Springer 2005
Hydrobiologia (2005) 543: 119–133
DOI 10.1007/s10750-004-6950-0
Primary Research Paper
Species composition and assemblage structure of chironomid larvae (Diptera:
Chironomidae) attaching to the artificial substrates in a Japanese temperate
basin, in relation to the longitudinal gradient
Eiso Inoue*, Koichiro Kawai & Hiromichi Imabayashi
Laboratory of Aquatic Ecology, Graduate School of Biosphere Science, Hiroshima University, 1-4-4 Kagamiyama,
Higashihiroshima, Hiroshima 739-8528, Japan
(*Author for correspondence: Tel.: +81-(0)82-424-7895, Fax: +81-(0)82-422-7059,
E-mail: inoueiso@hiroshima-u.ac.jp)
Received 15 March 2004; accepted in revised form 30 November 2004
Key words: Chironomidae, artificial substrate, topographic type, riparian canopy, bank protection, longitudinal
distribution
Abstract
The relative importance of natural and anthropogenic factors, especially topographic type, riparian canopy,
altitude, temperature and bank protection, on larval chironomid assemblage was investigated in a Japanese
basin. To focus on the macro-scale factors, a concrete block, as an artificial substrate, was used for chironomid
collection so that sampling regime may be identical among the sites. Partial CCA using sampling month as a
covariable revealed that topographic type, riparian canopy coverage, water temperature and altitude were the
main factors influencing species distribution. Stempellinella tamaseptima, Polypedilum tamanigrum and five
Rheotanytarsus species showed positive, whereas five Cricotopus species showed negative associations with
canopy coverage. Some traditional longitudinal zonations of species were still shown. Chironomus flaviplumus
and Chironomus yoshimatsui were merely associated with lower reaches. Stepwise multiple regressions of the
assemblage indices on the environmental variables were applied. Bank protection and depth showed negative
correlations with Shannon diversity H0 . Both topographic type and depth showed negative correlations with
Pielou equitability J. Topographic type (lower reach) and specific conductance showed positive, while bank
protection showed a negative correlation with abundance. Species richness was not explained by any variables.
As a whole, topographic type was the most directly related factor to chironomid assemblages.
Introduction
Horton (1945) developed stream order, a hierarchical classification system of streams, subsequently
modified by Strahler (1952, 1964), that remains in
wide use (Allan, 1995). That is, the smallest, permanently flowing stream is termed first order, and
the union of two streams of order n creates a stream
of order n þ 1. This provides a convenient system in
which rivers increase in order as they increase in size
(Allan, 1995). Vannote et al. (1980) used stream
order as physical template and proposed the River
Continuum Concept, a bold attempt to construct a
single synthetic framework to describe the function
of lotic ecosystems from source to mouth, and to
accommodate the variation among sites that results
from differences in their terrestrial setting. Particularly at headwater streams, condition of riparian
vegetation is likely to impact strongly on aquatic
communities and may be the single most important
anthropogenic factor influencing the structure,
function and water quality (Hynes, 1975; Sweeney,
1992). Recently, Roy et al. (2003) have shown that
increase in urban land cover resulted in less diverse
120
Figure 1. Schematic drawings showing three configuration
types of riffle (R)/pool (P) unit, longitudinal cross section.
Streams flow rightward: (a) type a (topographic type Aa); (b)
type b (Bb); (c) type c (Bc) (After Kani, 1944).
and more tolerant stream macroinvertebrate communities.
On the other hand, as a landscape scale
approach, Kani (1944) has proposed a system for
ecological classification of Japanese streams,
topographic type, by riffle/pool distribution and
configuration patterns as the following. The
distribution patterns can be divided into two types,
A and B. In type A, one reach from a meandering
point to the next meandering point comprises two
or more riffle/pool units. In type B, one reach
comprises a single unit. The unit configuration has
three types: a, b and c (Fig. 1). In type a, the riffle/
pool units are arranged as if in a series of stairs,
and the stream is more or less torrential and forms
small waterfalls. In type b, no waterfalls are found,
but riffles are partially wavy and foaming. In type
c, successive stream units are almost leveled, and
the water surface is usually waveless. Kani (1944)
found that distribution type A is always accompanied by the configuration type a, while B is
combined with either b or c. Therefore, three
binary combinations of distribution and configuration types are found: Aa, Bb and Bc. Even a
stream that at first sight does not seem well fit to
any one of the three can be classified as a transition in between Aa and Bb or Bb and Bc, and
Aa–Bb is frequently found in natural streams
(Fig. 2; Kani, 1944, 1981). Stream order system is
the most widely used in the world for describing
longitudinal position of streams (Allan, 1995).
However, the topographic type system has so far
been well established as one of the most useful
measure of riverine landscape features and used in
studies on distribution of lotic fishes and macroinvertebrates as well (Mizuno & Gose, 1972;
Okino, 2002; Ohgushi, 2004).
Chironomidae are the most widely distributed,
frequently the most abundant insects in freshwater
ecosystems (Pinder, 1986; Cranston, 1995), and the
Figure 2. Schematic drawings showing distribution types of riffle (scratches)/pool (dashed circles) unit, top view. Streams flow
downward: (a) type A (topographic type Aa); (b, c) transition types between types A and B (Aa–Bb); (d) type B (Bb); (e) type B (Bc).
(After Kani, 1944.)
121
most species-rich group in lotic macroinvertebrate
communities (Coffman, 1995). Chironomid species
diversity and assemblage structure have been
considered to be one of the most useful indicators
of various environmental factors, e.g. water quality (Wilson & Bright, 1973; Sæther, 1979; Wilson,
1987; Kawai et al., 1989), catchment urbanization
(Roy et al., 2003) and past environmental change
(Walker et al., 1991; Walker, 1995). Kawai &
Takahashi (1986) investigated chironomid fauna
of the Ohta River basin (Hiroshima, Japan) and
described longitudinal distribution of 96 species.
In our previous study, it has been shown that
intrageneric longitudinal zonations of the four
genera, i.e. Cricotopus, Polypedilum, Rheotanytarsus and Tanytarsus, were attributable to altitude,
specific conductance, topographic type and riverbed (riffle/pool) type (Kawai et al., 1998, 1999).
However, the relative importance of the factors
that compose river continuum on lotic chironomid
distribution remained uncertain.
The aim of the present study was to evaluate
influences of the macro-scale variables on larval
chironomid communities. Since substrate type is
an important micro-scale factor for macroinvertebrate community (e.g. Hynes, 1970; Minshall,
1984; Ruse, 1994; Allan, 1995; Lindegaard &
Brodersen, 1995), the condition was equalized by
the use of an artificial substrate for larval sampling. A direct ordination technique and stepwise
multiple regressions were used to determine which
variables explain the variations in abundance,
species richness, diversity, equitability and composition of the assemblages among sites. Finally,
usefulness of the artificial substrate and function
of river continuum on chironomids were discussed.
Materials and methods
Figure 3. Geographical position of the Ohta River basin and
locations of the 23 sampling sites (solid circles).
upper and middle reaches are often used as residential and/or agricultural land, but natural
deciduous forests are relatively rich, especially in
the north and west regions. The Hosomidani
riparian forest, locating around the origin of the
basin, is one of the most nature-rich districts in
Japan although now it has been becoming a social
concern as to conservation, due to plans of constructing a large scale traffic road. Conifer plantations which are mostly composed of Japanese
cedar (Cryptomeria japonica D. Don) occupy
many other forests. Contrariwise, areas surrounding the downstream and the river mouth,
where the river runs through central Hiroshima
City, are highly urbanized, and many buildings
and pavements cover the land entirely. In general,
water pollution has now been less severe except
downstream and river mouth, but physical habitat
modification and canopy removal which are
mainly caused by watercourse modification is still
the major disruption along the basin.
Sampling sites
Sampling methods
A total of 23 sampling sites in the Ohta River,
Hiroshima Prefecture, Japan (Fig. 3), were
selected, covering its entire basin with various
environmental conditions (Table 1). The main
stream originates in Mt. Kanmuri (1339 m a.s.l.)
of the western Chugoku Mountains and has
110 km of stretched main flow length and
1690 km2 of catchment area. Catchments of the
For quantitative sampling, a concrete construction
block (Fig. 4) was submerged in the stream, as a
standardized artificial substrate, at a riffle in each
site. The block is difficult to be flowed away and
provides sufficient microhabitats for various
species of chironomid larvae (Kawai et al., 1999).
The blocks were set between 5 and 15 July 1994
122
Table 1. Sampling records and the environmental variables for the 23 sampling sites in 1994
Site
Setting
Sampling
Altitude
(m a.s.l.)
a
Water
Surface
temperature
velocity
(C)
(ms )
)1
Depth
(cm)
Specific
Topographic
Canopy
Bank
conductance
typea
coverageb
protectionc
(lS cm )
)1
1
11 July
13 Oct.
560
17.8
0.25
20
50
2
2
0
2
7 July
13 Oct.
740
15.9
0.65
14
31
1
0
1
3
11 July
4 Oct.
770
15.6
0.42
20
35
2
0
2
4
5
14 July
14 July
4 Oct.
4 Oct.
700
720
13.7
16.2
0.16
0.93
50
30
41
39
1
2
2
0
0
2
6
14 July
4 Oct.
640
16.5
0.20
15
40
2
0
2
7
5 July
4 Oct.
640
15.5
0.41
23
29
1
0
1
8
11 July
15 Nov.
260
14.1
0.21
20
62
2
0
1
9
11 July
18 Oct.
230
14.2
0.45
9
60
2
0
2
10
7 July
15 Nov.
200
14.1
0.39
15
65
2
0
1
11
15 July
18 Oct.
100
15.5
0.49
9
46
2
1
1
12
13
7 July
7 July
4 Oct.
4 Oct.
180
220
12.1
19.9
0.41
0.65
5
22
42
88
1
1
2
1
1
1
14
7 July
4 Oct.
180
19.6
0.49
25
108
2
0
2
15
5 July
21 Nov.
50
14.2
0.57
35
61
2
0
2
16
15 July
4 Oct.
160
18.4
0.46
20
71
1
1
1
17
7 July
15 Nov.
100
13.3
0.36
7
77
1
0
2
18
15 July
15 Nov.
290
13.0
0.13
16
55
1
2
1
19
15 July
21 Nov.
360
10.8
0.45
24
74
1
2
1
20
21
15 July
15 July
21 Nov.
21 Nov.
180
250
12.1
11.0
0.53
0.27
24
8
98
36
1
1
0
2
1
0
22
5 July
21 Nov.
15
15.0
0.23
20
80
3
0
2
23
5 July
21 Nov.
10
14.5
0.53
6
71
3
0
2
1: Aa (headwater type); 2: Aa–Bb (upper-middle transitional reach type); 3: Bb (middle reach type), based on Kani (1944).
Coverage above the channel, 0: = 0%; 1: <50%; 2: >50%, based on Fujitani (2002).
c
Degree of protected bank in 50 m, 0: = 0%; 1: <50%; 2: >50%.
b
123
Figure 4. Concrete construction block, used as the standardized artificial substrate. The blocks were placed at riffles so that
the longer axis was parallel to the water flow.
(Table 1) and kept submerged on the riverbed for
at least 2 months. The collection was replicated
between 4 Oct. and 21 Nov. 1994 (Table 1). All
materials attached to and caught in the block, e.g.
moss, mud and litter which would be food for
larvae, were collected and transported to our laboratory with gentle cooling. The samples containing larvae were put into a container
(ø15 · H9 cm). The container was filled with
dechlorinated water to a level of about 6 cm,
covered with nylon net and aerated at room temperature (10–20 C). All emerging imagines were
collected every other day for at least a month and
for 2 months at the longest, until the emergence
ceased.
Identification to species
Only male imagines were used for identification
because of the difficulty in accurate identification
of females and immature stages (Inoue et al.,
2004). Male imagines were preserved as dried
specimens and mounted on microscopic slides with
gum chloral under a binocular dissecting microscope, following the method described by Sasa
et al. (1980). They were identified to species using
mainly the taxonomical keys provided by Pinder
(1978), Wiederholm (1989) and Sasa & Kikuchi
(1995).
Environmental assessments
At the same time of larval collection, the following
environmental conditions were measured at each
site: water temperature, depth, specific conductance, surface velocity, topographic type, canopy
coverage and degree of bank protection (Table 1).
Altitude was measured by reference to a 1:50 000
scale topographic map. Specific conductance was
measured using a salinometer (YSI model 33,
Yellow Springs Instrument Co., Inc. Ohio, USA),
as an indicator of water quality due to its high
sensitivity to a low level of trophic enrichment
(Sasa et al., 1980). Surface velocity above the
block was measured with a fishing float as a sign.
The topographic type of the stream was classified
on the basis of a criterion (Figs. 1 and 2) provided
by Kani (1944), as mentioned above. In Japanese
streams, types Aa, Bb and Bc are usually observed
in the mountainous headwaters, middle and lower
reaches, respectively, and type Aa–Bb is usually
observed as a transition between headwaters and
middle reaches. Canopy coverage and degree of
bank protection were measured by visual observation, both as indicators of vegetation richness
and anthropogenic modification of riparian terrain
(Yates & Noel, 1988; Fujitani, 2002).
In general, Japanese streams are extremely
short and steep compared to continental ones
(Okino, 2002). Combined with heavy rains, which
are brought by stationary front in early summer
and typhoons from late summer to autumn, many
basins have often suffered from flooding from way
back, and therefore bank protections were inevitably executed for flood control (Okino, 2002).
Two types of bank protection were observed,
crude and primitive one composed of irregular
natural rocks, and massive one in which banks
were covered firmly with concrete. The former was
frequently found in upper reaches and sometimes
difficult to be distinguished from natural rocky
banks. Therefore, only the latter ones were
regarded as protected banks in this study.
Data analysis
Preliminary to the analysis, chironomid assemblage data that contains less than 10 males were
excluded. In the beginning, a Detrended Correspondence Analysis (DCA; Hill, 1979) was used to
determine gradient length for the data set, which
indicated that unimodal rather than linear techniques were more appropriate (Ter Braak &
Prentice, 1988). Therefore, relationships between
the assemblage composition and the environmental variables were assessed by Canonical Correspondence Analysis (CCA; Ter Braak, 1986), using
124
the software CANOCO (version 4.5) for Windows
(Ter Braak & Šmilauer, 2002). CCA is a widely
used direct ordination technique that assumes a
unimodal model for the relationships between the
responses of each species to the ordination (environmental) axes (Johnson et al., 1993). A partial
CCA was then applied where the collecting month
was used as a covariable to eliminate temporal
effects on chironomid distribution pattern. The
forward selection option provided by CANOCO
was applied to the rest of the variables. Variables
with p < 0.05 (Monte-Carlo permutation test, 199
runs) were selected for the analysis. Scaling focused
on inter-species distances and no transformation of
data was made except that downweighting of rare
species option was set on. The significances
of the first ordination axis and of the sum of all
the canonical axes were evaluated with the
Monte-Carlo permutation test (199 runs).
Two community indices, Shannon diversity H0
(Shannon & Weaver, 1949) and Pielou evenness
component diversity (equitability) J (Pielou, 1969)
of the chironomid assemblage were calculated.
Then stepwise multiple regressions (forward
method) of these indices, species richness and total
abundance on the environmental variables were
performed using the software StatView (version
4.5) for Mac PowerPC (Abacus Concepts, Inc.
California, USA). F-to-enter and F-to-remove
were 4.000 and 3.996, respectively. All tests are
two-tailed with 5% a level.
Results
Table 2. A list of species and abundance collected by rearing of
larval samples
Species
Subfamily Tanypodinae
Conchapelopia pallidula (Meigen)
Paramerina divisa (Walker)
Rheopelopia eximia (Edwards)
Rheopelopia maculipennis (Zetterstedt)
Table 2 shows abundance of male imagines collected for each species. A total of 673 males
belonging to 69 species were recorded. The subfamilies Tanypodinae and Orthocladiinae, and
tribes Chironomini and Tanytarsini of the subfamily Chironominae composed 7.6, 19.0, 33.4 and
40.0% of total males, respectively. The most
abundant species was Chironomus yoshimatsui,
followed by Tanytarsus tamaundecimus and
Cladotanytarsus vanderwulpi. No species of the
subfamily Diamesinae were collected, and species
of Tanypodinae and Orthocladiinae were relatively rare, with the exception of the following 3
5
6
6
26
Rheopelopia ornata (Meigen)
7
Trissopelopia oyabetrispinosa Sasa,
1
Kawai et Ueno
Subfamily Orthocladiinae
Brillia japonica Tokunaga
Corynoneura celtica Edwards
Corynoneura lobata Edwards
Cricotopus (Cricotopus) bicinctus (Meigen)
Cricotopus (Cricotopus) bimaculatus
2
7
9
20
2
Tokunaga
Cricotopus (Cricotopus) metatibialis Tokunaga
3
Cricotopus (Cricotopus) tokunagai Hirvenoja
3
Cricotopus (Cricotopus) tremulus (Linnaeus)
Epoicocladius chuzeundecimus (Sasa)
2
1
Eukiefferiella sp.
1
Nanocladius quadrivittatus Niitsuma
2
Nanocladius seoulensis (Ree et Kim)
3
Nanocladius tamabicolor Sasa
3
Neobrillia longistyla Kawai
3
Orthocladius (Orthocladius) makabensis Sasa
2
Orthocladius (Orthocladius) tamarutilus Sasa
Parametriocnemus stylatus (Kieffer)
3
8
Paratrichocladius rufiventris (Meigen)
2
Psectrocladius aquatronus Sasa
1
Psectrocladius yunoquartus Sasa
Rheocricotopus chalybeatus (Edwards)
Chironomid data
Abundance
1
40
Rheocricotopus tamabrevis (Sasa)
1
Synorthocladius tamaparvulus Sasa
1
Thienemanniella nipponica Tokunaga
Thienemanniella vittata (Edwards)
1
2
Tvetenia tamaflava (Sasa)
5
Subfamily Chironominae
Tribe Chironomini
Chironomus flaviplumus Tokunaga
Chironomus yoshimatsui Martin et Sublette
Microtendipes britteni (Edwards)
7
100
5
Microtendipes pedellus (De Geer)
Microtendipes truncatus Kawai et Sasa
3
14
Polypedilum (Polypedilum) akisplendens
2
Kawai, Inoue et Imabayashi
Polypedilum (Polypedilum) asakawaense Sasa
1
125
Table 2. Continued.
Species
Polypedilum (Polypedilum) fuscovittatum
Abundance
3
Kawai, Inoue et Imabayashi
Polypedilum (Polypedilum) pedestre (Meigen)
4
Polypedilum (Polypedilum) takaoense Sasa
Polypedilum (Polypedilum) tamaharaki Sasa
4
20
Polypedilum (Polypedilum) tamahosohige Sasa
5
Polypedilum (Polypedilum) tamanigrum Sasa
7
Polypedilum (Polypedilum) tsukubaense Sasa
5
Polypedilum (Tripodura) unifascium (Tokunaga) 28
Polypedilum (Tripodura) sp. cf. asoprimum
1
Polypedilum (Uresipedilum) aviceps Townes
2
Polypedilum (Uresipedilum) convictum (Walker) 1
Polypedilum (Uresipedilum) cultellatum
1
Goetghebuer
Polypedilum (Uresipedilum) hiroshimaense
2
Five species of Rheotanytarsus were recorded and
all were relatively rare.
Species richness, male abundance, Shannon
diversity and equitability for each sampling site are
shown in Table 3. The abundance was the highest
at the site 22 followed by 23, both had the topographic type Bb and were located at the lowest
reach (Table 1, Fig. 3). Contrariwise, only low
abundances (n < 10) were recorded at the sites
2–4, 8, 10, 14, 19, which were usually located at the
upper reaches. Thus, assemblage data of these sites
were excluded for analyses hereafter.
Analysis of the chironomid assemblage composition:
CCA
Four variables were selected by CANOCO: topographic type, canopy coverage, water temperature
Kawai et Sasa
Polypedilum (Uresipedilum) paraviceps Niitsuma 2
Polypedilum (Uresipedilum) surugense Niitsuma 7
Polypedilum (Uresipedilum) tamasemusi Sasa
1
Tribe Tanytarsini
Cladotanytarsus vanderwulpi (Edwards)
65
Rheotanytarsus fluminis Kawai et Sasa
2
Rheotanytarsus rivulophilus Kawai et Sasa
5
Rheotanytarsus tamaquintus Sasa
6
Rheotanytarsus tamasecundus Sasa
2
Site Species
Rheotanytarsus tamatertius Sasa
4
Stempellinella tamaseptima (Sasa)
9
Tanytarsus takahashii Kawai et Sasa
Tanytarsus tamaduodecimus Sasa
3
16
Tanytarsus tamagotoi Sasa
17
Tanytarsus tamakutibasi Sasa
Tanytarsus tamaoctavus Sasa
2
6
Tanytarsus tamaundecimus Sasa
80
Virgatanytarsus arduennensis (Goetghebuer)
52
Total
Table 3. Species richness, total abundance and community
indices for chironomid assemblages at the 23 sampling sites
673
richness
Shannon
Pielou
abundance
diversity H0
equitability
(bit)
J
1
17
86
3.34
0.82
2*
3*
5
4
8
4
–
–
–
–
4*
3
3
–
–
5
5
23
1.47
0.63
6
8
17
2.44
0.81
7
8
17
2.13
0.71
8*
6
8
–
–
9
9
14
2.70
0.85
10* 3
11 8
5
15
–
2.74
–
0.91
12
8
12
2.62
0.87
13
7
10
2.72
0.97
8
–
–
14* 4
species: Rheopelopia maculipennis (8 sites, 26
males), Cricotopus bicinctus (2 sites, 20 males) and
Rheocricotopus chalybeatus (3 sites, 40 males). As
to Chironominae, no less than 18 species of the
genus Polypedilum were recorded whereas no more
than 3 species of one of its affined genera, Microtendipes, were recorded. A total of 14 species of
Tanytarsini were recorded and T. tamaundecimus
was the most dominant species, in turn
C. vanderwulpi and Virgatanytarsus arduennensis.
Total
15
6
15
1.87
0.72
16
16
55
3.40
0.85
17
8
11
2.85
0.95
18 11
19* 1
24
1
3.13
–
0.90
–
20
18
77
3.45
0.83
21
12
23
3.29
0.92
22
10
121
1.24
0.37
23
21
116
3.45
0.79
*Data excluded from analyses.
126
Table 4. Result of forward selection of environmental variables
used for partial CCA
Variable
F
p
Topographic type
1.911
0.010**
Canopy coverage
Water temperature
1.643
1.566
0.015*
0.030*
Altitude
1.495
0.035*
Bank protection
1.414
0.065
Surface velocity
1.366
0.140
Depth
1.189
0.245
Specific conductance
1.100
0.320
*p < 0.05, **p = 0.01 (Monte-Carlo permutation test).
and altitude, in order of the significance (Table 4).
The Monte-Carlo permutation test showed that
the analysis was significant (p ¼ 0.005) for both
the first axis and the sum of all the canonical axes.
Therefore, the first two axes were used for the
interpretation of the results. The first two axes
explained 28.0 and 71.0% of the variance in the
species data and species–environment relationship,
respectively (Table 5).
Topographic type was the environmental factor most strongly related to chironomid distribution, followed by canopy coverage, water
temperature and altitude (Table 4, Fig. 5). The
first canonical axis was interpreted as representing longitudinal gradient of the basin, since it
showed a positive association with altitude and
was negatively associated with topographic type
and with water temperature. Species associated
with the sites 22 and 23, e.g. C. bicinctus,
R. chalybeatus, Chironomus flaviplumus and
C. yoshimatsui were the more abundant species at
the Bb reaches and at higher temperature.
Cricotopus tokunagai, Psectrocladius yunoquartus,
Microtendipes britteni and Polypedilum aviceps
were abundant at the high-altitude sites.
Canopy coverage showed highly a negative
association with the canonical axis II. Therefore,
the axis was interpreted as representing the degree
of degradation of riparian vegetation and/or
anthropogenic disturbances. Corynoneura celtica,
Stempellinella tamaseptima and Polypedilum
tamanigrum, for which coordinates were located
on the lower part along the axis II, were regarded
as the abundant species at the sites with welldeveloped riparian canopy. All the species of
Rheotanytarsus were positively associated with
canopy coverage, whereas all the Cricotopus species showed negative responses in some degree.
Species, for which coordinates were located on the
center part of the graph, e.g. Polypedilum unifascium, C. vanderwulpi and T. tamaundecimus,
occurred in wider range of the environmental
conditions.
Analyses of the chironomid assemblage indices,
species richness and abundance: stepwise multiple
regression
Overall, which couples of variables were the most
strongly associated with chironomid community
indices? The results of the stepwise multiple
regressions are shown in Table 6. With respect to
Shannon diversity H0 , three variables were selected as predictor and the regression model was
significant (adj. R2 ¼ 0.507, p < 0.01). Bank
protection was the best explanatory variable and
showed a significant negative correlation with H0
(std. R ¼ )0.581, p < 0.01). Depth also showed
a
significant
negative
correlation
(std.
Table 5. Eigenvalues for partial CCA
CCA axes
I
II
III
IV
Eigenvalue
0.624
0.539
0.250
0.225
Species–environment correlations
0.979
0.931
0.918
0.939
15.0
38.1
28.0
71.0
34.0
86.2
Cumulative % variance
Of species data
Of species-environment relation
Total inertia = 4.586.
39.4
100.0
127
Figure 5. Partial CCA triplot showing the ordination of environmental variables (arrows), chironomid assemblages (open circles) and
species (crosses) on the axes I and II.
R ¼ )0.451, p < 0.05) while specific conductance
was positively correlated (std. R ¼ 0.389),
although the correlation was not significant. As
to Pielou equitability J, two variables were
selected and the model was significant (adj.
R2 ¼ 0.523, p < 0.005). Both topographic type
(std. R ¼ )0.621, p < 0.01) and depth (std.
R ¼ )0.436, p < 0.05) showed a significant negative correlation with J.
No significant regression model was found to
relate species richness to environmental variables
(data not shown). With regard to abundance,
three variables were selected and the model was
significant (adj. R2 ¼ 0.631, p < 0.005). Both
topographic type (std. R ¼ 0.835, p < 0.001)
and specific conductance (std. R ¼ 0.513,
p < 0.01) showed significant positive correlations whereas bank protection showed a significant negative correlation (std. R ¼ )0.517,
p < 0.05).
Discussion
From micro to macro-scale, many environmental
factors are known to influence distribution of
larval chironomid species and their local colonization (e.g. Wilson & Bright, 1973; Wilson, 1987;
Kawai et al., 1998; Ruse, 1994; Lindegaard &
Brodersen, 1995). In the present study, an artificial
substrate was placed at a riffle and used to eliminate differences in substrate condition. In addition
to topographic type, altitude, specific conductance, water temperature and depth, some riparian
conditions, i.e. canopy coverage and bank protection, were shown to influence community
composition.
The use of the artificial substrate
Chironomid assemblages, sampled by the use of
the artificial substrate, may be selective and are
128
Table 6. Result of stepwise multiple regression of Shannon diversity H¢, Pielou equitability J and total abundance on environmental
variables. Variables selected by forward method are shown in the table
Diversity (H0 )
Equitability (J)
Total abundance
R2
SE
df Adj. R2
Cumulative Total F
p
0.606
0.490
0.507
6.144
0.009**
Predictor variables
Partial R
SE
Std. partial R
F-to-remove
p
Bank protection
)0.576
0.185
)0.581
9.707
0.009**
Depth
Specific conductance
)0.035
0.013
0.014
0.006
)0.451
0.389
6.058
4.340
0.030*
0.059
Regression statistics
R2
SE
df Adj. R2
Cumulative Total F
p
0.586
0.102
9.216
0.003***
Predictor variables
Partial R
SE
Std. partial R
F-to-remove
p
Topographic type
)0.128
0.037
)0.621
12.130
0.004***
Depth
)0.070
0.003
)0.436
5.966
Regression statistics
R2
SE
df Adj. R2
0.705
23.478
Partial R
SE
Regression statistics
Predictor variables
Topographic type
Bank protection
Specific conductance
0.523
0.631
Std. partial R
0.030*
Cumulative Total F
p
9.539
0.002***
F-to-remove
p
44.906
9.845
0.835
20.805
<0.001***
)28.377
10.296
)0.517
10.106
0.017*
0.978
0.308
0.513
7.596
0.008**
Fin = 4.000; Fout = 3.996 (forward method); n = 16.
*p < 0.05, **p < 0.01, ***p < 0.005.
unlikely to represent natural communities (Glime
& Clemons, 1972). We would sample only a subset
of the entire chironomid community and many
species in these streams may not colonize on hard
substrates used in this study. In fact, only a small
numbers of males were collected at some of the
sites (Table 3) and 69 species were collected in
total from riffles only. Kawai & Takahashi (1986)
reported a total of 96 species from the Ohta River
basin, although they collected larvae at both riffles
and pools. A comparison of species composition
should be carried out between larval communities
inhabiting in situ bottom substrates of a certain
part of a riverbed and those inhabiting in an
artificial substrate placed at the same part, in order
to confirm the usefulness of the block as a substrate for lotic Chironomidae (Kawai et al., 1999).
Nevertheless, artificial substrates are useful for
comparing that subset of taxa across sites. The aim
of this study was to clarify the influence of macroscale factors, rather than micro-scale ones such as
substrate type. Even at an identical site, local
heterogeneity in substrate condition may cause
bias or distortion in assemblages that were sampled from natural substrates. On the other hand,
artificial substrates can provide a degree of sampling replicability not otherwise available, especially when they are placed in comparable
macrohabitats (Cairns & Pratt, 1993). In this respect, it was advantageous to apply the artificial
substrate to chironomid sampling in the present
study. Several micro-scale factors, e.g. depth and
surface velocity, still remained not identical among
the sites to some extent (Table 1), and therefore,
setting methods should be improved. In addition,
an interesting future study would be to use the
present methods to compare leaf pack and concrete block substrates.
129
Assemblage composition and species distribution of
chironomids
Partial CCA showed that chironomid assemblage
composition reflects topographic type, canopy
coverage, water temperature and altitude (Table 4,
Fig. 5). Kawai & Takahashi (1986) investigated
chironomid assemblages of the Ohta River and
reported that M. britteni, Brillia japonica, P. unifascium, were found only at Aa to Aa–Bb sites
whereas C. yoshimatsui and C. flaviplumus
occurred at only Bb to Bc sites. They also reported
that R. maculipennis, P. convictum, R. chalybeatus,
C. bimaculatus, M. truncatus and C. vanderwulpi
occurred at all the topographic type. On the other
hand, Kawai et al. (1998, 1999) showed that
P. hiroshimaense, P. takaoense, P. tamaharaki,
P. tamahosohige, P. tamanigrum, P. tsukubaense,
R. tamasecundus and T. tamaoctavus occurred
more frequently at Aa or Aa-Bb sites than Bb and
Bc sites. The present study largely supports the
results of Kawai & Takahashi (1986) and our
previous studies (Kawai et al., 1998, 1999) in
species distribution patterns along topographic
type and altitude.
Lindegaard & Brodersen (1995) have reviewed
faunistic studies which cover 104 sites in 80 different watercourses and showed the succession of
chironomid communities on altitude, latitude and
temperature. According to their criterion, the Ohta
River basin can be regarded as either of the following: lower mountain or lower montane streams
for the headwaters and upper to middle reaches,
and summer-warm lowland streams for the middle
to lower reaches. Compared with their results in
which percentages of Diamesinae and Orthocladiinae were 1–6 and 36–59%, respectively, it is
remarkable that Diamesinae was not collected and
Orthocladiinae was less abundant, composed
19.0% of total abundance (Table 2). These results
suggest that temperature fluctuations in the rearing room might have reduced the survival of cold
stenothermic
taxa,
especially
Diamesinae
(Lindegaard & Brodersen, 1995). However, imaginal community collected by light trap in a headwater of the same basin was composed of 11.6,
30.6, 33.9 and 24.0% of Tanypodinae, Orthocladiinae (excluding a terrestrial genus, Smittia
(Wiederholm, 1989)), Chironomini and Tanytarsini, respectively, and Diamesinae was not
collected (Kawai et al., 2002). These facts indicate
that the absence or rarity of Diamesinae and relative scarcity of Orthocladiinae in the basin, and
imply that larval rearing conditions could cause
small bias so that its influence on our general
conclusions could be substantially negligible.
The vector direction of water temperature in
the CCA graph was close to that of topographic
type, indicating that the ordination could not
clearly separate the influence of the two factors.
However, the association in vector direction
among topographic type, altitude and water temperature were consistent with longitudinal structure of river continuum (Vannote et al., 1980). On
the other hand, the impact of canopy coverage on
chironomid distribution was comparable to or
greater than the three factors mentioned above
(Fig. 5). It could be inferred that species distribution and community structure also reflects
anthropogenic factors, such as riparian forest
degradation.
There was a great variety of species distribution
pattern (Fig. 5). The assemblage composition for
the types Aa and Aa--Bb also showed some variations. On the other hand, assemblage compositions for the type Bb sites were quite similar and
some species occurred only at the sites. These
results may indicate that the environments of
upper reaches are more heterogeneous in microhabitats for chironomid larvae. However, possibilities that the species of the lower reaches are
ecological generalists and/or are more tolerant of
environmental variation cannot be discarded.
S. tamaseptima, P. tamanigrum and all the
species of Rheotanytarsus were positively associated with canopy coverage. Among these, larvae of
Rheotanytarsus are known as collector-filterers
and to construct conical catchnet for filtering food
materials from the water column (Pinder & Reiss,
1983; Berg, 1995). Rheotanytarsus can not
use large particles of litter (Coarse Particulate
Organic Matter: CPOM) directly, but other species, especially shredders colonizing in leaf packs,
should provide fine particles (Fine Particulate
Organic Matter: FPOM) that can be used by collector-filterers. For instance, a North American
stonefly, Pteronarcys scottii Hagen, known as a
shredder (Cummins et al., 1973; Short & Maslin,
1977), was reported to fragment CPOM as much
as 16% of its body weight per day (McDiffett,
130
1970). The other congeneric stonefly, P. californica
Newport, was shown to inhabit resulted in significant increase in FPOM (Short & Maslin, 1977).
By providing FPOM, shredders should contribute
to link litter with collectors (Cummins, et al.,
1989). The positive associations of Rheotanytarsus
with canopy coverage suggest that riparian vegetation functions as an allochthonous food source
for the species. In contrast, all the Cricotopus
species were negatively associated with canopy
coverage. The larvae are known as scrapers and to
feed on benthic and/or epiphytic algae (Cranston
et al., 1983; Berg, 1995). These results suggest that
riparian vegetation functions as a light barrier for
aquatic algae, resulting in effective decrease in
food source for Cricotopus larvae.
On the other hand, distributional patterns of all
the species with respect to riparian vegetation
cannot necessarily be associated with their feeding
mode. Some species were even reported to exhibit
considerable flexibility in larval feeding mode
(Berg, 1995). Hawkins et al. (1982) showed that
streams without any canopy shading had higher
abundances in the collector–gatherer, filter feeder,
herbivore shredder and piercer, and predator
guilds of total macroinvertebrate communities and
that Baetidae and Chironomidae were the most
abundant taxa in each of guilds. Considering that
there was no significant correlation between chironomid assemblage composition and specific
conductance, riparian vegetation could merely
function as a refuge of imagines for some species,
rather than as allochthonous food source. Riparian
canopy prevents the imagines from wind, dehydration and predators, and provides appropriate
swarming and oviposition sites (Downes, 1969;
Sweeney, 1993; Harrison et al., 2000). Structurally
complex riparian vegetation leads to a greater
supply of insect recruits to all habitats (Glime &
Clemons, 1972; Mauer & Brusven, 1983).
In this study, larval sampling was practiced
between Oct. and Nov. (Table 1). More univoltine
species might be collected if it was repeated in
springtime and other seasons.
Abundance, species richness, diversity and
equitability of chironomids
Shannon diversity H0 showed a significant negative
correlation with bank protection and depth while
it showed a positive correlation with specific conductance although the correlation was not significant (Table 6). Bank protection may be associated
with watercource modification, which in turn
reduces habitat heterogeneity. In general, concrete
bank protections are executed by using large
construction machinery into stream and riparian
terrain. For the purpose of flood control, dredging
of the channel and/or even straightening of stream
meandering are often accompanied (Mizuno &
Gose, 1972). Moreover, the possibility that bank
protection not only decreases in water movement
between surface water and hyporheos, but also
inhibits growth in structurally complex herbaceous
vegetation along bankside cannot be discarded. In
fact, Harrison & Harris (2002) showed that taxon
richness and Shannon diversity H0 of benthic
communities were significantly higher in ungrazed
streams compared with grazed ones by cattle
breeding, meaning that simplified bankside vegetation significantly decreases richness and diversity
of benthic communities. Future study should be
done to clarify if bank protection negatively
associates with complexity of bankside vegetation.
Both topographic type (lower reach) and depth
showed a significant negative correlation to J
(Table 6), indicating that lower reaches of the river
or deeper sites tend to be dominated by a small
number of species.
However, no predictor variable was selected
for species richness. The result suggests that
richness is a more or less stochastic variable that
reflects local variation in species composition,
rather than environmental factors. In fact, partial CCA showed that assemblage compositions
of some neighboring sites were relatively resemble each other (e.g. sites 5, 6 and 7; 11, 13 and
16; 18 and 21; 22 and 23) and vice versa.
Therefore, richness itself may not be a good
indicator of chironomid community structure.
Nevertheless, the negative correlation between H0
and bank protection, and that the number of
species positively associated with canopy coverage was larger than that of species negatively
associated with canopy coverage, suggesting a
substantial influence of these factors on chironomid richness. Comparisons of richness among
assemblages in which the local variation in
richness is minimized should be needed to clarify
the influence.
131
Both topographic type and specific conductance showed significant positive correlations with
abundance, whereas bank protection showed a
significant negative correlation (Table 6). Kawai
et al. (1989) showed a significant positive correlation between abundance and BOD. A possible
explanation to this is that trophic enrichment in
lower reaches may result in massive occurrence of
nuisance species. In fact, C. bicinctus and C. yoshimatsui, both reported to emerge in massive
swarms from polluted waters (Tabaru et al., 1987),
were collected in large numbers only at the two
lowest sites (Table 2, Fig. 5).
Surface velocity was not a predictor variable of
any analyses in the present study. This supports
the results in Peckarsky & Penton (1990) and Ruse
(1994) that surface current on the substrate was a
poor predictor of the interstitial flows experienced
by chironomid larvae.
Acknowledgements
We are very grateful to Mr. Shin-ichiro Nakama
and Ms. Keiko Anami, formerly Tokushige, both
former collaborators, for their kind help in chironomid collection. We also thank the two anonymous reviewers for their helpful comments and
suggestions on the early manuscript. Figures 1 and
2 were redrawn from Kani (1944). We regret his
early death in the same year in the World War II
and acknowledge permission of the publisher,
Kenkyusha Co., Inc. The senior author (E. I.)
would like to thank Mr. Hiroshi Ashiwa of Kyoto
University, Mr. Seiju Inaba and Ms. Fumie
Kikkawa for their constant encouragement which
inspired me to complete this work.
References
Conclusion
As a whole, the variable most highly associated
with chironomid assemblages was topographic
type. Longitudinal species succession has been
shown in overall benthic communities and in many
taxa (Hynes, 1970; Kawai & Takahashi, 1986;
Kawai et al., 1998, 1999; Fujitani, 2002). In addition, the significant correlation of the assemblages
with altitude and water temperature in partial
CCA and its vector direction also emphasize the
longitudinal gradient. The river continuum concept predicts changes in structure and function of
stream habitats along a transition from headwaters
to lower reaches (Vannote et al., 1980). Increase in
the other variables associated with anthropogenic
modification, such as bank protection and simplification of riparian vegetation, could result in a
reduction in the richness or number of intolerant
or endemic species. Moreover, the present study
was based on the use of the standardized artificial
substrate, indicating that riparian vegetation and
bank protection, other than the association between the river continuum and substrate type, also
function as determinants of larval chironomid
community composition. Our results are also
consistent with other studies indicating impacts of
riparian terrain on macroinvertebrate communities
(Harrison & Harris, 2002; Roy et al., 2003).
Allan, J. D., 1995. Stream Ecology. Structure and function of
running waters. Chapman & Hall, London.
Berg, M. B., 1995. Larval food and feeding behaviour. In
Armitage, P. D., P. S. Cranston, & L. C. V. Pinder (eds), The
Chironomidae: Biology and Ecology of Non-biting Midges.
Chapman & Hall, London: 136–168.
Cairns, J. Jr. & J. R. Pratt, 1993. A history of biological monitoring using benthic macroinvertebrates. In Rosenberg, D. M.
& V. H. Resh (eds), Freshwater Biomonitoring and Benthic
Macroinvertebrates. Chapman & Hall, New York: 10–27.
Coffman, W. P., 1995. Conclusions. In Armitage, P. D.,
P. S. Cranston, & L. C. V. Pinder (eds), The Chironomidae:
Biology and Ecology of Non-biting Midges. Chapman &
Hall, London: 436–447.
Cranston, P. S., D. R. Oliver & O. A. Sæther, 1983. The larvae
of Orthocladiinae (Diptera: Chironomidae) of the Holarctic
region. Keys and diagnoses. In Wiederholm, T. (ed.), Chironomidae of the Holarctic Region. Keys and Diagnoses.
Part 1. Larvae. Entomologica Scandinavica Supplement 19:
149–291.
Cranston, P. S., 1995. Introduction. In Armitage, P. D., P. S.
Cranston, & L. C. V. Pinder (eds), The Chironomidae:
Biology and Ecology of Non-biting Midges. Chapman &
Hall, London: 1–7.
Cummins, K. W., R. C. Petersen, F. O. Howard, J. C.
Wuycheck & V. I. Holt, 1973. The utilization of leaf litter by
stream detritivores. Ecology 54: 336–345.
Cummins, K. W., M. A. Wilzbach, D. M. Gates, J. B. Perry &
W. B. Taliaferro, 1989. Shredders and riparian vegetation.
Bioscience 39: 24–30.
Downes, J. A., 1969. The swarming and mating flight of Diptera. Annual Review of Entomolology 14: 271–298.
Fujitani, T., 2002. Species composition and distribution patterns of baetid nymphs (Baetidae: Ephemeroptera) in a
Japanese stream. Hydrobiologia 485: 111–121.
132
Glime, J. M. & R. M. Clemons, 1972. Species diversity of
stream insects on Fontinalis spp. compared to diversity on
artificial substrates. Ecology 65: 455–468.
Harrison, S. S. C., I. T. Harris, A. Croeze & R. Wiggers, 2000.
The influence of bankside vegetation on the distribution of
aquatic insects. Verhandlungen der Internationalen Vereinigung für Theoretische und Angewandte Limnologie 27:
1480–1484.
Harrison, S. S. C. & I. T. Harris, 2002. The effects of bankside
management on chalk stream invertebrate communities.
Freshwater Biology 47: 2233–2245.
Hawkins, C. P., M. L. Murphy & N. H. Anderson, 1982. Effects
of canopy, substrate composition, and gradient on the
structure of macroinvertebrate communities in cascade range
stream of Oregon. Ecology 63: 1840–1856.
Hill, M. O., 1979. DECORANA, a FORTRAN Program for
Detrended Correspondence Analysis and Reciprocal Averaging. Section of Ecology and Systematics, Cornell University, New York.
Horton, R. E., 1945. Erosional development of streams and
their drainage basins: hydrophysical approach to quantitative morphology. Bulletin of the Geological Society of
America 56: 275–350.
Hynes, H. B. N., 1970. The Ecology of Running Waters.
Liverpool University Press, Liverpool.
Hynes, H. B. N., 1975. The stream and its valley. Verhandlungen der Internationalen Vereinigung für Theoretische und
Angewandte Limnologie 19: 1–15.
Inoue, E., K. Kawai & H. Imabayashi, 2004. A new species of
the genus Stempellinella (Diptera: Chironomidae) from
Hiroshima, Japan. Limnology 5: 141–147.
Johnson, R. K., T. Wiederholm & D. M. Rosenberg, 1993.
Freshwater biomonitoring using individual organisms, populations, and species assemblages of benthic macroinvertebrates. In Rosenberg, D. M. & V. H. Resh (eds), Freshwater
Biomonitoring and Benthic Macroinvertebrates. Chapman
& Hall, New York: 40–158.
Kani, T., 1944. Ecology of torrent-inhabiting insects. In
Furukawa, H. (ed.), Insects, I. Kenkyusha, Tokyo: 171–317
(in Japanese).
Kani, T., 1981. Stream classification in ‘‘Ecology of torrentinhabiting insects’’ (1944): An abridged translation. Physiology and Ecology Japan 18: 113–118.
Kawai, K. & M. Takahashi, 1986. Distribution of larvae of
non-biting midges chironomids in the Ohta River. Japanese
Journal of Limnology 47: 307–314 (in Japanese with English
abstract).
Kawai, K., E. Inoue & H. Imabayashi, 1998. Intrageneric
habitat segregations among chironomid species of several
genera in river environments. Medical Entomology and
Zoology 49: 41–50.
Kawai, K., E. Inoue & H. Imabayashi, 1999. Differences in
occurrence patterns in relation to three environmental
factors among the lotic chironomid species of a genus,
Polypedilum. Medical Entomology and Zoology 50:
233–242.
Kawai, K., E. Inoue & H. Imabayashi, 2002. Temporal changes in
male chironomid midges attracted to black-light in the Yoshiki
River. Medical Entomology and Zoology 53: 281–284.
Kawai, K., T. Yamagishi, Y. Kubo & K. Konishi, 1989. Usefulness of chironomid larvae as indicators of water quality.
Japanese Journal of Sanitary and Zoology 40: 269–283.
Lindegaard, C. & K. P. Brodersen, 1995. Distribution of Chironomidae (Diptera) in the river continuum. In Cranston, P. S.
(ed.), Chironomids: From Genes to Ecosystems. CSIRO,
Melbourne: 257–271.
Mauer, M. A. & M. A. Brusven, 1983. Invertebrate abundance and colonization rate in Fontinalis neomexicana
(Bryophyta) in an Idaho Batholith stream, USA. Hydrobiologia 98: 9–16.
McDiffett, W. F., 1970. The transformation of energy by a
stream detritivore, Pteronarcys scotii (Plecoptera). Ecology
51: 975–988.
Minshall, G. W., 1984. Aquatic insect–substratum relationships. In Resh, V. H. & D. M. Rosenberg (eds), The Ecology
of Aquatic Insects. Praeger Publishers, New York: 358–400.
Mizuno, N. & K. Gose, 1972. The Ecology of Streams. Tsukiji
Shokan, Tokyo (in Japanese).
Ohgushi, R., 2004. The World of Aquatic Insects. Lives Connecting Freshwater and Land. Tokai University Press,
Kanagawa (in Japanese).
Okino, T., 2002. The Ecology of Streams. Kyoritsu Shuppan,
Tokyo (in Japanese).
Peckarsky, B. L. & M. A. Penton, 1990. Effects of enclosures on
stream microhabitat and invertebrate community structure.
Journal of the North American Benthological Society 9:
249–261.
Pielou, E. C., 1969. An Introduction in Mathematical Ecology.
Wiley-Interscience, New York.
Pinder, L. C. V., 1978. A key to adult males of British Chironomidae. Vol. 1 The key, Vol. 2 Illustrations of the hypopygia. Freshwater Biological Association Scientific
Publication 37: 1–169, Figs. 77–184.
Pinder, L. C. V., 1986. Biology of freshwater Chironomidae.
Annual Review of Entomology 31: 1–23.
Pinder, L. C. V. & F. Reiss, 1983. The larvae of Chironominae
(Diptera: Chironomidae) of the Holarctic region. Keys and
diagnoses. In Wiederholm, T. (ed.), Chironomidae of the
Holarctic region. Keys and diagnoses. Part 1. Larvae. Entomologica Scandinavica Supplement 19: 293–435.
Roy, A. H., A. D. Rosemond, M. J. Paul, D. S. Lei &
J. B. Wallace, 2003. Stream macroinvertebrate response to
catchment urbanization (Georgia, U.S.A.). Freshwater
Biology 48: 329–346.
Ruse, L. P., 1994. Chironomid microdistribution in gravel of an
English chalk river. Freshwater Biology 32: 533–551.
Sæther, O. A., 1979. Chironomid communities as water quality
indicators. Holarctic Ecology 2: 65–74.
Sasa, M. & M. Kikuchi, 1995. Chironomidae of Japan. University of Tokyo Press, Tokyo.
Sasa, M., M. Yasuno, M. Ito & T. Kikuchi, 1980. Studies on
chironomid midges of the Tama River. Part 1. The distribution of chironomid species in a tributary in relation to the
degree of pollution with sewage water. Research Report from
the National Institute for Environmental Studies, Japan 13:
1–8.
Shannon, C. E. & W. Weaver, 1949. The Mathematical Theory
of Communication. University of Illinois Press, Chicago.
133
Short, R. A. & P. E. Maslin, 1977. Processing of leaf litter by a
stream detritivore – effect on nutrient availability to collectors. Ecology 58: 935–938.
Strahler, A. N., 1952. Hypsometric (area–altitude) analysis of
erosional topography. Bulletin of the Geological Society of
America 63: 1117–1142.
Strahler, A. N., 1964. Quantitative geomorphology of drainage
basins and channel networks; section 4-II. In Chow, Van te
(ed.), Handbook of Applied Hydrology: A Compendium of
Water-resources Technology. McGraw-Hill, New York:
4.39–4.76.
Sweeney, B. W., 1992. Streamside forests and the physical,
chemical, and trophic characteristics of Piedmont streams in
eastern North America. Water Science and Technology 26:
2653–2673.
Sweeney, B. W., 1993. Effects of streamside vegetation on
macroinvertebrate communities of White Clay Creek in
eastern North America. Proceedings of the Academy of
Natural Sciences of Philadelphia 144: 291–340.
Tabaru, Y., K. Moriya & A. Ali, 1987. Nuisance midges
(Diptera: Chironomidae) and their control in Japan. Journal
of American Mosquito Control Association 3: 45–49.
Ter Braak, C. J. F., 1986. Canonical correspondence analysis: a
new eigenvector technique for multivariate direct gradient
analysis. Ecology 67: 1167–1179.
Ter Braak, C. J. F. & I. C. Prentice, 1988. A theory of gradient
analysis. Advances in Ecological Research 18: 271–317.
Ter Braak, C. J. F. & P. Šmilauer, 2002. CANOCO reference
manual and CanoDraw for Windows user’s guide: Software
for canonical community ordination (version 4.5). Microcomputer Power, New York.
Vannote, R. L., G. W. Minshall, K. W. Cummins, J. R. Sedell &
C. E. Cushing, 1980. The river continuum concept. Canadian
Journal of Fisheries and Aquatic Sciences 37: 130–137.
Walker, I. R., 1995. Chironomids as indicators of past environmental change. In Armitage, P. D., P. S. Cranston, & L. C.
V. Pinder (eds), The Chironomidae: Biology and Ecology of
Non-biting Midges. Chapman & Hall, London: 405–422.
Walker, I. R., J. P. Smol, D. R. Engstrom & H. J. B. Birks,
1991. An assessment of Chironomidae as quantitative indicators of past climatic change. Canadian Journal of Fisheries
and Aquatic Sciences 48: 975–987.
Wiederholm, T. (ed.), 1989. Chironomidae of the Holarctic
region. Keys and diagnoses. Part 3. Adult males. Entomologica Scandinavica Supplement 34: 1–532.
Wilson, R. S. & P. L. Bright, 1973. The use of chironomid
pupal exuviae for characterizing streams. Freshwater Biology 3: 283–302.
Wilson, R. S., 1987. Chironomid communities in the River
Trent in relation to water chemistry. In Sæther, O. A. (ed.),
A Conspectus of Contemporary Studies in Chironomidae
(Diptera). Contributions from the IXth symposium on
Chironomidae, Bergen, Norway. Entomologica Scandinavica
Supplement 29: 387–393.
Yates, S. & S. Noel, 1988. Adopting a stream. A Northwest
Handbook. University of Washington Press, Washington.