Download DIVERGENT PREY SELECTION IN TWO SPECIES OF

DIVERGENT
PREY
SELECTION
IN TWO
SPECIES
OF
WATERTHRUSHES (SEIURUS)
ROBERTJ. CRAIG
Quinebaug
ValleyCommunity
College,Danielson,
Connecticut
06239USA
ABSTRACT.--I studied Louisiana (Seiurusmotacilla) and Northern (S. noveboracensis)
water-
thrushesduringar• exceptionally
dry springto determineif environmental
stresselicited
interspecificcompetition.Previously, I had found little evidence for competition between
these speciesdespite wide overlap in foraging methods,use of foraging microhabitats,and
characteristicsof breeding habitat. I observedbreeding adults forage by placing them in a
portableflight cagelocatedin natural habitat,and concurrentlygathereddata on the influenceof waterthrushforagingon aquaticinvertebrateprey and prey abundance.The species
selecteddifferent prey. LouisianaWaterthrushesfed predominantly on Trichoptera larvae
and on larger averageprey than did Northern Waterthrushes,which fed predominantlyon
Diptera larvae. The specieshad similar foragingmethodsand attackrates,indicating that,
unlike many Parulinae,their principal foraging differenceswere in prey selectionrather
than in meansof locatingprey. Experimentswith foragingexclosures
gaveno clearevidence
that waterthrushesaffected prey biomassor composition.There was no significant relation
between territory size vs. prey biomassand water cover. No interspecificaggressionwas
observed,and territories overlapped widely, indicating that interference competition did
not occur.Divergencein prey selectionimpliesexploitlyecompetition,but invertebrateand
habitat data indicate that prey were not limiting, thus making competitionfor prey unnecessary.Therefore,from thesedataI cannoteliminatethe alternativehypothesisthat observed
differencesbetween the speciesmay only reflect independent specialization.Received
4 April
1986,accepted
29 September
1986.
Both are closely associatedwith wooded wetlands, and from spring arrival (mid-April) to
the end of incubation (late May) about 80% of
foraging occurs in water (Craig 1984, 1985).
Previously,I found that they overlap widely in
foraging methods,use of foraging microhabitats,and characteristics
of their breeding habitat, but they are not interspecificallyterritorial. When precipitationin December1984to May
1980, Emlen 1981). Moreover, Wiens (1977) 1985was 14 cm below average,one of the driest
assertedthat in variable environmentscompe- winter-spring seasonson record (NOAA 1985),
tition plays an important ecologicaland evo- there was a particularly strongpotential for relutionary role only during periodsof environ- sourcelimitation. I expectedthat if exploitive
mental stresswhen resourcesare limiting. In competition occurs, I would observe diversupportingthis view, Mountainspringand Scott gencein resourceuse,and if interferencecom(1985) found evidence for only intermittent petition occurs,it would produce interspecific
competition in Hawaiian bird communities. aggressionand spatial divergence (seeMaurer
Similarly, Robinson (1981) found increased 1984).
aggressionbetween two vireo species(Vireo:
Vireonidae) only when population density inSTUDY AREA AND METHODS
creased and, presumably, resources became
limiting.
I studied waterthrushes at Yale Forest, Ashford,
I studied resourceuse by Louisiana (Seiurus Tolland Co., Connecticut. The habitat was described
INCREASINGLY,
the role of interspecific competition in shapingthe ecologyof avian species
has been questioned.Ecologicalpatterns traditionally interpreted in terms of competition
may have alternative explanations(e.g. Confer
and Knapp 1981, Jehl and Parkes 1983), and
therefore the importanceof autecologyand behavioral limitations in influencing patternshas
been assessed
more carefully.(e.g. Rotenberry
motacilla) and Northern (S. noveboracensis)
wa-
by Craig (1985).Briefly, it consistsof a ravine, Boston
terthrushesto determine whether interspecific Hollow, vegetated by mixed hardwood-hemlock
pine (Pinusstrobus)
forestwith
competition occursbetween the two species. (Tsugacanadensis)-white
180
The Auk 104: 180-187. April 1987
April 1987]
Waterthrush
PreySelection
181
sectionsof stream and swamp along its base.In 1985
much of the area was noticeably drier than I had
observed it since 1978; normally rushing spring-
index varies from 0, no overlap, to 1, complete overlap. To obtain a relative measureof prey preference,
brooksin the vicinity were nearly dry.
Early in the breeding seasonforaging overlap is
ranking the frequency of resourceuse and resource
abundance and computing the difference in ranks.
The values obtained indicate preference relative to
other prey items present,with the highest negative
valuesbeing mostpreferred and the highestpositive
values being most avoided.
Prey abundance.--Isampled invertebrates in the
swampy upstreamhalf (2 ha) of BostonHollow, an
area normally used by both waterthrush species.
Doing so minimized effectsof downstreaminvertebrate drift (Hynes 1970) becausethe wetland origi-
greatest,foraging in water predominates, and avail-
able prey are largely limited to water. I gathered
foraging and prey data in aquatic habitats from
mid-April to the end of May, the period when competition seemedmostlikely to occur.However, I continued to make qualitative observationsthroughout
the breeding season.
Foraging.--To observe foraging I constructed a
portable 3.5 m2 x 2 m tall flight cage of 1.3-cm mesh
crop netting on a modified aluminum tent frame. After mist netting a breeding adult, I erected the flight
cagein the birds' territory over typical feeding habitat, which included water, mossyhummocks,mud,
leaf packets,and shrubs,and observedforaging from
a blind about 1 m away. I recordedon tape the duration of eachforaging episode,number of predation
attempts (attacks),foraging method used (i.e. pick,
leaf pull, hawk, hover; see Craig 1984), prey size,
and, when possible,prey identity. I limited observations
on each bird to 90 rain to minimize
interfer-
I used the method of Johnson (1980), which involves
nates in this area and has little
stream
flow.
On 16-
17 April I sampled 10 swamp poolsto determine the
compositionof invertebrate populationsat the time
of waterthrush
arrival.
At this time
I also anchored
1 m2 x 15 cm high wire screeningexclosures(1.5ram mesh) 8 cm into the substrateof 10 pools. Exclosureseliminatedall vertebratepredatorsexceptdusky
salamanders(Desmognathus
fuscus),which are also
waterthrush prey. However, waterthrushes were
probably the only important vertebratepredatorson
aquatic organisms;small mammals and frogs in the
habitat take primarily terrestrial and flying prey, no
fish are present, and there are no other important
avian predators.On 21-22 May I sampledinverte-
ence with breeding activitiesand to limit alteration
in type of prey available.About 45 rain elapsedfrom
the time of capture to placing a bird in the flight
cage,whichprovidedsomeuniformityin hungerstate brates in the 10 exclosures and also those at 10 difamong the birds studied.
ferent uncovered swamp pools.
Despite variation in behavior among individuals,
To sample,I presseda 0.5-m2 wooden frame 5 cm
I pooled data in analysesof foraging behavior, as- into the pool substrate, removed with a flat metal
suming that data setswere thus more representative scoopthe top 3 cm of bottom debris,and then swept
of the population present. This seemed reasonable the water column (generally <4 cm deep) 10 times
because I studied 50% of the breeders in Boston Holwith !.5-ram meshwire net to removeany remaining
low. However, subsets of the data were tested to asorganisms.At 2 sites I repeated the sampling procesessindividual variation. For analysisof attackrate dure to assessthe completenessof invertebrate reand picking frequency I used data from observation moval. Once collected, I rinsed mud from the samperiods with a minimum length of 30 s to obtain a ples through a !.5-ram mesh screen and froze the
measure of variation in behavior (Robinson and
remaining debris and organisms.Later, I sorted orHolmes 1982). Becausenearly all attackswere by ganismsfrom debris, identified and measuredthem,
either picking or leaf pulling, I did not analyzesep- and weighed each sample after drying to constant
completearately data on relative proportionsof other foraging massat 60øC.I resorted2 samplesto assess
methods.
ness of removal.
I believe this method of studyingdiet is preferable
to analysesof stomachcontents(e.g. Robinsonand
Holmes 1982)becauseit providesdata not only on
prey size and type but also on foraging methods.
Territorialityand habitat.--I color-banded and measured !3 of !8 breeding adults present, mapped territories,and observedover 26 daysfor intra- and interspecific aggression. To assessseparately the
importanceof amount of feeding habitat present, I
further analyzeddata on water coverand prey biomasscollectedfrom 1978 to !980 using methodsdescribedby Craig (!984, !985).
Moreover,
it circumvents bias that is due to differ-
ential digestabilityof prey, and eliminatesthe danger of bird mortality causedby administrationof
emetics (Zach and Falls 1976).
To calculateoverlapin foragingmethodsand prey
selectionI usedthe equation:
overlap= 1 - 0.5 • IPxi- Pyil,
where pxand pyare the frequenciesof resourceuse
of speciesx and y in categoryi (Schoener1970).The
RESULTS
Foraging.--Igathereddataon 4 of 6 Louisiana
Waterthrush
and 5 of 12 Northern
Waterthrush
breederspresent in BostonHollow.
182
ROBERT
J.CRAIG
TABLEI. Selectionof principal invertebrateprey taxa
as a percentageof total prey. Observed captures
are in parentheses.
[Auk,Vol. 104
60-
LOUISIANA
WATERTHRUSH
• NORTHERN
WATERTHRUSH
50-
Seiurus
motacilla
Diptera
Trichoptera
Ephemeroptera
Oligochaeta
Isopoda
Total captures
20.3 (13)
40.7 (24)
13.6 (8)
13.6 (8)
3.4 (2)
59
35
Seiurus
noveboracensis
54.7 (66)
11.0 (13)
13.6 (16)
5.1 (6)
4.2 (5)
118
40-
59
54
Louisiana Waterthrushes(59 total prey) ate
10Trichopteralarvaesignificantlymore often than
3
7
did Northern Waterthrushes (118 total prey,
adjustedX2 = 22.6, P < 0.01; Table 1). In con-<7
>7-13
>13-19
>19
trast, Northern Waterthrushes preyed on DipPREY SIZE (mm)
tera larvae (mostly Chironomidae)significantFig. I. Selection of prey sizes. Sample sizes are
ly more often than did LouisianaWaterthrushes given at the top of each bar.
(adjustedX2 = 17.1, P < 0.01). Overlap in selection of the 5 most common prey groups (Tadae larvae, and diplopodsby LouisianaWaterble 1) that include 91.5% of Louisiana WaterLepidoptera(moth)
thrushprey and 88.1%of Northern Waterthrush thrushes,and Desrnognathus,
adults,
Sialidae
nymphs,
Tipulidae
larvae, Forprey was 0.59. In comparisonsof 3 males of
eachspeciesI found significantdifferencesbe- micidae adults, diplopods,chilopods,and gastween individuals in use of Trichoptera (Lou- tropods by Northern Waterthrushes. In addition, both species chased but did not catch
isiana Waterthrush: X2 = 8.3, 2 df, P < 0.05;
Northern Waterthrush: X2 = 12.4, 2 df, P < 0.01) flying adult Ephemeropteraand Diptera. Only
and Diptera (LouisianaWaterthrush:X2= 16.5; 7 of 177 identifiable prey were nonaquatic.
By comparingbill to prey length in the field,
Northern Waterthrush: X2= 25.1; both 2 df, P <
0.01). I could not separatethe effectsof flight- I could identify 4 prey-sizecategories:<7 mm,
>7-13 mm, >13-19 mm, and >19 mm. Northcage placement from those of individual preferences.
ern Waterthrushestook a significantly lower
! consideredother prey too infrequently se- range of prey sizesthan did Louisiana Waterlectedfor trendsto be analyzedmeaningfully; thrushes(X2= 16.9,3 df, P < 0.01;Fig. 1). Overlarger samplesare necessaryto compensatefor lap between the speciesin selectionwas 0.73.
Moreover, LouisianaWaterthrusheshad signifthe heterogeneous distribution of invertebrates. Because! could not identify some prey icantly longer bills (1.43 + 0.10 cm, n = 10)
in smallsize classes,
my findings are alsobiased than Northern Waterthrushes (1.29 + 0.06 cm,
n = 20; t = 4.9, P < 0.01). In comparisonsof 3
in favor of larger taxa.
Other prey types eaten included Odonata males/speciesI found significantindividual dif(dragonfly) nymphs, Dytiscidaelarvae, Tipuli- ferences in prey-size selection by Louisiana
Waterthrushes (x 2 = 18.5, 3 df, P < 0.01) but
no significant differences between Northern
TABLE
2. Percentageuseof foragingmethodsby two
waterthrush species.Observed foraging episodes Waterthrushes(X2 = 6.5, 3 df, P > 0.05). Again,
I could not assessthe relative importance of
are in parentheses.
individual preferencevs. flight-cageplacement
Seiurus
Seiurus
in determining individual differences.
motacilla
noveboracensis
I found little difference between the species
Pick
50.5 (435)
53.4 (945)
in attack rate (t = 1.4, 144 df, P > 0.05, logLeaf pull
Hawk/hover
Total
48.7 (419)
0.8 (7)
861
45.9 (812)
0.7 (12)
1,779
transformed
data; Louisiana Waterthrush:
10.1 + 8.0 s/attack, n = 67; Northern Water-
thrush: 9.1 + 8.9 s/attack, n = 79). Analysis of
April 1987]
Waterthrush
PreySelection
183
TABLE
3. Taxonomiccompositionof aquaticinvertebratesin samplesfrom swamppoolscollected5 weeks
apart (mean no./m 2 + SD).
Sample
Taxon
Isopoda
Plecoptera
Ephemeroptera
Trichoptera
Diptera
Coleoptera
Megaloptera
Gastropoda
Oligochaeta
April
May control sites
May exclosuresites
167.4 -+ 117.3
54.2 + 81.6
24.2 + 32.3
50.4 + 45.2
32.6 + 23.6
14.0 + 6.8
7.2 + 5.2
7.8 + 8.7
5.2
150.8 + 82.2
94.2 -+ 121.2
64.2 + 72.9
36.4 + 27.3
41.6 + 36.1
11.0 + 6.1
12.8 + 14.4
18.0 -+ 12.6
0.8
83.0 + 45.9
43.0 + 32.6
33.3 + 51.4
35.8 + 32.0
80.5 -+ 85.3
3.3 + 2.4
12.5 + 12.9
2.5 + 1.8
0.2
Odonata
Hirudinea
0.4
0.2
0.2
0.6
0.2
0.8
Desmognathus
0.4
0
0.2
variance on a randomly selected,balancedsub- of data from 3 males/speciesagain showedno
set of data from 3 males/species,which facili- significantinterspecificdifference(F < 0.1, 1,30
tated performing F testsand eliminated possi- df, P > 0.05) but a significantdifference beble
variance
because of sexual
differences,
tween individuals (F = 12.0, 4,30 df, P < 0.01).
confirmedthe nonsignificanceof interspecific The relative importanceof flight-cage placedifferences (F < 0.1, 1,78 df, P > 0.05) and also ment and individual preference in determinshowed no significantdifferencebetween in- ing individual differences could not be asdividuals (F = 3.1, 4,78 df, P > 0.05).
sessed.
In the flight cage the speciesused foraging
methods (Table 2) in proportions similar to
those of free-ranging birds in early spring
(Craig 1984). Overlap in use of methods was
0.97, compared with 0.98 calculated from data
collected on free-ranging birds from 1978 to
1980 (Craig 1984). The species did not differ
significantly from each other in their use of the
methods(X2 = 2.0, 2 df, P > 0.05), and the percentage of picking for observations >30 s
showed no significantinterspecificdifference
(t = 1.5, 132 df, P > 0.05, arcsineX/--trans-
Although dataon foragingmethodssuggested that cagedbirds behavedsimilarly to freeranging waterthrushes, I compared data
gathered from the first 45 min of flight-cage
formed data; Louisiana Waterthrush: 68.4 +
30.5%, n = 55; Northern Waterthrush: 64.7 +
observations with those of the last 45 min to
determine if duration of observations affected
results.Resultswerenonsignificantfor prey size
(Louisiana Waterthrush: X2 = 0.7; Northern Wa-
terthrush:X2 = 6.8; both 3 df, P > 0.05), attack
rate (Louisiana Waterthrush: t = 0.3, 35 df;
Northern Waterthrush: t = 0.7, 64 df; both P >
0.05, log-transformeddata) and foraging methods (Louisiana Waterthrush: t = 0.2, 23 df;
Northern Waterthrush: t = 0.2, 62 df; both P >
28.6%,n = 77). Analysisof varianceon a subset 0.05,arcsinex/--transformed
data).Resultswere
TABLE4. Size distribution of invertebrate samples(mean no./m 2 + SD).
Sample
Size class(mm)
April
May controlsites
4.0 + 3.0
147.2 + 122.2
170.4 + 69.8
May exclosuresites
-<4
>4-7
>7-10
8.0 + 5.6
84.2 + 90.4
118.4 + 76.2
3.5 + 2.6
90.5 + 77.4
109.3 _+ 32.1
>10-13
>13-16
66.8 _+ 47.5
24.4 + 17.2
66.2 + 37.0
25.0 ___20.4
56.0 + 24.2
16.5 + 9.4
>16-19
11.0 + 10.2
7.4 _+ 8.3
7.8 + 6.5
>19-22
>22
8.4 + 10.1
11.2 + 10.1
8.0 + 9.2
13.8 + 13.6
8.3 _+ 7.2
12.8 + 7.0
184
ROBERT
J. CRAIG
TABLE
5. Ranking of prey selectionby waterthrushes.
Seiurus
Seiurus
Abundance
rank a
novebora-
motacilla
DifferUse b ence c
censis
Use b
Differencec
-<7
>7-13
2
1
3
2
+1
+1
2
1
0
0
>13-19
>19
3
4
1
4
-2
0
3
4
0
0
1
5
+4
5
+4
2.5
3.5
+1
2
-0.5
2.5
4
5
1
2
3.5
- 1.5
- 2
- 1.5
3
1
4
+ 0.5
- 3
- 1
Taxa
tera
Trichoptera
Diptera
Oligochaeta
• Abundance
pods (T = 48), and Coleoptera(T = 39, all P >
0.05)were significantlylesscommoninsidethe
exclosuresthan at unprotected sites.The Diptera averaged considerablymore common inside the exclosure,but not significantlyso (Table 3). For size classes,
only the >7-10-ram size
classwas significantly lesscommon inside the
screen exclosures (Table 4), which was due
Prey size (mm)
Isopoda
Ephemerop-
[Auk, Vol. 104
rank in relation
to other sizes or taxa, with
most fre-
quent classesreceiving lowest numbers.
bFrequencyof usein relation to other sizesor taxa.
• Computedby subtractinguse from abundanceranks;positivevalues indicaterelative avoidance,negativevaluesindicaterelative pref-
principallyto reducednumbersof isopods(T =
51, P < 0.05), the most numerousgroup in this
class.
Analysisof resortedinvertebratesamplesindicated that initial sorting removed >95% of
the individuals and biomass present. Resampling the same site showed that >95% of organismsand biomasswere in the initial sample, and therefore biomassfigures should be
increasedby at least 10%to obtain absoluteestimates.
By averagingdata from April and May control samples,I determined the abundancerank
of invertebrate
size classes and the 5 most fre-
quent waterthrush prey (Table 5). Because
Ephemeropteraand Trichoptera were nearly
also nonsignificantfor LouisianaWaterthrush equally abundantin samples,I consideredtheir
useof Trichoptera(adjustedx• = 2.1, P > 0.05), rank to be tied. Including only the mostcombut significant for Northern Waterthrush use mon prey is valid in this nonparametricmethof Diptera (adjustedx2 = 6.9, P < 0.01). During od becauseonly relative preferencesare deterthe first 45 rain 35.7% of prey were Diptera, mined (Johnson1980). Comparisonof use and
whereasduring the second45 rain 63.9%of the abundance ranks showed that Louisiana Waprey were Diptera.
terthrushesavoidedsmallprey relativeto large
Prey.--Basedon averagesfrom April and May prey, most strongly preferred Diptera larvae,
samplestakenoutsideexclosures,
the principal and strongly avoided isopods. In contrast,
erence.
aquatic invertebrates present as potential prey
Northern Waterthrushesshowed no preference
during the spring were isopods (Asellus),
Ephemeropteranymphs,Diptera (mostlyChironomidae)larvae, and Trichopteralarvae. No
significantchangeoccurredin the frequencyof
taxabetweenApril and May (Table 3), nor in
biomass(April: 1.46 _+ 0.85 g, May control:
1.21 -+ 0.51g; t = 0.3, 18 df, P > 0.05,log-transformed data) or size frequency (Table 4). Although data were not comparablequantitative-
for prey size, moststrongly preferred Diptera
larvae, and strongly avoided isopods.
Territoryandhabitat.--In 1985 2 pairs of Louisiana Waterthrushesand 6 pairs of Northern
ly because
samplingmethodsdiffered,Culicidae
larvae, Helodidae larvae, and Gastropoda(Physa) were much less common but Plecoptera
nymphswere much more commonthan they
were from
1978 to 1980.
Waterthrushes
inhabited
Boston Hollow.
This
was the same as in 1984, and differed little from
numbers present in 1979 (Louisiana Water-
thrush:3 pairs,Northern Waterthrush:6 pairs)
and 1980 (Louisiana Waterthrush: 3 pairs,
Northern Waterthrush:5 pairs).As in previous
years, in 1985 both speciesexhibited intense
intraspecific aggression, no interspecific
aggression,and extensiveterritorial overlap.
Multiple regressionanalysisof data on prey
Comparisonof unprotectedsiteswith exclo- biomass,water cover, and territory size from 8
sures(1.00 _+0.25 g) showedno significantdif- territories per speciescollected from 1978 to
ference in invertebrate biomass (t = 0.9, 18 df, 1980 (data reportedin Craig 1984) showed no
P > 0.05, log-transformeddata). However, is- significantassociationsfor either species(Louopods (Wilcoxon two-sample T = 52), gastro- isianaWaterthrush:F = 3.8, multiple r = 0.78,
April 1987]
Waterthrush
PreySelection
185
P > 0.05; Northern Waterthrush: F = 1.5, mul-
siana Waterthrush (Craig 1985) increases its
tiple r = 0.60,P > 0.05). LouisianaWaterthrush
territories had the highest correlation with
metabolicrequirements(Calderand King 1974),
and likely drivesits preferencefor larger prey.
This greater selectivity may also explain the
preference by Louisiana Waterthrushes for
habitats with fast-moving water (Craig 1985),
which contain a higher density of large invertebrates than do habitats with slow-moving
water cover (r = -0.63), but those of Northern
Waterthrusheswere mosthighly correlated(r =
-0.53) with prey biomass.
DISCUSSION
water (Craig 1984).
The similarity in useof foraging methodsbetween captive and free-ranging birds strongly
suggeststhat the flight cagehardly altered foraging behaviorof residentsand thereforewas
useful in observingthe detailsof foraging. The
similarityof behaviorrecordedduring the first
and second halves of each observation period
supported this view. However, predation was
limited to invertebratespresent in the cage.Because invertebrate distribution is heterogeneous (Hynes 1970, this study), additional investigationsare required to assessthe influence
In both speciespreferencefor prey type is
probably in large part a consequenceof preference for prey size. For example, average
April-May data on invertebratesshowed that
63.8% of Trichoptera larvae available are >13
mm
and
100% of Chironomidae
larvae
avail-
able were -<13 mm. However, certain prey, notably isopods,were strongly avoided regardlessof size. Isopodsburrow, which may make
them more difficult to catch,and they alsomay
be relatively unpalatable. Their burrowing
ability also may explain why they were less
of invertebrate
microdistribution
on individual
common inside the exclosures.Similarly, the
behavior. The lone significant difference for slow-moving,thick-bodiedTrichopteraare very
Northern Waterthrush predation on Diptera profitable prey (Ormerod 1985a),and thus parduring the first and secondhalves of the ob- ticularly suitable for LouisianaWaterthrushes.
servationperiod may be a reflectionof this hetMy findings on prey selectionmay provide
a mechanismfor coexistenceby these ecologierogeneity.
Despite this unquantified sourceof variance, cally similar species.Indeed, the findings illusseveralmajor differencesin foraging were ev- trate a degree of ecologicalseparationbetween
ident. The speciespreferreddifferent prey types the speciesundemonstratedpreviously,but for
and different rangesof prey size. Consequent- partitioning to be necessaryresourcesmust be
ly, overlap between the speciesin selectionof limiting. In exclosureexperimentsI found that
prey size and type was considerablyless than the waterthrushesappearedto have little clear
that previously calculated for foraging meth- effect on the biomassor compositionof their
odsor selectionof foraging microhabitats(Craig prey populations.A general decline in biomass
1984). Differencesin prey use are probably re- and numbers comparedwith control samples
lated to the prey handling capacity of the indicated possiblenegative effectsof the exclospecies; the larger bills of Louisiana Water- sures on invertebrates, but such effects were
thrushes should make them more effective at
limited principally to groupsnot heavily used
handling larger prey (Hespenheide1973)than by waterthrushes. Certainly, waterthrushes
Northern Waterthrushes. In contrast to prey must affect invertebrate levels to some extent,
use, measuresof attack rate and picking fre- evidence of which is provided by the increase
quency showed these behaviors to be highly (nonsignificant) of Diptera inside exclosures.
variable and with little interspecificdifference. However, only majorinfluencesare detectable
Unlike many other speciesof Parulinae (e.g. because,as indicated by the large standarddeMacArthur 1958) foraging differences in wa- viations in Tables 3 and 4, the distribution of
terthrushes are principally in prey selection invertebrateswas very heterogeneous.Neverrather than in method of locating prey.
theless,if the speciesreducedprey to limiting
The two speciesexploited the same inverte- levels, such a major drop in invertebrate popbrate prey resourcedifferently in relation to its ulations most likely would be observed.
abundance.
The Louisiana
Waterthrush
was
Other studies (e.g. Holmes et al. 1979,
relatively selectivecomparedwith the North- Schneider and Harrington 1981) have shown
ern Waterthrush. The larger size of the Loui- major impactsof birds on invertebrate prey,
186
ROI•ERT
J. CRAIG
even in experiments where prey could leave
foragingexclosures(Askenmoet al. 1977).The
impact disappearsduring peaksof invertebrate
populations,however,when birdscanno longer trackthe increasein availability(Otvos 1979).
I proposethat this mechanismaccountsfor my
failure to detectappreciableimpactby the waterthrusheson their prey. Eventhough I report
that population levels of at least several prey
taxa were particularly low in Boston Hollow
during 1985, the high productivity of wetlands
is well known (Keefe 1972). During the 5
weeks between sampling datesCulicidae matured from eggsto adultsand there were major
hatchesof Plecoptera,Ephemeroptera,Simuliidae, and Chironomidae.
Furthermore,
water-
thrushesalsoeat the foliage insectsthat begin
to appear toward the end of the period, so it
appears that the few birds present had little
impact on their prey.
The weak associationof prey-related factors
with territory size,which shouldexpandasprey
or feeding habitat decline if they are limiting,
provides additional support for the idea that
neither speciesis limited by prey. Instead it
appearsthat territory size is more closelyrelated to social factors; I have observed territories
of both speciesbeing halved after the arrival
of new territorial
males. The constant
number
[Auk,Vol. 104
observationsit is impossibleto rule out the hypothesisthat observedpatternsare the product
of independentevolutionaryspecialization,as
opposedto the result of interspecificcompetition.
ACKNOWLEDGMENTS
I thank G. A. Clark, Jr., S. W. Pacala,and P. H. Rich
for their many useful commentson this project. D.
M. Smith gave permissionto use Yale Forestas my
studyarea,J.A. Kuchleprovidedaccess
to laboratory
facilities, and C. Henry reviewed the manuscript.I
obtainedfinancialsupportfrom the E. A. Bergstrom
Memorial Fund and the ConnecticutDepartment of
EnvironmentalProtection.I especiallythankmy wife,
Susan,for her assistance
with invertebrateanalyses
and field measurements.
LITERATURE CITED
ASKENMO, C., A. VON BROMSSEN,J. EKMAN, & C.
JANSSON.1977. Impact of somewintering birds
on spider abundance.Oikos 28.'90-94.
CALDER,W. A., & J. g. KING. 1974. Thermal and
calorie relations of birds. Pp. 259-413 in Avian
biology,vol. 4 (D. S. Farherand J. R. King, Eds.).
New York, Academic Press.
CONFER,
J. L., & K. KNAPP. 1981. Golden-winged
Warblersand Blue-wingedWarblers:the relative
successof a habitat specialistand a generalist.
of waterthrush pairs present over 4 years despite the dry winter and spring of 1985 further
Auk 98: 108-114.
argues against close coupling of populations CRAIG,R.J. 1984. Comparativeforagingecologyof
Louisiana and Northern
waterthrushes.
Wilson
with prey abundance.This is in contrastto the
Bull.
96:
173-183.
findings of Ormerod (1985b),who found that
1985. Comparative habitat use of Louisiana
the population density of the ecologicallyconand Northern
waterthrushes.
Wilson
Bull. 97:
vergent Curasian Dipper (Cincluscinclus)was
347-355.
related to prey density. The Dipper is a permanent resident, however, so its populations
are more likely to be closely associatedwith
local conditionsthan are those of the neotropical migrant waterthrushes.
The
Louisiana
and
Northern
waterthrushes
EMLœN,
J. T. 1981. Divergencein the foraging responsesof birdson two Bahamaislands.Ecology
62: 289-295.
HESPENHEIDE,
H.A. 1973. Ecologicalinferencesfrom
morphologicaldata.Ann. Rev. Ecol.Syst.4: 213229.
showedimportant differencesin prey usageat
BostonHollow, although they used similar be-
HOLMES,R. T., J. C. SCHULTZ,& P. NOTHNAGLE.1979.
Bird predation on forest insects:an exclosureex-
haviors to locate prey. These differences indicate that exploitive competition can occur under certain conditions, and that it may be
avoided by divergence in resource use. Evidence for prey limitation was weak, however,
so there was no clear need to avoid competition. Moreover,as in earlier studies(Craig 1984,
periment. Science206: 462-463.
HYNES,H. B. N. 1970. The ecology of running
1985), there was no indication of interference
competitionbetween the species.Basedon these
waters. Toronto,
Univ.
Toronto
Press.
JEHL,J. R., JR.,& K. C. PARKES.1983. "Replacement"
of landbird specieson SocorroIsland, Mexico.
Auk 100: 551-559.
JOHNSON,
D. H. 1980. The comparisonof usageand
availability measurements for evaluating resourcepreference.Ecology61: 65-71.
KEEFE,
C.W. 1972. Marsh production:a summaryof
April1987]
Waterthrush
Prey
Selection
187
the literature. Contrib. Marine Sci., Univ. Texas
Kroll, and J. A. Jackson,Eds.). New York, Aca-
16: 163-181.
demic
Press.
S.K. 1981. Ecologicalrelationsand social
MACARTHUR,
R.H. 1958. Populationecologyof some ROBINSON,
warblers of northeastern
coniferous
forests.
interactionsof Philadelphiaand Red-eyedvireos. Condor 83: 16-26.
Ecology39: 599-619.
ß& R. T. HOLMES.1982. Foragingbehaviorof
1V[AURER,
B. A. 1984. Interference and exploitation
in bird communities. Wilson Bull. 96: 380-395.
forestbirds:the relationship
of searchtactics,
diet,
MOUNTAINSPRING,
S., & J. M. SCOTT. 1985. Interspe-
cific competitionamongHawaiian forestbirds.
Ecol.Monogr. 55: 219-239.
NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRA-
TION. 1985. Climatological data, annual sum-
and habitat structure.Ecology63: 1918-1931.
ROTENBERRY,
J.t. 1980. Dietary relationshipsamong
shrubsteppe
birds:competitionor opportunism
in a variableenvironment.Ecol.Monogr. 50:93110.
SCHNESDER,
D.C., & B. A. I-•ARRINGTON.1981. Timmary:New England97: 13.
ing shorebirdmigrationin relationto prey deORMEROD,
S.J. 1985a.Optimalforagingby breeding
pletion. Auk 98: 801-811.
Dipperson a Lancashirehill-stream.Naturalist
110: 99-103.
SCHOENER,
t. W. 1970. Non-synchronousspatial
overlapin lizardsin patchyhabitats.Ecology51:
1985b. Factorsinfluencing the abundance
of breedingDippersCinclus
cinclus
in the catchmentof the RiverWye,mid-Wales.Ibis 127:332340.
OTVOS,I. S. 1979. The effectsof insectivorousbird
408-418.
WI•s, J.A. 1977. On competitionand variableenvironments.
Amer. Sci. 65: 590-597.
ZACH,R., & J. B. FALLS.1976. Biasand mortality in
the use of tartar emetics to determine the diet of
activitiesin forestecosystems.
Pp.341-374in The
role of insectivorousbirds in forest ecosystems
Ovenbirds (Aves: Parulidae). Can. J. Zool. 54:
(J. G. Dickson, R. N. Conner, R. R. Fleet, J. C.
1599-1603.