ChapterPDF Available

Relationships Between Bird Density, Vegetation Characteristics, and Grasshopper Density in Mixed-Grass Prairie of Western North Dakota

Authors:

Abstract

We monitored the abundance of grassland birds and the density of then-principal prey (grasshoppers, Acrididae) on 20 transects in mixed-grass prairie of western North Dakota from 1987-1990. Bird and grasshopper densities were estimated on 10 occasions during this period, 4 times in 1987 and once in late May and June every year thereafter. Twelve transects were treated for grasshopper control in late June 1987. During all of the census periods except July 1987 and June 1990, there was a negative relationship between grasshopper density and total bird density; the relationship was significant (P>0.05) during June 1987 and May 1989. Of the four most common bird species recorded on the transects, 2 species, western meadowlarks (Sturnella neglecta) and grasshopper sparrows (Ammodramus savannarum) ,were negatively correlated with grasshopper densities during most of the census periods. The relationship was significant for western meadowlarks in June 1987, May 1989, and June 1990. However, density of the other 2 species, horned larks (Eremophila alpestris) and vesper sparrows (Pooecetes gramineus) ,were not correlated with grasshopper density. We explore two possible explanations for the negative relationship between bird abundance and grasshopper abundance: habitat association and bird predation. Available evidence supports both of these nonexclusive hypotheses. However, it is unlikely that bird predation has a significant impact when grasshopper densities are high. Our results suggest that habitat manipulation may be effective in decreasing grasshopper numbers and increasing bird populations in these grasslands.
Wiley
is collaborating with JSTOR to digitize, preserve and extend access to
Ecology.
http://www.jstor.org
Effects of Avian Predation on Grasshopper Populations in North Dakota Grasslands
Author(s): Ada C. Fowler, Richard L. Knight, T. Luke George and Lowell C. McEwen
Source:
Ecology,
Vol. 72, No. 5 (Oct., 1991), pp. 1775-1781
Published by: Wiley
Stable URL: http://www.jstor.org/stable/1940976
Accessed: 20-03-2016 18:57 UTC
REFERENCES
Linked references are available on JSTOR for this article:
http://www.jstor.org/stable/1940976?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
Ecology, 72(5), 1991, pp. 1775-1781
? 1991 by the Ecological Society of Amenica
EFFECTS OF AVIAN PREDATION ON GRASSHOPPER
POPULATIONS IN NORTH DAKOTA GRASSLANDS'
ADA C. FOWLER, RICHARD L. KNIGHT, T. LuKE GEORGE,
AND LOWELL C. MCEWEN
Department of Fishery and Wildlife Biology, Colorado State University,
Fort Collins, Colorado 80523 USA
Abstract. We experimentally studied the effects of avian predation on grasshopper
abundance in western North Dakota during the summers of 1988 and 1989. Grasshopper
densities in 15 x 15 m plots from which birds were excluded (NO BIRDS) were compared
with similar-sized plots where birds were allowed to forage (BIRDS). Plots were established
in early June at 16 sites (8 per year), and grasshopper densities were estimated from hoop
counts in NO BIRDS and BIRDS plots at 2-wk intervals until the end of July. There were
significantly more grasshoppers in NO BIRDS plots than in BIRDS plots in 1989 (P <
.0001), but not in 1988 (P .137). The difference between treatments in 1989 was first
detected 6 wk after the exclosures were erected. Between years there were no differences
in initial grasshopper abundance in all treatments (P > .388). After the final hoop count
each year, sweep-net sampling was also used to estimate grasshopper densities. By this
method, average grasshopper density in late July was 26% and 37% lower in BIRDS plots
than in NO BIRDS plots in 1988 and 1989, respectively. Average length, total biomass of
grasshoppers, species richness, and species diversity, however, did not differ between the
treatments. In 1988, 2 of 15 grasshopper species were significantly more abundant in the
NO BIRDS plots. There was no difference between the treatments among 16 species
identified in the 1989 samples, but the power of our tests to detect differences for individual
species was low. Our results support the hypothesis that avian predation reduces insect
populations at low and moderate densities.
Key words: Acrididae; avian predation; experimental enclosure; grasshopper populations; grass-
lands; integrated pest management; North Dakota.
INTRODUCTION
The effects of avian predation on insect population
dynamics are not well understood. Some investigators
have suggested that avian predators are important in
regulating insect populations, especially at endemic and
moderate levels of insect abundance (Otvos 1979, Ta-
kekawa et al. 1982), while others have suggested avian
predators have no significant impact (Wiens 1973, 1974,
Wiens and Rotenberry 1979, Rotenberry 1980a, b).
The few experimental studies that have investigated
the impact of avian predators on insect populations
have shown that birds may depress insect populations
(Holmes et al. 1979, Joern 1986, Greene 1989).
Effects of avian predators on insect populations in
rangeland habitats have not been well documented
(McEwen 1987). Based on bioenergetics models, Wiens
and Dyer (1975) and Rotenberry (1980a) concluded
that avian predation had no significant impact on ar-
thropod populations. Models used by Rotenberry
(1 980a), for example, indicated that birds took <0.7 7%
of the arthropod standing crop per day, even though
arthropod prey comprised 80% of the predicted bio-
I Manuscript received 15 May 1990; revised 12 December
1990; accepted 19 December 1990.
mass consumption by birds (Wiens and D yer 1975).
Wiens (1973), however, suggested that avian predation
may dampen oscillations of certain insect groups such
as grasshoppers.
In contrast to previous predictions, Joern (1986) and
C. E. Bock (personal communication) found that bird
predation reduced grasshopper density by 30-50%.
Also, species diversity of grasshoppers in plots from
which birds were excluded was significantly higher.
Many grassland birds preferentially select grasshoppers
as prey; Baldwin (1972) found that while grasshoppers
comprised 10.1 /% of available insects, they made up
48.3% of Lark Buntings' (Calamospiza melanocorys)
diet. Boyd (1976) estimated that a pair of Horned Larks
(Eremophila alpestris) consumed 156 grasshoppers per
day in a short grass plains habitat. Clearly, additional
investigations of the interrelationships between avian
predators and prey populations are needed (Wiens and
Dyer 1975, Joern 1986, Price 1987).
Our objective was to determine the effects of avian
predators on grasshopper abundance in North Dakota
grasslands. Using an enclosure experiment, we tested
the null hypothesis that there were no effects of avian
predation on grasshopper abundance against the alter-
native hypothesis that avian predation reduced grass-
hopper abundance.
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
1776 ADA C. FOWLER ET AL. Ecology, Vol. 72, No. 5
METHODS AND ANALYSES
Methods
We conducted this study during the summers of 1988
and 1989 in the Little Missouri National Grasslands,
McKenzie County, North Dakota (Fig. 1). The vege-
tation is a mixed-grass prairie dominated by blue gra-
ma (Bouteloua gracilis), western wheatgrass (Agropy-
ron smithii), needle and thread (Stipa comata), prairie
junegrass (Koeleria pyramidata), and threadleaf sedge
(Carexfilifolia). The grasslands are dissected by ravines
that are wetter than the surrounding prairies and sup-
port some shrubs, such as silver buffaloberry (Shep-
herdia argentea), choke cherry (Prunus virginiana),
snowberry (Symphoricarpos occidentalis), several Ar-
temisia species, and trees, principally green ash (Frax-
inus pennsylvanica) and cottonwood (Populus del-
toides). Using United States Forest Service Range Maps
(McKenzie District Office) and field reconnaissance,
we selected 16 study sites (8 per year) that were native
prairie (Fig. 1). To avoid interference with or destruc-
tion of plots by livestock, summer-grazed areas were
not used for study sites. Excluding livestock from ex-
closed areas during the summer would have created
islands of ungrazed habitat within grazed areas, which
might have affected grasshopper densities. All sites were
grazed by livestock in the winter. Sites ranged between
1.5 and 49 km apart.
At each site we chose an area of homogeneous grass-
land vegetation to include three 15 x 15 m plots. Each
plot was at least 25 m from the other two and 100 m
from any ravine. Treatments were randomly assigned
to the plots as NO BIRDS, SIDES, or OPEN. NO
BIRDS plots were constructed of 1 m high chicken-
wire (2.5-cm mesh) sides attached to metal fence posts.
The plots were covered with a black polypropylene 2.5-
cm mesh roof to exclude birds. SIDES plots were sim-
ilarly constructed but without the mesh roof. OPEN
plots were marked only with corner stakes. SIDES and
OPEN plots will be collectively referred to as "BIRDS."
We constructed both SIDES and OPEN plots to de-
termine if chicken wire affected grasshopper move-
ment. In 1988, seven replicates of the three types of
plots were erected between 2 and 8 June, and the last
replicate was erected on 18 June. In 1989 all eight
replicates were erected between 4 and 8 June.
We estimated grasshopper densities using 0.25-rM2
aluminum hoops (Onsager and Henry 1977). In each
plot 16 hoops were positioned in four rows 3 m apart.
We counted the number of grasshoppers within each
hoop as they flushed at the observer's approach. Veg-
etation within each hoop was brushed to flush any
remaining grasshoppers; all counts were made by A.
Fowler. Plots were established in early June when cows
were taken off the winter range. This corresponds to
the peak of hatching for first broods of common grass-
land birds in this area (T. L. George, A. C. Fowler, R.
L. Knight, and L. C. McEwen, unpublished manui-
script). Initial counts were taken when the plots were
N
Watford
City
* ~~~~~~~0 10 20
A ~~~~~~km
North
Study Sites Dakota
A 1989 z Little Missouri -
* 1988 National Grasslands
FIG. 1. Location of study sites in McKenzie County, North
Dakota.
established, and subsequent counts were made 2 wk
apart on 20 June, 5 July, and 19 July 1988, and 21
June, 4 July, and 19 July 1989. The final counts were
made between 21 and 29 July 1988 and 24 and 30 July
1989 after the common grassland bird species had fin-
ished nesting (T. L. George, A. C. Fowler, R. L. Knight,
and L. C. McEwen, unpublished manuscript). In 1988
seven of eight study sites were used in the count anal-
ysis; the eighth site was erected 2 wk after the other
sites, so it was dropped from the count analysis. In
1989 all eight study sites were used in the count anal-
ysis.
We used a second method (Joern 1986, G. E. Be-
lovsky, personal communication) to sample grasshop-
per densities at the end of the study each year. Heavy
plastic or fiberglass screen (1 m high) was stretched
around the outside of each plot to minimize grasshop-
per movement in and out of the plot; the sides on the
NO BIRDS plots and the SIDES plots were left stand-
ing. The mesh roof was removed from the NO BIRDS
plots, and the number of grasshoppers in the hoops
was counted on all plots as described above. Next,
grasshoppers within each plot were collected with an
insect sweep net. Sweep-net sampling represents the
minimum number of grasshoppers per plot. In 1988
A. Fowler systematically swept entire plots until < 10
individuals were collected on a single pass through the
plot. All grasshoppers collected at a plot were com-
bined into a sample. In 1989 the same sampling meth-
od was used as in 1988 except two or three observers,
instead of one, sampled each plot. Observers were ran-
domly assigned to a plot and completed one pass
through the plot before moving to the next plot. Each
plot was systematically swept 4-6 times. The duration
of the sampling at each plot at a site was kept the same.
Grasshoppers in sweep samples were frozen, counted,
and measured, and the adults were identified to species.
Grasshopper species richness was the number of grass-
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
October 1991 AVIAN PREDATION ON GRASSHOPPERS 1777
-1'r a
' 3 - 1988
NO BIRDS
C 2 - ---BIRDS
0~
cc 5 June 20 June 5 July 19 July 25 July
b
E -5 -1989 *
c4
7 3.
0 --I ---' ------I
6 June 21 June 4 July 19 July 27 July
Date of count
FIG. 2. Average grasshopper abundance (X ? 1 SE) esti-
mated from hoop counts in 1988 (a) and 1989 (b). Asterisks
denote significant differences (P < .05) between plots with
bird predation (BIRDS, N= 14 in 1988, N= 16 in 1989) and
without bird predation (NO BIRDS, N = 7 in 1988, N = 8
in 1989).
hopper species per plot. Species diversity was calcu-
lated as -2 ln(pdpi, where pi is the proportion of each
species in the sample. The grasshopper samples from
each plot were oven-dried at 650C, and weighed.
In 1989 we recorded plant species and cover data
after removing the plot sides. At each plot the plant
species composition was measured using 20 x 50 cm
Daubenmire frames (Daubenmire 1959). We system-
atically sampled 56 frames (four rows of 14) at each
plot. Species were grouped into six cover categories at
each frame: 1-5%, 6-25%, 26-50%, 51-75%, 76-95%,
and 96-100% (Daubenmire 1959). The height of veg-
etation was measured at five standard points within
each frame.
Analyses
Repeated-measures ANOVA was used to analyze the
hoop count data because sequential counts were not
independent (Beal and Khamis 1990). Randomized
complete block ANOVA, blocked on sites, was used
to analyze sweep-net abundance estimates, biomass,
length, species richness, species diversity, and species
composition of grasshoppers in plots. P values pre-
sented for count and sweep-net estimates are one-tailed
because we predicted that NO BIRDS plots would have
a greater density of grasshoppers than BIRDS plots.
Sweep-net and final hoop count estimates were com-
pared using a paired t test. We tested the effect of year
on grasshopper densities with a 2 x 2 factorial ANO-
VA design (main effects: year and treatments) using
the initial and final counts. Sweep-net samples, hoop
counts, biomass, length, and species composition of
grasshoppers in plots were transformed using logO(X
+ 1) to meet normality and homogeneous variance
assumptions of parametric analysis. The treatment ef-
fect from the ANOVA was analyzed further using 1 -df
contrasts, SIDES vs. OPEN plots, and NO BIRDS vs.
BIRDS plots. Because of the categorical nature of the
species composition of plants in plots, and the impos-
sibility of normalizing the vegetation height data, these
data were analyzed by site using rank transformation.
All analyses were conducted using SAS (1988). The
significance level was P < .05.
RESULTS
There were no significant differences in grasshopper
densities between SIDES and OPEN plots (P > .2, for
all comparisons), so these treatments were combined
using a linear contrast and are referred to as "BIRDS."
Density of grasshoppers in the hoop counts ranged from
0 to 4 insects/M2 in 1988 and 0 to 10.5 insects/M2 in
1989. For sweep-net samples, density ranged from 0.04
to 1.25 grasshoppers/M2 in 1988 and 0.16 to 4.13 grass-
hoppers/M2 in 1989.
Overall during 1988, there were no differences in
grasshopper densities determined by hoop counts be-
tween NO BIRDS plots and BIRDS plots (repeated-
measures ANOVA, P = .137, Fig. 2). We found no
date of count effect (P = .1161, repeated-measures
ANOVA) or date of count x treatment interaction (P
= .469).
In 1989 we found a significant difference between
plots in grasshopper densities estimated by hoop counts
(repeated-measures ANOVA, P < .0001); overall more
grasshoppers were found in NO BIRDS plots than
BIRDS plots (Fig. 2). Initially there was no significant
difference between NO BIRDS and BIRDS plots (P =
.388), but differences were detected at the third count
on 4 July (P = .0008) and remained significant for the
rest of the summer (P < .000 1 and P = .008 for count
4 on 19 July and final count, respectively). At the final
count there were 33% fewer grasshoppers in BIRDS
plots than in NO BIRDS plots. Densities in NO BIRDS
plots peaked in mid-July in 1989 (Fig. 2), but the BIRDS
plots remained constant over time. There was a sig-
nificant date of count effect (P = .003) and a significant
date of count x treatment interaction in 1989 (P =
.002).
In 1988 and 1989 grasshopper densities in all treat-
ments in early June were similar (P = .42). Grasshopper
density for all treatments in 1988 and 1989 was 0.96
? 0.15 and 1.06 ? 0.11 individuals/m2, respectively
for the initial counts (means ? 1 SE). At the final count,
grasshopper densities were significantly higher in 1989
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
1778 ADA C. FOWLER ET AL. Ecology, Vol. 72, No. 5
TABLE 1. Number of grasshoppers of two size classes col-
lected in sweep-net samples from plots without bird pre-
dation (NO BIRDS, N = 8) and with bird predation (BIRDS,
N = 16), compared by ANOVA.
Size Grasshopper density (no./m2, X 1 SE)
Year classes NO BIRDS BIRDS P
1988 Large 0.17 ? 0.04 0.10 ? 0.02 .008
Small 0.42 ? 0.09 0.31 ? 0.06 .123
Total 0.59 ? 0.12 0.41 ? 0.08 .053
1989 Large 0.59 ? 0.18 0.42 ? 0.07 .014
Small 0.77 ? 0.26 0.56 ? 0.14 .037
Total 1.36 ? 0.42 1.01 ? 0.19 .016
(1.47 ? 0.33 individuals/) when compared with the
final count in 1988 (0.65 ? 0.11 individuals/) (P =
.009). Most of this difference appeared to be due to
higher grasshopper densities on the NO BIRDS plots
in 1989 although there was no year x treatment in-
teraction (P = .248).
An examination of grasshoppers collected in the
sweep-net samples for both years suggested a bimodal
distribution of sizes. Consequently, grasshoppers were
separated into large (> 8 mm) and small (< 8 mm) size
classes. In 1988 the sweep-net estimates had signifi-
cantly more large grasshoppers in the NO BIRDS plots
than in the BIRDS plots, 37% more on average. There
was no difference in the densities of total grasshoppers
or small grasshoppers (Table 1).
Up to three observers did the sweep-net collecting
in 1989. The ANOVAs for total, large, and small grass-
hoppers showed a significant observer effect, but no
observer-treatment interaction; observers were differ-
ent but consistent over treatments. After accounting
for the observer effect, the contrast between NO BIRDS
and BIRDS plots was significant (Table 2); there were
significantly more total, large, and small grasshoppers
in the NO BIRDS plots than in the BIRDS plots (Table
1). There were, on average, 26% more total grasshop-
pers, 34% more large grasshoppers, and 23% more small
grasshoppers in the NO BIRDS plots.
We compared the difference between the final hoop
counts and sweep-net estimates using all treatments in
both 1988 and 1989. In 1988 the sweep-net estimates
were significantly lower than the final hoop counts (P
= .015). Grasshopper densities for 1988 final counts
and sweep-net estimates were 0.75 ? 0.10 and 0.47 +
0.07 individuals/, respectively (means ? 1 SE). In
1989 the sweep-net estimates did not differ from the
final hoop counts (P = .120). Grasshopper densities for
1989 final counts and sweep-net estimates were 1.47
+ 0.33 and 1.12 ? 0.19 individuals/, respectively.
We identified 15 adult grasshopper species in our
samples in 1988 and 16 in 1989; nymphs were not
identified to species. In 1988 two species, Trachyrha-
chys kiowa and Phlibostroma quadrimaculatum, had
significantly greater densities in the NO BIRDS plots
than in the BIRDS plots, but there was no difference
in density between treatments for any species in 1989
(Table 3). However, our ability to detect differences
between the BIRDS and NO BIRDS plots was limited
due to the low power of the tests. The average power
for the ANOVAs was 30% in 1988 and 16% in 1989.
There was no significant difference in mean length
of all grasshoppers between the NO BIRDS and BIRDS
plots in either 1988 or 1989 (Table 4). There was no
significant difference in the biomass of grasshoppers in
the treatments in 1988 or in 1989 (Table 5). Because
grasshoppers were combined into one sample, the sig-
nificant observer effect in 1989 may have obscured the
difference in biomass among treatments. Neither spe-
cies richness (Table 6) nor species diversity (Table 7)
differed between the BIRDS and NO BIRDS plots.
The amount of cover and the species composition
of plants were similar between treatments within a site
in 1989 (P > .138 for all sites). There was no significant
difference in vegetation height between treatments (P
= .752). The mean plant height was 120.00 ? 7.37
mm for NO BIRDS plots and 118.10 ? 5.21 mm for
BIRDS plots (means ? 1 SE).
DISCUSSION
Our results suggest avian predators may have sig-
nificant negative impacts on grasshopper population
size. Similar experiments performed by Joern (1986)
in western Nebraska and C. E. Bock in southern Ari-
zona (personal communication) also found densities of
grasshoppers were 30-50% higher in treatments from
which birds were excluded.
Grasshoppers may be attracted to the NO BIRDS
plots because the mesh roof provides a favorable mi-
croclimate. However, Joern (1986) found little differ-
ence in wind speed or solar radiation inside exclosures
that were very similar to ours. In addition, we did not
detect any difference in vegetation height or compo-
sition between our BIRDS and NO BIRDS plots in
1989. Thus, it is unlikely that grasshoppers were re-
sponding to microhabitat differences between the plots.
It is possible that the differences in grasshopper den-
sities between the BIRDS and NO BIRDS plots may
not be due to predation per se but simply reflect dif-
ferential movement of grasshoppers into and out of the
TABLE 2. ANOVA table for total grasshoppers collected in
sweep-net samples in 1989 from plots with bird predation
(BIRDS, N = 16) and without bird predation (NO BIRDS,
N= 8).
Source of variation df P
Site 7 .0001
Plot 2 .101
BIRDS (SIDES vs. OPEN) 1 .793
NO BIRDS vs. BIRDS 1 .016
Observer 2 .000 1
Observer x plot 4 .943
Error 101
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
October 1991 AVIAN PREDATION ON GRASSHOPPERS 1779
TABLE 3. Grasshopper species collected in sweep-net samples from plots without bird predation (NO BIRDS, N= 8) and
with bird predation (BIRDS, N = 16). Densities were compared using ANOVAs.
Grasshopper density (no./M2, X ? 1 SE)
1988 1989
Species NO BIRDS BIRDS P NO BIRDS BIRDS P
Aeropedellus clavatus 0.010 ? 0.006 0.008 ? 0.002 .529 0.030 ? 0.009 0.024 ? 0.004 .843
Ageneotettix deorum 0.014 ? 0.003 0.016 ? 0.004 .738 0.034 ? 0.013 0.026 ? 0.006 .227
Amphitornus coloradus 0.017 ? 0.005 0.016 ? 0.003 .584 0.036 ? 0.011 0.022 ? 0.004 .245
Aulocara elliotti 0.022 ? 0.009 0.011 ? 0.020 .378 0.024 ? 0.007 0.019 ? 0.006 .427
Melanoplus infantilis 0.019 ? 0.010 0.017 ? 0.005 .906 0.056 ? 0.020 0.064 ? 0.019 .722
Melanoplus sanguinipes 0.027 ? 0.017 0.016 ? 0.005 .640 0.115 ? 0.094 0.073 ? 0.034 .942
Metator pardalinus * * * 0.009 ? 0.004 0.010 ? 0.003 .992
Opeia obscure 0.030 ? 0.007 0.032 ? 0.007 .712 0.031 ? 0.010 0.033 ? 0.008 .512
Phlibostroma quadrimaculatum 0.071 ? 0.027 0.034 ? 0.022 .021 0.099 ? 0.054 0.040 ? 0.008 .255
Trachyrhachys kiowa 0.040 ? 0.008 0.011 ? 0.002 .004 0.033 ? 0.006 0.059 ? 0.012 .254
Other 0.016 ? 0.010 0.006 ? 0.001 .407 0.007 ? 0.001 0.009 ? 0.003 .335
* No individuals of this species were caught in 1988.
plots. C. E. Bock (personal communication) tested this
by constructing barriers around some of his plots from
which birds were excluded to keep grasshoppers from
moving in and out. Densities at the end of the exper-
iment in the plots with barriers were even higher than
in the plots without barriers. He suggested that the plots
from which birds were excluded were not attracting
grasshoppers, but that grasshoppers were dispersing
from them. Consequently, differences between our NO
BIRDS and BIRDS plots could have been greater than
estimated if dispersal from the NO BIRDS plots was
occurring.
The lack of a difference in grasshopper density be-
tween the SIDES and OPEN plots supports observa-
tions that grasshopper movements were not affected
by the chicken wire and suggests that predation was
similar in the BIRDS plots. Joern (1986) constructed
bird enclosures out of a polypropylene mesh and ob-
served grasshoppers moving through the polypropyl-
ene mesh without difficulty. We also observed grass-
hoppers freely moving through the chicken wire. Other
investigators (C. E. Bock, personal communication, G.
E. Belovsky, personal communication) found chicken
wire did not deter grasshopper movement.
TABLE 4. Lengths of grasshoppers of two size classes col-
lected in sweep-net samples from plots without bird pre-
dation (NO BIRDS, N = 8) and with bird predation (BIRDS,
N = 16). Comparisons were by ANOVAs.
Size Length (mm, X ? 1 sE)
Year classes NO BIRDS BIRDS P
1988 Large 17.84 ? 0.48 16.24 ? 1.11 .144
Small 6.32 ? 0.16 6.56 ? 0.13 *
Total 10.35 ? 0.76 9.33 ? 0.38 .184
1989 Large 16.97 ? 0.51 17.23 ? 0.41 .584
Small 5.62 ? 0.16 5.29 ? 0.36 .865
Total 11.32 ? 1.04 11.53 ? 0.78 .843
* There was a significant difference (P < .001) between the
BIRDS plots (SIDES vs. OPEN), so NO BIRDS and BIRDS
plots were not contrasted.
Belovsky et al. (1990) found that birds prey selec-
tively on large grasshopper species, while we found
little evidence of size-selective predation in our study.
The difference between our results and those of Belov-
sky et al. (1990) may be due to differences in meth-
odology and to differences in the species composition
of grasshoppers at the two sites. They estimated pre-
dation rates on grasshoppers that were tethered with
monofilament line. Tethered grasshoppers cannot hop
or fly great distances to avoid predators and therefore
may have been more susceptible to predation than free-
roaming grasshoppers. Belovsky et al. (1990) attempt-
ed to correct for this bias by adjusting the predation
rates based on capture success of untethered grasshop-
pers by birds. They used the same correction factor for
all species, however, and therefore their results may
still be biased if large-bodied species use different pred-
ator avoidance tactics than small species. Another fac-
tor that may have contributed to a lack of size-depen-
dent predation in our study was the distribution of
grasshopper sizes. In our study sites large species of
grasshoppers were not common. Metator pardalinus
was most common (male length: 26 mm, female length:
34 mm), but caught infrequently (four individuals cap-
tured in two BIRDS plots). Thus, there were few large
grasshoppers for the birds to prey on, and therefore
there may be no advantage to preying selectively on
large individuals in our study area.
Joern (1986, 1988) and Belovsky et al. (1990) also
found that birds preyed selectively on different species
of grasshoppers. Joern (1988) measured predation by
watching Grasshopper Sparrows (Ammodramus sa-
vannarum) forage on four species of grasshoppers in
small aviaries placed in grassland vegetation. He found
that the sparrows selected one species (Amphitornus
coloradus) more often than expected. This species oc-
curred in our study area in both years, but we failed
to find any difference in density between the BIRDS
and NO BIRDS plots (Table 3). Joern (1986) also found
evidence for selective predation in enclosure experi-
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
1780 ADA C. FOWLER ET AL. Ecology, Vol. 72, No. 5
TABLE 5. Biomass of grasshoppers collected in sweep-net
samples from plots without birds (NO BIRDS, N = 8) and
with bird predation (BIRDS, N = 16). Comparisons were
by ANOVAs.
Biomass (g/m2, X ? 1 SE)
Year NO BIRDS BIRDS P
1988 0.012 ? 0.003 0.007 ? 0.001 .179
1989 0.064 ? 0.022 0.047 ? 0.007 .187
ments. He found 6 of 24 grasshopper species had sig-
nificantly lower densities in the plots with bird pre-
dation. A seventh species had significantly greater
densities in the plots with bird predation. We found
little evidence for differences in species density as a
function of bird predation. None of the species differed
significantly in 1989 and only two species were signif-
icantly higher in NO BIRDS plots in 1988. The in-
ability to detect differences in species composition be-
tween treatments may have been a result of low power;
the average power for the ANOVA was 30% in 1988
and 16% in 1989. The heterogeneity of grasshopper
species among our sites could have contributed to the
lack of power. Joem's (1986) and Belovsky et al.'s (1990)
studies were conducted in small areas (several hectares)
with relatively homogeneous grasshopper species com-
position, while our sites were distributed over a much
larger area and there was greater heterogeneity among
sites. Local differences in prey selection by the birds
may therefore have been obscured by differences in
grasshopper density and species composition among
sites.
Joern (1986) found grasshopper species diversity and
richness were significantly higher in plots with bird
predation. In our experiment there was no difference
in species diversity or richness in either year. Unlike
Joern (1988), our results suggests that avian predation
on grasshopper assemblages has no apparent com-
munity effect. This conclusion should be tempered be-
cause of the low power associated with individual spe-
cies densities.
The evidence that avian predation depresses grass-
hopper populations was stronger in 1989 than in 1988.
T. L. George et al. (unpublished manuscript) studied
the impact of a drought during the summer of 1988
on the grassland birds in the Little Missouri National
Grasslands. They used permanent bird transects to fol-
low population trends during 1987-1989. The drought
correlated with a decline of 61% in the density of grass-
land birds between June 1987 and June 1988. Total
bird density increased between June 1988 and June
1989, although it was still 8% lower than in June 1987.
Also, nesting productivity of Vesper Sparrows (Pooe-
cetes gramineus) was significantly reduced in 1988 rel-
ative to 1987 and 1989. Grasshopper densities were
also reduced during the drought, but their numbers
recovered only slightly in 1989 (K. Winks and R. J.
Dysart, personal communication). Thus the relative
impact of bird predation may have been greater in 1989
than in 1988.
Our results suggest that avian predation can depress
prey numbers at the grasshopper densities observed in
our study. Grasshopper densities at our study sites ( 1-
4 individuals/) are considered low for this area (K.
Winks and R. J. Dysart, personal communication). This
is consistent with other studies of avian predation on
insect populations. Otvos (1979) and Takekawa et al.
(1982) reviewed the literature on the effects of avian
predation on their prey and concluded that avian pre-
dation in forest ecosystems had more impact on pop-
ulations of insect pests at low and moderate insect
densities. Greene (1989), in a study of forest birds,
found between-year differences that are consistent with
this interpretation. In the first year of his study insect
biomass was reduced outside bird enclosures, but in
the second year there was an outbreak of caterpillars
and he found no difference between the insect biomass
inside and outside exclosures. Both he and Holmes et
al. (1979) indicated that avian predation can depress
prey at low numbers.
Some studies of birds in rangelands (Wiens 1973,
1974, Wiens and Rotenberry 1979, Rotenberry 1 980a,
b) report that food is generally superabundant and birds
do not have significant effects on insect populations.
These studies were based on energetic models of avian
consumption and assumed no selectivity by the birds.
Our results, however, and experiments performed by
others (Joern 1986, C. E. Bock, personal communi-
cation), demonstrate that grassland birds depress grass-
hopper densities. This indicates that birds may have a
significant impact on some prey types, and that the
conclusions based on energetic models may be mis-
leading.
The effect of avian predation on grasshopper pop-
ulations at high densities in western North Dakota has
TABLE 6. Species richness of grasshoppers collected in sweep-
net samples from plots without birds (NO BIRDS, N = 8)
and with bird predation (BIRDS, N = 16). Comparisons
were by ANOVAs.
Species richness (X ? 1 SE)
Year NO BIRDS BIRDS P
1988 6.38 ? 0.89 6.13 ? 0.67 .73
1989 11.4 ? 0.39 11.5 ? 0.19 .81
TABLE 7. Species diversity of grasshoppers collected in sweep-
net samples from plots without birds (NO BIRDS, N = 8)
and with bird predation (BIRDS, N = 16). Comparisons
were by ANOVAs.
Species diversity (X ? 1 SE)
Year NO BIRDS BIRDS P
1988 1.50 ? 0.149 1.48 ? 0.127 .86
1989 4.95 ? 0.394 4.89 ? 0.746 .90
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
October 1991 AVIAN PREDATION ON GRASSHOPPERS 1781
not been determined, but evidence from other studies
suggests that birds have little impact on large-scale
outbreaks. Forest ecosystem ecologists (see Otvos [1979]
and Takekawa et al. [1982] for reviews) agree that dur-
ing large-scale insect outbreaks the effect of avian pred-
ators is insignificant. Crawford and Jennings (1989),
in a study of bird predation on spruce budworm, found
birds consumed 84% of larva and pupa at low densities
and 22% at intermediate densities, but bird predation
was ineffective at high densities (> 106 larva/ha).
Between-years effects (i.e., long-term effects) of bird
predation cannot be determined from our study be-
cause seasonal abundances of grasshoppers are related
to a variety of factors including: weather, parasite and
pathogen abundance, and grazing intensity (see
Dempster [1963] for review). Otvos (1979), Takekawa
et al. (1982), and Wiens (1973) concur that birds may
be able to delay insect population buildups, thus in-
creasing time between outbreaks. Only by incorporat-
ing the effects of avian predation into models of grass-
hopper population dynamics may we begin to answer
what long-term effects avian predators have on grass-
hopper populations. Bird predation, however, may have
important short-term economic benefits by reducing
forage loss to grasshoppers during the growing season.
Avian predators could be of importance as regulators
of grasshopper density in integrated pest management
(IPM) programs. The effects of avian predation and of
insecticides that did not harm birds could complement
each other (McEwen et al. 1986, McEwen 1987). More
field experiments are needed to determine the effects
of predation on insect pest population dynamics and
ecosystem function and to develop ways of incorpo-
rating predation by wild vertebrates into specific IPM
practices.
AcKNowLEDGMENTrs
We thank Jane Adams, Linda Arnold, Alan Warner, Scott
Miller, and Clark Nelson for help in constructing enclosures
and collecting data in North Dakota. William F. Andelt and
Beatrice van Horne provided valuable advice throughout the
project and made many helpful suggestions on the manu-
script. Our study was supported by the Grasshopper Inte-
grated Pest Management Project of USDA/APHIS.
LITERATURE CITED
Baldwin, P. H. 1972. The feeding regime of granivorous
birds in short grass prairie in Colorado USA. Pages 237-
247 in S. C. Kendeigh and J. Pinowski, editors. Productiv-
ity, population dynamics, and systematics of granivorous
birds. Polish Scientific Publications, Warszawa, Poland.
Beal, K. G., and K. J. Khamis. 1990. Statistical analysis of
a problem data set: correlated observations. Condor 92:
248-25 1.
Belovsky, G. E., J. B. Slade, and B. A. Stockhoff. 1990.
Susceptibility to predation for different grasshoppers: an
experimental study. Ecology 71:624-634.
Boyd, R. L. 1976. Behavioral biology and energy expen-
diture in a Homed Lark population. Dissertation. Colorado
State University, Fort Collins, Colorado, USA.
Crawford, H. S., and D. T. Jennings. 1989. Predation by
birds on spruce budworm Choristoneurafumiferana: func-
tional, numerical, and total responses. Ecology 70:152-
163.
Daubenmire, R. 1959. A canopy coverage method of veg-
etational analysis. Northwest Science 33:43-64.
Dempster, J.P. 1963. The population dynamics of grass-
hoppers and locusts. Cambridge Philosophical Society Bi-
ological Reviews 38:490-529.
Greene, E. 1989. Food resources, interspecific aggression,
and community organization in a guild of insectivorous
birds. Dissertation. Princeton University, Princeton, New
Jersey, USA.
Holmes, R. T., J. C. Schultz, and P. Nothnagle. 1979. Bird
predation on forest insects: an exclosure experiment. Sci-
ence 206:462-463.
Joern, A. 1986. Experimental study of avian predation on
coexisting grasshopper populations (Orthoptera: Acrididae)
in a sandhills grasslands. Oikos 46:243-249.
. 1988. Foraging behavior and switching by the
Grasshopper Sparrow Ammodramus savannarum search-
ing for multiple prey in a heterogeneous environment.
American Midland Naturalist 119:225-234.
McEwen, L. C. 1987. Function of insectivorous birds in a
shortgrass IPM system. Pages 324-333 in J. L. Capinera,
editor. Integrated pest management on rangeland: a short-
grass prairie perspective. Westview, Boulder, Colorado,
USA.
McEwen, L. C., L. R. DeWeese, and P. Schladweiler. 1986.
Bird predation on cutworms (Lepidoptera: Noctuidae) in
wheat fields and chlorpyrifos effects on brain cholinesterase
activity. Environmental Entomology 15:147-151.
Onsager, J. H., and J. H. Henry. 1977. A method for esti-
mating the density of rangeland grasshoppers (Orthoptera:
Acrididae) in experimental plots. Acrida 6:231-237.
Otvos, I. S. 1979. The effects of insectivorous bird activities
in forest ecosystems: an evaluation. Pages 341-374 in J. G.
Dickinson, R. N. Connors, R. R. Fleet, J. A. Jackson, and
J. C. Kroll, editors. The role of insectivorous birds in forest
ecosystems. Academic Press, New York, New York, USA.
Price, P. W. 1987. The role of natural enemies in insect
populations. Pages 287-312 in P. Barbosa and J. C. Schultz,
editors. Insect outbreaks. Academic Press, New York, New
York, USA.
Rotenberry, J. T. 1980a. Bioenergetics and diet in a simple
community of shrubsteppe birds. Oecologia (Berlin) 46:7-
12.
. 1980b. Dietary relationships among shrubsteppe
passerine birds: competition or opportunism in a variable
environment? Ecological Monographs 50:93-110.
SAS Institute. 1988. SAS/STAT user's guide. 6.03 edition.
SAS Institute, Cary, North Carolina, USA.
Takekawa, J. Y., E. 0. Garton, and L. A. Langelier. 1982.
Biological control of forest insect outbreaks: the use of avian
predators. Transactions of the North American Wildlife
and Natural Resources Conference 47:393-402.
Wiens, J. A. 1973. Pattern and process in grassland bird
communities. Ecological Monographs 43:237-270.
1974. Climatic instability and the "ecological sat-
uration" of bird communities in North American grass-
lands. Condor 76:385-400.
Wiens, J. A., and M. I. Dyer. 1975. Rangeland avifaunas:
their composition, Energetics, and role in the ecosystem.
Pages 140-181 in D. R. Smith, editor. Proceedings, Manage-
ment of Forest and Rangeland Habitat for Non-game Birds.
United States Forest Service General Technical Report
WO-1.
Wiens, J. A., and J. T. Rotenberry. 1979. Diet niche rela-
tionships among North American grassland and shrub-
steppe birds. Oecologia (Berlin) 42:453-492.
This content downloaded from 129.82.28.144 on Sun, 20 Mar 2016 18:57:30 UTC
All use subject to JSTOR Terms and Conditions
... Individual species are more common in areas with specific habitat characteristics, primarily related to vegetation structure (Walk and Warner 2000, Herkert 1994, George and McEwen 1992, Schramm et al.1984, Wiens and Rotenberry 1981, Rotenberry and Wiens 1980, Wiens 1969. Often, but not always, breeding and wintering habitats of birds are similar in structural characteristics. ...
... A critical first step to formulate habitat management techniques is to examine habitat associations. Individual species of grassland birds, along with birds in many other habitats, are more common in areas with specific habitat characteristics, usually associated with vegetation structure (Wiens 1969, Rotenberry and Wiens 1980, Wiens and Rotenberry 1981, George and McEwen 1992, Schramm et al. 1984, Herkert 1994, Walk and Warner 2000. Often, but not always, breeding and wintering grounds share similar structural characteristics. ...
... Abundances of other grassland species, such as Savannah sparrow and grasshopper sparrow, have been positively associated with percent of grass (Schneider 1998). Other parameters such as percent of bare ground, shrubs, vegetation height, and litter depth have also shown significant effects on the abundances of some species such as chestnut-collared longspur, bobolink, western meadowlark, clay-colored sparrow and vesper sparrow (George and McEwen 1991, Schneider 1998, Sutter and Brigham 1998. ...
Thesis
Loss and degradation of native prairie are the most likely causes of dramatic recent grassland songbird population declines. In southwestern Manitoba, most native prairies have been replaced by intensive cultivation. Remaining grasslands are found in a combination of native and non-native grasslands, including Wildlife Management Areas (WMAs).We surveyed grassland bird diversity and associated vegetation to determine how (1) WMAs and private lands contribute to the conservation of grassland birds, and (2) how different land uses affect prairie bird diversity. We sampled 229 plots on 38 prairies, including 13 private lands, 13 crown lands and 12 WMAs, which were characterized by different land uses or management practices (hayed, idle, burned, continuously and rotationally grazed). WMAs supported the higher number of nongrassland birds and the lowest richness of grassland species, compared to private and crown lands. The diversity of grassland birds was also affected by the different land management practices. Overall, the abundance and richness of grassland species was higher on continuously and rotationally grazed fields than hayed sites. The absence of active management on some areas like WMAs may affect the habitat suitability for grassland species by increasing the woody vegetation encroachment. Some active management practices, however, may not contribute to the conservation of bird species. Our results suggest that some management practices like haying may not contribute to a high richness and abundance of grassland birds, and other practices like grazing may play a more important role to provide suitable habitat in southwestern Manitoba.
... The GAP and CWHR programs were built upon one of the most successful and widely used means of defining species habitat relationships: the categorization of the landscape into land cover classes based upon vegetation composition. The science of wildlife-habitat relationships was developed (and continues) with the use of birds as model species and vegetation as the habitat predictor (Merriam, 1890;Adams, 1908;Lack, 1933;Svardson, 1949;Hilden, 1965;Verner et al., 1986;George and McEwen, 1992;Kellner et al., 1992;Scott et al., 1993;Morrison et al., 1998;Scott et al., 2002). Though many other factors have been included more recently in modeling efforts, vegetation, because of its historical usage, universal availability, and success with many species, has remained the primary variable used to predict animal habitat. ...
Article
The macro-habitat preferences of three conspicuous and widely distributed species of lizards (Aspidoscelis tigris, Callisaurus draconoides, and Uta stansburiana) were examined across four geomorphic landforms (sandy wash, rocky wash, alluvial plain, and alluvial deposit) in the southern Mojave Desert, California. All three species were non-randomly distributed across the four geomorphic landforms. The goal of this study was to develop less ecologically generalized habitat models (LizLand) than the vegetation-based wildlife–habitat relationship models in the California Gap Analysis Program (CA-GAP). Conceptually, LizLand is a geomorphological approach to habitat modeling in arid environments. Specifically, LizLand is a series of spatially explicit habitat models that define and predict habitat for A. tigris, C. draconoides, and U. stansburiana in Joshua Tree National Park and the Marine Air Ground Task Force Training Command, Marine Corp Air Ground Combat Center. LizLand models resulted in higher resolution habitat models with minimal reduction in model accuracy. These models more accurately captured the complexity of the Mojave Desert ecosystem and offered greater ecological resolution in identifying habitat in contrast to the CA-GAP models.
Article
Full-text available
In grassland ecosystems, invasions by nonnative grasses typically decrease floristic diversity and structural heterogeneity in ways that alter the quantity and quality of habitat for animals. Grassland arthropods that rely directly on herbaceous plants for food, shelter, or as substrates for reproduction are especially vulnerable to these invasions because many have evolved specialized relationships with host plants that might be displaced. We evaluated how invasions by nonnative grasses affected abundance and richness of foliage-dwelling arthropods in semidesert grasslands of Arizona, USA. On 90, 3.1-ha plots established along a gradient of invasion where dominance of nonnative grasses ranged from 0 to nearly 100% of grass cover, we captured > 90,000 arthropods from 11 orders during 270 surveys in 2014 and 2015. Although the invasion by nonnative grasses (primarily Eragrostis lehmanniana and secondarily E. curvula) increased the amount of herbaceous foliage available to arthropods, richness of arthropods decreased by an average of 2% and total abundance by an average of 7% for every 10% increase in nonnative-grass dominance. Responses to the plant invasion, however, varied among taxa and functional groups. As dominance of nonnative grasses increased, abundances of most predators and specialist herbivores decreased, whereas abundances of most generalist herbivores were lowest at intermediate points of the invasion gradient. The changes we observed in the arthropod community have potential to alter broad-scale ecological processes, including energy flow and nutrient cycling, and to reduce food resources for insectivores, which can have adverse, cascading effects on imperiled grassland ecosystems.
Article
Full-text available
Many bird species consume large quantities of invertebrates, some of which are considered agricultural pests. However, relatively few quantitative studies demonstrate that avian predators can reduce agricultural insect pests below a level at which unchecked populations would cause economic damage to crops or require treatment with pesticides. In fact, since the subject of economic ornithology (“all relations of birds that has any bearing upon the material welfare of [humans]”— McAtee, 1933) was last reviewed (e.g., McFarlane, 1976; Sweetman, 1958), there have been only a handful of published studies on the role of birds in agricultural ecosystems, and most of these refer to small-plot effects, with no attempt to aggregate the results at the level of an agricultural economic unit, i.e., the farm.
Article
Full-text available
This report is one in a series of literature syntheses on North American grassland birds. The need for these reports was identified by the Prairie Pothole Joint Venture (PPJV), a part of the North American Waterfowl Management Plan. The PPJV recently adopted a new goal, to stabilize or increase populations of declining grassland- and wetland-associated wildlife species in the Prairie Pothole Region. To further that objective, it is essential to understand the habitat needs of birds other than waterfowl, and how management practices affect their habitats. The focus of these reports is on management of breeding habitat, particularly in the northern Great Plains. Suggested citation: Dechant, J. A., M. L. Sondreal, D. H. Johnson, L. D. Igl, C. M. Goldade, A. L. Zimmerman, and B. R. Euliss. 1999 (revised 2002). Effects of management practices on grassland birds: Western Meadowlark. Northern Prairie Wildlife Research Center, Jamestown, ND. 33 pages.
Article
Full-text available
We experimentally studied the effects of avian predation on grasshopper abundance in western North Dakota during the summers of 1988 and 1989. Grasshopper densities in 15 x 15 m plots from which birds were excluded (NO BIRDS) were compared with similar-sized plots where birds were allowed to forage (BIRDS). Plots were established in early June at 16 sites (8 per year), and grasshopper densities were estimated from hoop counts in NO BIRDS and BIRDS plots at 2-wk intervals until the end of July. There were significantly more grasshoppers in NO BIRDS plots than in BIRDS plots in 1989 (P < .0001), but not in 1988 (P = .137). The difference between treatments in 1989 was first detected 6 wk after the exclosures were erected. Between years there were no differences in initial grasshopper abundance in all treatments (P > .388). After the final hoop count each year, sweep-net sampling was also used to estimate grasshopper densities. By this method, average grasshopper density in late July was 26% and 37% lower in BIRDS plots than in NO BIRDS plots in 1988 and 1989, respectively. Average length, total biomass of grasshoppers, species richness, and species diversity, however, did not differ between the treatments. In 1988, 2 of 15 grasshopper species were significantly more abundant in the NO BIRDS plots. There was no difference between the treatments among 16 species identified in the 1989 samples, but the power of our tests to detect differences for individual species was low. Our results support the hypothesis that avian predation reduces insect populations at low and moderate densities.
Chapter
This chapter discusses the fundamental principles of predator-prey interaction out of the enormous amount of information so far accumulated. Representative studies are presented and the data are reanalyzed to interpret comparatively the characteristics and actual status of the predator-prey interaction under a range of situations in the light of the theoretical principles. The chapter discusses that the predator-prey interaction is fundamentally a process that brings about disturbance to the overall system rather than stabilization, having an innate tendency to generate reciprocal oscillations in both populations. The predator-prey interactive system is fundamentally a delicate entity filled with contradictions, which arise from the fundamental paradox that predators are obliged to exploit and suppress continuously the very population of their prey that are the indispensable resource to assure survival and propagation of the predators themselves. The role of predators as mortality factors in natural animal populations is important in general, though it of course varies with species, with a general tendency to become less significant in higher animals, which usually have developed effective anti-predator defense strategies. The importance of predators as key factors in population fluctuations is, however, usually minor even among insect populations, and it seems to be virtually negligible in most populations of birds or mammals.
Article
Examines the influence of vertebrate and invertebrate predators and parasitoids as natural enemies of insects, ie as regulators of the populations of potential outbreak species. If regulation occurs before a particular population threshold, neither depletion of food resources nor economic damage will be found. Evidence is drawn from studies on attempts at biological control, and note is taken of how pesticides may affect a situation (eg after pesticide use target pests may actually increase and new pests may materialise). Population models and life table analysis often have limited value, but some experimental studies are proving to be of worth. The notion of more than a single population equilibrium is explored, and implications of examining trophic interactions at 3 levels are considered. -P.J.Jarvis
Article
Examined the hypothesis that predation by forest birds restricts expansion of low-density populations of spruce budworm (Lepidoptera: Tortricidae). Bird populations were censuses in spruce-fir stands of low to moderate budworm density in N New Hampshire and W Maine. In plots where spruce budworm densities ranged from low to transitional, the entire bird community as well as specific groups of birds (overstory warblers and golden-crowned kinglets Regulus satrapa; sparrows, vireos, and juncos; and nuthatches, thrushes, and purple finches Carpodacus purpureus) showed significant functional responses to increasing budworm density. When transitional budworm densities were eliminated from the analysis, only the warbler-kinglet group exhibited a significant functional response. Only Canada warbler Wilsonia canadensis and golden-crowned kinglet showed numerical responses to increasing budworm numbers. The percentage of the budworm population consumed by birds decreased as budworm numbers increased. Birds consumed 84% of the larvae and pupae where budworm populations were low, 22% where budworm populations were intermediate between low and high density. Once 4th instar larvae exceeded 106 individuals/ha, bird predation was ineffectual. The hypothesis that birds can effectively limit budworm increase is compatible with theories to explain budworm outbreaks. -from Authors