black dot
  Navigation & learning
  The general topic of locomotion is divided into sections of navigation & learning, considered here, and CRAWLING and JET PROPULSION considered elsewhere.
Research study 1

Octopuses, like other animals, navigate by retracing specific routes, by orientating to features of the landscape, or a combination of both.  An octopus’ suckers are extremely sensitive to chemicals and touch, and may be most used to navigate around at night, whereas vision may be more important during the day; Studies on daytime foraging by octopuses in the field show that homeward trips to the den do not retrace outgoing paths, suggesting either a good chemotactile memory of the habitat, full reliance on vision, or a combination of the two.  The diagram on the Left shows the route taken schematic of daily movement of an octopus from its den, Bermudaby an Octopus vulgaris in Bermuda during 7 daylight excursions from its den.  No equivalent information is yet available for night excursions for any species of octopus. 

schematic showing learning ability of octopuses Octopus rubescensLaboratory studies on visual navigation in Octopus rubescens at the Seattle Aquarium, Washington reveal a quick ability to associate landscape objects with food and to orientate quickly to a “reward”-object if it is moved around in the habitat.  In one experiment, an individual in its den is presented with a piece of plastic tubing containing a crab, sucker-contact with which will yield reward of a small shore crab Hemigrapsus oregonensis as food.  On the first presentation the octopus comes out to investigate and gets a food reward (see Trial 1 in diagram on Right). On the following day the “reward”-tubing with crab is moved 90o schematic showing learning ability of octopuses Octopus rubescenscounterclockwise.  The octopus starts out, spots the tubing, goes to investigate, and gets its reward (see Trial 2).  On subsequent days of testing, the octopus finds the “reward”-tubing, on one occasion (Trial 3) after first going incorrectly to the Trial 1 location from 2d previously. 

The author does several other experiments using different types and configurations of landscape novelties and obtains similar results.  In one such experiment, an octopus already experienced in finding the reward-tubing, is presented in successive daily tests with the “reward”-tubing placed between a larger black plastic box and a white dish.  In half the trials the octopus orientates first to the black box, likely the most visually obvious of the objects, and then moves to the reward-tubing; in the other half, it goes directly to the tubing.  The smaller visually less obvious dish is ignored.  Overall, the study shows that octopuses can learn to orientate to visual landmarks and, like rats, have a working memory of where they have been. Mather 1991 J Comp Physiol A 168: 491.

Research study 2

diagram of test arena with burrow locations shown in blueOctopuses in the field navigate from temporary home dens to foraging areas located many meters away and return using inbound routes that often differ from outbound routes.  Spatial orientation of this sort involves visual input on landmark recognition, possible use of solar and lunar cues, and chemotactile perception of substratum features.  Discrimination learning, that is, an ability to respond selectively to one of several stimuli (big or small cliff-face, boulder piles, anemone cluster) will be important in this and, while much work has been done on visual and tactile discrimination of objects by octopuses, much less is known about how discrimination learning is used in spatial orientation.  Keep in mind that perception of the same cues by an individual will differ on outbound (from the den) and inbound trips even if the route followed is identical.  Thus decisions must be made flexibly.  Whether octopuses Octopus bimaculoides are capable of spatial learning is investigated in a study at the Marine Biomedical Institute, Galveston.  The researchers devise a test arena (1.5-1.8m diameter) containing several prominent “landmark” objects (rock pile, artificial plants, coil of rope, and the like), and 6 potential openings into refuge “burrows”, each partially covered by a terra-cotta saucer (see schematic on Left) and all but one sealed during training experiments with a removable bung.  The arena is brightly lit to encourage the test octopus to enter the open burrow and hide away in the darkness.  An octopus entering a burrow will commonly pull the saucer over its head to block out the light.  Spatial learning is tested as follows.  Once or twice a day a selected octopus subject is placed in the centre of the arena and allowed to explore until it finds the open burrow.  Distance traversed measured in “circumferences” is recorded.  After a series of graph showing averaged learning results before and after reversal of burrow opening 24 such tests, the octopus is rested for a week (see graph on Right; subsequent tests show that the octopus can remember the burrow location for at least 1wk).  At this stage in testing the burrow location is reversed by 180o and the same octopus is re-tested over a further 18 trials.  Note in the graph that search paths increase significantly after reversal, but performance improves with repetition showing relearning.  The authors comment that the “hands-free” approach used here helps to minimise problems of stimulus presentation, handling, inadvertent experimenter cueing, and food rewards/punishment that have complicated other learning studies with octopuses.  Boal et al. 2000 J Comp Psychol 114 (3): 246.

NOTE  also known as central-place foraging

NOTE  the authors have combined data from 2 separate experimental series done in different locations in this graph, so a precise description of the protocol is not possible.  Note, however, that the trials are grouped into blocks of 3 and averaged

Research study 3

diagram of results of a series of 10 separate but successive tests with a single octopusIn a later paper a different set of researchers at Millersville University, Pennsylvania refer to this type of spatial learning as conditional discrimination, hitherto only known in vertebrates and in a few invertebrates such as honeybees and sea hares.  Whether octopuses are also capable of conditional discrimination is investigated with Octopus bimaculoides using methodology similar to that used in the preceding Research Study 2.  Experiments involve a circular platform or maze with 2 octopus-sized (8cm diameter) “burrows” potentially available in each maze (see accompanying illustration). Two mazes (A & B) are configured differently for each octopus tested using various removable “landmarks” such as stones, simulated algae, rope coils, and the like that are glued onto replaceable wall and floor panels.  A movable plug allows one burrow or the other to be made operational during a trial.  A trial involves an octopus being placed at the centre of the maze and allowed to explore until it finds and escapes via the open burrow.  On a subsequent test the following day and for 10 such cycles thereafter, a different maze configuration is used, and the burrow opening is located opposite to its position in the  previous test.  Conditional discrimination is demonstrated when an octopus, having successfully found and entered the burrow in a first trial, can determine the correct escape route when placed in the differrent maze configuration.  Thus, an octopus has to locate a new escape route from a different direction using a different configuration of cues, and do this without error over the last 5 trials (in order that statistical significance be shown).  Results show that this type of discrimination is achieved by 6 out of 10 of the test O. bimaculoides. One of the 6 individuals solved the maze in all 10 trials, one in its last 8 trials, two in their last 6 trials, and two in their last 5 trials. Although initial headings and explorations often seem to be haphazard, conspicuous landmarks may sometimes be visited frequently.  Interestingly, unlike in the preceding Research Study, no individual learns to run the maze more quickly with greater experience, but whether this is a feature unique to the particular type of conditional discriminations being tested is not discussed by the authors.  The study is valuable in that it expands our understanding both of the navigational abilities of O. bimaculoides, specifically, and the  “higher-level” learning potential of cephalopods, generally.  Hvorecny et al. 2007 Anim Cogn 10: 449.

NOTE  also included in the study are 2 species of cuttlefishes Sepia, results of which are similar to those for octopuses (although not so good) and are not included here

NOTE  the sample results presented here have been modified to correct what appears to be a small error in the authors’ original presentation