' AN ECOLOGICAL; ANALYSIS OF - . ’{  ‘ . 7 '1 9373::
’ FOREST EDGESUITABILETY V =
*- . ' FDR AVIAN POPULATlOng. ; (fl; - '55:: ~ l :

' Dissertation for the‘Degre‘e OfZPh'. D.‘ N- r
7fVTMICHIGAN'STATE‘UNIVERSITY :   ’ H
} "JOHN EDWARD GATES I
' 1976; - ' I

 

   
 

  

[JBRARY
Ngchigan State
University

  

“\

   

 

.5

This is to certify that the
thesis entitled /
AN ECOLOGICAL ANALYSIS OF FOREST EDGE
SUITABILITY FOR AVIAN POPULATIONS

presented by

“John'Edward Gates

has been accepted towards fulfillment
of the requirements for

Ph D. Fisheries & Wildlife

degree in

  

Major professor

  

. Date September 21, 1976

0-7639

 

 

IlHl”WWII!!!IllllfllllfllllIll!!!lllflllllllll

 

 

 

 

 

ABSTRACT

AN ECOLOGICAL ANALYSIS OF FOREST EDGE
SUITABILITY FOR AVIAN POPULATIONS

By
John Edward Gates

Observations of twenty-one species of open-nesting passerines
breeding in contiguous field and forest habitats at Rose Lake Wildlife
Research Area, Michigan,were made during l974 and l975. Data were col-
lected on nest dispersion and fledging success in relation to the field-
forest edge. Losses of eggs or nestlings were attributed to predation,
inclement weather, Brown-headed Cowbird parasitism, nest desertion,
hatching failure, or adult death.

Nests were not evenly distributed on the area, but had a definite
pattern. Over one-half of the nests were found within l5 m of the habitat
discontinuity. Seventy-five percent of the nests belonged to birds
characteristic of mixed breeding habitats, i.e., birds requiring open
areas with dense cover for nesting and feeding and elevated singing and
observation perches. These birds also accounted for the increase in
species nesting near edges. Nest dispersion seemed to be influenced by
the openness of the forest and the amount of woody vegetation in the
field.

Based on Kendall rank and partial rank correlation tests, numbers
of nests were found to be negatively correlated with distance from a

habitat discontinuity. Correlation between fledging success and distance

John Edward Gates
from the edge was positive and highly significant. 0f the several mor-
tality factors investigated, predation was found to be the major one
influencing the observed trend in fledging success. Although losses due
to cowbird parasitism were also significantly correlated with fledging
success, number of nests, and distance from the edge in the same manner
as predation loss, its influence was not as important. The increased
predation was attributed primarily to a functional response to higher
numbers of nests and a greater activity of potential nest predators in
the vicinity of the habitat discontinuities.

ReSults indicated that habitat suitability decreased with increas-
ing numbers of nests toward the narrow field-forest edges. Although
such abrupt habitat discontinuities did attract a variety and abundance
of birds characteristic of habitats with mixed life-form, they seemed to
function as ecological traps for birds nesting in their vicinity. Iron-
ically, the cowbird was also a victim of the increased predation rate.
As these man-made forest edges are of recent origin, they are perhaps
unrepresentative of the ecological niche in which these species evolved,
and thus they may be poorly adapted to cope with the increased nest

predation.

AN ECOLOGICAL ANALYSIS OF FOREST EDGE
SUITABILITY FOR AVIAN POPULATIONS

By

John Edward Gates

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Fisheries and Wildlife

1976

ACKNOWLEDGMENTS

I acknowledge the support and encouragement of my graduate committee:
Drs. Leslie W. Gysel, Rollin H. Baker, Donald L. Beaver, and Stephen N.
Stephenson. I especially thank my committee chairman, Dr. Leslie W.
Gysel, whose generous advice, constructive criticism, and interest in the
project was most welcomed. I am also indebted to Alan Tipton, Bruce Batt,
Wayne Schmidt, Dennis Fijalkowski, Richard Kasul, Phillip Davis, and
Gary Friday for valuable discussion, encouragement, and assistance during
different phases of the study. This work could not have been completed
without the cooperation of the staff of the Rose Lake Wildlife Research
Center. I especially thank Keith Heezen, Biologist in Charge, and Robert
Ankney, Research Technician, for their initial advice on selection of
study sites and continued interest in the project. Thanks go to Pat
Friday for quick, careful typing of the manuscript. A very warm-hearted
thanks goes to my wife, Donna, for her moral support and assistance dur-
ing the field work, and to my parents for their encouragement during
these years. Financial support was provided from McIntire-Stennis funds

and a Sigma Xi Grant-in-Aid of Research.

ii

TABLE OF CONTENTS

 

 

 

 

 

 

 

Page

LIST OF TABLES ......................... iv
LIST OF FIGURES ......................... v
INTRODUCTION .......................... 1
MATERIALS AND METHODS ...................... 6
Field Studies ....................... 6
Analytical Procedures ................... 9

STUDY SITES ........................... ll
RESULTS ............................. 22
Nest Distribution Patterns ................ 22
Pledging Success and Mortality Patterns .......... 27
Differential Mortality Patterns .............. 34
Seasonal Effects ..................... 37
Predator Activity ..................... 37
DISCUSSION ...................... , ..... 40
APPENDIX ............................ 47
LITERATURE CITED ........................ 49

iii

Table

LIST OF TABLES

Page
Density, basal area, and frequency of trees (> 7. 62 cm dbh)
by species in the forest habitat on each study site based
on five 0. 1- -acre (0. 04- ha) circular plots per site ..... l2

Density and basal area of trees (> 7. 62 cm dbh) by size
class in the forest habitat on each study site based on
five 0. 1- -acre (0. 04- ha) circular plots per site ....... l3

Means, standard deviations, and sample size ( ) of variables

used to describe habitats on each study site. Chi- -squared

(X2 ) and Kruskal -Wallis (H) tests were used to determine

whether or not observed differences were significant. . . . l4

Characteristic life-form of nesting habitat and numbers of

nests of open-nesting passerines located on the study area.
Numbers of nests in which eggs were determined to be pre-

sent are in parentheses ........... A ........ 23

Matrix of Kendall rank correlation coefficients, 1 (tau),
comparing the interrelationships between distance from a

habitat discontinuity (edge), nest distribution index (NDI),
fledging success, and the different mortality factors affect-

ing fledging success .................... 31

Kendall partial rank correlation coefficients, Tx .2, of
the relation between the §_and Y variables when t e effect
of variation of a third variable; 7, is eliminated ..... 33

Differential effects of mortality factors on eggs and
nestlings at Rose Lake ................... 36

Comparison of losses due to different sources of morta-

lity by month of nest initiation at Rose Lake. Sample
sizes are in parentheses .................. 38

iv

Figure

5.

LIST OF FIGURES

Page
A model representing the important components determin-
ing whether or not a bird will be successful in a certain
patch of vegetation ..................... 3

Representative vegetation profiles of the interface be-
tween field and forest habitats on the study sites

derived from the maximum height of overstory (3_7.62 cm
dbh), dominant understory (32.54 cm, <7.62 cm dbh), and
dominant herbaceous and woody ground-layer (<2.54 cm dbh)
vegetation. The zone at which dominant overstory trees
(30.48-38.l0 cm dbh) first occurred in the stand is re-
presented by cross-hatched areas, which designate i 1 SD
of the mean. Sample sizes are 10 (site 1), 30 (site 2),
and l5 (site 3). Differences between sites were not signi-
ficant (fl_= 3.9l, p_>0.05) .................. l6

Cumulative nest distribution patterns of open-nesting

passerines on the study sites according to characteristic
life—form of nesting habitat and number of species (numerals

above histograms) ...................... 25

Graphical presentation of the relationship of different

nest variables with distance from the habitat discontinu-

ity. Distances covered by each distance category are as

follows: (I) 0.0-0.82 m;(2) 0.82-2.19 m; (3) 2.19-4.34 m;

(4) 4.34-6.86 m; (5) 6.86-l0.06 m; (6) l0.06-l4.18 m;

(7) l4.18-26.74 m; (8) 26.74-46.24 m; (9) 46.24-65.58 m;

and (l0) 65.58-123.00 m. Sample sizes are in parentheses. . 29

A summary of the biologically important as well as statis-
tically significant pathways by which the habitat discon-

tinuity (edge) affected nest distribution, fleding success,

and losses due to different mortality factors in open-

nesting passerines. The width of connecting lines desig-

nate the significance of the relationship. The direction

and sign of the arrows indicate probable cause-and-effect
relationships and whether or not it is direct or inverse . . 4l

INTRODUCTION

The ecological relationships which exist at the natural junction
of different plant communities form the basis of the edge effect concept.
Simply stated, it is a tendency for increased variety and density of
organisms at the junction of different plant communities (Odum l97l).
This phenomenon is due in part to the presence of species characteristic
of each of the overlapping communities plus species inhabiting only the
region of overlap. These edge species may find habitats at the community
junction not present in either community alone, or require two or more
adjacent communities differing greatly in structure.

The idea of bringing different vegetation types into close associa-
tion through habitat manipulation has also become an important principle
to the wildlife biologist. It is called interspersion, and the border-
line, considered so favorable to game, is referred to as edge (Leopold
l933, Allen l962, Dasmann l964). It is generally assumed that the close
association of diverse plant communities results in optimum conditions
for reproduction and survival of certain wildlife species.

Although many terrestrial vertebrates of economic importance are
considered to be edge species or utilize edges to a large extent, the
concept of edge effect is especially applicable to avian populations.
Studies have shown that greater numbers of avian species exist in areas
of mixed habitat, and consequently more edge, than in generally uniform
vegetation (Johnston l947, Johnston and Odum l956). Other investigators

have demonstrated a greater density of birds in mixed habitats, or in

2
the margin, as compared with the interior of large tracts of uniform
vegetation (Lay 1938, Good and Dambach 1943, Edeburn 1947, Johnston 1970).
Nest densities have also been shown to be affected by proximity to habi-
tat discontinuities. Beecher (1942) attempted to locate all bird nests
on an area containing several marsh and upland communities. He found
the greatest density of nests in sample plots containing two or more
community types, and consequently more edge.

Because of the importance of the edge effect concept to the field
of wildlife management and the increasing need and demand for the manage-
ment of nongame species, this study was initiated to evaluate the effect
of habitat discontinuities on avian populations. The main objectives
of the research were to investigate the spatial distribution patterns of
bird nests in relation to habitat discontinuities, and to determine some
of the mechanisms operating at such boundaries which might influence
their suitability as breeding habitat for avian populations.

If we assume that the conceptual model presented in Fig. 1 is cor-
rect, then one of the most important components determining the suitabil-
ity of a habitat is nesting success rather than size of territory,
population density, or any other factor considered alone (see Dow 1969,
Fretwell 1968, 1969, Nettleship 1972), i.e., those individuals selecting
a better quality patch or nest site will have increased nesting success
over those selecting a poorer quality one. Nettleship (1972) has indi-
cated that, if habitat selection is an adaptive activity, then the only
measure of habitat suitability during the breeding season is nesting
success and the degree of productivity attained in it. Other researchers
have presented data supporting the view of increased nesting success in

optimum habitats (Krebs 1971, Heppleston 1972, Robertson 1972, Holm 1973).

Fig. l. A model representing the important components determining
whether or not a bird will be successful in a certain patch
of vegetation.

INDIVIDUAL
BIRD SPECIES

l

PROXIMATE FACTORS
(Different features of
the habitat)

l

HABITAT SELECTION

1

PATCH, OR NEST SITE
QUALITY

PREDATION FOOD INCLEMENT
WEATHER, ETC.

\1/

NESTING SUCCESS
(Nests, eggs, or
fledglings)

SURVIVAL

,.
I
I
I
I
I
I
I
I
I

_-—--_——__-_—_—_———4

NATURAL
SELECTION

t----_-—-_-_——---’

5
Although previous studies have primarily been concerned with overall
success in relatively uniform habitats rather than with the effect of
habitat discontinuities on nest dispersion and nesting success within
such habitats, the same logic should be applicable to the latter situa-
tion. Here I appraise the suitability of narrow man-made forest edges
for open-nesting passerines breeding in contiguous grassy field and oak-

hickory forest habitats in central Michigan.

MATERIALS AND METHODS

Field Studies

 

Study sites were surveyed and gridded with wire-stake flags set at
50 m intervals using the cardinal points as a reference. The wire—stake
flags were coded with black letters and numerals. The grids were later
used to map bird nests for determining spatial distribution patterns in
relation to the junction between field and forest, and to locate sample
points for analyzing the composition and structure of vegetation on each
site. A somewhat detailed vegetation analysis was done in order to
adequately describe the field and forest habitats and the interface used
by the breeding avifauna on the study sites.

Observations of avian populations on each study site were begun in
March 1974 and continued through September 1975. An intensive search for
all nests was made by systematically walking through the gridded study
sites. To determine if nests had been missed, a 30 m rope was dragged
through the field habitats (Wiens 1969), i.e., incubating birds should
be flushed from the nest. A combination pole and mirror device was used
to check the contents of nests in the canopy (Parker 1972). Open canopy
nests above 12 m were ignored for the most part due to the difficulty of
examining these nests on a daily basis. Nests in this category made up
a very small percentage of nests located on the study area. Throughout
most of the breeding season, daily trips to each site were made to
check nest contents and to locate new nests. Several precautionary

measures were undertaken to reduce the likelihood of predation due to

6

7
the frequency of trips to check nest contents. When checking the con-
tents of nests located below eye level, the observer was careful to
stand as far from the nest as possible. Care was also taken not to
create paths leading to nest sites, or to approach a nest when possible

predators or Brown-headed Cowbirds (Molothrus ater) were observed nearby.

 

All nests were marked in the field by placing a small strip of plastic
flagging on vegetation 3 m north of the nest. Nest locations were plot-
ted on scale maps of the study sites. Distances to habitat discontinu-
ities were determined by range finders or measuring tapes for nearby
nests.

Since open-nesting birds in the same locality generally have the
same nest success regardless of their own species density (Nice 1957,
Nolan 1963, Ricklefs 1970, Fretwell 1972), I decided to combine all
open nests and to consider differences in fledging success and mortality
patterns at the community level. Only those nests in which at least one
egg had been laid were used in these analyses. This restriction was
imposed so that all nests would initially have the same potential of
being successful.

Measurements of environmental variables associated with avian
breeding habitats on the study sites were made in July and August 1975.
The methods used to collect data in each habitat were generally similar
to that of James and Shugart (1970) using a modification of the range
finder circle technique. Sample points were located randomly within
each habitat type. Due to the possible influence of adjacent habitats,
I decided to avoid all points lying within 30 m of a habitat discontinu-
ity. Each sample point served as the center of a 0.1-acre (0.04-ha)

circular plot, which was used for recording data on the composition and

8
structure of the vegetation associated with each habitat. Adequacy of
the sample was determined by the method outlined by James and Shugart
(1970).

Within the 0.1-acre circle at each sample point, the species and
diameter at breast height (dbh) of trees (3_7.62 cm) were recorded on
data sheets. Canopy cover was determined from point samples (James and
Shugart 1970, Shugart and Patten 1972). The age of the dominant tree in
each circular plot was estimated by counting growth rings on an increment
core. Total height of vegetation directly over the sample plot was meas-
ured directly, or obtained from a Haga altimeter.

The number of species of understory plants (< 7.62 cm dbh, 3_1.0 m
in height) within 4 m of the sample point was determined in each plot.
Density of understory vegetation was estimated by averaging two 0.01-acre
(0.004—ha) arm length rectangular plots centered on the sample point
(Penfound and Rice 1957). Dominant understory vegetation cover (3_2.54
cm, < 7.62 cm dbh) was estimated from point samples (James and Shugart
1970, Shugart and Patten 1972).

Within 1 m of the sample point, the number of species of ground-
level vegetation (< 1.0 m in height) was determined by counting. Plot-
1ess point samples were made with an ocular tube to estimate percent
cover (Winkworth and Goodall 1962).

A vegetation profile of the interface between field and forest
habitats was estimated at each site by use of a measuring rod and a range
finder mounted on a tripod. Transects were located at a representative
portion of the interface and directed perpendicular to the habitat dis-
continuity at each site. They extended 20 m into each habitat type.

Height of overstory, dominant understory, and dominant herbaceous and

9
woody ground-layer (< 2.54 cm dbh) vegetation were recorded at each sam-

ple point. These data were used to construct vegetation profiles.

Analytical Procedures

 

Since I demonstrated that the assumptions of the parametric statis-
tical model were not met in the data under analysis, especially that of
normality (i.e., too many zeros), nonparametric techniques were used to
determine the significance of observed differences and trends in the
data. These distribution-free techniques result in conclusions which
require fewer qualifications, and thus have greater generality than
corresponding parametric tests. Siegel (1956) adequately discusses the
advantages and disadvantages of both techniques of inference.

Chi-squared values (K?) shown with tables have been derived from
contingency tables testing the significance of the difference between
the proportions indicated, unless otherwise stated. For those variables
which met the assumptions of ordinal measurement and underlying continu-
ity, the Kruskal-Wallis test statistic (H) (Siegel 1956, 184-193) was
calculated to determine whether differences among means in the k_popu-
lations were significant.

Kendall rank correlation coefficients (Siegel 1956, 213—223) were
used to determine the effect of distance from the habitat discontinuity
on nest dispersion, fledging success, and the different mortality factors
affecting fledging success and their interrelationships. When two vari-
ables are correlated, there is always the possibility that the correlation
may result from both being associated with a third variable. The observed
correlation may not reflect any direct association between the two vari-

ables. Partial correlation provides a statistical method of attacking

10
this problem. The association between K_and 1_is found with the third
variable ;_kept constant. The correlation between these adjusted vari-
ables is the partial correlation. Kendall partial rank correlation
coefficients (Siegel 1956, 223-229) were calculated for certain observed
correlations to understand and clarify the relationships between the
variables.

Because these nonparametric correlation techniques require ordinal
measurement and assume that the variables have underlying continuity,
data were grouped in the following manner to meet these assumptions.
Nests were first ranked according to their distance in meters from the
edge, which is here defined as the point at which woody vegetation is
first encountered in proceeding from field to forest habitats. Based
on a cumulative frequency distribution of distances, nests were next
partitioned into ten distance categories, each containing approximately
ten percent of the total number of nests. Each category, therefore, had
approximately an equal number of nests. The percent fledged and lost to
different mortality factors were then determined relative to the total
number of eggs laid. A nest distribution index (NDI) was finally calcu-
lated to correct for unequal distance intervals in each category:

5,71,
NDI = X 100

 

M1:

(lg/1,.)
l

c-Je
II

where 54 is the number of nests in the ith distance category, 14 is the
number of meters in the ith distance category, and g_is the number of
distance categories. Calculation of N01 allowed a comparison to be made of

the interrelationships between numbers of nests and the other variables.

STUDY SITES

The study sites were located on the Rose Lake Wildlife Research
Area in Clinton and Shiawassee counties, Michigan. The Research Area
is 19.3 km northeast of Lansing, just north of highway M-78 (I-69). It
consists of 1350.3 ha of moderately undulating farmland, abandoned fields
and pastures, oak and swamp woods, and marsh. Rose Lake Wildlife Re-
search Center staffed by several wildlife biologists and technicians is
located on the area. Three sites were selected from among a larger
series to represent the abrupt habitat discontinuity characteristic of
edges created by many agricultural, forestry, and wildlife management
programs, and because of the control I would have over all management
operations on the sites. General features of the sites and their vegeta-
tion are given in Tables 1, 2, and 3 and Fig. 2.

Site l.--This site consisted of a forest stand (woodlot #1, 4.28
ha) and an abandoned field (4.46 ha) located immediately north (woodlot
numbers are those designated by the biologists at the Research Center).

The forest habitat was dominated by Quercus velutina with Prunus serotina

 

 

and Carya glabra present as subdominants. Prunus serotina was important

 

 

due to an observed higher density of individuals in the smaller size
classes, especially those trees between 3-6 inches (7.62-15.24 cm) dbh.

The field habitat was dominated by Bromus inermis (95% cover, N = 40)

 

with Agropyron repens present in smaller proportion (5%, N = 40). It was

 

also composed of isolated patches of goldenrod (Solidago graminifolia,

 

11

 

12

mw.nm m¢.mm mm.wm qpm omv mmo Papop

 

 

 

 

 

 

 

 

 

 

 

 

 

1 ON - 1 No.0 - - OP 1 acau_amsm a___H

- - ON - - No.0 1 - m sense wasp:
om oN - 4m.o NN.o - Ne ON - Ezewa_a maeoammam
ON ON oo_ ON.N me.m om.e_ mm as N__ .mmflmmflm» “a
ON - - NN.o - - o, - 1 .mflflmmumm am
oo_ oo_ ON mo.m_ No.N Nm.N mm_ Ne 04 mafia madamao
cop oo_ 00— mm.o _N.N NN.¢ ¢N_ NN NNN .Nmflmmumm am

.1 1. ON .. 1 mod 1 1 m mcquNwEm mzczgm
oN ON - No.0 No.0 - m m - .am.wmmmmmmum
ON ON - Nm.o Nm._ - m ON 1 .Nmmmm hm
om oo_ oo_ 0N.o om.N NN.N SN mm, eo_ meaa_m mNeau

1 om 1 1 qo.o 1 1 m 1 acmLCNFoLmo mzcwaewu
ON ON ON m_.o a_.o N_.o _ ON m mN mamcwuacau Lawsocaams<
ON ON ON Nm.o NN.N NN._ o_ 44 SN ssanse ama<

N N P N N F N N P mmwoaam

Aav Ar1ms.s cmv A 1m;.msmpmv
zucmaomeu mew _mmwm xpwmcmo

 

 

.mpwm Leg myopa przogwu Amg1np.ov mLUM1_.o m>wm co vmmmn mpwm auapm some so
umpwnm: pmmgoo ecu cw mmwumam zn Acne Eu No.~ Av meme“ Lo NocmzcmLL use .wmgm memn .Xpwmcmo ._ mfinmp

13

Table 2. Density and basal area of trees (3_7.62 cm dbh) by size class
in the forest habitat on each study site based on five 0.1-
acre (0.04-ha) circular plots per site.

 

 

 

 

 

Size Density Basal area
c1355 (stems-ha'l) (sq m.ha-1)
(dbh) 1 2 3 1 2 3
3-6 inches 306 138 257 2.95 1.40 2.07

(7.62-15.24 cm)

6-9 inches 138 99 69 3.83 2.70 1.97
(15.24-22.86 cm)

9-15 inches 128 109 104 8.91 8.09 8.29
(22.86-38.l cm)

15-21 inches 69 69 64 10.52 10.12 9.66
(38.1-53.34 cm)

21-27 inches 10 5 15 2.57 1.16 4.02
(53.34-68.58 cm)

27-33 inches 0 0 5 0.00 0.00 1.82
(68.58-83.82 cm)

Total 653 420 514 28.78 23.47 27.83

 

14

 

 

 

Aoo_v Loose Aoo_v

_oo.ov - N¢.N_ _m Nm mm ANV La>oa steameaaes pea:_eoo
Ao_v Ao_v Ao_v

_o.ov mm.__ - m.Nm H mm m.©_ N _m o.N N NE A_-a; soo.o.meaamv Noamcau Neopmeaac:
Amy Amv Amv

Fo.oV mm.N - N.F a m N.o a m e.o a N A,-a; moo.o. NV maLomam Naoomemeca
Aoo_v Aoo_v LOOPS

mo.ox - No.4 0N LN mm ANS La>oo Naocau
Ame Amy Ame

mo.ox NN._ - 4N.m a N.mN NN.N H N.NN om.m a m.mN Lev smegma Lo “Lora; E=e_xaz
Ame Amy Amv

No.0, Nm._ - oNN.o H N_._ NmF.o a mm.o 4mm.o a NF._ A_-a; 40.0.5 cmv amen Lammm
Amv Ame Amy

mo.ov NN.N - L.o a _N m.N a Na ©.N a 0N A_-e; No.0.meaamv Nawmcmn mate
Amv Ame Amv

No.0. OP._ - N.NP a No m.__ a No_ ¢.m a No_ Ameaasv venom L0 was E=E_xaz

Susana; “mated
1. 1. m N _
m. I Nx muwmlwuzum mpnmwem>

 

 

Lo cmgpmgz mewscmpwv op vow: mew: mummu sz mWFFm31mem=L¥ vcm Amxv umemzam1wcu
some :0 mpmpwnm; mnwgummu op new: mmpnmwgm> Co A v mNFm mpgsmm use .mcowng>wu uemucmum .mcmmz

.pcmowmwcmwm mam: mmucwgmmmwn cw>gmmno no:

.mpwm zuzum

.m mpamh

 

AOOPO AOO_V AOO_O

 

 

 

 

OO.OA - OO.N OO NO NO AOL LO>OO OOOOOOOLOI
AOO_V AOO_V AOO_O
_OO.OV OL.NO - O.NN O OO O.ON O ON O.ON O OO_ AEOV OOLOO CO NOOOOO EOOOXOz
OOOOOOO O_OOL
AOO_O AOO_V AOO_V
_OO.Ov - ON.OO OO NO NO “NO LO>OU PO>O_1OOOOLO
.13
.1 LOO AOO AOL
PO.OV O_.OP - O.N O O O.N O __ O._ O O AN-O 0.0v OOOOOOO OO>OO-OOOOOO
A.O.OOOOV OOOOOOO OOOOOL
1. 1. N N _
.m I x O_OOOLO>
N OOOO OOOOO

 

 

.A.O.OOOOO N O_OOO

Fig. 2.

16

Representative vegetation profiles of the interface between
field and forest habitats on the study sites derived from the
maximum height of oVerstory (37.62 cm dbh), dominant understory
(32.54 cm, <7.62 cm dbh), and dominant herbaceous and woody
ground-layer (<2.54 cm dbh) vegetation. The zone at which dom-
inant overstory trees (30.48-38.10 cm dbh) first occurred in

the stand is represented by cross-hatched areas, which designate
i 1 SD of the mean. Sample sizes are 10 (site 1), 30 (site 2),
and 15 (site 3). Differences between sites were not significant
(fl_= 3.91, E.>0°05)'

HEIGHT IN METERS

3O

20

20

20

O

20

17

 

 

 

 

 

 

 

 

 

 

 

 

IO

 

SITE I Field Forest .
\\
\\
- AOverstory \§1
El Understory \S‘
oWoody Ground-layer . \N
o Herbaceous ,
Ground-layer
\
' \
I \
I I l I" . \s I l
'\
\
\‘
SITE 2
o
01
‘0
Lu
V
SITE 3

 

 

 

DISTANCE FROM EDGE 1N METERS

20

18

S, canadensis), aster (Aster pilosus), burdock (Arctium minus), milkweed

 

 

 

(Asclepias syriaca), and other forbs. A discontinuous row of rose (Rosa

 

multiflora) paralleled the northern edge at about 60 m into the field.

 

Site 2.--About 0.4 km due east of site 1, site 2 was chosen. It
was a forest stand (woodlot #5, 8.98 ha) with an abandoned pasture

(9.09 ha) located just north of the forest. Carya glabra was dominant

 

and g, velutina and Acer rubrum were subdominant species in the stand.

 

Prunus serotina and Quercus alba were also important constituents in the

 

 

stand, but primarily in the smaller size classes. Bromus inermis was

 

again the dominant species in the field habitat (95%, N = 40) with

Phleum pratense composing a small percentage of dominant herbaceous cover

 

(5%, N = 40). Furthermore, the field habitat contained several isolated

patches of goldenrod (S, graminifolia, S, canadensis), aster (A, pilosus),

 

 

sweet clover (Melilotus alba, M, officinalis), wild carrot (Daucus cardta),

 

 

 

thistle (Cirsium arvense, S, vulgare), and other forbs. Several isolated

 

large trees and a few scattered shrubs were located throughout the field.

Site 3.--This site was located 0.2 km east of site 2. It consisted
of a forest stand (woodlot #6, 5.27 ha) with an old field (5.79 ha) sit-
uated immediately east. The tree species dominant in the forest stand
was 9, glgg_with g, glgS§g_and g, velutina associated with it as sub-

dominants. Prunus serotina was again an important species in the 3-6

 

inch size class. Dominance in the field habitat was shared by S, inermis
(55%, N = 40) and A, repens (45%, N = 40). Besides the dominant grasses,

isolated patches of solidago (S, graminifolia, S, canadensis), aster

 

 

(A, pilosus), burdock (Arctium minus), milkweed (Asclepias syriaca),

 

wild carrot (Daucus carota), hedge-parsley (Torilis japonica), and other

 

 

forbs were also present.

19

Although structurally the habitats on the three sites were quite
similar, there were several significant differences. The forest on site
2 was noted to be quite open beneath the canopy with dominant understory
cover of 37% and fewer trees per plot, especially individuals between
3-9 inches (7.62-22.86 cm) dbh, than at sites 1 or 3. Light penetration
to the forest floor was observed to be greater here than at sites 1 or
3. This situation is perhaps reflected in the increased forest ground-
level cover at site 2. Other differences included greater numbers of
forest understory and ground-level species, and an increase in forest
understory density at sites 2 and 3 than at site 1. These differences
are most likely attributed to man's past influence on grazing practices
and lumbering operations. Herbaceous vegetation was also found to be
much taller in the field habitat at site 1 than at sites 2 or 3.

Vegetation profiles of the interface between field and forest
habitats on the study sites indicate fairly abrupt structural changes
between the two habitats. The edge represents a rather distinct
discontinuity between short herbaceous field vegetation and tall woody
forest vegetation. However, along the edge of the forest habitats,

patches of blackberry (Rubus allegheniensis), raspberry (A, occidentalis,

 

 

A, striQOSUS), gray dogwood (Cornus racemosa), hawthorne (Crataegus sp.),

 

as well as saplings of black cherry (E, serotina), red maple (A, rubrum),
hickory (S, glabra, S, ovata), and oak (Q, velutina, Q, alba) created a
somewhat intermediate shrubby vegetation zone of varying width. Reflect-

ing a wildlife habitat improvement program of the early 1950's (Robert

 

Ankney, pers. comm.), the presence of rose (A, multiflora) along the
northern edge of forests at sites 1 and 2 and honeysuckle (Lonicera

tartarica) and rose along the eastern edge of the forest at site 3added

20
to the shrubby appearance. This shrubby vegetation zone was for the
most part quite narrow. Trees characteristic of dominant forest over-
story (30.48-38.10 cm dbh) were found on the average 11.1-12.3 m from
the edge. The major structural difference between sites appeared to be
a generally more gradual vertical transition from field to forest habi-
tat at site 2 than at sites 1 or 3.

Soils on the study sites were primarily Fox sandy loam (Typic
Hapludalfs) with a small percentage of Washtenaw loam (Typic Haplaquents).
Fox soils are on nearly level to sloping areas of glacial outwash plains,
terraces, and moraines. These are well-drained soils developed in loamy
outwash overlying stratified calcareous sand and gravel at depths of
51—102 cm commonly on a slope of 1-12%. The small areas of Washtenaw
loam were located in the field sections of sites 2 and 3. These soils
are on nearly level areas in depressions or moraines, till plains, and
outwash plains. They are poorly-drained soils formed in 51-102 cm of
recent alluvium from adjacent soils and in drift of Wisconsin age.

Due to the influence of the Great Lakes, the climate of the region
alternates between continental and semimarine, depending upon the force
and direction of the wind (summaries presented here are from U.S. Dept.
of Commerce, Local Climatological Data, 1936-1975). The increased
cloudiness associated with the prevailing westerly winds during the fall
and winter months helps to moderate minimum temperatures. When winds
diminish or change to the north or northeast, skies often clear to give
the cooler temperatures more commonly associated with an inland location.
The prevailing wind is southwesterly averaging about 16.6 km-hr“. Mean
maximum temperatures are above freezing from March through December and

between 23 and 28° C in the summer. Mean minimum temperatures are above

21
freezing from April through October. Precipitation is well-distributed
throughout the year and averages about 7,861 mm annually, with 50%
occurring from April through August. May and June, with 874 and 876
mm respectively, are the wettest months of the year. Summer precipita-

tion is mainly in the form of afternoon showers or thundershowers.

RESULTS

Nest Distribution Patterns

 

During the two-year study 194 nests of twenty-one species of open-
nesting passerines were found on the area (Table 4). These nests were
not evenly dispersed on the study sites, but had a definite pattern.
Most of the nests were found within about 10-20 m of the habitat discon-
tinuity at each site (Fig. 3). Seventy-five percent of the nests be—
longed to birds characteristic of mixed breeding habitats, i.e., birds
requiring open areas with dense cover for nesting and feeding and ele-
vated singing and observation perches (Kendeigh 1944, Johnston 1947,
Johnston and Odum 1956). These birds also accounted for the increase
in number of bird species found nesting near the edge. Indeed, each
species seemed to have a definite preference as to nest site location
relative to the habitat discontinuity at each site (Appendix).

The distribution patterns of nests almost represent a symmetrical
bell-shaped curve with some variation in height between sites. Nest
distribution patterns at study sites 1 and 3 were quite similar. How-
ever, site 1 had an increased frequency of nests toward mid-field be-
cause of Song Sparrows, generally considered a mixed-habitat species,
nesting in some multiflora rose. Song Sparrows also used the multiflora
rose for singing and observation posts and cover. The somewhat less
clumping of nests near the habitat discontinuity at site 2 when compared

to sites 1 and 3 is attributed to the relatively gradual change in

22

23

Table 4. Characteristic life-form of nesting habitat and numbers of
nests of open-nesting passerines located on the study area.
Numbers of nests in which eggs were determined to be present
are in parentheses.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Species Life-form # Nests

Acadian Flycatcher Tree 1 (l)
(Epidonax virescens)

Eastern Wood Pewee Tree 3 (1)
(Contopus virens)

Blue Jay Mixeda l (l)
(Cyanocitta cristata)

American Crow Tree 1 (0)
(Corvus brachyrhynchos)

Gray Catbird Mixed 4 (3)
(Dumetella carolinensis)

Brown Thrasher Mixed 1 (1)
(Toxostoma rufum)

American Robin Mixed 3 (3)
(Turdus migratorius)

Wood Thrush Tree 7 (4)
(Hylocichla mustelina)

Red-eyed Vireo Tree 9 (5)
(Vireo olivaceus)

Blue-winged Warbler Mixed 1 (1)
(Vermivora pinus)

Chestnut-sided Warbler Mixed 4 (2)
(Dendroica pensylvanica)

Eastern Meadowlark Herb 1 (1)
(Sturnella magna)

Red-winged Blackbird Herb 26(22)
(Agelaius phoeniceus)

Northern Oriole Mixed 1 (O)

(Icterus galbula)

 

24

Table 4 (cont'd.)

 

 

 

 

 

 

 

 

 

 

Species Life-form # Nests
Scarlet Tanager Tree 1 (l)
(Piranga olivacea)
Cardinal Mixed 15(10)
(Cardinalis cardinalis)
Indigo Bunting Mixed 32(27)
(Passerina cyanea)
American Goldfinch Mixed 2 (2)
(Spinus tristis)
Rufous-sided Towhee Mixed 5 (5)
(Pipilo erythrophthalmus)
Field Sparrow Mixed 59(59)
(Spizella pusilla)
Song Sparrow Mixed 17(15)
(Melospiza melodia)
Total 194(164)

 

aMixed--Herb-Shrub, Herb-Tree, or Herb-Shrub-Tree.

25

.AOEOLOOOOOO O>OOO OOOLOEOOV mmwumam mo LOOEOO OOO pOpOOO; mOOpmoc OO ELOO1OOOF OOOOOLOOOOLOOO Op
mcwvcouom mmprm OOOOO mg» co mmOOmemOO mcwpmmc1cmao Oo OOLOOOOO :oOOOOOmeOO pmm: m>OOOOOE=u

.N .O:

26

mmwhmz z_ moow 20mm moszE
o

 

 

 

 

 

on o . on ,0 on I .
1 I: O O 1 O FOE: -
L N
N M m. 1
v mmuz -
3 a: m Re I
FM. 352“ a
a for I 1
2.2-33

 

 

 

 

 

512. men mmuz -
£9 - Em. .

 

W\‘

 

thIm- 85... 22a Mg 85% 22... LEW-

E Os; E .Okg I_I._. \1
L. OLLE- O NOE -
mm I mm m m LL: -
m
v
25

 

 

695“.

 

 

SLSBN :10 lNBOHEId

 

27
vegetation profile (structure) at site 2. For example, open-canopied,
shrub-nesting passerines found a wider zone of acceptable nesting cover
which resulted in somewhat less clumping of nests near the habitat dis-
continuity. The presence of several isolated large trees and shrubby
vegetation as well as patches of stout, erect herbs scattered throughout
the field habitat at site 2 also provided singing posts or suitable nest-
ing cover farther into the field interior for several mixed-habitat
species (e.g., Indigo Bunting, Field Sparrow, and Song Sparrow). Indigo
Buntings, another species typical of mixed habitats, were attracted into
the fields on all study sites by patches of burdock, milkweed, goldenrod,
and sweet clover, all of which were used as nest sites. Indigo Buntings
were also observed nesting farther into the forest interior at site 2
than at sites 1 or 3. This phenomenon was probably due to the increased
openness of the stand which resulted in suitable, low vegetation cover
for nesting. Patches of blackberry and raspberry were used for the most
part as nest sites in the forest interior. Cardinals, which generally
nested in shrubs along the habitat discontinuity, were also found nest-
ing somewhat farther into the forest habitat at site 2 than at sites 1
or 3. Nests were located in multiflora rose, hawthorne, and a grapevine
(Vi£i§_sp.) tangle. Nests of mixed-habitat species were especially
concentrated near the edge at site 3. The closed forest stand and
complete absence of woody vegetation in the field seemed to restrict

these species to the interface.

Fledging Success and Mortality Patterns

 

Since overall nest distribution patterns were basically the same
between sites (i.e., high numbers of nests near the habitat discontinuity)

and the observed trends in fledging success and losses due to different

28
mortality factors relative to the habitat discontinuity were consistent
during the period of the study, the major emphasis was placed on the
effect of the habitat discontinuity on these factors and their inter-
relationships with each other. Graphical presentation of the relationship
of the different variables with distance from the edge revealed several
interesting trends (Fig. 4).

Nest distribution index, calculated as previously described, ini-
tially dropped quickly and then gradually fell to a low of around one
percent. Over one-half of the nests were found within 15 m of the habi-
tat discontinuity. This trend is comparable to the frequency distribu—
tions of the raw data presented in Fig. 3. Fledging success steadily
increased with distance from lows of 20-30% near the edge to highs of
60-70% away from the edge. Parallel but opposite to the trend in fledg-
ing success, losses due to predation dropped from 40-50% near the edge
to 0-10% away from the edge. Predation appeared to drop steadily with
increasing distance until about 45 m from the habitat discontinuity,
where it quickly fell. Brown-headed Cowbird parasitism more or less
followed the same trend as predation, but it was more variable. Para-
sitism was as high as 15-25% near the edge to as low as 0-5% farther
from it. Losses due to nest desertions appeared to be relatively
constant oscillating between 0-10% until about 65 m and then sharply
rising to 15-20%. Losses due to inclement weather, adult death, and
hatching failure did not show any definite trend with distance from
the habitat discontinuity.

Table 5 compares the degree of correlation between all variables
using the Kendall rank correlation test. Kendall partial rank correla-

tion coefficients were then calculated for certain correlations observed

Fig. 4.

29

Graphical presentation of the relationship of different nest
variables with distance from the habitat discontinuity. Dis-
tances covered by each distance category are as follows:

(1) 0.0—0.82 m; (2) 0.82—2.19 m; (3) 2.19-4.34 m; (4) 4.34—6.86 m;
(5) 6.86-10.06 m; (6) 10.06-14.18 m; (7) 14 l8—26.74 m;

(8) 26.74—46.24 m; (9) 46.24-65.58 m; and (10) 65 58-123.00 m.
Sample sizes are in parentheses.

30

(16)

IO

(I7) (I6)

(I6
(17)
(’6 (I6) (,7)
Una,

 

1—1

40' (17)

p-

23456789

 

20
0

3L _ 0 Z

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10

23456789

)
6
6%
pm—
I”
6
news
)
w?
(
W1
(
)
5
6..
61m
17,
6H
m.
M.
m.
- O — - P b p
0 O 0 O
6 4 2

3L mmwoonm GZEQMJO

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10

23456789

3.: meJ zo_._.<ommn_

 

 

D m

n 9

8

h 7

HS

5 5

n 4

H a

n 2

H 1

%.o
Act wm01_
Em_._._m<m<a

oEmBOo

2345678910

DISTANCE CATEGORY FROM EDGE

 

I I—Inn nrlril‘III

m 0

Ao\ov mmo1_

20:.mmmwo

mez

31

 

.moo.o um pcmom%w:mwm axes m5.0 pm ucmowmwcmwm axe

mmmod pm ucmowwmcmwm at mmo.o pm pcmuwmwcmwm «

 

 

 

ooo._ FON.01 ampv.01 wwm.01 oom.o —mm.ou omo.o Omo.o omo.ou cummv
u_:u<
ooo.— m—_.o omo.o ww—.01 mwo.o: omo.o omo.o omo.01 mL:_wmw
323% w
u.
ooo._. ommé: wvoé mwmdu Nwmd kamomdu rkoomd cowtmmmv W
O82 m.
coo._ mmo.o *mo¢.o aasoom.o- *va.o kmm¢.o1 EmOuwmmLmamfl
2538 m.
w
coo._ mmN.O1 mwo.on mo~.01 mo—.o Lmzpmmz s
quEwOOOH
ooo.— *¥%¥¢m©.ou kkrrqm©.o xrrr¢no.01 cowpmnwgm
000.? *rrrmmm.01 *rkrmmm.o mmmouzw
mOOmOOOO
ooo.— arkkooo.—1 Hoz
ooo.p omum Eoem
mocmpmwo
spawn mes—OOO OOOOLOOOO EmOuOmOLOO cmgpmwz :oOuOOwLO mmmousm Hoz mmvm Eogm
OOOO< mewsuumz ummz OLOOZOU “OmeOUOH mermOOOO mucmpmwo

 

mgouumO OOOFOpLoz

 

 

cmmzpmn mavgmcoprOOLprcw as» mcwgmqsoo .Azmpv p .mpcmwowmwmoo compo—mggou xcme Ofimucmx mo xwgumz

.mmwoozm mOOOOwOO mewpomOmO weepumw OOOFONLOE ucmgwmmwn any
new .mmmoozm OOOOOOOO .AHozv xmucw cowuznwgumvv “mm: .Ammumv Opwzcwpcoumwu HOOOOO: O EOLO wucmpmwc

.m mpnmh

32

in Table 5 to understand and clarify the relationships between the vari-
ables by uncovering spurious relationships and searching for intervening
linking variables (Table 6). Since distance from a habitat discontinuity
and nest distribution index (NDI) had a perfect negative correlation,
both will be considered together when presenting the results of partial
correlation. The most significant inferences are summarized as follows.

The significant correlation observed between fledging success and

distance (NDI) became T = 0.511 (-0.511) when the influence due to

xy-z
predation was held constant. Since this value is smaller than T =

XV
0.733 (-0.733), the relation between fledgingsuccess and distance (NDI)
is affected by losses due to predation. Controlling for distance (NDI),
the relationship betweenffiedgingsuccess and predation was not as
strong, which indicates that distance from a habitat discontinuity or
numbers of nests influences the predation rate which in turn affects
fledging success.

Partialling out losses due to Brown—headed Cowbird parasitism had
little effect on the relationship between fledgingsuccess and distance
(NDI). The correlation between fledging success and distance (NDI)
seems to be relatively independent of the influence of losses due to
parasitism. Furthermore, when the effect of distance (NDI) was held
constant, the correlation between fledging success and parasitism became
less important. This suggests that the relations between fledging success
and distance (NDI) and between parasitism and distance (NDI) are influenc-
ing the observed relation between fledging success and cowbird parasitism.
Moreover, the correlation between parasitism and predation was nearly

eliminated when distance (NDI) was kept constant, indicating that both

are affected by distance (NDI) in the same manner.

Table 6.

33

Kendall partial rank correlation coefficients, txy.z, of the

relation between the X and Y variables when the effect of

variation of a third VariabTe,_;, is eliminated.

 

 

 

.5 .X .Z Txy Txy-za
Fledging Distance Predation 0.733 **** 0.511
success (NDI) (-0.733)**** (-0.511)
Fledging Predation Distance -0.674 **** -0.358
success (NDI)

Fledging Distance Cowbird 0.733 **** 0.662

success (NDI) parasitism (-0.733)**** (-0.662)

Fledging Cowbird Distance -0.600 *** -0.472

success parasitism (NDI)

Predation Cowbird Distance 0.405 * 0.180
parasitism (NDI)

Nest Distance Predation 0.506 ** 0.392

desertion (NDI) (-0.506)** (-0.392)

Nest Predation Distance -0.349 -0.013

desertion (NDI)

Nest Adult Distance -0.412 * -0.449

desertion death (NDI)

 

aSince the sampling distribution of the Kendall partial rank cor-
relation is unknown, no tests of the significance of an observed rxy,z

are possible.

* Significant at 0.05.

** Significant at 0.025.
*** Significant at 0.01.
**** Significant at 0.005.

34

Partialling out the effect due to predation, the uncontaminated
relation between losses due to nest desertion and distance (N01) is
expressed by rxy,z = 0.392 (-0.392), which is less than Txy = 0.506
(-0.506). Controlling for the effect of distance (NDI), the association
between nest desertion and predation became essentially non-existent.

The conclusion is that the observed significant correlation between
losses due to nest desertion and distance (N01) is relatively dependent
on both being associated with losses due to predation. A logical explan-
ation is that the increased predation rate nearer the edge results in
loss of a certain proportion of nests which otherwise would have been
deserted by the adult birds. This is based on the assumption that nest
desertions are random and constant over the entire area. Losses due to
nest desertion would, thereby, be less nearer the edge than farther from
it, producing the observed correlation. Nevertheless, I cannot refute
the likelihood that a real difference might exist farther from the hab-
itat discontinuity caused by lack of secure cover or nearby conspecifics,
i.e., the need of some form of social facilitation (Johnsgard 1967), but
I find it doubtful.

When distance from a habitat discontinuity or NDI was held constant,
the relation between losses due to nest desertion and adult death was ob-
served to be similar. Therefore, it might be assumed that the relation
between these two sources of mortality is relatively independent of the
effect of distance (NDI). It appears that the constancy of incubation
and brooding increases the probability that an adult bird may be taken

by a predator while on the nest.

Differential Mortality Patterns

 

Because of the observation that several mortality factors may have

a differential effect on the eggs or nestlings, the chi-squared one-sample

35
test (Siegel 1956, 42-47) was employed to determine the significance of
any observed differences. The results are presented in Table 7.

Parental activity at the nest site was noted to increase greatly
once the eggs had hatched. Predators might be able to locate nests more
easily during the nestling stage due to increased activity of parents
feeding nestlings, especially where many nesting birds are concentrated.
However, there was no significant difference in percent predation between
the egg and nestling periods. In fact, predation appears to be evenly
distributed throughout the nesting cycle. Losses due to heavy storms
and death of one or both parents also seem to fall into this same cate-
gory.

Brown-headed Cowbird parasitism was found to be significantly more
destructive of eggs than nestlings. Most egg losses occurred from the
female removing an egg of the host after depositing her own egg. Losses
which occurred during the nestling period resulted either from the re-
maining nestling(s) of the host species being crowded out of the nest
by the generally larger and stronger cowbird nestling(s), or from slow
starvation due to the cowbird nestling(s) dominating the feeding se-
quence. The majority of parasitized nests (62%) had only one cowbird
egg, while the remaining nests had two or three. One exceptional nest
had a total of five cowbird eggs deposited in it.

Nest desertions which resulted in losses occurred early in the
nesting cycle, usually before incubation began. Eighty-two percent of
the deserted nests had only one or two eggs in them.‘ Only 18% of the
deserted nests had clutches of three or four eggs. Causes of nest
desertion were difficult to determine. Several nests may have been de—

serted because of cowbird egg deposition. Due to the difficulty in

36

Table 7. Differential effects of mortality factors on eggs and nestlings
at Rose Lake.

 

 

a
Variable Percent Loss 52 .2

Eggs ’Nestlings

 

 

Predation 17.1 15.5 0.44 >0.05
Inclement weather 3.2 5.0 2.17 >0.05
Cowbird parasitism 5.3 0.7 19.88 <0.001
Nest desertion 3.2 0.0 18.00 <0.001
Hatching failure 3.7 - - -

Adult death 0.9 0.0 - >0.05b

 

aSample size = 562.

bSince the expected frequencies were less than five, the binomial
test was used to determine the two-tailed probability.

37
ascertaining whether or not the parent was killed, several recorded
desertions could have been losses due to adult death rather than deser-
tion. Some nest desertions were possibly due to the disturbance at the
nest caused by checking nest contents. Others were perhaps due to in-

adequate cover near the nest.

Seasonal Effects

 

Seasonal differences in mortality were for the most part non-
existent (Table 8). However, Brown-headed Cowbird parasitism was found
to be important only to birds initiating nesting in late spring (June).
Birds breeding in early spring (May) or in the summer (July and August)
were essentially free from the depredations wrought by the cowbird.
Losses due to nest deSertion were also greater in June, which may be a

result of the higher incidence of cowbird parasitism.

Predator Activity

 

Predation, which was the major factor affecting fledging success,
accounted for 59.6% of all losses. In most instances, the cause of the
loss was impossible to ascertain.

Several nests in the study area may have been predated by eastern

chipmunks (Tamias striatus). Field Sparrows nesting in or near an edge

 

seemed especially vulnerable. On one occasion, I witnessed two adult
Field Sparrows chirping and flying about their nest site. On closer

inspection I saw one chipmunk in the small Juniperus communis in which

 

the nest was located, and another on the ground at the base of the shrub.
Only one nestling remained in the nest from a brood of four. The east—

ern chipmunk was the most frequently observed mammal on the study sites,

38

Table 8. Comparison of losses due to different sources of mortality
by month of nest initiation at Rose Lake. Sample sizes are
in parentheses.

 

 

 

 

Variable Percent loss A? :p
TMay June July-Augusta
(232) (223) (107)
Predation 29.7 36.8 29.9 2.97 >0.05
Inclement weather 8.6 7.2 9.3 0.55 >0.05
Cowbird parasitism 2.6 11.2 2.8 17.30 <0.001
Nest desertion 1.7 6.3 0.0 12.05 <0.01
Hatching failure 3.9 3.6 3.7 0.02 >0.05

Adult deathb 2.2 0.0 0.0 - -

 

aTwo nests occurred in August, each contained three eggs.

bExpected frequencies in each cell were too small; therefore, no
value for A? or probability could be calculated.

39
and could have been a very important predator on eggs and nestlings.
Nolan (1963) believed that the chipmunk was a significant predator of
nests of deciduous shrub birds. Other potential mammalian predators
observed or known to be in the area included shorttail shrew (Blarina

brevicauda), white-footed mouse (Peromyscus leucopus, E, maniculatus),

 

 

 

ground squirrel (Citellus tridecemlineatus), red squirrel (Tamiasciurus

 

 

hudsonicus), fox squirrel (Sciurus niger), longtail weasel (Mustela

 

 

frenata), striped skunk (Mephitis mephitis), opossum (Didelphis

 

virginiana), raccoon (Procyon lotor), domestic cat (Felis catus), and

 

 

 

red fox (Vulpes vulpes).

 

Other predators of open-nesting birds included the Blue Jay and

possibly the Common Crow, several raptors (Buteo jamaicensis, Falco

 

sparverius, 0tus asio), and various snakes (Lampropeltis doliata

 

 

triangulum, Coluber constrictor foxi, Thamnophis sirtalis sirtalis).

 

Blue Jays were considered the cause of broken eggs and dead nestlings

in several nests of Red-eyed Vireos and Cardinals. Based on the observed
presence of a Screech Owl (Q, SSjg) and blood-stained feathers at the nest
site, a Blue Jay and Wood Thrush were thought predated by an owl while

incubating and brooding, respectively. A milk snake (A, g, triangulum)

 

was observed near the nest of an Eastern Meadowlark, which was later
predated. No direct observations of predation were made during the

study, thus much of the above information is speculative.

DISCUSSION

A summary of the biologically important as well as statistically
significant pathways by which the habitat discontinuity at Rose Lake
affected nest distribution, fledging success, and losses due to different
mortality factors in open-nesting passerines is presented in Fig. 5.

Distance from a habitat discontinuity, which also represents
changes in habitat composition and structure, inversely affects numbers
of nests. Some nest sites, especially those associated with an open
canopy and dense shrubby vegetation (Cardinal, Gray Catbird, Brown
Thrasher, Rufous-sided Towhee), are not found in either adjoining habi-
tat interiors. Since availability of nest sites is probably limiting,
such birds are restricted to nesting near edges which effectively con-
centrates nests. Birds that may require an open canopy and high sing-
ing or observation perches near their nest sites but nest and feed close
to the ground in low vegetation (Field Sparrow, Song Sparrow, Indigo
Bunting) would also be limited to nesting near edges. The increase in
woody ground-layer vegetation resulting from the openness of the forest
stand and the presence of woody vegetation in the field habitat seemed
to influence the nest location of these mixed-habitat species.

According to the data, distance from a habitat discontinuity and
numbers of nests, separately or in concert, influence the percent pre-
dation which is the primary factor affecting fledging success. Because
predation was higher near edges than in habitat interiors, fledging
success was lower. Holling (1959) has demonstrated that predation is

40

Fig. 5.

41

A summary of the biologically important as well as statistically
significant pathways by which the habitat discontinuity (edge)
affected nest distribution, fledging success, and losses due

to different mortality factors in open-nesting passerines.

The width of connecting lines designate the significance of the
relationship. The direction and sign of the arrows indicate
probable cause-and-effect relationships and whether or not it

is direct or inverse.

42

cowamo (-) DISTANCE

PARASITISM FROM
EDGE

/ _

 

 

 

PREDATION L 1-1 ,

 

 

(+)\ I t

 

 

NUMBERS
OF
NESTS

ADULT
DEATH

NEST
DESERTION

FLEDGING
SUCCESS

 

43

affected mainly by prey and predator density. The increase in numbers
of prey consumed per predator as prey density increases is termed the
functional response. An increase in the numbers of predators with in-
creasing prey density is the numerical response. Fretwell (1972) has
also shown that, within a habitat, avian breeding success is mainly
density-dependent, due to density-dependent nest predation. In fact,
the nest mortality of each open nest is dependent upon the density of
all open nests, regardless of species (Fretwell 1972). A major portion
of the increased predation is attributed to a functional response of
predators to increased nest numbers near edges. Because losses due to
predation did not significantly differ during the breeding season, it
is improbable that predator density was able to increase sufficiently by
reproduction or immigration in order to have a direct numerical response.

There is some indication that the number of prey consumed per pre-
dator is not just directly proportional to prey density, but that the
activity of predators searching for prey is greater near habitat dis-
continuities. Bider (1968) observed that such small mammals as red
squirrels, eastern chipmunks, shrews, and weasels were significantly
more active in the vicinity of the forest edge. Raccoons were also
found to extend their activity to a greater ecotonal depth than smaller
animals, and to utilize forest interiors only\sparingly (Bider,
Thibeault, and Sarazin 1968, Bider 1968). Bider (1968) thought that
the forest edge may function as a biological barrier which causes animals
to move parallel to it for some distance. This would result in animals
whose home ranges are delineated by barriers to be more active along it,
thus creating an activity vacuum at some distance away from the barrier

(Bider 1968). This fact could also explain the increased activity of

44
eastern chipmunks in the ecotone between field and forest habitats stud-
ied by Forsyth and Smith (1973). They observed that 69-90% of all home
ranges were entirely within, or extended into, the ecotone. It seems
likely that higher predation observed near habitat discontinuities at
Rose Lake can be attributed to both an increase in numbers of nests and
a greater activity of potential nest predators in the vicinity of edges.

Brown-headed Cowbird parasitism was also observed to be higher near
edges. It too was affected by distance from a habitat discontinuity or
nest numbers, separately or in combination. Although its effect upon
fledging success was not as important as predation, McGeen (1972) has
demonstrated that as cowbird pressures increase fledging success of
open-nesting birds can decrease. Because cowbird pressure is dependent
upon parasite and host nest density (McGeen 1972), a parasite can also
be considered to have a response to prey density (egg laying), which can
be identified as a functional response (Holling 1959). However, cow-
birds would not be expected to show a direct numerical response during
one breeding season for obvious reasons. Therefore, the increased inci-
dence of parasitism can be attributed in part to a functional response
of Brown-headed Cowbirds to increased nest numbers in the vicinity of
edges.

Furthermore, there is evidence that cowbirds exhibit increased
activity near habitat discontinuities due to their nest searching be-
havior (Berger 1951, Johnston 1960, Payne 1973, Norman and Robertson
1975).. Cowbirds at Rose Lake were often observed on a high perch near
the habitat discontinuity in a position to see nest building activity
of potential host species, or on several occasions moving about in the

vegetation along the field-forest edge presumably searching for nests

45
(see Norman and Robertson 1975). I seldom observed cowbirds in the for—
est or field interiors on the study sites. This indicates that cowbirds
may find it difficult to locate nests in habitats where their view of
nesting birds is obscured by dense foliage. It would also be a more
economical use of time and energy to concentrate their search in areas
where host nest density is high. Because cowbirds in northern, lower
Michigan begin egg laying in late May or early June and terminate in
July (Payne 1965, 1973), birds nesting in June and near a habitat dis-
continuity at Rose Lake were especially susceptible to cowbird parasitism.
Ironically, the cowbird also became a potential victim of increased pred-
ator activity due to its parasitism of nests near habitat discontinui-
ties. In fact, cowbirds had a fledging success of only 11.8% during the
study period.

In conclusion, the creation of narrow forest edges by agricultural,
forestry, or wildlife management programs does seem to attract a variety
and abundance of open-nesting passerines. This phenomenon is what is
classically known as the edge effect. Based on these observations,
earlier workers assumed that edges had high suitability for certain
avian species. However, this study has shown that birds nesting near
habitat discontinuities at Rose Lake are subject to high rates of preda-
tion and Brown-headed Cowbird parasitism due in part to large numbers of
nests near such areas. I would furthermore add that total production
of independent avian off-spring appears to be comparatively constant
over the entire area, or slightly higher near edges (unpublished data).

I conclude that actual habitat suitability decreases with increasing
nest frequencies toward relatively narrow field-forest edges. But why

should open-nesting passerines nest in the vicinity of edges in such

46
numbers? I propose that the narrow man-made forest edges studied in
central Michigan could well function as ecological traps for avian species
characteristic of habitats with mixed life-form. All the normal struc-
tural cues, indicating good nest sites, singing and observation perches,
cover, and food availability, are presumably present for these species
near edges. Those birds selecting to nest in the vicinity of the edge
perceive it as a suitable nesting area. However, they are perhaps
"tricked" by the juxtaposition of field and forest into settling near
the narrow zone of shrubby vegetation bordering these forested areas.
Extensive areas of narrow field-forest edge have only been in existence
for a brief period relative to the evolutionary time-scale. Forests,
which once covered this part of Michigan almost entirely, have been
reduced about 90% since logging first began around 1865 (information
derived from soil surveys of Clinton and Shiawassee counties, Michigan).
Therefore, these man-made situations are perhaps unrepresentative of
the ecological niche in which these mixed-habitat species evolved, and
thus they may be poorly adapted to cope with the increased predation

near them.

APPENDIX

Means, standard deviations, and sample size ( ) of distances from the
habitat discontinuity used by open-nesting passerines as nest locations
on study sites 1, 2, and 3. Positive values are forest locations, while
negative values are field.

 

 

 

 

Species Distance in Meters
l 2 3
Acadian Flycatcher - - 70.7
(1)
Eastern Wood Pewee 11.7 36.4 72.0
(1) (l) (1)
Blue Jay - 7.6 -
(1)
American Crow 32.0 - —
(1)
Gray Catbird 5 6 i 1.95 2 6 6.6
(2) (l) (1)
Brown Thrasher - - 0.0
(1)
American Robin -6.7 -l3.4 i 4.45 -
(l) (2)
Wood Thrush - 28.0 i 22.81 48.4 i 2.34
(4) (3)
Red-eyed Vireo 41.0 45.4 i 42.61 53.7 t 15.63
(1) (2) (6)
Blue-winged Warbler - 6.3 -
(l)
Chestnut-sided Warbler - 9.2(i)4.40 -
4

47

48

APPENDIX (cont'd.)

 

 

 

 

Species Distance in Meters
l 2 3
Eastern Meadowlark - - -41.l
(l)
Red-winged Blackbird -34.1 i 22.93 -86.0 i 21.21 -54.5 i 27.04**
(8) (4) (14)
Northern Oriole - - 0.0
(1)
Scarlet Tanager - 31.5 -
(1)
Cardinal 0.2 i 3.25 1.2 i 6.55 0.1 i 3.01
(4) (9) (2)
Indigo Bunting -9.6 i 26.65 8.2 : 32.93 -5.0 i 5.28
(8) (l7) (7)
American Goldfinch - -38.7 i 10.37 -
(2)
Rufous-sided Towhee - 0.0 -0.4 i 2.20
(l) (4)
Field Sparrow -4.9 i 5.37 -29.5 i 21.16 -5.7 i 6.40****
(7) (23) (29)
Song Sparrow -33.7 1 27.80 -32.2 i 51.62 -2.4 i 3.45
(10) (5) (2)

 

**

fl_
***1: fl

2. <0.025.

8.02,
26.78, p <0.005.

LITERATURE CITED

 

LITERATURE CITED

Allen, D. L. 1962. Our wildlife legacy. Funk and Wagnalls, New York.
422 p.

Beecher, W. J. 1942. Nesting birds and the vegetation substrate.
Chicago Ornithological Society, Chicago, Illinois. 69 p.

Berger, A. J. 1951. The cowbird and certain host species in Michigan.
Wilson Bull. 63:26-34.

Bider, J. R. 1968. Animal activity in uncontrolled terrestrial
communities as determined by a sand transect technique. Ecol.
Monogr. 38:269-308.

, P. Thibeault, and R. Sarrazin. 1968. Schemes dynamiques
spatio-temporels de l'activite de Procyon lotor en relation avec
le comportement. Mammalia 32:137-163.

 

Dasmann, R. F. 1964. Wildlife biology. John Wiley and Sons, Inc.,
New York. 231 p.

Dow, D. D. 1969. Home range and habitat of the Cardinal in peripheral
and central populations. Can. J. 2001. 47:103-114.

Edeburn, R. M. 1947. A study of the breeding distribution of birds in

a typical upland area. Proc. W. Va. Acad. Sci., Vol. 18, W. Va.
Univ. Bull. Ser. 47, no. 9-I:34-47.

Forsyth, D. J., and D. A. Smith. 1973. Temporal variability in home
ranges of eastern chipmunks (Tamias striatus) in a southeastern
Ontario woodlot. Am. Midl. Nat. 90:107-117.

Fretwell, S. 1968. Habitat distribution and survival in the Field
Sparrow (Spizella pusilla). Bird-Banding 39:293-306.

 

Fretwell, S. D. 1969. On territorial behavior and other factors in-
fluencing habitat distribution in birds. III. Breeding success
in a local population of Field Sparrows (Spizella pusilla Wils.).
Acta Biotheor. 19:45-52.

 

1972. Populations in a seasonal environment. Princeton
Univ. Press, Princeton, New Jersey. 217 p.

49

50

Good, E. E., and C. A. Dambach. 1943. Effect of land use practices on
breeding bird populations in Ohio. J. Wildl. Manage. 7:291-297.

Heppleston, P. B. 1972. The comparative breeding ecology of oyster-
catchers (Haematopus ostralegus L.) in inland and coastal habitats.
J. Anim. Ecol. 41:23-51.

 

Holling, C. S. 1959. The components of predation as revealed by a
study of small—mammal predation of the European sawfly. Can.
Entomol. 91:293-320.

Holm, C. H. 1973. Breeding sex ratios, territoriality, and reproductive
success in the Red-winged Blackbird (Agelaius phoeniceus). Ecology
54:356-365.

 

James, F. C., and H. H. Shugart, Jr. 1970. A quantitative method of
habitat description. Audubon Field Notes 24:727-736.

 

Johnsgard, P. A. 1967. Animal behavior. Wm. C. Brown Co., Inc.,
Dubuque, Iowa. 156 p.

Johnston, 0. W., and E. P. Odum. 1956. Breeding bird populations in
relation to plant succession on the piedmont of Georgia. Ecology
37:50-62.

Johnston, 0. W. 1970. High density of birds breeding in a modified
deciduous forest. Wilson Bull. 82:79-82.

Johnston, R. F. 1960. Behavioral and ecologic notes on the Brown-
headed Cowbird. Condor 62:137-138.

Johnston, V. R. 1947. Breeding birds of the forest edge in east-
central Illinois. Condor 49:45-53.

Kendeigh, S. C. 1944. Measurement of bird populations. Ecol. Monogr.
14:67-106. _

Krebs, J. R. 1971. Territory and breeding density in the Great Tit,
Parus major L. Ecology 52:2-22.

 

Lay, D. W. 1938. How valuable are woodland clearings to birdlife.
Wilson Bull. 50:254-256.

Leopold, A. 1933. Game management. Charles Scribner‘s Sons, New York.
481 p.

McGeen, D. S. 1972. Cowbird-host relationships. Auk 89:360-380.
Nettleship, D. N. 1972. Breeding success of the Common Puffin

(Fratercula arcticus L.) on different habitats at Great Island,
Newfoundland. Ecol. Monogr. 42:239-268.

 

Nice, M. M. 1957. Nesting success in altricial birds. Auk 74:305-321.

51

Nolan, V., Jr. 1963. Reproductive success of birds in a deciduous
shrub habitat. Ecology 44:305-313.

Norman, R. F., and R. J. Robertson. 1975. Nest-searching behavior in
the Brown-headed Cowbird. Auk 92:610-611.

Odum, E. P. 1971. Fundamentals of ecology. 3rd ed. W. B. Saunders
Co., Philadelphia, Pa. 574 p.

Payne, R. B. 1965. Clutch size and numbers of eggs laid by Brown-
headed Cowbirds. Condor 67:44-60.

. 1973. The breeding season of a parasitic bird, the Brown-
headed Cowbird, in central California. Condor 75:80-99.

Parker, J. W. 1972. A mirror and pole device for examining high nests.
Bird—Banding 43:216-218.

Penfound, W. T., and E. L. Rice. 1957. An evaluation of the arms-length
rectangular method in forest sampling. Ecology 38:660-661.

Ricklefs, R. E. 1969. An analysis of nesting mortality in birds.
Smithsonian Cont. 2001. 9. 48 p.

Robertson, R. J. 1972. Optimal niche space of the Red-winged Blackbird
(Agelaius phoeniceus). I. Nesting success in marsh and upland
habitat. Can. J. Zool. 50:247-263.

Shugart, H. H., Jr., and B. C. Patten. 1972. Niche quantification and
the concept of niche pattern, p. 283-327. lfl B. C. Patten (ed.).
Systems analysis and simulation in ecology. Vol. 2. Academic
Press, New York. 592 p.

Siegel, S. 1956. Nonparametric statistics for the behavioral sciences.
McGraw-Hill Book Co., New York. 312 p.

Wiens, J. A. 1969. An approach to the study of ecological relationships
among grassland birds. Ornithol. Monogr. 8. 93 p.

Winkworth, R. E., and D. W. Goodall. 1962. A crosswire sighting tube
for point quadrat analysis. Ecology 43:342-343.

 

 

 

 

 

 

L. ....