutionary Biology

12, 423–429.Hoggarth, M. A. (1999). Descriptions of some of the glochidia of the Unionidae

(Mollusca: Bivalvia). Malacologia 41, 1–118.Hove, M. C. & Neves, R. J. (1994). Life history of the endangered James spinymussel

Pleurobema collina (Conrad, 1837) (Mollusca:Unionidae). American Malacological Bulletin

11, 29–40.Howard, A. D. & Anson, B. J. (1922). Phases in the parasitism of the Unionidae.

Journal of Parasitology 9, 68–82.Jansen, W., Bauer, G. & Zahner-Meike, E. (2001). Glochidial mortality in

freshwater Mussels. In Ecology and Evolution of the Freshwater Mussels Unionoida,Ecological Studies (Volume 145, eds G. Bauer and K. Wachtler), pp. 185–211.Springer-Verlag, Berlin.

Jokela, J. & Mutikainen, P. (1995). Phenotypic plasticity and priority rules forenergy allocation in a freshwater clam: a field experiment. Oecologia 104, 122–132.

Jones, J. W. & Neves, R. J. (2002). Life history and propagation of the endangeredfanshell pearly mussel, Cyprogenia stegaria Rafinesque (Bivalvia: Unionidae). Journal of

the North American Benthological Society 21, 76–88.Jones, J. W. & Neves, R. J. (2011). Influence of life-history variation on demographic

responses of three freshwater mussel species (Bivalvia: Unionidae) in the ClinchRiver, USA. Aquatic Conservation: Marine and Freshwater Ecosystems 21, 57–73.

Jones, J. W., Neves, R. J., Ahlstedt, S. A., Hubbs, D., Johnson, M., Dan, H. &Ostby, B. J. K. (2010). Life history and demographics of the endangered birdwingpearlymussel (Lemiox rimosus) (Bivalvia:Unionidae). American Midland Naturalist 163,335–350.

Jones, J. W., Neves, R. J., Ahlstedt, S. A. & Mair, R. A. (2004). Life history andpropagation of the endangered dromedary pearlymussel (Dromus dromas) (Bivalvia:Unionidae). Journal of the North American Benthological Society 23, 515–525.

Kennedy, T. B. & Haag, W. R. (2005). Using morphometrics to identify glochidiafrom a diverse freshwater mussel community. Journal of the North American Benthological

Society 24, 880–889.McMahon, R. F. & Bogan, A. E. (2001). Mollusca: Bivalvia. In Ecology and

Classification of North American Freshwater Invertebrates (eds J. H. Thorp andA. P. Covich), pp. 321–429. Academic Press, San Diego.

Moles, K. R. & Layzer, J. B. (2008). Reproductive ecology of Actinonaias ligamentina

(Bivalvia: Unionidae) in a regulated river. Journal of the North American Benthological

Society 27, 212–222.Murphy, G. (1942). Relationship of the fresh-water mussel to trout in the Truckee

River. California Fish and Game 28, 89–102.

Nakaoka, M. (1994). Size-dependent reproductive traits of Yoldia notabilis (Bivalvia:Protobranchia). Marine Ecology Progress Series 114, 129–137.

O’Brien, C. A. & Williams, J. D. (2002). Reproductive biology of four freshwatermussels (Bivalvia: Unionidae) endemic to eastern Gulf Coastal Plain drainages ofAlabama, Florida, and Georgia. American Malacological Bulletin 17, 147–158.

Parker, R. S., Hackney, C. T. & Vidrine, M. F. (1984). Ecology and reproductivestrategy of a South Louisiana freshwater mussel, Glebula rotundata (Lamarck)(Unionidae:Lampsilini). Freshwater Invertebrate Biology 3, 53–58.

Paterson, C. G. (1985). Biomass and production of the unionid, Elliptio complanata

(Lightfoot) in an old reservoir in New Brunswick, Canada. Freshwater Invertebrate

Biology 4, 201–207.Paterson, C. G. & Cameron, I. F. (1985). Comparative energetics of two populations

of the unionid, Anodonta cataracta (Say). Freshwater Invertebrate Biology 4, 79–90.Perles, S. J., Christian, A. D. & Berg, D. J. (2003). Vertical migration, orientation,

aggregation, and fecundity of the freshwater mussel Lampsilis siliquoidea. Ohio Journal

of Science 103, 73–78.Pianka, E. R. (1970). On r- and K - selection. American Naturalist 104, 592–597.Pilarczyk, M. M., Stewart, P. M., Shelton, D. N., Heath, W. H. & Miller,

J. M. (2005). Contemporary survey and historical freshwater mussel assemblages insoutheast Alabama and northwest Florida and life history and host fish identificationof two candidate unionids (Quincuncina burkei and Pleurobema strodeanum). Unpublishedreport to US Fish and Wildlife Service, Panama City.

Poulin, R. (1995). Evolution of parasite life history traits: myths and reality. Parasitology

Today 11, 342–345.Price, J. E. & Eads, C. B. (2011). Brooding patterns in three freshwater mussels of

the genus Elliptio in the Broad River in South Carolina. American Malacological Bulletin

29, 121–126.Promislow, D. E. L. & Harvey, P. H. (1990). Living fast and dying young: a

comparative analysis of life-history variation among mammals. Journal of Zoology

(London) 220, 417–437.Regnier, C., Fontaine, B. & Bouchet, P. (2009). Not knowing, not recording, not

listing: numerous unnoticed mollusk extinctions. Conservation Biology 23, 1214–1221.Richard, P. E., Dietz, T. H. & Silverman, H. (1991). Structure of the gill during

reproduction in the unionids Anodonta grandis, Ligumia subrostrata, and Carunculina parva

texasensis. Canadian Journal of Zoology 69, 1744–1754.Roff, D. A. (1992). The Evolution of Life Histories: Theory and Analysis. Chapman

& Hall, New York.Rogers, S. O., Watson, B. T. & Neves, R. J. (2001). Life history and population

biology of the endangered tan riffleshell (Epioblasma florentina walkeri) (Bivalvia:Unionidae). Journal of the North American Benthological Society 20, 582–594.

van der Schalie, H. & van der Schalie, A. (1963). The distribution, ecology,and life history of the mussel, Actinonaias ellipsiformis (Conrad), in Michigan. Occasional

Papers of the Museum of Zoology, University of Michigan 633, 1–17.Sietman, B. E., Davis, J. M. & Hove, M. C. (2012). Mantle display and glochidia

release behaviors of five quadruline freshwater mussel species (Bivalvia: Unionidae).American Malacological Bulletin 30, 39–46.

Silverman, H., Kays, W. T. & Dietz, T. H. (1987). Maternal calcium contributionto glochidial shells in freshwater mussels (Eulamellibranchia:Unionidae). The Journal

of Experimental Zoology 242, 137–146.Stearns, S. C. (1992). The Evolution of Life Histories. Oxford University Press,

Oxford.Steingraeber, M. T., Bartch, M. R., Kalas, J. E. & Newton, T. J. (2007).

Thermal criteria for early life stage development of the winged mapleleaf mussel(Quadrula fragosa). American Midland Naturalist 157, 297–311.

Surber, T. (1915). Identification of the glochidia of fresh-water mussels. Report ofthe US Commissioner of Fisheries for 1914, Appendix 5 (Issued separately as USBureau of Fisheries Document 813), pp. 1–9.

Tankersley, R. A. (1996). Multipurpose gills: effect of larval brooding on the feedingphysiology of freshwater unionid mussels. Invertebrate Biology 115, 243–255.

Tankersley, R. A. & Dimock, R. V. (1993). The effect of larval brooding onthe filtration rate and particle-retention efficiency of Pyganodon cataracta (Bivalvia:Unionidae). Canadian Journal of Zoology 71, 1934–1944.

Thompson, R. J. (1984). Production, reproductive effort, reproductive value andreproductive cost in a population of the blue mussel Mytilus edulis from a subarcticenvironment. Marine Ecology Progress Series 16, 249–257.

Trouve, S. & Morand, S. (1998). Evolution of parasites’ fecundity. International Journal

of Parasitology 17, 1817–1819.Tuomi, J., Hakala, T. & Haukioja, E. (1983). Alternative concepts of reproductive

effort, costs of reproduction, and selection in life history evolution. American Zoologist

23, 25–34.Vaughn, C. C. & Hakenkamp, C. C. (2001). The functional role of burrowing

bivalves in freshwater ecosystems. Freshwater Biology 46, 1431–1446.Vaughn, C. C., Nichols, S. J. & Spooner, D. E. (2008). Community foodweb

ecology of freshwater mussels. Journal of the North American Benthological Society 27,409–423.

Weaver, L. R., Pardue, G. B. & Neves, R. J. (1991). Reproductive biology and fishhosts of the Tennessee clubshell Pleurobema oviforme (Mollusca:Unionidae) in Virginia.American Midland Naturalist 126, 82–89.


766 Wendell R. Haag

Weiss, J. L. & Layzer, J. B. (1995). Infestations of glochidia on fishes in the BarrenRiver, Kentucky. American Malacological Bulletin 11, 153–159.

White, C. R. & Seymour, R. S. (2004). Does basal metabolic rate contain auseful signal? Mammalian BMR allometry and correlations with a selection ofphysiological, ecological, and life-history variables. Physiological and Biochemical Zoology

77, 929–941.Whittingham, M. J., Stephens, P. A., Bradbury, R. B. & Freckleton, R. P.

(2006). Why do we still use stepwise modelling in ecology and behaviour? Journal of

Animal Ecology 75, 1182–1189.Whittington, I. D. (1997). Reproduction and host location among the parasitic

platyhelminthes. International Journal of Parasitology 27, 705–714.Williams, G. C. (1966). Natural selection, the cost of reproduction and a refinement

of Lack’s principle. American Naturalist 100, 687–690.Williams, J. D., Bogan, A. E. & Garner, J. T. (2008). Freshwater Mussels of

Alabama and the Mobile Basin in Georgia, Mississippi and Tennessee. Universityof Alabama Press, Tuscaloosa.

Winemiller, K. O. & Rose, K. A. (1992). Patterns of life-history diversification inNorth American fishes: implications for population and regulation. Canadian Journal

of Fisheries and Aquatic Science 49, 2196–2218.Wood, E. M. (1974). Development and morphology of the glochidium larvae of

Anodonta cygnea (Mollusca: Bivalvia). Journal of Zoology (London) 173, 1–13.Yeager, B. L. & Neves, R. J. (1986). Reproductive cycle and fish hosts of the

rabbit’s foot mussel, Quadrula cylindrica strigillata (Mollusca:Unionidae) in the UpperTennessee River drainage. American Midland Naturalist 116, 329–340.

Young, M. & Williams, J. (1984). The reproductive biology of the freshwater pearlmussel Margaritifera margaritifera in Scotland. I. Field Studies. Archiv fur Hydrobiologie

99, 405–422.

(Received 31 January 2012; revised 22 January 2013; accepted 25 January 2013; published online 28 February 2013)


the role of fecundity and reproductive effort in defining

Documents