Mass loss and macroinvertebrate colonisation of Pacificsalmon carcasses in south-eastern Alaskan streams

DOMINIC T. CHALONER,* MARK S. WIPFLI† and JOHN P. CAOUETTE†

*Department of Entomology, Michigan State University, East Lansing, MI, U.S.A.

†Pacific Northwest Research Station, USDA Forest Service, Juneau, AK, U.S.A.

SUMMARY

1. We examined the spatial and temporal dynamics of pink salmon (Oncorhynchusgorbuscha) carcass decomposition (mass loss and macroinvertebrate colonisation) in south-

eastern Alaskan streams. Dry mass and macroinvertebrate fauna of carcasses placed in

streams were measured every two weeks over two months in six artificial streams and

once after six weeks in four natural streams. We also surveyed the macroinvertebrate

fauna and wet mass of naturally occurring salmon carcasses.

2. Carcass mass loss in artificial streams was initially rapid and then declined over time

(k ¼ –0.033 day–1), and no significant differences were found among natural streams.

3. Several macroinvertebrate taxa colonised carcasses, but chironomid midge (Diptera:

Chironomidae) and Zapada (Plecoptera: Nemouridae) larvae were found consistently and

were the most abundant (on average 95 and 2%, respectively, of the invertebrates found).

Chironomid abundance and biomass increased over time, whereas Zapada abundance and

biomass did not. Significant differences in abundance were found among natural streams

for Baetis (Ephemeroptera: Baetidae) and Sweltsa (Plecoptera: Chloroperlidae) larvae, while

no significant differences were found for chironomid and Zapada abundance or biomass.

4. Our results suggest that salmon carcasses initially undergo a high rate of mass loss that

tapers off with time. Chironomid and Zapada larvae are likely to be important in mediating

nutrient and energy transfer between salmon carcasses and other components of the

freshwater-riparian food web in south-eastern Alaskan streams.

Keywords: carcasses, decomposition, invertebrate colonisation, mass loss, Pacific salmon

Introduction

The annual migration of Pacific salmon (Oncorhynchus

spp.) results in the movement of large amounts of

organic material into north Pacific rim catchments

(Cederholm et al., 1999). These salmon runs are

especially remarkable in Alaska and western Canada

where salmon populations have, until quite recently,

been stable (Baker et al., 1996). Marine-derived nitro-

gen and carbon delivered by salmon can be traced

into freshwater and terrestrial organisms (Schuldt &

Hershey, 1995; Bilby, Fransen & Bisson, 1996; Kline,

Goering & Piorkowski, 1997) and may be a resource

subsidy critical for maintaining freshwater and ripar-

ian productivity (Polis, Anderson & Holt, 1997).

However, the ecological processes by which salmon

influence ecosystems are not well understood

(Cederholm et al., 1999) and the overall influence of

marine resource subsidies probably underestimated

(Willson, Gende & Marston, 1998).

Little is known about carrion decomposition in

freshwater (Minshall, Hitchcock & Barnes, 1991;

Parmenter & Lamarra, 1991; Merritt & Wallace,

2000), including mass loss and invertebrate colonisa-

tion. Much more is known about the decomposition

of terrestrial carrion (Hanski, 1987) and other organic

inputs to freshwater, such as leaf litter (Webster &

Benfield, 1986). Terrestrial carrion, for example,

Correspondence: Dominic Chaloner, Department of Biological

Sciences, University of Notre Dame, Notre Dame, IN 46556-0369,

U.S.A. E-mail: chaloner.1@nd.edu

Freshwater Biology (2002) 47, 263–273

Ó 2002 Blackwell Science Ltd 263

undergoes a predictable sequence of mass loss and

invertebrate colonisation over time (Hanski, 1987;

Haskell et al., 1989), which can be influenced by

conditions associated with different biomes, such as

grassland versus woodland (Hanski, 1987). Although

obligate necrophagous invertebrate taxa have

evolved only on land where carrion supply is

frequent and predictable (Hanski, 1987; Haskell et al.,

1989), invertebrates are important processors of

marine as well as terrestrial carrion (Britton &

Morton, 1994). Freshwater invertebrates are generally

important agents in the breakdown of organic

material, nutrient cycling and energy flow within

food webs (Cummins, 1973), and can be abundant on

carrion in freshwaters (Vance, Van Dyk & Rowley,

1995; Keiper, Chapman & Foote, 1997; Kline et al.,

1997; Minakawa & Gara, 1999). Invertebrate coloni-

sation has been observed on adult salmon carcasses

throughout the Pacific north-west (Kline et al., 1997;

Minakawa & Gara, 1999), including south-eastern

Alaska where invertebrates have been found scaven-

ging on salmon carcasses (Wipfli, Hudson & Caou-

ette, 1998) and dead salmon alevins and eggs (Ellis,

1970; Elliott & Bartoo, 1981).

Our objectives in this study were to measure the

rates of salmon carcass decomposition, in terms

of mass loss and the abundance and biomass

of invertebrate colonists, and to compare rates of

decomposition among several streams. The results

of this study will contribute to a growing body of

knowledge concerning the influence of decom-

posing salmon on freshwater habitats. A better

understanding of the dynamics of Pacific salmon

decomposition should aid fisheries and ecosystem

management.

Methods

Study location

This study took place in the Margaret Creek catch-

ment on Revillagigedo Island in south-eastern Alaska,

U.S.A. Several anadromous salmonid species, inclu-

ding pink (Oncorhynchus gorbuscha), sockeye (O. nerka),

chum (O. keta) and coho salmon (O. kisutch), regularly

spawn in Margaret and Cobble Creeks, and since

installation of a fish ladder in 1989, in Sprout Fork and

Spike Creek (Bryant, Frenette & McCurdy, 1999).

Spawning takes place from July until late November,

but mostly during August and early September.

Our study combined artificial stream experiments

with natural stream experiments and surveys (cf.

Wipfli et al., 1998, 1999). This enabled us to study

temporal changes in artificial streams and spatial

differences among natural streams, as well as com-

pare patterns under artificial and more realistic

natural conditions. Artificial stream experiments were

conducted in six channels of a larger ‘mesocosm’

(Wipfli et al., 1998 for full description) supplied with

water from a portion of Spike Creek inaccessible to

salmon. Natural stream experiments took place in

Margaret Creek, Cobble Creek, Spike Creek and

Sprout Fork, which represented contrasts in dis-

charge, riparian management, stream habitat, and

history of salmon spawning (Table 1). Surveys of

naturally occurring salmon carcasses were limited to

Cobble Creek and Margaret Creek as only a few

carcasses were found in Sprout Fork and Spike Creek.

Surveys provided information about the invertebrate

abundance on carcasses and the amount of carcass

material present in streams during a natural salmon

run.

Discharge

Dissolved

oxygen Temperature Temperature Riparian

Stream (m3 s–1) (mg L–1) (°C, periodic) (°C, continuous) forest type

Margaret 3.83 10.6 10.4 9.2 Mainstem – mixed

(1.26) (9.9–11.6) (8.6–11.6) (6.2–11.6) old/young growth.

Cobble 0.58 11.1 8.5 n/a Tributary – mixed

(0.17) (10.5–11.7) (7.3–10.1) old/young growth.

Spike 0.19 10.7 9.4 8.8 Tributary – old

(0.12) (10.3–11.4) (7.9–11.0) (5.5–11.4) growth.

Sprout 0.06 10.5 9.1 n/a Tributary – young

(0.01) (10.3–10.8) (7.4–9.8) growth.

Table 1 Baseflow discharge [average

(±1 SEM)], dissolved oxygen [average

(range)], water temperature [periodic

and continuous: average (range)] and

riparian forest type of the four streams

used in this study

264 D.T. Chaloner et al.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

Carcasses placed in artificial streams

On 1 September 1997, 24 female pink salmon (48–

52 cm fork length, 1.25–1.49 kg wet mass, 0.35–0.42 kg

dry mass) were taken live from Margaret Creek and

killed. Initial dry mass of each fish was estimated by

multiplying the original wet mass by 0.28 (Wipfli

et al., 1998). Fresh, egg-bearing rather than spawned-

out salmon were used to ensure uniformity among

and between experiments. Salmon carcasses in similar

condition are commonly found in and around

streams. Four carcasses were tethered to the floor of

each artificial stream channel, on top of fibreglass

mesh (mesh size ¼ 1.6 mm) to facilitate their removal.

Channels were dammed to create pools 250 cm long,

18 cm wide and 13 cm deep. During the experiment,

discharge through the channels averaged 0.488 ±

0.001 L s–1 and water temperature, recorded continu-

ously at 1-h intervals with Optic StowawayÒ1 datal-

oggers, averaged 8.9 °C (range 2.4–12.9 °C). Dissolved

oxygen was measured semiweekly using a YSIÒ meter

(model 50B, Yellow Springs Instruments Inc., Yellow-

Springs, OH, USA), and averaged 11.74 mg L–1 (range

10.86–12.89 mg L–1).

After days 17, 28, 42 and 56, a single carcass was

randomly selected and removed from each channel.

Carcasses were rinsed several times with water, and

the washings put through 2-mm and 250-lm nested

sieves. The coarser sieve separated out large pieces of

salmon tissue, from which invertebrates were extrac-

ted without magnification and preserved in 95%

EtOH with material from the finer sieve. Tissue pieces

were frozen and later dried to a constant mass at

80 °C to determine the carcass mass remaining. The

material from the finer sieve consisted primarily of

invertebrates and silt, with insufficient salmon tissue

to justify determining its dry mass.

All invertebrates were extracted from each sample,

except large numbers (>1000) of chironomid midge

(Diptera: Chironomidae) larvae that were subsam-

pled (one-quarter). Invertebrates were preserved

with 70% EtOH and identified to family or genus

using Merritt & Cummins (1996). The lengths of

chironomid midge and Zapada (Plecoptera: Nemouri-

dae) larvae, the dominant carcass invertebrates, were

used in conjunction with established mass-length

equations (Smock, 1980; M.S. Wipfli, unpublished

data) to estimate total invertebrate biomass for each

carcass.

Carcasses placed in natural streams

Salmon carcasses were placed into four streams

(Sprout Fork, Spike, Cobble and Margaret Creeks),

between 2 and 10 September 1998. The pink female

salmon used were similar in condition and size

(48–52 cm fork length, 1.25–2.02 kg wet mass and

0.35–0.56 kg dry mass) to those used in artificial

streams. Carcasses were placed in wire mesh cages

(55 · 25 cm, galvanised fencing wire, mesh

size ¼ 5 cm) with a base of fibreglass mesh (mesh

size ¼ 1.6 mm) to facilitate collection and reduce loss

of material during retrieval. In each stream, four cages

were placed within two log jams where carcasses

were already present (D.T. Chaloner, personal obser-

vation). Cages were secured with parachute cord and

positioned so that they remained submerged and

were reasonably protected from vertebrate scaven-

gers. Temperature dataloggers recorded water tem-

perature hourly, and dissolved oxygen and stream

discharge were measured periodically (Table 1). Data-

loggers in Sprout Fork and Cobble Creek recorded

temperature for only part of the time, but data suggest

that all streams were similar.

Carcass cages were left in place for six weeks, long

enough to allow invertebrates to reach their maximum

abundance, based on artificial stream experiments,

and short enough to avoid autumn floods. At the end

of the experiment, carcasses were removed from cages

and processed as described above. Of the original 32

cages, 25 were recovered: eight from Sprout Fork, five

from Spike Creek, five from Margaret Creek, and

seven from Cobble Creek. Bears had tampered with

the others. One cage from Sprout Fork was completely

buried by sand, and therefore was not included in

statistical analyses.

Carcasses occurring naturally in streams

Surveys of salmon carcasses were undertaken three

times between the first sighting of chum salmon (26

July 1998) and two weeks after live pink salmon were

last seen in the streams (28 September 1998). Surveys

took place along a single 200-m reach of Cobble Creek

1The use of trade or firm names in this publication is for reader

information and does not imply endorsement by the U.S.

Department of Agriculture of any product or service.

Mass loss and macroinvertebrate colonisation of salmon carcasses 265

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

and Margaret Creek, moving in an upstream direction

to avoid encountering the same carcass twice. During

surveys, relative abundance (0–9, 10–99, 100–999, or

>1000 individuals per carcass) was estimated for

invertebrate taxa found on 30 carcasses. Each taxa

was assigned to a particular category (absent, trace,

present, or common) based upon their overall relative

abundance scores and the number of carcasses on

which they were found. The wet mass of 100 carcasses

was also measured and the carcasses returned to the

stream after weighing.

Experimental design and statistical analyses

A randomised complete block design was used for the

artificial stream experiment, with six channels as

blocks and four time periods as treatments. Each time

period was randomly assigned to one of four carcasses

in each block. Response variables were macroinverte-

brate abundance and biomass. Data were natural

logarithm (ln) transformed to meet the assumptions of

analysis of variance (ANOVA) and analysed with

PROC GLM (SAS Institute, 1989) at a ¼ 0.05. Carcass

mass loss in artificial streams was quantitatively

modelled using percent carcass mass remaining and

with best-fit exponential regression curves, as used in

other studies (Minshall et al., 1991; Parmenter &

Lamarra, 1991).

For the natural stream experiment, we analysed the

data using a nested ANOVA, with carcasses nested

within log jams and log jams nested within streams.

Response variables were percent carcass mass

remaining, macroinvertebrate abundance, and macr-

oinvertebrate biomass. Data were ln-transformed to

meet ANOVA assumptions and analysed with PROC

GLM (SAS Institute, 1989) at a ¼ 0.05. Data from the

surveys of naturally occurring carcasses were not

analysed statistically because of the structure and

qualitative nature of the data.

Results

Carcasses placed in artificial streams

Salmon carcasses changed in appearance over time,

although distinct stages were not evident. A thick,

furry mould-like coating developed within two days

of carcasses being placed in channels, which gradu-

ally turned into a dark slime layer as organic material

and silt accumulated on carcasses. The muscle tissue

became paste-like within a month and more fluid with

time, while skin, eggs and internal organs persisted

throughout the experiment. The percent carcass mass

remaining reflected these observations and fitted a

single exponential decay curve (Fig. 1). A double

exponential regression curve was unsuitable given the

sampling intervals and the experimental duration (cf.

Parmenter & Lamarra, 1991).

Of the 23 invertebrate taxa found on carcasses,

chironomid midge and Zapada larvae were the most

abundant taxa, constituting 97% (range 83–99%) and

3% (range <1–13%) of invertebrates found, respect-

ively (Table 2). Invertebrates were found predomin-

antly on the carcass surface but were also found

among the eggs and in the gill and mouth cavities

when carcasses were removed after six to eight weeks.

Chironomid abundance and biomass increased sig-

nificantly over time (P < 0.001, Table 2), whereas

Zapada biomass and abundance did not (P > 0.05,

Table 2) (Fig. 2).

Carcasses placed in natural streams

Salmon carcasses recovered from a single log jam

exhibited a range of percent remaining values (e.g.

8–29%). Some carcasses had missing parts, while

others were fully intact, which partly reflected how

exposed carcasses were within the log jams. How-

ever, the average carcass mass percent remaining for

Fig. 1 Mean carcass mass remaining (percentage of original dry

mass) after sampling salmon carcasses placed in artificial

streams. Error bars represent ±1 SEM.

266 D.T. Chaloner et al.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

natural streams (26 ± 2%, mean ± SEM) was similar

to that found in artificial streams after six weeks

(20 ± 2%, mean ± SEM). As in artificial streams,

skin, eggs and internal organs had not decomposed

appreciably, while the muscle tissue was paste-like.

Carcasses were covered with a dense, furry, blue-

gray mould-like layer, firmly attached to the skin,

unlike the slime layer found in artificial streams.

Overall, there were no obvious differences in carcass

appearance among natural streams and no signi-

ficant differences amongst streams in percent

mass remaining (P > 0.05), suggesting that carcass

decomposition was similar amongst streams in

terms of mass loss.

Table 2 Functional feeding group (c-g, collector-gatherer; om, omnivore; p, predator; sc, scraper; sh, shredder; after Merritt &

Cummins, 1996), mean percent relative abundance (calculated by pooling across all time periods or streams), mean actual abundance

(number per carcass), mean biomass (mg per carcass), and A N O V A results (with time or stream as the treatment, n = 24, d.f. = 3, 15) for

benthic macroinvertebrates collected from carcasses placed in artificial or natural streams. Values for biomass given in parentheses

below those for abundance

Artificial streams Natural streams

Functional

feeding Relative AbundanceA N O V AA N O V A

Relative AbundanceA N O V AA N O V A

Taxon group abundance (biomass) F P abundance (biomass) F P

Arthropoda

Insecta

Ephemeroptera* c-g, sc 0.2 1.9 nt†

Baetidae

Baetis c-g, sc 0.8 18.4 10.22 0.024

Other Ephemeroptera‡ c-g, sc 0.4 6.0 nt

Plecoptera

Nemouridae

Zapada sh 2.5 33.5 0.67 0.581 2.2 32.2 0.42 0.749

(15.1) (2.86) (0.072) (31.5) (0.51) (0.696)

Chloroperlidae

Sweltsa p 0.6 8.0 18.26 0.009

Capniidae sh 1.7 19.0 2.56 0.193

Other Plecoptera§,– sh, p 0.1 1.5 nt <0.1 0.3 nt

Coleoptera** several <0.1 0.1 nt 0.1 0.8 nt

Trichoptera†† several 0.4 4.4 nt

Onocosmoecus sh 0.4 5.0 1.07 0.457

Other Trichoptera‡‡ several 0.4 6.2 1.89 0.273

Diptera

Chironomidae several 96.6 2155.8 22.11 <0.001 92.6 1526.5 2.91 0.164

(783.7) (75.85) (<0.001) (251.7) (1.49) (0.346)

Simuliidae cf 0.1 3.8 nt

Other Diptera§§,–– several 0.2 2.2 nt 0.4 5.2 nt

Arachnida

Acari om <0.1 0.3 nt <0.1 0.1 nt

Annelida

Oligochaeta c-g <0.1 0.3 nt 0.2 2.9 nt

All taxa 100.0 2200.2 19.63 <0.001 100.0 1635.6 2.66 0.184

*Baetis, Cinygmula, Paraleptophlebia.

†Not tested.

‡Cinygmula, Drunella doddsi, Ephemerella, Paraleptophlebia, Serratella.

§Artificial streams: Capniidae, Leuctridae, Sweltsa.

–Natural streams: Hesperoperla, Leuctridae.

**Dytiscidae, Elmidae.

††Ecclisomyia, Micrasema, Onocosmoecus, Polycentropus, Rhyacophila.

‡‡Ecclisomyia, Hydropsychidae, Micrasema, Polycentropus, Rhyacophila.

§§Artificial streams: Ceratopogonidae, Empididae, Simuliidae, Tipulidae.

––Natural streams: Ceratopogonidae, Empididae, Tipulidae.

Mass loss and macroinvertebrate colonisation of salmon carcasses 267

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

Of the 27 invertebrate taxa found on carcasses

(Table 2), chironomid midges, Zapada, Capniidae

(Plecoptera), Baetis (Ephemeroptera: Baetidae), Sweltsa

(Plecoptera: Chloroperlidae), and Onocosmoecus (Tri-

choptera: Limnephilidae) larvae were the most com-

mon. Some invertebrates, especially chironomids,

were found in the gills and mouth and buried within

carcass tissue. As in artificial streams, chironomids

and Zapada were the only taxa found on all carcasses

and constituted 93% (range 83–99%) and 2% (range

<1–5%) of the invertebrates found, respectively. The

low statistical power of this study design meant only

large differences among streams, relative to within-

stream variation (between log jams), would be detect-

able. So although there appeared to be considerable

variation among streams in the abundance of several

taxa (Fig. 3), these differences were only significant

for the abundance of Sweltsa (P ¼ 0.0085) and Baetis

(P ¼ 0.024) (Table 2, Fig. 3). For the abundance of all

other taxa and biomass of chironomids and Zapada, no

significant differences were found among streams

(P > 0.05) (Table 2, Figs 3 and 4).

Carcasses occurring naturally in streams

Naturally occurring carcass material was found in a

variety of forms, reflecting the many ways that fish-

died (e.g. natural death, predator kills) and carcasses

Fig. 2 Mean abundance and biomass estimates of chironomid midge and Zapada larvae found on salmon carcasses placed in

artificial streams. Error bars represent ±1 SEM.

268 D.T. Chaloner et al.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

were moved around streams (e.g. scavenging, spates).

Carcass material ranged from complete fish (both

fresh and well decomposed) to tissue fragments. The

more decomposed material was often covered with a

layer of dense, mould-like biofilm, similar to that

found on carcasses artificially placed in natural

streams. The average mass of carcass material found

during the first survey of Margaret Creek (Fig. 5)

reflected partial consumption of chum salmon by

predators. By the first survey of Cobble Creek,

predators left chum carcasses virtually untouched, in

contrast to the smaller pink salmon that began to

appear in streams at this time. Towards the end of the

salmon run, whole carcasses were seen less fre-

quently, while small to medium-sized (5–20-cm

diameter) tissue fragments were more common. The

average mass of carcass material found over time

(Fig. 5) reflected these observations.

Several invertebrate taxa were found on naturally

occurring salmon carcasses (Table 3), with the most

abundant and common being chironomid larvae,

consistent with both artificial and natural stream

experiments. Zapada larvae were also found, but not

as often as mayflies and caddisflies. Fewer inverte-

brates were observed on carcasses in Cobble Creek

than Margaret Creek, consistent with the invertebrate

abundance found on artificially placed carcasses. In

Margaret Creek, invertebrates were not found on

carcasses in early August, but were present in Octo-

ber. In Cobble Creek, invertebrates were observed on

carcasses on all surveying dates. In both streams,

invertebrates were found in the mouth and on gills,

fin rays, and the surface of carcasses.

Fig. 3 Mean abundance of the most abundant invertebrate

taxa found on salmon carcasses artificially placed in natural

streams for six weeks. Error bars represent ±1 SEM.

Fig. 4 Mean biomass estimates of chironomid midge (j) and

Zapada (h) larvae found on salmon carcasses artificially placed

in natural streams for 6 weeks. Error bars represent ±1 SEM.

Fig. 5 Mean wet mass of carcass material found naturally in

Cobble (–j–) and Margaret (–r–) Creeks. Error bars represent

±1 SEM.

Mass loss and macroinvertebrate colonisation of salmon carcasses 269

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

Discussion

In our study, salmon carcasses persisted for over

six weeks and exhibited an exponential loss of mass

loss over time. Such a pattern has previously been

found with plant and animal material (Parmenter &

Lamarra, 1991), including freshwater carrion (e.g.

Payne & King, 1972; Minshall et al., 1991; Parmenter

& Lamarra, 1991). Results reported by Wipfli et al.

(1998) indicate an exponential loss of mass for Pacific

salmon although a regression curve was not fitted.

The decomposition rate of fish in our study

(k ¼ –0.033 day–1 for a single exponential regression)

was slower than that found by Parmenter & Lamarra

(1991) (k ¼ –0.061 day–1 for the first part of a double

exponential regression) and Minshall et al. (1991)

(k ¼ –0.048 day–1 during summer), but was still great-

er than that found for most non-woody and woody

plants (range of k ¼ –0.001 to –0.05 day–1, Webster &

Benfield, 1986). Our results suggest that salmon

carcass decomposition in water is likely to be slow,

taking more than a month to complete.

Overall, salmon carcass mass loss differed little

amongst streams, but there was considerable variation

within each log jam. This suggests that microhabitat

conditions, such as water flow and dissolved oxygen,

might be important factors in carrion decomposition.

Kline et al. (1997) observed more rapid salmon carcass

decomposition close to shore and in nearby streams

than at the bottom of Alaskan Lake Iliamna where

anoxia and negligible water flow seems likely.

Another important factor in salmon carcass decom-

position appears to be tissue type. In both artificial

and natural streams, muscle tissue decomposed more

readily than eggs, internal organs, or skin, which is

consistent with other studies (Merritt & Wallace,

2000). For example, Minshall et al. (1991) found that

the skin of fish persisted longer than other body parts.

Persistence of salmon tissue fragments may result in

the sustained release of marine-derived nutrients

(MDN) in streams (cf. Parmenter & Lamarra, 1991)

because these fragments are likely to become trapped

in interstitial spaces.

Several macroinvertebrate species colonised salmon

carcasses in our study, most notably chironomid

midges. Several different invertebrate taxa have been

found on salmon carcasses (Kline et al., 1997;

Minakawa & Gara, 1999) and other freshwater carrion

(Vance et al., 1995; Keiper et al., 1997). Compared with

these studies, we found that chironomid abundance

on salmon carcasses was similar (cf. >4 larvae per g of

salmon tissue) (Minakawa & Gara 1999) or much

larger (cf. <1 larva per g of rat tissue) (Keiper et al.,

1997). However, the other invertebrate taxa reported

Table 3 Invertebrate taxa found on naturally occurring carcasses in Margaret and Cobble Creeks

Margaret Creek Cobble Creek

Taxon 7–8 August 21–22 August 11–12 October 18 August 10 September 3 October

Arthropoda

Ephemeroptera* –† – + + + +

Plecoptera

Nemouridae

Zapada – – + – – –

Other Plecoptera‡ – – + – – –

Trichoptera§ – – + + – –

Diptera

Chironomidae – ++ +++ + +++ ++

Simuliidae – – + – – –

Other Diptera– – – + – – –

Annelida

Oligochaeta – – + – – –

*Baetis, Cinygmula.

†Abundance categories: – absent, + trace, ++ present, +++ common.

‡Capniidae, Sweltsa.

§Hydropsychidae, Onocosmoecus, Rhyacophila.

–Empididae, Tipulidae.

270 D.T. Chaloner et al.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

by previous studies (e.g. Kline et al., 1997; Minakawa

& Gara, 1999), such as stoneflies, notably Zapada and

Sweltsa, and caddisflies, notably Ecclisomyia and

Onocosmoecus, we found at much smaller abundances.

Invertebrates were also much less abundant on

naturally occurring carcasses than on those artificially

placed in natural or artificial streams. Focusing

surveys on carcasses trapped in log jams for several

weeks, may have produced results more similar to

those for other parts of the study because inverte-

brates are likely to be dislodged from carcasses that

are washed downstream.

Salmon carcasses could be important microhabitats

for chironomid midges and Zapada, representing

potential ‘ecological hotspots’ of invertebrate abun-

dance in streams. Wipfli et al. (1998) found that

chironomid abundance in the mineral substrate of

mesocosm pools declined after an early peak to about

12 000 larvae m–2. In our study, using the same

mesocosm, chironomid abundance increased steadily

to about 84 000 larvae m–2 on carcasses. In contrast,

Zapada abundance progressively increased through-

out the study by Wipfli et al. (1998), reaching about

700 larvae m–2 in mineral substrate as compared with

about 3700 larvae m–2 on carcasses in our study.

Chironomid larvae can reach larger population den-

sities in organic substrates (e.g. 250 000 larvae m–2,

Armitage, Cranston & Pinder, 1995), even in the

presence of the low dissolved oxygen levels often

associated with submerged carrion (Merritt &

Wallace, 2000).

Organisms are believed to use carrion as a source of

food and/or shelter (Merritt & Wallace, 2000). Inver-

tebrate consumption of salmon carcass material has

been indicated by anecdotal observations (Elliott &

Bartoo, 1981; Minakawa & Gara, 1999) and stable

isotope analyses (Bilby et al., 1996; Kline et al., 1997),

and seems likely for invertebrates belonging to the

‘shredder’ functional feeding group, such as Zapada,

that chew, burrow, or mine into coarse particulate

organic material (Merritt & Cummins, 1996). Inver-

tebrate colonists could help fragment carcass material,

promote nutrient leaching, and oxygenate anoxic

subsurface areas, as has been shown with other

organic substrates (Parmenter & Lamarra, 1991).

Minshall et al. (1991) suggested that invertebrates

could also transfer decomposing microbes between

carcasses, and from the surface to the carcass interior.

The thick mould-like microbial development

observed on carrion (e.g. Minshall et al., 1991; Vance

et al., 1995; Kline et al., 1997; Merritt & Wallace, 2000;

this study) is probably involved in decomposition and

could sequester nutrients leaching from carcasses. The

biofilm associated with salmon tissue and its interac-

tion with carcass invertebrates warrants study, given

what little is known about biofilm associated with

animal tissue (Stevenson, Bothwell & Lowe, 1996).

Salmon carcass invertebrates may facilitate the

transfer of MDN to higher trophic levels, such as

predatory fishes and insectivorous birds. Carcass

tissue is an important food source for fish (Bilby et al.,

1996), mammals and birds (Willson et al., 1998) but

carcasses become fragmented, buried, or washed away

(Cederholm et al., 1989). In these situations, carcass

invertebrates could convert tissue and associated

biofilm into food for other invertebrates and higher

trophic levels, as invertebrate tissue and faecal

material. Chironomids can produce prodigious

amounts of faecal material and are important food

item for many predators (Armitage et al., 1995). Car-

cass invertebrates, because of their proximity to the

MDN source, may also assimilate more MDN than

invertebrates in the same stream but on different

substrate types. Early invertebrate colonists, although

small in biomass, might be disproportionately

important in this respect because of their presence

during the large initial loss of mass. Cederholm et al.

(1989) emphasised the importance of physical MDN

retention devices (e.g. large wood structures) in

streams. Carcass biofilm and invertebrates may repre-

sent biological MDN retention devices also important

in streams but at a smaller scale, similar to the chemical

sorption processes discussed by Bilby et al. (1996).

The processes and mechanisms by which Pacific

salmon influence ecosystems are complex and involve

many factors, including the rate at which carcasses

lose mass and become colonised by invertebrates. Our

study indicates that salmon carcasses lose mass in a

way similar to other types of organic material and are

colonised by invertebrates in high densities. Better

understanding of salmon decomposition in freshwa-

ter-riparian ecosystems will contribute to the devel-

opment of more holistic Pacific salmon management

and restoration strategies, which can include placing

hatchery carcasses in streams (Ashley & Slaney, 1997).

The design of such projects would ideally reflect the

influence that salmon have upon all trophic levels in

freshwater-riparian food webs.

Mass loss and macroinvertebrate colonisation of salmon carcasses 271

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

Acknowledgments

We thank the following: Ketchikan Ranger District

(USDA Forest Service) personnel for logistical support

and use of Margarita Bay camp facilities; Rachel

Baker, Derek Busch, Kim Frangos, Adam Herron,

John Hudson, Gary Lamberti, Kristine Martin, Rich-

ard Merritt, Warren Mitchell, Jeff Nichols, David

Sperry, and Brenda Wright for their enthusiastic help,

advice, and support; Timothy Max for helpful statis-

tical advice; Patrick Armitage, Mason Bryant, John

Hudson, Gary Lamberti, Richard Merritt, William

Perry, Deanna Stouder, and Roger Wotton for their

many helpful suggestions on the manuscript and

stimulating discussions about dead fish; Linda Daniel

for important editorial guidance; two anonymous

reviewers for their useful comments. This research

was supported by the Pacific Northwest Research

Station (USDA Forest Service) (Cooperative Research

Agreement: PNW 96-3012-2-CA) and by a grant from

the USDA-CSREES National Research Initiative Com-

petitive Grants Program (99-35101-8592).

References

Armitage P.D., Cranston P.S. & Pinder L.C.V. (Eds)

(1995) The Chironomidae: Biology and Ecology of Non-

Biting Midges. Chapman & Hall, New York.

Ashley K.I. & Slaney P.A. (1997) Accelerating recovery of

stream and pond productivity by low-level nutrient

enrichment. In: Fish Habitat Rehabilitation Procedures

(Eds P.A. Slaney & D. Zaldokas), pp. 239–262. Water-

shed Restoration Technical Circ., no. 9, British Colum-

bia Ministry of Environment, Lands and Parks and

Ministry of Forests, Vancouver.

Baker T.T., Wertheimer A.C., Burkett R.D., Dunlap R.,

Eggers D.M., Fritts E.I., Gharrett A.J., Holmes R.A. &

Wilmot R.L. (1996) Status of Pacific salmon and

steelhead escapments in southeastern Alaska. Fisheries,

21, 6–18.

Bilby R.E., Fransen B.R. & Bisson P.A. (1996) Incorpor-

ation of nitrogen and carbon from spawning coho

salmon into the trophic system of small streams:

evidence from stable isotopes. Canadian Journal of

Fisheries and Aquatic Science, 53, 164–173.

Britton J.C. & Morton B. (1994) Marine carrion and

scavengers. Oceanography and Marine Biology: An

Annual Review, 32, 369–434.

Bryant M.D., Frenette B.J. & McCurdy S.J. (1999) Colon-

ization of a watershed by anadromous salmonids

following the installation of a fish ladder in Margaret

Creek, Southeast Alaska. North American Journal of

Fisheries Management, 19, 1129–1136.

Cederholm C.J., Houston D.B., Cole D.L. & Scarlett W.J.

(1989) Fate of coho salmon (Oncorhynchus kisutch)

carcasses in spawning streams. Canadian Journal of

Fisheries Aquatic Science, 46, 1347–1355.

Cederholm C.J., Kunze M.D., Murota T. & Sibatani A.

(1999) Pacific salmon carcasses: essential contributions

of nutrients and energy for aquatic and terrestrial

ecosystems. Fisheries, 24, 6–15.

Cummins K.W. (1973) Trophic relations of aquatic

insects. Annual Review of Entomology, 18, 183–206.

Elliott S.T. & Bartoo R. (1981) Relation of larval Polyped-

ilum (Diptera: Chironomidae) to pink salmon eggs and

alevins in an Alaskan stream. Progressive Fish Culturist,

43, 220–221.

Ellis R.J. (1970) Alloperla stonefly nymphs: predators or

scavengers on salmon eggs and alevins? Transactions of

the American Fisheries Society, 99, 677–683.

Hanski I. (1987) Nutritional ecology of dung- and carrion-

feeding insects. In: Nutritional Ecology of Insects, Mites,

Spiders and Related Invertebrates (Eds F. Slansky & J.G.

Rodriguez), pp. 837–884. Wiley, New York.

Haskell N.H., McShaffrey D.G., Hawley D.A., Williams

R.E. & Pless J.E. (1989) Use of aquatic insects in

determining submersion interval. Journal of Forensic

Science, 34, 622–632.

Keiper J.B., Chapman E.G. & Foote B.A. (1997) Midge

larvae (Diptera: Chironomidae) as indicators of post-

mortem submersion interval of carcasses in a wood-

land stream: a preliminary report. Journal of Forensic

Science, 42, 1074–1079.

Kline T.C., Goering J.J. & Piorkowski R.J. (1997) The

effect of salmon carcasses on Alaskan freshwaters. In:

Freshwaters of Alaska: Ecological Syntheses (Eds A.M.

Milner & M.W. Oswood), pp. 179–204. Springer-Ver-

lag, New York.

Merritt R.W. & Cummins K.W. (Eds) (1996) An Introduc-

tion to the Aquatic Insects of North America. Kendall/

Hunt, Dubuque.

Merritt R.W. & Wallace J.R. (2000) The role of aquatic

insects in forensic investigations. In: Forensic Entomol-

ogy: The Utility of Arthropods in Legal Investigations (Eds

J.H. Byrd & J.L. Castner), pp. 177–222. CRC Press, Boca

Raton.

Minakawa N. & Gara R.I. (1999) Ecological effects of a

chum salmon (Oncorhynchus keta) spawning run in a

small stream of the Pacific Northwest. Journal of

Freshwater Ecology, 14, 327–335.

Minshall G.W., Hitchcock E. & Barnes J.R. (1991)

Decomposition of rainbow trout (Oncorhynchus mykiss)

carcasses in a forest stream ecosystem inhabited only

272 D.T. Chaloner et al.

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273

by non-anadromous fish populations. Canadian Journal

of Fisheries and Aquatic Science, 48, 191–195.

Parmenter R.R. & Lamarra V.A. (1991) Nutrient cycling

in a fresh-water marsh: the decomposition of fish and

waterfowl carrion. Limnology and Oceanography, 36,

976–987.

Payne J.A. & King E.W. (1972) Insect succession and

decomposition of pig carcasses in water. Journal of the

Georgia Entomological Society, 7, 153–162.

Polis G.A., Anderson W.B. & Holt R.D. (1997) Toward an

integration of landscape and food web ecology: the

dynamics of spatially subsidized food webs. Annual

Review of Ecology and Systematics, 28, 289–316.

SAS Institute Inc. (1989) SAS/STAT User’s Guide, Version

6, 4th edn., Vol. 2. SAS Institute, Cary.

Schuldt J.A. & Hershey A.E. (1995) Effect of salmon

carcass decomposition on Lake Superior tributary

streams. Journal of the North American Benthological

Society, 14, 259–268.

Smock L.A. (1980) Relationships between body size and

biomass of aquatic insects. Freshwater Biology, 10, 375–

383.

Stevenson R.J., Bothwell M.L. & Lowe R.L. (Eds) (1996)

Algal Ecology. Academic Press, San Diego, CA.

Vance G.M., Van Dyk J.K. & Rowley W.A. (1995) A

device for sampling aquatic insects associated with

carrion in water. Journal of Forensic Science, 40, 479–482.

Webster J.R. & Benfield E.F. (1986) Vascular plant

breakdown in freshwater ecosystems. Annual Review

of Ecology and Systematics, 17, 567–94.

Willson M.F., Gende S.M. & Marston B. (1998) Fishes and

the forest: expanding perspectives on fish–wildlife

interactions. Bioscience, 48, 455–462.

Wipfli M.S., Hudson J.P. & Caouette J.P. (1998) Influence

of salmon carcasses on stream productivity: response

of biofilm and benthic macroinvertebrates in south-

eastern Alaska, USA. Canadian Journal of Fisheries and

Aquatic Science, 55, 1503–1511.

Wipfli M.S., Hudson J.P., Chaloner D.T. & Caouette J.P.

(1999) The influence of salmon spawner densities on

stream productivity in Southeast Alaska. Canadian Jour-

nal of Fisheries and Aquatic Science, 56, 1600–1611.

(Manuscript accepted 28 May 2001)

Mass loss and macroinvertebrate colonisation of salmon carcasses 273

Ó 2002 Blackwell Science Ltd, Freshwater Biology, 47, 263–273