Content uploaded by Christopher Paul Heesy
Author content
All content in this area was uploaded by Christopher Paul Heesy on Jun 29, 2018
Content may be subject to copyright.
Seeing in Stereo: The Ecology and Evolution
of Primate Binocular Vision and Stereopsis
CHRISTOPHER P. HEESY
Primates are the most visually adapted order of mammals. There is a rich his-
tory within anthropology of proposed explanations for the adaptive significance
of binocular vision, especially pertaining to primate origins and evolution. Depth
perception and orbit morphology have been hypothesized to be functionally
related to specialized locomotor or feeding behaviors. Many of these arguments
continue to this day. An understanding of specific primate visual adaptations,
including binocular vision, can shed light on these long-term and heated
debates.
Primates exhibit highly derived
neurological processing of spatial
form and contrast, depth and motion
perception, and multiple variants of
color vision. One of the most con-
spicuous primate visual specializa-
tions is the large area of overlap
between the fields of vision of the
two eyes.
1,2
Indeed, the wide binocu-
lar field is functionally related to
many other specializations of the pri-
mate visual system, including high
visual acuity, specializations of visual
pathways in the brain, changes in
the visuomotor system (Box 1), and
elaboration and differentiation of the
visual cortex. Visual-field overlap
also correlates with orbital conver-
gence, that is the degree to which
the bony orbits face in the same direc-
tion (Fig. 1).
1,3–5
The relationship
between these two variables is linear
and roughly isometric in mammals,
with primates plotting at the high
end of both distributions.
6,7
For a
given body size, strepsirhine prima-
tes have somewhat less convergent
orbits than do anthropoids; neverthe-
less, strepsirhine orbits are more
convergent than are those of many
nonprimates.
3–11
This is also true of
the earliest omomyiforms and adapi-
forms.
6,8,11,12
High orbital conver-
gence and wide visual-field overlap
thus appear to be inherited from pri-
mate ancestry.
ADAPTIVE SIGNIFICANCE OF
PRIMATE BINOCULAR VISION:
ACROBATIC ARBOREAL
LOCOMOTION OR NOCTURNAL
VISUAL PREDATION?
For much of the twentieth century,
features of primates, including spe-
cializations of the visual system,
were explained as adaptations to the
occupation of an arboreal niche.
13–16
Collins
14
was the first to suggest that
binocular vision and stereopsis, the
cortical process that mentally recon-
structs a three-dimensional world
using binocular visual information,
were required for the accurate judg-
ment of distance in arboreal locomo-
tion, including leaping between arbo-
real substrates. Depth perception is
presumably required for successful
judgment of distances between sub-
strates, especially during leaping,
where a misjudgment can prove
fatal. Robert Martin later incorpo-
rated elements of these ideas in his
fine-branch niche hypothesis of pri-
mate origins in which visual special-
izations for depth judgment would
prove effective during acrobatic loco-
motion that navigates a network of
fine arboreal supports.
17,18
Cromp-
ton
19
expanded on this idea, suggest-
ing that stereoscopic depth percep-
tion and orbit convergence evolved
in early primates so that they could
accurately judge distance during noc-
turnal vertical clinging and leaping,
a specialized form of high-speed ric-
ochetal arboreal locomotion used by
several taxa of extant primates.
Importantly, Crompton invoked the
demands of manuevering through
branches and other substrates in a
low-light, visually complex environ-
ment to locate target food (fruit and
possibly insects). The common com-
ponent to Collins’, Martin’s, and
Crompton’s proposals is the need for
accurate judgment of distance
among arboreal substrates to avoid a
catastrophic error in distance estima-
tion during arboreal leaping.
The functional link between the vis-
ual demands of arboreal locomotion
and binocular vision has been
criticized, most notably by Cartmill.
3–5
Although Cartmill directed his argu-
ments against Collins and other pro-
ponents of the arboreal hypothesis of
primate origins, and predated the
ARTICLES
Christopher Heesy is an assistant profes-
sor in the Department of Anatomy at
Midwestern University, Glendale, Ari-
zona, where he teaches clinical anatomy
and neuroscience. His primary research
interest is in the functional and evolution-
ary morphology of the primate visual sys-
tem. E-mail: cheesy@midwestern.edu.
Key words: orbit orientation; da Vinci stereopsis;
crypsis; camouflage; activity pattern; primate
origins
V
V
C2009 Wiley-Liss, Inc.
DOI 10.1002/evan.20195
Published online in Wiley InterScience
(www.interscience.wiley.com).
Evolutionary Anthropology 18:2135 (2009)
Box 1. The Primate Retinotectal Pathway and Visual Targeting
There are four main targets of vis-
ual information from the retina. The
main terminus of retinal information
is the pathway to the primary visual
cortex via the lateral geniculate nu-
cleus of the thalamus. The supra-
chiasmatic nucleus of the hypothala-
mus and the pretectum, which are
subcortical targets, respectively influ-
ence or control circadian rhythms
and pupillary light responses. The
retinotectal pathway, which targets
fibers to the superior colliculus in the
midbrain portion of the brainstem, is
a critical component of the system
for planned and coordinated eye
movements.
Comparative studies by Pettigrew
and coworkers have demonstrated
that primates, dermopterans (‘‘flying
lemurs’’), and some megachiropter-
ans (fruit bats) all have, as compared
to most other mammals, a unique
morphology and sensorimotor repre-
sentation in the superior colliculus.
88
The superior colliculus is a compo-
nent of a major brainstem structure
known as the tectum, which integra-
tes visual, auditory, and somatosen-
sory information (Box 1, Fig. 1A).
The superior colliculus of most mam-
mals has a complete map of the visual
field seen by the contralateral eye
(that is, the left superior colliculus
contains information collected by the
right eye, Box 1, Fig. 1A). The supe-
rior colliculi of primates have segre-
gated information so that each side
contains information on the contra-
lateral visual field (Box 1, Fig. 1B).
This arrangement integrates infor-
mation from each eye. While fruit
bats and flying lemurs are similar to
primates in that visual field represen-
tation in the superior colliculus
comes from both eyes, they also
retain some overlap and redundancy
in visual field components that is dis-
similar to the situation in all prima-
tes. Integrated visual field topography
of the superior colliculus is not sim-
ply a function of increasing binocular
visual field overlap. For example,
cats, which have substantial binocu-
lar overlap, retain the typically mam-
malian pattern of contralateral eye
representation in the colliculus.
1,89
Although this point has been
debated, a major function of the supe-
rior colliculus in primates is to redirect
the eyes, especially the zone of maxi-
mal binocular visual field overlap, to
an object of interest.
90
The primate
tectum has precisely coordinated vis-
ual to visuomotor maps related to ex-
traocular eye muscle function.
91
All-
man
1
argues that the lack of redun-
dancy in the visual map of the primate
superior colliculus potentially elimi-
nates ambiguity of targeting, which
otherwise might make primates’ abil-
ity to fixate on an object of interest less
rapid and accurate.
Box 1, Figure 1. This schematic, reproduced from Preuss,
103
is based on the work of Pet-
tigrew and colleagues.
88
Reprinted with kind permission of the author and Springer Sci-
ence and Business Media.
22 Heesy ARTICLES
work of Crompton and Martin, his
objections are equally applicable to all
of these hypotheses. The strongest
argument against a link between
arboreal habits and orbit orientation
is that arboreality is not unique
to primates, and yet other groups
of arboreal mammals, such as rodents,
sloths, and some tree-dwelling hy-
raxes, all have divergent orbits and
panoramic visual fields.
3–5,7,9,20
Loco-
motion in a complex arboreal environ-
ment does not require large binocular
and stereoscopic fields. Many arboreal
taxa with panoramic vision success-
fully locomote in this environment.
3–5
For example, arboreal mammals that
leap across discontinuous substrates,
such as between branches, do not
require convergent orbits and expan-
sive binocular visual fields, as demon-
strated by the fact that nonprimate ar-
boreal leapers, such as the squirrel
Sciurus, successfully traverse discon-
tinuous arboreal substrates despite
having divergent orbits and limited
binocular visual fields.
3–5
Indeed,
many gliding rodents, including most
anomalurids (scaly-tailed flying squir-
rels) and petauristines (‘‘normal’’ fly-
ing squirrels), which successfully tra-
verse distances exceeding 100 m
between trees, have relatively diver-
gent orbits and, by extension, small
binocular visual fields.
Additional criticisms of the connec-
tion between arboreality and orbit
orientation can be based on studies of
the visual information available to
most animals during locomotion, irre-
spective of the size of their binocular
field or orbit orientation. Optic flow
fields, which are generated by the vis-
ual perception of global motion of
objects in the environment during
locomotion, are based on the move-
ments of projected images across the
retina
21
(Fig. 2a). In other words,
movement causes visual information
from the environment to flow or radi-
ate across the retina. These optic flow
fields provide information on locomo-
tor heading, velocity, and time-to-
impact relative to objects or environ-
mental substrates.
21–23
Optic flow
fields are generated by the visual
angles subtended by an object and are
inversely proportional to the distance
from that object to the observer.
22
During locomotion, perception of the
distance to contact with an object is a
function of the ratio of the perceived
distance relative to the rate of change
in visual angle of the object. Stated
more simply, a locomoting animal
perceives objects close to it as larger
and moving more quickly than are
those that are farther away, which it
perceives to be smaller (Fig. 2b). In
the simplified case of moving straight
forward, points in visual space radiate
outward and toward the observer.
The utility of optic flow during
locomotion has been most extensively
studied in humans and birds.
23–25
Although, to the best of my knowl-
edge, optic flow has not been studied
in strepsirhine primates, extrapolat-
ing from the available data on maca-
ques and humans, as well as avians,
supports the hypothesis that arboreal
strepsirhines and, for that matter,
most animals, probably use optic flow
perception to determine heading, ve-
locity, and time-to-impact during
locomotion. It is reasonable to ques-
tion whether the large binocular
fields of primates have evolved to use
optic flow. Although several studies
suggest that binocular cues improve
locomotor performance using optic
flow in humans,
26,27
it is notable that
the birds that have been demonstrated
Figure 1. Relationship between orbit orientation and visual field overlap. A. Panoramic visual fields are composed of monocular visual
fields (lighter shaded regions) that minimally overlap and are associated with small regions of binocular overlap (darker shaded region).
B. Skull of the squirrel Sciurus carolinensis, which has laterally facing orbits and a large panoramic visual field. C. Mammals with substan-
tial binocular visual fields are associated with relatively abbreviated monocular visual fields (lighter shaded regions) compared with the
regions of binocular overlap (darker shaded region). D. Skull of the strepsirhine primate Propithecus verreauxi, which has convergent
(similarly facing) orbits and possibly a large binocular visual field (skulls not to scale). E. The correlation between orbit convergence and
binocular visual field overlap is significant (Spearman’s rho ¼0.832, P<0.01, n ¼27); the confidence intervals of the reduced major axis
slope include isometry. Both variables are presented in degrees. The ellipse denotes the positions of the outliers. Sminthopsis crassicau-
data and Dasyurus hallucata.Key: ~- Artiodactyla, l- Carnivora, - Chiroptera, Lagomorpha, ^-Metatheria, !- Perissodac-
tyla, n- Primates, - Rodentia, &- Scandentia. (Reprinted from Heesy,
9
by permission of S. Karger AG, Basel).
ARTICLES Evolution of Binocular Vision and Stereopsis 23
to use optic flow fields for heading
judgments during flying have small
binocular visual fields, yet obviously
are very successful at locomoting.
28
The avian data certainly suggest that
large binocular fields are not neces-
sary to take advantage of optic flow
information during locomotion.
An alternative to the locomotion
component of the ‘‘arboreal theory’’
has been proposed by Cartmill,
3–5
Allman,
1
and Pettigrew
29–31
to
explain the functional and adaptive
advantages of orbit convergence and
binocular visual field overlap. Cart-
mill
3–5
proposed the ‘‘visual preda-
tion hypothesis’’ to explain an adap-
tively cohesive suite of morphologi-
cal, ecological, and behavioral traits
of primates. The influence of activity
pattern, specifically nocturnality or
light-limited environments, on the
optical system was explicated later by
Allman and Pettigrew.
1,29,30
Although
Cartmill’s hypothesis potentially
explains the adaptive significance of
multiple traits, here I review only
those components related to evolu-
tion of visual system. High orbit con-
vergence and binocular visual field
overlap are hypothesized to be adap-
tations to the nocturnal visual preda-
tory habits of the last common
ancestor of all primates. This hy-
pothesis is based on several key
points: comparative orbit morphol-
ogy among mammals, comparative
optics, and the functional attributes
of stereopsis.
In addition to his observation that
many nonprimate arboreal mammals
have divergent orbits, Cartmill noted
that various mammals possess orbit
convergence similar to that of extant
strepsirhine primates. In a series of
allometric studies that controlled for
relative orbit size, Cartmill demon-
strated that relatively high orbit con-
vergence values are characteristic of
predatory mammals that rely on
vision to target, track, and seize
prey.
3,4
He found that orbit conver-
gence values among these predatory
mammals were similar to those of
small-sized predatory strepsirhine pri-
mates such as some lorisiforms, as
well as the haplorhine Tarsius. Based
on these observations, Cartmill
hypothesized that traits of the primate
visual system could be best explained
as adaptations for a visually directed
predatory ancestral habit.
3–5
Further
consideration of some of the grasping
adaptations and dietary ecology of
small predatory strepsirhines led
Cartmill to suggest that ancestral vis-
ual predation habits were directed to-
ward insects of the forest canopy and
undergrowth.
5
However, Cartmill did
not explicitly state a nocturnal activity
pattern for primate origins.
Allman
1
and Pettigrew
29,30
contrib-
uted to the visual predation hypothesis
by providing a functional explanation
that links orbit convergence to the
alignment of the orbital, optical, and
visual axes, specifically in nocturnal
taxa. The optical axis is defined as the
axis of symmetry through the cornea
and lens, whereas the visual axis is
defined as the line that fits through the
point of fixation, nodal points, and area
centralis (or fovea in taxa that possess
one). The orbital axis is the line of sym-
metry of the orbit. Pettigrew’s work
was first directed toward avians, but
was later expanded to include mam-
mals.
29,30
Theoretical optical effects of
Figure 2. A. In the simplified case of moving straight forward, the points in visual space
flow outward from the center. When followed over time, the length and direction of
each line provide information on direction and velocity. (Reprinted by permission from
Macmillan Publishers Ltd: [Nature] from Cumming,
102
copyright 1998). B. In the pilot’s-eye
view of traveling forward, there is a gradient of increasing flow of visual cues from the ho-
rizon toward the viewer (redrawn and modified from Gibson
21
). [Color figure can be
viewed in the online issue, which is available at www.interscience.wiley.com.]
24 Heesy ARTICLES
visual axis alignment in taxa with di-
vergent and convergent orbits are illus-
trated in Figure 3. In taxa with diver-
gent orbits and optic axes (Fig. 3b),
the visual axes converge on an object
in front of the animal; light traveling
along these axes consequently traver-
ses the most curved portion of the lens
of the eye. High curvature of lenses
will alter the path of light, depending
on the angle of approach of light and
degree of curvature of the lens sur-
face.
32
Passage of light through the
more curved portions of the lens pro-
duces spherical aberration,
32
which
reduces image quality in optical sys-
tems and has a potentially detrimental
effect on visual acuity by decreasing
image focus on the retina.
32–34
The
reduction in image quality is lessened
in animals having a spherical-shaped
lens and, therefore, a uniform angle of
approach. Diurnal animals can avoid
the image aberration and poor visual
acuity caused by spherical aberration
by reducing the diameter of the pupil,
thereby allowing light to pass only
through the center of the lens. The lens
center is optically superior because it
is the least curved and allows light to
pass through this paraxial zone with
less reorientation of the light path than
is caused by light passing through
the peripheral portion of the lens. Noc-
turnal animals, however, need to cap-
ture as much light as possible from the
environment to produce a sufficiently
bright image on the retina and
therefore cannot reduce pupil size to
the same degree as can diurnal
taxa.
1,29,30,35–37
However, by reorient-
ing the eyes and orbits so that the vis-
ual and optic axes are more closely
aligned, light from the object of inter-
est can pass through the center of the
lens (Fig. 3c). This close alignment of
the optic and visual axes allows noc-
turnal animals to view objects or prey
within binocular fields without com-
promising light-gathering capabilities.
The nocturnal visual predation hy-
pothesis is a composite of the com-
parative work conducted by Cartmill
with the optical arguments of Allman
and Pettigrew.
5
The nocturnal visual
predation hypothesis is supported by
comprehensive analyses of extant
comparative data on orbit conver-
gence (Box 2). In addition, phyloge-
netically based character analyses of
activity pattern support another pre-
diction of the nocturnal visual preda-
tion hypothesis, that ancestral prima-
tes were nocturnal (Box 3). On the
other hand, the nocturnal visual pre-
dation hypothesis has been criticized
based on phylogenetic and ecological
arguments. In order to address these
criticisms, it is necessary to distin-
guish those depth cues that are avail-
able to all animals from those func-
tional qualities that are exclusive to
binocular vision.
OPTICAL MECHANISMS FOR
JUDGING DEPTH ARE AVAILABLE
TO ALL MAMMALS
Many researchers have concluded
that primate binocular vision evolved
to allow improved depth perception.
Figure 3. Visual axis reorientation. A. Schematic diagram illustrating the relationship
between orientation of the optic axis and visual axis. B. and C. show these relationships in,
respectively, an animal with divergent orbits and one with convergent orbits. See text for
a detailed discussion of this figure. Please note that refraction and aberration of the path
of light are not detailed here. (Figure provided by C. F. Ross).
ARTICLES Evolution of Binocular Vision and Stereopsis 25
However, multiple cues to perceiving
depth are used by animals with ei-
ther panoramic or binocular visual
fields. Any discussion of the rele-
vance of depth perception to primate
visual system evolution necessitates
distinguishing those depth cues
available to animals irrespective of
visual field form from those depth
and visual benefits available solely to
animals with substantial binocular
visual field overlap.
Box 2. Orbit Convergence and Nocturnal Visual Predation Revisited
One critical prediction of the noc-
turnal visual predation hypothesis is
that convergent orbits occur in
mammals that use visual cues to
detect, track, and capture prey at
night. Correlative data on orbit
morphology and feeding ecology in
a sample of mammals fueled Cart-
mill’s original exposition of the noc-
turnal visual predation hypothe-
sis.
3,4
Two recent studies have revis-
ited this issue using larger and
more diverse sampling across mam-
mals.
Ravosa and Savakova
11
reex-
amined the relationship between
orbit convergence and dietary ecol-
ogy in didelphid marsupials; herpes-
tid, procyonid, and felid carnivorans;
pteropodids (fruit bats); scanden-
tians (tree shrews); dermopterans
(‘‘flying’’ lemurs); plesiadapiforms;
and several extant and fossil pri-
mates. The authors found that when
allometric factors are taken into
account, nocturnal arboreal frugi-
vores such as pteropodids exhibit
relatively low levels of orbital con-
vergence, similar to values found in
tupaiids, dermopterans, herpestids,
and plesiadapiforms. Eocene prima-
tes are similar to nocturnal primate
and felid analogs in possessing rela-
tively convergent orbits.
I also recently studied the ecologi-
cal factors associated with high
orbit convergence in mammals.
9
Data on orbit orientation for over
321 extant taxa of marsupial and
eutherian mammals were combined
with data on activity pattern, degree
of faunivory, and substrate prefer-
ence. Taxa were sorted by consider-
ing activity pattern, degree of fauni-
vory, and locomotor substrate to-
gether. For example, all mammals
that could be categorized as diurnal,
terrestrial, and nonpredatory (no
animal matter in the diet) were
averaged into a single group regard-
less of taxonomic classification. I
found that, in general, mammals
that are nocturnal and cathemeral
(active during night or day) have
significantly more convergent orbits
than do diurnal taxa, both within
and across orders. Faunivorous
eutherians, both nocturnal and di-
urnal, have higher mean orbit con-
vergence than do opportunistically
foraging or nonfaunivorous taxa
(Box 2, Fig. 1). However, substrate
preference, including arboreality, is
not associated with higher orbit
convergence or, by extension,
greater visual field overlap. Strepsir-
hine primates have a range of orbit
convergence values similar to that
of nocturnal or cathemeral preda-
tory nonprimate mammals (Box 2,
Fig. 1).
The Ravosa and Savakova
11
and
Heesy
9
studies differ in species sam-
pling and methodology, but the
results are entirely consistent. Both
studies found that faunivorous
eutherian mammals living in light-
limited visual environments possess
relatively more convergent orbits.
These results entirely support the
predictions of the nocturnal visual
predation hypothesis.
Box 2, Figure 1. Ranked ecological composite categories based on orbit convergence
(nonprimate eutherian mammals). Groups are ranked by mean orbit convergence.
Strepsirhine taxa are added at top for comparison. Top, beginning with ANF, has the high-
est mean convergence; bottom, ADO has the lowest mean convergence. Filled squares
indicate mean bracketed by error bars. Abbreviations for group codes are: ADO, arbo-
real diurnal omnivore; ANF, arboreal nocturnal faunivore; TCF, terrestrial cathemeral fau-
nivorous; ANO, arboreal nocturnal omnivore; ANN, arboreal nocturnal nonpredator; TNO,
terrestrial nocturnal omnivore and arboreal diurnal nonpredator; FNN, aerial or flying noc-
turnal nonpredator, TNV, terrestrial nocturnal, variably faunivorous; TCV, terrestrial cath-
emeral, variably faunivorous; ADF, arboreal diurnal faunivore; TDF, terrestrial diurnal fauni-
vore; TCO, terrestrial cathemeral omnivore; TCN, terrestrial cathemeral nonpredator; TDN,
terrestrial diurnal nonpredator; TNN, terrestrial nocturnal nonpredator; TDV, terrestrial diur-
nal, variably faunivorous; ADV, arboreal diurnal, variably faunivorous. Data from Heesy.
9
26 Heesy ARTICLES
Box 3. The Nocturnal Origin of Primates
Activity pattern, defined as the
time of day that an animal is awake
and active, is considered to be a
critical component of several
hypotheses for the origin of prima-
tes because of the profound influen-
ces of light available to the mor-
phology and function of the visual
system.
36,37
(However, see Ankel-
Simons and Rasmussen
104
for an al-
ternative view on the importance of
activity pattern to evolution of the
primate visual system). Early work
by Martin and Charles-Dominique
suggested that, based on the distri-
bution of activity pattern in some
strepsirhines, the primitive activity
pattern of primates was nocturnal-
ity.
17,92
Recently, however, primates
have been hypothesized to be primi-
tively diurnal on the basis of two
lines of evidence, the distribution of
color-sensitive photoreceptor opsin
genes in several extant strepsir-
rhines,
93,94
and reconstruction of
the activity pattern of the fossil
omomyiform, Teilhardina asiatica.
12
The problems with the applicability
and interpretation of the genetic
data are discussed in detail else-
where.
82,95,96
The possibility of a diurnal ances-
try for primates has been raised
again recently by Ni and co-
workers
12
in their recent description
of the basal omomyiform Teilhar-
dina asiatica from the earliest
Eocene deposits of the Lingcha For-
mation, China. The skull of T. asiat-
ica (IVPP V 12357) is complete
enough to allow estimates of orbit
diameter and skull length, facilitat-
ing the use of orbit-to-skull length
scaling to infer activity pattern.
97
This activity pattern inference, with
the addition of the reconstruction of
faunivory in T. asiatica and a phylo-
genetically based character analysis
of activity pattern, led the authors
to suggest that the last common
ancestor of primates was a diurnal
and visually directed predator. In
addition, if the most primitive omo-
myiform described to date was diur-
nal, it differs from all later omomyi-
forms, which were nocturnal,
95,97
implying that nocturnality is derived
within omomyiformes.
Ross and Martin,
98
as well as
Bloch and Silcox,
99
have suggested
that the extremely small size of Teil-
hardina asiatica precludes the use of
extant scaling prediction equations
for the purpose of reconstructing ac-
tivity pattern because Teilhardina
falls so far below the size range of
extant primates. On the other hand,
two critical qualities of the tree to-
pology employed by Ni and cow-
orkers may have influenced the phy-
logenetically based character analy-
sis of diurnality at the basal primate
node. First, in the tree constructed
by Ni and coworkers, Malagasy strep-
sirhines are not monophyletic, with
Galagoides demidoff, a lorisiform and
Box 3, Figure 1. Re-analysis of activity pattern character optimization on the cladogram
from Ni and colleagues.
12
Key: Black nodes and branches are nocturnal, white nodes
and branches are diurnal, lined nodes and branches are equivocal. Taxa with boxes
at left are those for which character states are either known or inferred; they were
included in the character optimization.
ARTICLES Evolution of Binocular Vision and Stereopsis 27
Artists have long understood and
used several depth cues, although
these are not often cited in relation to
the evolution of primate vision. For
example, interposition, or object order-
ing, is a depth cue whereby an object is
judged to be farther away than another
object that is partly obscuring the first
object. Perspective cues are also famil-
iar to every art student. The most com-
mon example is of lines that appear to
converge in the distance, such as train
tracks, but are assumed by the visual
system to be parallel, thereby generat-
ing the impression of depth.
38,39
We
can additionally generate powerful
depth cues with motion parallax by
moving the head from side to side.
Closer objects appear to move rela-
tively quickly (and opposite to head
motion) when projected across the ret-
ina, whereas more distant objects
appear to move more slowly.
38,39
Accommodation, the process by which
the lens changes shape to focus on
near objects, can provide a strong
depth cue. Also, near-field depth cues
can be generated by vergence eye
movements, which reorient the eyes to
focus on close objects. All of these cues
for judging depth are available to ani-
mals regardless of orbit orientation.
Interposition, perspective, motion par-
allax, vergence eye movement cues,
accommodation, and optic flow are
almost certainly used by arboreal taxa
with divergent orbits to determine
depth during locomotion.
Advantages of Binocularity
In addition to the depth informa-
tion available to animals using these
cues, at least three potential visual
advantages are unique to animals
with large binocular visual fields,
such as primates. These advantages
are enhanced light sensitivity, con-
trast discrimination, and expanded
stereoscopic depth perception. Binoc-
ular visual fields increase the proba-
bility of capturing light within the
region of overlap by a factor of
approximately 1.25–2.
37,40
This can be
especially beneficial to nocturnal taxa
or ones that are otherwise scotopi-
cally (vision in low light) adapted. On
a moonless night, light levels can be
100 million times dimmer than day-
light; also, there is tremendous varia-
tion in light availability from dusk to
full night.
36,37,41
There are examples
of mesopelagic fishes (those dwelling
200–900 m under water) that have
evolved dorsally oriented tubular eyes
and dorsal binocular overlap to
increase visual sensitivity to the
region above their heads, from which
light is comparatively more abun-
dant.
42
Binocularity also improves
contrast discrimination, defined as
the ability to detect luminance differ-
ences in adjacent objects or multiple
parts of the same object. This
improved contrast discrimination is
made possible by physiological sum-
mation of the doubled visual informa-
tion extracted from the similar images
presented to each eye.
40,43
Binocular
contrast helps distinguish unwanted
‘‘noise’’ from useful information by
physiological summation along the
visual pathway.
40,43,44
Simply put, the
ability to detect contrast differences,
especially subtle differences, is sub-
stantially improved within the binoc-
ular visual field as compared to mo-
nocular performance.
STEREOPSIS: THE CORTICAL
COMPUTATION OF
THREE-DIMENSIONAL STRUCTURE
AND DEPTH FROM THE
BINOCULAR FIELD
Basic Mechanisms of Stereopsis
We inhabit a three-dimensional
world, but our retinae can only cap-
ture it in two dimensions.
45
Stereo-
psis is the cortical process that men-
tally reconstructs a three-dimensional
world that has been simplified into
two dimensions by the retinal cap-
ture of light from the environment.
More formally, stereopsis is the com-
putation of object solidity and depth
based on binocular disparity cues.
45–47
Stereoscopic depth perception is a
Microcebus murinus, a lemuriform,
forming a clade to the exclusion of
Lemur catta, a lemuriform, a topol-
ogy that is not supported by molecu-
lar analyses of strepsirrhine relation-
ships.
100
Malagasy strepsirrhines are
the most variable clade of primates
in terms of activity patterns; there
are numerous nocturnal, diurnal,
and cathemeral species. The position
of Lemur catta at the base of the
strepsirrhine clade forces a diurnal
character optimization at the basal
node of this clade. Given the influ-
ence that this node has on character
optimization at the basal primate
node, it is essential to examine topol-
ogies and character distributions
that accurately reflect known strep-
sirhine phylogenetic relationships.
Second, the only outgroup to prima-
tes included in the analysis is ‘‘Scan-
dentia,’’ which was coded by Ni and
coworkers coded as diurnal. How-
ever, tree shrews (the Order Scan-
dentia) include a nocturnal species,
the pen-tailed tree shrew Ptilocercus
lowii. Coding scandentians as diur-
nal implies that diurnality is primi-
tive for the order, an assumption for
which there is no supporting evi-
dence. In addition, at least for the
purposes of activity pattern charac-
ter evolution, it is important to note
that dermopterans, which are the sis-
ter taxon either to scandentians or
primates, are also nocturnal.
101
When activity pattern character data
are optimized onto Ni and co-
workers’ cladogram modified to
include Ptilocercus lowii and der-
mopterans, and constrained to
enforce strepsirhine monophyly,
nocturnality is reconstructed at the
basal primate node (Box 3, Figure 1).
At least three potential
visual advantages are
unique to animals with
large binocular visual
fields: enhanced light
sensitivity, contrast
discrimination, and
expanded stereoscopic
depth perception.
28 Heesy ARTICLES
physiological process founded on
neurons sensitive to binocular dis-
parities that are responsible for the
perception of depth, object solidity,
and binocular fusion. Binocular neu-
rons require input from each eye.
48
Cortical stereoscopic processing inte-
grates or fuses percepts of retinal
images from each eye into a singular
mental visual image.
45–48
Binocular visual field overlap is a
prerequisite for stereoscopic depth
perception because this is the portion
of the visual field from which binocu-
lar parallax cues are collected. Binoc-
ular parallax is generated by the
slightly different views of objects pro-
jected onto each retina (Fig. 4a). Gen-
erally speaking, points that project to
slightly discrepant or noncorrespond-
ing points on each retina provide
cues to relative object location within
the binocular field when compared to
the retinal positions of each eye’s vis-
ual axis. These slight discrepancies,
called binocular disparities, provide
the cues from which stereoscopic
depth is computed (Fig. 4a). In the
schematic case illustrated in Figure 4b,
the angular deviations between the
retinal positions for points F (repre-
senting the visual axis) and N differ
between eyes. Simple subtraction
between the angular value for angle
R and angle L gives the disparity
value for point N. Expansion of the
binocular field, as in primates, leads
to an increase in the number of
points seen by both eyes, facilitating
Figure 4. Basic geometry of binocular disparity. A. In an idealized case, the eyes are both directed toward a single fixation point that
projects onto corresponding positions on each retina. Closer points project onto more lateral (and noncorresponding) parts of the ret-
ina, whereas more distant points project onto more medial noncorresponding retinal positions. These noncorresponding points produce
binocular disparity. (Redrawn from Wurtz and Kandel
59
). B. Binocular disparity is a geometric calculation. In this illustration, both eyes
are looking toward vertically oriented bars. The appearance of these bars present to each eye are illustrated schematically behind the
eyes. Both eyes fixate on a bar, labeled F, and the image of the bar falls onto the fovea (or area centralis for nonhaplorhine mammals,
which are afoveate). These images for F fall onto corresponding points on each retina. Images from a bar located within the binocular
field closer to the eyes, labeled N, fall onto noncorresponding points. The angular distances between the fixated points for F and from
N are illustrated as alpha R and alpha L on the projected images of bars behind the eyes. The difference between these angles is bin-
ocular disparity (disparity ¼angle R angle l). Similar angles of disparity are calculated from other points, such as P, within the same
binocular field (redrawn from Cumming
47
).
ARTICLES Evolution of Binocular Vision and Stereopsis 29
richer stereoscopic perception. The
amount of parallax available to an
animal is partly a function of the dis-
tance between the two eyes: the
greater the interocular distance, the
greater the parallax. Interocular dis-
tance therefore limits the functional
utility of stereopsis to a short range.
Behavioral data show that the maxi-
mum range over which stereopsis
functions is quite limited in amphib-
ians and birds (<1m), and mammals
probably also have a short range over
which stereopsis functions.
49,50
The
earliest true primates were likely ex-
traordinarily small
51
and, for this
reason, probably had a small distance
between the eyes and, therefore, lim-
ited optical parallax and a short dis-
tance over which stereopsis would be
effective. This limitation applies to
the scaling of arboreal leaping rela-
tive to effective stereoscopic range.
Generally speaking, the distance an
animal can leap is probably nega-
tively allometric: bigger animals may
leap absolutely greater distances that
are otherwise shorter relative to body
length or mass. However, interorbital
distance and binocular parallax are
proportional to head size. At the ex-
traordinarily small body sizes of the
earliest primates, effective stereo-
scopic range was undoubtedly sub-
stantially shorter than the maximum
possible leaping distance. This dis-
crepancy between the ranges of ster-
eoscopic depth and leaping distance
contradicts the hypothesis that pri-
mate binocularity evolved to judge
distance for arboreal leaping.
Monocular Occlusion Within
the Stereoscopic Field:
Da Vinci Stereopsis
An additional form of stereoscopic
vision that has not previously been
introduced into discussion of the
evolution of primate binocular vision
nonetheless has possibly played a
vital role. Euclid, Galen, and, most
notably, Leonardo da Vinci com-
mented on the impression of depth
generated by the partial, or monocu-
lar, occlusion of an object.
52
Mono-
cular occlusion or, as it has been
called by some authors, da Vinci
stereopsis, generates depth informa-
tion from the differences in views
presented to each eye by the partial
occlusion of a background object by
a foreground object.
52–55
Da Vinci
stereopsis differs from interposition,
described earlier, because some por-
tion of the occluded object is viewed
by one eye. Figure 5 illustrates a sim-
plified case of the da Vinci stereopsis
phenomenon. When an object that is
relatively close to the viewer is posi-
tioned in front of a background
object, it blocks or occludes a por-
tion of that background to both eyes.
In addition, there are regions of the
background that are occluded to only
one eye, not the other. These par-
tially occluded zones are named half-
occlusions or monocular zones.
52,54
Each monocular zone is specific to
one eye. The right eye views a mo-
nocular zone to the right of the fore-
ground (occluding) object that is
blocked from the view of the left eye,
whereas the left eye views a monocu-
lar zone to the left that is blocked
from the view of the right eye (Fig. 5).
Da Vinci stereopsis produces an
impression of depth because our
brains interpret occlusion as strong
evidence for the existence of an
object.
56,57
The mammalian visual
system is highly attuned to detecting
edges and contours.
56,58,59
Edges or
contours effectively distinguish ob-
jects from the background because
our visual systems assume that an
occluded object must lie in the back-
ground.
56,57
Also, if an occluded
object lies in the background, then
logically an object in the foreground
most be occluding the background
object, even if that object is not nec-
essarily immediately apparent. The
power of object occlusion to generate
depth information is exemplified by
the famous Kanisza triangle illu-
sion
60
(Fig. 6). Our visual system
interprets the Kanisza illusion as a
white-edged triangle that is partially
obscuring a black-edged triangle and
three black filled circles, which we
Figure 5. Da Vinci stereopsis. The nearer object occludes the background surface, which
creates features visible to only one eye. The partially occluded regions are called here
‘‘half-occlusions’’ or monocular zones. The portions of the object and background visible
to both eyes contribute binocular disparity cues for stereoscopic perception, whereas
the monocular regions are free of disparity cues. Stereopsis and monocular zones com-
bine to generate the perception of object ‘‘pop out,’’ which rapidly draws attention to
a target of interest. Basic and da Vinci stereopsis facilitate object detection and object
ordering. (Redrawn from Anderson
54
).
30 Heesy ARTICLES
interpret as lying in the background.
Our brains even go so far as to imag-
ine and construct edges to the white
foreground triangle.
Da Vinci and standard stereopsis
also work simultaneously and con-
gruently to reconstruct depth and
object ordering,
54
which may add to
the speed of depth processing, as
well as enhanced object recognition.
Disparity cues for standard stereo-
psis are available from the fore-
ground object and common back-
ground zones, but not monocular
zones (Fig. 6). The combination of
processing standard stereopsis infor-
mation from the binocular portion of
the visual field while simultaneously
recognizing monocular or stereopsis-
neutral zones can cause the percep-
tion of an object to ‘‘pop out’’ relative
to the background.
61
Nakayama and
Joseph
61
have suggested that stand-
ard stereopsis and monocular occlu-
sion combine to delineate the geome-
try of objects relative to the back-
ground, especially when those
objects are hidden or camouflaged.
Disparity is critical to determine
whether image fragments are part of
a single object or are segregated as
parts of separate surfaces.
62
Stand-
ard stereopsis allows object detection
even when the object is similar in
color and texture to the background.
Da Vinci stereopsis assists in edge
detection, while standard stereopsis
links components of an object into a
percept of a single solid object.
61
Object ‘‘pop out’’ probably rapidly
draws or concentrates visual atten-
tion to that object, allowing the tar-
geting of an object of interest within
a densely populated or otherwise vis-
ually confusing scene.
63
The combi-
nation of standard and da Vinci ster-
eopsis is responsible for immediate
object detection and object ordering.
An introduction to how various
forms of camouflage work is neces-
sary to understand how stereopsis
effectively functions to detect
obscured objects.
HOW STEREOPSIS NEUTRALIZES
CAMOUFLAGE
Animals subjected to intense pre-
dation pressure may evolve multiple
defenses. Examples include camou-
flage, morphological and chemical
defenses, predator signaling, warning
displays, and Mu¨ llerian and Batesian
mimicry, among others.
64
Pertinent
to the current discussion is camou-
flage defense, which, as defined by
Endler,
65
has three types: crypsis,
disruptive coloration, and masquer-
ade. Cryptic prey evolve overall body
coloration, as well as patterning, that
resemble the visual background these
animals tend to inhabit.
65
Animals
with disruptive coloration evolve
body coloration patterns, some of
which can be conspicuous, but none-
theless function to obscure the edges
of their body outlines, making detec-
tion less probable.
65,66
Disruptive col-
oration obscures the edges of an ani-
mal to make the edges of the prey
animal blend into the background.
Masquerade camouflage is used by
insects that are shaped and colored
to resemble less palatable items,
such as sticks or leaves.
65
The effec-
tiveness of crypsis and disruptive col-
oration probably differs based on
environment; disruptive coloration of
animal edges is less dependent on
the background, whereas disruptive
and cryptic coloration away from the
edges, toward the center of the body,
tends to require better matching to
the background.
67
Although the func-
tional properties of crypsis and dis-
ruptive coloration differ, both have
been demonstrated to be effec-
tive.
66,68–70
Julesz first proposed that stereo-
psis evolved to counter camouflage.
71
The crypsis-countering effect of ste-
reopsis is a natural product of the
retinal disparity of viewing three-
dimensional objects in a binocular
field.
71
There is disparity between a
cryptic prey animal, which is nearer
to the viewer, and the background
on which it sits, which is relatively
further away. Therefore, stereopsis
allows the depth ordering of these
objects, and the relative and absolute
distances between the prey and back-
ground can be computed. The func-
tional properties of standard and Da
Vinci stereopsis can work in con-
junction for prey detection. Stereo-
scopic fusion of depth cues from a
cryptic colored animal will break
camouflage by the differential per-
ception of depth from the prey ani-
mal and background. Da Vinci ste-
reopsis is especially suited to the
detection of edges, even when those
edges are obscured through disrup-
tive coloration. Combining stereo-
scopic depth cues and the edge
Figure 6. Kanisza triangle demonstrates the phenomenon of illusory contours. Contours
are mentally constructed even when these are not present in the figure. In addition, the
subjectively constructed triangle acquires the appearance of being whiter than the
background.
ARTICLES Evolution of Binocular Vision and Stereopsis 31
detection properties of Da Vinci ste-
reopsis can generate the perception
of the camouflaged prey animal
‘‘popping out’’ relative to the back-
ground. Expanding the binocular
and stereoscopic field also widens
the zone over which cryptic or dis-
ruptively colored prey can be
detected. In addition, enhanced bin-
ocularity increases the ability to
gauge the distance to cryptic or eva-
sive prey items without unnecessary
head movements, which could alert a
prey animal to the presence of the
predator prior to ambush.
4
PRIMATE BINOCULAR VISION IN
PHYLOGENETIC AND
FUNCTIONAL PERSPECTIVE
The predictions of the nocturnal
visual predation hypothesis are
strongly supported by comparative
data on orbit convergence in mam-
mals (Box 2) and analyses of activity
pattern evolution (Box 3). These data
conform to the theoretical expecta-
tions of nocturnal optics as well as
the functional components of stereo-
psis, both of which would aid in
detecting camouflaged prey items in
a light-limited environment. How-
ever, criticisms of the nocturnal vis-
ual predation hypothesis have been
leveled on multiple grounds. The
most prominent of these criticisms
relate to either the phylogenetic ori-
gin of primates or dietary ecology.
Because these objections have been
linked by some, I will review them
together. I also summarize new ideas
on binocular vision in Box 4.
When Cartmill first proposed the
‘‘visual predation hypothesis,’’ he ex-
plicitly sought to provide an adaptive
explanation for a cohesive suite of
morphological, ecological, and behav-
ioral traits.
3–5
Specifically, he linked
orbit convergence and stereopsis for
prey detection, with grasping extrem-
ities and claw loss for prey capture.
Recent phylogenetic analyses of
newly recovered and described ple-
siadapiforms has reopened the
debate on the morphological origin
of primates.
72–74
These analyses have
led Bloch and colleagues to suggest
that plesiadapiforms are stem pri-
mates. Furthermore, these authors
have suggested that their phyloge-
netic reconstruction can be used to
test the predictions of the nocturnal
visual predation hypothesis, espe-
cially the order of trait evolution.
Their results suggest that grasping
and visual adaptations evolved asyn-
chronously in primates. Specifically,
they found that adaptations for man-
Box 4. New Views on the Evolution of Primate Binocular Vision
Changizi and Shimojo
105
recently
suggested that the ability granted by
a binocular visual field to see
around obscuring objects in a visu-
ally cluttered environment like a
forest has been underappreciated in
primate evolutionary studies. Their
argument is reminiscent of Da Vinci
stereopsis, but the emphasis is on
the potential to look around an
object in the foreground in order to
see visually interesting objects in
the background. This can be easily
demonstrated. If the reader holds
up one thumb about five inches in
front of the eyes (it will not work if
it is much farther away), a portion
of the background is still seen by
each eye despite the obscuring fin-
ger. This creates the illusion that
the thumb is partially transparent
because the background can be seen
by each monocular field. Changizi
and Shimojo
105
call this an X-ray
effect because the viewer can ‘‘see
through’’ an object. Specifically they
hypothesize that ‘‘X-ray vision’’
evolved to cut through the clutter of
leaves in a forested environment.
Changizi and Shimojo ignore one
unresolved allometric element:
Early primates were very small and
most environmental objects, like
leaves, would be so large relative to
their interocular distance (and par-
allax) as to negate early primates’
ability to see around them. Never-
theless, the ‘‘X-ray’’ effect may
indeed potentially be an added ben-
efit to high binocular visual field
overlap, especially in the portion of
visual space that is very close to the
animal.
In a multifaceted synthetic hy-
pothesis for primate brain evolu-
tion, Isbell
106
suggested that preda-
tion on primates by snakes applied
a major selective pressure on spe-
cializations of primate brains,
including multiple components of
the visual pathway, specifically for
the rapid visual detection of snakes.
Visual input to several brain struc-
tures, such as the amygdaloid com-
plex, which is responsive to fearful
stimuli, as well as the superior colli-
culus, which is important for
motion tracking, may be expanded
or specialized in primates specifi-
cally for predator detection. Input
to these structures may covary
across primate groups based on his-
torical exposure to snake predation.
Several parts of Isbell’s hypothesis
are difficult to evaluate at present
because the functions for one key
visual pathway (called koniocellular)
are poorly understood due to the
paucity of data on the functional
properties of this pathway, as well
as its distribution across prima-
tes.
107
More importantly, Isbell’s hy-
pothesis highlights the importance
of sorting salient visual stimuli from
the tremendous flood of sensory
data that are constantly acquired. A
major new emphasis in visual neu-
roscience is the vital role of visual
attention, which can be thought of
as selective processing of some vis-
ual information at the expense of
other sensory information, visual
and otherwise, at the neurophysio-
logical and conscious levels.
108
The
new paradigm is to determine how
only salient features are recognized,
captured, and then processed. Isbell’s
snake-brain covariation hypothesis
illustrates, correctly in my opinion,
the fundamental importance of eco-
logically relevant sensory stimuli on
primate sensory system evolution.
32 Heesy ARTICLES
ual and pedal grasping thought to
characterize primates evolved in ple-
siadapiforms.
74,75
Visual features
evolved only at the primate (that is,
crown clade) node. In addition,
Bloch and coworkers argue that in-
sectivorous dental adaptations are
found neither in plesiadapiforms nor
the early fossil primate groups such
as omomyiforms and adapiforms.
Instead, they argue that a dietary
shift toward increased herbivory is
characteristic of the origin of prima-
tes.
74,76
Although the paleontological evi-
dence for dietary and locomotor
adaptations in plesiadapiforms is
compelling, additional lines of evi-
dence are worth considering. Silcox
and coworkers suggest that the asso-
ciation between visual adaptations
and diet is tenuous because adapi-
forms are often considered frugivo-
rous and omomyiforms are consid-
ered omnivorous or frugivorous.
74
When omomyiforms and adapiforms
are considered as whole groups, this
appears to be the case. However, die-
tary reconstruction in the earliest
omomyiforms is a notably difficult
exercise, with results varying based
on the size surrogate used in the
analysis.
77
Depending on the meth-
odology employed, some of the phy-
logenetically and morphologically
most primitive omomyiforms, Teil-
hardina americana and Steinius ves-
pertinus,
78,79
had either insectivorous
or frugivorous mixed feeding diets.
Other small-bodied omomyiformes,
such as Shoshonius cooperi, Loveina
zephyri, Omomys lloydi, and Utahia
kayi, had undisputed insectivorous
diets.
77
Nevertheless, it is difficult to
conclude, based on the ambiguity in-
herent in some results, that omomyi-
forms were primitively omnivorous
or frugivorous. The phylogenetic and
comparative analyses conducted by
Ni and colleagues,
12
as part of their
description of Teilhardina asiatica,
suggest that not only is it the most
primitive omomyiform known, but
also that it most likely had an insec-
tivorous diet. It is also worth noting
that the cercamoniine adapiform
Donrussellia, which is often consid-
ered to be phylogenetically close to
the origin of all adapiforms, has
been interpreted as having also been
insectivorous.
78,80,81
Phylogenetically
based character reconstruction of
diet also finds faunivory or insecti-
vory at the basal primate node.
82
The
extremely small sizes of the earliest
undisputed primates strongly sug-
gests that an insect protein must
have been a major dietary compo-
nent. It is certainly possible, given
the preceding evidence, that primates
were primitively insectivorous and
later diversified into broader and
more complex diets incorporating
more plant matter.
83
At the very
least, the question of dietary special-
izations in the very earliest primates
must be considered as yet unan-
swered.
An additional point to consider is
the evidence of a contemporaneous
increase in temperature and herbivo-
rous insect diversity (and, presum-
ably, abundance) during the late
Paleocene-Eocene, a period that
roughly coincides with the diversifi-
cation of omomyiform and adapi-
form primates.
84,85
Much of the evi-
dence of increased herbivorous
insect abundance comes from the
western interior of North America as
well as Patagonia, although it is as
yet unclear whether global diversity
and abundance levels of insects were
similar. Although the continent of or-
igin for primates is currently
debated, it appears that the diversifi-
cation of early primates roughly
coincides with a dramatic increase in
insect diversity. The massive increase
in insect availability has also been
associated with the rapid diversifica-
tion of echolocating insectivorous
bats during the late Paleocene-
Eocene.
86,87
Evaluating whether the
timing and area of origin for pri-
mates (and bats) coincides with
increased insect resources seems a
worthwhile line of inquiry.
Orbit convergence and broad stere-
opsis are among the most important
traits that characterize Primates as a
monophyletic group. All of the data
on binocular vision and stereopsis
reviewed here support the functional
and adaptive predictions of the noc-
turnal visual predation hypothesis,
which is certainly applicable to
explanations relating to the origin of
primates. It may be that grasping
precedes the evolution of primate
visual adaptations, but it is difficult
to accept that the mosaic evolution
of visual and postcranial specializa-
tions invalidates the entire nocturnal
visual predation hypothesis.
74
At
best, this calls into question Cart-
mill’s theories linking manual grasp-
ing to visual traits as an adaptively
cohesive suite of morphological traits
that evolved simultaneously. As new
evidence emerges, evolutionary pri-
matologists must, of course, always
be willing to adapt theory. However,
it is also easy to theorize that the
evolution of grasping may have set
the stage for selective pressure for
improved primate vision, especially
for accurate grasping during prey
capture. The most important issue is
one of understanding the selective
factors at work when all adaptive
traits are linked together. The noc-
turnal visual predation hypothesis is
a useful tool to aid in interpreting
the ecological context surrounding
primate origins.
ACKNOWLEDGMENTS
Meg Hall, John Fleagle, Mark Cole-
man, Brandon Wheeler, Jeff Plo-
chocki, and two anonymous reviewers
provided invaluable comments on
earlier drafts of this paper. I am espe-
cially grateful to one reviewer for
pointing out the negative allometry of
It may be that grasping
precedes the evolution
of primate visual
adaptations, but it is
difficult to accept that
the mosaic evolution of
visual and postcranial
specializations
invalidates the entire
nocturnal visual
predation hypothesis.
ARTICLES Evolution of Binocular Vision and Stereopsis 33
leaping relative to stereoscopic range.
Todd Preuss and Callum Ross allowed
the use of figures from their work.
My research, reviewed in part here,
has been supported by the Leakey
Foundation.
REFERENCES
1Allman J. 1977. Evolution of the visual system
in early primates. In: Sprague JM, Epstein AN,
editors. Progress in psychobiology and physio-
logical psychology. New York: Academic Press.
p 1–53.
2Allman J, McGuinness E. 1988. Visual cortex
in primates. In: Steklis HD, Erwin J, editors.
Comparative primate biology, vol. 4. Neuro-
sciences. New York: Alan R. Liss, p 279–326.
3Cartmill M. 1972. Arboreal adaptations and
the origin of the Order Primates. In: Tuttle R,
editor. The functional and evolutionary biology
of primates. Chicago: Aldine de Cruyter p 97–
122.
4Cartmill M. 1974. Rethinking primate origins.
Science 184:436–443.
5Cartmill M. 1992. New views on primate ori-
gins. Evol Anthropol 1:105–111.
6Ross CF. 2000. Into the light: the origin
of Anthropoidea. Ann Rev Anthropol 29:147–
194.
7Heesy CP. 2004. On the relationship between
orbit orientation and binocular visual field over-
lap in mammals. Anat Rec 281A:1104–1110.
8Ross CF. 1995. Allometric and functional
influences on primate orbit orientation and the
origins of the Anthropoidea. J Hum Evol
29:201–227.
9Heesy CP. 2008. Ecomorphology of orbit ori-
entation and the adaptive significance of binoc-
ular vision in primates and other mammals.
Brain Behav Evol 71:54–67.
10 Ravosa MJ, Noble VE, Hylander WL, John-
son KR, Kowalski EM. 2000. Masticatory stress,
orbital orientation and the evolution of the pri-
mate postorbital bar. J Hum Evol 38:667–693.
11 Ravosa MJ, Savakova DG. 2004. Euprimate
origins: the eyes have it. J Hum Evol 46:355–
362.
12 Ni X, Wang Y, Hu Y, Li C. 2004. A eupri-
mate skull from the early Eocene of China. Na-
ture 427:65–68.
13 Jones F. 1917. Arboreal man. New York:
Hafner Publishing.
14 Collins ET. 1921. Changes in the visual
organs correlated with the adoption of arboreal
life and with the assumption of the erect pos-
ture. Trans Ophthal Soc UK 41:10–90.
15 Smith GE. 1924. The evolution of man. Lon-
don: Oxford University Press.
16 Le Gros Clark WE. 1959. The antecedents of
man. Edinburgh: Edinburgh University Press.
17 Martin RD. 1979. Phylogenetic aspects of
prosimian behavior. In: Doyle GA, Martin RD,
editors. The study of prosimian behavior. New
York: Academic Press. p 45–77.
18 Martin RD. 1990. Primate origins and evolu-
tion: a phylogenetic reconstruction. Princeton:
Princeton University Press.
19 Crompton RH. 1995. ‘‘Visual predation,’’
habitat structure, and the ancestral primate
niche. In: Alterman L, Doyle GA, Izard MK, edi-
tors. Creatures of the dark: the nocturnal prosi-
mians. New York: Plenum Press. p 11–30.
20 Heesy CP. 2005. The function of the mam-
malian postorbital bar. J Morphol 264:363–380.
21 Gibson JJ. 1986. The ecological approach to
visual perception. Hillsdale New Jersey: Law-
rence Eribaum.
22 Lee DN. 1980. The optic flow field: the foun-
dation of vision. Philos Trans R Soc London B
290:169–179.
23 Lee DN, Reddish PE. 1981. Plummeting gan-
nets: a paradigm of ecological optics. Nature
293:293–294.
24 Wylie DRW, Bischof WF, Frost BJ. 1998.
Common reference frame for neural coding of
translational and rotational optic flow. Nature
392:278–282.
25 Warren WH Jr, Kay BA, Zosh WD, Duchon
AP, Sahuc S. 2001. Optic flow is used to control
human walking. Nature Neurosci 4:213–216.
26 van den Berg AV, Brenner E. 1994. Why two
eyes are better than one for judgements of
heading. Nature 371:700–702.
27 Gray R, Regan D. 1998. Accuracy of estimat-
ing time to collision using binocular and mo-
nocular information. Vision Res 38:499–512.
28 Martin GR, Katzir G. 1999. Visual fields in
short-toed eagles, Circaetus gallicus (Accipi-
tridae), and the function of binocularity in
birds. Brain Behav Evol 53:55–66.
29 Pettigrew JD. 1978. Comparison of the reti-
notopic organization of the visual wulst in noc-
turnal and diurnal raptors, with a note on the
evolution of frontal vision. In: Cool SJ, Smith
EL, editors. Frontiers of visual science. New
York: Springer Verlag. p 328–335.
30 Pettigrew JD. 1986. The evolution of binocu-
lar vision. In: Pettigrew JD, Sanderson KJ, Lev-
ick WR, editors. Visual neuroscience. London:
Cambridge University Press. p 208–222.
31 Pettigrew JD. 1991. Evolution of binocular
vision. In: Cronly-Dillon JR, Gregory RL, edi-
tors. Vision and visual dysfunction, vol. 2. Evo-
lution of the eye and visual system. Boca
Raton: CRC Press. p 271–283.
32 Charman WN. 1991. The vertebrate dioptric
apparatus. In: Cronly-Dillon JR, Gregory RL,
editors. Vision and visual dysfunction, vol. 2.
Evolution of the eye and visual system. Boca
Raton: CRC Press. p 82–117.
33 Hughes A. 1977. The topography of vision in
mammals of contrasting lifestyle: comparative
optics and retinal organisation. In: Crescitelli F,
editor. The visual system in vertebrates. New
York: Springer-Verlag. p 613–756.
34 Land MF,Nilsson D-E. 2002. Animal eyes.
New York: Oxford University Press.
35 Land MF. 1981. Optics and vision in inverte-
brates. In: Autrum H, editor. Handbook of sen-
sory physiology, VII/6B. New York: Springer-
Verlag. p 471–592.
36 Warrant E. 2004. Vision in the dimmest habi-
tats on earth. J Comp Physiol A 190:765–789.
37 Warrant E. 2008. Nocturnal vision. In: Bas-
baum AI, Kaneko A, Shepherd GM, Westheimer
G, editors. The senses: a comprehensive refer-
ence, vol. 2, Vision II. San Diego: Academic
Press. p 53–86.
38 Rock I. 1984. Perception. New York: Scien-
tific American Library.
39 Kandel ER. 1991. Perception of motion,
depth, and form. In: Kandel ER, Schwartz JH,
Jessell TM, editors. Principles of neural science,
3rd ed. Connecticut: Appleton and Lange.
p 440–466.
40 Pirenne MH. 1943. Binocular and uniocular
threshold of vision. Nature 152:698–699.
41 Martin GR. 1990. Birds by night. San Diego:
Academic Press.
42 Warrant E, Locket N. 2004. Vision in the
deep sea. Biol Rev 79:671–712.
43 Campbell FW, Green DG. 1965. Monocular
versus binocular visual acuity. Nature 208:191–
192.
44 Blake R, Sloane M, Fox R. 1981. Further
developments in binocular summation. Percept
Psychophys 30:266–276.
45 Poggio GF, Poggio T. 1984. The analysis of
stereopsis. Ann Rev Neurosci 7:379–412.
46 Howard IP, Rogers BJ. 1995. Binocular
vision and stereopsis. New York: Oxford Uni-
versity Press.
47 Cumming BG, DeAngelis GC. 2001. The
physiology of stereopsis. Ann Rev Neurosci
24:203–238.
48 Poggio GF. 1984. Processing of stereoscopic
information in primate visual cortex. In: Edel-
man GM, Gall WE, Cowan WM, editors.
Dynamic aspects of neocortical function. New
York: John Wiley. p 613–635.
49 Collett T. 1977. Stereopsis in toads. Nature
267:349–351.
50 McFadden SA. 1994. Binocular depth percep-
tion. In: Davies MNO, Green PR, editors. Percep-
tion and motor control in birds: an ecological
approach. Berlin: Springer-Verlag. p 54–73.
51 Gebo DL. 2004. A shrew-sized origin for pri-
mates. Yearbook Phys Anthropol 47:40–62.
52 Howard IP, Duke PA. 2003. Monocular
transparency generates quantitative depth.
Vision Res 43:2615–2621.
53 Nakayama K, Shimojo S, Silverman G.
1989. Stereoscopic depth: its relation to image
segmentation, grouping, and the recognition of
occluded objects. Perception 18:55–68.
54 Anderson B. 1998. Stereovision: beyond dis-
parity computations. Trends Cogn Sci 2:214–222.
55 Anderson B, Nakayama K. 1994. Toward a
general theory of stereopsis: binocular match-
ing, occluding contours, and fusion. Psychol
Rev 101:414–445.
56 Ramachandran VS. 1990. Perceiving shape
from shading. In: Rock I, editor. The perceptual
world. New York: W.H. Freeman & Co. p 127–
138.
57 Kandel ER, Wurtz RH. 2000. Constructing
the visual image. In: Kandel ER, Schwartz JH,
Jessell TM, editors. Principles of neural
science, 4th ed. New York: McGraw-Hill. p 492–
506.
58 Hubel DH, Wiesel TN. 1959. Receptive fields
of single neurones in the cat’s striate cortex.
J Physiol 148:574–591.
59 Wurtz RH, Kandel ER. 2000. Central visual
pathways. In: Kandel ER, Schwartz JH, Jessell
TM, editors. Principles of neural science, 4th
Ed. New York: McGraw-Hill. p 523–547.
60 Kanizsa G. 1955. Margini quasi-percettivi in
campi con stimolazione omogenea. Rivista di
Psicologia 49:7–30.
61 Nakayama K, Joseph J. 1998. Attention, pat-
tern recognition, and pop-out in visual search.
In: Parasuraman R, editor. The attentive brain.
Cambridge: MIT Press. p 279–298.
62 Nakayama K. 1996. Binocular visual surface
perception. Proc Nat Acad Sci USA 93:634–639.
63 He Z, Nakayama K. 1995. Visual attention
to surfaces in three-dimensional space. Proc
Natl Acad Sci USA 92:11155–11159.
64 Ruxton G, Sherratt T, Speed M. 2004. Avoid-
ing attack: the evolutionary ecology of crypsis,
warning signals and mimicry. Oxford University
Press.
65 Endler J. 2006. Disruptive and cryptic colo-
ration. Proc R Soc London B273:2425–2426.
34 Heesy ARTICLES
66 Stevens M, Cuthill IC. 2006. Disruptive colora-
tion, crypsis and edge detection in earl y visual
processing. Proc R Soc L ondon B273:21 41–2147.
67 Schaefer H, Stobbe N. 2006. Disruptive colo-
ration provides camouflage independent of
background matching. Proc R Soc London
B273:2427–2432.
68 Merilaita S, Lind J. 2005. Background-
matching and disruptive coloration, and the
evolution of cryptic coloration. Proc R Soc Lon-
don B272:665–670.
69 Merilaita S, Lyytinen A, Mappes J. 2001.
Selection for cryptic coloration in a visually
heterogeneous habitat. Proc R Soc London
B268:1925–1929.
70 Stuart-Fox D, Moussalli A, Whiting MJ.
2008. Predator–specific camouflage in chame-
leons. Biol Lett 4:326–329
71 Julesz B. 1971. Foundations of Cyclopean per-
ception. Chicago: University of Chicago Press.
72 Bloch JI, Boyer DM. 2002. Grasping primate
origins. Science 298:1606–1610.
73 Bloch JI, Silcox MT, Boyer DM, Sargis EJ.
2007. New Paleocene skeletons and the relation-
ship of plesiadapiforms to crown-clade prima-
tes. Proc Natl Acad Sci USA 104:1159–1164.
74 Silcox M, Sargis E, Block J, Boyer D. 2007.
Primate origins and supraordinal relationships:
morphological evidence. In: Henke W, Tatter-
sall I, editors. Handbook of paleoanthropology,
vol 2: Primate evolution and human origins:
Springer Verlag. p 831–859.
75 Kirk EC, Lemelin P, Hamrick MW, Boyer
DM, Bloch JI. 2008. Intrinsic hand proportions
of euarchontans and other mammals: implica-
tions for the locomotor behavior of plesiadapi-
forms. J Hum Evol 55:278–299.
76 Szalay FS. 1968. The beginnings of prima-
tes. Evol 22:19–36.
77 Strait SG. 2001. Dietary reconstruction of
small-bodied omomyoid primates. J Vert Paleo
21:322–334.
78 Rose KD, Bown TM. 1991. Additional fossil
evidence on the differentiation of the earliest
euprimates. Proc Natl Acad Sci USA 88:98–101.
79 Rose KD. 1995. The earliest primates. Evol
Anthropol 3:159–173.
80 Covert HH. 1986. Biology of early Cenozoic
primates. In: Swindler DR, Erwin J, editors.
Comparative primate biology, vol. 1: System-
atics, evolution, and anatomy. New York: Alan
R. Liss p 335–359.
81 Gebo D. 2002. Adapiformes: phylogeny and
adaptation. In: Hartwig WC, editor. The pri-
mate fossil record. New York: Cambridge Uni-
versity Press. p 21–43.
82 Heesy CP, Ross CF. 2004. Mosaic evolution
of activity pattern, diet, and color vision in hap-
lorhine primates. In: Ross CF, Kay RF, editors.
Anthropoid origins: New visions. New York:
Kluwer Academic/Plenum Publishers. p 665–
698.
83 Gilbert C. 2005. Dietary ecospace and the di-
versity of Euprimates during the Early and Mid-
dle Eocene. Am J Phys Anthropol 126:237–249.
84 Wilf P, Labandeira CC. 1999. Response of
plant-insect associations to Paleocene-Eocene
warming. Science 284:2153–2156.
85 Currano ED, Wilf P, Wing SL, Labandeira
CC, Lovelock EC, Royer DL. 2008. Sharply
increased insect herbivory during the Paleo-
cene-Eocene thermal maximum. Proc Natl Acad
Sci USA 105:1960–1964.
86 Eick G, Jacobs D, Matthee C. 2005. A nu-
clear DNA phylogenetic perspective on the evo-
lution of echolocation and historical biogeogra-
phy of extant bats (Chiroptera). Mol Biol Evol
22:1869–1886.
87 Teeling E, Springer M, Madsen O, Bates P,
O’Brien S, Murphy W. 2005. A molecular phy-
logeny for bats illuminates biogeography and
the fossil record. Science 307:580–584.
88 Pettigrew JD, Jamieson BGM, Robson SK,
Hall LS, McAnally KL, Cooper HM. 1989. Phylo-
genetic relations between microbats, megabats
and primates (Mammalia: Chiroptera and Pri-
mates). Phils Trans R Soc London B325:489–
559.
89 Lane R, Kaas J, Allman J. 1974. Visuotopic
organization of the superior colliculus in nor-
mal and Siamese cats. Brain Res 70:413–430.
90 Grantyn AA, Moschovakis AK. 2004. Struc-
ture-function relationships in the superior colli-
culus of higher mammals. In: Hall WC,
Moschovakis A, editors. The superior colliculus:
new approaches for studying sensorimotor inte-
gration. New York: CRC Press. p 107–145.
91 Schiller P, Stryker M. 1972. Single-unit re-
cording and stimulation in superior colliculus
of the alert rhesus monkey. J Neurophysiol
35:915–924.
92 Charles-Dominique P, Martin RD. 1970.
Evolution of lorises and lemurs. Nature
227:257–260.
93 Tan Y, Li W-H. 1999. Trichromatic vision in
prosimians. Nature 402:36.
94 Tan Y, Yoder AD, Yamashita N, Li W-H.
2005. Evidence from opsin genes rejects noc-
turnality in ancestral primates. Proc Natl Acad
Sci USA 102:14712–14716.
95 Heesy CP, Ross CF. 2001. Evolution of activ-
ity patterns and chromatic vision in primates:
morphometrics, genetics and cladistics. J Hum
Evol 40:111–149.
96 Perry GH, Martin RD, Verrelli BC. 2007. Sig-
natures of functional constraint at aye-aye
opsin genes: the potential of adaptive color
vision in a nocturnal primate. Mol Biol Evol
24:1963–1970.
97 Kay RF, Cartmill M. 1997. Cranial morphol-
ogy and adaptations of Palaechthon nacimienti
and other Paromomyidae (Plesiadapoidea, ?
Primates), with a description of a new genus
and species. J Hum Evol 6:19–35.
98 Ross CF, Martin RD. 2007. The role of vision
in the origin and evolution of primates. In: Pre-
uss TM, Kaas J, editors. Evolution of nervous
systems. vol. 4: The evolution of primate nerv-
ous systems. Oxford: Elsevier. p 59–78.
99 Bloch J, Silcox M. 2006. Cranial anatomy of
the Paleocene plesiadapiform Carpolestes simp-
soni (Mammalia, Primates) using ultra high-re-
solution X-ray computed tomography, and the
relationships of plesiadapiforms to Euprimates.
J Hum Evol 50:1–35.
100 Karanth KP, Delefosse T, Rakotosamima-
nana B, Parsons TJ, Yoder AD. 2005. Ancient
DNA from giant extinct lemurs confirms single
origin of Malagasy primates. Proc Natl Acad
Sci USA 102:5090–5095.
101 Janecka JE, Miller W, Pringle TH, Wiens F,
Zitzmann A, Helgen KM, Springer MS, Murphy
WJ. 2007. Molecular and genomic data identify
the closest living relative of primates. Science
318:792–794.
102 Cumming B. 1998. A sense of direction:
going with the flow. Nat Neurosci 1:6–8.
103 Preuss TM. 2007. Evolutionary specializa-
tions of primate brain systems. In: Ravosa M,
Dagosto M, editors. Primate origins: adapta-
tions and evolution. New York: Springer Scien-
ceþBusiness Media. p 625–675.
104 Ankel-Simons F, Rasmussen DT. 2008.
Diurnality, nocturnality, and the evolution of
primate visual systems. Yearbook Phys Anthro-
pol 51:100–117.
105 Changizi MA, Shimojo S. 2008. ‘‘X-ray
vision’’ and the evolution of forward-facing
eyes. J Theor Biol 254:756–767.
106 Isbell LA. 2006. Snakes as agents of evolu-
tionary change in primate brains. J Hum Evol.
51:1–35.
107 Casagrande VA, Khaytin I, Boyd J. 2006.
The evolution of parallel visual pathways in the
brains of primates. In: Preuss TM, Kaas J, edi-
tors. Evolution of nervous systems, vol. 4: The
evolution of primate nervous systems. Oxford:
Elsevier p 87–108.
108 McAlonan K, Cavanaugh J, Wurtz RH.
2008. Guarding the gateway to cortex with
attention in visual thalamus. Nature 456:391–
395.
V
V
C2009 Wiley-Liss, Inc.
ARTICLES Evolution of Binocular Vision and Stereopsis 35
