Ecology, 61(1), 1980, pp. 108-118
© 1980 by the Ecological Society of America
TROPHIC BASIS OF PRODUCTION AMONG NET-SPINNING
CADDISFLIES IN A SOUTHERN
APPALACHIAN STREAM1
$
ARTHUR C. BENKE
School of Biology, Georgia Institute of Technology, Atlanta, Georgia 30332 USA
AND
J. BRUCE WALLACE
Department of Entomology, University of Georgia, Athens, Georgia 30602 USA
Abstract. Life histories and annual production were determined for six species of net-spinning
caddisflies in a headwater stream of the Tallulah River in north Georgia, USA. Five species in the
family Hydropsychidae were univoltine, whereas the sixth, a member of the Philopotamidae, had at
least two generations
per year. Combined annual production, as determined by the Hynes method,
was 1.0 g/m2 (ash-free dry mass). Seventy-five percent of the production was concentrated in the two
largest species, Arctopsyche irrorata and Parapsyche cardis. The remaining production from highest
to lowest percent, was contributed by Dolophilodes distinctus, Hydropsyche sparna, Diplectrona
modesta, and Hydropsyche macleodi.
Analysis of gut contents alone indicated that detritus was the most important food source. However, food preference and food-specific ecological efficiencies were utilized to calculate the amount
of production attributable to each major food category. Surprisingly, almost 80% of all caddisfly
production was attributed to animal food, 13% to detritus, and 8% to algae. Actual annual consumption
required to account for this production was 2.28 g/m2 animals, 2.54 g/m2 detritus, and 0.51 g/m2 algae.
We attempt to quantify the role that net-spinning caddisflies play in the "spiralling" of seston in
mountain streams. Our results show that the omnivorous caddisflies are not the major consumer of
detritus and algae, and that they produce more detritus in their feces than they consume, thus appearing to lower the food quality of the seston. Net-spinning caddisfly production in this mountain
stream appears to be limited by the amount of high quality food available in the seston.
Key words: caddisfly; food habits; growth; Hydropsychidae; life history; Philopotamidae; secondary production; spiralling; stream insects.
INTRODUCTION
Net-spinning caddisflies of the family Hydropsychidae, along with blackflies, are often the .most conspicuous groups of filter-feeding insects in streams and
rivers (Hynes 1970). Hydropsychids rely heavily on
organic drift such as particulate detritus, algae, and
small animals for food (e.g., Hynes 1970, Wallace et
al. 1977, Wiggins 1977). Thus, as a group, net-spinning
caddisflies are considered to be omnivores, but their
role in specific energy flow pathways is largely unknown. This is because feeding studies are rarely
quantitative and have not been done in conjunction
with the necessary production (energy flow) analyses,
of vice versa.
In order to understand the role of an animal population in energy flow pathways, it is desirable, if not
essential, to estimate the production of that population. This "secondary" production is defined as the
rate of tissue elaboration over some interval of time,
regardless of whether the production is lost to carnivores, to nonpredatory mortality, or leaves the system, .such as by emergence of aquatic insects. See
1
Manuscript received 25 September 1978; revised 2 March
1979; accepted 13 March 1979.
Waters (1977) for a detailed review and discussion of
aquatic secondary production.
Although hydropsychids are a large and predominant family of caddisflies in lotic systems throughout
much of the world (Wiggins 1977), there is a paucity
of information on their production (Waters 1977). The
few existing studies indicate a wide range of values
from quite low in mountain streams (Cushman et al.
1975, 1977), to very high in enriched streams (Hopkins
1976) and coastal plain rivers (Van Arsdall 1977). More
emphasis needs to be placed on studying the production of this important group of insects if we are to
understand energy flow in many lotic systems.
Several species of hydropsychids generally coexist
at a given locality in lotic ecosystems. Resource partitioning appears to center around preferences for various current velocities, capture net mesh openings
(thus, with differences both in volume of water filtered
per unit time and particle retention capability), and
temporal variations in life cycles (Wallace 1975, Malas
and Wallace 1977, Wallace et al. 1977). Although the
above studies have described food-partitioning mechanisms, they have not quantified the contribution that
the various food resources make to growth and maintenance of these omnivores.
February 1980
PRODUCTION OF NET-SPINNING CADDISFL1ES
The present study deals with the functional role of
six species of net-spinning caddisflies coexisting in the
same locality in a southern Appalachian stream. Five
species are members of the family Hydropsychidae:
Arctopsyche irrorata Ross, Parapsyche cardis Ross,
Hydropsyche macleodi Flint, Hydropsyche sparna
Ross, and Diplectrona modesta Banks. The sixth,
Dolophilodes distinctus (Walker), is in the Philopotamidae. Our purpose is to demonstrate that quantified
feeding analyses of these species, when combined wilh
production estimates, yield a much better understanding of caddisfly trophic dynamics than either kind of
information by itself. Production and food analyses
are integrated with food-specific ecological efficiencies
to (1) assess the portion of caddisfly production attributable to various food types, and (2) estimate the total
amount of each food type consumed.
STUDY SITE
The Tallulah River at the Georgia-North Carolina
state boundaries is a headwater stream of the Savannah River. The .stream at the study site drnins an area
of —16 km" in the southern Appalachian Mountains.
The stream is fourth-order, =4-7 m in width, and has
a gradient of =<34 m/km. Maximum stream depth is
=0.6 m and substrate consists of cobbles ranging from
3 cm in diameter up to large outcrops of granite bedrock. Water is soft with a pH range of =6-6.8 and
conductivity of 8-40 /iS/cm. From July 1975 to July
1976, temperature ranged from 4-17°C with an annual
mean of 11°C.
The drainage basin is located within the confines of
the Nantahala National Forest and the Standing Indian
Wilderness Area. Terrestrial vegetation consists of
mixed oak-hickory forest, including yellow poplar and
several oak and hickory species intermixed with hemlock, Tsuga canadensis (L.). Evergreen shrubs such
as .Rhododendron spp. form dense understories and
shade much of the stream margin. Eastern white pine,
Pinus strobus L., is also found scattered throughout
the basin. Allochthonous inputs from this vegetation
are important and leaf packs are usually apparent
within the stream during autumn and winter.
MATERIALS AND METHODS
In order to estimate production from field data, it is
necessary throughout the time period of interest to
collect quantitative samples from which densities of
size categories within each species can be calculated.
Since we were interested in estimating annual .production, three or four benthic samples were taken at approximately monthly intervals from riffle areas for a
full year (total = 39). On each sampling date, regions
representing at least three different velocities ranging
from 40-165 cm/s were sampled. A 52 cm wide drift
net (220-/nm mesh) was anchored with steel rods and
a 50 x 50 cm (1A m2) area upstream of the net opening
109
was delineated. The rocky substrate within this area
was cleaned in front of the net opening to collect the
fauna. Depth of sampling was variable (up to 10 cm),
depending upon whether bedrock or sand was underlying the inhabited substrate. Contents of the net were
placed in plastic bags and preserved in 7-8% formalin.
A small amount of the dye Phjoxine B was added to
the formalin to facilitate laboratory sorting (Mason and
Yevich 1967). After sorting and species identification
of each larva, head capsule width at the level of the,
eyes was measured with an ocular micrometer to the
nearest 10 fjan. Head width frequency histograms were
constructed to determine larval instar. Head widths,
color patterns, setal shapes, and setal patterns were
all used for specific identification of early instars (I
and II; Mackay 1978).
To estimate production, it is also essential to know
the individual biomass of organisms collected. Therefore, mean individual biomass (ash-free dry mass) was
determined for each instar present by season. Preserved animals were placed in a drying oven at 65°C
Tor 24 li, transferred to desiccators for 48 h, then
weighed and ashed in a muffle furnace (=500°C) for 1
h. The ash was cooled in a desiccator (with CaSO4
added) for 24 h before final weighing. It is possible
that some loss in dry mass may occur due to preservation in formalin, but we are unaware of such a determination being reported for macroinvertebrates as
large as hydropsychids. Studies of the effects of formalin on smaller invertebrates have been somewhat
inconsistent (e.g., Howmiller 1972, Dermott and Paterson 1974, Dumont et al. 1975). We are assuming
that any such losses will be small, but recognize that
our absolute production estimates may be subject to
minor adjustments for this reason. The major conclusions reached in this paper should be unaffected.
It would have been desirable to calculate production
between sampling intervals and assess the temporal
distribution of production among species. However,
without laboratory growth studies, this can only be
estimated from field data if the survivorship of an actual cohort can be followed through time (e.g., Gillespie and Benke 1979). Since some of our species did
not have well-defined cohorts, and since our number
of samples per date was small, we could not utilize
actual cohort methods. Therefore, we calculated production for each species separately by the Hynes
method (Hynes and Coleman 1968), as modified by
Hamilton (1969), which yields only total annual production. In the Hynes method, organisms are first
grouped into selected size categories. For each size
category, an annual mean numerical density is calculated from samples taken throughout the year. The
resulting mean densities of each size category together
comprise a mean size frequency distribution which has
been referred to as an "average cohort" (Hamilton
1969, Benke and Waide 1977, Benke 1979). This av-
no
Ivcology, Vol.61, No. I
ARTHUR c. BF.NKK AND j. BRUCE WAI.I.ACR
TABLE I. Mean individual biomass (mg ash-free dry mass*) for each instar of each caddisRy species. Note that I'or some
species, individual biomass for the final (V) instar increases with season.
Species
I
II
III
Arctopsyche irrorata
.0430
1.71
.233
(W
(7)
(3)
(3)
Parapsyche cardis
.0070
.0863
.474
(23)
(18)
(AO
(11)
Hydropsyche macleodi
.0090
.0350
.161
(N)
(2)
(6)
(7)
Hydropsyche sparna
.0148
.0908
.0229
(N)
(25)
(17)
(25)
Diplectrona modesta
.0080
.0300
.110
(N)
(62)
(134)
(125) .
Dolaphilodes distinctus
.0065
.0135
.0380
(N)
(17) . (187)
(231)
* Mean ash content = 10.4% dry mass (SE = ±1.1%).
t (N) = number of individuals weighed.
erage cohort is an approximation of survivorship over
the full year, in which animals are assumed to spend
an equal length of time in each size category. Production of the average cohort is estimated by first calculating the number of individuals lost (presumably to
mortality) between successive size categories. This,
mortality also represents a loss in population biomass,
and the sum of all biomass losses is an approximation
of average cohort production. The Hynes method is
thus analogous to the removal-summation method
used for actual cohorts (Waters 1977). To obtain annual production, one must assume that there are the
same number of average cohorts during the year as
there are size categories; therefore, the average cohort
production value must be multiplied by the number of
size classes. The rationale behind the method assumes
that individuals take a full year to complete development of the aquatic stages. If development time differs
significantly from a full year, it is necessary to correct
the Hynes estimate by multiplying by 365/CPI, where
CPI, the cohort.production interval in days,-is equal
to larval development time (Benke 1,1979). The Hynes
method requires more assumptions than actual cohort
methods, but it has provided reasonable results when
applied to single species (Waters 1977).
Throughout the 12-mo study period, extensive foregut analyses were conducted on larvae (instars II-V)
of all six species of n«£-spinning caddisflies. Slides for
gut analysis were prepared using a modification of
Cummins' (1973) membrane filter technique. This
technique involved outlining gut content fragments on
paper using a drawing tube and compound microscope. Gut contents were placed in animal, vascular
plant detritus, fine detritus, filamentous algae, and dia-
Instar
Summer
IV
V,
5.71
11.73
(25)
(4)
1.74
(21)
.516
(6)
.267
(25)
.518
(75)
.116
(126)
Autumn
V2
12.70
(25)
2.65
(14)
1.72
(3)
1.06
(15)
Winter
Spring
V,
V4
26.24
42.31
(10)
(4)
7.07
13.76
(4)
(9)
— 4.44 —
(6)
— 2.63 —
(24)
3.16 —
(19)
.426
(127)
tom categories. Individual particle sizes of each category were measured (on a projected area basis) by
methods described elsewhere (Malas and Wallace
1977).
RESULTS AND DISCUSSION
Life histories and larval growth
Two aspects of larval growth essential for animal
production analysis are (1) biomass of each size category, and (2) knowledge of life history or mean larval
development time. Mean individual biomass was calculated for each of the five instars of each species
(Table 1). For species in which final instars were observed for extended periods, the final instar increased
its biomass considerably as time of pupation approached. Seasonal biomass values were therefore
calculated for final instars of certain species, and these
were utilized in estimating production.
Life history patterns are presented here as a series
of instar frequency histograms (Fig. 1):
Arctopsyche irrorata has a distinctive univoltine life
cycle. Pupation begins in mid-March and is mostly
completed by April. First instars were found in May.
Arctopsyche grows rather rapidly throughout the summer and all larvae reach the final instar by October.
However, they continue to feed and grow in the final
instar throughout the winter months (Table 1).
Parapsyche cardis also has a univoltine life cycle
which lags slightly behind that of Arctopsyche (Fig.
1). Most Parapsyche larvae are entering the final instar in March and April. Most pupate in May and early
June at the study site. First instars of the new generation were most abundant in July, although they were
February 1980
PRODUCTION OF NET-SPINNING CADDISFLIES
Arctopsyche
111
irrorata
TT
Hydropsyche
macleodi
!
"
•
'
Hydropsyche sparna
v
IV
in
u
i
I
t
f
Diplectrona
T
AUG
T
T
i
modesta
Dolophilodes
JUL
H
SEP
distinctus
OCT
NOV
JAN
FEB
MAR
APR . MAY
JUN
JUL
FIG. 1. Monthly instar frequency distributions for six species of net-spinning caddisflies in the Tallulah River. Instars
are designated as I through V. Width of each bar represents percentage of total animals found in a given instar. Dashed
bars represent ins tars not caught in benthos samples, but found by qualitative hand-picking or dipnetting.
found through September. Parapsyche overwinters in
a range of instars (II-V) with most in the fourth. Those
overwintering in the final instar continue to feed and
grow throughout the winter (Table 1). This univoltine
life history is generally consistent with that described
by Mackay (1969) for Parapsyche apicalis.
Hydropsyche macleodi is also a univoltine species
with a fairly synchronous life cycle (Fig. 1). Overwintering as fifth instars, they pupate in late March and
early April. First-instar H. macleodi were found in
June and July. The larvae.grow rather rapidly during
the summer and early autumn: All larvae are in the
overwintering final instar by November. Again, as in
the case of Arctopsyche and Parapsyche, H. macleodi
continue to feed and grow throughout the winter (Table 1).
Hydropsyche sparna, the most abundant of the five
hydropsychid species, has perhaps the most difficult
pattern to decipher. The larvae overwinter in a wide
range of instars (Fig. 1) with final instars being the
most prevalent. Pupae were found from late April
through the summer months. The possibility exists
that some H. sparna have a second generation during
the summer and early autumn months. However, the
majority of the instar distributions follow a definite
synchronous trend during the summer months. Williams and Hynes (1973) found a similar annual distribution of instars for H, sparna in a Canadian stream,
and assumed that a small second generation developed
during the summer.
Diplectrona modesta is clearly univoltine at the
study site (Fig. 1). Pupae were found from May to
112
ARTHUR C. BENKE AND J. BRUCE WALLACE
Ecology, Vol. 61, No. 1
TABLE 2. Hynes production calculations for Arctopsyche irrorata.
Instar
I
II
III
IV
v,t
V2t
v3t
V4t
Annual
mean
density2
(No./m )
11.028
11 000
11.556
' 1.443
i1.558
i1.550
;1.225
.250
N = \2.6t
Mean
individual
biomass
(mg)
.043
233
1.7 1
5.71
11.73
12.70
26.24
42.31
Annual
mean
standing
stock
(mg/m2)
.044
233
2.661
25.352
18.275
19.691
32.143
10.578
No.
lost/m2
Individual
biomass
at loss
(mg)
Biomass
loss
(mg/mz)
Correction
factor
x8
.028
-.556
-2.887
2.885
.008
.325
.975
.250
.138
.972
3.708
8.718
12.217
19.472
34.275
42.310
.004
-.540
-10.704
25.151
.098
6.328
33.418
10.578
.032*
-4.320*
-85.632*
201.208
.784
50.624
267.344
84.624
B = 109.0*
P = 604.6*
* Not included in production summation; see text.
t V, = summer, V2 = autumn, V3 = winter, V4 = spring.
* N = mean density, B = mean standing stock biomass, P = annual production.
early August. Some gravid females were captured in
the field in September. The larvae overwinter in instars II I-V and by April most of the larvae are in the
final instar. Cushman et al. (1977) also found a univoltine life history for D. modesta in a Tennessee
stream.
The life history of Dolophilodes distinctus is also
difficult to decipher, having at least two and possibly
three generations per year (Fig. 1). Ross (1944) discussed the unusual winter generation of this species.
Pupae of the winter generation were found primarily
from November through January. First instar offspring
of the winter generation are prevalent in the Tallulah
in January and February. Throughout the remainder
of the year, a wide assortment of instars is found. Ross
(1944) noted "this species is remarkable because of
the production of adults during the entire year."
Production of caddisflies
Production calculations, using the Hynes method,
are illustrated for Arctopsyche irrorata (Table 2). The
first four size categories are represented by the first
four instars. The last four categories are represented
by seasonal biomass values (Table 1) of the final instar.
The sums of the second, fourth, and eighth columns
of Table 2 are annual mean density (TV), annual mean
standing stock biomass (B), and annual production (P),
respectively. Negative biomass losses for the first
three small size categories (see last two columns of
Table 2) were ..excluded from the production summation since the second column of the table represents
survivorship of an average cohort, and it is theoretically impossible to have lower densities for instars IIII than for later instars. Hamilton (1969) recommends
retaining negative values, since their exclusion generally results in a positive bias. However, we feel the
apparent "negative production" of early instars is an
anomaly due to both lower efficiency in sampling these
stages and a shorter amount of time spent in the early
instars. The exclusion of these first few values increased the production estimate for Arctopsyche by
= 17%, but when applied to the other species, the increase was never >5%.
For univoltine species in which cohorts are distinct
(Arctopsyche irrorata, Parapsyche cardis, Hydropsyche macleodi, and Diplectrona modesta), there
was not a significant amount of time spent in nonlarval
stages (Fig. 1). We therefore assumed that the cohort
production interval (CPI) is 12 mo for these species,
and the Hynes table estimate was not adjusted. For
H. sparna, which was primarily univoltine with a possible small second generation, we conservatively assumed that CPI was also equal to 12 mo. Dolophilodes
distinctus had at least two generations per year and
we assumed that CPI is 6 mo. The Hynes estimate for
Dolophilodes was therefore multiplied by two according to Benke (1979). Further elucidation of life history
may reveal that the CPI is shorter and production
higher than we calculated for Dolophilodes.
Annual mean density, mean standing stock biomass,
production, and P/B ratios are compared- for each
species (Table 3). Although densities of Arctopsyche
irrorata are very low, their large individual biomass
values (Table 1) result in their accounting for 60% of
the net-spinning Trichoptera production. The next
largest species, Parapsyche cardis, accounts for 14%
•of the total. The four smaller species, although comprising 83% by number, contribute < 26% to the total
production.
For all five species of Hydropsychidae, where we
assumed CPI = 12 mo, the cohort P/B ratio (the P/B
ratio over the life-span of a cohort) is equal to the
annual P/B ratio. Dolophilodes distinctus, for which
we assumed CPI to be only 6 mo, has a cohort P/B
ratio of 4.16, or half the annual P/B ratio. Cohort P/B
ratios for all species are thus between four and six,
February 1980
PRODUCTION OF NET-SPINNING CADDISFLIES
113
TABLE 3. Summary of mean values for production parameters.*
.
.
_
(Nord and Schmulbach 1973, Rhame and Stewart
1976), our personal observations of abundance in such
rivers as the Savannah and Altamaha in Georgia, and
N
B
(mg- Annual Fremling's (1960) description of hydropsychid popuSpecies
(No./m2) (mg/m 2 )m- 2 -yr-') P/B
lations in the upper Mississippi suggest that production of net-spinning caddisflies on solid substrate surArctopsyche irrorata
12.6
109.0
604.6 5.55
Parapsyche cardis
16.3
33.7
143.2
4.25
faces is high in many rivei^ (i.e., approaching or
Hydropsyche macleodi
5.44
7.2
4.9
26.7
exceeding 10 g/m2 [dry mass]). It appears, as Fremling
Hydropsyche sparna
46.0
20.3
86.3
4.25
(1960) originally suggested, that substrate availability
Diplectrona modesta
15.6
11.5
53.0
4.61
Dolophilodes distinctus
76.5
10.4
along with sufficient current seem to limit abundance,
86.4 1 8.31t
(and thus production) of these organisms in many large
Total
174.2
189.8 1000.2 5.27
rivers.
* All samples from riffle areas; see text.
In smaller streams, such as the upper Tallulah Rivt With two generations per year or CPI (cohort production
interval) = 6 mo (Benke 1979).
er, one often finds much unoccupied solid substrate
with relatively low densities of net-spinning caddisflies. However, high abundance does occur below impoundments where lake plankton increases the food
supplies (e.g., Hynes 1970, Oswood 1976), and rewithin the same range of values observed for many cently Hopkins (1976) reported production of Hydrospecies (Waters 1977).
psyche sp. to be as high as 15 g/m2 (dry mass) in the
Our species-specific estimates of caddisfly produc- organically enriched Horokiwi Stream, New Zealand.
tion from 27 to 605 mg/m2 (ash-free dry mass) roughly All of these observations suggest that production of
cover the same range of values presented by Waters net-spinning caddisflies in small streams is limited by
(1977) for all species of caddisflies in his review on the amount of high quality food available in the seston.
secondary production of inland waters. The lone exTrophic basis of production
ception was Oligophlebod.es sigma (Limnephilidae)
with an annual production of 2.88 g/m2 (dry mass) in
To calculate the amount of production attributable
a Utah stream (Pearson and Kramer 1972). No esti- to various food types, three kinds of information are
mates existed for production of total net-spinning Tri- needed: (1) annual production which is presented
choptera, or total Trichoptera at the time of Waters' above, (2) proportion of food categories consumed,
(1977) review. However, Nelson and Scott (1962) re- and (3) assimilation and production efficiencies for the
ported production of filter-feeding insects in the Oco- various food types.
nee River (Georgia) to be 20.2 g/m2 (dry mass) of
Some 480 larval guts and =270,000 foregut particles
which a large fraction consisted of hydropsychids. The from specimens of the six species collected over the
only estimate for hydropsychids was that of Cushman 12-mo period were examined (Table 4). Since the peret al. (1975, 1977): 98 mg/m2 (dry mass) (as converted centages in each food category were tabulated from
by Wajcrs 1977) for Diplfctrona modesta in a small area measurements, relative volume of larger food
Tennessee stream. This is quite comparable lo our particles, such as animal food, in reality would be
ash-free value of 53 mg/m2 for the same species. Two somewhat larger than the area percentage. Volume of
more recent studies by Hopkins (1976) and Van Ars- small food particles, such as fine detritus and algae,
dall (1977) discussed below present much higher val- would be somewhat less than the area percentage.
ues for hydropsychids.
However, we are assuming that the area percentages
Although our production values are comparable to are reasonable approximations of relative volumes.
most previous estimates for caddisflies in streams, Furthermore, large particles of animal origin are often
they are low to moderate in comparison to production fragmented upon feeding, and particle size of some
reported for aquatic insects in general (Waters 1977), foregut contents would be less than actually conand are much lower than what we are beginning to sumed. This also made specific identification of animal
find in rivers. Mann (1975) has suggested that second- prey very difficult, but it appeared that midge larvae,
ary production in small streams will generally be much pupae, and adults, and mayflies were the most nulower than in larger rivers, and this seems to be true merous. Less frequently found were copepods and
for net-spinning caddisflies based upon our present small individuals of craneflies, elmid beetles, stoneknowledge. In addition to Nelson and Scott's flies, caddisflies (including smaller hydropsychids),
.(1962) high value for filter-feeders mentioned and other common components of stream benthos.
above, Van Arsdall (1977) recently reported proThe caddisflies are arranged in order of decreasing
duction values of 11-27 g/m2 (dry mass) of woody percentage of animal food (Table 4). This corresponds
substrate (or snag) surface area for net-spinning cad- roughly with a decrease in capture net mesh opening
disflies in the blackwater Satilla River in Georgia. size, ranging from =403 x 534 /xm for fifth-instar ArcStanding stock values of hydropsychids in other rivers topsyche to 1 x 6 /urn for Dolophilodes (Wallace et a!.
•w:
114
ARTHUR C. BENKE AND J. BRUCE WALLACE
Ecology, Vol. 61, No. 1
TABLE 4. Foregut contents of each species (instars II-V) across all seasons.*
Percent (by projected area) in each food category
Vascular
Fine
Filamentous
plant
Species
detritus
algae
Animal
detritusf
73.1
2.0
Arctopsyche irrorata
4.2
17.9
Parapsyche cardis
62.3
11.0
18.6
5.3
Hydropsyche macleodi
37.5
11.6
37.6
4.5
Hydropsyche sparna
23.5
10.2
8.7
36.6
19.3
38.5
1.8
Diplectrona modesta
37.3
Dolophilodes distinctus
1.6
0.8
87.4
0.1
* Expanded and modified from Wallace et al. (1977).
t Probably represents egested particles from shredders (e.g., Cummins 1973).
1977). There is also a rough correspondence with a
decline in food particle size from 10718 /imz for Arctopsyche to 139 fan2 for Dolophilodes (Wallace et al.
1977). From Table 4 we can conclude that Arctopsyche
and Parapsyche are mostly carnivorous; Hydropsyche
macleodi, H. sparna, and Diplectrona consume mostly detritus with significant amounts of animal food;
Dolophilodes relies primarily on fine detritus.
Shapas and Hilsenhoff (1976) also measured gut
contents on a projected-area basis using a similar technique and obtained results comparable to ours for the
same or congeneric species. Rhame and Stewart (1976)
estimated volumetric proportions for plant and animal
food categories of three hydropsychids, but excluded
the large unidentifiable fraction. Thus, their results
cannot be compared directly with ours. However, they
did show that food selection was seasonal. Very little
seasonal variation by food type was observed within
instars for our species collected on a monthly basis.
However, mean animal food particle size did vary seasonally within instars. This may be a reflection of seasonal variation of prey size. Other feeding studies on
closely related species using numbers of particles rather than area (or volume) approximations attach less
significance to the animal fraction which often comprises the largest particles (e.g., Mecom 1972, Williams and Hynes 1973).
Observation of food selection alone can be very
misleading in interpreting the role that an organism
plays in stream metabolism. For example, examination of the foregut contents (Table 4) indicates that on
the average, net-spinning caddisflies in the Tallulah
River consume 36% animals, 52% detritus, and 12%
algae, indicating that detritus is the most significant
food source. If feeding analyses were weighted according to mean densjty of each species (with Dolophilodes the highest), or if number of particles In each
food category rather than volume was considered, the
importance of detritus would seem even greater. To
obtain a more accurate assessment of a consumer's
utilization of food types, production and food-specific
assimilation efficiencies of the consumer must be taken into account.
Diatoms
2.9
2.9
8.8
21.0
3.0
10.2
Since we did not measure assimilation and production efficiencies for our species, we have made assumptions for these values based on the limited avail; able literature. We have assumed a net production
efficiency (production/assimilation = NPE) of 50%
based upon Edington and Hildrew's (1973) study of
Hydropsyche fulvipes and Diplectrona felix. '• This is
somewhat higher than the average of 43% found by
Otto (1974) for the limnephilid Potamophylax, cingulatus, and is reasonably consistent with values from
a variety of freshwater invertebrates (see McCulJough
1975). For assimilation efficiency (assimilation/ingestion = AE), we have assumed 30% for algae and 70%
for animal food as suggested by Winterbourn (1971)
based upon his study of the phryganeid Banksiola
crotchi. The 30% for algae is comparable to the AE
for the grazing Glossosoma nigrior (Glossosomatidae)
(Cummins 1975), and also compares favorably to
McCullough's (1975, McCullough et al. 1979) AE for
the hydropsychid Cheumatopsyche analis feeding on
diatoms (if AE is based upon total dry mass). The 70%
value for animal food seems quite reasonable, and perhaps conservative, in .view of the limited information
available for carnivore assimilation efficiencies which
have approached 90% (Lawton 1970, Brown 1974,
Heiman and Knight 1976). Assimilation efficiencies for
aquatic detritivores are the lowest among consumers.
Reported, values generally range from 5 to 25%, with
many close to 10% (McDiffett 1970, Ladle et al. 1972,
Otto 1974; Anderson and Grafius 1975, McCullough
1975, Winterbourn and Davis 1976, McCullough et al.
1979). We have therefore assumed that the assimilation efficiency for the detritus fraction of the food is
10%. For each of these assimilation efficiencies, our
•assumptions are probably toward the conservative
(low) side (e.g., see Heal and MacLean 1975), but the
relative values of the three are consistent with the literature.
For each particular food category, the relative contribution made to caddisfly production is calculated as
the product of the gross production efficiency (AE x
NPE = production/ingestion) and the percentage of
food type eaten. The procedure is illustrated for Arc-
February 1.980
115
PRODUCTION OF NET-SPINNING CADDISFLIES
TABLE 5. Procedure for calculating production
attributed to each food type and the amount of food type consumed for
Arctopsyche irrorala (P = 605 mg m~2 yr"1).
Net
Food
Assimila- production
type in tion effieffiRelative
foregut
ciency* ciency* amount to
(%)
(AE)
(NPE) production
Animal
73.1 X
.70 X
.5 = 25.6
Vascular plant detritus
4.2 X
.10 X
.5 =
.21
Fine detritus
.90
17.9 X
.10 X
.5 =
Filamentous algae
2.0 X
.30 X
.5 =
.30
Diatoms
.30 X . .5 =
.44
2.9 X
* Based upon literature values; see text.
topsyche irrorata in Table 5. These relative contributions are converted to percentages which are each
multiplied by caddisfly production to obtain the
amount of production attributed to each food type.
Then, dividing by the factor AE x NPE will yield the
amount of each food type actually consumed. It is also
possible to calculate amount of each material egested,
and this information is summarized for the caddisflies
as a group in Table 6. The accuracy of these estimates
depends on our assumptions regarding AE and NPE,
but the latter ratio, which is probably equivalent for
all food types of a given species, would only affect the
estimate of amount of food type consumed.
The proportion of production attributed to the various food types was determined for each of the six
caddisflies (Table 6). As one would expect for those
species which are primarily carnivorous (Arctopsyche
and Parapsyche, Table 4), most of their production
can be attributed to animal food. However, even
three species which are primarily detritivores (Hydropsyche macleodi, H. sparna, and Diplectrona modestd) are dependent on animal food for more than half
of their production based on our assumed assimilation
efficiencies. Only for Dolophilodes, which feeds primarily on detritus, is animal food a relatively small
contribution to caddisfly production. Although the
Production attributed
to food
type
(%)
93.3
0.8
3.3
1.1
1.6
Production atGross
tributed produc- Amount
to food
tion ef- food type
type ficiency* consumed
2
(mg-m-^ (AE x
(mg-nr
yr')w
NPE)
yr1)
564 +
.35 = 1611
100
5 -^
.05 =
20 H.05 =
400
7; -r : .15 --. *.: 47 : ..:.
67
10 +
.15 =.
contribution to Dolophilodes production by animal
food appears to be small, it could provide an important
growth supplement as Anderson (1976) demonstrated
for the detritivorous caddisfly Clistorina magnified.
Whereas conclusions based upon feeding analyses
alone (Table 4) indicate detritus to be the food of greatest importance, the analyses illustrated in Table 5 and
summarized in Table 6 presents a very different picture. Almost 80% (797 mg/m2) of total net-spinning
caddisfly production is attributed to animal food, 2.2%
to vascular plant detritus, 10.5% to fine detritus, 2.2%
to filamentous algae, and 5.5% to diatoms. Although
most caddisfly production is attributable to animal
food, that based upon algal food (7.7%) is more than
half that based upon total detritus (12.7%). Furthermore, it should be recognized that an unknown proportion of the fine detritus, which accounts for at least
10% of caddisfly production, may be derived from autochthonous sources (e.g., Goldman and Kimmel
1978). If other grazer/detritivores are utilizing algae
and detritus in proportions similar to that of caddisflies, it could mean that autochthonous food is more
important to consumers than we have been led to believe is the case in small streams, as recently suggested
by Minshall (1978).
The relatively low contribution of detritus to total
TABLE 6. Proportion of annual production attributed to various food types.
Species
Arclnpxyche irrorata
Ptirapxyche canli.i
Hydropsyche macleodi
Hydropsyche sparna
Diplectrona modesta
Dolophilodes distinctus
Total
Consumption to
account for production
Egested as detritus
Vascular
Fine
Diatoms
plant
Filamentous
Animal
Producdetritus
2
mg-nr'-yr-'1
mg-m" -yr-'
detritus
algae
tion2
mg-m- 2 -yr-'
2
2
1
(mg-m- -yr-')
(%)
(%)
mg-m- -yr-'(%:)
mg-m- -yr (%)
(%)
10(1.6)
605
5 (0.8)
564 (93.3)
20 (3.3)
7(1.1)
3(1.8)
5 (3.3)
5 (3.8)
143
127 (S8.9)
3 (2.2)
2 (7.5)
1 (3.9)
27
20 (74.6)
1 (3.3)
3(10.7)
18 (21.0)
8(8.7)
3 (3.4)
86
47 (54.8)
10(12.2)
2 (4.0)
1(2.4)
53
32 (59.9)
9 (16.6)
9(17.1)
20 (23.5)
0 (0.3)
86
7 (8.6)
58 (67.6)
1 (0.6)
22 (2.2)
55 (5.5)
1000
797 (79.7)
22(2.2)
105 (10.5)
5331
3329
2277 (42.7)
683
440 (8.3)
396
2100 (39.4)
1890
147 (2.8)
103
367 (6.8)
257
116
ARTHUR C. BENKE AND J. BRUCE WALLACE
net-spinning caddisfly production may lead to the initial conclusion that, with the exception of Dolophilodes, detritus is not a very important food source for
these insects. However, this is not necessarily true
because several other factors should be considered.
We have found that guts of early instars 11-111 contain
primarily fine particulate detritus fragments and only
seldom are animal fragments found (Wallace et al.
1977). Schroder (1976) found similar results for Hydropsyche instabilis in a mountain stream in the southern Black Forest, Germany. The obvious size limitation of food particles to the early life stages
undoubtedly restricts the types of food choices available to them. The gut content particle sizes of these
early instars are very small and thus contribute only
a small portion of the total gut contents of all instars
of a given species on a yearly basis as summed in
Table 4. There is also ample documentation in the literature that early instars compose a very small amount
of the total production of a given species (e.g., Kimerle and Anderson 1971). Consequently, these small,
early life stages, while representing a small portion of
the total production, are very dependent on detritus
as a food source.
Two further precautionary notes regarding the relative importance of food types to production of netspinning caddisflies should be mentioned. One is that
the two species comprising most of our production
estimate, Arctopsyche irrorata and Parapsyche cardis, are two of the largest, and perhaps two of the
most carnivorous net-spinning species in eastern
North America. Such carnivorous species seem largely restricted to cool headwater streams, and the relative importance of animal food to production of netspinning caddisflies is likely to decrease as one moves
downstream. The other factor relating to our interpretation of food type importance is that we have assumed equal rates of digestion for each food type. If
one type is digested more rapidly than the others, it
will be underrepresented in gut contents (Cummins
1973), and in its contribution to caddisfly production.
Thus, information on differential digestion rates would
allow one to refine the estimates we have made in this
paper and for any future studies of a similar nature.
Role of net-spinning caddisflies in
processing of seston
Drifting particulate organic matter, or seston, is processed within streams in a series of stages in which it
is captured or colonized by heterotrophs, partially
processed and released in the current, and thus available for further processing downstream. This cascading of partially processed organic material and its reutilization has been termed "spiralling" (Webster
1975). Based upon our production and feeding analyses, we can make a first attempt at quantifying the role
that net-spinning caddisflies play in the spiralling of
seston in mountain streams (Fig. 2). This is possible
Ecology, Vol. 61, No. 1
VASCULAR
PLANTS
MICRO--AND SMALL
MACROINVERTEBRATES
ALGAE
^bacteria
-NET-SPINNING
CADDISFLIES
FIG. 2. Role of net-spinning caddisflies in "spiralling" of
organic matter
in the Tallulah River (all flows in mg dry
mass-m~ 2 -yr l ).
•
because we can estimate the caddisfly production attributable to each food category and the amount of
each food type consumed and egested as feces (see
last column of Table 5 and bottom two rows of Table
6). Although this information can be calculated separately for each species given the information in this
paper, we present only the totals for all species at the
bottom of Table 6. Of the 5331 mg/m2 of organic matter
consumed by the caddisflies, 42.7% is animal, 47.7%
detritus, and 9.6% plant material. These percentages
differ from the relative amount of caddisfly production
attributed to them because of the different assimilation
efficiencies used for various foods. Interestingly, Dolophilodes distinctus consumes more than half (55%)
of the 2.1 g/m2 of fine detritus consumed by all netspinning caddisflies in a year, even though it accounts
for <9% of caddisfly production. In fact, because of
the low quality food utilized by Dolophilodes, its total
.consumption of organic 'matter is =60% of that consumed by Arctopsyche irrorata, even though its production is only 14% of Arctopsyche production.
Only the major pathways of food processing are
shown in Fig. 2. The fate of net-spinner production,
such as tosses to predation and emergence, is not
shown, and thus "inflows" to the net-spinners do not
balance the "outflow" as feces. As a group, the caddisflies ingest roughly equivalent volumes of detritus
and animals. However, because of the higher food value of animals, most of their production is due to this
food source.
We can deduce that net-spinning caddisflies are
probably not the major consumers of detritus or algae
in the system since their production of 1 g/m2 is less
than the production of their animal food consumed,
2.3 g/m2 (Fig. 2, and Table 6, next-to-last line). This
latter value represents a minimum estimate of production by their prey since it does not include any other
February 1980
PRODUCTION OF NET-SPINNING CAOD1SFLIES
prey mortality or emergence. These prey, which consist largely of midges and mayflies, probably consume
detritus and algae as their major food source. Furthermore, there are other large primary consumers in
the system, especially late-instar shredders (e.g., large
craneflies and the stonefly Pteronarcys), that are not
consumed by the caddisflies. However, their feces do
contribute to the caddisfly diet (e.g., Short and Maslin
1977) and may represent the most nutritious part of
the detritus.
We can also see from Fig. 2 that as a group, the netspinning caddisflies actually egest more detritus (3329
mg/m2) as feces than they ingest (2540 mg/m8). This
comparison is exaggerated to some degree since some
of the egested material may actually consist of viable
unbroken algal cells. However it seems clear that this
group of caddisflies is a net producer of detritus in the
upper Tallulah River. By filtering animals', detritus,
and algae from the seston, and releasing only detritus
as feces, they essentially lower the food quality of the
seston. But in doing so, they increase the food supply
for the microconsumers that serve as their prey, and
create a high quality food (their own production)
which is available as considerably larger food particles
to large carnivores.
This paper has been concerned with the trophic dynamics of one group of filter-feeding collectors. We
have shown how feeding, production, and bioenergetics information can be integrated to help establish the
functional role of this group in stream metabolism. If
the role of other consumers in lotic ecosystems is to
be understood, it may be necessary to take a similar
approach for all major groups simultaneously. In some
groups, the feeding analysis may be considerably simpler, since many species are much more restricted in
their feeding than the omnivores considered here (e.g.,
many odonates and plecopterans are strictly carnivorous). Furthermore, the techniques developed for
feeding studies seem quite reliable and quantitative for
many groups of animals (Cummins 1973, Wallace et
al. 1977). Much greater emphasis needs to be placed
on production analyses, and with the availability of
new methods such as the Hynes method (Hynes and
Coleman 1968, with modifications by Hamilton 1969,
Benke and Waide 1977, Waters 1977, and Benke 1979),
it now seems much more feasible for aquatic biologists
to estimate production for separate components of entire benthic communities.
The major shortcoming of the present paper is lack
of species-specific bioenergetics data, including assimilation efficiencies and gut passage times for various
food types, as well as net production efficiencies. Our
final figures are thus somewhat speculative, but our
overall, conclusions allow for some variation in the
bioenergetics assumptions and should be sound. The
conclusions are in agreement with Coffman et al.
(1971) who found that small invertebrates were the
food responsible for much of the autumnal biomass of
117
macrohenthos in a small woodland stream in Pennsylvania. They suggested that these small detritivores
and grazers must have very short turnover times. Future work may determine whether literature values can
be generalized to other studies as we have done, or if
bioenergetics data are required on a species-byspecies basis.
a. '
ACKNOWLEDGMENTS
This work was supported by Grants Number DEB 74-00618
A01 and DEB 78-03143 from the National Science Foundation. We are indebted to Kay Campbell and Diane Mates .'for'
technical assistance. The constructive comments of Kenneth
W. Cummins and an anonymous reviewer are gratefully acknowledged.
LITERATURE CITED
Anderson, N. H. 1976. Carnivory by an aquatic detritivore,
Clistorina magnifica (Trichoptera: Limnephilidae). Ecology 57:1081-1085.
Anderson, N. H., and E. Grafius. 1975. Utilization and processing of allochthonous material by stream Trichoptera.
Verhandlungen der Internationale Vereinigung fur Theoretische und Angewandte Limnologie 19:3083-3088.
Benke, A. C. 1979. A modification of the Hynes method for
estimating secondary production with particular significance for multivoltine populations. Limnology and Oceanography 24:168-171.
Benke, A. C., and J. B. Waide. 1977. In defence of average
cohorts. Freshwater Biology 7:61-63.
Brown, A. V. 1974. Ecological energetics of the dobson fly,
Corydalus cqrnutus. Dissertation. North Texas State University, Dentoh, Texas, USA.
Coffman, W. P., K. W. Cummins, and J. C. Wuycheck.
1971. Energy flow in a woodland stream ecosystem. I.
Tissue support trophic structure of the autumnal community. Archiv fur Hydrobiologie 68:232-276.
Cummins, K. W. 1973. Trophic relations of aquatic insects.
Annual Review of Entomology 18:183-206.
. 1975. Macroinvertebrates. Pages 170-198 in B. A.
Whitton, editor. River ecology. University of California
Press, Berkeley, California, USA.
Cushman, R. M., J. W. Elwood, and S. G. Hildebrand.
1975. Production dynamics of-AJloperla mediana Banks
(Plecoptera: Chloroperlidae) and Diplectrona modesta
(Trichoptera: Hydropsychidae) in Walker Branch, Tennessee. Environmental Sciences Division Publication 785, Oak
Ridge National laboratory. Oak Ridge, Tennessee, USA.
Cushman, R. M., J. W. Elwood, and S. G. Hildebrand.
1977. Life history and production dynamics of Alloperla
mediana and Diplectrona modesta in Walker Branch, Tennessee. American Midland Naturalist 98:354-364.
Dermott, R. M., and C. G. Paterson. 1974. Determining dry
weight and percentage dry matter of chironomid larvae.
Canadian Journal of Zoology 52:1243-1250.
Dtimonl, H. J., I. Van de Velde and S. Dumont. 1975. The
dry weight estimate of biomass in a selection of Cladocera,
Copepoda, and Rotifera from the plankton, periphyton,
and benthos of continental waters. Oecologia 19:75-97.
Edington, J. M., and A. H. Hildrew. 1973. Experimental
observations relating to the distribution of net-spinning
Trichoptera in streams. Verhandlungen der Internationale
Vereinigung fur Theoretische und Angewandte Limnologie
18:1549-1558.
Fremling, C. R. 1960. Biology and possible control of nuisance caddisflies of the upper Mississippi River. Research
Bulletin 483, Iowa Agricultural Experiment Station, Ames,
Iowa, USA.
118
ARTHUR C. BENKE AND J: BRUCE WALLACE
Gillespie, D. M., and A. C. Benke. 1979. Methods of calculating cohort production from field data—some relationships. Limnology and Oceanography 24:171-176.
Goldman, C. R., and B. L. Kimmel. 1978, Biological processes associated with suspended sediment and detritus in
lakes and reservoirs. Pages 19-44 in J. Cairns, Jr., E. F.
Benfield, and J. R. Webster, editors. Current perspectives
on river-reservoir ecosystems. North American Benthological Society, Columbia, Missouri, USA.
Hamilton, A. L. 1969. On estimating annual production.
Limnology and Oceanography 14:771-782.
Heal, O. W., and S. F. MacLean, Jr. 1975. Comparative
productivity in ecosystems—secondary productivity. Pages
89-106 in W. H. van Dobben and R. H. Lowe-McConnell,
editors. Unifying concepts in ecology. Junk, The Hague,
Netherlands.
Heiman, D. R., and A. W. Knight. 1976. The influence of
temperature on the bioenergetics of the carnivorous stonefly nymph, Acroneuria californica Banks (Plecoptera: Perlidae). Ecology 56:105-116.
Hopkins, C. L. 1976. Estimate of biological production in
some stream invertebrates. New Zealand Journal of Marine and Freshwater Research 10:629-640.
Howmiller, R. P. 1972. Effects of preservatives on weights
of some common macrobenthic invertebrates. Transactions of the American Fisheries Society 101:743-746.
Hynes, H. B. N. 1970. The ecology of stream insects. Annual Review of Entomology 15:25-42.
Hynes, H. B. N., and M. J. Coleman. 1968. A simple method of assessing the annual production of stream benthos.
Limnology and Oceanography 13:569-573.
Kimerle, R. A., and N. H. Anderson. 1971. Production and
bioenergetic role! of the midge Glyptotendipes barbipes
(Staeger) in a waste stabilization lagoon. Limnology and
Oceanography 16:646-659.
Ladle, M., J. A. B. Bass, and W. R. Jenkins. 1972. Studies
on production and food consumption by the larval Simuliidae (Diptera) of a chalk stream. Hydrobiologia 39:429448.
Lawton, J. H. 1970. Feeding and food energy assimilation
in larvae of the damselfly Pyrrhosoma nymphula (Sulz.)
(Odonata: Zygoptera). Journal of Animal Ecology 39:669689.
'
Mackay, R. J. 1969. Aquatic insect communities of a small
stream, on Mont St. Hilaire, Quebec. Journal of the Fisheries Research Board of Canada 26:1157-1183.
. 1978. Larval identification and instar association in
some species of Hydropsyche and Cheumatopsyche (Trichoptera: Hydropsychidae). Annals of the Entomological
Society of America 71:499-509.
Malas, D., and J. Bi Wallace. 1977. Strategies for coexistence in three species of net-spinning caddisflies (Trichoptera) in second-order southern Appalachian streams! Canadian Journal of Zoology 55:1829-1840.
Mann, K. H. 1975. Patterns of energy flow. Pages 248-263
in B. A. Whitton, editor. River ecology. University of California Press, Berkeley, California, USA. .
Mason, W. T., Jr., and P. P. Yevich. 1967. The use of Phloxine B and Rose Bengal stains to facilitate sorting benthic
samples. Transactions of the American Microscopical Society 86:221-223.
McCullough, D. A. 1975. The bioenergetics of three aquatic
insects determined by^adioisotopic analysis. Thesis. Idaho
State University, Pocatello, Idaho, USA.
McCullough, D. A., G. W. Minshall, and C. E. Gushing.
1979. Bioenergetics of lotic filter-feeding insects Simulium
. spp. (Diptera) and Hydropsyche occidental (Trichoptera)
and their function in controlling organic transport in
streams. Ecology 60:585-596.
McDiffett, W. F. 1970. The transformation of energy by a
Ecology, Vol. 61, No. 1
stream detritivore, Pteronarcyx scatti (Plecoptera). Ecology 51:975-988.
Mecom, J. O. 1972. Feeding habits of Trichoplera in a
mountain stream. Oikos 23:401-407.
Minshall, G. W. 1978. Autotrophy in stream ecosystems.
BioScience 28:767-771.
Nelson, D. J., and D. C. Scott. 1962. Role of detritus in the
productivity of a rock-outcrop community in a Piedmont
stream. Limnology and Oceanography 7:396-413.
Nord, A. E., and J. C. Schmulbach. 1973. A comparison of
the macroinvertebrate aufwuchs in the unstabilized and
stabilized Missouri River. Proceedings of the South Dakota
Academy of Science 52:127-139. •'
Oswood, M. W. 1976. Comparative life histories of the Hydropsychidae (Trichqptera) in- a Montana lake outlet.
American Midland Naturalist 96:493-497.
Otto, C. 1974. Growth and energetics in a larval population
of Potamophylax cingulatus (Steph.) (Trichoptera) in a
south Swedish stream. Journal of Animal Ecology 43:339361.
Pearson, W. D., and R. H. Kramer. 1972. Drift and production of two aquatic insects in a mountain stream. Ecological Monographs 24:365-385.
Rhame, R. E., and K. W. Stewart. 1976. Life cycles and
food habits of three Hydropsychidae (Trichoptera) species
in the Brazos River, Texas. Transactions of the American
Entomological Society 102:65-99.
Ross, H. H. 1944. The caddisflies, or Trichoptera, of Illinois. Bulletin 23(1), Illinois Natural History Survey, Urbana, Illinois, USA.
Schroder, P. 1976! On the diet of larval instars of the caddisfly Hydropsyche instabilis (Trichoptera: Hydropsychidae). Entomologica Germanica 3:260-264.
Shapas, T. J., and W. L. Hilsenhoff. 1976. Feeding habits
of Wisconsin's predominant lotic Plecoptera, Ephemeroptera, and Trichoptera. Great Lakes Entomologist 9:175188.
Short, R. A., and P. E. Maslin. 1977. Processing of leaf
litter by a stream detritivore: effect on nutrient availability
to collectors. Ecology 58:935-938.
Van Arsdall, T. C., Jr. 1977. Production and colonization
of the snag habitat in a southeastern blackwater river. Thesis. Georgia Institute of Technology, Atlanta, Georgia,
USA.
Wallace, J. B. 1975. Food partitioning in net-spinning Trichoptera larvae: Hydropsyche venularis, Cheumatopsyche
etrona, and Macronema zebratum (Hydropsychidae). Annals of the Entomological Society of America 68:463-472.
Wallace, J. B., J. R. Webster, and W. R. Woodall. 1977.
The role of filter-feeders in flowing waters. Archiv fur Hydrobiologie 79:506-532.
Waters, T. F. 1977. Secondary production in inland waters.
Advances in Ecological Research 10:91-164.
Webster, J. R. 1975. Potassium and calcium dynamics in
stream ecosystems on three southern Appalachian watersheds of contrasting vegetation. Dissertation. University of Georgia, Athens, Georgia, USA.
Wiggins, G. B. 1977. Larvae of the North American caddisfly genera (Trichoptera). University of Toronto Press,
Toronto, Canada.
Williams, N. E., and H. B. N. Hynes. 1973. Microdistribution and feeding of the net-spinning caddisflies (Trichoptera) of a Canadian stream. Oikos 24:73-84.
Winterbourn, M. J. 1971. An ecological study of Banksiola
crotchi Banks (Trichoptera: Phryganeidae) in Marion
Lake, British Columbia. Canadian Journal of Zoology
49:637-645.
Winterbourn, M. J., and S. F. Davis. 1976. Ecological role
of Zelandopsyche ingens (Trichoptera: Oeconesidae) in a
beech forest stream ecosystem. Australian Journal of Marine and Freshwater Research. 27:197-215.