Primates

, Volume 51, Issue 3, pp 251–261

Handedness in captive gorillas (Gorilla gorilla)

Authors

    • Department of ArchaeologyUniversity of Sheffield
  • Pia Nystrom
    • Department of ArchaeologyUniversity of Sheffield
Original Article

DOI: 10.1007/s10329-010-0191-9

Cite this article as:
Harrison, R.M. & Nystrom, P. Primates (2010) 51: 251. doi:10.1007/s10329-010-0191-9

Abstract

Species-level right handedness is frequently presented as a marker of human uniqueness. Handedness also has implications for the evolution of language and cognition. In this study, we examined handedness in 22 captive gorillas (Gorilla gorilla) across a range of behaviours that were part of their daily routine. Ten individuals showed no preferences for any of the behaviours performed, and the majority of the remaining individuals showed a preference for only one behaviour. These results lend support to the theory that species-level handedness is unique to humans. It is hoped that these results will contribute to investigations into the evolution of handedness, which can ultimately be used to further our understanding of the evolution of human language and cognition.

Keywords

GorillaHandednessLaterality of function

Introduction

It is widely reported that modern humans show species-level right handedness (Seltzer et al. 1990; Peters and Murphy 1992; Annett 2002) and that this can be correlated with differential specialisation in the hemispheres of the brain (Annett 1992; Witelson and Kigar 1992; Jancke et al. 1994; Amuts et al. 1996). Handedness may be studied because it is considered to be a marker of human uniqueness (Corballis 1989), but more importantly, some researchers have suggested that right handedness was the initial step in the specialisation of the left hemisphere, which ultimately led to the evolution of language (Corballis 1989, 1991; Bradshaw and Rogers 1993). Even though we do not directly contribute to this debate, the data presented will contribute important information to the database on handedness. The archaeological record can only provide limited information about the origins of manual laterality and laterality in the hemispheres of the brain. It is therefore necessary to construct inferential models based on the behaviour of extant nonhuman primates (McGrew 1993). It is particularly important to investigate other apes due to their genetic, neural and anatomical similarities to humans (Parnell 2001).

The majority of research on handedness in nonhuman apes has focused on the chimpanzee (Pan troglodytes), yet the results produced are varied. Different studies found species-level handedness (e.g. Hopkins et al. 2004), individual-level handedness (Hopkins 1994; Colell et al. 1995; Hopkins et al. 1997) or no significant hand preferences (Heestand 1986; Hopkins 1993; Hopkins et al. 1993a; Marchant and McGrew 1996; McGrew and Marchant 2001; Marchant and McGrew 2007). The fact that different populations are observed, different tasks are measured, and often different methods of analysis are used may explain the discrepancies in the results of previous research. Species-level handedness is mostly found in studies of captive populations in which one or more complex, artificial feeding tasks are recorded (Hopkins et al. 2001, 2004, 2005a). In contrast a lack of preference tends to be found in studies on wild populations, in which a larger range of the behavioural repertoire is recorded (Marchant and McGrew 1996; McGrew and Marchant 2001). In an ethological study of captive chimpanzees, Fletcher and Weghorst (2005) found no population-level handedness for non-tool-use tasks; however, during tool use, more than half of the individuals showed a significant left-hand preference. Consequently it appears less likely that there will be a significant expression of hand preference when tasks are less cognitively demanding, but when tool-using behaviours are recorded, individuals appear more likely to express side preference. Research on bonobos (Pan paniscus) has found little evidence for species-level handedness, though population-level handedness has been found for certain tasks, for example leading limb, reaching for food and gesturing (e.g. Hopkins et al. 1993b; de Vleeschouwer et al. 1995; Hopkins and de Waal 1995; Harrison and Nystrom 2008), and individual preferences have been found during tool use and feeding (Harrison and Nystrom 2008).

Data available on handedness in gorillas are limited. Previous results from feeding studies, bimanual feeding and food manipulation in captive populations have produced varied results, showing population-level preferences for either left or right hand, individual-level preferences or a lack of preferences (Fagot and Vauclair 1988; Olson et al. 1990; Annett and Annett 1991; Hopkins et al. 2003). In studies of food processing in wild populations, individual but no population-level preferences have been found (Byrne and Byrne 1991; Parnell 2001). In a review of 22 studies of gorilla laterality, McGrew and Marchant (1993) found inconclusive evidence and stated that gorillas could be shown to be right- or left-handed depending on the tasks measured, and they suggested that more data were required to clarify the issue.

Gorillas, like other nonhuman primates, use their hands for both locomotor and manipulatory functions. They are predominantly terrestrial, and locomote using knuckle-walking (Robbins 2007). Gorillas use their hands when feeding and foraging and during social interactions such as chest beating and grooming. Evidence for tool use in captive western gorillas is limited and includes throwing objects, and using tools during feeding activities (Natale et al. 1988; Fontain et al. 1995; Nakamichi 1998, 1999; Parker et al. 1999). There is anecdotal evidence of tool use in wild western gorillas. In two separate instances, two adult females used a detached branch, firstly to provide postural support when crossing water, and secondly to provide postural support during the processing of aquatic herbs (Breuer et al. 2005). There are no reports of tool manufacture in wild populations (as reviewed by Panger 2007). Consequently, gorillas might be expected to show less lateralisation of behaviour than chimpanzees, who habitually manufacture and use tools in the wild (e.g. McGrew 1974; Boesch and Boesch 1990; Brewer and McGrew 1990), as it is during tool use that most pronounced specialisation is found, even in modern humans (Marchant et al. 1995).

In this study, we examines a large data set of gorilla behaviours with the aim of determining whether or not gorillas show levels of handedness comparable with bonobos and chimpanzees. Handedness is analysed at the individual level to prevent any errors due to data pooling. Pooled data on all individuals combined are provided in the “Appendix”. These data are provided for comparative purposes, as many studies in the past as well as several more recent studies have analysed pooled data when investigating handedness (e.g. Mosquera et al. 2007; Hulme and Matsuzawa 2009; Llorente et al. 2009).

The influence of sex, age, posture and type of manipulation (grip) on the expression of handedness is examined because of their potential influence on the expression of laterality. In humans, a higher level of right handedness has been found in females than males (Oldfield 1971; Annett 2002); consequently, we examined the possible differences in laterality between the sexes. The effect of age is investigated, as an increase in handedness with age may suggest that it is a learned behaviour (Annett 2002) rather than innate. Posture is examined, as it has been suggested that a dominant hand will be required to perform precise grasping, especially when an unstable posture is adopted (de Vleeschouwer et al. 1995). When an individual assumes an unstable posture, the resulting instability needs to be compensated for, and increased activity is required in the nervous system to maintain balance. This in turn might lead to an increase in the strength of lateralisation to increase efficiency (de Vleeschouwer et al. 1995). Alternatively, it is possible that it is the type of grip that influences hand preference, as different grips are used in different postures (Hopkins 1993). It is also possible that hand preference increases when precision grips and grips demanding complex movements are used (Jones-Engel and Bard 1996; Christel et al. 1998; Hopkins et al. 2005b).

In addition to examining individual- and population-level preference, the extent of laterality is evaluated using the 5-level classification system presented by McGrew and Marchant (1997, 2001). This classification system ranges from level 1, where all individuals in the group are ambipreferent; to level 5, where all individuals in the group are completely lateralised to either the left or right.

Methods

Subjects

Behavioural data were collected during May through August 2000, and July through August 2001. The gorillas came from five groups located in zoos in Germany (Berlin and Stuttgart) and the UK (Chessington, Edinburgh and London). Twenty-two individuals made up the study population, comprising 6 males (5 adult, 1 immature) and 16 females (13 adult, 3 immature). Information on the subjects studied is presented in Table 1. Each group had a habitat that consisted of at least one indoor and one outdoor enclosure. There was comparable arboreal habitat available for, and used by, gorillas at each location. Various objects were available for them to interact with, e.g. dipping cups, plastic bottles, balls, shredded paper and straw. The Stuttgart population also had access to food puzzles.
Table 1

Information on individual study animals: sex, age, observation hours and location

Individual

Sex

Age (year/month)

Age categorya

Observation time (h:min)

Location

Bokito

M

4/4

Immature

10:0

Berlin

Jeremiah

M

13/5

Adult

15:15

London

Knorke

M

27/0

Adult

10:15

Berlin

Sam Sam

M

29/0

Adult

9:45

Edinburgh

Kumba

M

32/0

Adult

37:15

Chessington

Banjo

M

35/0

Adult

9:0

Stuttgart

Mjukuu

F

1/7

Immature

25:0

Chessington

Buu

F

2/9

Immature

35:0

Chessington

Mutasi

F

6/2

Immature

10:0

Stuttgart

Asili

F

9/1

Adult

27:45

Chessington

Shani

F

10/6

Adult

33:0

Chessington

Kolo

F

13/10

Adult

10:0

Stuttgart

Momo

F

19/1

Adult

9:30

Edinburgh

Kaja

F

21/0

Adult

28:45

Chessington

Salome

F

21/1

Adult

25:15

Chessington

Zaire

F

22/10

Adult

20:0

London

Diana

F

26/0

Adult

15:0

London

Dina

F

29/0

Adult

10:0

Stuttgart

Undi

F

29/0

Adult

10:0

Stuttgart

Yinka

F

29/0

Adult

9:30

Edinburgh

Bafia

F

33/0

Adult

30:15

Chessington

Mimi

F

36/0

Adult

10:0

Stuttgart

Total hours

   

400:30

 

aSubjects were considered adult when they reached 10 years for males and 6 years 6 months for females (Harvey et al. 1986: Table 16.1)

Data collection

Data were collected from focal individuals over a period of 400 h, with individual observation time ranging from 9 to 35 h, as shown in Table 1. Individuals were recorded carrying out all aspects of their daily routine using all occurrence focal sampling (Altmann 1974). Data were recorded in bouts (the continued occurrence of a particular behaviour). Bouts were independent in terms of hand usage, with each bout being terminated when the individual switched to another activity, or by some other action or incident that might potentially change the use of one of the hands (Hopkins and de Waal 1995; Hopkins et al. 2001; McGrew and Marchant 2001). This method is comparable with the majority of other studies (e.g. Hopkins et al. 1993a; Hopkins and de Waal 1995; Marchant and McGrew 1996; McGrew and Marchant 2001).

Hand and foot use were recorded during different activities: (1) leading limb (use of hand or foot to initiate at least three consecutive steps of locomotion from a stationary position); (2) scratch (scrape digits through one’s own hair); (3) carry or manipulate object (convey an object for at least three consecutive steps; move, control or handle an object with the hand or foot for a purpose other than tool use or feeding); (4) feed (put a food item into the mouth); (5) manual gesture (the only gesture recorded was begging for food from another individual, a human keeper or a member of the public. No other manual gesture was performed sufficiently frequently for data analysis to be conducted); (6) tool use [use of an object to alter the form, position or condition of another object, another individual or the user themselves (Beck 1980)]. All instances of tool use involved reaching for food with a stick and were thus similar in terms of the movements and cognitive demands required.

Behaviours recorded varied in biomechanical and cognitive complexity, with leading limb being the least complex and tool use being the most (see, for example, Fagot and Vauclair 1991; McGrew and Marchant 1997; Hopkins 2006). Carrying and object manipulation were combined into one category, as they both involved manipulating an object for a purpose other than feeding or tool use, either whilst moving or in a stationary position. For each activity, posture and grip used were recorded. Postures recorded were: sitting, bipedal, quadrupedal, and suspensory. These postures were chosen to provide a variety of circumstances in which the individual was in a stable or unstable position, with both hands free to perform a particular task or one hand providing postural support. The different grips were classified as precision (object held using hand phalanges); power (gripping an object into the palm using the fingers); and miscellaneous (the thumb was not involved, and the grip required neither precision nor power). Grip categories were based on Marzke and Wullstein’s (1996) classification of grips used by nonhuman apes, which recognises different elements of precision and power, along with grips that are less complex biomechanically. More detailed information regarding methods of data collection and analysis used can be found in Harrison and Nystrom (2008).

Data analysis

Data were analysed at the individual-level using binomial tests, which are useful when investigating dichotomous variables and are widely employed in research on handedness (e.g. Hopkins et al. 1993b; Hopkins and de Waal 1995; Marchant and McGrew 1996; McGrew and Marchant 2001). A significance level of P < 0.01 was used to prevent type I errors. Given the contradictory nature of the results of previous studies, no direction was specified, and values given are two-tailed. No statistical analyses were conducted on pooled data, as such data are not deemed valid in the study of handedness in nonhuman primates (see Martin and Bateson 1993).

Results

Table 2 shows the number of bouts, for all behaviours, recorded for each individual using left and right hand or foot. Most individuals showed no significant preferences (either left or right) for most behaviours. Twelve individuals showed at least one significant preference for one behaviour, but only two individuals (both female) showed a preference over more than one behaviour, out of which only one individual was consistent in preference across behaviours (to the right) (Table 2).
Table 2

Number of bouts per individual for each behaviour studied (left/right), with significant preferences shown

Individual

LLa (hand)

LLa (foot)

Scratch

Feed

Carryb

Gesture

Tool use

Total

Males

 Immatures

  Bokito

33/29

4/11

4/9

61/82

9/14

11/16

 

122/161

 Adults

  Jeremiah

43/50

  

73/91

4/6

  

120/147

  Knorke

19/16

  

43/34

   

62/50

  Sam Sam

15/12

 

4/5

39/16*

   

58/33

  Kumba

106/109

 

9/11

139/95*

29/14

  

283/229

  Banjo

7/14

 

12/12

31/9*

  

9/6

59/41

Females

 Immatures

  Mjukuu

60/84

9/8

6/10

48/14*

20/10

8/12

 

151/138

  Buu

157/141

5/20*

2/4

83/48*

59/41

8/10

24/5*

338/269

  Mutasi

17/23

  

26/20

 

2/8

27/1*

72/52

 Adults

 

  Asili

92/91

6/10

9/11

60/102*

25/49*

18/37

1/21*

211/321

  Shani

143/144

4/6

5/6

99/92

53/37

7/7

22/8

333/300

  Kolo

39/31

12/10

5/7

14/16

5/4

 

2/15*

77/83

  Momo

17/18

 

3/5

20/43*

4/6

  

44/72

  Kaja

90/88

 

12/8

118/116

15/19

 

3/6

238/237

  Salome

57/62

  

40/57

4/8

  

101/127

  Zaire

86/116

  

58/84

4/6

1/12*

 

149/218

  Diana

61/61

  

89/22*

   

150/83

  Dina

19/7

  

16/16

  

2/9

37/32

  Undi

30/26

 

4/11

4/7

  

13/16

51/60

  Yinka

9/8

  

22/32

   

31/40

  Bafia

148/158

 

5/18

87/147*

11/25

  

251/348

  Mimi

26/36

 

8/10

42/54

13/10

 

19/25

108/135

Total

1274/1324

40/65

88/127

1212/1197

255/249

55/102

122/112

 

Binomial tests have only been conducted on individual data, not on pooled totals

* Significant bias when tested using a Binomial Test (P < 0.01)

aLeading limb

bIncludes manipulation

The number of behaviours performed by individuals ranged from 2 to 7, and the number of behaviours for which no significant preferences were found ranged from 1 to 7 per individual. Seven individuals showed a significant left-hand preference, but only 1 individual did so across more than one behaviour. Six individuals showed a significant right-hand preference, but only 1 individual did so across more than one behaviour. Ten individuals showed no preferences for any behaviours. Overall, 83% of all behaviours recorded across all individuals showed no significant preferences. There was a higher incidence of left-hand use in males and right-hand use in females. Three of the six males showed a significant left-hand preference for one behaviour. Amongst females, there was a significant preference towards the left hand for at least one behaviour in 4 individuals and towards the right in 6 individuals. Age did not provide any significant differences.

The two behaviours that elicited the highest number of significant preferences in individuals were feeding and tool use (Table 2). During feeding, 6 individuals showed a significant left-hand preference, 3 a significant right-hand preference and 13 no preference. For tool use, 2 individuals showed a significant left-hand preference, 2 a right-hand preference and 6 no preference. For the other behaviours recorded, the majority of individuals showed no significant hand preferences (see Table 2). Table 3 shows the number of bouts, for every posture used, recorded for each individual using left and right hand or foot. When sitting, 9 out of 22 individuals showed a significant preference: 5 to the left and 4 to the right; all were females. Most individuals showed no significant preferences (either left or right) for the other postures used.
Table 3

Number of bouts per individual for each posture (left/right), with significant preferences shown

Individual

Sitting

Bipedal

Quad

Suspensory

Males

 Immatures

  Bokito

18/31

8/10

64/80

 

 Adults

  Jeremiah

61/57

 

23/47*

 

  Knorke

12/14

 

35/41

 

  Sam Sam

18/16

 

36/8*

 

  Kumba

83/65

 

109/85

 

  Banjo

58/40

 

11/6

 

Females

 Immatures

  Mjukuu

65/28*

7/12

30/17

23/3*

  Buu

175/101*

7/4

53/46

29/10*

  Mutasi

84/23*

 

23/17

2/4

 Adults

  Asili

70/172*

14/32

66/111*

1/7

  Shani

161/107*

 

119/90

9/6

  Kolo

13/55*

 

16/15

 

  Momo

16/34

 

15/22

 

  Kaja

92/120

1/6

73/64

 

  Salome

38/44

 

25/36

 

  Zaire

37/59

1/14*

25/41

6/1

  Diana

61/14*

 

26/12

 

  Dina

9/39*

 

24/11

 

  Undi

43/55

 

7/4

 

  Yinka

23/15

 

8/21

 

  Bafia

44/173*

1/7

88/109

1/5

  Mimi

89/104

 

29/46

 

* Significant bias when tested using a Binomial Test (P < 0.01)

Table 4 shows the number of bouts, for every grip used, recorded for each individual using left or right hand. Most individuals showed no significant preferences (either to the left or right) for any of the grips used. However, 4 individuals (3 to the left, 1 to the right) showed preferences for both precision and power grips compared with 0 individuals for the miscellaneous grip.
Table 4

Number of bouts per individual for each grip (left/right), with significant preferences shown

Individuala

Precision

Power

Misc

Males

 Immatures

  Bokito

44/51

15/37

29/28

 Adults

  Jeremiah

  Knorke

6/13

20/10

21/14

  Sam Sam

16/15

24/2*

16/7

  Kumba

46/22*

16/14

14/17

  Banjo

41/21

16/7

17/21

Females

 Immatures

  Mjukuu

45/14*

58/20*

26/19

  Buu

54/30

85/33*

22/29

  Mutasi

82/17*

17/11

13/19

 Adults

  Asili

21/29

9/22

42/61

  Shani

26/26

17/15

27/21

  Kolo

18/52*

6/9

8/11

  Momo

14/25

8/14

9/17

  Kaja

37/39

13/12

29/22

  Dina

24/37

4/8

7/5

  Undi

44/47

2/6

6/12

  Yinka

13/11

12/20

3/3

  Bafia

19/22

11/5

13/29

  Mimi

57/95

29/27

34/29

* Significant bias when tested using a Binomial Test (P < 0.01)

aGrip was not recorded in the three individuals at London

Discussion

Previous research on captive gorillas has produced varied results. Fagot and Vauclair (1988) found significant population-level left-hand preferences for feeding tasks that were complex in terms of biomechanical and cognitive demands required. Significant preferences at an individual but not population level have been found for general feeding and for coordinated bimanual feeding tasks (Fagot and Vauclair 1988; Hopkins et al. 2003). In contrast other studies have found a lack of population-level preference while feeding (Olson et al. 1990; Annett and Annett 1991) and for manipulating food (Olson et al. 1990).

This study lends support to some of the results from previous research. We found that 12 out of the 22 individuals showed a significant preference for at least one behaviour, whereas 10 showed no significant preferences for any of the behaviours recorded. For the less complex tasks, i.e. leading limb (hand or foot), scratching, gesturing and carry/object manipulation, the majority of individuals showed no significant preferences, which supports previous findings (Fagot and Vauclair 1988; Hopkins et al. 2003). As predicted by previous research (Fagot and Vauclair 1988; Olson et al. 1990), a greater proportion of individuals were significantly lateralised when a more complex behaviour was performed, i.e. tool use. For tool use, 4 individuals (40%) showed a significant hand preference (2 to the left and 2 to the right). Thus, even for the most complex task, more than half of the individuals were unlateralised. In addition, there was no bias towards a particular side in individuals that were significantly lateralised. Previous research from captive populations, along with results of this study, lends no evidence to suggest that gorillas show species-level handedness.

Two major studies have been conducted on wild populations of gorillas. Byrne and Byrne (1991) studied the feeding behaviour of wild mountain gorillas (N = 44) at Karisoke in Rwanda. They reported several stages involved in plant processing, with the hands and mouth performing different tasks at the different stages. Individual-level preferences were found for tasks in which the hands were used in complementary roles in each stage of the food processing sequence. However the nature of certain bimanually coordinated feeding tasks meant that it was not always easy to distinguish which hand is the dominant one. The majority of individuals showed significant hand preferences for the three tasks involving processing of leaf food, but not the fourth, which involved stem food. These hand preferences tended to be consistent across tasks within individuals. At a population-level, Byrne and Byrne (1991) found that most individuals gripped stems with their left hand and performed fine manipulation with their right hand, although these data were not statistically significant.

Parnell (2001) studied the feeding behaviour of western lowland gorillas (N = 33) in northern Congo. Hand preferences were recorded during the procurement and processing of different plant species. Across all test conditions, there was 1 individual that showed an exclusive hand preference (towards the right), whereas more individuals showed a significant left preference than a significant right preference, but the population-level preference did not exceed a chance distribution. The number of individuals showing no preference was greater than the number of significant right and left individuals in all except one of the test conditions (Parnell 2001).

Consequently the results of wild populations are comparable with data on captive populations, i.e. that some individual preferences are found, but no population-level preferences occur in gorillas. The results of this study concur with Parnell’s (2001) finding that the number of individuals showing no significant preferences for any behaviour exceeded the number of individuals showing a significant preference. Some significant individual preferences were found during feeding in this study; however, these results differ from Byrne and Byrne’s (1991) study, where the majority of individuals showed a significant preference for food processing tasks and individuals showed consistent hand preferences across tasks.

Influence of age and sex

We examined the effect of age on handedness to investigate whether it may be a learned rather than innate behaviour. Some age-related effects have been found in nonhuman apes. Studies of wild chimpanzee populations have found that immature individuals show more consistent hand preferences than adults (Boesch 1991; McGrew and Marchant 1992). In contrast, there is also limited evidence that the strength of preference increases in adult chimpanzees (Hopkins 1994) but no evidence for an effect of age in bonobos (Harrison and Nystrom 2008). No age-related hand differences have been recorded in gorillas (Fagot and Vauclair 1988; Byrne and Byrne 1991; Parnell 2001). In this study, we found no noticeable differences in hand preferences between immature and adult subjects for any of the tasks studied, although the small sample size of immature subjects (N = 4) makes this a very tentative conclusion. More data are required before we can make firmer conclusions about the effect of age on handedness in gorillas.

A review by Hopkins (2006) found no significant effect of sex on handedness in any great ape species, and previous research on gorillas has not revealed any significant sex-related differences (Annett and Annett 1991; Byrne and Byrne 1991; Parnell 2001; Hopkins et al. 2003). Even though we could not determine a significant difference between males and females in this study due to small sample size, especially for the males (N = 6), we did find a trend. The males that showed a significant preference during any of the behaviours always favoured the left hand, whereas more females tended to show a preference for the right hand. If these results can be replicated in a larger population, this would be the same pattern as that shown in humans (e.g. Annett 2002).

Influence of posture and grip

It has been suggested that posture has an effect on handedness and that as posture becomes more unstable, lateralisation increases as a compensatory measure (de Vleeschouwer et al. 1995). The use of a particular posture also raises the issue of which limbs are being used for support and thus cannot be used for other activities (McGrew and Marchant 1997). In arboreal situations, one forelimb is frequently required to provide some support due to the increased risk from falling (McGrew and Marchant 1997); consequently, this may have instigated an advantage towards hand specialisation (see also MacNeilage et al 1987 for a discussion of the potential role of posture in handedness). Olson et al. (1990) found no population- or individual-level preferences when gorillas were feeding from the ground; however, a population-level right-hand preference was found when feeding in an unstable tripedal arboreal posture. In wild lowland gorillas, Parnell (2001) found that hand preference was not significantly altered by posture, though more individuals were significantly lateralised when standing rather than when seated. Results from the study presented here are inconsistent with previous research. In this study, more individuals showed a significant preference when sitting than in any other posture. It is difficult to speculate why and warrants further investigation. However, it may be explained by the fact that unequal numbers of bouts were recorded. That is, individuals did not necessarily have equal numbers of bouts for each behaviour pattern represented within each category of posture and grip, which consequently could explain the lack of consistency with previous research.

Previous research on wild gorillas produced varying results as to the effect of grip and the necessity for fine manipulation on handedness. In Byrne and Byrne’s (1991) study, most individuals showed a significant hand preference for tasks in which fine manipulation was required. However, Parnell (2001) found no evidence to suggest that fine processing tasks had an effect on handedness (Parnell 2001). Grip has been shown to influence handedness in other species of ape. In an extensive study of grips in young captive chimpanzees, Jones-Engel and Bard (1996) found significant individual-level (but not population-level) preferences when precision and power grips were used but not during imprecise grips. Hopkins et al. (2005b) found preferential use of the right hand in chimpanzees when grasping objects between the thumb and index finger but not when other types of grip were used. In the this study, there were few significant individual preferences for any grip used. However, those that occurred did so when a precision (N = 4) or power (N = 4) grip was employed, i.e. no significant preferences were found when a miscellaneous grip was used. This concurs with Byrne and Byrne’s (1991) study and findings on chimpanzees in which significant preferences were found in more complex grips in terms of motor demands. Consequently, there is some evidence that hand preferences increase in nonhuman apes in grips that demand complex movements (Jones-Engel and Bard 1996; Christel et al. 1998; Hopkins et al. 2005b).

Classification within McGrew and Marchant’s (1997) system

McGrew and Marchant (1997) established a five-level system, later modified by Fletcher and Weghorst (2005), for classifying results of studies on handedness in a comparative framework. When the results of individual data from this study were placed into McGrew and Marchant’s five-level system, leading limb (hand, foot), scratch, carry/object manipulation and gesture all showed level 1 laterality, i.e. the majority of individuals were ambipreferent and the minority were lateralised to either side to various degrees. Feeding showed level 241 laterality, i.e. 41% of the population was significantly lateralised but with no combined distribution towards left or right. Tool use showed level 240 laterality, i.e. 40% of the population was significantly lateralised but with no overall bias towards left or right.

Harrison and Nystrom (2008) conducted a study of bonobos performing the same behaviours recorded in this study. Bonobos showed level 1 laterality for leading limb (hand, foot) and scratch. Gesture shows level 229, carry/object manipulation level 238, feeding level 264 and tool use level 258. In each of these cases, there was no overall bias towards left or right. Whereas there is no evidence for species-level handedness in bonobos, individual preferences are found in more behaviours studied than in this study of gorillas. Also, a higher proportion of individuals is significantly lateralised.

Results from chimpanzees are more variable but again show a higher level of lateralisation than the gorillas recorded in this study. Wild chimpanzee populations tend to show level 1 laterality (McGrew and Marchant 2001). Naturalistically housed populations show various levels of preference: level 1 laterality for most nontool use tasks; level 3 for a spontaneous bimanual gesture (clapping), i.e. most individuals show exclusive hand preference but no population-level preference; and level 4a for tool use, i.e. half of the population is significantly lateralised, with a significant bias in this case towards the left (Fletcher and Weghorst 2005; Fletcher 2006). In contrast, captive chimpanzees show level 4b laterality for a bimanual feeding task, i.e. significant population-level laterality to the right (Hopkins et al. 2001, 2004, 2005a). There are various theories as to the reason for these discrepancies between studies, notably, differences in sample sizes and ypes of measures or methods used (Hopkins 2006). There are often substantial differences between the types of tasks recorded in wild and captive studies. Studies of wild populations tend to record all naturally occurring behaviours (e.g. Marchant and McGrew 1996; McGrew and Marchant 2001), whereas in studies of captive populations, there is often one complex, artificial feeding behaviour being recorded (e.g. Hopkins et al. 2001, 2004, 2005a). It is difficult to compare the results of studies that have recorded very different behaviours using different methods. Comparisons between studies of captive and wild gorillas are equally problematic. In Byrne and Byrne’s (1991) study, most individuals showed significant lateralisations though the population did not. In contrast, in Parnell’s (2001) study, the majority of individuals were not lateralised, which concurs with the results of this study.

In conclusion, captive gorillas show no evidence for species-level handedness. When placed into McGrew and Marchant’s (1997) 5-level classification system, all behaviours showed level 1 laterality, i.e. ambipreference, with the exception of feeding and tool use, which were level 2. There was no evidence that handedness was affected by age or sex, though the sample sizes of the various subgroups was small. These results concur with the results of studies of wild populations of mountain gorillas in which individual preferences were found during complex feeding tasls (Byrne and Byrne 1991; Parnell 2001). This field of research would benefit greatly from further data on captive and wild populations of gorillas performing a range of behaviours and including any possible effects of age, sex and posture. It would also benefit from further investigation into the implications of individual-level preferences. These data support the theory that species-level handedness is unique to humans.

Acknowledgments

This research was supported by grants from the Wenner-Gren Foundation for Anthropological Research (Grant number Gr. 6581) and the Leakey Trust (UK). We are extremely grateful to the personnel at Berlin, Stuttgart, Chessington, Edinburgh and London zoos for granting permission to study the subjects and for all of their assistance. We would also like to thank the two reviewers for their constructive comments and William McGrew for his encouragement and advice.

Copyright information

© Japan Monkey Centre and Springer 2010