University of Washington Tacoma
UW Tacoma Digital Commons
&(918<9'1.(&8.437 (-4414+38*6).7(.51.3&6<687&3) (.*3(*7
(e Ecology of Individuals: Incidence and
Implications of Individual Specialization
Daniel I. Bolnick
Richard Svanbäck
James A. Fordyce
Louie H. Yang
Jeremy M. Davis
University of Washington Tacoma/*6)&:.79;*)9
See next page for additional authors
4114;8-.7&3)&)).8.43&1;4607&8 -@57).,.8&1(4224378&(42&9;*)9.&7%59'
?.768.(1*.7'649,-884<49+46+6**&3)45*3&((*77'<8-* (-4414+38*6).7(.51.3&6<687&3) (.*3(*7&8"#!&(42&.,.8&14224378-&7
'**3&((*58*)+46.3(197.43.3 &(918<9'1.(&8.437'<&3&98-46.=*)&)2.3.786&8464+"#!&(42&.,.8&1422437
*(422*3)*).8&8.43
413.(0&3.*1 :&3'>(0.(-&6)46)<(*&2*7$&3,49.*&:.7*6*2<917*<&66.3&3)46.78*6&@-*;
?*(414,<4+3).:.)9&173(.)*3(*&3)251.(&8.4374+3).:.)9&1 5*(.&1.=&8.43 SIAS Faculty Publications
-@57).,.8&1(4224378&(42&9;*)9.&7%59'
vol. 161, no. 1 the american naturalist january 2003
The Ecology of Individuals: Incidence and Implications
of Individual Specialization
Daniel I. Bolnick,
1,
*
Richard Svanba¨ck,
2,†
James A. Fordyce,
1
Louie H. Yang,
1
Jeremy M. Davis,
1
C. Darrin Hulsey,
1
and Matthew L. Forister
1
1. Section of Evolution and Ecology, Center for Population
Biology, Storer Hall, University of California, Davis, California
95616;
2. Department of Ecology and Environmental Science, Umea˚
University, SE-901 87 Umea˚, Sweden
Submitted August 6, 2001; Accepted June 11, 2002;
Electronically published December 11, 2002
abstract: Most empirical and theoretical studies of resource use
and population dynamics treat conspecific individuals as ecologically
equivalent. This simplification is only justified if interindividualniche
variation is rare, weak, or has a trivial effect on ecological processes.
This article reviews the incidence, degree, causes, and implications
of individual-level niche variation to challenge these simplifications.
Evidence for individual specialization is available for 93 species dis-
tributed across a broad range of taxonomic groups. Although few
studies have quantified the degree to which individuals are specialized
relative to their population, between-individual variation can some-
times comprise the majority of the population’s niche width. The
degree of individual specialization varies widely among species and
among populations, reflecting a diverse array of physiological, be-
havioral, and ecological mechanisms that can generate intrapopu-
lation variation. Finally, individual specialization has potentially im-
portant ecological, evolutionary, and conservation implications.
Theory suggests that niche variation facilitates frequency-dependent
interactions that can profoundly affect the population’s stability, the
amount of intraspecific competition, fitness-function shapes, and the
population’s capacity to diversify and speciate rapidly. Our collection
of case studies suggests that individual specialization is a widespread
but underappreciated phenomenon that poses many important but
unanswered questions.
Keywords: individual specialization, adaptive variation, niche width,
resource partitioning, frequency dependence, niche variation hy-
pothesis, individual ecology.
* Corresponding author; e-mail: dibolnick@ucdavis.edu.
†
Present address: Department of Limnology, Evolutionary Biology Centre,
Uppsala University, Norbyva¨gen 20, SE-752 36 Uppsala, Sweden.
Am. Nat. 2003. Vol. 161, pp. 1–28. 䉷 2003 by The University of Chicago.
0003-0147/2003/16101-010277$15.00. All rights reserved.
Ecologists have long used niche theory to describe the
ecology of a species as a whole, treating conspecific in-
dividuals as ecologically equivalent. For example, most
models of intraspecific competition, predator-prey dynam-
ics, and food web structure assume that conspecific in-
dividuals are identical (but see Lomnicki 1988; DeAngelis
and Gross 1992). Similarly, the majority of articles on
measuring species’ niche width make no mention of the
fact that individuals of the same species may use different
resources (e.g., Hutchinson 1957; Colwell and Futuyma
1971; Pielou 1972; Abrams 1980; Feinsinger et al. 1981;
Linton et al. 1981). This omission persisted despite a well-
developed literature on niche width variation, originating
with Van Valen’s (1965) niche variation hypothesis. On
the basis of his observations of island and mainland bird
populations, Van Valen proposed that niche expansion in
the absence of interspecific competition was achieved by
increased between-individual variation in resource use.
The role of between-individual niche variation in niche
evolution was further supported by theoretical work by
Roughgarden (1972, 1974). The ensuing flurry of empirical
tests varied between supportive (Fretwell 1969; Rothstein
1973; Grant et al. 1976; Bernstein 1979), inconclusive
(Willson 1969), and negative (Soule´ and Stewart 1970;
Soule´ 1972; Patterson 1983; Diaz 1994). Other empirical
studies downplayed the importance of interindividual diet
variation. Analyzing diet data for five species of Anolis
lizards, Roughgarden (1974) showed that between-
individual variation was generally small, a conclusion that
subsequently received theoretical support (Taper and Case
1985). On reviewing this debate, Grant and Price (1981,
p. 797) concluded that “the status of the adaptive variation
hypothesis hangs in the balance, and it is in danger of
death through neglect as a result of confusion in the em-
pirical tests and theoretical inadequacies.” As predicted,
discussion of individual variation trailed off in the 1980s
but has revived recently with renewed interest in adaptive
radiation and ecological speciation (Mousseau et al. 2000;
Schluter 2000; Halama and Reznick 2001).
Given the contentious history of the niche variation
This content downloaded from 128.095.104.109 on September 14, 2016 14:46:04 PM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
2 The American Naturalist
Figure 1: A schematic diagram of how individuals can subdivide the
population’s niche (thick curve). The total niche width (TNW, black ar-
row) is the variance of total resource use of all individuals (thin curves).
, where WIC (dotted arrow) is the average of in-TNW p WIC ⫹ BIC
dividual niche widths, and BIC (gray arrow) is the variance in mean
resource use among individuals. A, In a population of generalist indi-
viduals, WIC is a large proportion of TNW; B, WIC/TNW is small in a
population of individual specialists. Although the idealized Gaussian
curves used here are a poor description of niche shapes for many real
organisms, they usefully convey the concept of between-individual var-
iation. Real populations are likely to contain both generalized and spe-
cialized individuals, unlike the schematic diagrams shown here. Bolnick
et al. (2002) describe alternative indices that do not rely on assumptions
about resource distribution shapes and that can identify variation in
individual niche widths.
Figure 2: Individual specialization is part of a continuum from where
the within-individual component equals the total niche width (on the
solid diagonal , ; WIC/ ) to where WICWIC p TNW BIC p 0 TNW p 1
is a small proportion of TNW (close to the X-axis). A–D represent the
approximate position on the diagram of hypothetical populations with
(A) high WIC/TNW, (B) medium WIC/TNW, (C) low WIC/TNW, and
(D) high WIC/TNW but small total niche width. Schematic diagrams
represent the niche-use curves of two individuals from each of these four
populations.
hypothesis and its supporting theory, it is perhaps not
surprising that interindividual variation has been ignored
in many ecological studies. Two sources of skepticism seem
particularly common. First, many ecologists believe that
individual specialization is rare and/or weak (Case 1981;
Patterson 1983; Taper and Case 1985; Schoener 1986).
Second, even if interindividual variation is commonplace,
it may have a trivial impact on ecological processes so that
population averages are sufficient for understanding eco-
logical dynamics. The primary goal of this article is to
challenge both views by showing that individual special-
ization is widespread and that it can profoundly affect a
population’s ecological and evolutionary dynamics. In re-
viewing the incidence of interindividual niche variation,
we present a summary of available case studies and discuss
the range of mechanisms that can lead to individual
specialization.
Defining Individual Specialization
Roughgarden (1972, 1974) provided a quantitative frame-
work for thinking about intrapopulation niche variation.
Consider an idealized niche distribution in which indi-
viduals from a population consume prey that can be de-
scribed by a single continuous variable such as prey length
(fig. 1). The total niche width (TNW) of the consumer
population is simply the variance in the size of all captured
prey and can be partitioned into two components. The
within-individual component (WIC) is the average vari-
ance of resources found within individuals’ diets, while
the between-individual component (BIC) is the variation
among individuals, such that . Inter-TNW p WIC ⫹ BIC
individual variation is large when BIC is a large proportion
of TNW, such that WIC/TNW is small (fig. 2).
Intrapopulation niche variation can occur by subdivid-
ing the population’s niche in a number of different ways.
We often expect individuals of different age, sex, or ob-
This content downloaded from 128.095.104.109 on September 14, 2016 14:46:04 PM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Individual Specialization 3
viously distinct morphology to have different niches, as
reflected in Schoener’s (1986, p. 119) statement that “for
the most part, the important between-phenotype variation
in populations occurs between sex and age classes.” Con-
sequently, a researcher might investigate niche variation
in a population by constructing a statistical model testing
the effects of sex, age, and morphology on diet, most likely
dividing age and morphology into discrete age classes and
arbitrary ranges of morphology (e.g., Roughgarden 1974).
If a morphological trait is polymorphic (bimodal), one
might reasonably choose to use categories corresponding
to each morphotype, in which case the ANOVA model
would look like the following:
diet p sex ⫹ age class ⫹ morph ⫹ ,(1)
ii i ii
where the error term
i
is the residual diet variation not
attributed to these three a priori ways of classifying
individuals.
We believe that a description of intrapopulation niche
variation is facilitated when we can communicate the dis-
tinction between variation caused by each effect of the
model. Terms are already available to describe the three
main effects, and “ecological sex dimorphism” (Shine
1989, 1991), “ontogenetic niche shift” (Keast 1977; Polis
1984), and “resource polymorphism” (Wimberger 1994;
Skulason and Smith 1995; Smith and Skulason 1996) have
all previously been reviewed. The goal of this review is to
demonstrate that there can also be important niche var-
iation within the residual error term (among individuals),
which also deserves a unique designation. We therefore
define an “individual specialist” as an individual whose
niche is substantially narrower than its population’s niche
for reasons not attributable to its sex, age, or discrete (a
priori) morphological group. The phrase “individual spe-
cialization” can designate either the overall predominance
of individual specialists in a population or the degree to
which individuals’ diets are restricted relative to their pop-
ulation. It is important to note that these definitions con-
cern the relative width of individual and population niches,
not their absolute values. Consequently, individual spe-
cialization is characterized not by a low WIC per se but
by a low WIC relative to TNW.
Individual specialization is one of many factors con-
tributing to intrapopulation niche variation. Although the
case studies collected here are restricted to examples of
individual specialization, much of our discussion of the
causes and consequences of individual specialization is also
pertinent to other forms of niche variation. However, there
are some good biological reasons to distinguish between
sex- or age-based variation and individual-level niche var-
iation. Ecological differences between males and females
can arise as side effects of sexual selection, breeding be-
havior (Magurran and Garcia 2000), or nutritional or en-
ergetic requirements associated with reproduction (Belov-
sky 1978), mechanisms that are potentially (but not
necessarily) different from those generating individual spe-
cialization. Similarly, age-based niche shifts may arise as
a necessary consequence of body-size changes and devel-
opment so that niche partitioning is an incidental by-
product of ontogeny.
In contrast, our distinction between polymorphism and
individual variation is primarily semantic. We follow Smith
and Skulason (1996, pp. 111–112) in defining a poly-
morphism as “discrete intraspecific morphs,” implying
that the morphological distribution has more than one
mode and that members of the population can generally
(though not necessarily always) be unambiguously as-
signed to a particular group. By taking this definition, we
are ensuring that the word “polymorphism” is not simply
synonymous with the term “variation.”
In reality, individual variation and polymorphism are
ends of a continuum of increasingly discrete variation. This
review focuses on the less discrete end of this continuum,
in which individuals cannot clearly be assigned to distinct
morphotypes because either morphological variation is
continuous or resource use variation is not clearly tied to
morphology. We do so because individual-level variation
has been neglected rather than because it is fundamentally
different from polymorphism. Where examples cited in
this review overlap with Smith and Skulason’s (1996), ei-
ther it reflects our feeling that the case in question is not
composed of discrete morphs (e.g., Werner and Sherry
1986) or we are referring to populations in which the
variation is less discrete than those used for their review.
Examples of the latter include three-spine sticklebacks Gas-
terosteus aculeatus and arctic char Salvelinus alpinus.In
each species, benthic/limnetic variation is continuous in
some populations (Amundsen 1995; Robinson 2000; D. I.
Bolnick, unpublished manuscript) and discrete in others
(Schluter and McPhail 1992; Skulason et al. 1993; Snor-
rason et al. 1994).
Incidence of Individual Specialization
We surveyed the literature for examples of individual spe-
cialization on resources, such as prey taxa, host plants, or
oviposition sites, collecting a list of examples from 93 an-
imal species (table 1). We excluded cases where ecologically
differentiated individuals could not be considered sym-
patric or where diet groups showed significant reproduc-
tive isolation, because such variation cannot be said to
occur within a population. We also omitted cases of
within-colony niche variation in eusocial insects (Heinrich
1976; Rissing 1981; Johnson 1986; Wells and Wells 1986;
This content downloaded from 128.095.104.109 on September 14, 2016 14:46:04 PM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).