Hand preferences and other lateral biases in rehabilitated orang-utans,Pongo pygmaeus pygmaeus

Anim. Behav., 1996, 51, 13–25

Hand preferences and other lateral biases in rehabilitated orang-utans, Pongo pygmaeus pygmaeus L. J. ROGERS* & G. KAPLAN† *Department of Physiology, University of New England †Department of Aboriginal and Multicultural Studies, University of New England (Received 7 March 1994; initial acceptance 18 May 1994; final acceptance 3 May 1995; MS. number: 4610)

Abstract. The hand preferences of orang-utans have been studied to a lesser extent than those of the other great apes and there has been no previous investigation of free-ranging orang-utans. Left and right hand use by rehabilitated orang-utans living in their natural environment was scored during feeding and self-touching of the face, head and body. The scores were ranked according to three categories designating the degree of finger flexion and manipulation. A total of 43 subjects ranging in age from 1.5 to 28 years were scored, although the sample size varied for each condition. A significant left-hand preference within a subgroup of 31 subjects was found for scratching and for fine manipulation of parts of the face, the strength of this preference increasing with age. Left-hand preferences were present also for some of the categories of head and body touching, but not to the same degree as those for face touching. No group bias for either a left- or right-hand preference was found for hand use in feeding, although some individuals displayed significant left- or right-hand preferences for unimanual holding of food when in either the hanging or propped position and half of the subjects had a preferred hand for fine manipulation in bimanual feeding. Undifferentiated hand use during bimanual feeding was more characteristic of young orang-utans, and older females were more likely to have a right-hand preference. No hand preferences were found for touching conspecifics and sucking fingers or toes, but two out of four mothers displayed a preferred side for carrying the baby. ?

number of species: left handedness for visually guided acts of manual prehension and right handedness for fine manipulation. The recent volume edited by Ward & Hopkins (1993) arrived at a similar conclusion, but most of the contributors to this volume agreed that there is a great need for more data collection on handedness in primates, using improved methodology and more creative experimental paradigms. There is also a need to increase sample sizes and to use more demanding tasks, as handedness appears to be more obvious in tasks of greater difficulty (Fagot & Vauclair 1991). In fact, Fragaszy & Adams-Curtis (1993) emphasized the value of using bilateral tasks to assess the relative use of either hand. Moreover, we should stress the need for assessing handedness in primates living in their natural environment and in species-typical patterns of social interaction. Primates living in captivity have relatively limited access to objects that can be manipulated, or very little object choice, and they

Preferential hand use or handedness is one of the most accessible measures reflecting lateralization of brain function, an aspect of brain organization now known to be characteristic of a wide range of species (Bradshaw & Rogers 1993). The earlier studies of hand use in primates failed to find a population bias for left or right handedness, even though individual primates preferred to use one hand or the other (Warren 1980). However, an important review by MacNeilage et al. (1987) pointed out the inadequacy of the literature particularly in terms of the number of primate species sampled and in the sample sizes of those species studied. Re-assessment of some of the earlier data led MacNeilage et al. (1987) to the conclusion that there was more than simply suggestive evidence for handedness at a group or population level in a Correspondence: L. J. Rogers, Department of Physiology, University of New England, Armidale, NSW 2351, Australia (e-mail: [email protected]). 0003–3472/96/010013+13 $12.00/0

1996 The Association for the Study of Animal Behaviour

?

1996 The Association for the Study of Animal Behaviour

13

Age (years)

1.5 1.5 3 3 4 4 4.5 4.5 5 5.5 6 6.5 6.5 6.5 6.5 7 8.5 10 10 11 13 14 14.5

Individual

Females Rana Marico Rampig Rafida Gina Boon Merotai Uji Bobbie Alice Shena Abbie Judy Mosuli Leila Nora Noren Masneh Jessica Julie Gandipot Siduku* Clementine NS NS LL NS NS NS LL NS L

NS NS NS NS NS NS NS

NS

NS LL

NS NS L LL

LL LL L LL LL RR NS

Body touching 1991+1992

L NS

L

Face+head touching 1991+1992

L

R

L

R

NS NS R R NS

NS NS L NS

L

NS NS

NS

Unimanual holding when propped

NS NS R

NS

NS

NS

NS NS

NS NS L NS NS

NS R

Unimanual holding when hanging

NS

NS

NS

NS NS

Face touching only 1992

Table I. Hand preferences calculated from data collected in 1991 and 1992

R

R

NS L NS

NS

L NS

NS

Bimanual hand use when propped

NS

NS NS NS R

NS NS NS R

NS

NS NS NS NS

Touching conspecifics

NS

R R NS

Baby carrying

NS NS NS

NS NS

NS

NS

Finger or toe sucking

14 Animal Behaviour, 51, 1

2 2 2 3 3.5 3.5 3.5 4 4 4 4 4.5 5 7.5 7.5 9 9.5 10 13 28 NS LL NS NS NS LL NS

NS LL NS NS

NS LL LL

NS NS

NS

LL NS

LL

L LL

NS NS

L LL NS NS NS

NS

NS

L

NS NS

NS NS

NS

NS NS NS

L NS NS L L NS

NS NS NS

NS

NS R L

R

R NS

NS

NS

L

NS

NS NS NS

NS L

NS NS NS NS NS

NS

NS indicates that there was no significant left or right bias determined in a z-score analysis for the individual, L indicates a significant left-hand preference and R indicates a significant right-hand preference. The absence of a letter indicates that insufficient scores were collected. Age is in years to nearest 0.5 year on January 1992. For face+head touching and body touching, a significant bias using LH2 +LH3 versus RH2 +RH3 is indicated by two letters (LL or RR), whereas a significant bias for LH3 versus RH3 only is indicated by L or R. For unimanual holding, when hanging or in the propped position, L and R represent significance for LH2 +LH3 versus RH2 +RH3. For bimanual holding when in the propped position, L and R indicate significant z-scores for LH3RH2 versus LH2RH3. The sample sizes for touching conspecifics and finger or toe sucking were too small for statistical analysis. The ‘baby carrying’ is for the side of the mother on which the baby is carried. *Siduku had the index finger of her left hand missing. †Died before the 1992 recording period and thus the age given is for January 1991.

Males Along Lobolo Gemasil† Josh Mike Micky Ringo Saidon Tom Benedit Jojo CID Gino BJ† Korashi† Edwin Boy King Raja Simbo

Rogers & Kaplan: Hand preferences in orang-utans 15

16

Animal Behaviour, 51, 1

frequently live in isolation or are reared without adequate mother–infant interactions. These factors might well influence the development of hand preferences. Hence, we decided to study hand use in orang-utans living in their natural habitat. As completely wild orang-utans are so elusive, it was necessary to study rehabilitated orang-utans living in a free-ranging situation and habituated to the presence of humans (Kaplan & Rogers 1994). Although the hand preferences of rehabilitated orang-utans may be influenced by their contact in early life with humans, no matter how brief, they have also had considerable free-ranging experience as independent individuals in an environment that allows them a full range of possibilities for climbing, brachiating, breaking branches for nest building and for gathering food. Previous studies of handedness in orang-utans have been confined to captive environments. Heestand (1986, cited in Morris et al. 1993) reported a bias for leading out with the right limb in terrestrial locomotion in some zoo orang-utans, but this result may have limited relevance since orang-utans are primarily arboreal. Cunningham et al. (1989) studied the development of hand preferences in a single, captive orang-utan, and found a right-hand preference in reaching for food and a left-hand preference for touching the face. This left-hand preference for face touching is consistent with an earlier report of the same bias in a larger sample of orang-utans, as well as in monkeys, gorillas, Gorilla gorilla, chimpanzees, Pan troglodytes, and humans (Dimond & Harries 1984). However, this result was obtained from a scoring period of only 20 min, and it has been criticized on a number of other grounds including sampling methods and statistical analysis (Suarez & Gallop 1986). Brésard & Bresson (1983) conducted a similar detailed study of hand preference in a single, captive orang-utan on a number of tasks, and found consistent use of the right hand particularly in tasks requiring bimanual coordination. Olson et al. (1990) tested orang-utans on three different retrieval tasks, including eight subjects tested on a task requiring them to unfasten a hasp to obtain a food reward. Most of the individuals exhibited hand preferences but there was no group bias for left or right handedness. The authors attributed this apparent lack of handedness in orang-utans to their limited performance of bipedal locomotion and hence their lesser need to adopt an unstable

posture, which might lead to handedness. In fact, Hopkins (1993) has shown that orang-utans exhibit hand preferences when they reach for food in a standing posture (six out of the nine were righthand preferent) but not when they reach from a quadrupedal posture. This result is consistent with results for prosimians (Ward et al. 1993). Given that the posture adopted while using the hands has an influence on hand preferences, comparisons between captive and free-ranging or wild primates have become imperative. There has been only limited investigation of handedness in any of the apes, and only one conducted in free-living individuals (for a summary of handedness in all apes see Morris et al. 1993). The latter study was of wild mountain gorillas scored for hand use in feeding (Byrne & Byrne 1991). The researchers found evidence for individual preferences, but no overall bias within the sample of 44 subjects. It is also important to investigate hand preferences in more than one paradigm. In the present study we investigated hand use in touching as well as feeding.

METHODS Subjects and Data Recording We observed 43 subjects (23 females and 20 males) ranging in age from 1.5 to 28 years at the Orang-utan Rehabilitation Centres in Sarawak (Semengoh) and Sabah (Sepilok), East Malaysia. The main part of the study was in Sabah (38 of the subjects). All of the subjects were free-ranging, rehabilitated orang-utans with various degrees of independent living in the rainforest reserve. For example, the oldest male (28 years: Simbo in Table I) had been fully rehabilitated for many years and he visited the feeding platform very rarely, at times months apart and only before the main fruiting season. The youngest orang-utans fell into two extremes, those born to rehabilitated mothers and essentially with only indirect and fleeting contact with humans when the mother visited the feeding table (Rana and Marico in Table I) and those that were released into the forest only during the day and slept under shelter at night (the remaining 2–3 year olds in Table I). Most of the older orangutans no longer had close contact with humans but visited the provisioning sites regularly. Most of these orang-utans had suffered the trauma of being separated from their mothers when young

Rogers & Kaplan: Hand preferences in orang-utans (for more details of the individuals in this study see Kaplan & Rogers 1994). All of the subjects were recognized as individuals with their appropriate names. Although data were recorded for all of these animals, not all provided sufficient scores on every measure for analysis. The sample sizes for each data set are stated in the Results and in Table I. Most of the data were collected in Sabah. The scoring occurred over 2-month periods in 2 consecutive years, 1991 and 1992 (total of 4 months). Observations were made between 0930 and 1630 hours mostly prior to or after the provisioning of food and after the tourists had left the site. Most of the scores were taken when the orang-utans were within 50 m of the feeding platform but not actually on it. At Sepilok the observations were made in the vicinity of two feeding stations, one located further into the rainforest and visited only on occasions by orang-utans that had otherwise been fully rehabilitated (for more detail see Kaplan & Rogers 1994). We collected the data by both pen and paper recording in the field, using binoculars for observations, and videotaping for later replay and analysis. We recorded 250 h of data. Before scoring commenced we learned to recognize the individuals. Twelve of the individuals scored in 1991 were scored again in 1992. Table I lists the individuals providing sufficient scores to be included in the data analysis for each measure of hand use. Face, Head and Body Touching Touching of the face and head by the left or right hand was scored in three categories: (1) with the back of the hand or flat palm without finger flexion (usually used to swat insects), (2) with the palm of the hand towards the face or head using the finger tips with finger flexion (usually for scratching), and (3) as in (2) but using fine finger movements to manipulate a part of the anatomy (usually to clean the nose, ears, eyes or teeth). In 1991 no discrimination was made between touching the face or head, but in 1992 these two types of touching were scored separately. Touching the face or head with both hands simultaneously or with the feet was scored. Food Holding and Manipulation Handedness for feeding was more difficult to assess because of the variety of forms it takes,

17

bimanual as well as unimanual. Hand use in feeding was ranked from 1 to 3, according to the degree of food manipulation. Category 1 referred to reaching out and taking hold of food (bananas or breadfruit given at the feeding platform or leaves picked from trees by the orang-utans). Category 2 referred to holding the food object in a prehensile grasp together with taking it to the mouth, and category 3 scores required manipulation of the food object with fine finger movement. The posture of the individual, hanging or in the propped position, was also scored. The propped posture included being seated, usually on the feeding platform or ground, or standing using one hand to support the body. Combined use of both hands occurred only in the seated posture or occasionally when the orang-utan was hanging by its feet. When both hands were used, the differential use of each hand was assessed. For example, if a piece of food was held in the left hand and peeled by the right hand, it received a score of LH2RH3. Usually, bimanual holding and manipulation of food involved differential use of either hand, but equivalent use of both hands also occurred (LH2RH2 or LH3RH3). Use of the feet to hold food was scored as for the hands, but it occurred relatively infrequently. Other Lateralized Behaviour Recorded The hand(s) used to touch another orang-utan was also recorded throughout the observation periods. We did not score touching during bouts of play, but only brief tactile interactions made, for example, when one individual met another. Touching of conspecifics was not categorized for degree of hand and finger use. Four mothers were observed with their babies, and the side of the mother to which the baby was clinging, usually when she moved through the canopy, was scored each time a new sighting was made. The part of the mother’s body to which the baby was clinging was also recorded. A record was made at each new sighting of the mother and if the position of the baby changed during a sighting period. The hand used for thumb or finger sucking was also noted, as was the foot for toe sucking. Most examples of this behaviour occurred at Semengoh just prior to the time for the provisioning of food.

Animal Behaviour, 51, 1

18 50

50

(a)

Mean scores (%)

40

40

*

*

30

30

20

20

10

10

0

(b)

1

2

3

0

*

1

2

3

Category Figure 1. The mean scores for (a) face plus head touching and (b) body touching are presented as percentages of the total number of such acts scored in each case separately (either total a or total b), together with standard error bars. Categories 1, 2 and 3 refer respectively to (1) use of the hand without finger flexion either touching with the back of the hand or a flat palm, (2) use of the front of hand with finger flexion, and (3) use of the hand with fine finger movements to manipulate a part of the face. .: Use of the left hand; /: use of the right hand. The asterisks indicate significant differences within a given category (see text). A total of 31 subjects were scored for face plus head touching, and 27 for body touching. The data collected in 1991 and 1992 have been lumped.

RESULTS Face, Head and Body Touching In 1991 no discrimination was made between the acts of touching the head and face. Therefore, the data for 1991 plus 1992 were analysed first with these two categories lumped (Fig. 1a). A mean of 28.5 acts (&5.0 ) was scored per individual (N=17 females and 14 males). An individual’s score was not included unless 10 or more acts had been recorded. For each of the categories (1, 2 or 3) z-score tests (Edwards 1963) were applied to the scores of each individual to assess the significance of the hand preference against an expected binomial distribution of the total responses. Each binomial z-score was calculated with respect to chance (Table I). In addition, a two-tailed, paired t-test was applied to the group data to test the significance of the difference in left and right hand use. For category 1, use of the hand with a flat palm, no significant left–righthand bias was present for any of the individuals or for the group. However, this category occurred the least frequently. A significant left-hand bias was present for category 2, use of the finger tips of the hand with finger flexion to scratch the head or face (for the group, t=2.618, df=30, P=0.014). For category 3, use of the hand and fingers to manipulate a part of the face or head, there was

also a significant preference to use the left hand (t=5.71, df=30, P=0.001). Thus, a preference to use the left hand was present when the fingers were flexed to touch the face or head, and the group bias for the left hand was stronger with increased fine finger movement (Fig. 1a). The only individual scoring a significant right-hand preference (z-score) for touching the face was Siduku, who had lost the index finger on her left hand. The hand preferences for touching the face and head may be stronger in females than males: 11 of the 17 females but only four of the 14 males had a significant hand preference (z-scores) (Table I). This gender difference is not due simply to different sample sizes for the males and females (the mean number of acts per individual& was 29.9&8.7 for females and 21.4&5.0 for males). In addition, older orang-utans were more likely to exhibit stronger hand preferences for touching the head and face than were younger ones (r=0.555, N=31, P=0.001, for a correlation analysis of percentage preference, calculated by lumping category 2 and 3 acts, and arcsine transformed against age: Fig. 2). For touching the body sufficient data were collected over 1991 plus 1992 for 27 individuals (16 females and 11 males) with a mean& of 31.1&5.6 acts per individual (Table I, Fig. 1b). There was a non-significant tendency for the same

Rogers & Kaplan: Hand preferences in orang-utans

Preference (%)

100 90 80 70 60 50 0

2

4

6 8 10 Age (years)

12

14

16

Figure 2. The percentage preference for left- or righthand use in face and head touching. The scores in categories 2 and 3 (see text) were lumped. The preferences plotted take no account of the direction of the preference: left and right preference are treated the same.

left-hand preference in category 2, the most common manner of hand use for touching the body (t=1.786, df=26, P=0.086), and a significant lefthand preference for category 3 (t=3.44, df=26, P=0.002). As for touching the face and head, the degree of lateralization increased with increased fineness of the finger movement. Incidentally, most of the category 3 acts of touching the body involved manipulation of the genitalia. In 1992 a distinction was made between touching the head and face, and compatible data sets for both of these scores were obtained for 13 orang-utans (Table I). No hand preference emerged for the group in touching the head with category 2 or 3 hand use, but there was a significant left-hand preference for category 1 hand use

Mean scores (%)

35

(with a flat palm or the back of the hand, usually for swatting insects: P=0.007). A strong left-hand preference occurred for touching the face, category 3 (t=2.58, df=12, P=0.024; Fig. 3), the most common category for face touching. Therefore, the left handedness was most marked for touching the face with fine finger movement to clean the teeth, eyes or ears. Touching the face, head or body with both hands simultaneously was very rare: the face was touched thus on only one occasion and the body on only five occasions. Use of the feet for self-touching was equally rare. Food Holding and Manipulation Food could be held and manipulated by the hands or feet. In the hanging position use of the feet made up only 5.77&1.55% (X&) of the total acts of food holding (mean number of acts of holding food with the feet per individual was 1.6&0.5). In the propped (or seated) position, use of the feet comprised 1.30&0.48% of the total acts (mean number per individual was 0.6&0.2) and, as expected, this was significantly lower than the amount when hanging (t=3.065, df=48, P=0.0036). These sample sizes for holding food with the feet were too small to be analysed for laterality. There was a great amount of category 3 manipulation with unimanual holding both when hanging and propped (Fig. 4). This involved manipulating the piece of food using a hand and 35

(a)

30

30

25

25

20

20

15

15

10

10

5 0

(b)

*

5

* 1

19

2

3

0

1

2

3

Category Figure 3. The mean scores for (a) head touching and (b) face touching as a percentage of the total number of acts scored (a+b). The same individuals are included in each data set (N=13 for each). .: Use of the left hand; /: use of the right hand. Categories as in Fig. 1.

Animal Behaviour, 51, 1

20

Mean scores (%)

40

(a)

40

30

30

20

20

10

10

0

1

2

3

0

(b)

1

2

3

Category Figure 4. Unimanual hand use (a) in the propped position and (b) when hanging. .: Use of the left hand; /: use of the right hand. Categories as in Fig. 1. N=22 and 27, respectively.

the mouth. There was no group bias for preferential use of the left or right hand in either case, although some individuals did show a significant preference for one or the other hand. In the hanging posture, three individuals had significant left-hand preferences (determined by z-score analysis of category 2 plus category 3 scores), three right-hand preferences and the rest had no preference (Table I). Five of the six individuals with hand preferences were females. In the propped posture, the pattern was the same but in some cases different individuals were involved (Table I). Hand preferences for unimanual holding would appear to be rather variable. In this context, it is worth mentioning that, when social interaction occurred during feeding while hanging, it frequently involved two orang-utans hanging with one arm so that they faced each other and adopting the use of opposite hands for holding and hanging. We have called this ‘mirror imaging’ behaviour (see Kaplan & Rogers 1994, page 128). It occurred frequently between a mother and her offspring, but also between consorting pairs and in other social interactions. Differential use of the hands could be assessed when the food was held simultaneously with both hands. It occurred for only 7.36&2.23% of the acts scored when the orang-utans were hanging, but in the propped position significantly more instances of hand use involved bimanual holding and manipulation (29.10&4.70% of the acts involved use of both hands; P=0.004). Thus, bimanual holding when in the propped posture

allowed a better assessment of handedness in feeding. Figure 5 shows examples for individuals feeding in the propped posture including both unimanual (LH0 and RH0) and bimanual hand use. When holding food with one hand and manipulating it with the other (i.e. LH3RH2 and LH2RH3), Boon had a strong left-hand preference and Gemasil an absolute left-hand preference for fine manipulation (LH3RH2 is the largest of these categories). Mosuli was ambidextrous and CID had an absolute right preference. All of these individuals showed no left or right preference in unimanual hand use (LH0RH3/LH0RH2 versus LH3RH0/ LH2RH0; Fig. 5). Thus hand preference in the propped posture may be revealed in bimanual acts of feeding even when it is not present in unimanual acts of feeding. Hand preferences were thus determined for each individual using only the bimanual scores for the propped posture. Again the significance of the differential left–right bias for each individual was determined by applying a standard binomial z-score. Of the 18 individuals (nine females and nine males) for which there were sufficient data, nine had a significant hand preference (four left and five right) and 10 had no preference. In most cases the individuals displaying a hand preference in bimanual feeding were not those showing it in unimanual feeding (Table I). However, females tended to be more consistently right-hand preferent across food holding/manipulating measures than males (Table I).

Rogers & Kaplan: Hand preferences in orang-utans Boon

21

Gemasil

30

40

20

30

10

20

0 LH3

10

30

40

20

30

10

20

0 LH3

10

0

LH2 LH1 LH0

RH0

RH1

RH2

0

LH2

RH3

% Observation

40

% Observation

40

LH1 LH0

RH0

RH1

RH2

RH3

CID

Mosuli

40

30

40

20

30

20

30

10

20

10

20

0 LH3

10

0 LH3

10

% Observation

30

0

LH2 LH1 LH0

RH0

RH1

RH2

RH3

% Observation

40

40

0

LH2 LH1 LH0

RH0

RH1

RH2

RH3

Figure 5. Three-dimensional plots of hand use by four individuals in the propped position. The mean number of observations of a particular category of hand use (see Fig. 1 and text) is plotted as a percentage of the total observations of hand use on the vertical axis. On the X- and Y-axes the various categories are represented. For example, LH3RH0 refers to unimanual holding by the left hand including manipulation and taking the food to the mouth, whereas LH3RH2 refers to bimanual holding of food with use of the right hand to grasp it and the left hand to manipulate with fine finger movements.

Note in Fig. 5 the occurrence of undifferentiated use of both hands (LH2RH2 and LH3RH3) by Gemasil, CID and Boon. For females, but not males, undifferentiated use of the hands when in the propped posture (expressed as a percentage of total bimanual hand use) correlated negatively with age (r=0.534, N=9, P=0.041, correlation analysis).

Other Lateralized Behaviour Recorded Touching of conspecifics occurred far less often than self-touching; over the same period for which we scored hand use in self-touching and feeding, we scored 22 individuals performing acts of touching a conspecific but the number of acts per individual was only 10.9&8.0 (X&).

Animal Behaviour, 51, 1

22

Table II. The side of the mother to which the baby clung when she moved through the canopy Mother

Age of baby

Total observations

% Right side

z-score P-value

Clementine Gandipot Julie Jessica

1 year 1.5 years 1 year 1 week

11 15 39 90

55 47 72 74

  <0.01 <0.001

Nevertheless, we must emphasize that this behaviour was not our main focus and we could therefore have over-looked some instances. The acts of touching conspecifics were not ranked according to the categories 1–3, but they rarely involved category 3 with fine finger movements to manipulate another individual, unless the orang-utans were playing and our scores of touching conspecifics did not include scores for play interactions. Category 3 touching of a conspecific was, however, observed occasionally in brief encounters; for example, when one orang-utan pinched the chin of another. Considering the data for touching conspecifics as a whole, there was no group bias for hand preference in touching conspecifics; the % right handedness was 52.7&7.4% (X&). For most individuals there were not sufficient scores to justify calculation of a z-score, but out of the seven with sufficient scores only three had a significant bias (two right and one left: Table I). The most common position for the baby was clinging to the hip of the mother, although a newborn baby was cradled on the mother’s chest. Four mothers were observed carrying babies. No significant side bias for carrying was found for two of the mothers, although there was a low sample size for both of these. A significant (z-score) right-side preference was present for the other two mothers (Table II). Thumb or finger sucking was recorded for 11 individuals and a total of 17 events, mostly prior to the provisioning of food at Semengoh, when the orang-utans exhibited other frustration behaviour, or after food had been placed on the platform and the particular individual showed reluctance to approach it. The finger placed in the mouth was either the thumb or the index finger, with equal occurrence. Toe sucking was also observed in five of the same individuals which exhibited thumb or finger sucking. Out of 13

observed events of toe sucking, all involved sucking the first (largest) toe. Although the sample size is too small to draw a conclusion, there was no indication of a group preference for the digits of the left or right hand or foot (17 left to 13 right). Finger and toe sucking may be more common in younger orang-utans: in our study it was observed in males aged less than 4 years and in females aged less than 7 years.

DISCUSSION We have found a strong group bias for a left-hand preference in touching the face and to a lesser degree for touching the head and body. The left-hand bias of our sample was stronger when fine finger movements were used to manipulate a part of the face. This hand preference present within the group as a consistent lateral bias indicates that orang-utans have lateralization of hemispheric control. These are the first data on hand preferences available for a large sample size of orang-utans and for orang-utans living in their natural environment. As humans show similar left handedness for touching the face (British people: Dimond & Harries 1984; but not Japanese people: Hatta & Dimond 1984), it is possible that these rehabilitated orang-utans acquired their left-hand preference by observing humans. This could be answered only by studying completely wild orangutans. In humans the left-hand preference does not occur as an obligatory consequence of right handedness for other tasks requiring fine motor control, such as writing, and so too in orang-utans the left-hand preference for face touching occurs independently of any group bias for a preferred hand for holding and manipulating food. Therefore, hand use to manipulate parts of the face may be an independent motor laterality, determined by

Rogers & Kaplan: Hand preferences in orang-utans right-hemisphere dominance. Dimond & Harries (1984) suggested that the left hand might be used to heighten the effectiveness of facial expressions of emotion which are also controlled by the right hemisphere and expressed more strongly on the left side of the face, as known for humans and for rhesus monkeys, Macaca mulatta (Hauser 1993). As there has been very limited study of facial expressions in orang-utans (see Kaplan & Rogers 1994), we cannot confirm or reject this as an explanation for our results. However, much of the face touching that we observed occurred when the orang-utans were in relative isolation, and the left-hand preference was most apparent when the face touching involved fine finger movements to manipulate parts of the face (teeth, eyes, ears). The hand was not being used as a pointer for facial expressions, as suggested by Dimond & Harries (1984), but to perform self-directed manipulation. Self-directed behaviour could also be used to describe sucking of digits. There was, however, no indication of a group bias for preferring a particular hand in these scores, although our data are insufficient for drawing a conclusion. Nevertheless, it may indicate that orang-utans differ from chimpanzees in this behaviour, as Hopkins & Bard (1993) have reported a right-hand preference for ‘hand-to-mouth’ behaviour in infant chimpanzees (see also Bard et al. 1990). However, the hand-to-mouth behaviour of chimpanzees is not exactly equivalent to finger sucking, and it could perhaps be compared to face touching. If so, the direction of laterality for this behaviour in chimpanzees is opposite to that of orang-utans and humans. For touching the body there was a less marked, but significant, left-hand preference within the group for manipulation of parts of the body, usually the genitalia. This preference is consistent with dominance of the right hemisphere for controlling self-touching. Touching the head, as opposed to the face, rarely involved fine finger movement for manipulation and the data showed no group bias of handedness for this form of touching or for scratching. However, a significant left-hand bias was present for touching the head with the back of the hand or a flat palm without finger flexion, again reflecting right hemispheric dominance. Touching of another orang-utan showed no handedness bias in our data but, as we had few

23

observations per individual, we hesitate to draw any firm conclusions on this behaviour. Nevertheless, it is interesting to speculate that touching the self and touching others may be controlled by different neural mechanisms. Hand preferences for holding and manipulating food clearly had no relationship to hand preferences for self-touching. There was no group bias for a consistently preferred hand irrespective of whether it was assessed by scoring unimanual holding or by differential use of the hands when the piece of food was held bimanually. Some individuals exhibited right-hand preferences across some of the types of feeding behaviour, others left-hand preferences, but the majority of orang-utans were ambidextrous. Hopkins (1993) might argue that a group bias for handedness would emerge only when the orang-utans have to reach for food from an unstable posture, but hanging must be such a posture and no more right handedness emerged for unimanual holding when hanging. Perhaps the captive orang-utans tested by Hopkins (1993) had learned their right-hand preferences from their human carers. If orang-utans can develop hand preferences by learning, the ‘mirror image’ feeding that we observed may be a reason why there is no population bias for handedness. This pattern of behaviour, in which two orang-utans face each other using opposite hands to hang and to hold the food, might be a form of social interaction involving learning (particularly between mother and offspring) or bond confirmation. If such mother– infant learning has a lasting effect on handedness, one might predict that offspring would have hand preferences opposite to their mother’s hand preference. This would be contrary to some evidence for crab-eating macaques, M. fascicularis, showing that offspring tend to have the same hand preferences as their mothers (Brinkman 1984). Alternatively, owing to their arboreal existence, orang-utans may maintain malleability of hand use when hanging. As found previously for other primates (Ward et al. 1993), age was an important factor contributing to the strength of hand preferences. Older orang-utans had stronger preferences when touching the face and, when feeding with both hands, older females were more likely to exhibit a righthand preference for fine manipulation. Forsythe & Ward (1988) have noted a similar increase in handedness with age for food reaching and

24

Animal Behaviour, 51, 1

holding in the black lemur, Lemur macaco, although Heestand (1986, cited in Morris et al. 1993) found no effect of age on the lateralization of leading limb preference in orang-utans. In our study, the stronger hand preferences (right or left) that occurred with increasing age suggest that the preferences are learned by individuals. Gender might influence the strength and direction of hand preferences because older females were right-hand preferent for feeding, whereas this was a less obvious trend in males. Ward and her colleagues have reported similar effects of gender in lower primate species (Milliken et al. 1989; Ward et al. 1993). The younger orang-utans in our study were less likely to have a hand preference for food holding and manipulation, although they had a left-hand preference for touching the face. They were also more likely to feed by using both hands equally, in an undifferentiated manner. This apparently reflects a lesser degree of manipulative ability in young individuals. The side of the mother to which her baby was clinging as she moved was lateralized in two out of the four mother–infant pairs recorded. This bias might influence the development of lateralized hand use in the offspring. Manning & Chamberlain (1990) reported a significant left-side preference for cradling a baby in a mixed group of captive chimpanzees, gorillas and orang-utans, as also is the case for humans. There were four orang-utans in their sample, three with a left-side preference and one with a right-side preference, but there were methodological problems in the collection of their data (see Hopkins et al. 1993). Furthermore, it must be stressed that, whereas captive apes apparently hold or cradle the baby, all of our observations were for infants clinging to the mother while she used her limbs to brachiate. In fact, only Jessica with a 1-week-old baby was seen cradling her baby at times. It is possible that cradling of older infants, as observed by Manning & Chamberlain (1990), occurs more frequently in captivity where brachiation is less possible and where other orang-utans, or even humans, may potentially have more access to the baby (Hatta & Koike 1991). In the mother–infant pairs for which we found a right-side preference, the position of the baby appeared to be determined by the mother, as we frequently observed the mother using a hand to move the baby to her right side. This result

contrasts with that of Hatta & Koike (1991) who reported that Old World monkeys use their left hand to take up the baby when they are frightened. These variations indicate the need for more information on these lateralized behaviour patterns in different species as well as orang-utans. Hand preferences vary with the task being performed and the investigation of hand preferences within a species should not be limited to measurement of hand use for retrieving and holding food.

ACKNOWLEDGMENTS L.J.R. gratefully acknowledges support from a University of New England Research Grant for this project and G.K. acknowledges the same from the Queensland University of Technology. We thank G. Krebs for assistance in analysing the data and two anonymous referees, who provided most useful suggestions for improving the manuscript.

REFERENCES Bard, K. A., Hopkins, W. D. & Fort, C. L. 1990. Lateral bias in infant chimpanzees (Pan troglodytes). J. comp. Psychol., 104, 309–321. Bradshaw, J. L. & Rogers, L. J. 1993. The Evolution of Lateral Asymmetries, Language, Tool Use, and Intellect. San Diego: Academic Press. Brésard, B. & Bresson, F. 1983. Handedness in Pongo pygmaeus and Pan troglodytes. J. hum. Evol., 12, 659–666. Brinkman, C. 1984. Determinants of hand preference in Macaca fascicularis. Int. J. Primatol., 5, 325. Byrne, R. W. & Byrne, J. M. 1991. Hand preferences in the skilled gathering tasks of mountain gorillas (Gorilla g. berengei). Cortex, 27, 521–546. Cunningham, D., Forsythe, C. & Ward, J. 1989. A report of behavioral lateralization in an infant orangutan (Pongo pygmaeus). Primates, 30, 249–253. Dimond, S. & Harries, R. 1984. Face touching in monkeys, apes and man: evolutionary origins and cerebral asymmetry. Neuropsychologia, 22, 227–233. Edwards, A. L. 1963. Statistical Methods for the Behavioral Sciences. New York: Holt, Reinhart and Winston. Fagot J. & Vauclair, J. 1991. Manual laterality in nonhuman primates: a distinction between handedness and manual specialization. Psychol. Bull., 109, 76–89. Forsythe, C. & Ward, J. P. 1988. Black lemur (Lemur macaco) hand preference in food reaching. Primates, 29, 369–374. Fragaszy, D. M. & Adams-Curtis, L. E. 1993. An exploration of manual preference and performance in

Rogers & Kaplan: Hand preferences in orang-utans crab-eating macaques. In: Primate Laterality: Current Behavioral Evidence of Primate Asymmetries (Ed. by J. P. Ward & W. D. Hopkins), pp. 75–105. New York: Springer-Verlag. Hatta, T. & Dimond, S. J. 1984. Differences in face touching by Japanese and British people. Neuropsychologia, 22, 531–534. Hatta, T. & Koike, M. 1991. Left-hand preference in frightened mother monkeys in taking up their babies. Neuropsychologia, 29, 207–209. Hauser, M. D. 1993. Right hemisphere dominance for the production of facial expression in monkeys. Science, 261, 475–477. Heestand, J. 1986. Behavioral lateralization in four species of ape. Ph.D. thesis, University of Washington, Seattle. Hopkins, W. D. 1993. Posture and reaching in chimpanzees (Pan troglodytes) and orang-utans (Pongo pygmaeus). J. comp. Psychol., 107, 162–168. Hopkins, W. D. & Bard, K. A. 1993. The ontogeny of lateralized behavior in nonhuman primates with special reference to chimpanzees (Pan troglodytes). In: Primate Laterality: Current Behavioral Evidence of Primate Asymmetries (Ed. by J. P. Ward & W. D. Hopkins), pp. 251–265. New York: Springer-Verlag. Hopkins, W. D., Bard, K. A., Jones, A. & Bales, S. L. 1993. Chimpanzee hand preference in throwing and infant cradling: implications for the origin of human handedness. Curr. Anthropol., 34, 786–790. Kaplan, G. & Rogers, L. J. 1994. Orang-Utans in Borneo. Armidale: University of New England Press.

25

MacNeilage, P. F., Studdert-Kennedy, M. G. & Lindblom, B. 1987. Primate handedness reconsidered. Behav. Brain Sci., 11, 748–758. Manning, J. T. & Chamberlain, A. T. 1990. The left-side cradling preference in great apes. Anim. Behav., 39, 1224–1227. Milliken, G. W., Forsythe, C. & Ward, J. P. 1989. Multiple measures of hand-use lateralization in the ring-tailed lemur (Lemur catta). J. comp. Psychol., 103, 262–268. Morris, R. D., Hopkins, W. D., Bolser, G. & Washburn, D. A. 1993. Behavioral lateralization in languagetrained chimpanzees. In: Primate Laterality: Current Behavioral Evidence of Primate Asymmetries (Ed. by J. P. Ward & W. D. Hopkins), pp. 207–233. New York: Springer-Verlag. Olson, D., Ellis, J. E. & Nadler, R. D. 1990. Hand preferences in captive gorillas, orang-utans, and gibbons. Am. J. Primatol., 20, 83–94. Suarez, S. D. & Gallup, G. G. 1986. Face touching in primates: a closer look. Neurospsychologia, 24, 597–600. Ward, J. P. & Hopkins, W. D. (Ed.) 1993. Primate Laterality: Current Behavioral Evidence of Primate Asymmetries. New York: Springer-Verlag. Ward, J. P., Milliken, G. W. & Stafford, D. K. 1993. Patterns of lateralized behavior in prosimians. In: Primate Laterality: Current Behavioral Evidence of Primate Asymmetries (Ed. by J. P. Ward & W. D. Hopkins), pp. 43–74. New York: Springer-Verlag. Warren, J. M. 1980. Handedness and laterality in humans and other animals. Physiol. Psychol., 8, 351–359.