ORIGINAL PAPER

Abiotic factors controlling the establishment and abundanceof the invasive golden mussel Limnoperna fortunei

Marcia D. Oliveira • Debora F. Calheiros •

Claudia M. Jacobi • Stephen K. Hamilton

Received: 9 November 2008 / Accepted: 24 August 2010 / Published online: 29 September 2010

� Springer Science+Business Media B.V. 2010

Abstract Limnoperna fortunei (Bivalvia, Mytilidae),

a freshwater bivalve native to Southern Asia, has been

an invasive species in South America since 1991. It

spread upstream in the La Plata basin reaching the

Paraguay River in the vicinity of the Pantanal wetland,

Brazil, around 1998. The role of abiotic factors in

controlling establishment and abundance of this spe-

cies is not well known, making projections of its risk of

further spread difficult. This study evaluates the

importance of abiotic factors to L. fortunei populations

established in rivers of the Pantanal, focusing on larval

and juvenile densities and taking advantage of a wide

range of seasonal variability in water temperature,

flow, dissolved oxygen, and suspended material.

Temperature, river stage (influencing several water

characteristics) and water velocity are the main

variables related to the larval and juvenile densities.

In the Pantanal, environmental variables vary over a

broader range compared with other South American

locations, subjecting L. fortunei to oxygen depletion,

low calcium, low pH, and high water velocity and

suspended solids, associated with low chlorophyll

a concentrations. The combined effect of several of

these conditions may explain the relatively low

densities in some Pantanal sites. However, they

probably will not prevent the persistence of popula-

tions in the Pantanal and the eventual establishment of

viable populations in upriver systems connected to the

Pantanal. These results are pertinent not only to this

species but also to other aquatic invasive invertebrates

whose expansion may be limited by thermal extremes,

episodic oxygen depletion, and waters that are too

dilute or acidic for optimal biocalcification.

Keywords Aquatic invasive species � Freshwater

mollusks � Wetlands � South America � Oxygen �Calcium

Introduction

The freshwater mytilid bivalve Limnoperna fortunei

(Dunker, 1857), the golden mussel, is an invasive

species in South America that in many respects is

comparable to the invasive dreissenid zebra mussel,

Dreissena polymorpha (Pallas), in North America

(Ricciardi 1998). L. fortunei was first detected in

South America in 1991 in brackish waters of the La

M. D. Oliveira (&) � D. F. Calheiros

Embrapa Pantanal, Rua 21 de Setembro, 1880, CP 109,

CEP 79320 900 Corumba, MS, Brazil

e-mail: mmarcia@cpap.embrapa.br

C. M. Jacobi

Federal University of Minas Gerais, Av. Antonio Carlos,

6627, CEP 31270-901 Belo Horizonte, MG, Brazil

S. K. Hamilton

W.K. Kellogg Biological Station and Department

of Zoology, Michigan State University, 3700 E. Gull Lake

Drive, Hickory Corners, MI 49060-9516, USA

e-mail: hamilton@kbs.msu.edu

123

Biol Invasions (2011) 13:717–729

DOI 10.1007/s10530-010-9862-0

Plata River, Argentina (Pastorino et al. 1993). It was

carried upstream in subsequent years, presumably

aided by ship and barge traffic in the Paraguay-Parana

waterway, reaching the Pantanal wetland in Brazil by

1998 (Oliveira et al. 2006). Several authors have

noted the potential for L. fortunei to invade North

American waters, where it could occupy regions that

are too warm for the dreissenid mussels (Ricciardi

1998; Oliveira et al. 2010a).

In spite of some controversy about the taxonomy

of genus Limnoperna, and consequently about its

distribution (Lee and Morton 1983; Morton 1996),

L. fortunei is thought to be native to two major river

basins in China, the Pearl (Zhujiang) and Yangtze

(Changjiang) rivers (Morton 1996). In the Yangtze

River basin, where it has a broad distribution

(Lee and Morton 1983), L. fortunei has been referred

to as Limnoperna lacustris (Xie et al. 1999; Wang

et al. 1999, 2007). Introductions into Hong Kong,

Japan, and South America were mentioned by

Morton (1975, 1977), Magara et al. (2001), and

Darrigran et al. (1999), respectively. Morton (1987)

defined L. fortunei as an r-selected species, adapted to

the invasive colonization of a wide range of aquatic

environments, with important characteristics of a

successful invader such as being short-lived, capable

of rapid growth, and having a high fecundity, aside

from broad physiological tolerance to several abiotic

factors that often limit aquatic invertebrates.

L. fortunei is recognized as a problem since it

promotes large-scale fouling in water supply systems

and power plants, attaining densities greater than

50,000 ind m-2 (Darrigran et al. 2003). Morton

(1975) suggested that L. fortunei has a potential for

causing problems even in its native environments,

being considered a pest that fouls boats and jetties in

the Pearl River. There are no reports of fouling by

Limnoperna in the Yangtze River, where only

1.0–40.0 ind m-2 of L. lacustris (L. fortunei accord-

ing to Morton 1996) were recorded by Wang et al.

(2007).

In general, exotic species with broad distribution

in their native range should also achieve broad

distributions in comparable introduced habitats if

given enough time (Guo 2006). Modeling the distri-

bution potential based on their ecological niche has

been widely used to predict the spread of invasive

species (Drake and Bossenbroek 2004; Bossenbroek

et al. 2007; Kluza and McNyset 2005). Therefore, to

build realistic models and validate them it is funda-

mental to have reliable environmental data and

scientific understanding of the species biology

(Peterson and Vieglais 2001). Information on abiotic

limiting factors is essential to predict where an

invasive species is likely to flourish. In the case of

L. fortunei its broad tolerance to chemical and

physical variables that usually inhibit other aquatic

invertebrates makes it difficult to identify limiting

abiotic factors that would control or prevent its spread.

Ricciardi (1998) suggested some limits of L. fortunei

tolerance to chemical and physical aquatic variables,

based on characteristics of habitats invaded by this

species in eastern Asia and southern South America.

Inferences about these abiotic limits are mostly based

on observations of adult mussels. However, larvae and

juveniles may be more sensitive than adults to some of

these variables, and there is very little information on

how abiotic factors affect immature stages in natural

environments, with direct implications for adult

mussel density and invasive potential.

The objective of this study is to test our current

understanding of the environmental tolerance of

L. fortunei, with an emphasis on initial colonization

and immature stages. We examine the population

dynamics of L. fortunei in rivers of the Pantanal

wetland, a tropical floodplain-river complex where

abiotic factors such as temperature, oxygen and

calcium concentrations vary over a greater range than

has been reported for any other South American

location containing this mussel. We describe the

environment where L. fortunei is established in the

Pantanal, compare water conditions with the tolerance

limits presented in the literature, and examine the

influence of abiotic factors on L. fortunei colonization

and density. These results are pertinent not only to this

species but also to other aquatic invasive invertebrates

whose expansion may be limited by thermal extremes,

episodic oxygen depletion, and waters that are too

dilute or acidic for optimal biocalcification.

Methods

Study area

The Brazilian portion of the upper Paraguay River basin

covers about 362,000 km2, within which the Pantanal

wetland, the largest contiguous floodplain area in the

718 M. D. Oliveira et al.

123

world, extends over about 140,000 km2 in Brazil

(PCBAP (Plano de Conservacao da Bacia do Alto

Paraguai) 1997) (Fig. 1). The climate in the region is

tropical with marked wet and dry seasons. The maxi-

mum daily temperature averages 32�C in the Pantanal,

ranging from 28 to 34�C. The minimum monthly

temperature averages about 20�C (PCBAP 1997).

The Paraguay River (mean depth 8.6 m and width

varying from 100 to 500 m) flows over 3,800 km

from north to south collecting the waters of large

tributaries on its left margin, including the Cuiaba,

Taquari and Miranda rivers (Fig. 1). Its mean annual

discharge is about 300 m3 s-1 in the northern Pant-

anal and 1,555 m3 s-1 after it receives water from its

major tributaries in the Pantanal.

The rivers and floodplains of the region exhibit a

marked and protracted seasonal flood pulse, with

water levels varying seasonally over a range of 3–5 m

in the southern Pantanal. Flooding occurs as a result

of river overflow, local precipitation or a combination

of both. The contact between the Paraguay River and

its tributaries with the extensive floodplains at the

beginning of the rising water phase often causes

dramatic changes in water chemistry, a natural

phenomenon labeled as ‘oxygen depletion events’

by Hamilton et al. (1997) and Calheiros and Hamilton

(1998) in reference to one of its main consequences.

Oxygen depletion to hypoxic or anoxic conditions,

accompanied by a decrease in pH and other chemical

changes, extends to the Paraguay River and some-

times results in massive fish kills in the rivers and

associated floodplain waters.

The Miranda River is one of the main tributaries of

the Paraguay River within the Pantanal and drains

about 47,000 km2 (PCBAP 1997), including upland

tributaries with clear waters that originate in lime-

stone formations of the Bodoquena hills. The lower

Miranda river is about 150 m wide and its discharge

is around 86 m3 s-1 (PCBAP 1997). Its flow regime

responds more directly to rain events than the

Paraguay River, so the flooding phase takes place

earlier, from November or December to April, and is

influenced by backwater effects of the Paraguay

River as well as rainfall in its drainage basin.

Data collection and analysis

The study area included two reaches of the Paraguay

River (Porto Esperanca site/PR-PE, 19�360S and

57�260W, and Corumba site/PR-CR, 18�590S and

57�420W), and one section in the lower Miranda

River (Passo do Lontra, MR, 19�340S and 57�14.70W)

close to the confluence with the Paraguay River.

These sites represent a gradient of L. fortunei density

and calcium concentration, with the highest densities

of juveniles and adults at Porto Esperanca, and the

lowest in the Miranda river, which is rich in calcium

compared to the Paraguay River (Oliveira 2009). The

establishment of L. fortunei was confirmed in 1998 in

the Paraguay River and 2002 in the Miranda River.

River stage (water level) data for the Paraguay River

(19�020S, 57�330W) were obtained from daily readings

carried out by the Brazilian Navy and by the National

Water Agency (ANA), and stage data for the Miranda

River were obtained from the Federal University of

Mato Grosso do Sul. Substrata for larval settlement

were available at all sites, in the form of hard natural

rock surfaces or artificial structures, and thus substrata

were not considered a limiting factor.

Samples for analysis of limnological variables and

larval and juvenile density were taken at each site

monthly from January 2004 to November 2007. For

water analysis, subsurface water samples were taken

with a Van Dorn sampler. Water temperature,

dissolved oxygen concentration (DO), electrical con-

ductivity and pH were measured in situ with a water

quality sonde (YSI). Water transparency was esti-

mated using a Secchi disc. Total alkalinity was

analyzed by the Gran titration method (Gran 1952)

and free CO2 was calculated following Kempe

(1982), with modifications by Hamilton et al.

(1995). Chlorophyll a was analyzed following

Marker et al. (1981). Total suspended solids (TSS),

inorganic (IM) and organic matter (OM), and organic

matter percentage were determined by gravimetric

and loss-on-ignition methods, and calcium concen-

tration was analyzed by atomic absorption, all based

on APHA (1998). Total organic carbon (TOC) was

analyzed using a Shimadzu Carbon Analyzer. Water

velocity was measured with a portable flowmeter

(Marsh-McBirney, Inc, model 2000).

To determine juvenile density, we counted indi-

viduals of 0.3–3.0 mm shell length settled after about

30 days onto artificial substrata. Each substratum was

a 100 cm2 Nylon net (1.0 mm mesh) installed inside

a PVC tube with a diameter of 12 cm (or a PET

bottle), closed with a 5.0-mm mesh net to prevent fish

predation (predation represents up to 35% of total

Abiotic factors controlling the establishment 719

123

Brazil

Argentina

Peru

Bolivia

Chile

Colombia

Venezuela

Paraguay

Ecuador

Uruguay

GuyanaSuriname

Chile

French Guiana

Fig. 1 Location of the Pantanal wetland in South America (left panel), Upper Paraguay Basin drainage and floodplain (gray area;

right panel). Black circles indicate study sites

720 M. D. Oliveira et al.

123

losses after settlement; Oliveira, unpublished data).

Six of these units were attached to supports of

existing bridges or docks in each site. The Nylon net

was replaced each month and settled organisms were

preserved in 70% ethanol for counting.

For larval counts, three samples of 300 l were

concentrated in a 36-lm mesh plankton net, and

preserved in 70% ethanol. Five aliquots of 2-ml

subsamples were taken using a Hensen-Stempel pipette

and all larvae in each subsample were counted in a

Fig. 1 continued

Abiotic factors controlling the establishment 721

123

reticulated camera, with the aid of an 809 stereomicro-

scope (Wetzel and Likens 2001).

The adult condition index, shell length:weight

ratio, and shell growth rate were measured to

compare the condition of adult mussels in the

Miranda river with those in the Paraguay River-PE,

where adult density is much higher. To calculate the

Condition Index (CI = tissue weight 9 100/shell

length) we measured shell length of mussels larger

than 13.0 mm (average length of 14.2 mm in the

Paraguay River-PE and 17.2 mm in the Miranda

River). Mussel tissue was separated from the shell

and both were dried to constant weight at 80�C. We

also conducted in situ assays of shell growth rate,

placing 30 mussels ranging between 6.0 and 10.0 mm

in shell length in a cage protected with Nylon mesh in

the Paraguay River-PE and Miranda River. Mussel

shell growth was were measured every 30 days over

210 days in the Paraguay River and 320 days in the

Miranda River, from June 2006 to July 2007.

To examine how the flood pulse affected mussel

densities we estimated the Spearman correlation

between juvenile density and limnological variables,

combining data from all sites. We also examined the

relationship between river stage and other variables.

We were particularly interested in how water temper-

ature and oxygen depletion events affect L. fortunei

recruitment (measured as juvenile density). These two

variables were combined in a two-factor logistic

regression to evaluate the probability of L. fortunei

recruitment. Analyses and graphs were performed

using SYSTAT 11 software (Wilkinson 2004).

Results

The Paraguay River stage (at Corumba and Porto

Esperanca) varied seasonally over a range of 3–4 m,

and the Miranda River stage varied over 2–3 m

(Fig. 2). In the Paraguay River, floodplain inundation

was generally observed from April to August, whereas

in the Miranda River the flood pulse began earlier but

was more variable. During much of the year, dissolved

O2 concentrations were above 5.0 mg l-1, or 50–60%

of atmospheric equilibrium, in all three sites. The

oxygen depletion events, in which dissolved oxygen

fell as low as 0.0 mg l-1 on specific measurement

dates, were observed in all years at the start of flooding

(rising water) in the Paraguay River, and during the

high water phase in the Miranda River (Fig. 2).

In the Miranda River pH values (6.0–7.8, mean 7.2)

were higher than at the Paraguay River sites (5.4–7.1,

mean 6.4 at Corumba site) (Fig. 2). The Porto Espe-

ranca site was not included in Fig. 2 because river stage

and dissolved oxygen seasonal patterns were similar to

the Corumba site, and the Porto Esperanca site was

Fig. 2 Seasonal variation

of river stage, dissolved

oxygen (mg l-1), pH and

calcium (mg l-1)

concentration in the

Paraguay River (Corumba

site) and Miranda River.

Lines represent the mean

(±SE) of monthly samples

from 2004 to 2007. Note

that the scale is different for

river stage and calcium

between the left and right

panels

722 M. D. Oliveira et al.

123

intermediate in pH and calcium values (5.9–7.0 and

1.8–8.9 mg l-1, respectively) because it is below the

confluence of the Miranda River. A clear decrease of

pH to 6.0 was observed in the Miranda River at the time

of its lowest oxygen concentration. In the Paraguay

River the minimum pH was 5.4 at Corumba and 5.9 at

Porto Esperanca. Some acidification was also observed

during falling water (November to December) in the

Paraguay River.

Free CO2 was highest during oxygen depletion

events and is the predominant factor controlling pH

changes. The maximum free CO2 concentrations

were 68 mg l-1 (Porto Esperanca site) and 87 mg l-1

(Corumba site) in the Paraguay River, and

153 mg l-1 in the Miranda River. The normal CO2

concentration throughout the year, i.e. excluding

oxygen depletion events, was below 20 mg l-1.

Calcium concentrations were relatively high in the

Miranda River, varying between 10 and 23 mg l-1.

Calcium concentrations were much lower over the

course of the year in the Paraguay River (between 1.0

and 6.0 mg l-1 at the Corumba site, Fig. 2), and were

lowest from October to February.

Water transparency in the Miranda River

(10–90 cm Secchi depths) was lower than in the

Paraguay River (20–190 cm). Concentrations of TSS

up to 120 mg l-1 occurred from October to December

in the Miranda River, although the mean values were

between 10 and 50 mg l-1. Both Paraguay River sites

showed TSS concentrations of less than 20 mg l-1

along the year, and reached around 50 mg l-1 from

October through December. Ratios of organic matter

(OM) to inorganic matter (IM) indicate predominance

of particulate inorganic material throughout the year

(Paraguay River = 4.5 ± 2.4 mg l-1 at Corumba

and 4.3 ± 2.1 mg l-1 at Porto Esperanca; Miranda

River = 5.2 ± 1.9 mg l-1).

The concentration of chlorophyll a was low at all

sites, averaging less than 1.0 lg l-1 (Paraguay River-

PE: mean 0.79 ± 0.76 lg l-1; CR: 0.78 ± 0.98 lg l-1

and Miranda River: 0.98 ± 0.98 lg l-1). TOC was

slightly higher in the Miranda River (mean

1.7 ± 0.9 mg l-1) compared to the Paraguay River at

Porto Esperanca (1.3 ± 1.0 mg l-1) and Corumba

(1.2 ± 0.8 mg l-1). Ratios of TOC to chlorophyll

a indicate predominance of detritus in both Miranda

(82,204 ± 21,084) and Paraguay rivers (115,033 ±

36,128 at Corumba and 17,715 ± 75,565 at Porto

Esperanca).

Water velocity, usually between 0.6 and

0.8 m s-1, was relatively high in the Miranda River.

At Porto Esperanca, on the other hand, mean monthly

values were low (\0.2 m s-1) and fairly uniform.

Comparatively, water velocity at Corumba fluctuated

more, ranging between 0.1 and 0.6 m s-1 (Fig. 3).

Both planktonic larval density and juvenile settle-

ment lasted around seven to 8 months (Fig. 4) in both

Paraguay River sites, associated with seasonally high

water temperature, but showed a poorly defined

pattern in the Miranda River. From February to

March no larvae or juveniles were recorded in the

Miranda River. In general, larval settlement was

observed from September to April with peaks from

October to February. Between May and August

settlement usually was not observed. After Septem-

ber, when water temperature rose above 8�C, larvae

and juveniles were again observed, except in the

Miranda River, where larvae occurred in August and

September. In general, the reproductive period was

coincident with temperatures between 28.0 and

32.5�C (maximum water temperatures of 35�C were

recorded in the Paraguay River in October and

November).

The relationship between juvenile density and

most limnological variables was weak, as shown by

Spearman’s correlation (Table 1). The best correla-

tions with juvenile density were observed for water

velocity, water temperature and river stage. River

stage covaries with variables such as dissolved

oxygen, pH, free CO2, total suspended solids and

organic matter percentage, as shown by the correla-

tions in Table 1.

Fig. 3 Water velocity in the Paraguay and Miranda rivers.

Lines are the means of monthly samples from 2004 to 2007,

indicated by symbols: PR-PE (open square), PR-CR (fulltriangle) and MR (cross)

Abiotic factors controlling the establishment 723

123

Although dissolved oxygen appears not to be

correlated with juvenile density, graphical analysis

(Fig. 2) and evidence of adult mussel mortality

during oxygen depletion events (Oliveira et al.

2006) suggest that this variable is important to larval

and juvenile density. The results from the logistic

regression (Fig. 5) suggest that recruitment increases

with dissolved oxygen concentration (t-ratio =

-5.245, P \ 0.001, q2 = 0.255, n = 145). The

probability of recruitment was less than 10% when

DO was close to 0.0 mg l-1, even at the temperature

of 28�C.

Larval densities were similar among Paraguay

River sites, and lower in the Miranda River. Juvenile

density in the Paraguay River-PE was much higher

than at the other sites (Table 2, Fig. 4). Ratios of

mean juvenile to larval densities suggested that

settlement is lower in the Miranda River and

Paraguay River at CR, compared to the Paraguay

River at PE (Table 1).

Higher condition index values (IC = 6.2 ± 2.9)

and slightly smaller shell length:weight ratios (0.2 ±

0.1) were recorded in the Miranda River compared to

Paraguay River-PE (IC = 4.6 ± 1.2 and shell

length:weight = 0.3 ± 0.1). Shell length monitoring

suggested that shell growth was similar in the

Miranda River and Paraguay River-PE with

increments of about 0.05 mm day-1, resulting in

shell lengths of about 17.0–18.0 mm in the first year

for both sites.

Discussion

According to Morton’s studies of L. fortunei in Hong

Kong (1975, 1977, 1982), this subtropical freshwater

species has broad tolerance to chemical and physical

variables. The results of this study and related studies

we have conducted in the Pantanal region (Oliveira

2009; Oliveira et al. 2010b) broaden the known range

of abiotic factors that this invasive species can

tolerate. We also document reproductive and settle-

ment patterns in relation to seasonal changes in

temperature and other abiotic factors, revealing how

this species maintains its populations in rivers subject

to seasonal variability in flow and turbidity as well as

episodic oxygen depletion.

As is common in ecological field studies, it is

difficult to point to any single variable as the most

important controlling factor because the potential

stressors are variable in time and space, and tend to

coincide or overlap in time. Furthermore, the popu-

lations under study varied in their time of establish-

ment and the antecedent conditions to which they had

Fig. 4 Mean (±SE) seasonal variation of water temperature (�C) and L. fortunei larval concentrations (ind m-3) and juvenile

densities (ind m-2) in the Paraguay and Miranda rivers from 2004 to 2007. Note the difference in scales

724 M. D. Oliveira et al.

123

been exposed. Nonetheless, documentation of sur-

vival, growth and reproduction in the face of

potentially stressful conditions reveals ranges of

tolerance and thereby helps establish thresholds.

We concur with Ricciardi (1998) and Montalto

and Marchese (2003) that adults might tolerate water

temperatures as high as 35�C because this was the

maximum temperature observed in the Paraguay

River. In the Paraguay River L. fortunei is exposed

to relatively high temperatures (above 30�C) com-

pared to its native environment in China (maximum

monthly mean was up to 27.8�C) (Ho et al. 2003) and

other environments in South America, where maxi-

mum monthly water temperature is usually below

30�C (Boltovskoy et al. 2009). No information was

found in the literature about the upper limit of larval

tolerance related to water temperature. Cataldo et al.

(2005) tested the larval development in temperatures

up to 30�C, and faster development was observed at

temperatures around 28.0–30.0�C than 25.0�C. In the

Pantanal larval development may take place at

temperatures around 32.0�C, a common temperature

during the reproductive period.

Both water temperature and river stage signifi-

cantly influenced L. fortunei density fluctuations,

although in different ways. While temperature seems

to govern the reproductive cycle, the seasonality of

flood pulse might be related to changes in chemical

variables such as oxygen depletion events, which isTa

ble

1S

pea

rman

’sq

corr

elat

ion

coef

fici

ents

bet

wee

nju

ven

ile

den

sity

and

lim

no

log

ical

var

iab

les

Juv

enil

e

den

sity

Wat

erv

elo

city

tem

per

atu

re

Wat

er

tem

per

atu

re

Riv

er

stag

e

OM

(%)

Alk

alin

ity

TS

Sp

HC

alci

um

Fre

e

CO

2

Ch

loro

ph

yll

aD

isso

lved

ox

yg

en

Juv

enil

ed

ensi

ty1

––

––

––

––

––

–

Wat

erv

elo

city

-0

.46

82

1–

––

––

––

––

–

Wat

erte

mp

erat

ure

0.3

55

50

.04

01

1–

––

––

––

––

Riv

erst

age

-0

.31

75

-0

.15

54

0.0

54

71

––

––

––

––

OM

(%)

-0

.02

68

0.3

43

5-

0.0

06

5-

0.5

74

01

––

––

––

–

Alk

alin

ity

-0

.25

36

0.5

31

3-

0.1

07

6-

0.3

62

20

.33

03

1–

––

––

–

TS

S0

.15

75

0.2

67

80

.01

14

-0

.70

49

0.8

46

20

.20

00

1–

––

––

pH

-0

.14

51

0.3

84

8-

0.2

26

6-

0.6

18

20

.50

29

0.4

34

50

.53

43

1–

––

–

Cal

ciu

m-

0.1

44

70

.55

08

0.0

12

9-

0.4

41

40

.34

97

0.9

00

90

.26

17

0.4

43

31

––

–

Fre

eC

02

-0

.08

10

-0

.10

05

0.1

18

90

.46

49

-0

.42

09

0.0

01

5-

0.5

12

6-

0.7

48

2-

0.0

22

21

––

Ch

loro

ph

yll

a-

0.0

09

40

.20

55

-0

.27

53

-0

.08

62

-0

.07

75

0.2

81

8-

0.0

83

40

.05

31

0.2

33

90

.12

14

1–

DO

-0

.00

50

0.1

37

7-

0.4

72

8-

0.6

24

30

.48

10

80

.11

32

0.6

12

30

.51

41

0.1

17

4-

0.4

67

40

.16

71

1

All

dat

afr

om

the

Par

agu

ayR

iver

(at

Co

rum

ba

and

Po

rto

Esp

eran

casi

tes)

and

Mir

and

aR

iver

wer

eco

mb

ined

(n=

14

5sa

mp

lin

gs)

Fig. 5 Simulations of L. fortunei recruitment as a function of

water temperature and DO based on a two-factor logistic

regression. Concentrations of DO were fixed and recruitment

probability was calculated for different temperatures. All data

from the Paraguay River (PE and CR sites) and Miranda River

were combined for this analysis

Abiotic factors controlling the establishment 725

123

known to cause adult mussel die-offs (Oliveira et al.

2006), directly affecting density.

Both larval and juvenile densities indicated almost

continuous reproduction for about 7–8 months in the

Pantanal, a span that resembles that of other popu-

lations established in South America (Boltovskoy and

Cataldo 1999, Darrigran et al. 2003; Santos et al.

2005). Continuous occurrence of planktonic larvae

6–10 months of the year was observed by Boltovskoy

et al. (2009) in most lentic water bodies. These

authors suggested that larval density might be

regulated by the time of reproduction, which in turn

is influenced by water temperature and food supply.

The reproductive pattern of L. fortunei in South

America is quite different from results described by

Iwasaki and Uryu (1998), who compared data from

central China, Hong Kong and Japan. The processes

of gametogenesis and spawning of L. fortunei have

been related to maximum and minimum temperature

or to abrupt changes in water temperature (Morton

1982; Darrigran et al. 2003; Santos et al. 2005). We

suggest that abrupt changes, rather than absolute

values, may elicit the spawning of L. fortunei in the

Pantanal, since mean winter temperatures (which

remained above 20�C) were not limiting.

Although river stage has no direct effect on

density, unless the mussel habitat is exposed to air

during low waters, the seasonal flood pulse causes

important changes in the chemical variables. Oxygen

depletion events usually occur at the end of the

reproductive period in the Paraguay River and in the

middle of the reproductive period (February to

March) in the Miranda River, when low densities or

absence of larvae and juveniles were recorded. In the

latter case, the effect might be more severe -stopping

reproduction- and could explain the low densities in

the Miranda River; we observed that the probability

of reproduction was less than 10% when DO was

close to 0.0 mg l-1. In the Pantanal low DO, possibly

exacerbated by high free CO2, was considered the

main cause of fish mortality during oxygen depletion

events (Hamilton et al. 1995; Calheiros and Hamilton

1998). Information about free CO2 concentration is

not provided for any place where L. fortunei is

present. Oxygen depletion events have been shown to

affect L. fortunei density via direct mortality of adults

(Oliveira et al. 2006). Effects of anoxic conditions on

larval development have not been documented in the

literature, but low dissolved oxygen levels affecting

L. fortunei gametogenesis and spawning were cited

by Morton (1982). The tolerance limits of each

variable were difficult to estimate because of the

interaction among factors, so the minimum dissolved

oxygen requirements of young and adults of

L. fortunei in the Pantanal wetland remain unclear.

L. fortunei is well established in the Paraguay

River (Corumba site) where mean pH is around 6.4,

the lower limit suggested by Ricciardi (1998),

although it may drop to 5.5 for periods of days to a

few weeks during oxygen depletion events. Kara-

tayev et al. (2007) suggested 5.5 as the lower pH limit

to L. fortunei but provided no example of established

populations in habitats with similar characteristics,

and Montalto and Marchese (2003) recorded 80% of

adult survival in pH 5.0 after 96 h of laboratory tests,

but this period is too short to extrapolate to a natural

environment.

The Pearl River in China, which is thought to be a

native environment of L. fortunei, carries about

33 mg l-1 of calcium (Zhang et al. 2007). The lower

limit of 3.0 mg l-1 proposed by Ricciardi (1998) was

likely based on data from Sum Chum Reservoir (China),

the site where L. fortunei has been reported at the lowest

calcium concentrations: 2.4 and 4.8 mg l-1 with a mean

Table 2 Larval, juvenile, and adult density, and juvenile to larva ratio

Larval density Juvenile density (30-day recruitment) Adult density* Juvenile:

larval ratioN ind m-3 N ind m-2 N ind m-2

Paraguay, Corumba 46 1.629 ± 4.571 46 3.149 ± 6.928 12 6.539 ± 5.904 1.93

Paraguay, Porto

Esperanca

41 1.772 ± 4.033 46 88.523 ± 153.349 12 143.134 ± 62.692 49.96

Miranda 42 990 ± 2.292 46 377 ± 682 12 585 ± 358 0.38

Data are mean ± SD of monthly samples from Paraguay River, at Corumba and Porto Esperanca sites, and Miranda River, from 2004

to 2007

*Source: Oliveira (2009)

726 M. D. Oliveira et al.

123

pH of 6.4 (Morton 1975; Ricciardi 1998). In the

Paraguay River at Corumba calcium concentrations

were quite low, varying between 1.0 and 6.0 mg l-1,

and pH can occasionally be more acidic than at Sum

Chum. Indications that mussels are tolerant to ionically

dilute water over short time intervals come from lab

experiments by Deaton et al. (1989) who verified that

mussels retained byssal attachments and filtering activ-

ities by the end of 1 week in deionized water.

In the relatively calcium-rich water of the Miranda

River, other negative factors such as high water

velocity, low availability of food, and high concen-

tration of suspended sediments might explain the low

larval settlement. The recent colonization in the

Miranda River (around 5 years) might contribute to

low juvenile density, but does not explain low density

in the Corumba site (colonization time about

7–10 years). Besides that, no clear increases in larval

or juvenile densities observed from 2004 to 2007

(Oliveira 2009).

The effects of turbidity and low food quality and

quantity on the establishment of Dreissena polymor-

pha are well known (Madon et al. 1998; Schneider

et al. 1998; Allen et al. 1999; Baines et al. 2005,

2007). Schneider et al. (1998) suggested that an

organic:inorganic matter ratio lower than 0.5 can be

considered low food quality. This threshold is related

to the inability of D. polymorpha to selectively ingest

organic material as food quality declines. Availability

of food was proposed by Boltovskoy et al. (2009) to

be a major factor controlling the reproductive activity

of L. fortunei, although other factors such as pH and

calcium were not explored. These authors also

reported low larval density in turbid water (transpar-

ency around 0.1–0.4 m) with scarce phytoplankton.

In the Miranda River mussel reproduction (October to

February) might have been negatively affected by

high concentration of suspended solids, low

organic:inorganic matter ratio (0.2), and low chloro-

phyll a occurring at the same time. However,

L. fortunei attains high densities in the Guaiba Lake

(Southern Brazil) where suspended solids concentra-

tions range up to 214 mg l-1 and water transparency

is around 0.4 m. The difference in settlement could

be because chlorophyll a is higher in Guaiba Lake

(0–23.9, mean 3.5 lg l-1; Santos et al. 2005) than in

Pantanal waters.

Shell growth at rates similar to those reported for

Argentine populations (Boltovskoy and Cataldo 1999)

and similar condition indices in the Miranda and

Paraguay River at PE suggest that the environmental

conditions are suitable for adults in both sites. In the

Miranda River the condition index is higher than those

of mussels from the Itaipu reservoir in Brazil, where

chlorophyll a ranges between 4 and 22 lg l-1 (Silva

2006), suggesting that L. fortunei uses other sources of

energy besides phytoplankton in the Pantanal rivers,

where this resource is scarce (Oliveira and Calheiros

2000). Detrital organic carbon can provide another

source of energy to L. fortunei in the Miranda River,

considering that concentrations of TOC in the Pantanal

waters are around 1–2 mg l-1, analogous to the Parana

River (Depretris and Kempe 1993) where L. fortunei is

successfully established.

In the Miranda River, the good mussel condition

index and greater shell length support the idea that

low food combined with high concentration of

inorganic sediments are probably not a limiting

factor for growth of adults. However, they might be

detrimental to larval stages, where survival seems to

be low in the Miranda River compared to the

Paraguay River. Another negative factor in the

Miranda River might be the higher water velocity,

which ranged from 0.4 to 1.1 m s-1. Although not

harmful to adults it seems to have hindered settlement

onto artificial substrata. Nagaya et al. (2001) found

that a fluid velocity around 1.0 m s-1 affected the

attachment of L. fortunei juveniles.

We confirmed that the environments where

L. fortunei lives in the Paraguay and Miranda rivers

are quite extreme compared to other places where

L. fortunei occurs. Larval density was relatively

lower in the Pantanal compared to data summarized

in Boltovskoy et al. (2009), which likely contribute to

the lower juvenile and adult densities. The effects of

multiple abiotic stressors on larval development in

natural environments are poorly known, and further

research is necessary. Most studies are more con-

cerned with adults than larval stages. Additional

information about larval tolerance or the effects of

environmental variables on reproductive phases such

as gametogenesis and spawning is critical in order to

formulate effective measures to control L. fortunei.

We suggest that L. fortunei can withstand lower

limits of pH and calcium than those proposed by

Ricciardi (1998). The upper water temperature limits

were in the range suggested by Ricciardi (1998), but

tolerance to low dissolved oxygen was not possible to

Abiotic factors controlling the establishment 727

123

determine because oxygen depletion events coincided

with changes in other variables. Extreme conditions

in the Pantanal are likely responsible for maintaining

relatively low densities of L. fortunei, but the

populations persist and can lead to further expansion

into the more hospitable waters in upper Paraguay

River basin locations where constant vessel traffic

exerts a considerable propagule pressure and seasonal

extremes may be less marked.

Acknowledgments This work was supported by The National

Council for Scientific and Technological Development (PELD

and CTHIDRO programs), and Embrapa Pantanal. We are

grateful to the Pantanal Matogrossense National Park for the field

assistance, and the Federal University of Minas Gerais/ECMVS

and US Fish and Wildlife Service for support. We are also

grateful to Maria D. Oliveira, Egidia do Amaral, Waldomiro L.

Silva, and Isac Teixeira for technical assistance, and to the

students Viviane Eilers, Izabella Xavier, Claudiane Santos, and

Tatiane Cielo for the great collaboration in field work and

laboratory analyses. The valuable suggestions of two anonymous

referees are gratefully acknowledged.

References

Allen YC, Thompson BA, Ramcharan CW (1999) Growth and

mortality rates of the zebra mussel, Dreissena polymor-pha, in the lower Mississippi River. Can J Fish Aquat Sci

56:748–759

APHA (1998) Standard methods for the examination of water

and wastewater, 20th edn. American Public Health

Association, American Water Works Association, and

Water Environment Federation, Washington, DC, 1268 p

Baines SB, Fisher NS, Cole JJ (2005) Uptake of dissolved

organic matter (DOM) and its importance to metabolic

requirements of the zebra mussel, Dreissena polymorpha.

Limmnol Oceanogr 50:36–47

Baines SB, Fisher NS, Cole JJ (2007) Dissolved organic matter

and persistence of the invasive zebra mussel (Dreissenapolymorpha) under low food conditions. Limnol Ocea-

nogr 52:70–78

Boltovskoy D, Cataldo D (1999) Population dynamics of

Limnoperna fortunei, an invasive fouling mollusc, in

the Lower Parana River (Argentina). Biofouling 14:

255–263

Boltovskoy D, Sylvester F, Otaegui A et al (2009). Environ-

mental modulation of reproductive activity of the invasive

mussel Limnoperna fortunei: implications for antifouling

strategies. Austral Ecol. doi:10.1111/j.1442-9993.2009.

01974.x

Bossenbroek JM, Johnson LE, Peters B et al (2007) Forecast-

ing the expansion of Zebra Mussels in the United States.

Conserv Biol 21(3):800–810

Calheiros DF, Hamilton SK (1998) Limnological conditions

associated with natural fish kills in the Pantanal Wetland

of Brazil. Verh Internat Verein Limnol 26:2189–2193

Cataldo D, Boltovskoy D, Hermosa JL et al (2005) Tempera-

ture-dependent rates of larval development in Limnopernafortunei (Bivalvia:Mytilidae). J Moll Stud 71:41–46

Darrigran G, Penchaszadeh P, Damborenea MC (1999) The

reproductive cycle of Limnoperna fortunei (Dunker, 1857)

(Mytilidae) from a neotropical temperate locality.

J Shellfish Res 18(2):361–365

Darrigran G, Damborenea MC, Penchaszadeh P et al (2003)

Reproductive stabilization of Limnoperna fortunei (Biv-

alvia Mytilidae) after ten years of invasion in the Amer-

icas. J Shellfish Res 22:141–146

Deaton LE, Derby JGS, Subhedar N et al (1989) Osmorregu-

lation and salinity tolerance in two species of bivalve

mollusk: Limnoperna fortunei and Mytilopsis leucophae-ta. J Exp Mar Biol Ecol 133:67–79

Depretris PJ, Kempe S (1993) Carbon dynamics and sources in

the Parana River. Limnol Oceanogr 38:382–395

dos Santos CP, Wurdig NL, Mansur MCD (2005) Fases larvais

do mexilhao dourado Limnoperna fortunei (Dunker)

(Mollusca, Bivalvia, Mytilidae) na Bacia do Guaıba, Rio

Grande do Sul, Brasil. Rev Bras Zool 22:702–708

Drake JM, Bossenbroek JM (2004) The potential distribution

of Zebra Mussels in the United States. Bioscience

54(10):931–941

Gran G (1952) Determination of the equivalence point in

potentiometer titrations, Part II. Analyst 77:661–671

Guo Q (2006) Intercontinental biotic invasions: what can we

learn from native populations and habitats? Biol Invasions

8:1451–1459

Hamilton SK, Sippel SJ, Melack JM (1995) Oxygen depletion

and carbon dioxide and methane production in waters of

the Pantanal wetland of Brazil. Biogeochemistry

30:115–141

Hamilton SK, Sippel SJ, Calheiros DF et al (1997) An oxygen

depletion event and other biogeochemical effects of the

Pantanal wetland on the Paraguay River. Limnol Ocea-

nogr 42:257–272

Ho KC, Chow YL, Yau JTS (2003) Chemical and microbio-

logical qualities of the East River (Dongjiang) water, with

particular reference to drinking water supply in Hong

Kong. Chemosphere 52:1441–1450

Iwasaki K, Uryu Y (1998) Life cycle of a freshwater mytilid

mussel, Limnoperna fortunei, in Uji River, Kyoto. Jpn J

Malacol 57(2):105–113

Karatayev AY, Boltovskoy D, Padilla D et al (2007) Theinvasive bivalves Dreissena polimorpha and Limnopernafortunei: parallels, contrasts, potential spread and invasion

impacts. J Shellfish Res 26:205–213

Kempe S (1982) Long-term records of CO2 pressure fluctua-

tions in fresh waters. SCOPE/UNEP Sonderband

52:91–332

Kluza DA, McNyset M (2005) Ecological niche modeling of

aquatic invasive species. Aquat Invaders 16(1):1–7

Lee SY, Morton B (1983) The Hong Kong Mytilidae. In:

Morton B, Dudgeon D (eds) The malacofauna of Hong

Kong and Southern China II. Proceedings of the second

international workshop on the malacofauna of Hong Kong

and Southern China. Hong Kong University Press. Hong

Kong, pp 49–76

Madon SP, Schneider DW, Stoeckel JA et al (1998) Effects of

inorganic sediment and food concentrations on energetic

728 M. D. Oliveira et al.

123

processes of the zebra mussel, Dreissena polymorpha:

implications of growth in turbid rivers. Can J Fish Aquat

Sci 55:401–413

Magara Y, Matsui Y, Goto Y et al (2001) Invasion of the non

indigenous nuisance mussel, Limnoperna fortunei, into

water supply facilities in Japan. J Water Supply Res

Technol - AQUA 50:113–124

Marker AFH, Nush EA, Rai H et al (1981) The measurement of

photosynthetic pigments in freshwaters and standardiza-

tion of methods: Conclusions and recommendations. Arch

Hydrobiol Beih Ergebn Limnol 14:91–106

Montalto L, Marchese M (2003) Limnoperna fortunei (Dunker,

1857) (Bivalvia:Mytilidae) tolerance to temperature and

pH in experimental conditions. Neotropica 49:26–34

Morton B (1975) The colonization of Hong Kong’s water

supply system by Limnoperna fortunei (Dunker 1857)

(Bivalvia: Mytilacea) from China. Mal Rev 8:91–105

Morton B (1977) The population dynamics of Limnopernafortunei (Dunker 1857) (Bivalvia: Mytilacea) in Plover

Cove reservoir, Hong Kong. Malacologia 16:165–182

Morton B (1982) The reproductive cycle in Limnoperna for-tunei (Dunker 1857) (Bivalvia: Mytilidae) fouling Hong

Kong’s raw water supply system. Oceanol Limnol Sinica

13:312–325

Morton B (1987) Comparative life history tactics and sexual

strategies of the fresh and brackish water bivalve fauna of

Hong Kong and Southern China. Am Malacol Bull

5:91–99

Morton B (1996) The aquatic nuisance species problem: a

global perspective and review. In: D’Itri FM (ed) Zebra

mussels and other aquatic nuisance species. Ann Arbor

Press, Chelsea, pp 1–54

Nagaya K, Matsui Y, Ohira H et al (2001) Attachment strength

of an adhesive nuisance mussel, Limnoperna fortunei,against water flow. Biofouling 17:263–274

Oliveira MD (2009) Factors controlling the density and

potential distribution of golden mussel (Limnoperna for-tunei Dunker 1857) in the Upper Paraguay Basin. Ph.D.

thesis, Universidade Federal de Minas Gerais, Belo Hor-

izonte, Brazil

Oliveira MD, Calheiros DF (2000) Flood pulse influence on a

phytoplankton community in south Pantanal floodplain,

Brazil. Hydrobiologia 427:101–112

Oliveira MD, Hamilton SK, Jacobi CM (2010a) Forecasting

the expansion of the invasive golden mussel Limnopernafortunei in Brazilian and North American rivers based on

its occurrence in the Paraguay River and Pantanal wetland

of Brazil. Aquat Invasions 5(1):59–73

Oliveira MD, Hamilton SK, Calheiros DF, Jacobi CM (2010b)

Oxygen depletion events control the invasive golden mus-

sel (Limnoperna fortunei) in a tropical floodplain. Wetlands

30(4):705–716. doi:10.1007/s13157-010-0081-3

Oliveira MD, Takeda AM, Barros LF et al (2006) Invasion by

Limnoperna fortunei (Dunker, 1857) (Bivalvia Mytilidae)

of the Pantanal wetland, Brazil. Biol Invasions 8:97–104

Pastorino G, Darrigran G, Martin SM et al (1993) Limnopernafortunei (Dunker, 1857) (Mytilidae), nuevo bivalvo in-

vasor en aguas del Rio de La Plata. Neotropica 39:34

PCBAP (Plano de Conservacao da Bacia do Alto Paraguai)

(1997) Analise Integrada e prognostico da Bacia do alto

Paraguai.Programa Nacional do Meio Ambiente. Brasılia:

PNMA. V.3. 367 p

Peterson AT, Vieglais DA (2001) Predicting species invasions

using ecological niche modeling: new approaches from

bioinformatics attack a pressing problem. Bioscience

51:363–371

Ricciardi A (1998) Global range expansion of the Asian mussel

Limnoperna fortunei (Mytilidae): another fouling threat to

freshwater systems. Biofouling 13:97–106

Schneider DW, Madon SP, Stoeckel JA et al (1998) Seston

quality controls zebra mussel (Dreissena polymorpha)

energetics in turbid rivers. Oecologia 117:331–341

Silva DP (2006) Aspectos bioecologicos do mexilhao dourado

Limnoperna fortunei (Bivalvia, Mytilidae) (Dunker,

1857). Ph.D. thesis. Universidade Federal do Parana.

Curitiba, 138 p

Wang HZ, Xie Z, Wu XP et al (1999) A preliminary study of

zoobenthos in the Poyang Lake, the largest freshwater

lake of China, and its adjoining reaches of Changjiang

River. Acta Hydrobiol Sinica 23:132–138

Wang HZ, Xu QQ, Cui YD et al (2007) Macrozoobenthic

community of Poyang Lake, the largest freshwater lake of

China, in the Yangtze floodplain. Limnology 8:65–71

Wetzel RA, Likens GE (2001) Limnological analyses.

Springer, New York 391 p

Wilkinson L (2004) Systat 11. SYSTAT Software Inc, San Jose

Xie Z, Liang YL, Wang J et al (1999) Preliminary studies of

macroinvertebrates of the Mainstream of the Changjiang

(Yangtze) River. Acta Hydrobiol Sinica 23:148–157

Zhang SR, Lu XX, Higgitt DL et al (2007) Water chemistry of

the Zhujiang (Pearl River): natural processes and anthro-

pogenic influences. J Geophys Res Earth Surf 112:1–17

Abiotic factors controlling the establishment 729

123