shredders: species richness, abundance, and role in litter breakdown in tropical hong kong streams

J. N. Am. Benthol. Soc., 2009, 28(1):167–180� 2009 by The North American Benthological SocietyDOI: 10.1899/08-043.1Published online: 9 December 2008

Shredders: species richness, abundance, and role in litter breakdownin tropical Hong Kong streams

Aggie O. Y. Li1AND David Dudgeon2

Division of Ecology and Biodiversity, School of Biological Sciences, The University of Hong Kong,Pokfulam Road, Hong Kong SAR, China

Abstract. Shredders appear to be scarce in many tropical streams, but few data are available about theirabundance or role in litter breakdown in these systems. Shredder species richness, abundance, and biomasswere investigated in 10 small streams (5 shaded, 5 unshaded) in Hong Kong as a step toward understandingthe role of shredders in tropical Asian streams. In addition, breakdown of Liquidambar formosana(Hamamelidaceae) leaves was investigated to determine if breakdown rates changed in response toshredder species richness, abundance, or biomass in these 10 streams. Shredders were scarce in terms ofspecies richness (a total of 6 obligate shredders and 2 facultative shredders) and abundance (mean¼ 2.0%,range¼ 0–7.7% of total macroinvertebrate abundance). Obligate shredder biomass was not high across the10 streams (mean¼ 13% of total biomass), and it was variable in shaded streams (0.6–38.7%); all high valueswere attributable to the presence of a few large individuals (,0.2% of total abundance). Relative abundanceand biomass of obligate shredders were higher in shaded than in unshaded streams (abundance: 1.9% vs0.1%; biomass: 22.3% vs 4.3%). Shading did not affect the densities or relative abundances of functionalfeeding groups (FFGs) other than shredders. Litter breakdown rates did not vary in response to speciesrichness, abundance, or biomass of shredders among the 10 streams. However, when the 2 moderatelynutrient-enriched streams were excluded, breakdown rates were positively related to obligate shredderdensities, indicating a possible contribution of shredders to litter breakdown. The potential importance ofmicrobes was indicated by a relationship between breakdown rates and stream PO4 concentrations.

Key words: functional feeding groups, leaf litter, decomposition, macroinvertebrates, microbes, phos-phate.

Allochthonous organic matter, particularly leaf litter,can be the major source of energy in forest streams(e.g., Fisher and Likens 1973, Wallace et al. 1997), andchanges in litter input affect macroinvertebrate growthrates (Johnson et al. 2003) and production (Wallace etal. 1997, 1999). Biological breakdown of leaf litteroccurs mainly through the activities of detritivorousshredders and microbial action (Anderson and Sedell1979, Webster and Benfield 1986). Dead leaves areusually a low-quality food for shredders (high C:N;Friberg and Jacobsen 1999), but litter can serve as asubstrate or a trap for fine particulate organic matter(FPOM) that is eaten by nonshredding macroinverte-brates (e.g., Dudgeon and Wu 1999, Mathuriau andChauvet 2002). Litter breakdown rates are positivelyrelated to shredder abundance or biomass in temper-ate streams (e.g., Sponseller and Benfield 2001, Hagen

et al. 2006), and a positive relationship betweenshredder species richness and breakdown rates hasbeen detected in field studies (Jonsson et al. 2001,Lecerf et al. 2006) and some laboratory experiments(e.g., Jonsson and Malmqvist 2000).

The River Continuum Concept (RCC; Vannote et al.1980) includes a description of expected longitudinalchanges in the relative abundances of benthic macro-invertebrate functional feeding groups (FFGs). Inparticular, the RCC predicts that shredders andcollector–gatherers codominate benthic communitiesin shaded headwater streams that receive largeamounts of litter from riparian forest and typicallyhave a heterotrophic community metabolism. Incontrast to RCC predictions, a paucity of shreddersseems typical of streams outside the north-temperatezone, e.g., in Australia and New Guinea (Bunn 1986,Yule 1996), New Zealand (Winterbourn et al. 1981),tropical Asia (Dudgeon 1989, 2000), East Africa(Tumwesigye et al. 2000, Dobson et al. 2002), and the

1 E-mail addresses: [email protected] [email protected]

W:/xml/jnabs/jnbs_2801/jnbs-28-01-15.3d � Monday, 24 November 2008 � 9:35 pm � Allen Press, Inc. � 67

167

Neotropics (Rosemond et al. 1998, Greathouse andPringle 2006). However, there are reports of abundantshredders in leaf packs and benthic samples fromtropical Australian streams (Pearson et al. 1989,Cheshire et al. 2005).

At present, data about shredder representation inthe tropics (especially in tropical Asia) are scarcecompared to data from the temperate region. Morestudies of the richness and standing stocks ofshredders are needed to improve the robustness ofany generalizations about shredder scarcity or abun-dance in tropical streams. Our study, undertaken inHong Kong within tropical East Asia, was designed todocument interstream variation in the richness andabundance of shredders and to investigate the role ofshredders in litter breakdown. Our study had 2 maincomponents. First, the species richness, abundance,and biomass of shredders and other benthic macroin-vertebrate FFGs were quantified within pools andriffles in each of 10 small streams. Five shaded and 5unshaded streams were sampled to determine whethercommunity functional organization and, in particular,the density and biomass of shredders were affected byriparian shading conditions. We expected that shred-der species richness, abundance, and biomass wouldbe higher in shaded than in unshaded streams. Second,the effects of shredder species richness, abundance,and biomass on litter breakdown rates in pools andriffles among each of the 10 streams were investigatedwith standardized litter bags containing Liquidambarformosana Hance (Hamamelidaceae) leaves. We hopedthat the combination of quantitative information

(species richness, standing stocks) with data on therole of shredders in litter processing across a range ofstreams would provide unequivocal evidence of theabundance and importance of shredders in tropicalEast Asian streams.

Methods

Study sites

The study was carried out in Hong Kong (lat228090N–228370N), which has a tropical climate with2 distinct seasons (Dudgeon and Corlett 2004): the wetseason (April–September, when .80% of the annualrainfall occurs) and the dry season (October–March).The study included 9 streams in the mainland NewTerritories and a single site on Lantau Island (Fig. 1).Five sites were shaded (.40% cover, S1–S5), and 5were unshaded (,40% cover, U1–U5). All sites wereinside or immediately adjacent to protected areas/nature reserves. Two of the study streams (S3 and U3)were moderately enriched by nutrients from small-scale agriculture (market gardens) in their catchments,but the others were relatively undisturbed with soft,nutrient-poor water and low conductivity (Table 1; Li2008). Other information on the study streams andtheir benthic macroinvertebrate communities is givenby Li and Dudgeon (2008).

Sample collection and laboratory procedures

Field samples were collected in the dry season(December–February), when shredders are more

FIG. 1. Ten study sites in Hong Kong Special Administrative Region, southern China. (see Table 1 for site codes).

W:/xml/jnabs/jnbs_2801/jnbs-28-01-15.3d � Monday, 24 November 2008 � 9:35 pm � Allen Press, Inc. � Page 168

168 [Volume 28A. O. Y. LI AND D. DUDGEON

TA

BL

E1.

Ch

arac

teri

stic

so

fth

e10

Ho

ng

Ko

ng

stu

dy

stre

ams.

Dat

aar

efr

om

Li

(200

8)an

dL

ian

dD

ud

geo

n(2

008)

.S¼

shad

ed,

U¼

un

shad

ed,

Ag¼

agri

cult

ura

lla

nd

,A

a¼

aban

do

ned

agri

cult

ura

lla

nd

,P

l¼

pla

nta

tio

n,

Sh¼

shru

bla

nd

,D

M¼

dry

mas

s,28¼

seco

nd

ary.

Mu

iT

zeL

amS

hin

gM

un

Tsi

ng

Lu

ng

Tau

Tai

Po

Kau

Wu

Kau

Tan

gC

hu

enL

un

gH

auT

on

gK

aiP

akN

gau

Sh

ekT

aiL

amC

hu

ng

Wo

ng

Lu

ng

Han

g

Sit

eco

de

S1

S2

S3

S4

S5

U1

U2

U3

U4

U5

%ri

par

ian

cov

erag

e65

(S)

60(S

)55

(S)

75(S

)85

(S)

20(U

)20

(U)

25(U

)25

(U)

15(U

)L

atit

ud

ean

dlo

ng

itu

de

2282

40 N

1148

140 E

2282

40 N

1148

090 E

2282

30 N

1148

030 E

2282

50 N

1148

110 E

2283

00 N

1148

150 E

2282

40 N

11480

70 E

2282

60 N

1148

190 E

2282

60 N

1148

070 E

2282

40 N

1148

030 E

2281

60 N

1138

570 E

Alt

itu

de

of

stu

dy

reac

h(m

)10

023

011

018

090

370

110

9070

90S

trea

mo

rder

33

33

33

43

44

Wet

wid

th(m

)3.

92.

94.

24.

02.

94.

15.

82.

74.

13.

1D

epth

(m)

0.26

0.17

0.19

0.20

0.17

0.15

0.23

0.14

0.16

0.18

Su

bst

rate

com

po

siti

on

%B

ou

lder

(.25

6m

m)

4530

2040

530

655

5025

Co

bb

le(6

4–25

6m

m)

1535

3525

4025

2525

2525

Gra

vel

(2–6

4m

m)

1510

2520

2510

520

1015

San

d(.

2m

m)

05

2015

100

025

50

Bed

rock

2520

00

2035

535

,10

35C

urr

ent

vel

oci

tyin

riff

les

(m/

s)0.

40.

20.

50.

40.

30.

40.

30.

50.

40.

2T

emp

erat

ure

(8C

)17

.011

.017

.016

.011

.016

.512

.514

.516

.516

.5C

on

du

ctiv

ity

(lS

/cm

)24

.040

.437

.421

.610

.220

.216

.734

.841

.518

.5D

isso

lved

O2

(mg

/L

)9.

310

.48.

99.

410

.19.

010

.59.

99.

28.

8p

H6.

66.

77.

06.

76.

76.

86.

76.

76.

85.

7N

H3-N

(lg

/L

),

0.1

,0.

10.

30.

1,

0.1

0.1

,0.

10.

60.

10.

2N

O3-N

(lg

/L

)56

.871

.046

6.6

104.

18.

913

7.1

48.5

1101

.126

2.5

348.

3N

O2-N

(lg

/L

)1.

11.

03.

61.

01.

10.

81.

12.

03.

00.

5P

O4-P

(lg

/L

)15

.115

.546

.415

.815

.510

.69.

144

.410

.78.

5C

PO

Mst

and

ing

sto

ck(g

DM

/m

2)

65.1

619

.062

.66

23.9

16.1

62.

348

.56

14.8

44.7

67.

96.

96

1.0

12.9

63.

730

.66

13.2

39.8

611

.230

.76

10.2

Lan

du

se28

fore

st,

Aa

28fo

rest

Sh

,P

l,A

g28

fore

stS

h,

Pl,

Aa

Sh

,A

gS

hP

l,A

gS

h,

Pl

28fo

rest

,S

h


2009] 169SHREDDERS IN TROPICAL STREAMS

abundant than in the wet season (Dudgeon 1999a) andstreams have higher retention capacities for coarseparticulate organic matter (CPOM). Quantitative sam-ples of benthic macroinvertebrates and associated litter(CPOM) were collected from randomly selected 30 cm3 30 cm quadrats from pools and riffles in each of thestudy streams using a 300-lm-mesh D-framed net. Allsubstrates within each quadrat were washed inside theD-framed net to remove macroinvertebrates, and allmaterial washed into the net was retained. Thiscollection method was chosen to be directly compara-ble with that of Cheshire et al. (2005). Twelve quadratsamples (6 from pools, 6 from riffles) for abundancedetermination were collected in December 2005–January 2006 from each stream (total ¼ 120 samples)and were immediately preserved in 5% formaldehydesolution. Six quadrat samples (3 from pools, 3 fromriffles) for biomass measurements were collected fromeach stream in January–February 2007, transported tothe laboratory in a cool box, and stored at�208C beforefurther processing. Samples from the same mesohabi-tat in each stream were pooled (total ¼ 20 pooledsamples, 1 from pools and 1 from riffles in each of the10 streams). CPOM standing stocks (g dry mass [DM]/m2) in each sample were determined by oven-drying(608C, 48 h) and weighing CPOM from each sample.

Macroinvertebrates were handpicked from all ben-thic samples under a Leica MZ8 stereomicroscope(Leica Microsystems, Wetzlar, Germany), identified tospecies or morphospecies (or chironomid subfamily),and counted. If samples were exceptionally large, afixed-area/fixed-count subsampling procedure (Kingand Richardson 2002) was used after large or rare taxahad been removed and counted. Morphospeciescounts were used to calculate the abundance ofshredders and other FFGs from pools and riffles ineach stream. FFG assignment was determined by gut-content analyses of abundant taxa (Li and Dudgeon2008) or based on literature for uncommon taxa(Dudgeon 1989, 1999b). FFGs were obligate shredders,facultative shredders, collector–gatherers, scrapers,filter-feeders, and predators. For biomass measure-ments, macroinvertebrates in each FFG from the samemesohabitat in each stream were pooled, oven-dried(608C, 48 h), and weighed. Snail shells and gutcontents of large macroinvertebrates (.5 mm bodylength) were removed prior to weighing.

In situ litter breakdown

The role of shredders in litter breakdown wasstudied in the dry season (December 2005–February2006). Recently abscised leaves of L. formosana werecollected from Tai Po Kau Nature Reserve and air-

dried to constant mass. Eight litter bags (4 mm mesh)containing a known mass (;5 g) of air-dried leaveswere placed in each of the 10 study streams (4 in pools,4 in riffles). Litter bags were retrieved after 28 d, andthe contents of each bag were rinsed, oven-dried at608C for 48 h, and weighed. Four bags of leaves fromeach stream were combusted at 5008C for 4 h toprovide a conversion factor from oven dry mass (DM)to ash-free dry mass (AFDM). In addition, 10 sets ofair-dried leaves (;5 g each) were oven-dried (608C, 48h), weighed, combusted (5008C, 4 h), and reweighed toprovide a conversion factor from initial air-dry mass toAFDM. Macroinvertebrates from litter bags were notexamined because data on the densities and biomass ofFFGs in the benthic samples were already availableand used for subsequent analyses.

Data analyses

Three-way nested analysis of variance (ANOVA)was used to compare densities (individuals/0.09 m2)of shredders and other macroinvertebrate FFGs.Shading conditions (shaded vs unshaded) and meso-habitats (pools vs riffles) were fixed factors, and stream(streams 1–10) was a random factor nested withinshading (total n ¼ 120). Biomass values (mg DM/0.54m2) of shredders and total macroinvertebrates werecompared between shaded and unshaded streamsusing Student’s t-tests with streams as replicates (totaln¼ 10). Bivariate linear regression analyses were usedto test for relationships between CPOM standingstocks (g DM/m2) and species richness, abundance(ind./m2), or biomass (mg DM/m2) of shredders (orother FFGs) among streams, with separate regressionsundertaken for pool and riffle mesohabitats (n ¼ 10each).

Nonmetric multidimensional scaling (NMDS) ordi-nation was used to reveal patterns in macroinverte-brate assemblages and relative abundances of FFGsbetween shading conditions and mesohabitats, and 2-way nested analysis of similarity (ANOSIM) and theBray–Curtis similarity index were used to detectsignificant differences in assemblage structure causedby these factors. Streams and mesohabitats werecombined as a single factor (streams þmesohabitats),which was nested within shading conditions inANOSIM to detect any shading or streams þ meso-habitats effects (total n ¼ 120). SIMPER (PRIMER,version 6; PRIMER-E Ltd., Plymouth, UK) was used toreveal the contributions by each taxon or FFG to anyobserved differences. Square-root(x) and arcsine(x)transformations were applied to nonpercentage dataand percentage data (respectively) for NMDS, ANO-SIM, and SIMPER to prevent overrepresentation of



dominant taxa/FFGs. All ordinations and relatedmultivariate statistics were undertaken using PRIMER 6.

Three-way nested ANOVA was used to compare the% leaf mass loss of L. formosana leaves among shadingconditions and mesohabitats using the same model.Bivariate linear regression analyses were used to detectany significant relationships between mean % leafmass loss and abundance, biomass, or species richnessof shredders. Separate tests were undertaken for poolsand riffles with streams as replicates (n¼ 10 each test).Stepwise multiple regression analyses (forward–back-ward: a-to-enter � 0.05, a-to-remove � 0.10) wereused to detect significant relationships between mean% leaf mass loss and potentially relevant environmen-tal variables (i.e., CPOM standing stock, currentvelocity, dissolved inorganic N concentration, PO4

concentration, water temperature, shredder abun-dance, shredder biomass, and shredder richness), andseparate analyses were undertaken for pools and riffles(n ¼ 10 each). Shredder abundance and biomass persquare meter and per gram natural leaf (estimated byassuming all shredders were found on leaf litter) wereincluded in the regression analyses. These regressionanalyses were repeated after exclusion of the 2nutrient-enriched streams (i.e., S3 and U3) (n¼ 8 each).

All parametric calculations (ANOVA, t-tests, regres-sions) were undertaken with SPSS 15.0 (SPSS Inc.,Chicago, Illinois) or Microsoft Excel 2003 (MicrosoftCorporation, Redmond, Washington) at a significancelevel of a ¼ 0.05. Statistical power was calculated forANOVA, and no correction was applied for multipletesting. Log10(x) transformation or arcsine(x) transfor-mation (for percentage data) was applied prior to anyparametric statistical tests wherever transformationsnormalized the data sets or reduced the difference invariance among data sets.

Results

Benthic macroinvertebrate assemblages: FFGs

In total, 115 morphospecies in 60 families werefound in the 10 study streams (Appendix 1; availableonline from http://dx.doi.org/10.1899/08–043.1.s1).The dominant FFG in numerical terms was collector–gatherer, which accounted for .45% of total macroin-vertebrate abundance in all shaded and unshadedstreams (Table 2). In terms of biomass, obligate andfacultative shredders (51% of total biomass) weredominant in shaded streams, and collector–gather-ers/scrapers (59%) were dominant in unshadedstreams (Table 3), partly because of the high biomassof large-bodied Brotia hainanensis snails (Pachychili-dae), which were present in low densities (0.9% of totalabundance). These snails were assigned to the facul-

tative shredder FFG in shaded streams and wereassigned to the collector–gatherer/scraper FFG inunshaded streams because they had different dietsunder different shading conditions (Li and Dudgeon2008).

Shredder species richness and abundance were low.Only 6 obligate shredder taxa and 2 facultativeshredder taxa (mean richness 6 95% CI ¼ 4.4 6 1.0)were found in the 10 study streams, and these taxaaccounted for 0–7.7% (mean ¼ 2.0 6 1.7) of totalmacroinvertebrate abundance. Obligate shredders ac-counted for only 1.9% of total abundance in shadedstreams and 0.1% in unshaded streams (Table 2).Anisocentropus maculatus and Ganonema extensum (Cal-amoceratidae) were the most abundant obligateshedder species, but they constituted only 1.2 6 1.1%and 0.7 6 0.9% (mean 6 95% CI) of total abundance inshaded streams, and ,0.2% in unshaded streams.Eulichas dudgeoni (Eulichadidae), Goerodes doligung(Lepidostomatidae), Rhopalopsole sp. (Leuctridae), andTipula sp. (Tipulidae) were uncommon and accountedfor ,0.3% (mean¼,0.1%) of total abundance in the 10study streams. Facultative shredders (Amphinemurasp., and B. hainanensis in shaded streams) accountedfor 1.8% of total abundance in shaded streams and0.2% in unshaded streams (Table 2). At most, 5obligate shredders and 2 facultative shredders werepresent in any study stream (maximum in S4).

In terms of biomass, obligate shredders constituted13% of total macroinvertebrate biomass in the 10 studystreams: 22% in shaded streams and 4% in unshadedstreams (Table 3). This biomass mainly reflected thepresence of a few large larvae of the beetle E. dudgeoni(which often are .100 mg DM and can reach ;280mg). In S2, for example, a single E. dudgeoni larva,which accounted for 0.05% of macroinvertebrateabundance and 2.6% of obligate shredder abundance,contributed 12% of macroinvertebrate biomass and91% of shredder biomass. The same trend of highbiomass contributed by a few E. dudgeoni was seen inS4, S5, and U4, where the beetle larvae constituted,0.2% of total abundance but .15% of total biomass.In shaded streams, small numbers of large B. haina-nensis (1.2% of total macroinvertebrate abundance)constituted, on average, 29% of total macroinvertebratebiomass and .99% of facultative shredder biomass.

Effect of shading on abundance of FFGs

Total macroinvertebrate densities did not differsignificantly between shaded and unshaded streams(3-way nested ANOVA, F1,8 ¼ 0.39, p ¼ 0.549), butmacroinvertebrate densities were significantly higherin riffles than in pools (F1,8¼ 27.97, p¼ 0.001; Table 2).



Total macroinvertebrate biomass did not differ signif-

icantly between shaded and unshaded streams (t8 ¼0.98, p ¼ 0.178; Table 3). Facultative shredders were

significantly more abundant in shaded than in

unshaded streams (density: F1,8 ¼ 9.98, p ¼ 0.013;

relative abundance: F1,8 ¼ 8.83, p ¼ 0.018), and their

density (but not relative abundance) was higher in

riffles than in pools (density: F1,8 ¼ 8.52, p ¼ 0.019;

relative abundance: F1,8 ¼ 1.54, p ¼ 0.250; Table 2).

Obligate shredders were significantly more abundant

in shaded than in unshaded streams (relative abun-

dance: F1,8¼ 5.67, p¼ 0.044) and in pools compared to

riffles (relative abundance: F1,8¼ 7.97, p¼ 0.022; Table

2), but their density did not differ significantly

between shaded and unshaded streams (density: F1,8

¼ 4.93, p ¼ 0.057). However, the density of obligate

shredders was higher in pools than in riffles in shaded

streams but not in unshaded streams (Table 2), as

shown by a significant interaction between shading

and mesohabitat (F1,8 ¼ 7.25, p ¼ 0.027). Species

richness of obligate shredders did not differ signifi-

cantly between shaded and unshaded streams (t8 ¼0.81, p ¼ 0.222), but absolute biomass (t8 ¼ 2.36, p ¼0.023) and relative biomass (t8 ¼ 2.31, p ¼ 0.025) of

obligate shredders were higher in shaded than in

unshaded streams (Table 3). CPOM standing stocks

also were significantly higher in shaded than in

unshaded streams (t8 ¼ 2.44, p ¼ 0.020; Tables 2, 3).

Among streams, there were positive relationships

between CPOM standing stocks (g DM/m2) and

obligate shredder richness (OShrich) and density

(OShden; ind./m2) in pools (richness: R2 ¼ 0.51, F1,8

¼ 8.23, p ¼ 0.021, [OShrich] ¼ 0.63 þ 0.04[CPOM];

density: R2 ¼ 0.45, F1,8 ¼ 6.42, p ¼ 0.035, log[OShden]

TABLE 2. Mean (695% confidence interval) density (ind./m2) and relative abundance (%) of functional feeding groups (FFGs) inpools and riffles in shaded and unshaded streams in Hong Kong. DM ¼ dry mass, – ¼ FFG not present.

FFG

Pools Riffles Pools þ riffles

DensityRelative

abundance DensityRelative

abundance DensityRelative

abundance

Shaded streamsObligate shredders 117.0 6 100.7 4.6 6 3.7 28.1 6 26.3 0.5 6 0.4 72.6 6 62.2 1.9 6 1.8Facultative shredders 51.1 6 45.8 2.2 6 2.1 132.6 6 119.7 1.6 6 0.9 91.9 6 74.9 1.8 6 1.1Collector–gatherers 1836.3 6 424.6 71.4 6 8.2 4647.8 6 4051.2 52.1 6 12.4 3242.0 6 2015.9 58.6 6 9.2Scrapers 251.1 6 78.1 10.3 6 4.1 1103.3 6 574.0 20.8 6 10.7 677.2 6 279.4 16.4 6 7.4Filter-feeders 15.9 6 3.2 0.7 6 0.1 1292.6 6 1360.2 12.9 6 7.2 654.3 6 679.1 9.8 6 6.2Predators 274.4 6 72.4 10.8 6 2.6 771.5 6 283.6 12.1 6 4.2 523.0 6 118.0 11.5 6 3.4Others 2.2 6 1.4 0.1 6 0.0 4.4 6 7.0 ,0.1 3.3 6 3.4 ,0.1Total densities (ind./m2) 2548.1 6 384.8 7980.4 6 5442.3 5264.3 6 2656.1CPOM standing stock

(g DM/m2) 35.4 6 12.8 59.3 6 26.4 47.4 6 17.2

Unshaded streamsObligate shredders 5.2 6 5.4 0.2 6 0.1 10.7 6 13.0 0.1 6 0.1 8.0 6 6.8 0.1 6 0.1Facultative shredders – – 14.4 6 17.8 0.2 6 0.3 7.2 6 8.9 0.2 6 0.2Collector–gatherers 1868.9 6 627.6 67.8 6 9.0 6814.1 6 3437.1 53.5 6 6.7 4341.5 6 1945.2 57.0 6 5.3Scrapers 419.6 6 309.6 13.4 6 4.6 2249.3 6 1224.0 20.2 6 10.6 1334.4 6 733.3 18.1 6 7.4Filter-feeders 21.5 6 18.8 0.6 6 0.4 2431.5 6 1612.7 16.2 6 7.0 1226.5 6 814.8 13.3 6 6.0Predators 501.9 6 311.4 18.0 6 6.4 990.4 6 305.5 9.7 6 4.9 746.1 6 246.0 11.3 6 4.8Others 0.4 6 0.7 ,0.1 0.7 6 1.5 ,0.1 0.6 6 0.7 ,0.1Total densities (ind./m2) 2817.4 6 1121.3 12,511.1 6 6048.6 7664.3 6 3439.8CPOM standing stock

(g DM/m2) 12.6 6 7.2 35.8 6 21.7 24.2 6 12.0

Overall (all streams)Obligate shredders 61.1 6 60.0 2.4 6 2.3 19.4 6 14.9 0.3 6 0.2 40.3 6 36.3 1.0 6 1.0Facultative shredders 25.6 6 27.3 1.1 6 1.2 73.5 6 68.9 0.9 6 0.6 49.5 6 45.0 1.0 6 0.7Collector–gatherers 1852.6 6 357.4 69.6 6 5.9 5730.9 6 2602.5 52.8 6 6.7 3791.8 6 1368.5 57.8 6 5.0Scrapers 335.4 6 160.3 11.8 6 3.1 1676.3 6 739.1 20.5 6 7.1 1005.8 6 427.7 17.3 6 4.9Filter-feeders 18.7 6 9.2 0.6 6 0.2 1862.0 6 1061.8 14.6 6 4.9 940.4 6 533.8 11.6 6 4.3Predators 388.1 6 168.0 14.4 6 4.1 880.9 6 209.1 10.9 6 3.1 634.5 6 147.8 11.4 6 2.8Others 1.3 6 0.9 ,0.1 2.6 6 3.6 ,0.1 1.9 6 1.9 ,0.1Total densities (ind./m2) 2682.8 6 565.7 10,245.7 6 4111.3 6464.3 6 2193.6CPOM standing stock

(g DM/m2) 24.0 6 10.2 47.6 6 17.8 35.8 6 12.5



¼�0.91þ1.59log[CPOM]) and in riffles (richness: R2¼0.49, F1,8 ¼ 7.80, p ¼ 0.023, [OShrich] ¼ 0.47 þ0.04[CPOM]; density: R2 ¼ 0.56, F1,8 ¼ 10.26, p ¼0.013, log[OShden] ¼ �0.84 þ 1.17log[CPOM]), asshown by bivariate linear regressions. Obligateshredder biomass (OShbio, mg DM/m2) also waspositively related to CPOM in riffles (R2 ¼ 0.43, F1,8 ¼6.06, p¼ 0.039, log[OShbio]¼�1.92þ 2.62log[CPOM])and among streams (R2 ¼ 0.51, F1,8 ¼ 8.40, p ¼ 0.020,log[OShbio] ¼ �2.01 þ 2.94log[CPOM]). A within-stream (pools þ riffles) relationship between obligateshredder density and CPOM standing stocks wasdetected in only one shaded stream (S1: R2 ¼ 0.35,F1,10 ¼ 5.44, p ¼ 0.042, log[OShden] ¼ 0.07 þ0.82log[CPOM]), but obligate shredder density waspositively related to CPOM standing stocks withinriffles in S1 (R2 ¼ 0.88, F1,4 ¼ 30.67, p ¼ 0.005,log[OShden]¼�0.13þ 0.81log[CPOM]), S2 (R2¼ 0.87,F1,4 ¼ 27.67, p ¼ 0.006, [OShden] ¼ �0.13 þ

0.27[CPOM]), and U4 (R2 ¼ 0.75, F1,4 ¼ 12.10, p ¼0.025, [OShden] ¼ �0.42 þ 0.18[CPOM]), and within

pools in S4 (R2 ¼ 0.75, F1,4 ¼ 11.93, p ¼ 0.026,

log[OShden] ¼ 0.20þ 2.19log[CPOM]). The only other

significant relationships between FFGs and CPOM

standing stocks were for scraper species richness

(Scrich: R2¼ 0.53, F1,8¼ 9.04, p¼ 0.017, [Scrich]¼ 6.79

þ 0.71[CPOM]) and filter-feeder biomass (Fbio: R2 ¼0.68, F1,8 ¼ 17.37, p ¼ 0.003, log[Fbio] ¼ �0.62 þ2.04log[CPOM]) in riffles.

Shading did not affect the densities or relative

abundances of any FFGs other than shredders (F1,8 �2.90, p � 0.127). Density of scrapers seemed higher in

unshaded streams (1334.4 6 733 ind./m2 vs 677.2 6

279 ind./m2), but this difference was not significant

(F1,8 ¼ 1.06, p ¼ 0.334). Nonetheless, densities and

relative abundances of scrapers, collector–gatherers,

and filter-feeders were higher in riffles than in pools, as

TABLE 3. Mean (6 95% CI) biomass (mg dry mass [DM]/m2) and relative biomass (%) of functional feeding groups (FFGs) inpools and riffles in shaded and unshaded streams in Hong Kong. – ¼ FFG not present.

FFG

Pools Riffles Pools þ riffles

BiomassRelativebiomass Biomass

Relativebiomass Biomass

Relativebiomass

Shaded streamsObligate shredders 887.9 6 816.2 29.1 6 13.3 670.1 6 1072.2 12.8 6 15.4 778.2 6 673.3 22.3 6 14.4Facultative shredders 1186.6 6 929.9 43.9 6 22.8 1017.1 6 1057.5 23.2 6 23.1 1101.8 6 918.3 28.7 6 20.5Collector–gatherers/scrapers 262.5 6 197.7 19.8 6 20.2 389.3 6 110.2 19.6 6 19.0 325.9 6 72.3 19.0 6 19.6Filter-feeders 12.9 6 23.5 0.3 6 0.5 276.4 6 267.3 9.7 6 10.1 144.6 6 145.1 4.6 6 3.6Predators 75.6 6 31.9 6.9 6 6.1 1214.4 6 789.6 34.4 6 15.8 645 6 379.7 25.2 6 18.7Others 0.2 6 0.4 ,0.1 13.6 6 14.0 0.4 6 0.4 6.9 6 7.1 0.2 6 0.2Total biomass (mg DM/m2) 2425.8 6 1647.6 3580.8 6 1994.4 3003.3 6 1422.1CPOM standing stock

(g DM/m2) 32.8 6 28.9 31.3 6 7.3 32.0 6 15.5

Unshaded streamsObligate shredders 14.2 6 11.6 1.9 6 2.8 130.7 6 234.8 6.4 6 10.6 72.5 6 120.9 4.3 6 6.0Facultative shredders – – ,0.1 ,0.1 ,0.1 ,0.1Collector–gatherers/scrapers 1801 6 1501.5 73 6 20.0 1087.7 6 828.7 52.2 6 27.4 1444.3 6 1146.9 58.9 6 25.3Filter-feeders 0.1 6 0.3 ,0.1 157.1 6 124.9 11.3 6 13.4 78.6 6 62.5 8.1 6 10.6Predators 264.4 6 233.1 20.7 6 14.5 434 6 231.3 29.4 6 22.8 349.2 6 117.8 28.0 6 20.9Others 7.8 6 13.8 4.4 6 8.5 11.0 6 12.8 0.6 6 0.7 9.4 6 7.8 0.8 6 1.1Total biomass (mg DM/m2) 2087.5 6 1702.7 1820.6 6 779.4 1954.1 6 1179.1CPOM standing stock

(g DM/m2) 12.2 6 5.6 13.2 6 6.0 12.7 6 5.5

Overall (all streams)Obligate shredders 451.1 6 479.1 15.5 6 11.0 400.4 6 546.6 9.6 6 9.1 425.8 6 396.6 13.3 6 9.4Facultative shredders 593.3 6 585.2 22.0 6 17.9 508.5 6 599.1 11.6 6 13.3 550.9 6 563.0 14.3 6 13.5Collector–gatherers/scrapers 1031.8 6 873.1 46.4 6 21.9 738.5 6 455.4 35.9 6 19.0 885.1 6 653.4 38.9 6 19.9Filter-feeders 6.5 6 11.8 0.2 6 0.3 216.8 6 144.4 10.5 6 7.9 111.6 6 77.5 6.3 6 5.4Predators 170.0 6 126.9 13.8 6 8.7 824.2 6 464.1 31.9 6 13.2 497.1 6 210.9 26.6 6 13.2Others 4.0 6 7.0 2.2 6 4.3 12.3 6 9.0 0.5 6 0.4 8.2 6 5.0 0.5 6 0.6Total biomass (mg DM/m2) 2256.7 6 1122.4 2700.7 6 1161.7 2478.7 6 935.9CPOM standing stock

(g DM/m2) 22.5 6 15.4 22.3 6 7.4 22.4 6 10.0



was the density (but not relative abundance) ofpredators (F1,8 � 8.30, p � 0.020; Table 2).

Multivariate analyses of macroinvertebrate assemblages

The biplot for the first 2 axes of NMDS could notdistinguish between macroinvertebrate assemblages inshaded and unshaded streams (Fig. 2A), but a clearseparation between pools and riffles was evident (Fig.2B). Two-way nested ANOSIM confirmed the absenceof a shading effect (global R¼�0 .013, p¼ 0.464), but asignificant difference was detected for the combined

factor (streams þmesohabitats) (global R ¼ 0.791, p ,

0.001). Because no significant shading effect waspresent, a 2-way ANOSIM was undertaken usingmesohabitats and streams as factors (i.e., with shadingconditions excluded). This analysis yielded a signifi-cant difference between mesohabitats (global R ¼0.702, p , 0.001) and among streams (global R ¼0.725, p , 0.001). SIMPER was used to determinetaxon contributions to differences in assemblagestructure between pools and riffles (Table 4). With40% as a cutoff for cumulative contributions to thedissimilarity, the top 11 contributors to the dissimilar-ity among macroinvertebrate assemblages were main-ly filter-feeders and collector–gatherers/scrapers, andtheir densities were higher in riffles than in pools.

A clear separation between pools and riffles alsowas evident from an NMDS biplot of the relativeabundances of FFGs (Fig. 3B), but not between shadedand unshaded streams (Fig. 3A). Two-way nestedANOSIM revealed a significant difference amongstreams þ mesohabitats (global R ¼ 0.575, p , 0.001),but none between shading conditions (global R ¼�0.012, p¼ 0.497). A subsequent 2-way ANOSIM withshading conditions excluded (see previous) revealed adifference in FFG relative abundances between poolsand riffles (global R ¼ 0.672, p , 0.001) and amongstreams (global R¼ 0.401, p , 0.001). SIMPER showedthat filter-feeders, which were relatively abundant inriffles (15% vs 0.6%; Table 2), were the major

FIG. 3. Nonmetric multidimensional scaling (NMDS)ordination of the relative abundances of macroinvertebrateFFGs in 5 shaded and 5 unshaded streams (A) and betweenmesohabitats (pools/riffles) (B).

FIG. 2. Nonmetric multidimensional scaling (NMDS)ordination of macroinvertebrate assemblages (i.e., counts ofmorphospecies) in 5 shaded and 5 unshaded streams (A) andbetween mesohabitats (pools/riffles) (B).

TABLE 4. Top contributors to dissimilarity between mac-roinvertebrate assemblages in pools and riffles revealed bySIMPER. FFG ¼ functional feeding group, Co ¼ collector–gatherers, Sc ¼ scrapers, F ¼ filter-feeders, P ¼ predators.

Taxon FFG % contribution Cumulative %

Chironominae Co 7.28 7.28Orthocladiinae Co 4.87 12.15Baetidae Sc 4.35 16.5Hydrocyphon sp. Co 4.31 20.81Cheumatopsyche spp. F 3.54 24.36Eusimulium sp. F 3.46 27.82Cinygmina sp. Sc 3.32 31.14Chimarra sp. F 2.88 34.03Agapetus spp. Sc 2.79 36.82Tanypodinae P 2.55 39.37Isca purpurea Co 2.42 41.79



contributors to this mesohabitat difference and ac-counted for 31% of the dissimilarity. Collector–gather-ers and scrapers contributed to 25% and 21% of thedifference, whereas obligate and facultative shredderscontributed the least to the difference (8% and 4%).

Factors determining litter breakdown rates

The breakdown rates of L. formosana in terms of %leaf mass loss were significantly higher in riffles (mean% loss 6 SE¼ 39.0 6 3.1%) than in pools (25.2 6 2.2%;F1,8 ¼ 65.50, p , 0.001). No significant shading effectwas detected (shaded: 33.1 6 3.4%, unshaded: 31.1 6

4.2%; F1,8 ¼ 0.12, p ¼ 0.740), but % leaf mass lossdiffered among streams (F8,8 ¼ 10.39, p ¼ 0.002).Therefore, 2-way ANOVA was conducted usingstreams and mesohabitats as fixed factors. Thisanalysis confirmed the significantly higher breakdownrates in riffles compared to pools (F1,53 ¼ 86.38, p ,

0.001) and a significant difference among streams (F9,53

¼ 10.96, p , 0.001). Tukey tests indicated significantlyhigher breakdown rates in U3 (mean % loss ¼ 46.5%;responsible for 7 of 15 significant pairwise differences)and lower breakdown rates in S5 (21.2%; involved in 5pairwise differences) compared to other streams.

Bivariate linear regressions yielded no relationshipbetween litter breakdown rates (in terms of % leafmass loss) and abundance, biomass, or species richnessof obligate shredders for pool data only, for riffle dataonly, and for streams with pool and riffle datacombined (F1,8 � 2.99, p � 0.122). However, afterexcluding nutrient-enriched S3 and U3, litter break-down rates were positively related to obligate shred-der densities (OShden: both per m2 and per g naturalleaf) in pools (OShden/m2: R2 ¼ 0.66, F1,6 ¼ 11.69, p ¼0.014, [% leaf mass loss] ¼ 19.49 þ 0.05 [OShden];OShden/g leaf: R2 ¼ 0.56, F1,6 ¼ 7.60, p ¼ 0.033,arcsine[% leaf mass loss] ¼ 0.44 þ 0.14 log[OShden])and among streams (OShden/m2: R2¼ 0.52, F1,6¼ 6.61,p ¼ 0.042, [% leaf mass loss] ¼ 25.80 þ 0.07 [OShden];OShden/g leaf: R2 ¼ 0.53, F1,6 ¼ 6.86, p ¼ 0.040,arcsine[% leaf mass loss] ¼ 0.51 þ 0.23log[OShden]).

When all 10 study streams were included, stepwisemultiple regressions incorporating stream nutrientconcentrations indicated that litter breakdown waspositively related to PO4 (lg/L) only, as seen in the bestmodels for pools (adjusted R2 ¼ 0.36, F1,8 ¼ 6.02, p ¼0.040; arcsine[% leaf mass loss]¼0.255þ0.217log[PO4])and riffles (adjusted R2 ¼ 0.61, F1,8 ¼ 14.87, p ¼ 0.005;arcsine[% leaf mass loss]¼ 0.266þ 0.330log[PO4]) andamong streams (poolsþ riffles: adjusted R2¼0.54, F1,8¼11.72, p ¼ 0.009; arcsine[% leaf mass loss] ¼ 0.265 þ0.272log[PO4]). After excluding S3 and U3, litterbreakdown rates were related to obligate shredder

densities (per m2) and total shredder richness (TShrich)only in the best model for pools (adjusted R2¼0.91, F2,5

¼ 35.53, p ¼ 0.001; [% leaf mass loss] ¼ 24.47 þ0.06[OShden] – 2.78[TShrich]), and with obligateshredder densities (per g natural leaf) only amongstreams (adjusted R2 ¼ 0.46, F1,6 ¼ 6.86, p ¼ 0.040;arcsine[% leaf mass loss] ¼ 0.51 þ 0.23 log[OShden]).Leaf mass loss was ;17–44% faster in the stream withhighest contribution of obligate shredders (i.e., 178ind./m2 in S4) compared to those with low shredderabundance (,25 ind./m2 in, e.g., S5 and U1).

Discussion

Shredder species richness was low in streams inHong Kong compared to streams in temperatelatitudes. Only 8 shredders (obligate þ facultative)were found in the 10 study streams, whereas �10shredder species have been reported in north-temper-ate streams (e.g., Angradi 1996, Stone and Wallace1998, Murphy and Giller 2000). For comparativepurposes, data on shredder richness were compiledfrom a literature survey of studies that includedbetween 1 and 9 streams sampled using methods thatwere broadly comparable to those used in our study(Appendix 2; available online from http://dx.doi.org/10.1899/08–043.1.s2). Authors of north-temperatestudies reported an average of ;15 shredder speciesper study, whereas authors of studies in tropicalregions reported only ;6 species per study. Shreddersaccounted for ;20% of total species richness in benthicsamples in north-temperate streams (Appendix 2),compared to only ;12% in tropical regions. In HongKong, shredders constituted only ;8% (on average) ofspecies in shaded streams and as few as ;4% inunshaded streams (Appendix 2).

Major shredding taxa in north-temperate streamsinclude amphipods, isopods, filipalpian stoneflies,case-building caddisflies, and shredding craneflies(Anderson and Sedell 1979, Merritt and Cummins1996). Some of these taxa (amphipods and isopods) donot occur in tropical Asian streams (Dudgeon 1999b).Among the others, shredding Plecoptera seem to beparticularly diverse in temperate streams. For exam-ple, Angradi (1996) reported 8 genera (and 5 families)of shredding Plecoptera from a few streams in the US,and these families are also well represented inPalearctic Asia (Kobayashi and Kagaya 2002, Hoangand Bae 2006). However, diversity of Plecopteradeclines toward tropical latitudes (Dudgeon 1999b),and some families of shredding Plecoptera, such asPteronarcyidae and Peltoperlidae, are not present inHong Kong or over most of the rest of the Orientalregion. Likewise, the mainly Palearctic Limnephilidae,



Phryganeidae, and Sericostomatidae, which includemany shredding genera (Merritt and Cummins 1996),are absent (Dudgeon 1999b). Only 13% of theTrichoptera genera recorded during our study wereshredders, whereas the proportion in north-temperatestreams is much higher (34–37%; summarized inWinterbourn et al. 1981). The high representation ofshredders in north-temperate streams should beinterpreted with caution. Some taxa classified asshredders might be trophic generalists under certainconditions and consume significant amounts of non-CPOM food in the field and laboratory feeding trials(e.g., Friberg and Jacobsen 1999, Dangles 2002).Despite this caveat, our study shows clearly that therichness of shredders in Hong Kong (and elsewhere inthe Asian tropics; see Dudgeon 2000) is lower than inmost north-temperate streams.

As was the case for species richness, the density andrelative abundance of obligate shredders in the studystreams were low (,5% of total macroinvertebrateabundance) and comparable to values reported insome other tropical regions where shredders consti-tuted, on average, ;5% of total macroinvertebrateabundance (Appendix 3; available online from http://dx.doi.org/10.1899/08–043.1.s3). In contrast, shred-ders account for ;24% of total macroinvertebrateabundance in north-temperate streams and reach 70–80% of total abundance in some streams (e.g., Voughtet al. 1998, Lecerf et al. 2006; Appendix 3). Shreddersaccounted for only 2–9% of total macroinvertebrateabundance in some North Carolina (US) streams (e.g.,Stone and Wallace 1998, Wallace et al. 1999; Appendix3). This low proportion apparently reflects high totalmacroinvertebrate densities, as shown by the fact thatshredder densities in these US streams (754-2292 ind./m2) were much greater than the highest density in ourstudy (i.e., 178 ind./m2 in S4) and ;19–573 greaterthan the mean density of obligate shredders (40 ind./m2) across the 10 Hong Kong study streams (Appendix3).

Low retention of leaf litter or CPOM has been citedas a possible cause of the paucity of shredders in somestreams (e.g., in New Zealand; Winterbourn 1995), butlow retention was not the case in Kenya (Dobson et al.2002) nor in our study, where CPOM standing stocksin shaded streams were comparable to levels in sometemperate streams (Appendix 3). Leaf litter is oftenabundant in tropical streams, but much of this materialmight be unsuitable food for shredders. Tropical plantsare highly diverse and have leaves that are generallymore recalcitrant than leaves from temperate decidu-ous trees, and tropical leaves have high levels ofsecondary compounds that provide a defense againstintensive terrestrial herbivory in the tropics (Coley and

Aide 1991). Toughness and the presence of chemicalcompounds reduce litter breakdown rates in streams(e.g., Stout 1989, Quinn et al. 2000), and these defenses,which affect the food quality of CPOM, presumablyreduce shredder feeding rates and abundance (Irons etal. 1988, Stout 1989).

Some authors have suggested that decapod macro-consumers (crabs and shrimps) contribute to leaf litterbreakdown in tropical streams that lack insect shred-ders (e.g., Crowl et al. 2001, Dobson 2004). Decapodsmight evade benthic samplers so that their abundanceis underestimated (Dobson 2004), as confirmed byPringle and Ramırez (1998) in Costa Rica andDudgeon (2006) in Sulawesi. However, undersamplingof decapods would not have accounted for the scarcityof shredders during our study because the only 2decapods common in the study streams are notshredders. Macrobrachium hainanense (Palaemonidae),a widespread and common Hong Kong shrimp, is apredator (Mantel and Dudgeon 2004), whereas thelocal atyids (Caridina spp.) are small (,2.5 cm totallength) and assimilate FPOM and periphyton but littleleaf litter (Yam and Dudgeon 2005).

Obligate shredders were scarce in Hong Kong interms of both density and relative abundance, but thebiomass of obligate shredders could be substantial(22% of total macroinvertebrate biomass in shadedstreams, up to 39% in S4 and S5), and both biomassand relative biomass of obligate shredders in shadedHong Kong streams (except S3) were comparable tovalues reported from temperate streams (Appendix 3).Obligate shredders were scarce in S5 (0.2% abun-dance), which, like S3, had a streambed consisting ofsand and small stones, but the presence of large sand-dwelling larvae of E. dudgeoni meant that obligateshredders constituted a significant proportion of totalbiomass despite their low abundance. This resultunderscores the fact that obligate shredder biomassin Hong Kong streams was dominated by a few largeindividuals and lacked the range of taxa and sizespresent in temperate streams.

Shredder abundance and biomass were positivelyrelated to CPOM standing stocks among Hong Kongstreams (see also Friberg 1997) and were generallylower in unshaded streams that contained less CPOM(18 vs 41 g DM/m2 in shaded streams). Similardifferences in shredder abundance and biomassbetween shaded and unshaded streams (or sites) havebeen reported in Hong Kong (Dudgeon 1989) andelsewhere (e.g., Gray and Johnson 1988, Greathouseand Pringle 2006), although a few studies have failedto find such differences (Hawkins et al. 1982, Bojsenand Jacobsen 2003). However, the difference inobligate shredder density between shaded and un-



shaded streams was only marginally significant (at p¼0.057 in our study), perhaps because statistical power(¼ 0.494) was weakened by low sample size (5streams/shading type) and high interstream variation.Despite the positive relationship between shreddersand quantities of CPOM among streams, a within-stream relationship (i.e., obligate shredder abundancevs CPOM at the patch level) was detected in only oneof the shaded streams (S1), in contrast to some studiesin temperate streams (e.g., Murphy and Giller 2000). Apossible explanation is that much of the litter wasunpalatable and refractory to shredders (see previous),and, thus, CPOM did not strongly influence shredderdistribution within Hong Kong streams.

A notable result of our study is that shadingconditions did not affect the abundance of any FFGsother than shredders. Scrapers were expected toincrease in abundance and biomass as light and,hence, algal biomass (i.e., food availability) increasedwith decreases in shading (Vannote et al. 1980), as hasbeen reported elsewhere (Reed et al. 1994, Maridet etal. 1998, Greathouse and Pringle 2006). The reasons forthis lack of response are not known, but Dudgeon(1989) found that the difference between the relativeabundance of scrapers in shaded (11–16%) andunshaded (18–25%) Hong Kong streams was small.The generally unspecialized feeding mode of mostconsumers, which tended to consume at least 2 mainfood types (Li and Dudgeon 2008), might account forthis weak response to shading conditions.

Breakdown rates of L. formosana leaves were rapid,in accordance with a previous study in Hong Kong(Dudgeon 1982). Litter breakdown rates were notrelated to biomass or species richness of shredders, incontrast to previous findings, which found breakdownrates to be positively related to shredder biomass(Sponseller and Benfield 2001) or richness (Jonsson etal. 2001, Lecerf et al. 2006) in temperate streams. Whenthe 2 nutrient-enriched streams were excluded fromthe regressions, a positive relationship between litterbreakdown rates and obligate shredder densities inpools and among streams was apparent, as seen also insome temperate streams (Sponseller and Benfield 2001,Hagen et al. 2006). These data indicate that shredderscan make some contribution to litter breakdown inshaded Hong Kong streams despite their low abun-dance. Note, however, that only shredder abundance/biomass in the benthos was examined, and, despite thecorrection for CPOM standing stocks in each stream,abundance/biomass of benthic shredders might not bea good predictor of shredder abundance/biomass inlitter bags. Nonetheless, previous studies in HongKong have shown that shredders were similarly scarcein litter bags/leaf packs (e.g., Dudgeon 1982, Dudgeon

and Wu 1999, Li et al. 2008). The positive relationshipbetween stream PO4 concentrations and litter break-down rates points to a possible role of microbes inlitter breakdown, in accordance with other studiesreporting effects of nutrients on microbial activity andlitter breakdown rates (e.g., Elwood et al. 1981, Niyogiet al. 2003, Ardon et al. 2006). Our data suggest thatthe contribution made by shredders might be evidentonly when microbial activity is limited by nutrientavailability. This relationship might be a generalfeature of tropical streams because high prevailingtemperatures stimulate microbial activity (e.g., Padgett1976, Irons et al. 1994). Our study was conductedduring the cool dry season in Hong Kong, but watertemperatures averaged 15–168C and were consider-ably higher than values recorded during many studiesof litter breakdown in temperate streams. The relativeimportance of microbes requires detailed investigationbecause it has implications for understanding foodwebdynamics and the trophic base of production intropical forest streams where shredders are scarce.

Acknowledgements

The authors are grateful to Chen Young (CarnegieMuseum) for his advice on Tipula larvae, and LuzBoyero (James Cook University) for her advice onprocedures and data in Cheshire et al. (2005), as well asPamela Silver, Bruce Chessman, and 2 anonymousreferees for their constructive comments on themanuscript. Special thanks are due to Lily Ng andJames Hui for their technical support and Teresa Mafor her assistance in sample sorting. The workdescribed in this paper was partially supported by agrant from the Research Grants Council of Hong KongSpecial Administrative Region, China (Project No.[HKU] 7509/06M), and by a postgraduate studentshipawarded to AOYL during her M.Phil. studies at theUniversity of Hong Kong.

Literature Cited

ABELHO, M., AND M. A. S. GRACA. 1996. Effects of eucalyptusafforestation on leaf litter dynamics and macroinverte-brate community structure of streams in central Portu-gal. Hydrobiologia 324:195–204.

ANDERSON, N. H., AND J. R. SEDELL. 1979. Detritus processingby macroinvertebrates in stream ecosystems. AnnualReview of Entomology 24:351–377.

ANGRADI, T. R. 1996. Inter-habitat variation in benthiccommunity structure, function, and organic matterstorage in 3 Appalachian headwater streams. Journal ofthe North American Benthological Society 15:42–63.

ARDON, M., L. A. STALLCUP, AND C. M. PRINGLE. 2006. Does leafquality mediate the stimulation of leaf breakdown by



phosphorus in Neotropical streams? Freshwater Biology51:618–633.

BASAGUREN, A., A. ELOSEGUI, AND J. POZO. 1996. Changes in thetrophic structure of benthic macroinvertebrate commu-nities associated with food availability and stream flowvariations. Internationale Revue der Gesamten Hydro-biologie 81:79–91.

BOJSEN, B. H., AND D. JACOBSEN. 2003. Effects of deforestationon macroinvertebrate diversity and assemblage structurein Ecuadorian Amazon streams. Archiv fur Hydro-biologie 158:317–342.

BUNN, S. E. 1986. Spatial and temporal variation in themacroinvertebrate fauna of streams of the northernjarrah forest, Western Australia: functional organization.Freshwater Biology 16:621–632.

BUSS, D. F., D. F. BAPTISTA, M. P. SILVEIRA, J. L. NESSIMIAN, AND L.F. M. DORVILLE. 2002. Influence of water chemistry andenvironmental degradation on macroinvertebrate assem-blages in a river basin in south-east Brazil. Hydro-biologia 481:125–136.

CHESHIRE, K., L. BOYERO, AND R. G. PEARSON. 2005. Food websin tropical Australian streams: shredders are not scarce.Freshwater Biology 50:748–769.

COLEY, P. D., AND T. M. AIDE. 1991. Comparison of herbivoryand plant defenses in temperate and tropical broad-leaved forests. Pages 25–49 in P. W. Price, T. M.Lewinsohn, G. W. Fernandes, and W. W. Benson(editors). Plant–animal interactions: evolutionary ecolo-gy in tropical and temperate regions. John Wiley andSons, New York.

CROWL, T. A., W. H. MCDOWELL, A. P. COVICH, AND S. L.JOHNSON. 2001. Freshwater shrimp effects on detritalprocessing and nutrients in a tropical headwater stream.Ecology 82:775–783.

CUMMINS, K. W., AND M. J. KLUG. 1979. Feeding ecology ofstream invertebrates. Annual Review of Ecology andSystematics 10:147–172.

DANGLES, O. 2002. Functional plasticity of benthic macroin-vertebrates: implications for trophic dynamics in acidstreams. Canadian Journal of Fisheries and AquaticSciences 59:1563–1573.

DANGLES, O. J., AND F. A. GUEROLD. 2000. Structural andfunctional responses of benthic macroinvertebrates toacid precipitation in two forested headwater streams(Vosges Mountains, northeastern France). Hydrobiologia418:25–31.

DOBSON, M. 2004. Freshwater crabs in Africa. FreshwaterForum 21:3–26.

DOBSON, M., AND A. G. HILDREW. 1992. A test of resourcelimitation among shredding detritivores in low orderstreams in southern England. Journal of Animal Ecology61:69–77.

DOBSON, M., A. MAGANA, J. M. MATHOOKO, AND F. K. NDEGWA.2002. Detritivores in Kenyan highland streams: moreevidence for the paucity of shredders in the tropics?Freshwater Biology 47:909–919.

DUDGEON, D. 1982. An investigation of physical andbiological processing of 2 species of leaf litter in Tai Po

Kau Forest Stream, New Territories, Hong Kong. Archivfur Hydrobiologie 96:1–32.

DUDGEON, D. 1989. The influence of riparian vegetation on thefunctional organization of four Hong Kong streamcommunities. Hydrobiologia 179:183–194.

DUDGEON, D. 1999a. The population dynamics of threespecies of Calamoceratidae (Trichoptera) in a tropicalforest stream. Pages 83–91 in H. Malicky and P.Chantaramongkol (editors). Proceedings of the 9th

International Symposium on Trichoptera. Faculty ofScience, University of Chiang Mai, Chiang Mai, Thai-land.

DUDGEON, D. 1999b. Tropical Asian streams: zoobenthos,ecology and conservation. Hong Kong University Press,Hong Kong SAR, China.

DUDGEON, D. 2000. The ecology of tropical Asian rivers andstreams in relation to biodiversity conservation. AnnualReview of Ecology and Systematics 31:239–263.

DUDGEON, D. 2006. The impacts of human disturbance onstream benthic invertebrates and their drift in NorthSulawesi, Indonesia. Freshwater Biology 51:1710–1729.

DUDGEON, D., AND R. T. CORLETT. 2004. The ecology andbiodiversity of Hong Kong. Friends of the Country Parkand Joint Publishers, Hong Kong SAR, China.

DUDGEON, D., AND K. K. Y. WU. 1999. Leaf litter in a tropicalstream: food or substrate for macroinvertebrates? Archivfur Hydrobiologie 146:65–82.

DUNCAN, W. F. A., AND M. A. BRUSVEN. 1985. Benthicmacroinvertebrates in logged and unlogged low-ordersoutheast Alaskan streams. Freshwater InvertebrateBiology 4:125–132.

ELWOOD, J. W., J. D. NEWBOLD, A. F. TRIMBLE, AND R. W. STARK.1981. The limiting role of phosphorus in a woodlandstream ecosystem: effects of P enrichment on leafdecomposition and primary producers. Ecology 62:146–158.

FENOGLIO, S., T. BO, AND M. CUCCO. 2004. Small-scalemacroinvertebrate distribution in a riffle of Neotropicalrainforest stream (Rio Bartola, Nicaragua). CaribbeanJournal of Science 40:253–257.

FISHER, S. G., AND G. E. LIKENS. 1973. Energy flow in BearBrook, New Hampshire: an integrative approach tostream ecosystem metabolism. Ecological Monographs43:421–439.

FRIBERG, N. 1997. Benthic invertebrate communities in sixDanish forest streams: impact of forest type on structureand function. Ecography 20:19–28.

FRIBERG, N., AND D. JACOBSEN. 1999. Variation in growth of thedetritivore-shredder Sericostoma personatum (Trichop-tera). Freshwater Biology 42:625–635.

GJERLØV, C., AND J. S. RICHARDSON. 2004. Patchy resources in aheterogeneous environment: effects of leaf litter andforest cover on colonisation patterns of invertebrates in aBritish Columbian stream. Archiv fur Hydrobiologie 161:307–327.

GRAY, L. J., AND K. W. JOHNSON. 1988. Trophic structure ofbenthic macroinvertebrates in Kings Creek. Transactionsof the Kansas Academy of Science 91:178–184.

GREATHOUSE, E. A., AND C. M. PRINGLE. 2006. Does the river



continuum concept apply on a tropical island? Longitu-dinal variation in a Puerto Rican stream. CanadianJournal of Fisheries and Aquatic Sciences 63:134–152.

HAGEN, E. M., J. R WEBSTER, AND E. F. BENFIELD. 2006. Are leafbreakdown rates a useful measure of stream integrityalong an agricultural landuse gradient? Journal of theNorth American Benthological Society 25:330–343.

HALL, R. O., G. E. LIKENS, AND H. M. MALCOM. 2001. Trophicbasis of invertebrate production in 2 streams at theHubbard Brook Experimental Forest. Journal of theNorth American Benthological Society 20:432–447.

HAWKINS, C. P., M. L. MURPHY, AND N. H. ANDERSON. 1982.Effects of canopy, substrate composition, and gradient onthe structure of macroinvertebrate communities inCascade Range streams of Oregon. Ecology 63:1840–1856.

HOANG, D. H., AND Y. J. BAE. 2006. Aquatic insect diversity ina tropical Vietnamese stream in comparison with that ina temperate Korean stream. Limnology 7:45–55.

IRONS, J. G., M. W. OSWOOD, AND J. P. BRYANT. 1988.Consumption of leaf detritus by a stream shredder:influence of tree species and nutrient status. Hydro-biologia 160:53–61.

IRONS, J. G., M. W. OSWOOD, R. J. STOUT, AND C. M. PRINGLE.1994. Latitudinal patterns in leaf litter breakdown: istemperature really important? Freshwater Biology 32:401–411.

JOHNSON, B. R., W. F. CROSS, AND J. B. WALLACE. 2003. Long-term resource limitation reduces insect detritivoregrowth in a headwater stream. Journal of the NorthAmerican Benthological Society 22:565–574.

JONSSON, M., AND B. MALMQVIST. 2000. Ecosystem process rateincreases with animal species richness: evidence fromleaf-eating, aquatic insects. Oikos 89:519–523.

JONSSON, M., B. MALMQVIST, AND P. O. HOFFSTEN. 2001. Leaflitter breakdown rates in boreal streams: does shredderspecies richness matter? Freshwater Biology 46:161–171.

KING, R. S., AND C. J. RICHARDSON. 2002. Evaluatingsubsampling approaches and macroinvertebrate taxo-nomic resolution for wetland bioassessment. Journal ofthe North American Benthological Society 21:150–171.

KOBAYASHI, S., AND T. KAGAYA. 2002. Differences in littercharacteristics and macroinvertebrate assemblages be-tween litter patches in pools and riffles in a headwaterstream. Limnology 3:37–42.

LECERF, A., P. USSEGLIO-POLATERA, J. Y. CHARCOSSET, D.LAMBRIGOT, B. BRACHT, AND E. CHAUVET. 2006. Assessmentof functional integrity of eutrophic streams using litterbreakdown and benthic macroinvertebrates. Archiv furHydrobiologie 165:105–126.

LI, A. O. Y. 2008. Shredders and leaf litter breakdown inHong Kong streams. MPhil Thesis, The University ofHong Kong, Hong Kong SAR, China.

LI, A. O. Y., AND D. DUDGEON. 2008. Food resources ofshredders and other benthic macroinvertebrates inrelation to shading conditions in tropical Hong Kongstreams. Freshwater Biology 53:2011–2025.

LI, A. O. Y., L. C. Y. NG, AND D. DUDGEON. 2008. Effects of leaftoughness and nitrogen content on litter breakdown and

macroinvertebrates in a tropical stream. Aquatic Sciences(in press).

LUGTHART, G. J., AND J. B. WALLACE. 1992. Effects ofdisturbance on benthic functional structure and produc-tion in mountain streams. Journal of the North AmericanBenthological Society 11:138–164.

MANTEL, S. K., AND D. DUDGEON. 2004. Dietary variation in apredatory shrimp Macrobrachium hainanense (Palaemoni-dae) in Hong Kong forest streams. Archiv fur Hydro-biologie 160:305–328.

MARIDET, L., J. G. WASSON, M. PHILIPPE, C. AMOROS, AND R. J.NAIMAN. 1998. Trophic structure of three streams withcontrasting riparian vegetation and geomorphology.Archiv fur Hydrobiologie 144:61–85.

MATHURIAU, C., AND E. CHAUVET. 2002. Breakdown of leaf litterin a Neotropical stream. Journal of the North AmericanBenthological Society 21:384–396.

MERRITT, R. W., AND K. W. CUMMINS (EDITORS). 1996. Anintroduction to the aquatic insects of North America. 3rd

edition. Kendall/Hunt Publishing Co., Dubuque, Iowa.MISERENDINO, M. L., AND L. A. PIZZOLON. 2004. Interactive

effects of basin features and land-use change onmacroinvertebrate communities of headwater streamsin the Patagonian Andes. River Research Applications20:967–983.

MURPHY, J. F., AND P. S. GILLER. 2000. Seasonal dynamics ofmacroinvertebrate assemblages in the benthos andassociated with detritus packs in two low-order streamswith different riparian vegetation. Freshwater Biology43:617–631.

NIYOGI, D. K., K. S. SIMON, AND C. R. TOWNSEND. 2003.Breakdown of tussock grass in streams along a gradientof agricultural development in New Zealand. FreshwaterBiology 48:1698–1708.

PADGETT, D. E. 1976. Leaf decomposition by fungi in a tropicalrainforest stream. Biotropica 8:166–178.

PEARSON, R. G., R. K. TOBIN, R. E. W. SMITH, AND L. J. BENSON.1989. Standing crop and processing of rainforest litter ina tropical Australian stream. Archiv fur Hydrobiologie115:481–498.

PRINGLE, C. M., AND A. RAMIREZ. 1998. Use of both benthic anddrift sampling techniques to assess tropical streaminvertebrate communities along an altitudinal gradient,Costa Rica. Freshwater Biology 39:359–373.

QUINN, J. M., G. P. BURRELL, AND S. M. PARKYN. 2000. Influencesof leaf toughness and nitrogen content on in-streamprocessing and nutrient uptake by litter in a Waikato,New Zealand, pasture stream and streamside channels.New Zealand Journal of Marine and Freshwater Re-search 34:253–271.

REED, J. L., I. C. CAMPBELL, AND P. C. E. BAILEY. 1994. Therelationship between invertebrate assemblages andavailable food at forest and pasture sites in threesouth-eastern Australian streams. Freshwater Biology32:641–650.

ROSEMOND, A. D., C. M. PRINGLE, AND A. RAMIREZ. 1998.Macroconsumer effects on insect detritivores and detri-tus processing in a tropical stream. Freshwater Biology39:515–523.



SPONSELLER, R. A., AND E. F. BENFIELD. 2001. Influences of landuse on leaf breakdown in southern Appalachian head-water streams: a multiple-scale analysis. Journal of theNorth American Benthological Society 20:44–59.

STONE, M. K., AND J. B. WALLACE. 1998. Long-term recovery ofa mountain stream from clear-cut logging: the effects offorest succession on benthic invertebrate communitystructure. Freshwater Biology 39:151–169.

STOUT, R. J. 1989. Effects of condensed tannins on leafprocessing in mid-latitude and tropical streams: atheoretical approach. Canadian Journal of Fisheries andAquatic Sciences 46:1097–1106.

TUMWESIGYE, C., S. K. YUSUF, AND B. MAKANGA. 2000. Structureand composition of benthic macroinvertebrates of atropical forest stream, River Nyamweru, western Ugan-da. African Journal of Ecology 38:72–77.

VANNOTE, R. L., G. W. MINSHALL, K. W. CUMMINS, J. R. SEDELL,AND C. E. CUSHING. 1980. The river continuum concept.Canadian Journal of Fisheries and Aquatic Sciences 37:130–137.

VELASQUEZ, S. M., AND M. L. MISERENDINO. 2003. Habitat typeand macroinvertebrate assemblages in low-order Pata-gonian streams. Archiv fur Hydrobiologie 158:461–483.

VOUGHT, L. B. M., A. KULLBERG, AND R. C. PETERSEN. 1998. Effectof riparian structure, temperature and channel morphom-etry on detritus processing in channelized and naturalwoodland streams in southern Sweden. Aquatic Conser-vation: Marine and Freshwater Ecosystems 8:273–285.

WALLACE, J. B., S. L. EGGERT, J. L. MEYER, AND J. R. WEBSTER.1997. Multiple trophic levels of a forest stream linked toterrestrial litter inputs. Science 277:102–104.

WALLACE, J. B., S. L. EGGERT, J. L. MEYER, AND J. R. WEBSTER.

1999. Effects of resource limitation on a detrital-based

ecosystem. Ecological Monographs 69:409–442.

WEBSTER, J. R., AND E. F. BENFIELD. 1986. Vascular plant

breakdown in freshwater ecosystems. Annual Review of

Ecology and Systematics 17:567–594.

WHILES, M. R., AND J. B. WALLACE. 1997. Leaf litter

decomposition and macroinvertebrate communities in

headwater streams draining pine and hardwood catch-

ments. Hydrobiologia 353:107–119.

WINTERBOURN, M. J. 1995. Rivers and streams of New

Zealand. Pages 695–716 in C. E. Cushing, K. W.

Cummins, and G. W. Minshall (editors). Ecosystems of

the world. Volume 22: river and stream ecosystems.

Elsevier Press, New York.

WINTERBOURN, M. J., J. S. ROUNICK, AND B. COWIE. 1981. Are

New Zealand stream ecosystems really different? New

Zealand Journal of Marine and Freshwater Research 15:

321–328.

YAM, R. S. W., AND D. DUDGEON. 2005. Stable isotope

investigation of food use by Caridina spp. (Decapoda:

Atyidae) in Hong Kong streams. Journal of the North

American Benthological Society 24:68–81.

YULE, C. M. 1996. Trophic relationships and food webs of the

benthic invertebrate fauna of two aseasonal tropical

streams on Bougainville Island, Papua New Guinea.

Journal of Tropical Ecology 12:517–534.

Received: 10 March 2008Accepted: 9 October 2008



shredders: species richness, abundance, and role in litter breakdown in tropical hong kong streams

Documents