J. Field Ornithol., 55(2):160-180
THE EFFECT OF DITCHING
ON
SALT
FOR MOSQUITO CONTROL
MARSH
IN ROWLEY,
USE
BY BIRDS
MASSACHUSETTS
BY Jo ANN CLARKE,BRIAN A. HARRINGTON,THOMASHRUBY,
AND FRED E. WASSERMAN
Historically, mosquitopopulationshave been reducedby destroying
their breedinghabitat in saltmarshesand freshwaterwetlandsthrough
drainageor constructionof impoundments.During the 1930's,90% of
the Atlantic coastalmarsheswere grid-ditchedto drain surfacewater
where mosquitosbreed (Bourn and Cottam 1950). Ditcheswere dug at
50 to 100 m intervalsthroughoutthe marsh,connectingperpendicularly
to common
tidal channels.
Additional
ditches
were often
added
to the
systemto drain specificpoolsand pans.
By increasingthe area of marsh that is regularly flooded, ditches
increaseanimal populationsthat exploit the intertidal habitat (Teal
1962, Ferrigno 1970, Kuenzlerand Marshall1973).Effectsfarther from
the ditches,however,are not sowell understood.Someauthorsreport
increasedpopulationsof invertebrates(Rockel 1969, ShislerandJobbins
1975, Lesseret al. 1976),ducks,and muskrats(Corkran 1938) on ditched
sites.Others (Bradbury 1938, Headlee 1939, Travis et al. 1954) have
claimedlittle or no changein wildlife. Bourn and Cottam (1950) reported a significantdecreasein invertebrate populationson Delaware
marshesafter drainageof marshpools.Negative effectson marshuse
by shorebirds,waterfowl,and wadingbirds havebeen reported in midAtlanticstates(Urner 1935, Ferrignoet al. 1975),andin southernNew
England(Reinert et al. 1982).
The purposeof thisstudywasto determineif grid-ditchingaltersbird
use of Massachusetts salt marshes, and if so, whether the alteration in
bird use correspondsto changesin invertebrate abundances.Data and
conclusions
presentedhere are basedon quantitativeobservations
made
during the summermonthsof 1982 on 2 ditchedand 3 unditchedsites
in the Rowleytownship,EssexCounty.
STUDY
AREAS
AND
METHODS
Studyareas.--The studyareasselectedrepresent:(1) saltmarshesclose
to coastalbaysand oceaninlets, with little fresh water input, and (2)
backwatersalt marsheswith a lesssalinecharacter(Fig. 1). Vegetation
on the firstmarshtype wasrestrictedto Spartinaalterniflora(cordgrass),
S. patens(salt marsh hay), Distichlisspicata(spikegrass),and Limonium
carolinianurn
(sealavender)on higher spots,and Salicorniaeuropea
(glass
wort) around pools.The only encroachingupland vegetationwasoccasionalIvafrutescens
(marshelder) on elevatedditch spoil or Chenopodiumalbum(lambsquarter)on marshedges.Vegetationon the backwater
160
Vol.55,No.2
Effect
ofMosquito
Ditching
onBirds
[ 161
FIGURE1. Upland (stippled)and marshland(open) habitat in Rowley, Massachusetts.
Blackrectanglesindicatelocationsof five 3-haplotsstudiedduring the summer,1982.
Plots1A, lB, and 2 (coastalmarshes)are part of the RowleyRiver drainagesystem;
Plots 3 and 4 (backwatermarshes)drain at the head of the Mill River. For scaleeach
plot is 100 x 300 m.
marsheswas more varied. S. alterni,
fiora, S. patens,and D. spicatawere
still characteristic,but many areassupporteddensestandsof Potentilia
egedei
(silverweed),
Juncusgerardi(blackgrass),Eleocharis
parvula(spikerush),Plantagooliganthos
(seasideplantain),and variousgrasses.Long
rows of I. frutescens
and the sedgesScirpusamericanus
and Cyperus
polystachyos
bordered creeksand ditches,and Typhusangustifolia(narrowleaved cattails)encroachedon the marsh edges.
Observationswere carriedout within 100 x 300 m plotsduringJuneAugust, 1982; 3 plots were laid out on the coastalmarsh (1A, lB, and
2 in Fig. 1) and 2 on the backwatermarsh (3 and 4 in Fig. 1). Plot 1A
had beengrid-ditchedin the past,but not recently.In the centerof the
plot, ditcheshad cloggedsufficientlyfor a small,shallowpool systemto
form. In Plot 1B, whichwascontiguouswith 1A, we found evidenceof
old ditches,but mostwere completelyfilled and a systemof deeppools
stretchedapproximately% the length of the plot. Plot 2 was crossed
aboutevery50 m by well-maintainedditchesand containedonly2 small
pools,although round sunkenareasfilled with S. alternifloraindicated,
as suggestedby Rockel (1969), that poolshad existedprior to recent
ditch maintenance.In our analyses,we consideredthe conditionof Plot
1A as neglectedditches,Plot lB as no ditches,and Plot 2 as recently
maintained
ditches.
The backwatermarshplotshad more complexditchinghistories.The
162]
J. A. Ctarke
etat.
j. Field
Ornithol.
Spring 1984
ditchesin Plot 3 were irregularly placed, often clogged,and failed to
drain its 2 smallpool systems.Ditching in Plot 4 was more regular.
Although standingwater wasoften found in the depressions
between
grasshillocks,there were no true pools.These siteswere consideredas
unditchedand ditched, respectively.
Bird censuses.--From
30 June to 10 July birds were censusedonce a
week on each plot. From 11 July to 31 August, the major migration
periodfor mostshorebirdsin the area,countswereincreasedto 5 times
a week, then reducedto twicea week for the first 2 weeksof September.
Countswere madeat different timesof the day and at varyingstagesof
both daily and seasonaltidal cycles.
A typicalcensusconsisted
of walkinga rectangularpath 30 m inside
each plot edge at approximately21 m/min averagewalking rate (25
min per census).
All birdsspottedor heardsingingwithinthe plot during
the censuswere identified by speciesand classifiedas (1) foraging or
non-foraging,and (2) in-poolsor not-in-pools.Birdsflying over the plot
were not recordedunlessforaging on the wing.
Invertebrate
censuses.--Invertebrates
movingaroundon the surfaceof
marsheswere trappedin plasticpitfall traps(10 cm wideat the mouth,
7 cm wide at the base,and 15 cm tall) with 3 cm of water in the bottom.
The trapswere buried flushwith the soilsurfaceat approximately25
m intervalsalongtransectson the 50, 150, and 250 m linesof eachplot.
This resultedin 9 pitfall samplesper plot. When a trap site locatedby
this method ended in a permanent pool, it was moved to the nearest
pool edge.Traps were placedfor 48 h during neaptidesfrom 14 to 16
July and 15 to 18 August.In spiteof the lowertides,someAugusttraps
were flooded,resultingin fewer than 9 pitfall samplesper plot at that
time.
Soildwellinginvertebrateswere collectedfrom substratecoresusing
a sectionof 10 cm diameter polyvinylchloride (PVC) pipe with a sharpened lower edge. The pipe wasdriven 4 cm into the marsh. Inasmuch
asthe majority of infaunalmarshinvertebrateslive in the top 1 to 4 cm
(Wieserand Kanwisher1961, Bell et al. 1978, Coull and Bell 1979), this
depth wasconsideredsufficient.Half of the sod plug wasplacedin a
Berlesefunnel for 48 to 82 h where heat and light drove the invertebratesout of the soilinto a collectingjar containingformalin.The other
half of the sodplug wasdiscarded.Core siteswere locatedalternately
10 m to the left and right of the pitfall sitesin eachplot. Sampleswere
collectedfrom 24Juneto 12July,andagainfrom 2 Augustto 17August.
This resultedin 18 core samplesper plot.
Pool dwellinginvertebrateswere capturedusinga 10-cmdiameter
PVC pipe with a slidingbottom panel. The pipe wassubmergedin a
pool and allowedto settlethrough about2 cm of bottom flocculus.The
slidingdoor wasthen pulledshutby an attachedline and the sampleof
bottom flocculusand standingwater lifted out and poured through a
0.5 mm mesh. Invertebratescaught in the sievewere collectedand
preserved.The sizeof the meshprecludedquantitativesamplesof nematodes,minute copepods,and smallmite species.Pool sampleswere col-
Vol.55,No.2
Effect
ofMosquito
Ditching
onBirds
[ 163
T^BLE 1. Speciesand abundanceof birds sightedon the 5 Rowley plots during 38
censuses,
July-September 1982. Common namesfollow the American Ornithologists'
Union (1983).
Plot
Species
1A
Great Blue Heron
--
Great Egret
SnowyEgret
Little Blue Heron
Green-backed
Heron
1B
2
Total#
3
4
sighted
I
12
1
4
120
3
27
38
7
--
4
-34
4
67
---
-18
--
I
10
26
24
---
--
-1
Black-crownedNight
Heron
--
GlossyIbis
American
Black
Red-tailed
American
Merlin
Hawk
Kestrel
Biack-bellied
4
--
--
--
4
6
10
--
13
--
29
1
6
--
3
--
10
12
2
1
29
15
1
Duck
---
--
2
11
13
--
--
14
--
SemipaimatedPlover
--
51
--
Killdeer
--
--
--
Greater Yellowlegs
LesserYellowlegs
SpottedSandpiper
SemipalmatedSandpiper
LeastSandpiper
22
46
12
3
95
Reeve
--
Ring-billedGull
Herring Gull
3
--
Common
Plover
4
-....
Tern
39
55
25
125
336
--
14
58
41
--
41
51
139
4
173
178
--
112
241
43
301
616
--
4
8
--
--
I
1
I
1
-1
2
--
2
I
3
3
--
12
12
131
--
--
--
134
Least Tern
--
20
--
--
--
20
Mourning Dove
Chimney Swift
Belted Kingfisher
-3
--
---
----
2
-4
3
---
5
4
4
Northern
--
--
--
I
--
I
Flicker
EasternKingbird
Tree Swallow
Bank Swallow
Barn Swallow
BlueJay
American
3
--I
-6
7
11
1
4
30
9
140
90
31
142
89
17
--
117
58
33
119
35
14
60
12
21
--
--
--
I
--
47
--
7
--
60
1
14
--
14
1
Crow
6
Marsh Wren
Brown Thrasher
....
--
--
--
EuropeanStarling
Sharp-tailedSparrow
SongSparrow
16
30
3
20
16
--
51
13
1
Common
Bobolink
---
---
-50
Yellowthroat
Red-wingedBlackbird
Common
American
Grackle
Goldfinch
27
I
--
6
---
54
578
284
116
1
235
50
31
-94
11
322
203
46
2
11
1
53
3
114
33
76
73
215
I
52
3
29
1
83
4
--
164]
J. A. Clarke
etal.
IA
lB
2
J.Field
Ornithol.
Spring 1984
3
4
283
IA
(20.7)
lB
2
239
(•?.8)
36.8
(•8.6)
22.6
(37.3)
lOB
(!0.4)
(24.2)
16.6
I•.0
• (•4.5)
•.6
••
•0.5
51.•
51.•
(•z.5)(•4)
4
([•.6,) (14.5)(24.7)
Birds
FIGURE2. The mean similarity(Jaccardcoefficient)of bird censuses
paired within and
betweenplots.Numbersindicatemean valuesand standarddeviationsare in parentheses.Asterisksindicatebetweengroup similaritiesthat were statisticallydifferent
(P < .001) from within groupsimilarities.Comparisons
were madebetweensimilar
marshtypeswith different ditchinghistoriesor betweendifferent marshtypeswith
similarditchinghistories.Comparisons
betweenditchedand unditchedsiteson dissimilarmarshtypeswere not made,as indicatedby the blank squares.
lectedfrom 8 to 26 July; 16 sampleswere collectedfrom poolsin Plot
1A, 13 from Plot lB, and 13 in Plot 3. Random measurements of
temperature,salinity,andpH of the poolsystems
weremadethroughout
the periodof invertebratecollectionusinga Yellow SpringsInstruments
meter (33 set).
Predationbybirdsin saltmarshpools.--One pool from Plot 1A and one
from
Plot
lB were used for exclosure
studies to examine
the effect
of
bird predationuponfishandinvertebratepopulationsof the pools.The
1A pool wasshallow(<25 cm deep) with graduallyslopingsidesand a
surfacearea of approximately12 m2. The pool on Plot lB wasover 50
cm deep with steepsidesand a surfacearea of 10 m2. Both poolswere
divided into approximatelyequal sectionsby extending 1 mm mesh
fiberglassscreeningacrossthe pool from the pool bottomto 5 cm above
the water'ssurface;thispreventedmovementof minnowsand macroinvertebratesbetweenhalves.Dupont VEXAR 2 cm plasticutility netting
wasplacedapproximately50 cm abovethe water surfaceto exclude
birds from one half of eachpool. The closedhalvesservedas controls
for the open experiments.
Vol.55,•o.2
Effect
ofMosquito
Ditching
onBirds
[ 165
T^BLE2. Useof saltmarshwith pools(unditched)andwithoutpools(ditched)in Rowley,
Massachusetts
by birdsof sevenforagingguilds,June-September1982.
Total # per
ha in plots
with poolsa
Guild
Herons and ibises
Shorebirds
Gulls
Terns
25
158
2
17
Total # per
ha in plots
Total exwithout
pected#
poolsb
in eachplotc
1
7
1
0
13
80
2
9
Chi-square
value
23.1 ***
152.5***
0.6 ns
17.1 ***
Swallowsand kingbirds
80
52
66
5.7*
Blackbirds
49
48
49
0.01 ns
Other songbirds
22
23
22
0.01 ns
Plots 1A, lB, and 3; a total of 9 hectares.
Plots 2 and 4; a total of 6 hectares.
Calculatedaccordingto the assumption,Ho: pools(unditched)=
no pools(ditched).
*** Differenceat significance
level of P • .001.
* Differenceat significance
level of P • .05.
nsDifferencenot statistically
significant.
To measure initial invertebrate abundance, water column samples
were collectedfrom each experimental and control area. The numbers
of minnowspresentin eacharea were estimatedby placingan 11 x 15
cm piece of light gray plexiglasson the pool bottom and countingall
minnowsswimmingacrossit during 10 min. Minnowswere countedin
eachpool half. Netting wasleft over the poolsfor 18 days(25 July to
13 August). At the end of this time, we again counted minnowsand
sampledcontrol and experimental halvesfor invertebrates.
Similarityand statisticalanalysis.--The invertebrate and bird populationsfrom the studyplotswere comparedusingsimilarityanalysis.This
procedure calculatesa similarity coefficientfor each pair of samples
yieldinga final matrix of coefficientsthat compareseach samplewith
every other samplein the analysis(Field and McFarlane 1968). To
compareinvertebratesamples,the similaritycoefficientdescribedby
Brayand Curtis(1957) wascalculatedafter logarithmictransformation
of the abundancedata. The very wide abundancerange of somespecies
madelog transformationnecessary.
To comparebird populationsfrom
plot to plot, the Jaccard coefficient was used (•accard 1902) which
considersonly speciespresenceas the criterion of similarity.
The mean similaritiesbetweenreplicatesampleswithin eachplot and
between samplesfrom different plots were calculatedfrom the matrix
of coefficients.
The
Student's
t-test was used to determine
if the mean
similarityof samplesfrom 2 different plots wassignificantlylessthan
either within-plot mean. If this were true, then the two siteswere consideredecologicallydistinct.
Differencesin total invertebrateabundancesin eachplot were tested
usinga one-wayanalysisof variance.Pair comparisonswere made with
166]
J. A. Clarke
etal.
j. Field
Ornithol.
Spring 1984
v
I
v
I
v
v
v
v
I
I
I
I
Vol.
55,No.2
Effect
ofMosquito
Ditching
onBirds
[ 167
168]
J. A. Clarke
etat.
j. Field
Ornithol.
Spring 1984
a Student NewmanoKeulstest (SNK) (Zar 1974). Total seasonalabundancesof birdson the plotswere compareddirectly,and apparentpreferencesof different avian guildsfor the ditchedor unditchedcondition
weretestedusingchi-squareanalysis.In the poolexclosurestudiesand
associatedcontrols,the mean populationof invertebrate and fish taxa
beforeexclosurewascomparedto that after 1$ daysof exclosure.The
Student's
t-test was used to determine
if an observed
difference
was
significant.The significancelevel appliedin all testswasP • .05 unless
otherwise
indicated.
RESULTS
Bird censuses.--Thebird speciesrecorded in each plot are listed in
Table 1. The similarityanalysis(Fig. 2) indicatesthat speciesspotted
on Plot lB were different from those on Plots 1A and 2; likewise, the
birds on Plot 3 were distinctas a group from thoseon Plot 4. Thus,
bird useof activelyditchedsites(no pools)differed from useof sites
with no ditches or neglectedditches (pools).Similarity analysisalso
revealed differencesbetween speciesof birds spotted on coastaland
backwatermarshes.Bird populationson Plots1A and 1B were different
from those on Plot 3, and Plot 2 was different from Plot 4.
When speciesabundanceis included in the analysis,the differences
betweenbird populationsof ditchedand unditchedsitesare evenmore
apparent.The greatestnumberof individuals,regardlessof guild, was
seenon the unditchedplot with pools.When a specieswasalsoobserved
on Plot 1A (neglectedditches)the numberswere alwayslessthan on
lB. Most specieswere absentfrom Plot 2, the activelyditchedsitewith
no pools.The exceptionswere Leastand Spottedsandpipersand Ringbilled Gulls. In these3 species,however,the number of birds seenon
Plot 2 wassmallcomparedto the number foragingon Plots1A and 1B.
A comparisonof shorebird, heron, tern, and gull abundanceson the
backwater marsh indicates that, as on the coastal marsh, these birds
foraged almostexclusivelyon Plot 3, the unditchedplot with pools.
Exceptionswere the Green-backedHeron, whichfrequentedthe tall S.
alternifloraalong the ditchesof Plot 4, and the Ring-billedGull.
A chi-squareanalysis(Table 2) alsoindicatesthat shorebirds,herons,
and other marine-orientedbirds were attractedby marshpools.Aerial
insectivores
suchasswallows
and kingbirdsalsopreferredthe unditched
plotsfor foraging.The patternof occurrenceof hawks,songbirds,
and
other specieson the plots,however,did not appearcorrelatedwith the
ditchedor unditchedcondition.The onlyexceptionswerethe American
Crow, which was observedonly in plots with pools, and the Belted
Kingfisher.
Pitfall traps.--The invertebrateanimalscollectedfrom pitfall traps
on the 5 plots are listed in Table 3. Significantlymore invertebrates
were collectedper sampleon Plot lB than on any of the other plots
(ANOVA, F = 9.46, df = 4,60, P < .001; SNK, P < .05). On Plots 1A
and 1B (coastalmarsheswith pools),largenumbersof the minutesnail,
Vol.55,•o. 2
Effect
ofMosquito
Ditching
onBirds
IA
IA
lB
lB
2
$
IA
4
lB
2
3
4
42.7
42.5
IA
(•7.5)
45.5
51.0
(19.4)
(21.1)
44.8
(15.7)
46.4
(18.2)
51.7
(15.4)
44.3
48.6
44.3
lB
49.7
$
(16.8)(18.5) (16•.5)(14.9)
45.2
[ 169
49,1
48.4
50.5
57.0
4
(15.4)(16.7)(14•.2)(15.0) (11.7)
(20.0
45.6
(i5.5)
66.5
(13.4)
51.4
60.0
68.9
(16.9)
(11.8)
(11.6)
42,3
5;5,4
57.5
(17,9)
(17.4)
(16.1)
'51,6
(18.4)
40.3
50.9
5;5.1
50.;5
(18.8)
(20.8)
(2o.I)
;(32.6)
46.5
(22.9)
Pitfalls
Cores
IA
IA
lB
$
lB
55.8
(15.1)
55.7
66.8
(i3.o)
2_0. I
:>3.4
(14.5)
(14,8)
:50.5
Pools
F•ct•E 3. The meansimilarity(Bray-Curtis
coefficient),
afterlogarithmic
transformation
of invertebrateabundancedata, between and within plots for all samplestaken by
core,pitfall,andpoolsampling
methods.Numbersindicatemeanvaluesandstandard
deviationsare in parentheses.Asterisksmark the betweengroup meansthat were
significantly
different(P < .05) from the withingroupmeans.
Hydrobiaminuta,were collectedin the traps,while spiderswere common
on Plots3 and 4 (backwatermarshes).Amphipodsand miteswere abun-
dant in samplesfrom all marshtypes.
Similarityanalysisof the pitfall samples(Fig. 3) revealedthat the type
and number
of invertebrates
in Plot
lB were
different
from
those in
Plot 2. There were no other significantdifferencesbetweenplots.
Substrate
cores.--Invertebrates
collectedfrom the coresampleson the
plotsare listed in Table 4. The mean number of invertebratescollected
per core did not differ significantlyfrom plot to plot, regardlessof the
ditchedor unditchedcondition(ANOVA, P > .20). Mitesand springtails were the most commonlycollectedinvertebratesat all sites.Dipterans,both larvaland adult stages,and amphipodswerealsoabundant.
Comparisonof the mean similaritiesof core samplespaired within
and betweenplots(Fig. 3) revealedthat invertebratetaxa collectedon
170]
J. A. Clarke
etal.
j. Field
Ornithol.
Spring 1984
Vol.
55,•o.2
Effect
ofMosquito
Ditching
onBirds
v
v
v
[ 17 1
172]
J. A. Clarke
etal.
j. Field
Ornithol.
Spring 1984
T^BLE5. Mean abundance,i (SD), and percentageof invertebratetaxa in pool samples
from Plots 1A, 1B, and 3 in Rowley,Massachusetts,
summer 1982.
Plot
1A
lB
3
(n= 16)
(n= 13)
(n= 13)
•
Adult
%
•
%
•
%
insects:
Collembola
--
0.2
0.3
--
(0.6)
Coleoptera
0.1
0.1
--
--
0.1
(0.3)
0.1
0.2
(0.8)
0.1
0.1
--
--
0.1
--
0.1
(0.3)
0.9
0.8
7.5
(0.3)
Diptera
Hemiptera
0.3
0.2
(0.4)
1.9
(0.3)
Homoptera
Notonectidae
0.1
(0.3)
33.9
24.1
(115.3)
Hymenoptera
6.4
8.8
(4.9)
--
(0.9)
--
0.1
0.9
(0.3)
Larvae:
Diptera
Neuroptera/Coleoptera
3.4
(5.3)
2.4
4.5
(6.4)
0.1
0.1
--
6.2
(0.3)
3.2
(4.9)
30.2
0.2
1.9
(0.4)
Arachnids:
Acarina
Araneae
0.6
0.4
7.4
(1.1)
(6.2)
--
--
10.1
0.1
0.9
(0.3)
0.1
0.9
(0.3)
Crustaceans:
Amphipoda
Isopoda
Ostracoda
5.1
(17.3)
3.6
-0.1
0.1
(0.3)
Gastropods:
Hydrobiaminuta
96.7
(103.7)
0.1
(0.3)
0.1
--
0.1
(0.3)
0.1
--
2.9
4.0
(5.0)
68.8
2.9
50.7
(34.5)
69.5
0.2
0.3
0.9
(1.8)
Worms:
Polychaetes
--
(0.4)
27.4
(5.1)
--
8.5
Vol. 55, No. 2
[173
EffectofMosquito
DitchingonBirds
T^BI•E 5.
Continued.
Plot
Capitellids
1A
lB
3
(n -- 16)
(n -- lS)
(n -- 1s)
•
%
0.1
0.1
•
%
(0.3)
Opheliids
0.1
•
%
2.0
18.9
(•.s)
0.1
(o.s)
Ascidians
0.3
--
0.4
(0.6)
Mean
no. of inverte-
bratesper sample
140.5
(166.0)
100
73.0
(47.0)
100
10.6
100
(8.3)
Plot 2 were significantlydifferent from thosecollectedfrom Plots lB,
3, and 4. No other differenceswere significant.
Poolsampling.--The invertebratescollectedfrom water column and
bottomflocculusof the poolsystems
on Plots1A, 1B, and 3 are listedin
Table 5. Temperature, salinity,and pH of thesepoolsare presentedin
Table 6. The poolson Plot 3 were warmer (ANOVA, F = 59.27, df-2,39, P(.001;
SNK, P(.01)
and less saline (ANOVA, F=
130.14, df = 2,39, P ( .001; SNK, P (.01) than thoseon Plots 1A and
lB. The pH on plot 3 wasalsohigher than that on Plot lB (ANOVA,
F = 9.10, df = 2,39, P ( .001; SNK, P (.05). Fewer invertebrateswere
collected per sample from pools on plot 3 than from other pools
(ANOVA, F = 5.42, df = 2,39, P ( .01; SNK, P ( .01 for 1A to 3, P (
.001 for lB to 3). Hydrobiaminutawasby far the mostabundantinvertebratecollectedon Plots 1A and lB (approximately70% of sample),
while Plot 3 had 3 equallyabundanttaxa: dipteran larvae, ostracods,
and capitellidworms.Homopteransof the Family Notonectidae(backswimmers),Hydrobiaminuta,and dipteran larvae were commonon all
plots,but the first 2 animalspredominatedin the more salinepoolsof
Plots1A and lB. Dipteran larvaewere equallydistributedin all pools.
In addition to the invertebrateslisted, numerousgrassshrimp (Palaemonetes
pugio)were seenin Plot 1A and lB pools.The only vertebrate
identified,the commonmummichog(Fundulusheteroclitus),
waspresent
in large numbersin all pool systems.
Similarityanalysisof invertebratesin the pool samplesis summarized
in Fig. 3. The invertebratesfrom poolsin both Plot 1A and Plot lB
were significantlydifferent from thoseof Plot 3's pool system.
Predationbybirdsin saltmarshpools.--Invertebrateabundancesin the
netted and open sidesof the 2 test poolsbefore and after exclosureof
174]
J. A. Clarke
etal.
J.Field
Ornithol.
Spring 1984
T^B1.E6. Mean temperature,pH, and salinityof poolson Plots 1A, lB, and 3 asmonitored with YSI instruments,Rowley,Massachusetts,
summer 1982.
Temperature(C)
pH
Salinity(ppt)
• (SD)
• (SD)
• (SD)
Plot
1A (n -- 16)
lB (n -- 13)
3 (n = 13)
25.1 (3.0)
23.6 (1.3)
32.9 (2.3)**
6.8 (0.4)
6.5 (0.5)
7.5 (1.0)*
21.6 (2.1)
23.9 (2.0)
10.6 (2.6)**
* IndicatespH of Plot 3 wassignificantlydifferent from that of Plot lB (P (.05).
** Indicatestemperatureand salinityof Plot 3 were significantlydifferent from those
of Plots 1A and lB (P (.01).
birds are comparedin Table 7. Although almostall the invertebrate
populationsfluctuated,only the one changemarked with an asteriskis
statistically
significant(P (.05). Fluctuationsin the nettedhalf of a pool
generallycorrespondedwith fluctuationsin the open side.The relative
minnowabundancesin pools1A and 1B before and after the exclosure
period are shownin Fig. 4. While minnownumberswere stablein the
exclosedareas, numbers decreasedin the areas exposedto bird predation. The decreasewassignificant,however,only in the shallowpool
in Plot 1 A.
During the exclosureperiod,birdswereseenfrequentlyin andaround
the openhalvesof both testpools.SnowyEgretsand Greater and Lesser
yellowlegswere repeatedlyobservedforagingin pool 1A, while terns
and yellowlegswere seenfishingin the deeperpool on Plot lB. Semipalmated and Least sandpipers,and SemipalmatedPloverswere observedforagingand probing alongthe edgesof both pools1A and 1B.
DISCUSSION
Ditching for mosquitocontroldoesnot appearto havehad a marked
effecton the invertebratefauna of the Rowleystudysites.On all plots,
whether coastalor backwater,similar numbersand typesof invertebrates were collectedfrom core samplesand pitfall traps. The invertebratescollectedfrom pool samples,however,were different on the
coastaland backwatermarshes.The warmer,lesssaline,and higherpH
environmentin the poolsof Plot 3 predictablywouldsupportdifferent
speciesthan would the poolsof coastalplots 1A and lB. Both the
ditched
and unditched
backwater
marsh sites had similar
numbers
and
ordersof invertebrates.The invertebratepopulationsof the coastalplots
with activeand neglectedditches,althoughdifferent from thoseon the
unditchedplot with pools,seemedroughly comparableto each other
in termsof the quantityof potentialprey for birds. Epifaunalinvertebratestaken in pitfall trapswere more abundanton unditchedPlot 1B
than on ditched Plot 2, but, except for the rarer taxa, no order was
found exclusivelyon a ditchedor unditchedsite.With the exceptionof
backswimmers
(Family Notonectidae),invertebratesfound in poolson
Vol.55,No.9
Effect
ofMosquito
Ditching
onBirds
[ 175
TABLE7. Mean abundance,X (SD), of invertebrate taxa in samplesfrom experimental
(netted)andcontrol(open)saltmarshpoolsbefore(1) andafter (2) an 18-daybird exclosure
period, Rowley, Massachusetts,
summer 1982.
Pool
1A
Pool
Netted
Taxa
1
Open
2
1
1B
Netted
2
1
Open
2
1
2
Insects:
Hemiptera
Notonectidae
2.0
4.0
0.7
1.0
5.8
4.0
9.7
1.7
(2.6)
(6.9)
(1.2)
(1.0)
(4.3)
(3.6)
(5.7)
(1.5)
--
--
Diptera
Adults
Larvae
0.3
0.3
0.3
0.3
(0.6)
(0.6)
(0.6)
(0.6)
13.3
(4.2)
--
1.7
(2.1)
11.3
1.3
4.7
1.5
3.7
10.0
6.3
(11.7)
(0.6)
(2.5)
(1.3)
(4.7)
(12.2)
(10.1)
Arachnids:
0.3
2.7
0.3
(0.6)
(4.6)
(0.6)
1.0
(1.7)
0.3
(0.6)
--
Isopoda
--
--
--
0.3
(0.6)
Ostracoda
--
--
0.7
0.3*
21.3'
(1.2)
(0.5)
(13.3)
Acarina
--
5.3
--
(4.6)
10.7
1.0
(6.1)
(1.7)
Crustaceans:
Amphipoda
0.7
(1.2)
Gastropods:
Hydrobiaminuta
37.7
79.0
(7.2) (136.8)
Total
Total
increases
decreases
Unchanged
0.7
0.3
(0.6)
10.7
31.3
(9.3) (30.5)
(1.2)
(0.5)
--
6.7
(6.4)
51.5
57.0
46.3
53.0
(31.0)
(25.6)
(34.6)
(41.0)
4+
2-
6+
1-
4+
3-
4+
3-
2
1
1
1
* Indicatesthesetwo valuesare significantlydifferent, P < .05.
Plots 1A, lB, and 3 were found under moist thatch or in wet sunken
areasof the ditchedmarshes.The minnowsabundantin poolswere also
abundantin ditchesat the siteswithout pools.
We concludethat while grid-ditchingfor mosquitocontrolmayhave
someeffecton marshinvertebratepopulations,it doesnot significantly
reduceinvertebrateabundancesin Massachusetts.
The effectof ditching
on invertebrates
hasbeenpoorlyresearched
in the New Englandarea,
so we have no local comparisonfor this conclusion.Invertebrate research on ditched and unditched
sites in mid-Atlantic
and southern states
hasproducedcontradictory
results,althoughmostrecentlypublished
reports(Rockel1969,ShislerandJobbins1975,Lesseret al. 1976,Chick
176]
J. A. Clarke
etal.
j. Field
Ornithol.
Spring 1984
140
II
120
2
0•100
2
,,i,-
õ 80
open
o
•
60
netted open**
netted
6 4O
2
• 20
0
t
Plot
1A
(< .25m)
Plot
lB
(>.5 m)
FIGURE
4. Changesin minnowabundancein shallow(1A) and deep(lB) saltmarshpools
at controlsitesclosedto preventbird predationfor 18 daysand at experimentalsites
open to predation. Bars represent 1 standarddeviationfrom the mean value of 3
counts;thoselabelled 1 representinitial counts,those labelled 2, final counts.
1979) agreethat ditchingdoesnot havethe markedeffecton invertebrate populationsreported by earlier researchers(Bourn and Cottam
1950). It is important to note that the distinctionbetweenditchedand
unditchedplots in our studydependedonly on the current condition
of eachstudysite.All areashadat one time beenditched,and thismay
have affectedaspectsof the plots' current invertebrate fauna.
Ditching of the saltmarshfor mosquitocontrol appearsto affectbird
useof the Rowleysites.The unditchedplotsof both coastaland backwatermarshtypeshad a greatervarietyand and numberof birdsthan
did the correspondingditchedsites.Shorebirds,herons,terns,and swallowsexhibiteda preferencefor foragingon the unditchedplots with
pools.
Analysisalsoindicatedthat the composition
of bird species
seenon
Vo•.
55,•o.2
Effect
ofMosquito
Ditching
onBirds
[ 177
the Rowleyplotswasdifferentfor backwaterandcoastalmarshes.Many
songbirdsnestingin the large woodedarea surroundingthe backwater
sitesusedthe nearbymarshfor foraging.The many standsof marsh
elder, sedge, and cattails on the backwater marshesalso provided a
habitat for Red-wingedBlackbirdsand Sharp-tailedSparrows.The differencebetweencoastaland backwatermarshtypes,therefore,is more
likely attributableto the greatervarietyof passerinespeciesfrequenting
the backwaterplots than to differencesin the presenceof shorebirds,
herons, terns, and swallows.
The pool exclosureexperimentssuggestedheavyfish predationby
yellowlegs
andSnowyEgrets.GreatandLittle Blueherons,GlossyIbises,
and terns were also seenforaging in poolsduring the summer.The
smallersandpipers
and ploversforagedheavilyaroundpool edges.The
exclosurestudiesseemedto indicatethat bird predation had little effect
on marshpool invertebratepopulationsduring an 18-dayperiod, but
the edgesof the poolswerenot well sampled,and it ispossiblethat these
smaller birds had an effect there
that we did not detect.
In addition
to
thesespecies,kingbirdsand swallowsforagedfor insectsin the air above
the poolsystems.
All of thesebirdsfocusedtheir foragingin andaround
salt marsh pools.
Decreases
in invertebratepopulationsdid not consistently
correspond
to decreasesin bird use. While ditching may have someeffect on the
invertebratefaunaof a marsh,it doesnot stronglylimit the abundance
of potentialpreyfor birdsthat mightforagethere.Therefore, the effect
of ditchingon bird useappearsto be controlledby factorsother than
prey abundance.
Ditching may not significantlydecreaseinvertebratepopulationsof
the marsh,but it doesdrain pools.Pool drainagewasa primary goal of
the originalgrid-ditchingsystems
(Smith 1975), and the effecthasbeen
documentedin New Englandmarshes(Reinert et al. 1982). In the present study,shorebirds,wadingspecies,terns, swallows,and crowswere
stronglyattractedto the saltmarshpools.Given this strongattraction,
weconcludethat ditchingadverselyaffectsavianpopulations
by draining
their foragingareas.
We speculatethat the effectof ditchingon birdsdoesnot resultfrom
a reductionin the abundanceof potentialprey, but rather from the
reductionin accessibility
of prey asa resultof pool drainage.Although
invertebratesand fishmay still be plentiful on the marsh,withoutpools
and pool edges,they may be more difficult to spot and capture. The
foragingareaavailablein ditchesislimited by their narrowwidth, steep
sides,and constantlyfluctuatingwater levels(Reinert et al. 1982). In
addition, poolsoften supportalgal matswhich, when castto the marsh
surfaceon flood tides,smotherand kill smallpatchesof marshvegetation, opening up new spotsfor foraging. Without pools,fewer surroundingbare spacesare created.These matsalsocarry trappedinvertebratesout of the water makingthem availableto smallsandpipersand
plovers.
Our hypothesisconcerninglimited accessibility
of prey is consistent
178]
J. A. Clarke
etal.
j. Field
Ornithol.
Spring 1984
with the resultsof Reinertet al. (1982) in southernNew England.That
studyindicatedthat birdswere more abundanton marsheswith pools,
but the reasonsfor this preferencewere unclear.No other studieshave
beenpublishedthat addressthe effectof ditchingon avianpopulations
in New England.Researchelsewhere(Urner 1935, Ferrigno 1970, Burger and Shisler 1978, Shisler and Jobbins 1975) is limited, and has
generallyconcentratedon bird speciesthat breed on the marsh,not on
migratoryor transientflocks.
The marsh managementcurrently practicedby Massachusetts
mosquitocontrolassociations
consists
almostentirelyof maintainingthe old
grid-ditchingsystems
that drain pools(McGlathery1982).We are unable
to directlyjudge the effectiveness
of this practicein limiting area mosquitobreeding,becausemosquitopopulationswere not censused
on the
Rowleyplots.Openmarshwatermanagement(OMWM), however,which
treatsonlyselectedareasof the marshandoftenretainsor createspools,
is the techniquestronglyrecommendedin the mid-Atlanticstatesfor
both effectivemosquitocontrol and minimum impact on salt marsh
fauna(FerrignoandJobbins1968, Ferrigno 1970, Ferrignoet al. 1975,
McGlathery 1982). Total implementationof modern OMWM systems
is presentlycost-prohibitive
in Massachusetts,
and may not be entirely
applicableto northernmarshesbecauseof the tidal regimen,problems
from winter icing,anddifferences
in species
diversity(N. Dobson,pers.
comm.).It is probable,however,that existinggrid ditchescan be modified to allow somepool formationon saltmarsheswithoutincreasing
mosquitolevels(N. Dobsonand W. Montgomery,pers.comm.),thereby
enhancingthe value of the marsh to shorebirds,herons,and terns.
SUMMARY
Observationsof bird populationson salt marsheswith and without
well-maintainedgrid ditchesshowedthat this form of mosquitocontrol
decreaseshabitat use by shorebirds,herons,ibises,terns, and aerial
insectivores.
Thesebirdsnormallyforagein or aroundsaltmarshpools
that are absenton well-drained,ditchedmarshes.Foragingby passerines
and other speciesseemedunaffectedby the ditched or unditchedcondition of the marshes.A concurrentstudyof invertebratepopulations
suggeststhat ditching doesnot causea reduction in the abundanceof
invertebrateprey. It mayreducethe accessibility
of that prey to certain
avian predators,however,by eliminatingthe poolswhere thesebirds
forage.If shorebird,heron, and tern populationsusingthe salt marsh
are to be encouraged,marshpoolsshouldbe maintained.Experimental
modificationof existing Massachusetts
ditching systemsis recommended.
ACKNOWLEDGMENTS
Funding for this studywasprovided under a contract to Manomet
Bird Observatoryfrom the Town of Rowley,Massachusetts,
via a Community Assistance
grant from the Massachusetts
CoastalZone Manage-
Vol.
55,1•o.
2
Effect
ofMosquito
Ditching
onBirds
[ 179
mentProgram.Cooperativeservices
wereprovidedby the EssexCounty
MosquitoControl Commission
and the Massachusetts
AudubonSociety,
Resourcesfor CapeAnn Program. Technicaladviceand assistance
were
providedby a number of individuals.In particular, we wish to thank
Norman DobsonandWalter Montgomeryof the EssexCountyMosquito
Control Commission,Dr. ErnestRuber, and Dr. Gary Benson.We are
alsogratefulto the citizensof Rowley,especially
John Ewelland Donald
and Ruth Alexanderfor their encouragementof our work.
LITERATURE
CITED
AMERICAN ORNITHOLOGISTS' UNION. 1983. Check-list of North American birds, 6th
Edition.
BELL,S.S., M. C. WATZIIN,ANDB.C. COULL.1978. Biogenicstructureand its effecton
the spacialheterogeneityof the merofaunain a saltmarsh.J. Exp. Mar. Biol. 35:99107.
BOURN,W. S., ANDC. COTTAM. 1950. Some biologicaleffectsof ditching tidewater
marshes.U.S. Dep. Int. Fishand Wildl. Serv.Res.Rep. 19.
BRADBURY,
H. M. 1938. Mosquitocontrol operationson the tide marshesin Massachusetts
and their effecton shorebirdsand waterfowl.J. Wildl. Manage.2:49-52.
BRA¾,J.R., ANDJ.T. Ct•RTIS.1957. An ordinationof the uplandforestcommunitiesof
southernWisconsin.Ecol. Monogr. 27:325-349.
BURGER,
J., ANDJ.K. SHISLER.1978. The immediateeffectsof ditchinga salt marshon
nestingHerring Gulls,Larusargentatus.
Biol. Conserv.15:85-105.
CHICK,B. E. 1979. A report detailingthe Rye saltmarshmanagement
pilot studyfor
mosquitopreventionand marshrejuvenation.Rye MosquitoControlDistrictComm.
Rep.
CORKRAN,
W. S. 1938. New developments
in mosquitocontrol in Delaware.Proc. N.J.
MosquitoExterm. Assoc.25:130-137.
COULL,B.C., ANDS.S. BELL.1979. Perspectives
of marinemerofaunalecology.Pp. 189216, in Ecologicalprocesses
in coastaland marinesystems,
Livingston,R.J., ed., Mar.
Sci. Vol. 10. Plenum Press, N.Y.
FERR•GNO,
F. 1970. Preliminaryeffectsof openmarshwater managementon the vegetation and organismsof the salt marsh. Proc. N.J. MosquitoExterm. Assoc.57:
79-94.
, ANDD. M. JoB•NS. 1968. Open marshwater management.Proc. N.J. Mosquito
Exterm.
Assoc. 55:104-115.
, P. SLAVIN,and D. M.Jo•INS. 1975. Saltmarshwater managementfor mosquito
control. Proc. N.J. MosquitoExterm. Assoc.62:30-38.
FIELD,J.G., ANDG. MCFARLANE.
1968. A quantitative"similarity"analysis
of rockyshore
samplesin FalseBay, SouthAfrica. Zool. Afr. 3:119-137.
HEADLEE,
T.J. 1939. Relation of mosquitocontrol to wildlife. Proc. N.J. Mosquito
Exterm.
Assoc. 26:5-12.
JACCARD,
P. 1902. L'ois de distributionflorale dansla zone alpine.Bull. Soc.Vaudoise
Sci. Nat.
38:6-130.
KUENZLER,
E. J., ANDH. L. MARSHALL.
1973. Effectsof mosquitocontrolditchingon
estuarineecosystems.
N.C. Water Resources
Res.Inst. Rep. 81.
LESSER,
C. R., F.J. MVR•'H¾,
ANDR. W. LAKE.1976. Someeffectsof gridsystemmosquito
controlditchingon saltmarshbiota in Delaware.Mosq.News36:69-77.
MCGLAT•SRY,K.J. 1982. A literary review of the effectsof mosquitocontrol ditching
on the saltmarshecosystem,
and itsimplicationon the managementpoliciesfor TNC
preserves.Nature Conservancy
EasternRegionalOfficeRep.
REINERT, S. E., F. C. GOLET, AND W. R. DERAGON. 1982. Avian use of ditched and
unditchedsaltmarshesin southeastern
New England:a preliminaryreport. URI Res.
Rep.
180]
J. A. Clarke
etal.
J.Field
Ornithol.
Spring 1984
ROCKEL,
E.G. 1969. Marsh physiography:
influenceon distributionof intertidal organisms.Proc. N.J. MosquitoExterm. Assoc.56:102-116.
SHISLER,
J. m., ANDD. M. JOBBINS.
1975. Aspectsof biologicalproductivityin mosquito
ditched salt marshes.Proc. N.J. Mosquito Exterm. Assoc.62:48-49.
Sm•rI4,L. W., JR. 1975. Training manual for the Northeastern Mosquito Control Association.
TEAL,J. M. 1962. Energyflow in a saltmarshecosystem
of Georgia.Ecology43:614624.
TR^VIS,B. V., G. H. BRADLEY,
^ND W. C. MCDUFFIE.1954. The effect of ditchingon
saltmarshvegetationin Florida. Proc. N.J. MosquitoExterm. Assoc.41:235-244.
URNER,C. A. 1935. Relationof mosquitocontrolin New Jerseyto bird life of the salt
marshes.Proc. N.J. MosquitoExterm. Assoc.22:130-136.
W1ESER,
W., ^•4DJ. K^•qW•SlqER.
1961. Ecologicaland physiologicalstudieson marine
nematodes from a small salt marsh near Woods Hole, Massachusetts. Limnol. Ocean-
ogr. 6:262-270.
Z^R,J. H. 1974. Biostatistical
analysis.Prentice-Hall,EnglewoodCliffs,N.J.
MahometBird Observatory,
Box 936, Mahomet,Massachusetts
02345, and
Boston
University,
Department
ofBiology,
Boston,
Massachusetts,
02215 (JAC);
ManometBird Observatory
(BAH); Resources
for CapeAnn, Massachusetts
AudubonSociety,159 34ain Street,Gloucester,
Massachusetts
01930 (TH)
(M.A.S. Contribution83-7);Biology
Department,
Boston
University
(FEW).
Received9 Aug. 1983; accepted20 Jan. 1984.