Anim Cogn (2012) 15:591–596

DOI 10.1007/s10071-012-0487-9

ORIGINAL PAPER

CuttleWsh rely on both polarized light and landmarks for orientation

Lelia Cartron · Anne-Sophie Darmaillacq · Christelle Jozet-Alves · Nadav Shashar · Ludovic Dickel

Received: 22 July 2011 / Revised: 15 March 2012 / Accepted: 15 March 2012 / Published online: 30 March 2012© Springer-Verlag 2012

Abstract CuttleWsh are sensitive to linear polarization oflight, a sensitivity that they use in predation and possibly inintraspeciWc communication. It has also been shown thatcuttleWsh are able to solve a maze using visual landmarks.In this study, cuttleWsh were trained to solve a Y-maze withthe e-vector of a polarized light and landmarks as redundantspatial information. The results showed that cuttleWsh canuse the e-vector orientation and landmarks in parallel to ori-ent and that they are able to use either type of cue when theother one is missing. When they faced conXicting spatialinformation in the experimental apparatus, the majority ofcuttleWsh followed the e-vector rather than landmarks.DiVerences in response latencies in the diVerent conditionsof testing (training with both types of cue, tests with singlecue or with conXicting information) were observed and dis-cussed in terms of decision making. The ability to use nearWeld and far Weld information may enable animals to inter-pret the partially occluded underwater light Weld.

Keywords Navigation · Polarized light · Landmarks · Cephalopods

Introduction

Navigation can be deWned as the process which enables acourse or path from one place to another to be identiWed

and maintained (Gallistel 1990). Spatial orientation allowsanimals to reach beneWcial places containing food, conspe-ciWcs or shelters, and to avoid risky places with predators.Hence, the ability to move from one place to another in aneYcient way is an important factor for survival. For thispurpose, animals can use a variety of spatial informationsuch as beacons, landmarks, overall spatial geometry, sunor magnetic compasses, celestial polarization and self-gen-erated cues for dead reckoning (reviewed in Rozhok 2008;Shettleworth 2010; see also Kraft et al. 2011 for honeybeenavigation using polarization). The way diVerent types ofspatial information interact has been extensively studiedaddressing spatial tasks (Healy 1998 and referencestherein). Historically, the theory deriving from Pavlovianconditioning was that multiple redundant spatial cues com-pete for associative learning so one cue can overshadow orblock learning about another one (Diez-Chamizo et al.1985; March et al. 1992; Redhead et al. 1997; Pearce et al.2001). However, it seems likely that the importance of spa-tial learning for survival favors the simultaneous acquisi-tion of redundant types of spatial information because itallows secondary cues to be used as backup when primarycues are unavailable (Shettleworth 2010). An increasingnumber of experiments support this idea in species rangingfrom insects (Wehner et al. 1996; Steck et al. 2011) torodents (Lavenex and Schenk 1996; Rossier et al. 2000;Gibson and Shettleworth 2003, 2005).

Spatial orientation in aquatic environments seems to beparticularly challenging since visual cues can undergomajor changes due to water turbidity and depth, and olfac-tory cues are highly fugitive due to currents or Xow direc-tions. Hence, it would be beneWcial to learn and memorizediVerent types of information and to orient according to thecues available under the speciWc environmental conditions.Many species of crustacean, Wsh, and sea turtles can rely on

L. Cartron · A.-S. Darmaillacq · C. Jozet-Alves · L. Dickel (&)Groupe Mémoire et Plasticité comportementale, Université de Caen Basse-Normandie, EA 4259, 14032 Caen cedex, Francee-mail: ludovic.dickel@unicaen.fr

N. ShasharDepartment of Life Sciences, Ben Gurion University of the Negev, Eilat Campus, Beer-Sheva, Israel

123

592 Anim Cogn (2012) 15:591–596

a large variety of spatial information to orient such asgeomagnetic Weld, celestial and polarized light cues, oceancurrents, olfactory, and other visual cues (Healy 1998;Odling-Smee and Braithwaite 2003; Parkyn et al. 2003;Horvath and Varju 2004; Lohmann et al. 2008; Shettle-worth 2010). However, far fewer studies have investigatedhow diVerent kinds of spatial information interact in a con-trolled spatial task.

CuttleWsh (Mollusca: Cephalopoda) are colorblind(Mäthger et al. 2006) but are sensitive to the e-vector orien-tation of linearly polarized light, a sensitivity which arisesfrom the orthogonal distribution of the microvilli of neigh-boring photoreceptor cells in their retina (Shashar et al.1996). In the cuttleWsh, polarization sensitivity is involvedin predation (Shashar et al. 1998, 2000) and possibly inintraspeciWc communication (Shashar et al. 1996; Boalet al. 2004). In a range of animals, polarization sensitivity isinvolved in orientation and may well serve for underwaternavigation (Sabbah et al. 2005; Lerner et al. 2011 and refer-ences therein). CuttleWsh are known for their remarkablelearning abilities (Messenger 1973; Darmaillacq et al.2004, 2006), notably in spatial learning where they havebeen shown to solve a maze using landmarks (Alves et al.2007; Jozet-Alves et al. 2008). However, their ability to usean e-vector orientation for a spatial task has never beeninvestigated.

The Wrst aim of the study was to determine whether cut-tleWsh could use the e-vector of polarization to Wnd a shel-ter. The second aim was to test whether cuttleWsh were ableto learn to Wnd a shelter using e-vector orientation and land-marks in parallel and then to rely on one type of cue whenthe other one became unavailable. Last, we addressedwhich type of cue was preferentially followed when cuttle-Wsh faced conXicting information in the experimental appa-ratus.

Materials and methods

Subjects

Eleven cuttleWsh, Sepia oYcinalis, were tested. They werereared from hatching to 7 months in laboratory conditionswith running seawater at 15 § 1 °C at the Centre deRecherches en Environnement Côtier (CREC, Luc-sur-Mer, France). They were Wrst housed in groups and then,1 week before behavioral experiments began, housed inindividual tanks. The animals were provided with enriched(or “semi natural”) habitats within the tank followingDickel et al. (2000) who showed that an enriched environ-ment has a positive eVect on cuttleWsh growth rates and theacquisition and the retention of information. They were fed

daily with live shrimps (Crangon crangon) and crabs (Car-cinus maenas) of suitable size.

Apparatus

Experiments were conducted in a Y-maze (Fig. 1) made ofspectrally Xat white PVC to reduce potential brightnessartifacts, which could arise from the interaction of thepolarized light source and the inside walls of the tank (Jan-der and Waterman 1960). One arm (30 cm long £ 10 cmwide £ 15 cm high) was used as a start box closed by asliding glass door. The other two arms (goal arms) eachended with a shelter (20 cm long £ 20 cm wide £ 15 cmhigh) which were closed by sliding glass doors. The shel-ters were made of black PVC and were covered by anopaque top, so as to be attractive to the cuttleWsh. Twoblack-and-white PVC rectangles (4 cm wide £ 8 cm high),a striped rectangle (horizontal stripes 2 cm wide), and aspotted rectangle (spots 1 cm in diameter) were used aslandmarks. They were placed 5 cm above the water surfaceat the junction of the Y (Fig. 1; Alves et al. 2007; Jozet-Alves et al. 2008). They could be swapped in position orremoved. Black opaque curtains were placed around andabove the maze to prevent cuttleWsh from seeing externalcues and light from the laboratory. The apparatus was illu-minated by a 300-W halogen lamp located 1 m above thecentre of the maze. The lamp was enclosed in a black baZe(25 cm long £ 25 cm wide £ 35 cm high), the bottom ofwhich was a layer of commercial wax paper above a linearpolarizing Wlter (American Polarizers, Inc. AP38-030T).The wax paper was used as a light diVuser to remove poten-tial inherent polarization in the optical system. The Wlterpolarized the light along a pre-set e-vector axis. This Wltercould be rotated to orient the e-vector parallel to one or theother arm (and perpendicular to the other) and could beremoved to illuminate the maze with a nonpolarized light.There was no diVerence in light intensity between the twogoal arms regardless of the e-vector orientation (300 luxmeasured with a luxmeter, Hanna Instruments). Water Xowwas provided between trials to reduce water heating andremove olfactory cues.

Procedure

The procedure consisted of three steps (Fig. 1). First, cuttle-Wsh were trained to Wnd a shelter (the reward; see Alveset al. 2007). A combination of a landmark and an e-vectororientation was randomly chosen for each cuttleWsh toindicate the goal arm (e.g., the black-and-white spottedrectangle and the e-vector perpendicular to the longitudinalaxis of the goal arm). The sliding glass door at the entranceof the shelter was opened only if the cuttleWsh made the

123

Anim Cogn (2012) 15:591–596 593

correct choice. The goal arm (i.e., correct choice) for eachtrial was randomly chosen, and the cues were changedaccordingly between trials (rotation of the linear polarizerand permutation of the right/left position of the landmarks).CuttleWsh were given Wve trials per session, with one train-ing session per day, until they reached the learning criterion(i.e., eight correct choices out of 10, signiWcant choice at10 % level of signiWcance with a binomial test; Jozet-Alveset al. 2008).

In step 2, cuttleWsh were given two single cue tests, eachcomposed of two consecutive sessions (Fig. 1). In thee-vector single cue test, cuttleWsh had to Wnd the shelterusing only the e-vector orientation, whereas in the land-marks single cue test, cuttleWsh had to Wnd the shelter usingonly the landmarks. The single cue tests were conductedto determine whether the cuttleWsh had learned to use thee-vector orientation, the landmarks, or both, to Wnd theshelter. The order of single cue tests was randomly chosenfor each cuttleWsh. Each single cue test was followed bytwo training sessions with both types of cue.

In step 3, cuttleWsh were given one probe trial with con-Xicting information: the e-vector orientation indicating onearm as rewarded while the landmarks indicated the otherarm as rewarded.

Statistical analyses

Data were analyzed using StatXact 7 (Cytel Studio soft-ware). All analyses used a signiWcance threshold of� = 0.05 and the tests were two-tailed. In step 1, the per-centage success between the Wrst and the last session wascompared with a permutation exact test for paired samples.In steps 2 and 3, the mean latencies to choose the goal armwere compared using an exact permutation test for pairedsamples. In step 3, the choice for the e-vector orientation orthe landmarks was compared using the exact Chi-squaretest.

Results

In step 1, all cuttleWsh (n = 11) reached the learning crite-rion in 6–12 sessions (i.e., 30–60 trials), with amean § SEM of 8 § 1 sessions. Percentage success(mean § SEM) was signiWcantly higher during the last ses-sion for each cuttleWsh (91 § 3 %) than during the Wrstsession (35 § 7 %; exact permutation test for paired sam-ples: n = 11, P = 0.002; Fig. 2). It made no diVerencewhich e-vector orientation and which landmark signaledthe goal arm.

In step 2, one cuttleWsh died after the e-vector single cuetest due to unrelated reasons so that the Wnal sample size forthe landmarks single cue test was 10. All cuttleWsh suc-ceeded in choosing the goal arm with only the e-vector ori-entation (n = 11, mean percentage success § SEM =87 § 2 %) and with only the landmarks (n = 10; 88 § 2 %;Fig. 2). For the cuttleWsh that were given the e-vector singlecue test Wrst (n = 5), and those that were given the landmarks

Fig. 1 a Y-maze apparatus. e-vector: illumination of the maze fromabove with a light polarized in a single e-vector axis (dashed line witharrow); landmarks: black-and-white spotted or striped rectangles; Ststart box; dashed line start box sliding glass door; Sh: shelter(=reward); dotted line shelter sliding glass door. b Chronology of theexperiments. Step 1: training with both the e-vector (light shaded) andthe landmarks (striped); it consists in one session (Wve trials) per dayuntil cuttleWsh reach the learning criterion. Step 2: single cue tests (twosessions) alternating with two sessions with the two types of cues. Halfof the cuttleWsh were presented with the e-vector single cue test Wrstand the other half with the landmarks single cue test Wrst. Step 3: probetrial with conXicting information between the e-vector orientation andthe landmarks (dark shaded)

Fig. 2 Mean percentages of correct choices per session (§SEM) dur-ing the Wrst and last sessions of training and during the single cue tests.Asterisks indicate signiWcant diVerences **P < 0.01; exact permuta-tion tests for paired samples

123

594 Anim Cogn (2012) 15:591–596

single cue test Wrst (n = 5), the latency to enter the goal armincreased, though not signiWcantly, in the Wrst trial of thesingle cue test compared with the Wrst trial of the last train-ing session (respectively: mean latency § SEM = 67 § 9 svs. 83 § 7 s and 55 § 15 s vs. 82 § 11 s; exact permutationtest for paired samples: P = 0.062 and P = 0.062; Fig. 3).

In step 3, eight out of 10 cuttleWsh followed the e-vectororientation and two followed the landmarks. They tended tofollow the e-vector rather than landmarks (exact Chi-squaretest, �² = 3.6, P = 0.058). The latency to enter the goal arm(mean § SEM) increased signiWcantly between the Wrsttrial of the training session (with both types of cue, per-formed just before the probe trial; 52.1 § 7.8 s) and theprobe trial (126.4 § 15.3 s; exact permutation test forpaired samples P = 0.007; Fig. 3). Furthermore, this latency(mean § SEM) increase (196.7 § 57.1 %) was signiW-cantly greater than the increase between the training and theWrst single cue test (55.6 § 19.3 %; exact permutation testfor paired samples P = 0.047).

Discussion

The aim of this study was threefold: (1) to determinewhether cuttleWsh could use the e-vector of polarized lightto Wnd a shelter; (2) to test whether cuttleWsh were able tolearn to Wnd a shelter using e-vector orientation and land-marks in parallel and then to rely on one type of cue whenthe other one became unavailable; (3) to test which type ofcue was preferentially followed when cuttleWsh faced con-Xicting information in the experimental apparatus. Here,we showed that cuttleWsh learned to solve a maze with twodistinct visual cues (step 1) but were able to solve it whenonly one type of cue was available (step 2). This demon-

strates that they are able to learn to use landmarks, asalready shown by Alves et al. (2007); this also shows forthe Wrst time that cuttleWsh can use an e-vector orientationto Wnd a shelter. To date, it was only known that polariza-tion sensitivity is used by cuttleWsh for predation (Shasharet al. 2000) and possibly in communication (Shashar et al.1996; Boal et al. 2004). Squid (Jander et al. 1963) exhibitspontaneous preferential swimming direction relative tothe e-vector orientation, like some species of crustaceans(Horvath and Varju 2004; Sabbah et al. 2005; Lerner et al.2011). CuttleWsh learn to use landmarks and the e-vectororientation in parallel. In the water, the ability to memorizeredundant spatial information at the same time when Wnd-ing a goal location should be advantageous since land-marks may not be visible from large distances due to waterturbidity or bottom topography. They should therefore beconsidered as near Weld visual cues. On the other hand,polarization information is based on the sun’s position inthe sky (far Weld). Thus, it may be altered, for example, bywaves or clouds (Horvath and Varju 2004; Sabbah et al.2005; Shashar et al. 2011). Luschi et al. (1997) demon-strated that the crab Dotilla wichmanni mainly relies onskylight polarization to Wnd food patches. However, whenthis polarization information is not available (e.g., under acloudy sky), the crab is able to switch to visual landmarkswhich can be more advantageous under those environmen-tal conditions.

During the probe trial, we observed that the latency tochoose an arm was signiWcantly greater under conXictinginformation than when the two types of cue were congru-ent. Additionally, the absence of one of the two types of cuein the maze induced a slower response which conWrms thatcuttleWsh used both landmarks and the e-vector orientationin step 1. The response was even slower when visual cues

Fig. 3 Mean latencies (§SEM) to enter the goal arm. a e-vector singlecue test Wrst (n = 5): comparison between the Wrst trial of the last train-ing session and the Wrst trial of each single cue test. b Landmarks sin-gle cue test Wrst (n = 5): comparison between the Wrst trial of the last

training session and the Wrst trial of each single cue test. c Comparisonbetween the Wrst trial of the last training session and the probe trial(n = 10). Asterisks indicate signiWcant diVerences ***P < 0.001; exactpermutation tests for paired samples

123

Anim Cogn (2012) 15:591–596 595

were in conXict (step 3), suggesting that cuttleWshresponded to the incongruence between the cues. Thechange of the response latency is what would be predictedif more complex cognitive processes were required beforethe cuttleWsh made a decision, though there may be otherexplanations. Nippak and Milgram (2005) suggested thatthe response latency in dogs was longer for more complexvisual discrimination tasks compared with simple tasks andreXected the cognitive strategy used by the animal. We alsofound that when cuttleWsh faced conXicting information inthe experimental apparatus (step 3), they tended to followthe e-vector orientation rather than the landmarks. Thedegree to which one cue rather than another controls behav-ior depends on their salience and their validity, that is, onboth the strength and the information content of the signals.Field measurements are needed to examine the relative sta-bility of diVerent cues available to cuttleWsh in the wild.Although little is known about cuttleWsh behavior in theirnatural habitat, we demonstrate here for the Wrst time thatcuttleWsh can use both polarized light and landmarks to Wndand return to speciWc sites, such as patches of food or safeplaces.

Acknowledgments We thank the staV of the CREC for their techni-cal assistance. This research was supported by a grant from the Minist-ère de l’Enseignement Supérieur et de la Recherche to L.C.

References

Alves C, Chichery R, Boal JG, Dickel L (2007) Orientation in the cut-tleWsh Sepia oYcinalis: response versus place learning. AnimCogn 10:29–36. doi:10.1007/s10071-006-0027-6

Boal JG, Shashar N, Grable M, Vaughan K, Loew E, Hanlon RT(2004) Behavioral evidence for intraspeciWc signaling with achro-matic and polarized light by cuttleWsh (Mollusca: Cephalopoda).Behaviour 141:837–861. doi:10.1163/1568539042265662

Darmaillacq A-S, Chichery R, Poirier R, Dickel L (2004) EVect of ear-ly feeding experience on subsequent prey preference by cuttleWsh,Sepia oYcinalis. Dev Psychobiol 45:239–244. doi:10.1002/dev.20034

Darmaillacq A-S, Chichery R, Shashar N, Dickel L (2006) Early famil-iarization overrides innate prey preference in newly hatched SepiaoYcinalis cuttleWsh. Anim Behav 71:511–514. doi:10.1016/j.an-behav.2005.04.019

Dickel L, Boal JG, Budelmann BU (2000) The eVect of early experi-ence on learning and memory in cuttleWsh. Dev Psychobiol36:101–110

Diez-Chamizo V, Sterio D, Mackintosh NJ (1985) Blocking and over-shadowing between intra-maze and extra-maze cues: a test of theindependence of locale and guidance learning. Q J Exp Psychol B37:235–253

Gallistel CR (1990) The organization of learning. The MIT Press,Cambridge

Gibson BM, Shettleworth SJ (2003) Competition among spatial cues ina naturalistic food-carrying task. Anim Learn Behav 31:143–159

Gibson BM, Shettleworth SJ (2005) Place versus response learningrevisited: tests of blocking on the radial maze. Behav Neurosci119:567–586. doi:10.3758/BF03195977

Healy S (1998) Spatial representation in animals. Oxford UniversityPress, Oxford

Horvath G, Varju D (2004) Polarized light in animal vision: polariza-tion patterns in nature. Springer, Berlin

Jander R, Waterman TH (1960) Sensory discrimination between polar-ized light and light intensity patterns by arthropods. J Cell CompPhysiol 56:137–159. doi:10.1002/jcp.1030560304

Jander R, Daumer K, Waterman TH (1963) Polarized light orientationby two Hawaiian decapod cephalopods. Z Vgl Physiol 46:383–394. doi:10.1007/BF00340466

Jozet-Alves C, Moderan J, Dickel L (2008) Sex diVerences in spatialcognition in an invertebrate: the cuttleWsh. Proc R Soc B275:2049–2054. doi:10.1037/0735-7044.119.2.567

Kraft P, Evangelista C, Dacke M, Labhart T, Srinivasan MV (2011)Honeybee navigation: following routes using polarized-lightcues. Phil Trans R Soc B 366:703–708

Lavenex P, Schenk F (1996) Integration of olfactory information in aspatial representation enabling accurate arm choice in the radialarm maze. Learn Mem 2:299–319. doi:10.1101/lm.2.6.299

Lerner A, Sabbah S, Erlick C, Shashar N (2011) Navigation by lightpolarization in clear and turbid waters. Phil Trans R Soc B366:671–679. doi:10.1098/rstb.2010.0189

Lohmann KJ, Lohmann CMF, Endres CS (2008) The sensory ecologyof ocean navigation. J Exp Biol 211:1719–1728. doi:10.1242/jeb.015792

Luschi P, Seppia CD, Crosio E (1997) Orientation during short-rangefeeding in the crab Dotilla wichmanni. J Comp Physiol A181:461–468

March J, Chamizo VD, Mackintosh NJ (1992) Reciprocal overshad-owing between intra-maze and extra-maze cues. Q J Exp PsycholB 45:49–63

Mäthger LM, Barbosa A, Miner S, Hanlon RT (2006) Color blindnessand contrast perception in cuttleWsh (Sepia oYcinalis) determinedby a visual sensorimotor assay. Vision Res 46:1746–1753

Messenger JB (1973) Learning in the cuttleWsh, Sepia. Anim Behav21:801–826

Nippak PM, Milgram MW (2005) An investigation of the relationshipbetween response latency across several cognitive tasks in thebeagle dog. Prog Neuropsychopharmacol Biol Psychiatry29:371–377

Odling-Smee L, Braithwaite VA (2003) The role of learning in Wsh ori-entation. Fish Fish 4:235–246

Parkyn DC, Austin JD, Hawryshyn CW (2003) Acquisition of polar-ized-light orientation in salmonids under laboratory conditions.Anim Behav 65:893–904. doi:10.1006/anbe.2003.2136

Pearce JM, Ward-Robinson J, Good M, Fussell C, Aydin A (2001)InXuence of a beacon on spatial learning based on the shape of thetest environment. J Exp Psychol 27:329–344

Redhead ES, Roberts A, Good M, Pearce JM (1997) Interactionbetween piloting and beacon homing by rats in a swimming pool.J Exp Psychol 23:340–350

Rossier J, Haeberli C, Schenk F (2000) Auditory cues support placenavigation in rats when associated with a visual cue. Behav BrainRes 117:209–214

Rozhok A (2008) Orientation and navigation in vertebrates. Springer,Berlin, Heidelberg

Sabbah S, Lerner A, Erlick C, Shashar N (2005) Under water polariza-tion vision—a physical examination. In: Pandalai SG (ed) Recentresearch developments in experimental and theoretical biology.TRN Press, Kerala, pp 123–176

Shashar N, Rutledge PS, Cronin TW (1996) Polarization vision in cut-tleWsh—a concealed communication channel? J Exp Biol199:2077–2084

Shashar N, Hanlon RT, deM Petz A (1998) Polarization vision helpsdetect transparent prey. Nature 393:222–223

123

596 Anim Cogn (2012) 15:591–596

Shashar N, Hagan R, Boal JG, Hanlon RT (2000) CuttleWsh use polar-ization sensitivity in predation on silvery Wsh. Vis Res 40:71–75.doi:10.1016/S0042-6989(99)00158-3

Shashar N, Johnsen S, Lerner A, Sabbah S, Chiao CC, Mäthger LM,Hanlon RT (2011) Underwater linear polarization: physical limi-tations to biological functions. Phil Trans R Soc B 366:649–654.doi:10.1098/rstb.2010.0190

Shettleworth SJ (2010) Getting around: spatial cognition. In: Shettle-worth SJ (ed) Cognition, evolution, and behavior, 2nd edn.Oxford University Press, New York, pp 261–312

Steck K, Hansson BS, Knaden M (2011) Desert ants beneWt fromcombining visual and olfactory landmarks. J Exp Biol 214:1307–1312. doi:10.1242/jeb.053579

Wehner R, Michel B, Antonsen P (1996) Visual navigation in insects:coupling of egocentric and geocentric information. J Exp Biol199:129–140

123