Hiromi Kobayashi & Unique morphology of the human eye and Shiro Kohshima its adaptive meaning: comparative studies Biological Laboratory, Faculty on external morphology of the primate eye of Bioscience and Biotechnology, Tokyo Institute In order to clarify the morphological uniqueness of the human eye of Technology (c/o Faculty of and to obtain cues to understanding its adaptive significance, we Science), 12-1, O-okayama compared the external morphology of the primate eye by measuring 2-chome, Meguro-ku, Tokyo nearly half of all extant primate species. The results clearly showed 152-8551, Japan. E-mail: exceptional features of the human eye: (1) the exposed white sclera is [email protected] void of any pigmentation, (2) humans possess the largest ratio of exposed sclera in the eye outline, and (3) the eye outline is extraor- Received 30 October 1998 dinarily elongated in the horizontal direction. The close correlation of Revision received the parameters reflecting (2) and (3) with habitat type or body size of 29 January 2001 the species examined suggested that these two features are adapta- and accepted 5 February tions for extending the visual field by eyeball movement, especially in 2001 the horizontal direction. Comparison of eye coloration and facial Keywords: primates, eye coloration around the eye suggested that the dark coloration of morphology, sclera colour, exposed sclera of nonhuman primates is an adaptation to camouflage communication, adaptation, the gaze direction against other individuals and/or predators, and that human evolution, theory of the white sclera of the human eye is an adaptation to enhance the gaze mind. signal. The uniqueness of human eye morphology among primates illustrates the remarkable difference between human and other primates in the ability to communicate using gaze signals. 2001 Academic Press
Journal of Human Evolution (2001) 40, 419–435 doi:10.1006/jhev.2001.0468 Available online at http://www.idealibrary.com on
Introduction the human eye. For example, in humans, the widely exposed white sclera (the white of the Recognizing others’ gaze direction is one of eye) surrounding the darker coloured iris the important cognitive bases for communi- makes it easy for others to discern the gaze cation in humans (Gibson & Pick, 1963; direction and has been said to be a charac- Kendon, 1967). To clarify the biological teristic of humans not found in other pri- basis of this ability, especially in relation to mate species (Morris, 1985). However, this the evolution of social intelligence, has not been examined in detail, partly researchers have experimentally examined because of the difficulty in measuring the the cognitive ability to detect gaze direction soft parts of living animals. of others in nonhuman primates (Gomez, In this study, we measured the external 1991; Itakura & Anderson, 1996; Tomasello eye morphologies of nearly half of all extant et al., 1998). However, little attention has primate species with video camera and been given to external morphology of the computer-aided image analysing techniques eye, although this ability of humans might to clarify the morphological uniqueness of be supported by a unique morphology of the human eye and to understand adaptive Address correspondence to: Hiromi Kobayashi, meanings of external eye morphology in Ph.D., 6–8, Nakanoshima-cho, Fukakusa, Fushimi-ku, Kyoto-city, Kyoto, 612-0049, Japan. Tel.: +81 75 644 primates. The results clearly showed excep- 1402; Fax: +81 75 644 1402. tional features of the human eye in both
0047–2484/01/050419+17$35.00/0 2001 Academic Press 420 . . shape and coloration. In our preceding coidea; 43, Hominoidea; 9) were studied paper (Kobayashi & Kohshima, 1997), we (Table 1). Facial images of 80 species were briefly reported the morphological unique- recorded by video camera at the Japan ness of the human eye and discussed its Monkey Centre. Facial images of eight adaptive meanings. In the present paper we species (Microcebus (1), Loris tardigradus (2), fully analysed the results and examined the Perodicticus potto (1), Tarsius (1), Saguinus following hypotheses on adaptive meanings imperator (1), Pithecia monachus (1), Cacajao of primate eye morphology. rubicundus (1), Cercopithecus hamlyni (1)) We measured width/height ratio of the eye were collected from books (Itani & Uehara, outline (WHR) and an index of exposed 1986; Yoshino, 1994). For humans, facial sclera size in the eye outline (SSI) to analyse images of 244 Japanese, 347 Caucasian and eye shape. These eye-shape parameters 68 Afro-Caribbean adults were studied. 244 closely correlated with habitat type or body Japanese, 280 Caucasian and 2 Afro- size of the species examined. To explain the Caribbean images were recorded by video correlation, we postulated a hypothesis that camera, and 67 Caucasian and 66 Afro- these two features are adaptations for Caribbean images were collected from extending the visual field by eyeball move- books (Ohara, 1970; Gomi, 1994). ment, especially in the horizontal direction. Two parameters were measured for each This hypothesis was examined and sup- species: the width/height ratio of the eye ported by analysing the eye movement of outline (WHR) and an index of exposed video-recorded primates and comparing the sclera size in the eye outline (SSI). Frontal way that gaze direction changes among full-face images without obvious facial species with various body sizes and habitat expression of subjects were recorded by types. video camera. These images were processed To explain the unique coloration of the and analysed on a Macintosh Quadra human eye with its exposed white sclera void 840AV computer using the public domain of any pigmentation, we postulated a NIH Image program. For each image, (A) hypothesis that only coloration of the human the distance between the corners of the eye, eye is adapted to enhance the gaze signal (B) the longest perpendicular line between while eye coloration of other primates is the upper and lower eyelid, (C) width of the adapted to camouflage the gaze direction exposed eyeball, and (D) diameter of the iris against other individuals and/or predators. were measured (Figure 1). WHR means This hypothesis was examined and sup- (A)/(B) and SSI means (C)/(D). Data of ported by analysing relationships among iris weight, crown–rump length and habitat type coloration, sclera coloration and facial col- of primates were collected from books (Itani oration around the eye. Our results sug- & Uehara, 1986; Napier & Napier, 1985) gested that unique features of the human eye since we could not get permission for started to evolve as adaptations to large body physical contact with primates. Walking- size and terrestrial life and were completed height and sitting-height of primates were as a device for communication using gaze measured in the Japan Monkey Centre using signal. marks on the wall of the cages.
Eye-coloration measurements Method Coloration of the exposed sclera (including Eye shape measurements the conjunctiva to be precise), iris and face A total of 874 adult animals (88 species: around the eye was recorded for each of 91 Prosimii; 10, Ceboidea; 26, Cercopithe- species by direct observation of living 421 animals (82 species) and of eyeball speci- To calculate the ratio of scanning per- mens (55 species, 124 animals) kept in the formed only by eyeball movement, move- Japan Monkey Centre (Table 1). The sclera ments of the eyeball and the head scanning colour included in the term ‘‘Pale brown’’ were counted (total observation time: was a paler one than yellow ochre: 10YR 10,037 sec) for 29 individuals of 18 species 6/7·5 of Munsell Colour System (see (Table 1). Figure 9). To calculate the ratio of horizontal scan- Eye coloration of 82 primate species were ning to vertical scanning frequency and classified into 4 types (see Type 1–4, Figure time duration of horizontal and vertical 11) by the differences of colour or contrast scanning were measured (total observation between sclera and iris/face. This classifica- time=12,579 sec) for 40 individuals of 26 tion was carried out by one person observing species (Table 1): arboreal species: Lemur living animals. To check reliability of this catta, Cebus apella, C. albifrons*, Pithecia classification, another person independently pithecia*, Ateles belzebuth*, A. geoffroyi, A. classified the face pictures of 76 primate paniscus, Cercocebus galeritus, Cercopithecus species (see Figure 11). The results agreed cephus*, Colobus angolensis, Presbytis cristata, in 70 species (92%). Disagreement was only P. vetulus, P. francoisi*, Nasalis larvatus, observed between Type 1 and Type 2 in 6 Hylobates lar and H. pileatus; semi-arboreal species (8%). species: Macaca fuscata*, Cercocebus torqua- Eyeball specimens of the Japanese tus, Mandrillus sphinx*, M. leucophaeus*, macaque (1 subject) and crab-eating Cercopithecus ascanius, Presbytis entellus and macaque (2 subjects) were supplied from a Pan troglodytes; terrestrial species: Papio co-operative program of the Primate hamadryas, Erythrocebus patas and Homo Research Institute, Kyoto University, sapiens (*: duration time only) (Table 1). Inuyama, Aichi, Japan. Eyeballs were fixed All these videotape analyses were carried with 4% paraformaldehyde in 0·1 M phos- out by one person. To check reliability of the phate buffer (pH 7·2) at 4"C overnight. The analyses, a second person scored every tissue including the conjunctiva and cornea 0·5 sec randomly sampled 20% of the video- separated from eyeballs was washed several tapes independently. Agreement between times in cold phosphate-buffered saline the persons was 82% on average. The (PBS), dehydrated in an ethanolic series Cohen’s kappa (Bakeman & Gobbman, finishing xylene and embedded in paraffin. 1997) was 0·63 on average. Serial sections with a 4 !m thickness were cut with disposable blades, floated on water Results and discussion and placed on slides. These sections were deparaffinized in xylene, washed in ethanol Unique shape of the human eye and PBS and studied by light microscopy Figure 2(a) shows that human eyes have the (see Figure 10). largest exposed sclera area and show extraordinary horizontal elongation of the Eye movement analysis eye outline among primates. SSI increased To analyse the movement of the eye and the in the following order: Prosimii (prim- head when the animals change the direction itive type)
Table 1A
Eye colour (Figure 9) Eye movement Eye shape Living animals Eyeball (Figure 7— (Figure 7— [Figure 2(a)] (Figure 11) specimens (Figure 5) time) frequency)
Prosimii Microcebus 13 Lemur catta 33 111 L. macaco 222 L. fulvus 22 Varecia variegata 361 Loris tardigradus 25 Perodicticus potto 12 Galago senegalensis 321 Otolemur crassicaudatus 122 Tarsius 12 Ceboidea Callimico goeldii 121 Callithrix jacchus 332 C. argentata 22 C. humeralifer 22 G. geoffroyi 332 C. penicillata 21 Cebuella pygmaea 346 Saguinus midas 122 S. weddelli 11 S. imperator 111 S. labiatus 113 S. mystax 222 S. oedipus 143 Cebus capucinus 333 C. albifrons 1412 C. apella 23 111 C. nigrivittatus 1 Aotus trivirgatus 342 Callicebus moloch 252 Saimiri sciureus 2102 Pithecia pithecia 2211 P. monachus 1 Cacajao rubicundus 1 Alouatta caraya 116 Ateles paniscus 12 111 A. belzebuth 24 1 A. geoffroyi 352222 Lagothrix lagothricha 24 Cercopithecoidea Macaca sylvanus 33 M. silenus 231 M. maurus 11 M. nemestrina 332 M. nigra 11 M. fascicularis 335 M. fuscata 671 1 M. fuscata yakui 332 M. mulatta 41817 423
Table 1B
Eye colour (Figure 9) Eye movement Eye shape Living animals Eyeball (Figure 7— (Figure 7— [Figure 2(a)] (Figure 11) specimens (Figure 5) time) frequency)
M. cyclopis 510 M. sinica 341 M. assamensis 111 M. radiata 291 M. thibetana 34 M. arctoides 12 Cercocebus torquatus 22 222 C. torquatus lunulatus 23 C. atys 111 C. galeritus agilis 111111 C. galeritus chrysogaster 23 C. albigena 1 Papio papio 332 P. anubis 12 50 2 P. cynocephalus 13 P. hamadryas 391111 Mandrillus sphinx 34 1 M. leucophaeus 241 2 Theropithecus gelada 111 Cercopithecus hamlyni 11 C. neglectus 33 C. mitis 25 C. ascanius schmidti 22 C. cephus 22 2 C. ascanius 23 222 C. mona 221 C. petaurista 111 C. aethiops 25 Miopithecus talapoin 22 Allenopithecus nigroviridis 13 Erythrocebus patas 44 222 Colobus angolensis 56 111 C. guereza 22 Presbytis cristata 287444 P. francoisi 241 1 P. pileata 111 P. vetulus 22 111 P. entellus 472111 P. obscura 2 Nasalis larvatus 22 222 Hominoidea Hylobates lar 22 111 H. agilis 112 H. pileatus 22 111 H. syndactylus 22 H. klossii 1 Pongo pygmaeus 231 Pan troglodytes 9132333 P. paniscus 11 Gorilla gorilla 46 Homo sapiens 659 247 2 2 2 424 . .
Relationship between eye shape and visual function Primates are mammals with well developed visual function. Many primate species have (1) cone cells for colour vision, (2) the fovea (a dense concentration of cones in the retina focused on the centre of gaze) for a high resolution visual image, (3) forward-facing Figure 1. (A): the distinace between the corners of the eyes for wide stereoscopic vision by both eye, (B): the longest perpendicular line to (A) between eyes, and (4) a well developed postorbital the upper eyelid and lower eyelid, (C): width of plate behind the eyes. It is possible that the the exposed eyeball, and (D): diameter of the iris were measured. WHR means (A)/(B) and SSI means difference in eye shape parameters among (C)/(D). phylogenic groups has some relationship with these anatomical structures for well developed visual functions. However, since variance (ANOVA) (SSI: F (3,90)=14·68, most primate groups except the prosimians P<0·01, WHR: F (3,90)=32·77, P<0·01). have all these anatomical structures, the In multiple comparison (LSD), the differ- difference in eye shape parameters among ence between phylogenic groups, except phylogenic groups cannot be explained by WHR between Cercopithecoidea and the difference in these visual functions. For Hominoidea, was significant (SSI: MSe= example, some nocturnal prosimian species 0·02, P<0·01, WHR: MSe=0·04, P<0·01)]. are the only primates which lack both cone Even in Hominoidea species, SSI and WHR cell and a fovea in their eyes (Wolin & of human were exceptionally high [Figure Massopust, 1970; Alfieri et al., 1976; 2(b): the difference between phylogenic Webb & Kass, 1976; Debruyn et al., 1980; groups was significant with ANOVA (SSI: F Castenholtz, 1984). Figure 3 shows the re- (4,675)=79·82, P<0·01, WHR: F (4,676)= lationship of eye shape parameters with 33·92, P<0·01). In multiple comparison orbital axis angle measured from the (LSD), the SSI difference between species cranium (Shigehara, 1996) which reflects of Hominoidea, except between orang-utan the ability of stereoscopic vision. In this and gorilla and between orang-utan and figure Prosimians with high values of orbital chimpanzee, was significant (MSe=0·02, axis angle showed low values of SSI and P<0·01). The WHR difference between WHR. However, among simians with more human and others was significant (MSe= forward-facing eyes and developed post- 0·19, P<0·01). We measured both sexes of orbital closure, SSI and WHR spread over three human races: Mongoloid, Caucasian various values and no significant correlation and Afro-Caribbean, and sexual and racial is observed. These facts suggest that the difference were slight in these parameters variation in the eye shape parameters relative to interspecies difference. of each phylogenic group does not reflect It is possible that the difference in these evolutionary trends in these visual functions. eye shape parameters reflect some difference in visual function and/or adaptation to Relationship between body size and SSI some environmental or physiological factor SSI correlates well with various body size such as the habitat and body size of the parameters (weight: r=0·59, P<0·001, species. Thus, we analysed relationships crown–rump length: r=0·59, P<0·001, between the eye shape parameters and these sitting height: r=0·65, P<0·001, walking factors. height: r=0·72, P<0·001). The best 425
Figure 2. Variation of WHR (width/height ratio of the eye outline) and SSI (index of exposed sclera size in the eye outline) among the phylogenic groups of primates (a), and in Hominoidea (b). 426 . .
Figure 3. Orbital axis angle and WHR/SSI. Orbital axis angle is the angle formed when right and left orbital axes meet. Orbital axis is the line between the lowest point of the optic canal and the centre of the orbital width (Shigehara, 1996). Circles, prosimians; squares, simians.
correlation was observed with walking primates, to adjust the images to the central height (Figure 4). This means that species fovea. with larger body size have a larger exposed If we suppose that a larger exposed sclera scleral area. is an adaptation for extending the visual field A larger SSI means a smaller iris relative by eyeball movement, the correlation to the eye outline and probably a greater between SSI and body size can be explained ability for visual field extension by eyeball by the theory of scaling. This is because, as movement; in eyes with a large SSI the small body size becomes larger, visual field exten- iris has a wider space to move within the sion by eyeball movement becomes more open eye outline. In mammalian animals, effective than that by head or body move- only primates have central foveae necessary ment. This is so because as body height for fine vision. Therefore, eyeball and/or becomes greater, the weight of the head and head movement are important, especially for body increases proportionally to the cube of 427
Figure 4. SSI and walking height. body height. In contrast, the force required and counted the movements of head and for movement increases only with the square eyeball when they changed the direction in of body height as it depends on the size of which they were looking. The results the muscle cross-section. Moreover, since showed that the proportion of scanning per- the relative growth of the eyes to body height formed only by eyeball movement was cor- is smaller than that of head and body size, related with SSI (Figure 5, r=0·73, comparative eyeball size becomes smaller in P<0·001). It was exceptionally high in larger animals (Schultz, 1940). Thus, to humans (61#28% of horizontal scan, n=5) save energy when changing the direction of compared with other primates (4·3–24·4%, gaze, a large-sized species would move the mean=10·6%). The highest proportion eyeball more often than a small-sized in nonhuman primates was observed in species, and have a larger exposed scleral chimpanzees (20–35%, n=3), the largest area. Besides, in small species with com- nonhuman species examined. These results paratively large eyeballs in a small skull, agreed with our hypothesis. space for muscles moving the eyeball may be seriously limited. Relationship between habitat type and WHR To examine this hypothesis, we video- The mean value of WHR is greatest in recorded various primates (18 species, 29 terrestrial species, moderate in semi- individuals) eating food by hand in cages, arboreal species and lowest in arboreal 428 . .
Figure 5. SSI and eye movement. EyeMove/TotalMove=frequency of gaze direction change only by eye movement/frequency of all gaze direction change by head movement or/and eye movement.
Figure 6. WHR and habitat type. species [Figure 6: Difference among habitat is adaptive in extending the visual field hori- types was significant with ANOVA and with zontally by eyeball movement, and terres- multiple comparison (ANOVA: F (2,127)= trial life needs more horizontal scanning 24·63, P<0·01, LSD: MSe=0·052, than vertical scanning. P<0·01)]. This result suggests that a hori- To examine this hypothesis, we observed zontally elongated eye outline is adaptive to various primates eating food by hand in terrestrial life in some way. We speculated cages and measured the time and frequency that horizontal elongation of the eye outline of horizontal scanning and vertical scanning. 429
Figure 7. (a) Habitat type and the ratio of horizontal scanning to vertical scanning. (b) Ratio of horizontal scanning to vertical scanning and WHR.
The result shows that the ratio of horizontal adaptation for visual field extension by scanning to vertical scanning is higher in eyeball movement. terrestrial species than in arboreal species ff [Figure 7(a): the di erence among habitat Unique coloration of the human eye types was significant with ANOVA (time: F (2,23)=14·5, P<0·01, frequency: F (2,15)= The colour of exposed sclera. The following 4·53, P<0·05). The difference between four colorations of exposed sclera were terrestrial species and arboreal ones was observed (Figures 8 and 9); (a) in almost all significant in multiple comparison (LSD) nonhuman primates (85 species out of 92 (time: MSe=20·76, P<0·01, frequency: species, or 92%) the exposed part of the MSe=20·05, P<0·01)]. These ratios were sclera is uniformly brown or dark brown, (b) also correlated with WHR [Figure 7(b), Macaca sylvanus and M. nemestrina with a time: r=0·74, P<0·001, frequency: r=0·88, pale brown body colour had sclera coloured P<0·001]. The results support our hypoth- pale brown, (c) Saguinus midas, S. labiatus, esis. These investigations suggest that the Callithrix argentata and Callimico goeldii had shape of the eye outline and relative size of brown sclera with a white part in corner of the exposed scleral area are the result of an the eye, (d) humans were the only primates 430 . .
Figure 8. Three types of sclera coloration in nonhuman primates.
having white sclera without any pigmenta- tion. Microscopic analysis of the eyeball section specimens from the Japanese macaque and crab-eating macaque revealed Figure 10. Eyeball section of crab-eating macaque.
Figure 9. Variation of scleral colour. The shaded portion of the nonhuman primate eyeballs shows the general area where colour was noted. The solid line surrounding the cornea represents the eye outline. 431 that the brown coloration of the exposed for predators to know if the prey has sclera was due to pigmentation in the epi- them in their gaze. If prey animals can thelium cornea, conjunctiva and sclera make it known to the predator that (Figure 10). External observations of other they already know of its presence, their nonhuman primate eyes suggest that their chances of survival may increase (Sherman, dark coloration is also due to similar pig- 1977). mentation. Humans have transparent To examine the gaze camouflage theory, conjunctiva and white sclera without pig- we analysed the relationship between sclera mentation. The inner part of the sclera in colour, iris colour and face colour around nonhuman primates was also white like that the eye. If the dark coloration of exposed of humans. sclera is adaptation for gaze camouflage, the colour of exposed sclera should be similar to Adaptive meaning of dark-coloured sclera. the colour of iris and/or face around the Nonhuman primates have sclera coloured eyes, to make it difficult to detect the pos- brown. As pigmentation costs some energy, ition of iris in the eye outline and/or the eye the dark coloration of the exposed sclera outline in the face. probably has some adaptive function. Brown coloration of the exposed sclera was Relationship between sclera colour, iris colour observed in many other mammal species and face colour too, and the following two hypotheses have Figure 11 shows the relationship between been proposed on the adaptive meanings of sclera colour, iris colour and face colour sclera colour. (1) Anti-glare theory: it was around the eye. 82 primate species observed pointed out that the pigmentation may be an were classified into the following four types anti-glare device because it seemed to be by the difference of colour or contrast absent in nocturnal or crepuscular species between sclera and iris/face. (Duke-Elder, 1985). However, our results on primate species were contrary to this Type (1) faceYscleraYiris (43 species): expectation: nocturnal species (Galago sen- darkness of sclera colour is similar to that egalensis, Tarsius syrichta, Perodicticus potto, of face and iris; both eye outline in the face Nycticebus coucang and Aotus trivirgatus) also and iris position in the eye outline are had coloured sclera and diurnal humans had unclear. no pigmentation. Therefore, our results Type (2) face Figure 12. Difference of gaze stimulus between human and orang-utan. 433 Almost all nonhuman primate species Evolution of unique morphology of the human observed (80 out of 81 species) belonged to eye Type 1 or Type 2 coloration. In these Our results suggest that the unique shape of coloration types, the position of the iris human eye (largest SSI and WHR in pri- in the eye outline was unclear because of mates) is a result of adaptations to extend similarity between sclera colour and the visual field by eyeball movement, iris colour (‘‘Gaze camouflage type’’). In especially in the horizontal direction. In Type 1 coloration (43 species), the position human evolution, the ratio of exposed sclera of the eye outline in the face was also in the eye outline might increase because the unclear because of similarity between visual field extension by eyeball movement sclera colour and face colour around the became more effective as body size eyes. In addition, the ruffed lemur (Varecia increased. And the eye outline might be variegata), the only species that had horizontally elongated because the terrestrial Type 3 coloration, has a very small exposed life of humans needed more horizontal sclera area (SSI=1·08) and almost all the scanning than vertical scanning by eyeball area of its eye outline is occupied by movement. the iris. Thus, the ruffed lemur’s eye also Why have only humans discarded sclera can be seen as a ‘‘gaze camouflage pigmentation? It may be because the neces- type’’. The results thus support the ‘‘gaze sity for gaze camouflage decreased and that camouflage theory’’. of gaze-signal enhancement increased in In contrast, humans were the only species human evolution. The predation risk might that had Type 4 coloration, in which both decrease because of the enlarged body size eye outline in the face and iris position in and the use of tools and fire. Gaze-signal the eye outline were very clear because the enhancement might aid the conspecific colour of the exposed sclera is paler than communication required for increased that of the facial skin and iris (‘‘gaze co-operative and mutualistic behaviours to signalling type’’). The human was the allow group hunting and scavenging. only species with sclera much paler than Co-operative and mutualistic behaviours the facial skin. Because of this coloration, might need refined communication systems, it is very easy to discern the gaze direction such as language, to inform one’s intention in human, in contrast to the gaze- to other members of the group. The human camouflaging eyes of the other primates. eye, with a large scleral area surrounding the Figure 12 shows the contrast between sclera iris and a great ability of eyeball movement, colour and face/iris colour of human and would have provided a chance for a drastic orang-utan (Pongo pygmaeus). In this figure, change from ‘‘the gaze-camouflaging eyes’’ darkness of colours was shown by 256 steps into ‘‘the gaze-signalling eyes’’ through a grey scale number (white=0, black=255, small change in scleral coloration. The SSI Figure 12). Even great apes that have SSI and WHR of human eyes are even greater and WHR near that of humans [Figure than those of gorillas, the largest primate, 2(b)] had ‘‘gaze camouflage eye’’ type with which suggests adaptation for gaze-signal brown sclera, brown facial skin and brown enhancement. iris, in which the eye position and iris pos- ition is unclear. In contrast to the gaze- Eye morphology and gaze-signal camouflaging eyes of orang-utan, the human communication sclera is remarkably paler than the facial skin Baron-Cohen (1995) postulated a neural and the iris, and it is very easy to discern the mechanism (eye direction detector) in gaze direction. human brains specialized to detect others’ 434 . . eye direction, and discussed the possibility species examined suggested that these two that such a mechanism might be related with features are adaptations for extending the evolution of a ‘‘theory of mind’’. In the visual field by eyeball movement, recent years, many studies have been carried especially in the horizontal direction. out to examine the cognitive ability to detect Comparison of eye coloration and facial others’ gaze direction in various nonhuman coloration around the eye suggested that primates. The studies have suggested the the dark coloration of exposed sclera of limited ability to detect others’ gaze direc- nonhuman primates is an adaptation to tion in nonhuman primates (Itakura & camouflage the gaze direction against other Anderson, 1996; Tomasello et al., 1998). individuals and/or predators, and that the However, there seems to be some confusion white sclera of humans is an adaptation to in defining ‘‘gaze direction’’ in these studies. enhance the gaze signal. Since gaze direction can be changed by eyeball movement, by head movement and by body movement, it should be defined by Acknowledgements considering all those factors: eyeball direc- We would like to thank Shigetaka Kodera tion, head direction and body direction. for his interest in our study and for Most of these studies, however, defined allowing us to observe the animals of the ‘‘gaze direction’’ mainly by eyeball direction. Japan Monkey Centre. It is a pleasure Our results (Figure 5) suggest that the con- to acknowledge the hospitality and encour- tribution of eyeball movement to the change agement of the members of JMC. We in gaze direction is extremely high in wish to express our gratitude to Tetsuro humans compared with other primate Matsuzawa, Takashi Kageyama and species. The gaze signalling eye coloration of Nobuo Shigehara for stimulating discussion humans also suggests that the contribution under the co-operative research programme of eyeball direction to the gaze signal is in Primate Research Institute, Kyoto exceptionally high in humans. Therefore, for University, and to Manabu Ogiso, nonhuman primates, head direction and Nobuyuki Saitoh and Teruhiko Hamanaka body direction might be more important to for providing eyeball samples. We are detect others’ gaze direction than eyeball indebted to a number of our colleagues at direction. We should pay more attention to the Tokyo Institute of Technology and head direction and body direction in future the Primate Research Institute, Kyoto analyses of gaze-signals in nonhuman University, especially to Mitsue Nomura, primates. Michael A. Huffman, Sou Kanazawa, Masami Yamaguchi and Kazuhide Summary Hashiya, for their constructive criticism on this paper. Comparative analysis of the external mor- phology of the primate eye revealed excep- tional features of the human eye: (1) the References exposed white sclera is void of any pigmen- tation, (2) humans possess the largest ratio Alfieri, R., Pariente, G. & Sole, P. (1976). Dynamic electroretinography in monochromatic lights and of exposed sclera in the eye outline, and (3) fluorescence electroretinography in lemurs. Doc. the eye outline is extraordinarily elongated Ophthal. Proc. Ser. 10, 169–178. in the horizontal direction. The close corre- Andrew, R. J. (1964). The displays in the Primates. In (J. Buiettner-Janusch, Ed.) Evolutionary and Genetic lation of the parameters reflecting (2) and Biology of the Primates, vol. 2. New York: Academic (3) with habitat type or body size of the Press. 435 Bakeman, R. & Gottman, M. (1997). Observing inter- Kobayashi, H. & Kohshima, S. (1997). Unique mor- action: an introduction to sequential analysis (2nd ed.). phology of the human eye. Nature 387, 767–768. Cambridge: Cambridge University Press. Morris, D. (1985). Body Watching. Oxford: Equinox Baron-Cohen, S. (1995). Mindblindness: An essay on Ltd. autism and theory of mind. Cambridge: MIT Press. Napier, J. R. & Napier, P. H. (1985). The Natural Castenholtz, A. (1984). The eye of Tarsius. In History of the Primates. Cambridge, MA: MIT Press. (C. Niemitz, Ed.) Biology of Tarsiers,pp.303–318. Ohara, K. (1970). One.Tokyo:TsukuiShokan. Stuttgart: Gustav Fischer. Perrett, D. I. & Mistlin, A. J. (1990). Perception of Chance, M. R. A. (1962). An interpretation of some facial characteristics by monkeys. In (W. C. Stebbins ff agonistic postures: the role of ‘‘cut-o ’’ acts and & M. A. Berkley, Eds) Comparative Perception– postures. Symp. Zool. Soc. Lond. 8, 81–89. complex signals,pp.187–215. New York: John Wiley Debruyn, E. J., Wise, V. L. & Casagrande, V. A. & Sons, Inc. (1980). The size and topographic arrangement of Schultz, A. H. (1940). The size of the orbit and of the retinal ganglion cells in the Galago. Vision Res. 20, eye in primates. AJPA Old Series 26, 389–408. 315–327. Duke-Elder, S. S. (1985). The eye in evolution. In Sherman, P. W. (1977). Nepotism and the evolution of (S. S. Duke-Elder, Ed.) System of Ophthalmology,p. alarm call. Science 197, 1246–1253. 453. London: Henry Kimpton. Shigehara, N. (1996). Metrical study of the direction of Gibson, J. J. & Pick, A. D. (1963). Perception of the orbits in primates. Primates Res. 12, 165–178. another person’s looking behaviour. Am. J. Psychol. Tomasello, M., Call, J. & Hare, B. (1998). Five pri- 76, 386–394. mate species follow the visual gaze of conspecifics. Gomez, J. C. (1991). Visual behaviour as a window for Animal Behaviour 55, 1063–1069. reading the mind of others in primates. In (A. Van Hooff, J. A. R. A. M. (1962). Facial expressions in Whiten, Ed.) Natural Theories of Mind,pp.195–207. the higher primates. Symposia of the Zool. Soc. Lond. Oxford: Blackwell. 8, 97–125. Gomi, A. (1994). Americans 1.0 1994 Los Angeles. Webb, S. V. & Kaas, J. H. (1976). The size and Tokyo: Fuga Shobo. distribution of ganglion cells in the retina of the owl Itakura, S. & Anderson, L. R. (1996). Learning to use monkey, Aotus trivirgatus. Vision Res. 16, 1247–1254. experimenter-given cues during an object-choice task Wolin, L. R. & Massopust, L. C. Jr (1970). Mor- by a capuchin monkey. Cahiers De Psychologie phology of the primate retina. In (C. R. Noback & W. Cognitive-Current Psychology of Cognition 15, 103– Montagna, Eds) The Primate Brain,pp.1–27. New 112. York: Appleton Century Crofts. Itani, J. & Uehara, S. (1986). Primates. In (D. W. Yoshino, S. (1994). Animal Face. Tokyo: Nikkei Macdonald, Ed.) The Encyclopaedia of Animals, vol. Saiensu. 3. Tokyo: Heibonsha Limited.