Elsevier

Behavioural Processes

Volume 106, July 2014, Pages 187-192
Behavioural Processes

Touchscreen performance and knowledge transfer in the red-footed tortoise (Chelonoidis carbonaria)

https://doi.org/10.1016/j.beproc.2014.06.003Get rights and content

Highlights

  • Red-footed tortoises were successfully trained to use a touchscreen.

  • Two of the tortoises learned to use the touchscreen to solve a spatial task.

  • They were able to transfer their knowledge from the touchscreen to a physical arena.

Abstract

In recent years red-footed tortoises have been shown to be proficient in a number of spatial cognition tasks that involve movement of the animal through space (e.g., the radial maze). The present study investigated the ability of the tortoise to learn a spatial task in which the response required was simply to touch a stimulus presented in a given position on a touchscreen. We also investigated the relation between this task and performance in a different spatial task (an arena, in which whole-body movement was required). Four red-footed tortoises learned to operate the touchscreen apparatus, and two learned the simple spatial discrimination. The side-preference trained with the touchscreen was maintained when behaviour was tested in a physical arena. When the contingencies in the arena were then reversed, the tortoises learned the reversal but in a subsequent test did not transfer it to the touchscreen. Rather they chose the side that had been rewarded originally on the touchscreen. The results show that red-footed tortoises are able to operate a touchscreen and can successfully solve a spatial two-choice task in this apparatus. There was some indication that the preference established with the touchscreen could transfer to an arena, but with subsequent training in the arena independent patterns of choice were established that could be evoked according to the test context.

Introduction

The ability to navigate through space successfully and efficiently can be considered to bestow a survival advantage as it allows for the successful passage between feeding grounds, sleeping quarters, and so on. Most research on spatial cognition has concentrated on navigation by mammals and birds (reviewed by Healy, 1998). There has been less research with reptiles, and much of what exists has been concerned with the study of seasonal, large-scale movements of sea turtles (Dutton et al., 1999) which are guided by the use of a variety of cues, including geomagnetic (e.g., Lohmann et al., 2001, Lohmann et al., 2004), visual (Avens and Lohmann, 2003), and celestial cues (DeRosa and Taylor, 1980). However, the majority of reptiles do not face the challenge of navigation on such a scale. For example, when painted turtles (Chrysemys picta marginata) were displaced a mile from their home pond they became disorientated and failed to find their way back (Emlen, 1969). This species can, however, navigate successfully on a smaller scale. When the turtles were released 100 m from home they were able to return quickly, and did so on a direct route. The turtles appeared to be using landmarks, such as the edge of a wood near the home pond, to guide their choices. This finding is perhaps unsurprising as this species, like the majority of reptiles, spend their lives within a small area, with which they are familiar. Research investigating small-scale navigation (for a review see Mueller et al., 2011) has shown that in this case too, reptiles are able to use a range of different strategies to find a goal. These are exemplified in a series of studies of spatial learning in the red-footed tortoise (Chelonoidis carbonaria; Wilkinson et al., 2007, Wilkinson et al., 2009, Mueller-Paul et al., 2012a, Mueller-Paul et al., 2012b).

This species is a land-dwelling chelonian, native to Central and South America. It is food motivated and is considered an omnivore, although much of its diet is fruit (Strong and Fragoso, 2006). The red-footed tortoise is a relatively active species, and is capable of travelling up to 85 m/h (Moskovits 1985, cited by Strong and Fragoso, 2006). They are highly visual, appear to have good colour vision and, whenever possible, use vision to solve a task (Wilkinson and Huber, 2012). Their liveliness and food motivation, in addition to their visual abilities (Wilkinson and Huber, 2012) makes them an ideal species for studying visual based spatial learning.

Recent research has revealed that the red-footed tortoise is able to master an eight-arm radial maze, in which it is required to remember several different spatial locations within a single trial (Mueller-Paul et al., 2012a, Mueller-Paul et al., 2012b, Wilkinson et al., 2007). These tortoises appear to be able to use room cues for navigation in a cognitive map-like manner, but they also exhibit stereotypic response strategies if cues are less salient (Mueller-Paul et al., 2012a, Mueller-Paul et al., 2012b, Wilkinson et al., 2009). Odour, too, has been identified as a possible cue, but appears to be used only when other cues are not available (Mueller-Paul et al., 2012a, Mueller-Paul et al., 2012b). Although red-footed tortoises are able to use different mechanisms to reach a goal they appear to prefer the first successful strategy they used, even if another might be simpler under changed circumstances (Mueller-Paul et al., 2012a, Mueller-Paul et al., 2012b). More flexibility was observed in a study by Wilkinson et al. (2010a). They showed that red-footed tortoises can learn the path that leads to a goal by observing a demonstrator tortoise. But the tortoises did not learn simply about the exact route followed by the demonstrator as they were able to apply the principles of the task even when the path to food was altered by introducing additional turns (Wilkinson and Huber, 2012). To this extent, red-footed tortoises have demonstrated an ability to generalize knowledge across variations of a previously learned task.

To examine further the mechanisms controlling spatial learning in this species it will be informative to test the tortoise's performance on comparable tasks in different domains. In the study to be reported here we made use of a 2-dimensional (2-D) display presented on a touchscreen and a traditional testing arena in which “real” 3-dimensional (3-D) objects could be presented. Assessing differences and similarities of behaviour in such distinct domains has the potential to tell us about the generality of spatial cognitive processes. Spontaneous transfer of knowledge from one domain to another would indicate a high level of generality of the acquired spatial knowledge. In particular, transfer from the touchscreen to a 3-D arena might be taken to indicate that a kind of mental map could be derived from the overview of the entire set-up that was provided in the touchscreen situation. A series of studies investigating transfer in pigeons has revealed strong similarities between spatial learning performance on a touchscreen and in a 3-D arena (reviewed by Cheng et al., 2006). For example, Kelly and Spetch (2004) and Kelly et al. (1998) demonstrated that pigeons were able to use feature and geometric cues to a similar extent when presented in a 2-D schematic and in a navigable 3-D environment. Further, the birds appear to use the configuration of landmark arrays to do this (Spetch et al., 1996). This suggests that similar spatial learning mechanisms govern the performance in these different domains, at least in this species.

Efficient transfer on a task of this type requires the subject to recognize that a picture represents an object, and evidence of this ability in non-human animals is scanty (for a review see Fagot, 2000). Recently, however, picture–object recognition has been investigated in the red-footed tortoise (Wilkinson et al., 2013). The findings revealed that the tortoises were able to recognie a correspondence between real objects and 2-D images of them. The animals were trained to distinguish colour-matched food and non-food items and were later able to make the same distinction between colour photographs of similar food and non-food items, the tortoises confused the real food items with the corresponding photographs, finding it difficult to differentiate between a photograph and the 3-D item that it represented, suggesting similar processing of 2-D and 3-D stimuli.

The present study made use of the 2-D-image recognition ability of red-footed tortoises in order to further investigate the mechanisms underlying tortoise spatial navigation. The first stage involved training subjects on a spatial discrimination in a touchscreen task that provided small-scale stimuli and a full overview of the situation. (The ability of this species to touch a stimulus-defined location in order to receive a reward in a different feeder location, has yet to be demonstrated; however, the proficient use of a pecking key has been shown in terrapins, Chrysemys picta picta; Bitterman, 1964, Powers et al., 2009.) We then went on to study performance on a comparable test in an arena that required walking through space towards one of a pair of 3-D objects. This allowed us to assess the possibility of transfer from touchscreen to the arena. To investigate the possibility of transfer in the other direction, we then trained subjects in the arena (the rewarded spatial position being reversed from that selected in the first phase of touchscreen training) prior to a test with the touchscreen.

Section snippets

Subjects

Four juvenile red-footed tortoises (Chelonoidis carbonaria–formerly Geochelone) with plastron lengths of 13 cm (Esme), 13 cm (Molly), 12 cm (Quinn) and 11 cm (Emily), took part in the study. The tortoises’ sex was unknown, as unambiguous sexual dimorphism develops only later in the life of this species. The tortoises were housed as a group of four in a 120 × 70 cm arena, at 28 ± 2 °C and approximately 60% humidity, with permanent access to fresh water, shelter, UV light, and heat lamps. The tortoises

Acquisition of touchscreen operation

All four tortoises learned to operate the touchscreen and to collect rewards from the feeder. Table 1 shows the number of sessions required by each individual to reach the criterion in each pretraining phase.

Touchscreen training

Emily and Molly did not progress to this stage as they stopped working during the sequence 2 stage of pretraining. The reason for this is unknown, as up to this stage they had performed reliably and with levels of success comparable to those of Esme and Quinn. Esme and Quinn, however,

Discussion

The results of the present experiment show that red-footed tortoises are capable of learning to operate a touchscreen Skinner box. This was true of all animals and is the first demonstration of such behaviour in this species, but it is in line with evidence from Bitterman (1964) and Powers et al. (2009), showing that terrapins could learn to use a pecking key. This suggests that tasks involving this sort of response are within the behavioural repertoire of chelonia generally and opens up the

Acknowledgements

This work was supported by funding from a Royal Society International Joint Project (to A.W. and G.H.) and the Austrian Science Fund (FWF no. 19574, to L.H.). The authors would like to thank the cold-blooded cognition research group, and in particular Wolfgang Berger for his invaluable help with the construction of the apparatus.

References (27)

  • M.E. Bouton et al.

    Memory processes in classical conditioning

    Neurosci. Biobehav. Rev.

    (2004)
  • K. Cheng et al.

    Small-scale spatial cognition in pigeons

    Behav. Processes

    (2006)
  • L. Avens et al.

    Use of multiple orientation cues by juvenile loggerhead sea turtles Caretta caretta

    J. Exp. Biol.

    (2003)
  • M.E. Bitterman

    An instrumental technique for the turtle

    J. Exp. Anal. Behav.

    (1964)
  • K.M. Davis et al.

    Turtles (Pseudemys nelsoni) learn about visual cues indicating food from experienced turtles

    J. Comp. Psychol.

    (2011)
  • C.T. DeRosa et al.

    Homeward orientation mechanisms in three species of turtles (Trionyx spinifer, Chrysemys picta, and Terrapene carolina)

    Behav. Ecol. Sociobiol.

    (1980)
  • P.H. Dutton et al.

    Global phylogeography of the leatherback turtle (Dermochelys coriacea)

    J. Zool.

    (1999)
  • S.T. Emlen

    Homing ability and orientation in the painted turtle Chrysemys picta marginata

    Behaviour

    (1969)
  • J. Fagot

    Picture perception in animals

    (2000)
  • S. Healy

    Spatial Representation in Animals

    (1998)
  • D.M. Kelly et al.

    Reorientation in a two-dimensional environment: II, Do pigeons (Columba livia) encode the featural and geometric properties of a two-dimensional schematic of a room?

    J. Comp. Psychol.

    (2004)
  • D.M. Kelly et al.

    Pigeons’ (Columba livia) encoding of geometric and featural properties of a spatial environment

    J. Comp. Psychol.

    (1998)
  • K.J. Lohmann et al.

    Regional magnetic fields as navigational markers for sea turtles

    Science

    (2001)
  • Cited by (30)

    • Environmental enrichment impacts discrimination between familiar and unfamiliar human odours in snakes (Pantherophis guttata)

      2021, Applied Animal Behaviour Science
      Citation Excerpt :

      This is partly due to the perception that reptiles adapt well to minimalistic enclosures (Warwick, 1990), but also the misconception that they are sluggish and their behaviour is largely controlled by innate drives (Wilkinson and Glass, 2021). There is now evidence of impressive cognitive abilities in this group (see Matsubara et al., 2017, for a review), including complex social learning (e.g. Wilkinson et al., 2010; Kis et al., 2015), spatial learning (e.g. Mueller-Paul et al., 2012, 2014), visual discrimination and long-term memory retention (e.g. Soldati et al., 2017). With their cognitive abilities being comparable to those of many mammals and birds (Wilkinson and Huber, 2012), alongside the increasing popularity of reptiles as pets (Burman et al., 2016), there is a clear need to further our understanding of the impact that environmental enrichment has upon the cognition and welfare of reptiles.

    • A computerized testing system for primates: Cognition, welfare, and the Rumbaughx

      2018, Behavioural Processes
      Citation Excerpt :

      This type of work will allow for comparisons among the various bear species as there has been growing interest in cognition in this taxa (Johnson-Ulrich et al., 2016; Kelling et al., 2006; Perdue et al., 2009, 2011; Swaisgood et al., 1999; Vonk and Beran, 2012; Vonk et al., 2012; Vonk and Johnson-Ulrich, 2014; Zamisch and Vonk, 2012). Increasingly, touchscreen or similar systems are being introduced to zoo animals and have already been introduced to a range of species (e.g., Mueller-Paul et al., 2014; Leighty et al., 2011; Wagner et al., 2016). Second, there is a need for continued development of enrichment for zoo animals, and this technology has been demonstrated to effectively accomplish this in a laboratory setting.

    • Cold-blooded cognition: new directions in reptile cognition

      2017, Current Opinion in Behavioral Sciences
      Citation Excerpt :

      It remains unknown whether these conditions would impact on performance of a social cognition task. Much cognition research with Chelonia has investigated spatial navigation with the focus on the impressive long-range navigational feats of sea turtles [39–41] and short-range spatial learning in terrapins and tortoises [9,42–45]. There is evidence for use of geomagnetic cues, visual cues and response-based strategies in these species [46].

    View all citing articles on Scopus
    View full text