Patches and Structural Components for a Landscape Ecology

Author(s): Richard T. T. Forman and Michel Godron

Reviewed work(s):
Source: BioScience, Vol. 31, No. 10 (Nov., 1981), pp. 733-740
Published by: University of California Press on behalf of the American Institute of Biological Sciences
Stable URL: http://www.jstor.org/stable/1308780 .
Accessed: 18/05/2012 07:00

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

University of California Press and American Institute of Biological Sciences are collaborating with JSTOR to
digitize, preserve and extend access to BioScience.

http://www.jstor.org
 Patches and Structural
Components
For A Landscape
Ecology
Richard T. T. Forman and Michel Godron

Landscapes as ecologicalunitswith structureandfunctionarecomposedprimarily Between the stands of a cluster are

of patchesin a matrix.Patchesdifferfundamentally in originanddynamics,while
size, shape,and spatialconfigurationarealso important.Linecorridors,stripcorri-
dors,streamcorridors,networks,andhabitationsaremajorintegrativestructural
characteristicsof landscapes.(Acceptedfor publication29 May1981)
Landscapessurroundus, yet curiously
it is hard to find people with the same
concept of a landscape. Artists and hu-
manistscommonlyportraythe landscape
as what the eye can perceive and some-
times limit the idea to naturallandforms
or communities.Such a landscapegener-
ally includes a high degree of spatial
heterogeneity.In geographicalliterature,
the landscape plays a central role, with
most definitionsfocusing on the dynamic
relationship between two characteris-
tics-natural landformsor physiographic
regions and human cultural groups
(Grossman 1977, Mikesell 1968, Sauer
1963).In this article we ask whether the
landscape is a recognizable and useful
unit in ecology, with a distinctive struc-
ture and function that can be analyzed,
as is done for organismsor ecosystems.
What are the structuralcomponents of a
landscapeand their characteristics?Are
there interesting,indeed critical,ecologi-
cal questions facing us that may be
solved using a landscape approach?
Walkingin a small area of an agricul-
tural landscape, one might encounter a
corn field, a bean field, an abandonedold
field, an uplandoak stand, and a lowland
elm-ash-sycamore woods adjacent to
one another.If one studiedthis clusterof
five specific communitiesor stands, one
would find fluxes of energy, mineralnu-
trients, and species between adjacent
stands, indicating considerable interac-
Formanis with the Departmentof Botany, Rutgers
University, New Brunswick,NJ 08903. Godronis
with the Centre d'Etudes Phytosociologiques et
Ecologiques,CNRS, B. P. 5051, Route de Mende,
34033Montpellier,France.
tion among stands of the cluster. If one
moved several kilometers away within
the landscape, one would find a similar
cluster of stands with similar interac-
tions. Moving on, one would find this
clusterrepeateduntil enteringa different
geomorphologicalarea, or an area sub-
jected to differentnaturalor humandis-
turbances. Here, a different cluster of
interactingstands would be evident as
one entered, for example, a landscapeof
ridges and valleys, a suburban land-
scape, or a sandy forested landscape.
Such observationsare at the heart of the
landscapeconcept, which we describeas
follows.
A landscapeis a kilometers-widearea
where a cluster of interactingstands or
ecosystems is repeated in similar form.
The landscapeis formed by two mecha-
nisms operating together within its
boundary-specific geomorphological
processes and specific disturbances of
the component stands.
Landscapesvary considerablyin areal
extent, and a localized area of a few
metersor hundredsof metersacross is at
a finer level of scale than a landscape.
Because of the area's geomorphology,
the complex of landforms and parent
materials present is relatively constant
over a landscape.Each stand has a char-
acteristicdisturbanceregime (the sum of
the frequencies, intensities, and types of
individual disturbances). A cluster of
stands or a "stand cluster," therefore,
has a disturbanceregime cluster, which
in turn is fairly constant throughoutthe
landscape. Disturbances include both
naturalevents and humanactivities such
as fire, hurricanes, agricultural prac-
tices, or forest cutting.
transitionzones or ecotones, which may
vary from being abrupt to gradual and
wide. In less disturbedlandscapes,grad-
ual community gradients may be com-
mon or uncommon,dependingupon how
sharp environmental changes are with
distance. However, with greater distur-
bance, especially by human activity, a
landscape mosaic of ecosystem patches
with distinct boundaries comes into
sharperfocus.
We suggest that landscape' is a dis-
tinct, measurableunit with several inter-
esting ecological characteristics.Within
the landscape is a recognizable and re-
peated cluster of ecosystems and distur-
bance regimes. The boundary between
landscapes (which differin geomorphol-
ogy and disturbance) is relatively dis-
tinct, particularly in vegetation struc-
ture. Ecologically, landscapestructureis
measuredby the distributionof energy,
mineralnutrientsand species in relation
to the numbers, kinds, and configura-
tions of the component ecosystems.
Landscapedynamics is the flux of ener-
gy, mineralnutrients,and species among
the component ecosystems, and conse-
quent changes in those systems.
The key landscapestructurequestions
today center on the importanceof num-
bers, kinds, and configurationsof eco-
systems. Randomness is rare within a
landscape. The overwhelmingnumberof
'Relatedconcepts: A region is boundedby a com-
plex of physiographic,economic, social and cultural
characteristics(Dickinson1970,Isard1975).A stand
(ora localizedcommunity)is the groupof organisms
at a specificlocality, and is homogeneousenoughto
be considereda unit(Greig-Smith1964,Daubenmire
1968).The ecosystemconcept--organismsand their
encompassingabioticenvironment-may be applied
at any level of spatialscale (Odum1971).However,
in practiceone looks for relativehomogeneityso as
to characterizean ecosystem with a limitednumber
of measurements(Woodwell and Whittaker1968,
Forman1979a,Bormannand Likens 1980).Though
one may applythe ecosystemconceptto a heteroge-
neous regionor landscape,in this articlewe limitits
use to standswithina landscape.

November 1981 733

species exhibits an aggregated or clus-
tered distributionof individuals(Chessel
1978,Greig-Smith1964, Kershaw 1973),
and even in randomand regularspecies
distributions,some aggregationsof indi-
viduals normally are present in a sur-
roundingarea of lower density (Godron
1966, 1971). This basic aggregationpat-
tern of individualsof a species underlies
the patchiness of vegetation and animal
communities so commonly seen in
nature.
In simplestterms, patches are commu-
nities or species assemblagessurrounded
by a matrixwith a dissimilarcommunity
structureor composition.The matrixex-
hibits several characteristicsitself, such
as the degree of heterogeneityand con-
nectivity, but in this article we focus on
patches and the other structuralcompo-
nents, corridors, networks and habita-
tions. We further limit the analysis to
patches at a single level of scale, the
landscape, though most of the resulting
patternsappear to apply to all levels of
scale.
PATCHORIGINS
Causal Mechanisms
Five causal mechanisms predominate
and the five types of patches produced
differ strikingly in their dynamics and
stability (Forman 1979b).A spot distur-
bance patch results from disturbanceof
a smallarea in the matrix(Figure 1). For
example, patches are produced by a
small fire in a grassland, a large blow-
down in a forest, overgrazingby a local
explodingpopulationof rodents, or local
spraying of a generalized insecticide.
Otherexamplesare given by Heinselman
(1973), Levin and Paine (1974), Pickett
and Thompson (1978), and Forman and
Boerner (1981). Following the distur-
bance, succession proceeds until the
patch disappears by becoming like the
matrix;that is, population changes and
immigrationsand extinctions of species
take place until the relative abundances
of the species are similarto those of the
surroundingmatrix. In unusual cases,
especially where the intensity of distur-
bance is severe or the matrixis undergo-
ing rapidchange, succession may lead to
a semi-stable patch that differs signifi-
cantly from the matrix. The spot distur-
bance patch typically has high popula-
tion changes and species immigration
causes its ultimatedisappearance.
A remnant patch is caused by wide-
spread disturbancesurroundinga small
area, the inverse of the spot disturbance
734
A. SPOT DISTURBANCE D. INTRODUCEDPATCH
An environmental resource patch re-
PATCH
flects the normalheterogeneousdistribu-
DISTURBANCE INTRODUCEDBY PEOPLE
tion of resources in the environments
0000111110000 o000111*0000 and results from the environmentalre-
8. REMNANT PATCH
sources of a relatively permanent and
DISTURBANCE DISTURBANCE E. EPHEMERAL PATCH
discrete area which differ from the sur-
+++ 4
iTIME roundingarea. Unlike the other patches,
OOOOO1eenon=n OO oo0000000Qo the environmentalresource patch is not
C. ENVIRONMENTAL dependent on disturbance. Concentra-
RESOURCE PATCH tions of amphibians and reptiles in a
desert oasis, patches of heaths on an
t ft
RESOURCE
oooooooooooo exposed mountain ridge, acid-tolerant
mosses in a glacier-causedbog, and pol-
Figure 1. Patch origins. Species dynamics linatorsin a moist alpine gully are exam-
withina patch and turnoverof the patch itself ples. Other examples are described by
differsubstantiallyaccording to the mecha-
nisms causing a patch. 0 0 0 = matrix; Brown (1971), MacArthur and Wilson
0 0 0 = patch; disturbance = a sudden (1967), Simberloff(1976), Smith (1974),
severe environmentalchange. Stiles (1979),and Willis (1974). Since the
cause of the environmental resource
mechanism. This patch is a remnant of patch is relatively permanent,the patch
the previous communityembedded in a is permanent,and species changes sim-
matrix that has been disturbed. Exam- ply reflect those normal in the interac-
ples of remnantpatches are woodlots in tion between a small community in dy-
an agriculturalarea, a shrub-coveredis- namic equilibrium with a surrounding
land produced by flooding a valley, a matrixcommunity.
breedingwarblercommunityon a south- An introduced patch is dominated by
facing slope that survived a rare freeze, an aggregationof individualsintroduced
or a pocket of herbivores that escaped into a matrix by people. Pine and euca-
the invasion of an aggressive non-native lyptus plantations (Pinus, Eucalyptus),
species. Furtherexamples are described golf courses, fields of wheat and corn
by Galli et al. (1976), Gottfried (1979), (Triticum,Zea), or a largefeeding station
Helliwell (1976), Pollard et al. (1974), that attractsvertebratespecies to a small
Seignobos (1978), and Southwood area are examples. Introduced patches
(1961).If the disturbancein the matrixis remainas long as the humandisturbance
temporary,succession will proceed until regime maintainsthem. Thereafter,spe-
the matrix converges with the patch in cies from the matrix colonize, and like
species similarity.Here again the patch the spot disturbance patch, the patch
disappears.If this convergence is rapid, disappears as it converges with the
the patch may change relatively little in matrix.
species composition. However, if the An ephemeral patch is a transient ag-
disturbance of the matrix is chronic, gregation of species caused by normal
inhibitingthe normal successional rate short-livedfluctuationsiri resource lev-
and direction, the patch will remain. In els, that is, levels of biotic or abiotic
this case, a net loss of species may take environmentalchange that are frequent
place (Diamond 1972, Willis 1974). This enough and of a low enough intensity
hypothesizednet loss would be rapid at that species have adaptedto them. Espe-
first, finallydroppingto zero, a response cially rare or severe environmental
referredto as a relaxation period. The changes are considered disturbances,
species lost are those requiringa habitat which in turncause spot disturbanceand
larger than the remnant patch or those remnantpatches. Examples of ephemer-
sensitive to a modified microenviron- al patches are mammalsfeeding at dawn
ment within a patch. arounda large savannamudhole,a local-
Hence, remnant patches vary from ized bloom of annualsin the desert, or a
short-lived,as producedby a single natu- large shrubbyarea in field-to-forestsuc-
ral or humandisturbance,to long-lived, cession. However, the rapid-turnover
resulting from chronic human distur- ephemeral patch appears more promi-
bance. In the same way, spot distur- nent at finer levels of scale than the
bance patches may be short- or long- landscape.
lived. In remnant patches caused by In summary,patch is a spatialconcept
chronic disturbance,the net loss of spe- focused on a small area. Both the causal
cies duringa relaxationperiod results in mechanismsof patches and the resulting
the patch remaining,but with a species dynamics of patches differ greatly. As
composition differing from the original with most biological patterns, some
patch. overlapsexist amongthe five basic patch
BioScience Vol. 31 No. 10
origins. For example, a severe chronic
disturbance in the matrix might so
changethe matrixthat convergencewith
a remnantpatch community is prevent-
ed, and the remnantpatch develops into
an environmentalresource patch.
Other Characteristics
A multitude of possible community
types, namedprimarilyby physiognomy
or predominantspecies, may character-
ize patches in a landscape. The number
of such different community types is a
key structural characteristic of land-
scapes, not only for mapping, but to
provide an index of the range of biotic
richness, productivity,and nutrientand
water fluxes in a landscape.
In addition, each ecosystem compo-
nent is normallypatchy in the landscape.
For example, superimposingmaps of soil
types, tree communities, and herbivo-
rous mammal communities for a land-
scape may show several places where
boundaries coincide and many places
where they do not. The degree of con-
gruityin space amongthe units of differ-
ent components is useful in mapping,
land use planning,and analyses of land-
scape structure(Forman1979b,McHarg
1969).
PATCHSIZE
Productivity,nutrientand water flux,
and species dynamics are all affected by
the size of landscape patches. Island
biogeographic theory developed from
studying archipelagoes in a matrix of
water lends insight into the relationship
between species and area. The number
of species, S, (= species diversity)on an
island was related directly to three fac-
tors in order: the island area, its isola-
tion, and its age (MacArthurand Wilson
1967). The basic island area effect,
though, is mainly due to habitatdiversi-
ty; in most cases, larger islands simply
have more habitats which, therefore,
supportmore species. However, there is
also an area effect: When the habitat
diversity of large and small islands or
patchesdoes not differ,more species are
typicallyfound on the former(Formanet
al. 1976,Simberloff1976).Finally, one of
the major factors determiningdiversity
on an island or patch is the history and
present regime of disturbance(Carlquist
1974,Pickett and Thompson 1978).Sum-
marizingthe patternsfor islands, species
diversity is a function of certain island
characteristics listed in the suggested
order of overall importance:2
November 1981
S = f(habitatdiversity ? disturbance+
area - isolation + age)
Patches in the landscape, however,
differ significantly from islands sur-
rounded by water (Forman 1979b).
Above, we analyzed patch origins and
here note that average turnover rates
(the appearance and disappearance)of
landscapepatches may be high, whereas
islands are essentially permanent.Simi-
larly, the sharpnessof the patch bound-
ary varies greatlyin the landscape(Whit-
taker 1973), and gradual gradients may
be more conducive to the movement of
species between patch and matrix. The
heterogeneityof the landscape matrixis
often high, which implies a large source
of species in the matrixand strongdirec-
tional effects of the matrix on different
sides of the patch. The landscapematrix
may be used as a rest stop for many
species moving between patches, partic-
ularly in the limited area of a landscape
comparedwith extensive oceanic archi-
pelagoes. Here the importanceof isola-
tion, a central characteristic of island
biogeographictheory, is lessened.
Withina landscape,the "species rain"
appearsto be high, that is, most species
reach most patches within their life cy-
cle. Therefore,when species extinctions
take place in patches, rapid recoloniza-
tion is facilitatedand the effect of isola-
tion minimized.Whilethis species rainis
suggestedto be high for a community,as
measured by species diversity, a rela-
tively smallproportionof individualspe-
cies has a limited dispersabilitywithin
the landscape. Isolation in the landscape
may be critical for these individualspe-
cies, many of which are uncommon.
Thus, in conservationnot only must the
basic communitypatternbe considered,
but also the specific populations (Dia-
mondand May 1976,Formanet al. 1976,
Pickett and Thompson 1978, Simberloff
1976, Terborgh 1976). Summarizingthe
species diversity patterns for landscape
patches we suggest:
S = f(habitatdiversity + disturbance+
area + age + matrixheterogeneity
- isolation - boundary discreteness)
Landscapepatch area has been shown
to correlate strongly with species diver-
sity (Galli et al. 1976, Gottfried 1979,
Moore and Hooper 1975,Peterken 1974,
Robbins1980,Whitcomb1977),but rare-
ly has area been considered separate
2+ = positively related to diversity; - = negatively
related;+ = usuallynegatively,but sometimespos-
itively, related. Units are not considered in this
encapsulation.
from habitatdiversity. When patch area
alone is evaluated, we find it to be an
importantdeterminantof species diversi-
ty, and that species groups (such as
trees, seed-eating birds and insectivo-
rous birds) respond differentlyto patch
area(Elfstrom1976,Formanet al. 1976).
PATCHEDGEAND SHAPE
The microenvironmentin the center of
a tiny patch of woods differs strikingly
from the center of an extensive woods.
This results largely from penetrationof
airfrom the surroundingmatrixthrough-
out the tiny woods, whereas this air
penetrates only a limited distance into
the edge of the extensive forest. The
outer band of a patch, which has an
environmentsignificantlydifferentfrom
the interiorof the patch, is known as the
patch edge. This produces an edge ef-
fect, that is, a differencein species com-
position and abundancein the edge. For
example, differences between the edge
and interior of deciduous forests in
North America and Europe have been
documentedfor a host of meteorological
factors, vegetational characteristics(Ja-
kucs 1972,Wales 1967, 1972),andanimal
communities(Galli et al. 1976, Johnston
1947, Leopold 1933, Patton 1975). Soil
and fire characteristics probably also
differ.
Several factors affect the width of the
patch edge. The angle of the sun plays a
major role, with edges facing equator-
ward typically wider than those facing
poleward(Wales 1972),and those in tem-
perateareas wider than in tropicalareas.
Windalso exerts a majorinfluence,with
the prevailingwind direction duringthe
active or growth period having a wider
edge than other sides. The degree of
species differencebetween the patch and
matrixis significant,too.
The patch edge appears to vary in
width from a few meters to a few tens of
metersin patches at the landscapelevel.
Different groups of organisms respond
differentlyto the environmentallydeter-
mined edge width. For example, in
woodlots, avian and tree communities
appearto differfrom the interioronly in
the outer portion of a forest edge, while
herbs and mosses appear sensitive to
essentially the entire edge width.
Patch shape as a variableis important
in several ways, such as a target for
dispersalor home range suitability;here
we consider patch shape in the context
of the edge concept. A large isodiametric
patch is mostly interior, with a band of
edge in the outer portionof the patch. A
735
rectangularpatch of the same size has
proportionally less patch interior and
more patch edge. Finally, a narrowstrip
patch of the same size may be all edge.
Since communityand populationcharac-
teristics differ between the interior and
the edge, comparing these characteris-
tics with the interior to edge ratio of
patches may be useful in evaluatingthe
importance of patch shape in a
landscape.
Whitmore(1975) noted that plant spe-
cies composition and community struc-
ture varied according to the shape of
openings in tropical rain forests. Stiles
(1979) found sharp differences in wasp
nesting density in the New Jersey Pine
Barrens according to the width of the
habitat.In Idaho rockslides, small mam-
mal density correlated best with the
lengthof the rockslideperimeter(Bunnel
and Johnson 1974). Unpublished data
(Formanand Clay) on mushroomdiver-
sity in old New Jersey two-hectare oak
woodlots indicate a halving of species
diversity and a threshold response in
proceeding from isodiametric through
rectangularto strippatches. Patch width
or shape, therefore, appears to be a
major ecological variable in the
landscape.
Several special cases of shape bear
mention. Ring zones are belts of vegeta-
tion, commonlywithina particularaltitu-
dinal range, which extend around a
mountain, and contain a "hole" with
differentvegetationat a differentaltitude
(Hedberg 1955, MacArthurand Wilson
1967).The interiorto edge ratioindicates
that ring zones are more similarto strip
patches than isodiametricpatches. Lin-
ear patches and dendritic patterns con-
tain special characteristicsand are con-
sideredbelow.
The peninsula,where a narrowportion
projectsfrom a largepatch, is a common
shape, and species diversity commonly
decreases progressively toward the tip.
The reason for this pattern in major
continentalpeninsulasof North America
is hypothesizedto be species extinction
on the peninsula duringthe Pleistocene
and subsequent gradual recolonization
from the continent (Simpson 1964, Tay-
lor and Regal 1978). An alternativeex-
planationbased on the edge effect, that
the peninsularedge has a climate strong-
ly modified by the surroundingwater
leavinglittle if any interiorenvironment,
is well knownto farmerswho must grow
different crops on peninsulas (e.g.,
Squier 1877). Apparentlythe peninsular
effect has not been studied at the land-
scape patch level.
736
PATCHNUMBERS A. SIZE B. SHAPE C. BIOTIC TYPE D. NUMBER
AND CONFIGURATION 0o o
@ 0 @ o
o
So far we have focused on the charac-
teristics of individualpatches. Patches,
however, generally do not exist singly, E. CONFIGURATION
but vary in numbersand in their config- 1) Distance
Apart
Size Shope
Difference
Biotic Type
Difference
Origin
Difference
Difference
urationandjuxtapositionto one another.
Patches exhibitingany of the above de- 0 0- 0 Introduced
scribed patch characteristics may, of
course, vary from zero to many in a 2) Regular Random Aggregated
landscape. In understanding a land- 0000
0 . 0
00oo
00 o
0 o
0o
scape, determiningthe numberof patch-
0
0o ooo
S08 ooooo
introducd
es in each of at least four categories o 0 O
oo
) Disitance SiNegate ShapDendritic Liner Circuiginor
appears essential. How many patches % 0 0
are there of each of the patch origins? .0.
0
0 0
How many of each communitytype are 0
there for each of the patch origins? In in a
Figure 2. Patch characteristics
each category thus formed, what is the landscape.
size distribution of the patches? And
what is the distributionof patch shapes
in each of these? boundaries,drainageditches, and irriga-
Determiningthe numbers in each of tion channels, are narrowand typically
these four categories is not difficult in have only species characteristicof patch
some landscapes. A subsamplecan then edges. Strip corridors are wider bands
be selected for measurementof the spe- containinga patch interior environment
cies, energy, or nutrient component of in which interiorspecies may migrateor
interest,and by simple multiplicationthe live. Stream corridors, which border wa-
statusof the componentin the patches ofter courses and vary in width according
a landscape can be estimated with a to the size of the stream, control water
measure of variability. However, this and mineralnutrientrunoff, minimizing
estimation is inadequate, because the flooding, siltation, and soil fertility loss.
spatial configurationamong the patches Networks are formed by intersectingor
has been ignored. For example, a land- anastomosing corridors and therefore
scape with ten evenly-distributedlarge contain loops. Some overlap among the
patches differs fundamentally in most four basic types exists, such as edge
ecological fluxes from a landscape withspecies moving in all four, or a wide
the ten patches clustered at one end. stream corridor also functioning as a
strip corridor for movement of patch
Various spatial configurations(Figure
2) can be examinedusing standardstatis-interiorspecies.
tical techniques (Chessel 1978, Daget Line corridorsare particularlycharac-
1979, Godron 1971, Kershaw 1973) ap- teristic of landscapes dominatedby hu-
plied to the distribution of patches inman disturbance. They originate in the
each of the categories just described. same ways as patches, e.g., remnanttree
The patches of a category may be ran- lines left between fields from an earlier
dom, regular,or aggregated;or positive forest, paths as spot disturbancelines,
or negative associations among patches and introduced lines as shrub and tree
of differentcategories may be present. plantings for defense, enclosing live-
This provides insightinto both the causestock or decreasingwind (Kellogg 1934,
Rotzien 1963, Seignobos 1978, Van Ei-
of the patches and the potentialfor inter-
patchinteraction.For example, common mern et al. 1964).
nonrandompatternsof patches are seen The plant and animal species of line
in limestone karst topography, in den- corridors generally also characterize
dritic stream basins, along roads and patch edges (Pollardet al. 1974). These
corridors provide habitat and breeding
property lines, or encircling towns. Fi-
sites for species requiringthe surround-
nally, the actualdistance between patch-
es is an importantmeasure of potential ing matrixenvironmentfor protectionor
patch interactions. feeding. Introduced nonnative species
are common in line corridors,especially
the disturbance-causedcorridors.
CORRIDORS The microstructureof the line pro-
There are four types of corridors in vides insight into its potential functions
landscapes: Line corridors, such as (Les Bocages 1976, Lewis 1969, Pollard
paths, roads, hedgerows, property and Relton 1970, Pollard et al. 1974,
BioScience Vol. 31 No. 10
Southwood 1961). Hence, a path line
contains mainly disturbance-resistant
species and has compacted soil, often
with attendanterosion along the line. In
contrast, the hedgerowline of shrubs or
trees, which is higher than the matrix,
cuts wind velocity, shades the adjacent
matrix, and has a high evapotranspira-
tion rate. Irrigationchannels, and often
roads and hedgerows, include adjoining
ditches and embankmentswith consider-
able microhabitatdiversity where am-
phibians,reptiles, and moisture-tolerant
plantsare often favored. Changesin line
corridorsthroughtime are little known.
M. D. Hooper, however, found a linear
correlationbetween hedgerow age and
shrubspecies diversity in managedBrit-
ish hedgerows,with an averageone spe-
cies gained per century (Les Bocages
1976,Pollardet al. 1974).
General characteristics of the wider
strip corridors are reasonably well
known, despite a paucity of direct stud-
ies. The corridor must provide protec-
tive cover for species from naturalpred-
ators, domestic animals, and human
effects lining each side of the corridor.
The outer portions of the strip corridor
have the edge effect, while the central
portion contains the interior environ-
ment required for many patch interior
species (Andersonet al. 1977,Johnsonet
al. 1979).For this reason, the width of a
stripcorridoris critical, since the interior
environmentmust be present and suffi-
ciently wide itself to be used by interior
species.
In contrast to the line and strip corri-
dors, the stream corridor is normally a
dendriticpatternformed by intersecting
narrowfingersupstreamwhich gradually
widen downstream.The streamcorridor
is the most widespread corridor type,
andthe concept has developed fromcon-
siderationsof water and mineralnutrient
flows. This corridorstrongly affects the
erosion rate of the stream banks and
adjoininguplandand the absorptionrate
of water from precipitationand runoff.
These, in turn, control siltationand flood
levels in downstream ecosystems. The
streamcorridoris optimumwhen it dou-
bles as a strip corridorfor the migration
of interior species. Since many species
cannot survive the occasional floods of
the streamlowlandor the wet soils of the
lowland and adjoiningbanks, the corri-
dor must include a strip of interiorenvi-
ronment on well-drained soil atop the
streambank.
A corridor should be continuous for
maximumeffectiveness (Getz et al. 1978,
Schreiber and Graves 1977). In land-
November 1981
scapes with ample human activity, one
type of corridor, such as a road, com-
monly crosses another type, such as a
hedgerow. The degree to which such
crossings are effective barriers to the
migration of different species needs
study.
The corridormay exist as an isolated
unit or it may interconnectpatches in the
landscape. In patches, species become
extirpatedfor many reasons. Following
loss of a species in a patch, a connected
corridorfacilitatesrapidreestablishment
of certain species in the patch. A strip
corridor that links small patches may
enrich those patches with species that
otherwisecould not survive in small iso-
lated patches, because many species
have minimumpatch size requirements
for survival (Galli et al. 1976, Robbins
1980,Terborgh1976).Additionally,cor-
ridors facilitate gene flow across the
landscape.
Networks are particularlywidespread
in landscapes bearingthe heavy imprint
of human activity. Anastomosing line
corridors generally form networks,
though occasionally networks may be
composed of strip corridors. Familiar
examples are the interconnectedhedge-
rows or "bocage" and the networks of
roads and railroads. A few networks
reflect natural conditions, such as the
polygon soils of arctic tundra areas or
the reticulatetrails of large mammalsin
east African savannas.
As isolated units, single corridorsare
consideredto enhance the movement of
species. However, as a series of inter-
connected links and loops, a network
provides a more efficient migratorysys-
tem, since alternativepathways are pre-
sent. This structureis importantfor ani-
mal foraging efficiency, predator
avoidance,and minimizingthe barrieror
isolatingeffect of a local disturbanceor
break in a corridorlink. The frequency
of intersectionsof corridorsand the de-
gree to which such intersections are ex-
panded nodes or patches may play an
important role in migration efficiency.
Some networks, such as paths and
roads, are especially effective for move-
ment of people and domestic animals.
We hypothesize networks to be impor-
tant migrationroutes for native species,
but as yet, the evidence is meager (Pol-
lard et al. 1974).
In short, networks are prominentfea-
tures of most landscapes today. Their
functional importancerests not only in
movement along the links, but in their
impact on the matrix and patches in the
surroundinglandscape.
HABITATIONS
A final major structuralcharacteristic
of manylandscapesis humanhabitation,
including the house with its associated
yard, courtyard,farm buildings and im-
mediate surroundings. Habitations, of
course, are disturbance-caused,partially
or totally eliminatingthe naturalecosys-
tem at that spot. The continued exis-
tence of the habitationdepends on main-
taininga chronic disturbancelevel.
The primary ecological structure of
habitationsis based on the types of or-
ganismsthat have replacedthe naturally
occurring ones. Foremost are people,
who provide the continued disturbance
regime to maintainthe habitationarea.
Most of the plants, in turn, are intro-
duced by people. Some may be native
species, but humansexhibit a propensity
for surroundingthemselves with a di-
verse and exotic species assemblage.
People also generally introduce domes-
tic, rather than native animals into
homes, and both animal and plant pests
are inadvertently introduced. Native
species from the surroundingmatrix or
patches immigrateinto habitationareas,
but their success depends upon the level
of disturbancemaintained.
Distance between habitationsin effect
defines urban, suburban,town, village,
and various rural areas. The greatest
density and diversity of introducedspe-
cies appears, on the average, in subur-
ban areas, and indeed, in all areas with
contiguous homes, the ecosystem is
dominated by humans and introduced
species (Schmid 1975). In rural areas
with isolatedhomes, the entire borderof
the habitation interfaces with patches,
networks, corridors or the matrix, so
that interactionwith other landscapeele-
ments is at a maximum.This interaction
is the primaryecological importanceof
habitationsin rurallandscapes.
DYNAMICSOF THE LANDSCAPE
Ourprimaryobjective in this article is
to lend insight into the ecological struc-
ture of landscapes, particularlypatches
(Figure3). Yet, the structureis ultimate-
ly of importanceas it relates to function.
We have touched on the dynamics of
patchesthemselves. Here we brieflysug-
gest some examples of fluxes between
structuralcomponents of the landscape
(Forman 1981), that is, interactions be-
tween patch and matrix,patch and patch
of the same type, patches of different
types, networkand matrix, and the like.
737
 Interactionsbetween patch and matrix isolatedunits. Networks and streamcor- CONCLUSION
are importantin both directions. Heat ridors retardsurface water and nutrient The structureof a landscapeis primar-
energy carriedby wind from one to the runoff, and subsequent siltation and ily a series of patches surroundedby a
other accelerates the evapotranspiration floodsin downstreamecosystems, and in matrix. The origins of patches differ ac-
rate and desiccates the microenviron- a similarfashion, a networkmodifiesthe
cordingto the disturbanceregime in the
mentfor certainspecies. Similarly,wind flow of air and heat energy over the
patch, disturbancein the matrix, natural
carries moisture, ash, dust, and propa- landscape, which in turn alters evapo- distributionof environmentalresources,
gules back and forth. Fire and other transpirationand the moisture patterns species introductions by people, and
disturbancesstart in one and enter the of the soil. time. These differences in patch origin
other, and many types of animalsforage Finally, habitations, as species determinethe species dynamics and the
from one to another. sources, provide people and nonnative stability and turnover of patches
Corridors and networks facilitate plants and animals. They, in turn, har-themselves.
movementof species frompatchto patch vest species, form corridors and net- Patch area, and secondarily isolation,
in the landscape, but also play a major works, produce various disturbances, have traditionallybeen considered the
role in inhibiting migration of matrix and colonize both the surroundingpatch-
major variables indicating the species
species by subdividing the matrix into es and matrix.
diversityof a patch. We hypothesizethat
species diversity in a landscape patch is
47?
:: .
~ a function of the following patch varia-
~~UP, bles in orderof overall importance:habi-
tat diversity ?+disturbance+ area + age
+ matrix heterogeneity - isolation -
boundarydiscreteness. Patch shape as a
modifierof area is importantto species
diversity and is mediated through the
. . ....."" patch edge or edge effect.
The numbersof patches of each patch
origin,biotic patch type, size, and shape
.... .... .......--
.....
determine in part the landscape struc-
i7 ture. However, the spatial configuration
among the patches present may be just
as importantas the numbers.
Corridorsvary in width and function.
Line corridors,particularlythose result-
::7
f7 ing from humanactivities, are very nar-
row and used primarilyfor movement of
edge species or people. Strip corridors,
for effective movement of species char-
acteristic of the interior of a patch, are
wide enoughto includean interiormicro-
environmentas well as edge effect on
both sides. A special case is the stream
corridor,which also controls water and
nutrientflows across the landscape.
Networks composed of intersecting
corridorsare prominentfeatures of most
landscapes. Networks provide an effi-
cient migratoryroute as well as alter the
flow of nutrients, water, and air across
the landscape.
Figure 3. Portionof an agriculturallandscape in New Jersey. Farmingpracticesfor corn and The concept of repetitive patterns in
beans since 1701 have moldedthis landscape. Limitedsuburbanizationeffects are recent.The
geomorphologyis a level Triassicred shale, on whicha well-drainedsilt loamof the Penn series the structureof landscapes opens up a
predominates.The biotic patch types present are dominated by white, red and black oak host of ecological questions related to
(Quercusalba, Q. borealis,Q. velutina),except in streamcorridorsand wet spots where pinoak, both structure and function, and pro-
red maple,ash and elm (Q. palustris,Acer rubrum,Fraxinus,Ulmus)predominate.Photograph
taken May 29, 1970. A. Spot disturbancepatch (small opening in forest). B. Strip corridor vides a relatively simple frameworkfor
(powerlinecrossing stream corridor).C. Narrowpatch with no forest interior.D. Stripcorridor testing them. It also provides a land
(wooded).E. Tinypatches withno forest interior.F. Peninsula.G. Tinyremnantpatchaffectedby managementtool for helping to deter-
proximityto largerpatch. H. Introducedpatch (golfcourse). I. Introducedline corridor(Platanus mineprioritiesin the land use. Finally, it
plantedalong road).J. Large remnantpatch (well-developedforest interior;patch edge about emphasizesthat no patch stands alone.
twiceas wideto south as north).K. Roadnetwork.L. Dwellingsclustered(village).M.Introduced
patch(cemeteryconifersand grass). N. Environmental resource patch(lowlandtree species on
wet spot). 0. Temporalpatch (area of shrubs and successional trees undergoingrapidchange). ACKNOWLEDGMENTS
P. Widestream corridor(containingboth riverand canal). Q. Narrowstream corridor.R. Matrix
We thank Steward T. A. Pickett and
(corn and bean fields). S. Line corridor(road). T. Habitation(area of farm buildings). U.
Hedgerownetwork(connectingwoods patches). V. Small remnantpatch (containslimitedarea Mark J. McDonnell for significantlyim-
of forest interior). provingthis manuscript,and the Nation-

738 BioScience Vol. 31 No. 10

al Science Foundationfor grantDEB-80- within differentsized forest islands in cen-
04653 in support of a portion of this tral New Jersey. Auk 93: 356-364.
work. Getz, L. L., F. R. Cole, and D. L. Gates.
1978. Interstate roadsides as dispersal
REFERENCESCITED routes for Microtus pennsylvanicus. J.
Anderson, S. H., K. Mann, and H. H. Shu- Mammal. 59: 208-212.
gart, Jr. 1977. The effect of transmission Godron,M. 1966.Applicationde la th6oriede
line corridors on bird populations. Am. l'informationAl'6tude de l'homogdn6it6et
Midl. Nat. 97: 216-221. de la structure de la v6gdtation. Oecol.
Bormann, F. H., and G. E. Likens. 1980. Plant. 2: 187-197.
Pattern and Process in a Forested Ecosys- 1971. Essai sur une approche proba-
tem. Springer-Verlag,New York. biliste de l' cologie des vegetaux. These
Brown, J. H. 1971. Mammalson mountain- d'Etat, Univ. Sci. Tech. Languedoc,Mont-
tops: non-equilibriuminsular biogeogra- pellier, France.
phy. Am. Nat. 105:467-478 Gottfried,B. M. 1979.Smallmammalpopula-
Bunnell, S. D., and D. R. Johnson. 1974. tions in woodlot islands. Am. Midl. Nat.
Physical factors affectingpika density and 102: 105-112.
dispersal. J. Mammal. 55: 866-869. Greig-Smith, P. 1964. Quantitative Plant
Carlquist,S. J. 1974.Island Biology. Colum- Ecology. Butterworths,London.
bia Univ. Press, New York. Grossman, L. 1977. Man-environmentrela-
Chessel, D. 1978.Descriptionnon paramdtri- tionships in anthropologyand geography.
que de la dispersionspatiale des individus Assoc. Am. Geogr. Ann. 67: 126-144.
d'une esp~ce. Biometrie et Ecologie 1: 45- Hedberg, 0. 1955. Vegetation belts of the
135. East African mountains. Sven. Bot. Tidskr.
Daget, P. 1979. La nombre d'espces par 45: 140-202.
unit6d'6chantillonnagede taille croissante. Heinselman, M. L. 1973. Fire in the virgin
La Terre et la Vie 32: 461-470. forests of the Boundary Waters Canoe
Daubenmire, R. 1968. Plant Communities: A Area, Minnesota.Quat. Res. (NY) 3: 329-
Textbook of Plant Synecology. Harper and 382.
Row, New York. Helliwell, D. R. 1976. The effects of size and
Diamond,J. M. 1972. Biogeographicalkinet- isolation on the conservation value of
ics: estimationof relaxationtimes for avi- wooded sites in Britain. J. Biogeogr. 3:
faunas of southwest Pacific islands. Proc. 407-416.
Nat. Acad. Sci. USA 69: 3199-3203. Isard, W. 1975. Introduction to Regional Sci-
Diamond,J. M. and R. M. May. 1976.Island ence. Prentice-Hall,New York.
biogeographyand the design of naturalre- Jakucs, P. 1972. Dynamische Verbindung der
serves. Pages 163-186 in R. M. May, ed. Walder und Rasen. Akad. Kiado, Verlag
Theoretical Ecology. Saunders, Philadel- Ungarischen Akad., Wissenschaften,
phia, PA. Budapest.
Dickinson, R. E. 1970. Regional Ecology: Johnson, W. C., R. K. Schreiber,and R. L.
The Study of Man's Environment. John Burgess. 1979.Diversityof small mammals
Wiley & Sons, Inc., New York. in a powerline right-of-way and adjacent
Elfstrom, B. A. 1976. Tree species diversity forest in east Tennessee. Am. Midl. Nat.
and forest island size on the Piedmont of 101:231-235.
New Jersey. M.S. thesis, Rutgers Univ., Johnston, V. R. 1947. Breeding birds of the
New Brunswick,NJ. forest edge in Illinois. Condor49: 45-53.
Forman, R. T. T., ed. 1979a. Pine Barrens: Kellogg, R. S. 1934. The shelterbeltscheme.
Ecosystem and Landscape. Academic J. Forestry 32: 947-977.
Press, New York. Kershaw, K. A. 1973. Quantitative and Dy-
. 1979b. The Pine Barrens of New namic Plant Ecology. American Elsevier
Jersey: An ecological mosaic. Pages 569- Publishers,Inc., New York.
585 in R. T. T. Forman, ed. Pine Barrens: Leopold, A. 1933. Game Management.
Ecosystem and Landscape. Academic CharlesScribner'sSons, New York.
Press, New York. Les Bocages: Histoire, Ecologie, Economie.
. 1981. Interactions among landscape 1976. Universit6Rennes, Rennes, France.
elements: a core of landscape ecology. In Levin, S. A., and R. T. Paine. 1974. Distur-
Perspectives in Landscape Ecology. Proc. bance, patch formation, and community
Int. Congr.LandscapeEcol., 1981,Veldho- structure. Proc. Nat. Acad. Sci. USA 71:
ven. Pudoc Publ., Wageningen,The Neth- 2744-2747.
erlands,in press. Lewis, T. 1969. The distributionof insects
Forman, R. T. T., A. E. Galli, and C. F. neara low hedgerow.J. Appl. Ecol. 6: 443-
Leck. 1976.Forest size and avian diversity 452.
in New Jersey woodlots with some landuse MacArthur,R. H., and E. O. Wilson. 1967.
implications. Oecologia (Berl.) 26: 1-8. The Theory of Island Biogeography.
Forman, R.T.T., and R. E.J. Boerner. PrincetonUniversityPress, Princeton,NJ.
1981. Fire frequency and the Pine Barrens McHarg, I.L. 1969. Design with Nature.
of New Jersey. Bull. Torrey Bot. Club 108: NaturalHistory Press, Garden City, New
34-50. York.
Galli, A. E., C. F. Leck, and R. T. T. For- Mikesell,M. W. 1968. Landscape.Int. Ency-
man. 1976. Avian distribution patterns clo. Soc. Sci. 8: 575-580.
November 1981
Moore, N. W., and M. D. Hooper. 1975.On
the number of bird species in British
woods. Biol. Conserv. 8: 239-250.
Odum, E. P. 1971. Fundamentals ofEcology.
W. B. SaundersCo., Philadelphia,PA.
Patton, D. R. 1975. A diversity index for
quantifyinghabitat edge. Wildl. Soc. Bull.
394: 171-173.
Peterken,G. F. 1974.A method of assessing
woodlandflorafor conservationusing indi-
cator species. Biol. Conserv. 6: 239-245.
Pickett, S. T. A., and J. N. Thompson. 1978.
Patch dynamics and the design of nature
reserves. Biol. Conserv. 13: 27-37.
Pollard, E., M. D. Hooper, and N. W.
Moore. 1974. Hedges. W. Collins Ltd.,
London.
Pollard,E., and J. Relton. 1970.Hedges V. A
study of small mammals in hedges and
cultivated fields. J. Appl. Ecol. 7: 549-557.
Robbins,C. S. 1980.Effectof forest fragmen-
tationon birdpopulations.Pages 198-212in
R. M. DeGraafand K. E. Evans, compil-
ers, Management of North Central and
NortheasternForests for Nongame Birds.
U.S. Dep. Agric. For. Serv., Gen. Tech.
Rept. NC-51.
Rotzien, C. L. 1963. A cumulativereporton
winter bird populationstudies in eight de-
ciduous shelterbeltsof the Red River Val-
ley, North Dakota. Proc. ND Acad. Sci. 17:
19-23.
Sauer, C. 0. 1963. The morphologyof land-
scape. Pages 315-350 in J. Leighly, ed.
Land and Life: A Selection from the Writ-
ings of Carl Ortura Sauer. University of
CaliforniaPress, Berkeley.
Schmid, J. A. 1975. Urban Vegetation: A
Review and Chicago Case Study. Depart-
ment of Geography,Universityof Chicago,
Chicago, IL.
Schreiber, R. K., and J. H. Graves. 1977.
Powerlinecorridorsas possible barriersto
the movement of small mammals. Am.
Midl. Nat. 97: 504-508.
Seignobos, C. 1978. Les systhmes de d6fense
vdgdtaux pr6-coloniaux. Annales de l'Uni-
versite du Tchad. Serie Lettres, Langues
Vivantes et Sciences Humaines. Numdro
Spdcial.Tchad.
Simberloff,D. S. 1976. Experimentalzooge-
ography of islands: Effects of island size.
Ecology 57: 629-648.
Simpson, G.G. 1964. Species density of
North American recent mammals. Syst.
Zool. 13: 57-73.
Smith, A. T. 1974. The distributionand dis-
persal of pikas: Consequences of insular
population structure. Ecology 55: 1112-
1119.
Southwood, T. R. E. 1961. The number of
species of insect associated with various
trees. J. Anim. Ecol. 30: 1-8.
Squier, E. G. 1877. Peru: Incidents of Travel
and Exploration in the Land of the Incas.
Holt, Rinehart& Winston, New York.
Stiles, E. W. 1979.Animalcommunitiesof the
New Jersey Pine Barrens. Pp. 541-553 in
R.T.T. Forman, ed. Pine Barrens: Ecosys-
tem and Landscape. Acad. Press, NY.
739
Taylor, R. J., and P. J. Regal. 1978. The
peninsulareffect on species diversity and
the biogeographyof Baja California.Am.
Nat. 112:583-593.
Terborgh,J. 1976. Island biogeographyand
conservation:Strategyand limitations.Sci-
ence 193: 1029-1030.
Van Eimern, J., R. Karshon, L. A. Razu-
mova, and G. W. Robertson. 1964. Wind-
breaks and shelterbelts. World Meteorol.
Organiz.Tech. Note No. 59. 188 pages.
Wales,B. A. 1967.Climate,microclimateand
vegetation relationships on northern and
southernforest boundariesin New Jersey.
WilliamL. Hutcheson Mem. For. Bull. 2:
1-60.
1972. Vegetation analysis of north-
ern and southern edges in a mature oak-
hickoryforest. Ecol. Monogr. 42: 451-471.
Whitcomb,R. F. 1977. Island biogeography
and "habitat islands" of eastern forest.
Am. Birds 31: 3-5.
Whitmore,T. C. 1975. TropicalRain Forests
of the Far East. Oxford University Press,
New York.
Whittaker,R. H., ed. 1973. Ordinationand
Classificationof Communities.JunkPubl.,
The Hague.
Willis, E. 0. 1974. Populationsand local ex-
tinctionsof birdson BarroColoradoIsland,
Panama.Ecol. Monogr. 44: 153-169.
Woodwell, G.M., and R. H. Whittaker.
1968.Primaryproductionin terrestrialeco-
systems. Am. Zool. 8: 19-30.
ITURTOX ScienceCo.)
(successortoMacmillan
See new addressand phone
shorter
Although ourname
Making slides
quality
thesame
Employees
Courteous attention
toyourneeds
Offeringslidesreasonably
priced
Minimum ordernotrequired
Payingshipping on$25.00ormore
Automatic 5%discount onpre-payment
Nowatyourservice
You'lllikeus!Tryus!
5000 W. 128thPlace
Alsip,Illinois60658 312/371-5500
740
Informationfor Contributors
* Correspondence: Allcorrespondenceshouldbe directedto BioScience, AmericanIn-
stituteof BiologicalSciences, 1401 WilsonBlvd.,Arlington,VA 22209; 703/527-6776.
* Editorial Policy: The editors welcome manuscriptssummarizingimportantareas of
biologicalinvestigation,writtenfor a broad audience of professionalbiologists and ad-
vanced students. We also publish commentarieson BioScience articles, or on other
articles, or on currentissues in the life sciences. In addition,we publisheditorialsor
short statements of opinion.
* Editorials must not exceed 600 words.
* Articles and Commentaries must not exceed 5000 words, about five BioScience
pages. Keep titles as short as possible, consistent with clarity.The 5000-word count
does not allow for illustrations and tables, so text length must be adjusted to ac-
commodate them. The editors reserve the rightto edit the manuscripts;alterations
will not be made withoutauthorpermission.
Papers are accepted for publicationon the conditionthat they are submittedsolely to
BioScience and thatthey willnot be reprintedor translatedwithoutthe consent of the edi-
tors. As the publisher,the AIBS requiresan assignment of copyrightfromall authors.
Papers willusuallybe publishedin the orderof acceptance. The date of acceptance will
be publishedwiththe article.Aboutsix monthsusuallyelapse between receiptof manu-
scriptand publication.
Authorsmust obtain writtenpermissionto reprintany copyrightedmaterialthat has
been publishedelsewhere, includingtables and figures. Photocopies of the permission
letters must be enclosed withthe manuscriptand creditgiven to the source.
* Referees: Manuscriptswillbe reviewedby scientistscompetentin the field.Authorsare
requested to submitthe names and addresses of four potentialauthoritativereviewers
fromoutside theirown institutionbut in NorthAmerica.
* Preparationof manuscript: Manuscriptsmustconformto the Councilof Biology Edi-
tors Style Manual,4th ed., except for referencestyle (see below).Manuscriptsshould be
neatly typewritten,double-spaced throughout,including references, tables, footnotes,
captionsfor illustrations,etc., on one side onlyof 81/2x 11-inchwhitebond paper.Submit
originalplus two copies; the authorshould retaina copy. A separate titlepage should be
provided,and footnotes,figurecaptions, and tables should be typed on sheets separate
fromthe text. At least one copy must be complete withfigures, tables, and references.
Allweights and measures must be in the metricsystem.
* Abstract: An abstractof not more than 50 words must accompanyarticlesand com-
mentaries;therefore,a summaryshould not be included.
* Illustrations: Illustrations
such as photographs,maps, line drawings,and graphs must
be in "camera-ready" form(i.e., originalart drawnby a commercialartist)and submitted,
unmounted,with the manuscript.Numberfigures consecutively and identifyon the re-
verse side. Photographsmust be glossy, black-and-white, and from4 x 5 to 8 x 10 inches
in size. Drawingslargerthan 8 x 10 are usuallynot acceptable. Letteringon all illustra-
tions must be sufficientlylargeto allowreductionto a double or single columnwidth.Pho-
tomicrographsshould have a scale bar. Figurecaptionsfor illustrationsshould be typed
on separate pages. Colorphotographsfor the cover willalso be considered.
* Footnotes: Footnotesin text should be keptto a minimumand should be indicatedby
consecutive superscriptnumerals.Footnotes in tables are representedby symbols (see
p. 39, CBE Style Manual,4th ed.). "Personal communications" are footnotes and
mustincludename and affiliation of source as wellas monthand year of communication.
* References: "ReferencesCited"includesliterature,publishedor unpublished,whichis
retrievableby readers. Citationsin the text are designated by authorname and year of
publicationin parentheses (Link1928, McNaughtonand Wolf1973) in alphabeticalorder.
Use the firstauthor'sname and "et al."for workshaving more than two authors(Scho-
landeret al. 1950), butlistall authorsin the references.Allworkscited in the text must be
listed in the "ReferencesCited"and vice versa. Use the BIOSISListof Serials forjournal
abbreviations.BioScience referencestyle does notfollowthe CBEStyle Manual,4th ed.;
referto a recent issue of the journal.Some samples are:
Link,G. K. K. 1928. Bacteriain relationto plantdiseases. Pages 590-606 in E. O. Jordan
and I. S. Falk, eds. The Newer Knowledge of Bacteriology and Immunology.Uni-
versity of Chicago Press, Chicago, IL.
McNaughton,S. J., and L. L. Wolf.1973. GeneralEcology. Holt,Rinehartand Winston,
New York.
Scholander, P. F., V. Waiters, R. Hock, and L. Irving.1950. Heat regulationin some
arcticand tropicalanimalsand birds.Biol. Bull. 99: 236-258.
* Reprints: Ordersfor reprintsmay be placed eitherbefore or afterthe issue is printed,
but must be prepaidor accompaniedby an institutional purchaseorder.Allowfourweeks
for deliveryafter publication.Minimumorder:50 copies.
BioScience Vol. 31 No. 10