The Cephalopod Page Home
Subscribe to the Ceph Group

Ceph Mailing Groups

Choice of arm usage in a Giant Pacific Octopus, Enteroctopus dofleini (Wülker, 1910), is based on relative proximity to object of interaction

<< Cephalopod Articles | By

Abstract

The choice of arm usage in an individual adult male Enteroctopus dofleini was observed, recorded and analyzed. It was hypothesized that the octopus would use the arm that was closest to an object of interaction, and not show a preference for a particular arm as has been found in most other animal species. Although preliminiary, the results supported the hypothesis. Further study in this area will yield insight into evolution of handedness in animals.

Introduction

The natural world is filled with examples of lateralization. It can be recognized at several different levels ranging from non-living physical properties like the spins of sub-atomic particles, and the draining of water, up to those exhibited in living organisms. Objects and organisms that are in some way asymmetrical in two dimensions must be oriented differently. This difference may be spatial, as in an asymmetric object; relative, as in spirals which are laterally asymmetrical but obliquely radially symmetrical; or temporal, as seen when an organism interacts with its environment. In organisms, lateralization is evident in the orientation of structures of chemicals like the direction of spiral in DNA and RNA, of structures that are a larger part of an organism like the enlarged claw of the fiddler crab or the direction of spiral in the narwhal tooth, or in a larger aspect affecting the orientation of the entire organism. Examples of this last category include plants, where bindweed twirls right and honeysuckle left; and animals where the shape of many mollusk shells, and flounder are oriented preferentially in one direction. Lateralization can also be applied to behavior, as in the preference of human beings to usually use the right hand for most tasks (Annett, 1985).

This preference in using a structure on one side of the body over the one on the other side has been thoroughly examined in many different animal species in the interest of comparative biology, evolutionary history, and psychology. However, the choice of species for such interests has unfortunately been skewed phylogenetically. The subjects are often mammalian. In vertebrates, lateralization of neural control of vocalization has been studied in the frog (Rana pipiens), many species of birds, and mammals (monkeys, dogs, and humans). In these studies most species show a dominance of the left hemisphere of the brain in processing sound information (Rogers and Bradshaw, 1996). Circling and turning behavior has also been investigated. A preference for one side has been found when avoiding predators and displaying to mates in fishes, and when circling the perimeter of cages in dolphins and rats. In addition, impala, cows, and whales show preference in using one side for various activities (Rogers and Bradshaw, 1996). Limb use has been investigated intensively. "Pawedness" has been found in rodents and birds (Rogers and Bradshaw, 1996; Annett, 1985), as well as in apes, monkeys, and cats (Annett, 1985). Many species of birds and rodents have also been studied in depth to determine hemisphere responsibility for various actions (Rogers and Bradshaw, 1996). It is important to note that, in most species studied, a preference was often found in an individual, but rarely was that preference pervasive throughout that individual's population or species. Usually, within one group, some individuals would prefer to use one side, some would prefer the other, and some would be ambivalent. Human beings are one of the very few exceptions that consistently show a preference to use a specific side, the right, over the other (Rogers and Bradshaw, 1996).

Cephalopods (Sanders, 1975), especially octopuses (Wood, 1999), have attained a level of sensory and neural development that facilitates the performance of highly complex adaptive behaviors. This fact makes them the most intelligent of invertebrates. One of the most intensely studied facets of these animals is their behavior. They have been trained to grab only certain colored balls (Pechenik, 1996), to navigate through a maze (Moriyama, 1997), and to open a jar (Wood, 1999). They have also been observed watching others of their own kind, and learning from them by example (Fiorito, 1992). General observations have led some to believe that octopuses are 'handed' in the way that humans are (Hemdal, 2000, personal communication). Also, Elliot and Roy (1996) make the statement "All animals show a preference for using one limb in activities such as feeding". However, although a large amount of research has examined octopus arm movement (Gutfreund, et al., 1996; Mather, 1998), no scientific data supporting arm usage preference has ever been published for an octopus (Mather, 2000, personal communication).

Mather (1998) suggests that since all of the arms are structurally similar, any arm can be used for any task, but that this equal potential probably is not related to usage frequency. She also cited an unpublished manuscript by Faulkes (1988) that stated that some species of octopuses may use their arms with differential frequency while performing certain coordinated activities. This would be expected, as different arms would be assigned to different tasks in order to perform an integrated operation. The current analysis attempts to determine whether an individual octopus uses particular arms to reach for an object, resulting in a frequency different than if it simply reached with the arm nearest to the object. Of particular interest, it tests whether a male octopus uses the third arm on its right side less frequently than would be expected if the octopus simply uses arms based on proximity to the item of interest. The third arm in males, termed the hectocotylus, is modified to deliver spermatophores to a female when mating (Mather, 1998). This modification possibly makes it less useful in ordinary manipulations. The hypothesis for this test is delineated below:

H1= If an octopus desires to touch an object, it will reach out with whichever arm is nearest to the object.

H0= If an octopus desires to touch an object, it will reach out with particular arms more often than with other arms. This would require a preferred arm to reach into the vicinity of another, less preferred arm.

Due to the great phylogenetic distance between octopuses and organisms previously studied, knowledge of octopus arm choice would be extremely useful in better understanding the questionably universal nature of preference in arm usage. Is it simply a manifestation of necessity stemming from the nature of laterality, or is it a character that has been derived in relatively recent vertebrates? Cephalopods and vertebrates have been evolving separately for at least 500 million years. Their common ancestor probably had very little central nervous tissue. Convergently, there exist similarly detailed plans of nervous systems between these two groups (and also arthropods). The brains of octopuses are divided into lobes that are allocated to specific functions similarly to the lobes of vertebrates. This is probably because similar environmental conditions (light, gravity) require basic organizational structures to negotiate them. These lobes have evolved relatively recently in octopuses (about 10 million years ago), and are arranged differently in cuttlefish and squid. This indicates that if arm usage preference does exist in octopuses, then it may be a necessity of laterality, another example of convergence. However, the nervous systems of octopuses are less centralized than those of vertebrates. Three hundred of an octopus's 500 nerves are in its arms. Many of these arm nerves are computational rather than simply motor nerves. In fact, it has been shown that it partially thinks with its arms. A severed arm placed in close vicinity of food will reach out (sensing it chemically), grab the food, and move it in the direction of where the animal's mouth used to be (Young, 1964). This different aspect of an octopus's CNS implies that it might be possible that preference in arm usage is non-existent in octopuses.

Materials and Methods

Protocol

A series of interactions was performed to determine if an individual adult male Enteroctopus dofleini had an arm preference when reaching for objects. Two observations were made per week. Data were collected in this way for ten weeks, at the end of which time the animal was deemed ill by zoo staff and was no longer responding to stimulus.

Each Monday data was obtained by observing the octopus's interactions with a 'training stick'. This stick was constructed from a piece of PVC pipe approximately four feet long with four rows of small holes drilled down its length to make it porous. Glued on the interaction end of the stick was a red puck composed of smooth plastic. The stick was presented to the animal and its choice of arm used to reach for the stick was recorded on a count sheet (Martin and Bateson, 1993). The arms were designated as given by Wells (1978).

A similar procedure was also employed each Wednesday. In this case, a bamboo rod of similar size and shape to the training stick was used to offer food to the animal. This rod was the same object that was usually used to feed the octopus on Wednesdays and Saturdays, its normal feed days. Feeding continued as established with arm usage data being recorded during Wednesday feedings. A piece of food was pierced on the end of the rod and then the rod was submerged into the aquarium. Data were recorded each time the octopus grabbed food off of the feeding stick. The food offered included shrimp, capelin, and squid, and didn't differ from the regularly offered amounts or proportions. With both the training stick and the feeding stick, a data point recorded which arm was in the general vicinity the object when it was offered, as well as which arm reached to grab the object. If it appeared that two arms were reaching with equal intent, the stick placement was recorded as two separate instances, one representing the grab by each arm.

Data Analysis

A program was written in Pascal TURBO5 to analyze the data. Places were reserved in the computer's memory to store data. These can be seen in the VAR block below:

VAR {put} {grab}
putgrab: ARRAY[1..8,1..8] OF INTEGER;
simputgrab: ARRAY[1..8,1..8] OF INTEGER;
put: ARRAY[1..8] OF INTEGER;
grab: ARRAY[1..8] OF INTEGER;
grabtotal: ARRAY[1..9] OF INTEGER;
puttotal: ARRAY[1..9] OF INTEGER;
i,k,q: INTEGER;
ns: INTEGER;
sighi, siglo :ARRAY[1..8,1..8] OF INTEGER;
p : INTEGER;
g : INTEGER;
nop : INTEGER;
datafile: TEXT;
outfile: TEXT;

'p' represented the vicinity of the arm where the stick was placed. 'g' represented which arm reached out to grab the stick. 'putgrab' was the table that held the numbers of times each particular arm reached out to grab the stick when it was placed in the vicinity of a particular arm. 'simputgrab' was the computer's simulation of 'putgrab'. 'puttotal' was the cumulative number of times the stick was placed near a particular arm and 'grabtotal' was the cumulative number of times each particular arm grabbed the stick. The first step was to rearrange the data into a form that would be most useful when importing into TURBO5. This was accomplished by combining the data from all of the count-sheets onto one table in the form of a text file (Appendix A). This data was then read into the program and analyzed.

The analysis of arm usage was examined by two separate methods. Both of these simulated placing the stick into the aquarium and then simulated the octopus grabbing it.

Test 1:
The first analysis was accomplished by simulating stick placement in the vicinity of each arm with the same frequency as that with which it was placed near that arm in the dataset of actual interactions. For each simulated stick placement, an arm was chosen equiprobably to grab the stick. This was accomplished with the following Pascal code:

{simulate data table 1000 times}
nop := 37;

FOR ns := 1 TO 1000 DO BEGIN
FOR p := 1 TO 8 DO FOR g := 1 TO 8 DO simputgrab[p,g] := 0;
FOR i := 1 TO nop DO BEGIN
p := 1; {simulate putting the stick down}
WHILE i > puttotal[p] DO p := p + 1;
g := RANDOM(8) + 1; {simulate grabbing stick}
simputgrab[p,g] := simputgrab[p,g] + 1;
END; {FOR i}

Test 2:
For the second analysis the arm used to grab the stick was not chosen equiprobably. Instead, the chance of any arm being chosen in a simulated grab was dictated by the total frequency in which that arm was used in the original interaction. This was performed in Pascal TURBO 5 by:

{simulate data table 1000 times}
FOR ns := 1 TO 1000 DO BEGIN

FOR p := 1 TO 8 DO FOR g := 1 TO 8 DO simputgrab[p,g] := 0;
FOR i := 1 TO nop DO BEGIN
p := 1; {simulate putting the stick down}
WHILE i > puttotal[p] DO p := p + 1;
k := RANDOM(nop) + 1; {simulate grabbing the stick}
g := 1;
WHILE k > grabtotal[g] DO g := g + 1;
simputgrab[p,g] := simputgrab[p,g] + 1;
END; {FOR i}

As shown in Appendix A, the total number of actual stick placements was 37. For each test, these 37 stick placements were each simulated 1000 times. Since these placements were simulated in the exact same arm vicinities as in the actual data set, the set of placement vicinities was the independent variable. The dependant variable was the number of grabs per arm. For each arm, the observed number of grabs was compared to the number of grabs in each of the 1000 simulations.

Results
The number of simulated grabs out of the total 1000 that were equal or above the observed number for Test 1 can be seen in Table 1. The number of simulated grabs out of the total 1000 that were above the observed number for Test 2 can be seen in Table 2.

 

 

Puts

 

 

R4

R3

R2

R1

L1

L2

L3

L4

Grabs

R4

251

1000

1000

1000

1000

1000

1000

1000

R3

246

2

1000

1000

1000

1000

1000

1000

R2

1000

500

7

1000

1000

1000

1000

1000

R1

1000

1000

397

104

1000

1000

1000

1000

L1

1000

1000

1000

1000

0

707

665

1000

L2

1000

1000

1000

1000

1000

0

253

1000

L3

1000

1000

1000

1000

1000

700

12

1000

L4

1000

1000

1000

1000

1000

1000

649

139

Table 1. TEST 1: number of times out of 1000 that number of equiprobably
simulated grabs with a particular arm equaled or exceeded observed number
of grabs when the stick was put near a particular arm.


 

 

Puts

 

 

R4

R3

R2

R1

L1

L2

L3

L4

Grabs

R4

49

1000

1000

1000

1000

1000

1000

1000

R3

257

2

1000

1000

1000

1000

1000

1000

R2

1000

443

5

1000

1000

1000

1000

1000

R1

1000

1000

206

57

1000

1000

1000

1000

L1

1000

1000

1000

1000

0

921

884

1000

L2

1000

1000

1000

1000

1000

2

616

1000

L3

1000

1000

1000

1000

1000

733

15

1000

L4

1000

1000

1000

1000

1000

1000

354

56

Table 2. TEST 2: number of times out of 1000 that number of proportionally
simulated grabs with a particular arm equaled or exceeded observed number
of grabs when the stick was put near a particular arm.


Discussion

Test 1:
As can be seen in Table 1, the octopus tends not to use an arm that is not in the immediate vicinity of an object if it wishes to grab the object. In most cases, a significantly small number (50/1000 = .05) of equiprobably simulated grabs exceeded the number of actual grabs made by the arms that were closest to the interaction object. The simulated values for the most posterior arms were high compared to the observed grabs. This is possibly due to the fact that the animal had been used to accepting food from the stick and when the lid was opened it positioned itself facing upward toward the person opening the lid. Therefore, the vicinities of the anterior arms were more accessible to the stick with than those posterior, which were positioned beneath the octopus. This confounds the analysis and also led to a smaller sample size being obtained for the posterior arms. It is interesting to note, however, that, in this species, the more posterior the arms, the smaller they are. Possibly if this positioning situation were remedied this species would still show less frequent usage of the more posterior arms.Also, there is an insignificant usage of R1. This is probably due to the fact that throughout the entire experiment, the stick was placed in its vicinity only one time. This small sample size is probably responsible, and the insignificance is an artifact.

Test 2:
When the data is analyzed by simulating grabs of specific arms proportionally to the number of times they were actually used, the insignificance of the posterior arms falls out (Table 2). This indicates that when you consider the infrequency of how often those arms are used, the other arms don't seem to compensate as much. It is interesting to note that in all simulations the hectocotylus was utilized just as frequently as would be expected if R3 were structurally normal. The added altered morphology must not impair its operation at all.

In continuing research, three alterations from the present protocol are recommended. The first one that is suggested here is to utilize smaller animals in larger enclosures. If an interaction object is placed a greater distance from the octopus it would allow for greater access to the different arms of the octopus, then any contaminating variables stemming from the relative position of the octopus should be removed. This will give desired sample sizes for all arms.Annett (1985) states that feeding requires so little skill that humans often use the non-preferred hand. It is unknown at this point whether this is the case in octopuses or not. One way to alleviate influence from this possibility would be to use a more complex interaction from which to record data. Grabbing a stick is a very simple action. A more complex action, like opening a jar (Wood, 1999), might yield different results than the current study.

Finally, it is impossible to conclude from this study that octopuses in general do not exhibit lateralization do to the fact that only one individual was observed. In fact, in most species examined in previous studies, a considerable proportion of individuals have been ambidextrous. Lack of examining a substantial number of individual animals could very easily yield misleading results. Any future study should include a larger number of individuals.

Although the current study analyses arm usage preference, it does not look at side preference. Lateralization of behavior might exist but not be evident in arm usage analysis. Future study should also identify if octopuses preferentially use arms on a particular side when interacting with objects that are not in the immediate vicinity of any particular arm or side.

Larger sample size and side usage preference would make it possible to identify any existing generalization toward brain hemisphere - arm side correlation. If a large proportion of individuals sampled from a population show a consistent bias in the direction of lateralization, then hemispheric specialization is indicated (Rogers and Bradshaw, 1996).

The methods employed here lay out the groundwork for this system to be further studied. If this established protocol is followed, and the recommendations are heeded, then future studies will doubtlessly realize the tendencies that octopuses show when making choices regarding arm usage. The interpretation of this information will not only yield a better understanding of octopus behavior, but also of the evolution and questionable necessity of arm usage preferences and lateralization.

Acknowledgements

I would like to thank Jay Hemdal for granting me access to the octopus. Beth Stark served as an advisor on the experimental design. Ernest Dubrul made the program possible through the University of Toledo. George Estabrook advised me on the trying task of writing the computer programs that analyzed the data.

Appendix A

data totals
object placed in the vicinity of:
R4 R3 R2 R1 CENTER L1 L2 L3 L4 AWAY right side left side total
grabbed with R4 1 1 2
R3 1 4 5 10
R2 1 3 3 10 17
R1 1 1 9 5 1 17
L1 11 7 1 1 5 25
L2 2 7 2 1 1 6 19
L3 1 4 1 5 11
L4 2 1 1 1 5
total 2 5 4 1 27 7 9 8 1 2 22 18 106
data totals
object placed in the vicinity of:
R4 R3 R2 R1 L1 L2 L3 L4 total
grabbed with R4 1 1
R3 1 4 6
R2 1 3 10
R1 1 1 12
L1 7 1 1 21
L2 7 2 30
L3 1 4 35
L4 1 1 37
total 2 5 4 1 7 9 8 1 37

Literature Cited

Annett, Marian. 1985. Left, Right, Hand and Brain: The Right Shift Theory. Lawrence Associates, London.
Fiorito, Graziano; and Pietro Scotto. 1992. Observational learning in Octopus vulgaris. Science. Vol. 256(5056), Apr 1992, 545-547.
Gutfreund, Yorum; Flash, Tamar; Yarom, Yosef; Fiorito, Graziano; Segev, Idan; Hachner, Binyamin. 1996. Organization of octopus arm movements: A model system for studying the control of flexible arms. The Journal of Neuroscience. 16(22): 7297-7307.
Hemdal, Jay. 2000. Personal communication. Curator, Toledo Zoo aquarium.
Martin, Paul and Patrick Bateson. 1993. Measuring Behaviour, an introductory guide. 2nd ed. Cambridge Univ. Press, Cambridge, UK.
Mather, Jennifer A. 1998. How do octopuses use their arms? Journal of Comparative Psychology. Vol. 112(3), Sep 1998, 306-316.
Mather, Jennifer A. 2000. Personal Communication.
Moriyama, Tohru; and Yukio-Pegio Gunji. 1997. Autonomous learning in maze solution by Octopus. Ethology. Vol. 103(6), Jun 1997, 499-513.
Rogers, Lesley J. and John L . Bradshaw. 1996. Motor asymmetries in birds and nonprimate mammals. In: Manual Asymmetries in Motor Performance. Digby Elliot and Eric A. Roy eds. CRC Press. Boca Raton.
Sanders, G.D. 1975. Chapter 11: The cephalopods. In Corning, Dyal, and Willows (editors): Invertebrate Learning, Vol. 3: Cephalopods and Echinoderms. Plenum Press, New York.
Wells, M J. 1978. Octopus: Physiology and behavior of an advanced invertebrate.Methuen, London.
Wood, James B. 1999. Enrichment for an advanced invertebrate. The Shape of enrichment. Vol. 8, No.3, August 1999.
Wood, James B. 2000. Personal communication.
Young, J.Z. 1971. A model of the brain. Oxford University Press, London.


» What's New?
» Cephalopod Species, Information, and Photographs
» Articles on Octopuses, Squid, Nautilus and Cuttlefish
» Cephalopod Lesson Plans by Wood, Jackson and Amity High School Teachers
» The Cephalopod Page F.A.Q.
Resources
CephBase Cephalopod database by Wood, Day and O'Dor
Upcoming Conferences
Sources of Live Cephalopods
Cephalopod Links
Want to learn more about Cephalopods?
References and Credits
Home

The Cephalopod Page (TCP), © Copyright 1995-2014, was created and is maintained by Dr. James B. Wood, Associate Director of the Waikiki Aquarium which is part of the University of Hawaii. Please see the FAQs page for cephalopod questions, Marine Invertebrates of Bermuda for information on other invertebrates, and MarineBio.org and the Census of Marine Life for general information on marine biology.