Limnol. Oceanogr.

, 47(3), 2002, 742–752

q 2002, by the American Society of Limnology and Oceanography, Inc.

Relative importance of carbon sources for macroinvertebrates in a

Rocky Mountain stream
James H. McCutchan, Jr.,1 and William M. Lewis, Jr.
Center for Limnology, Cooperative Institute for Research in Environmental Sciences, University of Colorado, Boulder,
Colorado 80309-0216

Abstract
Estimates of carbon sources, as determined from ratios of stable isotopes, were used in conjunction with estimates
of secondary production to determine the relative contribution of algal carbon to macroinvertebrate production
across a gradient of elevation in a Rocky Mountain stream (North St. Vrain Creek, Colorado). The relative contri-
bution of algal carbon to macroinvertebrate production was then compared to the relative availability of algal carbon.
Although algal production accounted for less than 2–40% of the combined sources of organic matter to North St.
Vrain Creek, the relative contribution of algal carbon to annual macroinvertebrate production ranged from approx-
imately 40% at a subalpine site to nearly 80% at a more open site in the foothills. Thus, the proportional contribution
of algal carbon to consumer production greatly exceeded the relative availability of algal carbon in North St. Vrain
Creek. Despite the disproportionate importance of algal carbon to consumers, most macroinvertebrates in North St.
Vrain Creek used some vascular plant carbon.

The abundance of terrestrial plant detritus in most shaded are nutritionally important, even when present in small
headwater streams has led ecologists to the conclusion that amounts relative to vascular plant carbon (Koslucher and
vascular plant detritus generally supports the growth of pri- Minshall 1973; Minshall 1978; Mayer and Likens 1987).
mary consumers in small, shaded streams. Although terres- Analysis of gut contents has been the primary method of
trial plant detritus often is of lower nutritional value than tracing pathways of organic matter across trophic levels in
other, less abundant food sources (e.g., algae), microbial col- streams and has been used successfully to study the com-
onization and processing increase its nutritional value (Cum- plexity of trophic interactions among consumers (e.g., Benke
mins and Klug 1979). Algae are known to support most of et al. 2001), but this method is not ideally suited for esti-
the consumer production in some streams (e.g., desert mation of the relative contribution of potential carbon sourc-
streams; Minshall 1978), but it is generally accepted that es to the growth of consumers. Because gut contents are not
consumers in small, canopied streams depend largely on ter- fully assimilated, analyses of gut contents can lead to over-
restrial carbon. estimation of the importance of some food items (Rosenfeld
Analyses of gut contents from macroinvertebrates often and Mackay 1987). Also, identification of gut contents is not
are consistent with the conclusion that vascular plant detritus always possible, especially for detritivores.
is the main source of nutrition for primary consumers in Stable isotope ratios of carbon and nitrogen change pre-
shaded streams (e.g., Minshall 1967; Anderson and Sedell dictably between diet and consumer (Peterson and Fry 1987;
1979; Hall et al. 2001). Also, recent work has shown that McCutchan 1999; McCutchan et al. unpubl. data) and have
amorphous detritus, which usually is assumed to be derived been used to study the use of potential food sources by con-
from terrestrial plant material, can be an important compo- sumers in several streams. Analysis of stable isotopes
nent of gut contents of many stream consumers (Wallace et showed that algae were an important food source for con-
al. 1987; Benke and Jacobi 1994; Hall et al. 2001). Some sumers in the Kuparuk River, Alaska (Peterson et al. 1997)
analyses of gut contents, however, have suggested that algae and in New Zealand streams (Rounick et al. 1982). In con-
trast, another study of stable isotopes showed that neither
1
Present address: Institute of Ecosystem Studies, Box AB, Mill- algae nor aquatic mosses were an important carbon source
brook, New York 12545-0129. for consumers in the lower reaches of the Koroc River, Que-
bec (Bunn et al. 1989). For the Montmorency River, Quebec,
Acknowledgments
The authors gratefully acknowledge the inspiration and help of Junger and Planas (1994) found that the contribution of algal
the late Robert C. Averett, who was involved in the early devel- carbon to consumers was greater at a fourth-order site than
opment of this study. We thank J. Saunders, T. Avant, W. Cross, M. at a second-order site. Winterbourn et al. (1986) showed that
May, S. Horn, D. O’Meara, C. Borth, D. Lipson, C. Kendall, S. consumers at shaded sites in two British streams were iso-
Silva, D. McKnight, R. Guererri, and many others for their valuable topically similar to terrestrial detritus, but many consumers
help. We also thank K. H. Macneale, P. J. Mulholland, and two at unshaded sites were isotopically similar to algae.
anonymous reviewers for their comments. This project was sup- No published studies have partitioned production of an
ported by the National Science Foundation (DEB 9318205), the
entire macroinvertebrate community according to its ultimate
U.S. Geological Survey, the National Park Service, and by fellow-
ships from the University of Colorado (Graduate School and the sources of organic matter. The goal of the work reported here
Cooperative Institute for Research in Environmental Sciences). A is to determine, from stable isotope ratios and simultaneous
grant from the Andrew W. Mellon Foundation to Gene E. Likens estimates of consumer production, the relative contribution
provided financial support to J.M. during the writing of this paper. of algal carbon to the production of macroinvertebrates in a
742
 Carbon sources for stream invertebrates 743

North St. Vrain Creek has only a small human population

and shows little human disturbance. Peak discharge typically
occurs in June, following snowmelt, and discharge is lowest
during winter. Discharge during snowmelt was much higher
in 1995–1996 than in 1994–1995, particularly at the foothills
site. Across all sites, dissolved oxygen remained near satu-
ration and pH was circumneutral. The mean concentration
of total dissolved phosphorus was similar across the highest
three sites (;0.2 mM) and was slightly higher at the foothills
site (0.3 mM; McCutchan 1999). The mean concentration of
dissolved inorganic nitrogen was less than 10 mM across all
sites (McCutchan 1999).
Samples and field measurements were taken at each site
as necessary to estimate (1) aquatic primary production, (2)
litterfall, (3) macroinvertebrate production, and (4) stable
isotope ratios for macroinvertebrates and their ultimate food
sources. Sampling for estimates of aquatic primary produc-
tion and macroinvertebrate production was monthly during
summer and fall and was less frequent during spring and
winter.
Aquatic primary production was estimated by the open-
Fig. 1. Locations of sampling sites on North St. Vrain Creek (a channel oxygen method (Odum 1956; Fellows et al. 2001;
5 alpine site; s 5 subalpine site; m 5 montane site; f 5 foothills
site).
McCutchan et al. 2002). However, temporal resolution for
these estimates was insufficient to provide reasonable annual
estimates for all sites. For this reason, annual net primary
Colorado mountain stream. The relative contribution of algal production (NPP) was estimated from temperature and ben-
carbon to macroinvertebrate production is then compared to thic chlorophyll a (Chl a) by use of the equation of Morin
the relative availability of algal carbon; the purpose of the et al. (1999), based on the assumption that NPP is half of
comparison is to test the null hypothesis that primary food gross primary production. The dry mass of algae was as-
resources in streams are used by consumers in proportion to sumed to be 50% carbon, and primary production under
their availability. snow cover was assumed to be zero. For each site, benthic
Chl a was estimated for each day of the year by linear in-
Methods terpolation from measured concentrations; on each sampling
date, scrapings were taken from three or four stones col-
This study was conducted on North St. Vrain Creek, lected from each of seven to ten transects (McCutchan
which is located in the Front Range of the Colorado Rockies; 1999). A Monte Carlo approach, similar to that described in
the four sampling locations vary in elevation and other fea- McCutchan et al. (1998), was used to estimate uncertainty
tures (Fig. 1; Table 1). Woody vegetation at the alpine site in annual estimates of NPP; estimates of uncertainty incor-
is sparse (Engelmann spruce, Picea engelmannii; willows, porate uncertainty in the coefficients for the regression equa-
Salix sp.). The subalpine site, which is the most heavily can- tion (Morin et al. 1999) and variation in chlorophyll con-
opied of the four, is dominated by subalpine fir (Abies la- centration on each sampling date.
siocarpa) and some Engelmann spruce. The montane site has Direct litterfall (units of litterfall are gDM m22 yr21) at
blue spruce (Picea pungens), ponderosa pine (Pinus pon- canopied sites was predicted from watershed area (units are
derosa), and Douglas fir (Pseudotsuga menzesii). The foot- hectares [ha]) according to a relationship derived from pub-
hills site is dominated by blue spruce and alder (Alnus ten- lished estimates of litterfall for temperate, canopied streams
uifolia) along the stream edge; ponderosa pine and Douglas (Benfield 1997), as follows:
fir are found beyond the stream margin. The watershed for

Table 1. Physical characteristics of the study sites. Discharge data are from USGS gage data or were estimated from watershed area
and gage data for nearby streams.

Alpine Subalpine Montane Foothills

Mean elevation (m) 3,405 3,152 2,426 2,231
Watershed area (ha) 166 854 9,060 20,500
Mean annual discharge (m3 s21) 0.05 0.15 1.50 2.38
Mean annual temperature (8C) 2.6 3.0 4.1 4.8
Mean channel width (m) 2.8 4.3 10.6 11.0
Mean canopy cover (%) 1 39 28 29
Stream order 1 2 3 4
744 McCutchan and Lewis

log litterfall 5 2.98 6 0.125 2 0.0542

6 0.00876(log watershed area)2 ;
R 2 5 0.94 (1)
where units of watershed area are ha. Parameters for Eq. 1
are significantly different from zero (p , 0.001). Estimates
of litterfall for uncanopied, temperate streams (Benfield
1997) were used to predict litterfall at the alpine site, as
follows:
log litterfall 5 1.48 6 0.92 (2)
Quantitative samples for estimation of macroinvertebrate
density and size frequency were collected at random X, Y
coordinates with a modified Surber sampler (225-mm mesh,
0.14 m 2 or 0.25 m 2 area; typically seven to ten per date) and
were preserved in 70% ethanol. Surber samples were divided Fig. 2. Graphical representation of growth (g) and turnover (t)
for a consumer. The parameter m0 is the initial mass of the consumer
into coarse (850-mm mesh) and fine (225-mm mesh) frac-
and mt (m0 1 g) is the mass at time t. The sum of growth and
tions, and a sample splitter similar to the one described by turnover equals assimilation over the interval.
Waters (1969) was used to subsample the fine fraction. Ma-
croinvertebrates were picked from samples under magnifi-
cation and sorted to the lowest possible taxonomic division. epilithon by centrifugation through colloidal silica and were
For each sampling date, body length or width of preserved collected on Whatman QM/A filters (Hamilton and Lewis
individuals was measured with an optical micrometer. Indi- 1992; Hamilton et al. 2001). All samples for isotope analysis
vidual dry mass was estimated by use of exponential equa- were lyophilized and stored in a desiccator until analysis.
tions calibrated empirically (McCutchan 1999) or, in a few Stable isotope ratios were measured on a Micromass Optima
cases, taken from the literature (Winberg 1971; Smock isotope-ratio mass spectrometer operated in conjunction with
1980). Organisms used in the calibration of size-to-mass re- an elemental analyzer. Isotope ratios are reported here in
lationships were held overnight in filtered water for clear- standard d notation (d13C or d15N; units are ‰; Peterson and
ance of gut contents, measured with an optical micrometer, Fry 1987).
lyophilized, and weighed to the nearest mg. The mean individual mass of a consumer taxon at time t
For each site, macroinvertebrate production (MP) for each (m t) is a function of initial mass (m0), net growth over the
taxonomic group was estimated over the intervals between interval (g 5 m t 2 m0), and the mean loss of tissue through
sampling dates. The instantaneous exponential form of the turnover (t; Fig. 2). The carbon isotope ratio of a consumer
increment-summation method (Gillespie and Benke 1979) at time t (d13C t) can be approximated from the initial isotope
was used to estimate production of taxa that could be iden- ratio (d13C0), m0, g, t, the isotope ratio of the diet (i.e., that
tified to species and followed over time as distinct cohorts. of ingested food, d13Cdiet), and the isotopic shift between diet
For taxa that could not be identified to species or could not and consumer (Dd13C) as follows (McCutchan 1999; Mc-
be followed over time as distinct cohorts (e.g., Chironomi- Cutchan and Lewis in press):
dae), production was estimated by the instantaneous growth
method (Gillespie and Benke 1979). Growth rates of non- d 13 C 0 m0 1 (g 1 t)(d 13 C Diet 1 Dd 13 C)
d 13 C t ø (4)
cohort taxa were estimated with a multiple regression equa- m0 1 (g 1 t)
tion developed from growth rates of species in North St.
Over each interval of time, m t was estimated from m0 and g
Vrain Creek that could be followed as cohorts:
as follows:
log g 5 22.2 6 0.28 1 0.061 6 0.0034T¯ 2 0.12
mt 5 m0e gt (5)
6 0.017 log m; R 5 0.42
2
(3)
For species with identifiable cohorts, g was estimated from
where g is the instantaneous rate of growth, T̄ is average measurements of dry mass of consumers for each sampling
temperature in 8C, and m is mass in mg. All coefficients of date. For other taxa, g was estimated from temperature and
Eq. 3 are significantly different from zero (p , 0.0001). dry mass (Eq. 3). For this study, t was set equal to g for
Because no significant differences in growth rate were found aquatic insects (i.e., net production efficiency 5 0.5) and
among orders or families (Tukey-Kramer Honestly Signifi- three times g for other macroinvertebrates (Humphreys
cant Difference test), Eq. 3 was used for all taxa. 1979; Oertli 1993).
Plant material (living and dead) was collected from the A model based on Eqs. 4 and 5 was used to estimate the
riparian zone at each site. When sufficiently abundant, algal proportional contribution of algal carbon (kalgal carbon) to the
filaments or bryophytes were collected from the stream with growth of each consumer taxon. The model predicts changes
forceps; samples were rinsed with water to remove fine par- over time in isotope ratios of consumers from (1) temporal
ticulate material and were picked clean under magnification. changes in isotope ratios of food sources, (2) estimates of
Epilithic material was scrubbed from the upper surfaces of growth and tissue turnover, and (3) the shift in isotope ratio
cobbles and small boulders. Algal cells were isolated from between diet and consumer. For each interval between sam-
 Carbon sources for stream invertebrates 745

Table 2. Annual net primary production (NPP; mean 6 SD, estimated from benthic chl a and temperature), litterfall (mean 6 SD,
predicted using relationships developed from Benfield 1997), and macroinvertebrate production (MP) at each sampling site; units of annual
production and litterfall are gDM m22 yr21.

Alpine Subalpine Montane Foothills

1994–1995 1995–1996 1994–1995 1995–1996 1994–1995 1995–1996 1994–1995 1995–1996
NPP 23614.3 5.262.8 24614.1 4.662.6 37620.7 24613.0 82648.9 36619.9
Litterfall 53683.4 3456119 148666.6 105652.8
MP 5.78 4.73 4.71 3.33 8.11 6.34 12.1 10.1

pling dates, changes in isotope ratios of consumers and food preliminary estimate of trophic position was calculated from
sources (i.e., algae and terrestrial plants) and the mean in- Eqs. 6 and 8; kalgal nitrogen was set equal to the preliminary
dividual mass of a given taxonomic group were predicted estimate of kalgal carbon; trophic shift for nitrogen was set to
over short (daily) intervals of time. It was assumed that iso- 1.6‰ for the first trophic transfer and to 2.6‰ for subse-
tope ratios of sources changed linearly over time between quent transfers (McCutchan et al. unpubl. data). The estimate
adjacent sampling dates. kalgal carbon was estimated by varying of trophic position (l) then was used to recalculate kalgal carbon.
the estimate of the fraction of algae in assimilated organic Based on the second estimate of kalgal carbon, trophic level also
matter to achieve the best fit between observed and predicted was recalculated. Recalculations were necessary because
isotope ratios for each consumer taxon. kalgal carbon affects estimates of trophic position when d15N dif-
If kA nitrogen, the relative contribution of nitrogen from one fers between sources, but more than two iterations had little
food source to the growth of a consumer, is known, trophic effect on final estimates.
position can be estimated from d15N of a consumer and of For each taxonomic group, the amount of secondary pro-
its potential food sources. It is not possible, however, to duction supported by algal carbon was estimated as the prod-
solve simultaneously for kA nitrogen and trophic position. There- uct of kalgal carbon and production for each interval between
fore, estimations of trophic position were based on the as- sampling dates. Production supported by terrestrial carbon
sumption that kA nitrogen 5 kA carbon. d15Ndiet, the d15N for assim- was estimated by difference. For each year of the study,
ilation of sources A (d15NA) and B (d15NB), was estimated by macroinvertebrate production supported by algal carbon was
a simple two-source mixing model: summed across taxa for each site (to estimate the contribu-
tion of algal carbon to production by the entire community),
d15Ndiet 5 d15NB 2 kA nitrogen(d15NB 2 d15NA) (6)
as was macroinvertebrate production supported by terrestrial
If the trophic shift between diet and consumer differs be- carbon.
tween low-protein and high-protein diets (McCutchan 1999;
Vander Zanden and Rasmussen 2001; McCutchan et al. un- Results
publ. data), d15N for a consumer differs from that of the diet
as follows:
Production and litterfall—Estimates of annual NPP
d15Nconsumer 5 d15Ndiet 1 Dd15Nlow 1 (l 2 2)Dd15Nhigh (7) ranged from ,10 gDM m22 yr21 at the alpine and subalpine
sites in 1995–1996 to over 80 gDM m22 yr21 at the foothills
where Dd15Nlow is the trophic shift associated with primary
site in 1994–1995 (Table 2). In each year of the study, es-
consumption (i.e., diets low in protein), Dd15Nhigh is the tro-
timates of annual NPP were similar for the alpine and sub-
phic shift associated with the assimilation of animal material,
alpine sites and increased with watershed area below the
and l is trophic position; l 5 2 for a strict herbivores and,
subalpine site. For a given sampling location, NPP was 2–5
for trophic positions above 2, noninteger values of l reflect
times as high in the first year of the study as in the second.
feeding at more than one trophic level (as in Vander Zanden
Estimated annual litterfall was about 100–350 gDM m22 yr21
et al. 1997). Trophic position can be estimated by rearrange-
at the canopied sites and was about 50 gDM m22 yr21 at the
ment of Eq. 7, as follows:
alpine site (Table 2). Predicted litterfall for the subalpine site
(345 gDM m22 yr21) was similar to measured rates of litter-
1 2
d 15 N consumer 2 d 15 N diet 2 Dd 15 N low
l5 12 (8) fall for a subalpine site near the University of Colorado
Dd 15 N high Mountain Research Station (370 gDM m22 yr21; L. Scott-
In the application of Eqs. 6 and 8 to the present study, mean Denton unpubl. data).
isotope ratios of sources were used to estimate kalgal nitrogen Averaged across the two years, estimates of total macroin-
from kalgal carbon where isotope ratios of sources did not vary vertebrate production ranged from just over 4 gDM m22 yr21
significantly over the year. Where isotope ratios of sources at the subalpine site to over 10 gDM m22 yr21 at the foothills
did vary seasonally, they were interpolated between dates of site and was higher in 1994–1995 than in 1995–1996 (paired
measurement. t-test, p , 0.01; Table 2). The mean estimate of MP for the
Preliminary estimates of kalgal carbon were based on the as- foothills site was higher than the mean for the alpine or
sumption that the consumer was one trophic level above pri- subalpine sites (Tukey-Kramer HSD, p , 0.05) but did not
mary producers and that the trophic shift for carbon was differ significantly from MP at the montane site. Across the
0.5‰ (McCutchan et al. unpubl. data). For each taxon, a four sampling locations, macroinvertebrate production av-
746 McCutchan and Lewis

Table 3. Secondary production for dominant macroinvertebrate taxa at each site (% of total; mean of both years 6 SE). Taxa listed
account for .80% of total production at each site.

Taxonomic group Alpine Subalpine Montane Foothills

Coleoptera: Elmidae ,0.1 0.760.49 7.360.24 4.860.92
Diptera: Chironomidae 26.263.87 31.461.27 12.460.81 9.661.55
Simuliidae 2.061.11 15.762.53 3.560.54 2.662.21
Ephemeroptera: Baetis sp. 2.861.81 9.361.53 11.461.06 10.563.1
Cinygmula sp. 10.665.73 2.960.44 1.760.13 0.460.18
Drunella sp. 0 3.660.41 7.162.49 10.160.92
Epeorus sp. 15.768.46 0.860.05 3.360.24 2.660.73
Ephemerella sp. 0 0 4.660.53 0.960.32
Rithrogena sp. 0 1.060.25 6.461.18 5.361.97
Plecoptera: Chloroperlidae 3.160.43 2.460.96 4.160.86 4.362.95
Zapada sp. 3.160.43 10.664.83 1.160.24 0.260.10
Trichoptera: Arctopsyche grandis 0 0 2.260.86 18.3610.8
Rhyacophila sp. 22.963.46 7.460.05 13.265.26 4.060.52
Oligochaeta 6.560.69 3.461.15 4.860.30 3.760.15

eraged 37% (69.9% SE) of aquatic NPP but only 5% at the foothills site. Shredders accounted for a higher fraction
(61.0% SE) of the sum of NPP and litterfall (Table 2). of production at the subalpine site (14% of total) than at
At each site, most of the annual macroinvertebrate produc- other sites (#6% of total). Predator production was similar
tion could be attributed to a small number of taxa (Table 3; (18–21% of total) across the alpine, subalpine, and montane
Web Appendix 1 at http://www.aslo.org/lo/toc/volp47/issuep3/ sites but was somewhat higher (34% of total) at the foothills
0742a1.pdf). Chironomidae accounted for about 30% of total site.
production at the alpine and subalpine sites and about 10% at
the montane and foothills sites. Baetis accounted for about 10% Stable isotope ratios—At each site, the mean d13C of algae
of total production at the canopied sites but was less important was significantly higher than the mean d13C of terrestrial
at the alpine site. Heptageniidae and Ephemerellidae were im- plants (t-test, p , 0.001), as was mean d15N (t-test, p , 0.05)
portant consumers at all sites except the subalpine site; Cin- except for the alpine site (Fig. 3). Isotope ratios of mosses
ygmula and Epeorus accounted for nearly 30% of production and liverworts were similar to those of terrestrial plants.
at the alpine site and Rithrogena and Drunella were important There were no significant seasonal trends in d13C of terres-
at the montane and foothills sites. Other important taxa includ- trial plants, but d13C of algae at the montane and foothills
ed Simuliidae (16% of total) and Zapada (11% of total) at the sites and d15N of algae at the montane site changed with
subalpine site and Arctopsyche grandis (18% of total) at the season. The trend in d13C of algae was similar for the mon-
foothills site. In terms of production, Rhyacophilia was an im- tane and foothills sites; d13C of algae increased abruptly dur-
portant predator (7–23% of total) at the upper three sites, but ing snowmelt, reached a maximum in late summer, and de-
not at the foothills site (4% of total). clined over the fall (Fig. 4). At the montane site, d15N of
Together, scrapers and collector-gatherers accounted for algae was highest just before snowmelt and decreased over
43–77% of macroinvertebrate production at each site (Table the summer.
4). Filtering collectors accounted for 11% of macroinverte- Isotope ratios of consumers are shown for each site in Fig.
brate production at the subalpine site and 16% of production 3. At each of the lower elevation sites, some consumers had

Table 4. Percentage of total macroinvertebrate production (MP) and mean (6SE) proportional contribution of algal carbon to production
(kalgal carbon) by functional feeding groups at each site. Assignments of functional feeding groups are from Merritt and Cummins (1984) and
Ward and Kondratieff (1992); for taxa belonging to multiple feeding groups, production was divided equally among groups.

Feeding group Alpine Subalpine Montane Foothills

Scraper % of total MP 33 26 32 23
kalgal carbon 0.6460.11 0.3060.12 0.6860.18 0.7360.27
Collector-gatherer % of total MP 44 31 37 20
kalgal carbon 0.5960.13 0.3460.22 0.6060.21 0.6960.27
Filtering collector % of total MP 1 11 3 16
kalgal carbon 0.1360.07 0.1260.03 0.7560.18 0.7160.24
Shredder % of total MP 2 14 7 6
kalgal carbon 0.4260.08 0.3160.17 0.5260.27 0.5760.29
Predator % of total MP 19 18 21 34
kalgal carbon 0.5160.17 0.3560.18 0.6560.25 0.8160.22
Carbon sources for stream invertebrates 747

Fig. 4. The d13C of algae (mean 6 SE) and terrestrial plant

material as a function of the day of the year at the montane and
foothills sites. The regression equations were used to predict d13C
of algae for days when no algae were collected. Mean daily dis-
charge (solid line; from the U.S. Geological Survey) at the montane
site is shown for reference.

higher d13C than the mean d13C of algae, but few consumers
were more enriched than the most enriched algal samples.
The d15N of consumers typically was 2–5‰ higher than that
of primary producers, but some consumers were enriched in
15
N by more than 10‰ relative to primary food sources (i.e.,
algae and terrestrial plant material). Seasonal variation in
isotope ratios of some taxa (e.g., Chironomidae at the mon-
tane site; Fig. 5) paralleled changes in isotope ratios of algae,
but, for many consumers, isotopic changes lagged behind
changes for algae (e.g., Baetis bicaudatus and Heptageni-
idae). Taxa with the most negative d13C values (i.e., near the
ratio of terrestrial plants) generally showed little seasonal
variation in d13C (e.g., Nemouridae; Fig. 5).
For some taxa, production-weighted means of kalgal carbon
were ,0.15 (e.g., Simuliidae at the alpine and subalpine
sites; Web Appendix 1) and, for shredders, means for esti-
mates of kalgal carbon ranged from 0.31 to 0.57 (Table 4). At

←
Fig. 3. The d13C and d15N of algae (solid diamonds; mean 6
SE), terrestrial plants (open diamonds; mean 6 SE), and individual
consumers (circles) at each sampling location. Means for sources
reflect seasonal changes in isotope ratios; values for individual con-
sumers represent single points in time.
748 McCutchan and Lewis

Fig. 6. Macroinvertebrate production supported by algal and

terrestrial carbon at each site. Areas of bars represent production
over intervals between sampling dates.

each study site, however, algal carbon supported a large pro-

portion of macroinvertebrate production. The proportional
Fig. 5. Seasonal variation in d C of selected taxa (circles; mean
13
contribution of algal carbon to macroinvertebrate production
6 SE) at the montane site. Mean daily discharge (solid line; from
the U.S. Geological Survey) and d13C of algae and terrestrial plants was highest at the foothills and alpine sites, but algae con-
(see Fig. 4; dashed lines) are shown for reference. tributed substantially to consumer production even at the
subalpine and montane sites, where canopy cover was high
(Fig. 6). On an annual basis, the absolute contribution of
terrestrial carbon to macroinvertebrate production (i.e., the
amount of macroinvertebrate production supported by ter-
 Carbon sources for stream invertebrates 749

brate production was closely correlated with aquatic NPP (r

5 0.91). Temperature or algal production may have limited
macroinvertebrate production in St. Vrain Creek, but other
factors probably played a role in the regulation of secondary
production. For example, food quality can differ substan-
tially between algae and terrestrial vascular plant material
(Fuller and Mackay 1981) or among plant species (Cummins
et al. 1973). Also, hydrologic differences (among sites and
between years) may have affected the availability of algal
and terrestrial sources of organic matter in a similar manner
(e.g., high flow conditions may have reduced the availability
of both algal and terrestrial sources of food).
Scrapers and collector-gatherers (mostly Chironomidae,
Baetidae, Ephemerellidae, and Heptageniidae) accounted for
much of the production at the upper three sites (Tables 3 and
4). Because of their high growth rates and high densities,
Fig. 7. Proportional contribution of algal carbon to total ma- Chironomidae and Baetidae often are among the most pro-
croinvertebrate production (kalgal carbon for the community or Kalgal carbon) ductive invertebrate taxa in small streams (Jackson and Fish-
as a function of watershed area in 1994–1995 and 1995–1996. An er 1986; Wallace and Gurtz 1986; Gaines et al. 1992; Benke
index of the relative availability of aquatic primary production (RA, 1998). Shredders contributed little to total production in
the ratio of NPP to the sum of NPP and litterfall; mean 6 SD) is North St. Vrain Creek, even where canopy cover was high.
shown for reference. Predators accounted for a large proportion of production at
all sites. Although relative production by shredders was
highest at the subalpine site and filtering collectors were im-
restrial carbon), averaged across all sites, was 2.5 (60.8 SD) portant at the foothills site, downstream changes in the rel-
gDM m22 yr21 and did not differ significantly among sites ative importance of functional feeding groups were small,
(Tukey-Kramer HSD); the absolute contribution of algal car- contrary to the predictions of the River Continuum Concept
bon to consumer production averaged 3.8 (62.7) gDM m22 (Vannote et al. 1980; Minshall et al. 1985).
yr21 (Fig. 6). Algal carbon supported a similar amount of
production across the higher elevation sites, but algae sup-
ported more production at the foothills site than at other sites Seasonal variation in isotope ratios of sources—Isotope
(Tukey-Kramer HSD, p , 0.05). ratios of algae were stable during summer months but
The relative contribution of algal carbon to annual ma- changed during snowmelt at two sites. When d13C of avail-
croinvertebrate production (kalgal carbon for the community or able food sources changes seasonally, d13C of consumers
Kalgal carbon) ranged from less than 40% at the subalpine site to may not reflect d13C of assimilated organic matter at the time
nearly 80% at the foothills site (Fig. 7). Because inputs of of sampling. At the montane and foothills sites, changes in
dissolved organic matter and lateral inputs of litter often isotope ratios of many consumers dependent on algal carbon
equal or exceed direct litterfall to streams (Webster and Mey- lagged behind changes in isotope ratios of algae (e.g., Baetis
er 1997), RA (the ratio of NPP to the sum of NPP and bicaudatus and Heptageniidae, Fig. 5). The lag was more
litterfall; Fig. 7) provides a minimum estimate of the relative pronounced for slowly growing taxa than for taxa with high
availability of algal carbon. Thus, in North St. Vrain Creek, growth rates (e.g., Chironomidae) and was more pronounced
algal production accounted for less than 2–40% of the or- during winter when growth rates were low.
ganic matter input across all sites, but the contribution of Estimates of the proportional contribution of food sources
algal carbon to consumer production averaged six (62.8 SE) to consumers by use of standard mixing models depend on
times the relative availability of algal carbon. the assumption that consumers are in isotopic equilibrium
with their food sources. Laboratory studies have shown that,
Discussion when consumers are switched to a new diet, isotopic equi-
librium usually is approached only after consumers have
Macroinvertebrate production—Annual macroinvertebrate quadrupled in mass (Fry and Arnold 1982; Herzska and Holt
production in North St. Vrain Creek fell slightly below that 2000). Equations 3 and 5 predict that it would take nearly 5
for most streams of similar watershed area (Benke 1993; months for a consumer of 10 mg to quadruple in mass at
Grubaugh et al. 1997), possibly because of low temperatures 58C. Thus, during winter or at other times when growth rates
relative to other streams (Table 1). Feeding and metabolic of consumers are low relative to the time scale for changes
processes generally are reduced at low temperature (Hauer in isotope ratios of food sources, assumptions implicit in
and Benke 1991). Ice cover interrupts the input of terrestrial standard two-source mixing models may not be met (Mc-
detritus to the stream and also can limit aquatic primary Cutchan and Lewis in press; McCutchan et al. unpubl. data).
production. Macroinvertebrate production was highest The special measures taken in this study (i.e., density gra-
downstream where temperature was highest and was higher dient centrifugation, frequent sampling of sources and con-
in the first year of the study when temperature was higher sumers, and use of Eqs. 4 and 5) made it possible to estimate
and discharge was lower. Across both years, macroinverte- the proportional contribution of algal carbon to macroinver-
750 McCutchan and Lewis

tebrate production over most of the year, even when isotope estimates of production integrated over time than for instan-
ratios of algae were changing. taneous estimates (Morin et al. 1999). Litterfall also was
predicted by relationships developed from published esti-
Contribution of food sources to consumers—Although the mates for a wide range of sites and uncertainty in predicted
extent of omnivory among stream consumers is widely ap- rates of litterfall was relatively large (Table 2). Also, some
preciated, it has been hypothesized that algae are the dom- of the terrestrial plant material that enters a stream is retained
inant source of nutrition for scrapers whereas shredders and only briefly before being exported downstream, where it col-
collectors are supported largely by carbon derived from ter- lects in depositional areas. Thus, the rate at which terrestrial
restrial plants, and predators depend indirectly on both algal organic matter is supplied to consumers varies spatially and
and terrestrial sources of carbon (Cummins and Klug 1979; temporally and may differ from average rates of input. How-
Merritt and Cummins 1984). Stable isotope data indicate that
ever, estimated litterfall, which is a conservative estimate of
few consumers in North St. Vrain Creek relied solely on
algal or terrestrial carbon (Web Appendix 1). At the subal- the input of terrestrial carbon to a stream, far exceeded
pine site, Agraylea (Hydroptilidae), a specialist on filamen- aquatic NPP at the subalpine and montane sites in both years
tous algae (Merritt and Cummins 1984), was the only taxon and at the foothills site in the second year of the study (Table
to obtain nearly all of its carbon from algae. Bibiocephala 2).
grandis (Blephariceridae), a scraper common at the montane During this study, algal production was not the dominant
site, also obtained nearly all of its carbon from algae. Nem- source of organic matter available to consumers, but algal
ouridae (mostly shredders) and Simuliidae (filtering collec- carbon supported most of the consumer production in North
tors) at the subalpine site were largely dependent on terres- St. Vrain Creek (Fig. 7). Nonetheless, algal production was
trial carbon. Most consumers, however, were omnivorous adequate to support even more than the observed macroin-
with respect to carbon sources, as shown by Mihuc (1997) vertebrate production. For the second trophic level, the max-
for other streams. imum amount of production that can be supported by algal
Despite the absence of tree canopy at the alpine site, ter- carbon is equal to the product of aquatic NPP (approximately
restrial plants supported the growth of many consumers 5–80 gDM m22 yr21 for North St. Vrain Creek) and maxi-
there. Terrestrial carbon that enters the stream as detritus or mum net production efficiency for primary consumers (about
as dissolved organic matter probably supports the growth of 39%; Humphreys 1979). The maximum possible production
consumers at times when algae are not able to grow because for each higher trophic level is equal to the product of pro-
of shading from snow cover (Wetzel 1995). Even though duction by the previous level and net production efficiency
there was no tree canopy at the alpine site, some leaf litter
for carnivores (about 56%; Humphreys 1979). The maxi-
from riparian herbs and willows reached the stream. Esti-
mum secondary production that could have been supported
mates of direct litterfall were higher at canopied sites than
at the alpine site, but aquatic NPP at the alpine site was low by algae was 61% of NPP if there were two trophic levels
(Table 2) and the nutritional quality of plant litter there may above primary producers and 73% of NPP if there were three
have been higher than at other sites. Terrestrial invertebrates, trophic levels above primary producers (4–17 gDM m22 yr21
which derive their support from vascular plants, have high at the alpine site; 26–60 gDM m22 yr21 at the foothills site).
nutritional value and may have been an important source of The actual amount of macroinvertebrate production sup-
organic matter for some consumers. Inputs of terrestrial in- ported by algal carbon was much lower (;1 gDM m22 yr21
vertebrates to streams vary seasonally but can be higher than at the subalpine site; ;8 gDM m22 yr21 at the foothills site).
the lowest estimates of aquatic NPP reported here (Wipfli Thus, the contribution of algal carbon to macroinvertebrate
1997; Shepard and McCutchan unpubl. data). Terrestrial in- production was well below its maximum potential, even
vertebrates can be an important trophic link between terres- though algal carbon contributed to macroinvertebrate pro-
trial plants and salmonids in some streams (Cada et al. 1994; duction with greater efficiency than did terrestrial carbon.
Wipfli 1997), but their importance as a food source for Vannote et al. (1980) suggested that changes over a stream
macroinvertebrates in St. Vrain Creek is unknown. continuum in the composition of the consumer community
result from changes in the physical environment and in the
Synthesis—Estimates of aquatic NPP and litterfall were relative input of terrestrial organic matter to the stream. This
necessary to test the assumption that food resources contrib- assertion has become an important organizing principle in
ute to macroinvertebrate production in proportion to their
the study of stream ecosystems and provides a basis for un-
availability. For some sites, low rates of aquatic photosyn-
derstanding changes with stream size in community structure
thesis and high reaeration rates led to low confidence in es-
timates of NPP based on the open-channel oxygen method and ecosystem function. It often has been assumed from this
(see McCutchan et al. 1998). For this reason and due to the assertion that the ultimate sources of organic matter in
remote locations of field sites, annual estimates of NPP were streams contribute to the growth of consumers in proportion
not possible with the open-channel method; instead, annual to their availability. The assumption that sources of organic
NPP was estimated from temperature and Chl a, based on matter contribute to consumers in proportion to their avail-
the regression equation of Morin et al. (1999). Although es- ability does not apply, however, to North St. Vrain Creek.
timates of primary production with the regression method Although terrestrial plants are the dominant source of or-
are not as precise as estimates based on direct measurements, ganic matter in this stream, algal carbon supports most of
uncertainty with the regression method is much lower for the macroinvertebrate production.
 Carbon sources for stream invertebrates 751

References position of red drum (Sciaenops occelatus) larvae in response

to dietary shifts: Potential applications to settlement studies.
ANDERSON, N. H., AND J. R. SEDELL. 1979. Detritus processing by Can. J. Fish. Aquat. Sci. 57: 137–147.
macroinvertebrates in stream ecosystems. Annu. Rev. Entomol. HUMPHREYS, W. F. 1979. Production and respiration in animal pop-
24: 351–377. ulations. J. Anim. Ecol. 48: 427–453.
BENFIELD, E. F. 1997. Comparison of litterfall input to streams, p. JACKSON, J. K., AND S. G. FISHER. 1986. Secondary production,
104–108. In J. R. Webster and J. L. Meyer. [eds.], Stream or- emergence, and export of aquatic insects of a Sonoran desert
ganic matter budgets. J. N. Am. Benthol. Soc. 16: 3–161. stream. Ecology 67: 629–638.
BENKE, A. C. 1993. Concepts and patterns of invertebrate produc- JUNGER, M., AND D. PLANAS. 1994. Quantitative use of stable car-
tion in running waters. Verh. Int. Ver. Limnol. 25: 15–38. bon isotope analysis to determine the trophic base of inverte-
. 1998. Production dynamics of riverine chironomids: Ex- brate communities in a boreal forest lotic system. Can. J. Fish.
tremely high biomass turnover rates of primary consumers. Aquat. Sci. 52: 52–61.
Ecology 79: 899–910. KOSLUCHER, D. G., AND G. W. MINSHALL. 1973. Food habits of
, AND D. I. JACOBI. 1994. Production dynamics and resource some benthic invertebrates in a northern cool-desert stream
utilization of snag-dwelling mayflies in a blackwater river. (Deep Creek, Curlew Valley, Idaho-Utah). Trans. Am. Microsc.
Ecology 75: 1219–1232. Soc. 92: 441–452.
, J. B. WALLACE, J. W. HARRISON, AND J. W. KOEBEL. 2001. MAYER, M. S., AND G. E. LIKENS. 1987. The importance of algae
Food web quantification using secondary production analysis: in a shaded headwater stream as food for an abundant caddisfly
Predaceous invertebrates of the snag habitat in a subtropical (Trichoptera). J. N. Am. Benthol. Soc. 6: 262–269.
river. Freshw. Biol. 46: 329–346. MCCUTCHAN, J. H., JR. 1999. Carbon sources for macroinverte-
BUNN, S. E., D. R. BARTON, H. B. N. HYNES, G. POWER, AND M. brates in St. Vrain Creek, Colorado. Ph.D. dissertation, Uni-
A. POPE. 1989. Stable isotope analysis of carbon flow in a versity of Colorado.
tundra river system. Can. J. Fish. Aquat. Sci. 46: 1769–1775. , AND W. M. LEWIS, JR. In press. Seasonal variation in stable
CADA, G. F., J. M. LOAR, AND D. K. COX. 1994. Food and feeding isotope ratios of stream algae. Verh. Internat. Ver. Limnol. 27.
preferences of rainbow and brown trout in southern Appala- , , AND J. F. SAUNDERS, III. 1998. Uncertainty
chian streams. Am. Midl. Nat. 117: 374–385. in the estimation of stream metabolism from open-channel ox-
CUMMINS, K. W., AND M. J. KLUG. 1979. Feeding ecology of stream ygen concentrations. J. N. Am. Benthol. Soc. 17: 155–164.
invertebrates. Annu. Rev. Ecol. Syst. 10: 147–172. , J. F. SAUNDERS, III, W. M. LEWIS, JR., AND M. G. HAYDEN.
, R. C. PETERSEN, F. O. HOWARD, J. C. WUYCHECK, AND V. 2002. Effects of groundwater flux on open-channel estimates
I. HOLT. 1973. The utilization of leaf litter by stream detriti- of stream metabolism. Limnol. Oceanogr. 47: 321–324.
vores. Ecology 54: 336–345. MERRITT, R. W., AND K. W. CUMMINS, EDS. 1984. An introduction
FELLOWS, C. S., H. M. VALETT, AND C. N. DAHM. 2001. Whole- to the aquatic insects of North America. Kendall/Hunt.
stream metabolism in two montane streams: Contribution of MIHUC, T. B. 1997. The functional trophic role of lotic primary
the hyporheic zone. Limnol. Oceanogr. 46: 523–531. consumers: Generalist versus specialist strategies. Freshw.
FRY, B., AND C. ARNOLD. 1982. Rapid 13C/12C turnover during Biol. 37: 455–462.
growth of brown shrimp (Penaeus aztecus). Oecologia 54: MINSHALL, G. W. 1967. Role of allochthonous detritus in the trophic
200–204. structure of a woodland springbrook community. Ecology 48:
FULLER, R. L., AND R. J. MACKAY. 1981. Effects of food quality
139–149.
on the growth of three Hydropsyche species (Trichoptera: Hy-
. 1978. Autotrophy in stream ecosystems. BioScience 28:
dropsychidae). Can. J. Zool. 59: 1133–1140.
767–771.
GAINES, W. L., C. E. CUSHING, AND S. D. SMITH. 1992. Secondary
, K. W. CUMMINS, R. C. PETERSON, C. E. CUSHING, D. A.
production estimates of benthic insects in three cold desert
BURNS, J. R. SEDELL, AND R. L. VANNOTE. 1985. Develop-
streams. Gt. Basin Nat. 52: 11–24.
GILLESPIE, D. M., AND A. C. BENKE. 1979. Methods of calculating ments in stream ecosystem theory. Can. J. Fish. Aquat. Sci.
cohort production from field data—some relationships. Limnol. 42: 1045–1055.
Oceanogr. 24: 171–176. MORIN, A., W. LAMOUREUX, AND J. BUSNARDA. 1999. Empirical
GRUBAUGH, J. W., J. B. WALLACE, AND E. S. HOUSTON. 1997. Pro- models predicting primary productivity from chlorophyll a and
duction of benthic macroinvertebrate communities along a water temperature for stream periphyton and lake and ocean
southern Appalachian river continuum. Freshw. Biol. 37: 581– phytoplankton. J. N. Am. Benthol. Soc. 18: 299–307.
596. ODUM, H. T. 1956. Primary production in flowing waters. Limnol.
HALL, R. O., JR., G. E. LIKENS, AND H. M. MALCOLM. 2001. Tro- Oceanogr. 1: 102–117.
phic basis of invertebrate production in 2 streams at the Hub- OERTLI, B. 1993. Leaf litter processing and energy flow through
bard Brook Experimental Forest. J. N. Am. Benthol. Soc. 20: macroinvertebrates in a woodland pond (Switzerland). Oecol-
432–447. ogia 96: 466–477.
HAMILTON, S. K., AND W. M. LEWIS, JR. 1992. Stable carbon and PETERSON, B. J., M. BAHR, AND G. W. KLING. 1997. A tracer in-
nitrogen isotopes in algae and detritus from the Orinoco River vestigation of nitrogen cycling in a pristine tundra river. Can.
floodplain, Venezuela. Geochim. Cosmochim. Acta 56: 4237– J. Fish. Aquat. Sci. 54: 2361–2367.
4246. , AND B. FRY. 1987. Stable isotopes in ecosystem studies.
, J. L. TANK, D. F. RAIKOW, W. M. WOLLHEIM, B. J. PETER- Annu. Rev. Ecol. Syst. 18: 293–320.
SON, AND J. R. WEBSTER. 2001. Nitrogen uptake and transfor- ROSENFELD, J. S., AND R. J. MACKAY. 1987. Assessing the food
mation in a midwestern U.S. stream: A stable isotope enrich- base of stream ecosystems: Alternatives to the P/R ratio. Oikos
ment study. Biogeochemistry 54: 297–340. 50: 141–147.
HAUER, F. R., AND A. C. BENKE. 1991. Rapid growth of snag-dwell- ROUNICK, J. S., M. J. WINTERBOURN, AND G. L. LYON. 1982. Dif-
ing chironomids in a blackwater river: The influence of tem- ferential utilization of allochthonous and autochthonous inputs
perature and discharge. J. N. Am. Benthol. Soc. 10: 154–164. by aquatic invertebrates in some New Zealand streams: A sta-
HERZKA, S. Z., AND G. J. HOLT. 2000. Changes in isotopic com- ble isotope study. Oikos 39: 191–198.
752 McCutchan and Lewis

SMOCK, L. A. 1980. Relationships between body size and biomass to the mountain stream insects of Colorado. University Press
of aquatic insects. Freshw. Biol. 10: 375–383. of Colorado.
VANDER ZANDEN, M. J., G. CABANA, AND J. B. RASMUSSEN. 1997. WATERS, T. F. 1969. Subsampler for dividing large samples of
Comparing trophic position of freshwater fish calculated using stream invertebrate drift. Limnol. Oceanogr. 14: 813–815.
stable nitrogen isotope ratios (d15N) and literature dietary data. WEBSTER, J. R., AND J. L. MEYER, EDS. 1997. Stream organic matter
Can. J. Fish. Aquat. Sci. 54: 1142–1158. budgets. J. N. Am. Benthol. Soc. 16: 3–161.
, AND J. B. RASMUSSEN. 2001. Variation in d15N and d13C WETZEL, R. G. 1995. Death, detritus, and energy flow in aquatic
ecosystems. Freshw. Biol. 33: 83–89.
trophic fractionation: Implications for aquatic food web stud- WINBERG, G. G. 1971. The linear dimensions and body mass of
ies. Limnol. Oceanogr. 46: 2061–2066. animals. J. Gen. Biol. 32: 714–723.
VANNOTE, R. L., G. W. MINSHALL, K. W. CUMMINS, J. R. SEDELL, WINTERBOURN, M. J., J. S. ROUNICK, AND A. G. HILDREW. 1986.
AND C. E. CUSHING. 1980. The river continuum concept. Can. Patterns of carbon resource utilization by benthic invertebrates
J. Fish. Aquat. Sci. 37: 130–137. in two British river systems: A stable carbon isotope study.
WALLACE, J. B., A. C. BENKE, A. H. LINGLE, AND K. PARSONS. Arch. Hydrobiol 107: 349–361.
1987. Trophic pathways of macroinvertebrate primary consum- WIPFLI, M. S. 1997. Terrestrial invertebrates as salmonid prey and
ers in subtropical blackwater streams. Arch. Hydrobiol. nitrogen sources in streams: Contrasting old-growth and
(Suppl.) 74: 423–451. young-growth riparian forests in southeastern Alaska, U.S.A.
, AND M. E. GURTZ. 1986. Response of Baetis mayflies Can J. Fish. Aquat. Sci. 54: 1259–1269.
(Ephemeroptera) to catchment logging. Am. Midl. Nat. 115: Received: 30 January 2001
25–41. Accepted: 14 January 2002
WARD, J. V., AND B. C. KONDRATIEFF. 1992. An illustrated guide Amended: 1 February 2002