Journal of Comparative Psychology
2009, Vol. 123, No. 3, 264 –274
© 2009 American Psychological Association
0735-7036/09/$12.00 DOI: 10.1037/a0015921
Memory of Visual and Topographical Features Suggests Spatial Learning
in Nautilus (Nautilus pompilius L.)
Robyn J. Crook
Roger T. Hanlon
Brooklyn College and City University of New York
Marine Biological Laboratory, Woods Hole, Massachusetts
Jennifer A. Basil
Brooklyn College and City University of New York
Previous studies demonstrate that soft-bodied (coleoid) cephalopods are adept at learning and remembering features of their environment, but little is known about their primitive relative, nautilus. Nautilus
makes nightly migrations from deep to shallow water along coral reef slopes, covering large areas of
varied substrate. Memory of its surroundings may be advantageous, but the nautilus brain is the simplest
among extant cephalopods, lacking dedicated neural regions that support learning and memory in other
cephalopods. The authors hypothesize that the absence of these regions in nautilus may affect memory
storage. Here the authors report the first evidence for spatial memory in 2- and 3-dimensional arenas. In
a small open-field maze, nautiluses learned the location of a goal within 3 trials, and memory was stable
for at least 2 weeks. In 3-dimensional environments, animals habituated within and across trials when
their surroundings were unchanged, but activity increased when the environment changed topographically, although not when the change was visual only. These results are comparable to performances of
coleoids in similar tasks and are surprising given the far simpler neuroanatomy of nautilus.
Keywords: nautilus, cephalopod, learning, memory, spatial
numerous discrete lobes with dedicated functions, at least one of
which (the vertical lobe) is dedicated to learning and consolidation
of memory, although other regions may play a role in memory
storage (Fiorito & Chichery, 1995; Hochner, Brown, Langella,
Shomrat, & Fiorito, 2003; Hochner, Shomrat & Fiorito, 2006;
Young, 1960, 1991). There is considerable evidence for spatial
learning in coleoids (Alves, Boal, & Dickel, 2007; Alves, Chichery, Boal, & Dickel, 2007; Boal, Dunham, Williams, & Hanlon,
2000; Karson, Boal, & Hanlon, 2003; Mather, 1991; Wells, 1964),
as well as numerous examples of associative and nonassociative
learning (reviewed by Hanlon & Messenger, 1996, and Mather,
2008).
By contrast, the primitive nautiloids have changed relatively
little over the course of their evolutionary history and retain the
external, chambered shell that is characteristic of both nautiloid
and coleoid ancestors (Teichert, 1988). Their brains are relatively
simple, containing fewer lobes and fewer neurons than do coleoid
brains (Young, 1965). The ring-shaped brain is less centralized
than that of the coleoids to accommodate the large esophagus that
passes through its center, and the main lobes are not differentiated
clearly from the surrounding tissue, in contrast to the defined,
lobular structure of the coleoid central nervous system. Crucially,
those regions implicated strongly in learning and memory in
coleoids are absent in nautiloids. Although there are some structural similarities between the cerebral nerve cord of nautilus and
the inferior frontal lobe in coleoids (Young, 1965), the function of
this region in nautiloids is undescribed.
There is some evidence that the neuroranatomy of nautiloids
may represent an ancestral condition (Shigeno, Kidokoro, Tsuchiya, Segawa, & Yamamoto, 2001; Shigeno et al., 2007; Young,
Cephalopoda (Mollusca) is a diverse and highly successful
taxon that today occupies a wide range of marine niches. All
modern cephalopods belong to one of two subclasses, the ancient
Nautiloidea or the more modern Coleoidea (Teichert, 1988). The
coleoids include those cephalopods with an absent or internalized
shell (octopuses, cuttlefishes, and squid), while the nautiloids
contain only the remnant genera Nautilus and Allonautilus (Ward
& Saunders, 1997). Coleoids have a large and centralized nervous
system that supports a range of complex and plastic behaviors
(reviewed by Hanlon & Messenger, 1996). Their brains contain
Robyn J. Crook and Jennifer A. Basil, Biology Department, Brooklyn
College, and Ecology, Evolution, and Behavior Subprogram, City University of New York Graduate Center; Roger T. Hanlon, Marine Biological
Laboratory, Woods Hole, Massachusetts.
This research was supported in part by the Marine Biological Laboratory
(MBL) and a Grass Foundation fellowship to Robyn J. Crook. We thank
Tina Kuroiwa for valuable input into the planning stages of this study and
assistance with animal care. Martin Schreibman and the members of the
Aquatic Research and Environmental Assessment Center at Brooklyn
College (AREAC) assisted with animal care and maintenance of our
animal housing system, as did the staff of the Marine Resources Center at
the MBL. We gratefully acknowledge the assistance that Robert Rockwell,
Peter Moller, Michael Kuba, Chris Soucier, and Frank Grasso provided
with statistical procedures, experiment design, and interpretation of our
results. Jean Boal provided helpful feedback on a later version of the
manuscript.
Correspondence concerning this article should be addressed to Robyn J.
Crook, who is now at the Department of Integrative Biology and Pharmacology, University of Texas Medical Branch, Houston, TX 77030. E-mail:
Robyn.Crook@uth.tmc.edu
264
MEMORY OF VISUAL CUES IN NAUTILUS
1965, 1991), and therefore identifying comparable behavioral expressions of learning in nautilus may provide evidence for an
evolutionary antecedent of the “memory” centers present in the
modern coleoid brain. Alternatively, it is plausible that the simple
nautiloid brain represents a secondary simplification and that neural evolution in cephalopods is bidirectional (e.g., Safi, Seid, &
Dechmann, 2005). Similar learning abilities in nautiloid and coleoid cephalopods would therefore suggest a trade-off between
reduction in brain volume or complexity and maintenance of
behavioral plasticity mediated by different neural substrates in
nautiloids.
Learning and memory in nautilus were unknown until recently.
Crook and Basil (2008) used an associative learning paradigm to
condition N. pompilius to associate a pulse of blue light paired with
food odor. Animals demonstrated temporally separated short- and
long-term associative memory, similar to results observed in cuttlefishes (Agin, Chichery, Maubert, & Chichery, 2003; Agin,
Dickel, Chichery, & Chichery, 1998; Messenger, 1971). The duration of long-term memory (LTM) was surprisingly short in these
experiments (under 24 hr), raising the possibility that brief LTM
expression is intrinsic to the simple nautilus brain. In the present
study we aimed to clarify whether this brief LTM expression was
stable across different learning contexts or was instead specific to
that associative learning paradigm. Here we focus on learning of
visual environmental features, an important component of spatial
learning. Such behavior is fundamental to survival in a range of
animal species and is potentially an ecologically relevant task for
N. pompilius.
Spatial learning is a widely distributed ability among evolutionarily distant taxa and is mediated in vertebrates and among invertebrates by nonhomologous neural structures, suggesting multiple
evolutionary origins for this behavior. The absence of coleoid-like
neural complexity in nautiloids is thus not necessarily a limitation
on spatial learning, although some differences in cue use and
strategy might be expected given the different visual systems
(Muntz, 1987) and behavioral ecology (Hanlon & Messenger,
1996) of the two subclasses.
Forming representations of space involves stages of exploration,
habituation, and memory maintenance (patrolling; Gallistel, 1990;
Healy, 1998; O’Keefe & Nadel, 1978; Shettleworth, 1998). If a
change to a familiar environment is perceived, then the level of
exploratory behavior should increase (dishabituation), presumably
because the previously gathered environmental information must
be updated (bees: Simonds & Plowright, 2004; cave fish: Teyke,
1989; fish: Welker & Welker, 1958; hamsters: Poucet, Chapuis,
Durup, & Thinus-Blanc, 1986). Patterns of staged exploratory
behavior have been observed in some coleoids in their natural
habitat (e.g., foraging excursions in reef octopuses: Forsythe &
Hanlon, 1997; Mather, 1988, 1991). Laboratory studies also suggest spatial learning in octopuses (Boal et al., 2000; Papini &
Bitterman, 1991; Walker, Longo, & Bitterman, 1970; Wells,
1964). While cuttlefishes rely primarily on crypsis to avoid detection by predators (Hanlon & Messenger, 1988), laboratory studies
indicate that they are also capable of spatial learning (Alves, Boal,
& Dickel, 2007, Alves, Chichery, Boal, & Dickel, 2007). There are
limited data on squid and on pelagic octopuses.
The possibility that nautilus also uses memory of its surroundings is an intriguing one. In contrast to cephalopod species in
which spatial learning has been described, nautilus is a deep-water
265
inhabitant that spends most of its time in darkness (Carlson,
McKibben, & DeGruy, 1984). While the brain is simple compared
with those of the other cephalopods, nautilus is nonetheless capable of both short- and long-term memory (Crook & Basil, 2008).
Demonstrating spatial learning in nautilus would add to our limited
understanding of how nautiluses may behave in the wild and,
secondarily, characterizing learning behavior across a range of
contexts in nautiluses provides an important comparative perspective on the remarkable abilities exhibited by other cephalopod
species.
The purpose of this study was to investigate learning and memory of visual environmental cues in wild-caught N. pompilius. We
examined learning in a small-scale environment (within 10 –30
body lengths), using two techniques: a two-dimensional open-field
maze that required animals to locate and remember a fixed goal
position in the arena with the aid of visual cues, and an artificial
reef construction that tested habituation and dishabituation to a
complex, three-dimensional environment that underwent changes
in shape and pattern.
Experiment 1: Memory in a Two-Dimensional
Environment
This experiment investigated whether animals could learn to
locate a goal point within their environment that was marked with
a proximate cue. Escape and bright lighting in the arena were used
as motivators, because nautiluses typically descend in the presence
of bright moonlight in the wild (Ward, Carlson, Weekley, &
Brumbaugh, 1984) and will avoid shallow areas in their home
tanks (Crook, 2008). Memory relied on a goal– cue association
being learned during training.
Method
Animals. This experiment was conducted at Brooklyn College
in March and April of 2007. Wild-caught, subadult Nautilus pompilius (N ! 10) from the Philippines were obtained through a
commercial supplier (Sea Dwelling Creatures, Inc., Los Angeles,
CA). Shell diameters ranged from 110 mm to 119 mm. Sexes were
undetermined because most animals this size are sexually immature. Animals were housed in two darkened, cylindrical tanks (1.5
m high " 0.8 m diameter) in a 530-L, recirculating, artificial
seawater system (Instant Ocean™). The tanks were connected in
tandem to a 95-L biofilter that supplied aeration and filtration for
the system. Water was maintained at 17°C by a chiller (Aqua
Logic 1.3-hp AE4) and was kept sterile by two ultraviolet (UV)
filters (Emperor Aquatics 80-W Model 02080) and two protein
skimmers (Red Sea™), operating constantly. The 12-hr light/dark
cycle (06:00 –18:00) alternated between very dim light and complete darkness. Tanks were covered, but two holes in the tank lids
(diameter ! 3 cm) let dim light into part of the tanks during the
light half of the day/night cycle. Animals were maintained in the
holding facility for at least 2 weeks before being used in any
experimental procedure. We fed each animal a 1.5-cm cube of
frozen Tilapia (Oreochromis niloticus) head by hand every 4 days.
Feeding schedule was altered during experiments to ensure that
animals taking part in experimental procedures were not fed immediately prior to the procedure but were always fed several hours
after the conclusion of testing. A delay of 3 to 4 hr between
266
CROOK, HANLON, AND BASIL
behavioral testing and delivery of food should have prevented any
association between the two events forming. A crucial component
of this experiment involved animals swimming downward through
the water column; thus only animals showing normal buoyancy
regulation behavior were used.
Apparatus. Trials were conducted in one of the two cylindrical home tanks (Figures 1a and 1b), which had been fitted with a
circular open-field maze that was immersed 10 –15 cm below the
water surface, about the depth of a nautilus-shell diameter. The
maze was a circular polyvinyl chloride (PVC) platform made from
a modified tank lid (Aquatic Ecosystems, Inc.), resized to fit
tightly within the cylindrical home tank. The upper surface of the
maze platform was smooth and black, identical in appearance and
texture to the inner walls of the home tank. A single hole (diameter ! 20 cm) cut into the platform close to one edge served as the
goal location for all memory trials (Figure 1b). The hole was the
only point of exit from above the platform into deeper and darker
water (130 cm deep, under the maze) within the home tank. A
visual and tactile “beacon” surrounded the goal, acting as both a
proximate (tactile) and a distant (visual) navigational aid. The
beacon was constructed from a ring-shaped piece of bubble wrap
with strips of white tape affixed, forming stripes that radiated from
the hole outward. This combination of pattern and texture con-
A
Camera
Escape
hole
Maze
Tank
B
20cm
L
E
100cm
S
Figure 1. The apparatus set-up for Experiment 1. (a) The maze was
immersed in one of the cylindrical home tanks, to a depth similar to the
height of the focal animal’s shell. A camera recorded the trial from
overhead. The arrow shows the escape path. (b) Maze shown from above.
S ! start position; E ! escape hole; L ! landmark. Not to scale.
trasted strongly with the smooth black surface of the platform. The
test arena was lit by fluorescent tubes located directly above the
tank. Within the tank, there was a gentle, anticlockwise flow of
water (#1 cm/s) from the inlet pipe below the maze platform to the
outlet pipe above it.
Trials were videotaped with a digital camera (Hitachi CCD,
Model KP-M2U) suspended above the tank, and a Sony DV
minirecorder (Sony DV walkman, Model GV-D900). The experimenter, using the minirecorder display, monitored the animal
from behind a blind.
Experimental procedures. Animals (N ! 10) were trained and
tested in random order. Five minutes before a trial began, 1 animal
was removed from its home tank and placed in an uncovered
bucket containing home tank water. This allowed the animal a
period of preexposure to ambient light levels in the experimental
room before the trial began. The water level above the platform
was adjusted to the same depth as the test subject’s shell height
such that the base of the animal’s shell would remain in contact
with the platform as it swam. The shallow depth of the water above
the platform and the bright light in the experiment room acted as
aversive stimuli.
Within the tank, the goal location (exit) was marked by the
beacon, but otherwise the interior of the tank was featureless,
because all surfaces were made from identical black PVC plastic.
Flow within the tank provided local hydrodynamic cues. A light
gradient within the room was created by an algal-culture stand in
one corner lit by broad-spectrum fluorescent lights, a black UV
filter casing projected above the tank rim to the left of the escape
point, and the wall opposite the algal culture stand painted a darker
shade than the other walls (yellow vs. white). The visual sensitivity
of nautilus is not well known (Muntz, 1987), and we address use
of specific cue types and intra-maze versus extramaze cues in other
experiments (Crook, 2008).
Video recording of the experimental arena began 30 s before the
animal was positioned in a consistent start location (Figure 1b),
which was located opposite (180°) the escape hole. The subject
was held in the start position for 5 s, facing into the center of the
tank, then released and allowed 10 min to locate and swim through
the exit hole and into deeper water. Animals that remained
stationary for more than 1 min or attached to the maze surface with
their tentacles during training trials were prompted with a tap
on the hood with a wooden rod to motivate renewed search
behavior. Once subjects descended through the hole, they remained undisturbed in the deeper, dark water for 10 min as a
positive reinforcement for completing the task. Animals that failed
to complete the task within the 10 min were guided gently by hand
into the escape hole as soon as the 10-min trial time expired. This
was done either by allowing the animal to attach its tentacles to the
experimenter’s fingers, who then towed the animal at its normal
swimming speed into the hole, or if the tentacles were retracted, it
was pushed from behind at swimming speed until it contacted the
hole and exited. Our animals were accustomed to being hand fed,
so contact with the experimenter’s hand and being moved manually were experiences with which our animals were familiar.
Contact did not appear to have been perceived as threatening,
given the normal coupling with feeding and the absence of typical
alarm behaviors in response to being moved by hand. Once inside
the hole, subjects descended usually without further prompting and
were left under the maze for 10 min. After 10 min, animals were
MEMORY OF VISUAL CUES IN NAUTILUS
retrieved with a net, the maze was removed and cleaned to remove
olfactory cues that may have been present on the maze surface, and
the procedure was repeated.
All subjects received a single block of five 10-min training
trials, with an intertrial interval (ITI) of 15 min. Retention of
memory was tested with a single probe trial, identical to training
trials, at 18 hr, 24 hr, 36 hr, 48 hr, 72 hr, 96 hr, or 7 days. Because
retention was still apparent at 7 days posttraining (see the Results
section), we performed a follow-up test of 6 animals at 21 days
posttraining. For each retention test, 6 of the 10 animals were
selected randomly to receive a probe trial (however, all 10 animals
were tested at 7 days, when the experiment was designed to
conclude). We used this sampling method to ensure that each
animal was not subject to intensive, daily handling over the 5 days
that included training and testing up to 96 hr, which may have
affected motivation, behavior, and feeding (R. Crook, unpublished
data, 2007). Animals were collected from the trial tank as soon as
they located the exit hole and did not receive the 10-min reward
period as occurred in training trials.
Data analysis and statistical procedures. Escape times and
swimming paths were recorded from video footage of each trial.
Escape paths were recorded from video footage by plotting the
position of the animal every 2 s. Time to escape was calculated
from the time of release at the start of the trial to the time when the
animal’s shell dropped into the escape hole. We used nonparametric statistics as our small sample size and the distribution of our
data indicated that the assumptions of an analysis of variance
(ANOVA) were not met. Escape times among training trials and
between the final training trial and test trials were compared with
nonparametric Friedman tests for repeated measures, and post hoc
Wilcoxon signed-ranks test were applied to pairwise comparisons.
A sequential Bonferroni correction (Rice, 1989) was used to cor-
267
rect for accumulating alpha error rates arising from multiple post
hoc comparisons. All probability values quoted are two-tailed.
Hedges’ g effect sizes and their 95% confidence intervals (presented as effect size; upper/lower CI) are quoted for selected
significant comparisons.
Results
Escape times. Escape times decreased rapidly across the five
training trials. There was an overall difference among trials (Freidman test, Q ! 16.71, df ! 4, p ! .002). In the first trial, 7 of the
10 animals failed to escape within the time allotted and were
moved to exit hole at the conclusion of the trial; an escape time of
600 s was recorded for these individuals. In the second trial only
3 animals failed to complete the task within 10 min. The mean
escape time decreased significantly from the first trial (Z ! $2.70,
df ! 1, p ! .007; Figure 2). but CLs of the effect size overlapped
zero (g ! $0.68; 0.22/$1.38). There was a further significant
decrease in the third trial (Z ! $2.55, df ! 1, p ! .01; g ! $2.04;
$0.96/$3.12), where the mean escape time reached an asymptote
at about 1 min, and all animals exited the maze successfully during
the trial period. There was no further decrease in Trial 4 (Z !
$1.02, df ! 1, p ! .31) or from Trial 4 to 5 (Z ! $1.37, df ! 1,
p ! .17). In Trial 4 and Trial 5, all animals escaped within 10 min.
There were no significant differences in comparisons of escape
times in the final training trial to any of the memory tests (see
Figure 2), indicating retention of memory across the 21-day test
period. Although a slight increase in escape times was apparent in
the MT at 7 days and 21 days, escape times remained significantly
lower than those of naı̈ve animals in Trial 1 (7 days: Z ! $2.39,
df ! 1, p ! .02; 21 days: Z ! $2.20, df ! 1, p ! .03).
Figure 2. Mean escape times (%1 standard error) in successive training and testing trials in Experiment 1.
Nautiluses (N ! 10) received five successive, 10-min training trials (intertrial interval, ITI ! 15 min), then 6
subjects were selected randomly for memory tests at each of the retention periods (all 10 received a memory test
at 7 days posttraining). Escape times decreased rapidly and remained consistently low from the third training trial
to at least 21 days posttraining. Points labeled with a different letter are significantly different ( p & .05,
two-tailed).
268
CROOK, HANLON, AND BASIL
Movement patterns. We computed a simple statistic of movement patterns by comparing the number of times per trial animals
changed heading direction by more than 15°, an angle we considered to represent intentional change and not merely oscillations
created by the natural swimming motion made during movement
across the surface of the maze (see Tourtellot, Collins & Bell,
1991). Complete routes taken by each animal and the mean number of changes of direction per trial are shown in Figure 3.
Increasing linearity of escape paths during training indicates that
learning was responsible for the decrease in escape times rather
than an increase in search activity or swimming distance, producing an increased encounter rate with the exit location. A significant
decline in direction changes compared with naive animals occurred in Trial 4 (Z ! $2.14, p ! .03, g ! $1.40; $0.43/$2.38)
and persisted in Trial 5 (Z ! $1.97, p ! .04; g ! $1.36;
$0.39/$2.33). Paths remained relatively direct across tests at each
retention interval, with no significant increases in comparison with
the final training trial.
1
2
9.0 ±2.1
10.1±3.3
Experiment 2: Memory of Three-Dimensional
Environments
In Experiment 1 it was apparent that our animals learned to
associate the beacon with the reward; however, whether spatial or
associative learning was responsible was not clear. Therefore we
used a different experimental paradigm to illustrate learning of
spatial representations in three dimensions. The purpose of this
experiment was to investigate whether
Training
(a)
nautiluses would move about a new environment spontaneously;
(b)
nautiluses would show reduced activity over time consistent with habituation;
(c)
increased activity (dishabituation) would occur when
features of the environment were changed; and
3
4
8.9±1.7
4.0±0.7*
5
3.8±1.7*
Test Trials
18h
3.2±1.0
24h
4.1±0.6
36h
2.7±0.4
96h
7d
21 d
3.4±0.2
4.8±0.6
4.8±0.7
48h
3.8±0.6
72h
4.6±0.7
Figure 3. Route maps show the paths that nautiluses took in each trial of Experiment 1. Positions were plotted
every 2 s from the release point (lower right of each large circle) until animals reached the exit (inner circle, top
left) or until the 10-min trial expired. The route of one randomly selected animal has been highlighted in black:
This individual received five training trials and was tested at the intervals of 18 hr, 96 hr, 7 days, and 21 days
posttraining. Paths of all other individuals are shown in gray. Points show the final position of each subject. In
training trials (1–5, top panel), paths initially suggested a random exploratory pattern. In successive training trials
the paths taken to the exit became progressively straighter. Paths remained relatively direct throughout the
retention period (bottom panel), suggesting that memory was retained for at least 21 days after training. The
mean number of changes of swimming direction ('15° change in heading angle) are shown below each map.
Asterisks indicate significant differences ( p & .05, two-tailed) in the number of direction changes. Training trials
were compared to behavior of naı̈ve animals in Trial 1; a significant decline appeared in training Trials 4 and
5. Behavior in each memory test was compared to the final training trial, no significant differences were detected.
MEMORY OF VISUAL CUES IN NAUTILUS
(d)
269
whether changes both to the shape (topographical
change) of the environment and to the pattern (visual
change) produced an increase of exploratory activity.
We used habituation (decreasing activity levels) as a measure of learning in this experiment because the arena did
not include aversive shallow areas or a specific goal point.
Method
Animals. This experiment was conducted in July and August
of 2007 at the Marine Biological Laboratory (MBL; Woods Hole,
MA). Wild-caught animals (N ! 5) were obtained from Vanuatu
via a commercial supplier (Pacific Island Imports, Artesia, CA).
The housing system and experimental arena were cylindrical PVC
tanks that were connected to the open seawater system at the MBL.
Water temperature increased slowly over the course of the experiment from 19°C to around 21°C. The home tank was covered by
light-proof sheeting, which was lifted only during feeding and
daily maintenance. Animals were fed by hand with a whole uncooked shrimp every 4 days.
Apparatus. Experimental trials were conducted in a cylindrical PVC tank (diameter 120 cm, depth 120 cm) completely enclosed within a blind of light-blocking black plastic. The tank
contained a single inlet pipe about 10 cm below the water surface
and an outlet pipe about 2 cm above the floor of the tank. Flow was
slow (#10 cm/min) and circular. The arena was lit from directly
overhead with a single incandescent blue bulb (40 W), providing
consistent dim light throughout the experimental procedures.
Three artificial “reefs” were built into this test arena over the
course of the experiment. Each reef was constructed from 10
concrete cinder blocks (19.4 " 19.4 " 39.7 cm) and three short
lengths of white PVC plumbing pipe (inside diameter ! 10 cm,
length ! 20 cm). Three sides of each block had been painted with
inert acrylic paint: five painted white and five painted black. The
blocks and pipe lengths were arranged against one side of the
cylindrical experimental tank, in three different configurations
(Reef 1, Reef 2, and Reef 3; Figure 4) over the course of the
experiment. Each reef contained the same number of blocks and
pipe pieces, and the number of exposed surfaces and the number of
painted sides visible (not abutting another block) were constant
across the three different reef configurations. Reef 1 (Figure 4a)
was a step-like stack of blocks that contacted approximately a third
of the edge of the tank. Reef 2 and Reef 3 (Figures 4b and 4c) were
stacked in an identical H shape that contacted a narrow portion of the
tank wall but projected into the center of the water column. Reefs 2
and 3 were identical to each other structurally (but were a different
spatial configuration than that of Reef 1), but they differed with
respect to the arrangement of the black and white blocks. Thus an
animal exposed to Reef 1, then Reef 2, would have experienced a
structural change and a visual change to its environment (the shape
and the pattern of light/dark blocks were different), while an animal
exposed to Reef 2 and then to Reef 3 would have experienced a visual
(the pattern of light/dark blocks) change only.
All experimental trials were recorded from an oblique overhead
angle with a Sony mini-DV handycam (DCR-HC36) equipped
with night-shot capability, on a time-lapse recording protocol of 1
frame per 4 s (15 frames per minute). Small visual markers on the
inside walls of the tank created a virtual grid for video-analysis
Figure 4. In Experiment 2, nautiluses (N ! 5) were exposed to three
different “reef” configurations over successive trials. Each panel shows the
front and side views of each reef. (a) Reef 1 was a step-like stack of blocks
and pipes. (b) Reef 2 and (c) Reef 3 were identical structurally but different
visually—the black and white blocks were located in different positions
between the two reefs. In these views the third pipe is concealed behind a
block at the base of the stack, opposite to the upright piece shown. Diagram
is not to scale.
purposes, allowing us to estimate movement of the animal in
three-dimensional space from this recording position.
Experimental procedures. Before trials for each reef commenced, the trial tank was drained partially, and one of the three
reefs was assembled by hand. The tank was refilled with fresh
seawater from the open seawater system and at least 5 hr elapsed
with continuous flow before an animal was introduced. Animals
were exposed first to Reef 1, then to Reef 2, then to Reef 3. For
each trial, an animal was collected from the home tank in a
darkened bucket and transferred into the trial tank, at the surface of
the water and opposite the reef, then allowed to move about the
tank freely for 3 hr. At the conclusion of the 3-hr period the animal
was removed from the trial tank and returned to the home tank for
6 hr. This process was repeated for five trials, then during the
normal 6-hr ITI the tank was drained, and Reef 2 was assembled.
Another five trials were given on the same schedule, before the
final reef change was conducted (Reef 3), and again animals
received five 3-hr trials in the final reef configuration.
Data analysis and statistical procedures.
Behavioral data
were taken from tapes at the conclusion of the experiment. An
assistant who was unfamiliar with the experiment assigned a code
number to each tape that was unknown to the videoanalyst until
after the analysis was complete.
We recorded the distance that each animal swam to the nearest
10 cm per “active minute” (minutes during which some movement
was recorded) in each trial, across the 3 hr of trial time. We also
270
CROOK, HANLON, AND BASIL
calculated the percentage of time that the animals spent swimming
(“percentage activity”) for the whole trial: (minutes during which
some movement was observed/180)!100. Within-trial habituation
was identified by comparing behaviors between the first and third
hours of each trial. Across-trial habituation was identified by comparing behaviors between the first and last trials in each reef configuration. Comparisons between the last trial in each reef configuration
and the first trial of the next configuration provided evidence for
dishabituation in response to a change of environment. We scored the
subject’s proximity to the reef by taking point observations of its
position every 60 s (0 ! more than 30 cm away; 1 ! within 30 cm
but not touching the reef; 2 ! touching the reef). Proportional data
(number of times each score was recorded/180) were compared
among trials. The sample size in this experiment was small, and data
did not meet assumptions of a normal distribution; thus nonparametric
statistics were applied. All probability values are two-tailed.
Results
Within-trial habituation. For Reef 1, there was significant
decrease in distance swum between the first and third hours
(Wilcoxon signed-ranks test, Z ! $3.18, df ! 1, p ! .001; g !
$2.40; $0.78/$4.03; Figure 5a); neither Hour 1 nor Hour 3 was
different from Hour 2. There was also a decline in the percentage
of time animals were active between the first and second hours
(Wilcoxon signed-ranks test, Z ! $2.55, df ! 1, p ! .01; g !
$1.64; $0.21/$3.07), but there was no further decline in activity
in the third hour (Z ! $1.95, df ! 1, p ! .051).
In Reef 2 the pattern was similar. There was a significant decline
in distances swum between the first and second hours (Z ! $2.69,
df ! 1, p ! .007; g ! $2.02; $0.50/$3.55). In the third there was
a nonsignificant decline from Hour 2 (Z ! $1.82, df ! 1, p !
.06). The percentage of time that animals were active also decreased across the 3 hr. There was a significant decline from the
first hour to the third hour only (Z ! $3.41, df ! 1, p ! .001; g !
$1.56; $0.14/$2.97).
In the final reef configuration (Reef 3), the decline was derived
from the change in activity between Hour 1 and Hour 2 (Z ! $3.07,
df ! 1, p ! .002; g ! $2.14; $0.59/$3.70). There was no further
decline in the third hour (Z ! $0.29, p ! .76). The percentage of time
that animals were active declined slightly from the first trial to the
Figure 5. Within-trial habituation was apparent in each of the three reef environments. Bars show the mean %1
standard error for Hours 1, 2, and 3 for trials in each reef. (a) The distance that animals (N ! 5) swam across
the 3-hr trial time decreased in each of three reef configurations. (b) The percentage of time that animals (N !
5) were active also decreased across the 3-hr trial period. Asterisks show significant differences between hours
( p & .05, two-tailed).
MEMORY OF VISUAL CUES IN NAUTILUS
third trial, but this result was not significant after the Bonferroni
correction was applied (Z ! $2.11, df ! 1, p ! .04).
Across-trial habituation over repeated exposures to identical
reefs: Dishabituation in response to reef changes. There were
no significant differences across trials in the distances that animals
swam. However, the percentages of time that animals were active
did show evidence for across-trial changes (Figure 6). Activity
declined from Trial 1 to Trial 5 (Z ! $2.55, df ! 1, p ! .01; g !
$1.40; $0.01/$2.78) and increased when animals experienced
the topographic change, from Reef 1 to Reef 2 (Trial 5 to Trial 6;
Z ! $2.20, p ! .04; g ! 1.45; 2.84/0.06), indicating that some
dishabituation to the environment had occurred. By the final trial
271
in Reef 2, activity had declined again (Trial 6 to Trial 10; Z !
$2.32, p ! .02; g ! $1.46; $0.06/$2.85), but there was a
nonsignificant increase in activity when Reefs 2 and 3 were
exchanged (Trials 10 to 11), indicating that a visual change only
did not result in increased exploration (Z ! $0.14, df ! 1, p !
.89). There was no further change in activity from the first trial to
the final trial in Reef 3 (Z ! $0.82, df ! 1, p ! .41).
Movement patterns. Time spent both in proximity to (within
30 cm) and in contact with the reef structures typically increased
with increasing experience, although variation within subjects was
large, and significant differences were few (see Figure 7). There
was a significant increase in the proportion of time animals spent
Figure 6. There was no evidence for across-trial habituation in (a) the mean distances that animals (N ! 5)
swam per minute of each trial, but there were significant differences in (b) the percentages of time that animals
were active. Points show the mean %1 standard error for each trial. Percentage activity decreased across
successive trials in Reef 1, indicating that animals remembered features of the environment across the five-trial
period. When the reef was changed from Reef 1 to Reef 2 (a structural change), activity increased, then declined
again across successive trials. There was no increase in activity when the environment changed visually only.
Asterisks indicate significant differences ( p & .05, two-tailed) between behavioral measures in each trial.
272
CROOK, HANLON, AND BASIL
Figure 7. Proximity to the reef tended to increase with experience. Proximity was scored in point observations
every 60 s as 0 ('30 cm away), 1 (&30 cm away), or 2 (touching the reef). Time spent in contact with the reef
(dashed line) and time spent within 30 cm of the reef (solid line) increased from Trial 1 to Trial 5, but changes
beyond the fifth trial were nonsignificant. Linking bars show significant effects. Top bar: time spent within 30
cm of the reef. Bottom bar: time spent in contact with the reef. Asterisks indicate significant differences ( p &
.05, two-tailed).
within 30 cm of the reefs from Trial 1 to Trial 5 (Z ! $2.02, df !
1, p ! .04), and in the time spent in contact with the reef (Z !
$2.02, p ! .04). Although a slight decline in proximity was
evident at the change from Reef 1 to Reef 2, neither measure was
significant (within 30 cm: Z ! $1.48, p ! .13; in contact: Z !
$1.82, p ! .68). There were no significant changes between the
first and last trials in either of the subsequent configurations or at
the second change of reef.
Discussion
N. pompilius was able to learn about features of its spatial
environment rapidly and retained this memory for at least 14 days
after training. Despite possessing a considerably simpler brain
(Young, 1965) and behavioral repertoire (Basil et al., 2000;
Soucier & Basil, 2008) than do coleoid cephalopods, nautiluses
displayed comparable associative and nonassociative learning and
memory when tested under similar circumstances (e.g., Alves,
Chichery, et al., 2007; Boal et al., 2000; Hvorecny et al., 2007;
Long, Hanlon, Ter Maat, & Pinsker, 1989; Papini & Bitterman,
1991). We stress that the results presented here do not provide
unequivocal evidence for spatial learning in the nautilus, such as
the use of path integration vectors or formation of cognitive maps,
but instead suggest some mechanisms and types of cues that may
be learned and remembered by nautiluses as they navigate about
their environment. More study is required to resolve what tactics
and cues nautiluses are able to use to navigate and to learn about
their environment. In Experiment 1 it was likely that the animals
simply learned the association between the proximate visual cue
and the exit, although it is possible that extramaze visual cues or
other cue types such as hydrodynamic flow were used to navigate.
We investigate use of multiple visual cues and map-based navigation in the nautilus in another study (Crook, 2008).
Despite some uncertainties over the type of learning involved,
our study demonstrates several important new findings: very rapid
learning of visual cues and retention of memory that was stable for
a considerable period.
Acquisition times for nautiluses were similar to those recorded
in coleoids: Octopuses placed in a circular open-field maze similar
to the set-up in Experiment 1 learned the location of a burrow
within the first three trials and retained this memory for at least a
week (Boal et al., 2000). In similar spatial memory experiments
with cuttlefish, subjects showed a marked drop in escape time after
four training trials (Karson et al., 2003), whereas we observed a
significant decrease in escape time after a single training trial and
further improvement in the next. Habituation was also rapid, in
some cases within an hour of exposure. Habituation to stimulus
objects, indicated by decreased exploration, has been shown to
occur both within and across trials in octopuses. Kuba, Byrne,
Meisel and Mather (2006) found that octopuses decreased exploration of a novel object within 30 min of its presentation and
showed reduced exploratory behavior in repeated trials over 5
days. Similarly, Boal et al. (2000) found evidence for habituation
to familiar surroundings over the course of 24 hr and memory of
those surroundings for a further 24 hr, suggesting that habituation
occurs on comparable time schedules in the nautilus and in octopuses.
We found clear retention of memory at 21 days posttraining in
Experiment 1, although repeated testing of subjects may have
reinforced this memory over the week after the training trials: The
21-day test occurred after the test of all subjects at 7 days post-
273
MEMORY OF VISUAL CUES IN NAUTILUS
training, so memory persisted at least 14 days without any “reminders.” A 2- to 3-week period of retention is substantially longer
than that displayed by N. pompilius in a classical-conditioning task
(Crook & Basil, 2008) and is similar to retention times seen in
octopuses (Boal et al., 2000; Boycott & Young, 1955; Young,
1961) and among other invertebrates (gastropods: Lukowiak, Adatia, Krygier & Syed, 2000; honeybees: Menzel, 1968; wasps:
Muller, Collatz, Wieland, & Steidle, 2006).
Dishabituation in response to topographical changes of surroundings (Experiment 2) indicates that nautiluses can encode
memories of complex, three-dimensional spaces. The increased
activity at the start of each trial was not simply a handling effect;
activity decreased over successive trials in the same environment
but increased when the environment changed. We found no evidence for dishabituation when only the pattern of light and dark
blocks changed but the shape of reef remained constant (when
animals experienced the switch from Reef 2 to Reef 3), suggesting
that this visual information either was not learned or was not
sufficient in itself to prompt renewed exploration of the environment. It is possible that only gross visual features, such as the
shape of the block array, were learned and that the relatively minor
differences produced by changing the pattern of black and white
blocks were not as salient as were changes to the shape of the reef.
In this experiment we did not manipulate the order of reef
presentation for different individuals because the number of subjects available was very small. It is possible therefore that the lack
of dishabituation displayed when the pattern changed may have
been an effect of repeated testing, because at this time all subjects
had experienced 10 trials rather than the 5 that preceded the change
from Reef 1 to Reef 2. Alternatively, there may be a hierarchical
order of cues necessary to elicit increased exploration, and visual
pattern changes alone, in the absence of topographical changes,
were insufficient to prompt renewed exploration. Nautiluses were
able to use a static, visual cue when it was provided as a navigational aid in Experiment 1. The beacon provided a strong associative relationship between cue and goal that probably served to
enhance the perceived value of the visual cue, in contrast to the
visual components of the reefs in Experiment 2 that were not
associated with any reinforcing stimulus.
Whereas there are clear ecological roles for spatial learning in
coleoids, it is unclear what use spatial memory may be to nautilus
in its natural environment. The few studies of the ecology of
nautiluses suggest minimal value for learned behaviors in comparison with the clear benefits to many of the coleoid cephalopods.
Retaining memories of features of their environment may be useful
for identifying areas in which animals have foraged recently, to
identify the limits of a suitable foraging zone, or to return to
familiar shelter locations between foraging bouts. In Experiment 2
we noted that time spent in close proximity to the reefs increased
with experience, suggesting perhaps that animals learned the location of appropriate shelters during initial familarization when
close contact with the reefs was less apparent. Once familiar with
their surroundings, nautiluses showed some preference for more
enclosed spaces. Individuals were consistent in their choice of
settling position across trials, although different animals settled in
different locations. This suggests a combination of learned and
innate preferences for certain shelter locations in artificial environments and perhaps also in their natural environment.
Unlike other wide-ranging marine species (albatross: Jouventin
& Weimerskirch, 1990; turtles: Avens, Braun-McNeill, Epperly &
Lohmann, 2003; sharks: Edren & Gruber, 2005), nautiluses probably do not return over long temporal or geographical scales to
precise, known locations such as feeding or spawning grounds but
rather might use smaller-scale spatial memories in an episodic
manner during discontinuous foraging and sheltering bouts. Visual
cues are likely to be of most use during foraging excursions into
shallow water but perhaps also during deeper stages of the daily
migratory cycle to locate and occupy previously used shelters.
Further studies of both captive and wild nautiluses are needed to
shed more light on the roles for spatial learning and memory in the
life of this ancient species.
References
Agin, V., Chichery, R., Maubert, E., & Chichery, M. P. (2003). Timedependent effects of cycloheximide on long-term memory in the cuttlefish. Pharmacology Biochemistry and Behavior, 75, 141–146.
Agin, V., Dickel, L., Chichery, R., & Chichery, M. P. (1998). Evidence for
a short-term memory in the cuttlefish, Sepia. Behavioural Processes, 43,
329 –334.
Alves, C., Boal, J. G., & Dickel, L. (2007). Short distance navigation in
cephalopods: A review and synthesis. Cognitive Processing, 9, 239 –
247.
Alves, C., Chichery, R., Boal, J. G., & Dickel, L. (2007). Orientation in the
cuttlefish Sepia officinalis: Response versus place learning. Animal
Cognition, 10, 29 –36.
Avens, L., Braun-McNeill, J., Epperly, S., & Lohmann, K. J. (2003). Site
fidelity and homing behavior in juvenile loggerhead sea turtles (Caretta
caretta). Marine Biology, 143, 211–220.
Basil, J. A., Hanlon, R. T., Sheikh, S. A., & Atema, J. (2000). Threedimensional odor tracking by Nautilus pompilius. Journal of Experimental Biology, 203, 1409 –1414.
Boal, J. G., Dunham, A. W., Williams, K. T., & Hanlon, R. T. (2000).
Experimental evidence for spatial learning in octopuses (Octopus bimaculoides). Journal of Comparative Psychology 114, 246 –252.
Boycott, B. B., & Young, J. Z. (1955). A memory system in Octopus
vulgaris Lamark. Proceedings of the Royal Society of London B, 143,
449 – 480.
Carlson, B. A., McKibben, J. N., & DeGruy, M. V. (1984). Telemetric
investigation of vertical migration of Nautilus belauensis in Palau
(Western Caroline Islands, Pacific Ocean). Pacific Science, 38, 183–188.
Crook, R. (2008). Behavioural correlates of learning and memory in
chambered nautilus, Nautilus pompilius. Unpublished doctoral dissertation, City University of New York, New York.
Crook, R., & Basil, J. (2008). A biphasic memory curve in the chambered
nautilus, Nautilus pompilius L. (Cephalopoda: Nautiloidea). Journal of
Experimental Biology, 211, 1992–1998.
Edren, S., & Gruber, S. (2005). Homing ability of young lemon sharks,
Negaprion brevirostris. Environmental Biology of Fishes, 72, 267–281.
Fiorito, G., & Chichery, R. (1995). Lesions of the vertical lobe impair
visual discrimination learning by observation in Octopus vulgaris. Neuroscience Letters, 192, 117–120.
Forsythe, J. W., & Hanlon, R. T. (1997). Foraging and associated behavior
by Octopus cyanea Gray, 1849 on a coral atoll, French Polynesia.
Journal of Experimental Marine Biology and Ecology, 209, 15–31.
Gallistel, C. R. (1990). The organization of learning. Cambridge, MA:
MIT Press.
Hanlon, R. T., & Messenger, J. B. (1988). Adaptive coloration in young
cuttlefish (Sepia officinalis L.): The morphology and development of
body patterns and their relation to behaviour. Philosophical Transactions of the Royal Society of London B, 320, 437– 487.
274
CROOK, HANLON, AND BASIL
Hanlon, R. T., & Messenger, J. B. (1996). Cephalopod behaviour. Cambridge, England: Cambridge University Press.
Healy, S. (1998). Spatial representation in animals. New York: Oxford
University Press.
Hochner, B., Brown, E. R., Langella, M., Shomrat, T., & Fiorito, G.
(2003). A learning and memory area in the octopus brain manifests a
vertebrate-like long-term potentiation. Journal of Neurophysiology, 90,
3547–3554.
Hochner, B., Shomrat, T., & Fiorito, G. (2006). The octopus: A model for
a comparative analysis of the evolution of learning and memory mechanisms. Biological Bulletin, 210, 308 –317.
Hvorecny, L. M., Grudowski, J. L., Blakeslee, C. J., Simmons, T. L., Roy,
P. R., Brooks, J. A., et al. (2007). Octopuses (Octopus bimaculoides) and
cuttlefishes (Sepia pharaonis, S. officinalis) can conditionally discriminate. Animal Cognition, 10, 449 – 459.
Jouventin, P., & Weimerskirch, H. (1990). Satellite tracking of wandering
albatrosses. Nature, 343, 746 –748.
Karson, M. A., Boal, J. G., & Hanlon, R. T. (2003). Experimental evidence
for spatial learning in cuttlefish (Sepia officinalis). Journal of Comparative Psychology, 117, 149 –155.
Kuba, M. J., Bryne, R. A., Meisel, D. V., & Mather, J. A. (2006).
Exploration and habituation in intact free moving Octopus vulgaris.
International Journal of Comparative Psychology, 19, 426 –238.
Long, T. M., Hanlon, R. T., Ter Maat, A., & Pinsker, H. M. (1989).
Non-associative learning in the squid Lolliguncula brevis (Mollusca,
Cephalopoda). Marine Behaviour and Physiology, 16, 1–9.
Lukowiak, K., Adatia, A., Krygier, D., & Syed, N. (2000). Operant
conditioning in Lymnaea: Evidence of intermediate and long-term memory. Learning and Memory, 7, 140 –150.
Mather, J. A. (1988). Daytime activity of juvenile Octopus vulgaris in
Bermuda. Malacologia, 29, 69 –76.
Mather, J. A. (1991). Navigation by spatial memory and use of visual
landmarks in octopuses. Journal of Comparative Physiology A: Neuroethology, Sensory, Neural, and Behavioral Physiology, 168, 491– 497.
Mather, J. A. (2008). Cephalopod consciousness: Behavioural evidence.
Consciousness and Cognition, 17, 37– 48.
Menzel, R. (1968). Das Gedachtnis der Honigbiene fur Spektralfarben [The
memory of the honeybee for spectral color]. Zeitung für Vergleichende
Physiologie, 60, 82–102.
Messenger, J. B. (1971). Two-stage recovery in Sepia. Nature, 232, 202–
203.
Müller, C., Collatz, J., Wieland, R., & Steidle, J. L. M. (2006). Associative
learning and memory duration in the parasitic wasp Lariophagus distinguendus. Animal Biology, 56, 221–232.
Muntz, W. R. A. (1987). Visual behavior and visual sensitivity of Nautilus
pompilius. In W. B. Saunders & N. H. Landman (Eds.), Nautilus: The
biology and paleobiology of a living fossil (pp. 231–240). New York:
Plenum Press.
O’Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map.
Oxford, England: Oxford University Press.
Papini, M. R., & Bitterman, M. E. (1991). Appetitive conditioning in
Octopus cyanea. Journal of Comparative Psychology, 105, 107–114.
Poucet, B., Chapuis, N., Durup, M., & Thinus-Blanc, C. (1986). A study of
exploratory behavior as an index of spatial knowledge in hamsters.
Animal Learning and Behavior, 14, 91–100.
Rice, W. R. (1989). Analyzing tables of statistical tests. Evolution, 43,
223–225.
Safi, K., Seid, M. A., & Dechmann, D. K. N. (2005). Bigger is not always
better: When brains get smaller. Biology Letters, 1, 283–286.
Shettleworth, S. J. (1998). Cognition, evolution and behavior. New York:
Oxford University Press.
Shigeno, S., Kidokoro, H., Tsuchiya, K., Segawa, S., & Yamamoto, M.
(2001). Development of the brain in the oegopsid squid, Todarodes
pacificus: An atlas up to the hatching stage. Zoological Science, 18,
527–541.
Shigeno, S., Sasaki, T., Moritaki, T., Kasugai, T., Vecchione, M., & Agata,
K. (2007). Evolution of the cephalopod head complex by assembly of
multiple molluscan body parts: Evidence from Nautilus embryonic development. Journal of Morphology, 269, 1–17.
Simonds, V., & Plowright, C. M. S. (2004). How do bumblebees first find
flowers? Unlearned approach responses and habituation. Animal Behaviour, 67, 379 –386.
Soucier, C. P., & Basil, J. A. (2008). Chambered nautilus (Nautilus
pompilius pompilius) responds to underwater vibrations. American Malacological Bulletin, 24, 3–11.
Teichert, C. (1988). Main features of cephalopod evolution. In M. R.
Clarke & E. R. Trueman (Eds.), Paleontology and neontology of cephalopods (pp. 11–79). San Diego, CA: Academic Press.
Teyke, T. (1989). Learning and remembering the environment in the blind
cave fish Anoptichthys jordani. Journal of Comparative Physiology A:
Neuroethology, Sensory, Neural, and Behavioral Physiology, 164, 655–
662.
Tourtellot, M. K., Collins, R. D., & Bell, W. J. (1991). The problem of
movelength and turn definition in analysis of orientation data. Journal of
Theoretical Biology, 150, 287–197.
Walker, J. J., Longo, N., & Bitterman, M. E. (1970). The octopus in the
laboratory: Handling, maintenance, training. Behavioral Research Methods and Instrumentation, 2, 15–18.
Ward, P. D., Carlson, B. A., Weekley, M., & Brumbaugh, B. (1984).
Remote telemetry of daily vertical and horizontal movement by Nautilus
in Palau. Nature, 309, 248 –250.
Ward, P. D., & Saunders, W. B. (1997). Allonautilus: A new genus of
living Nautiloid cephalopod and its bearing on phylogeny of the Nautilida. Journal of Paleontology, 71, 1054 –1064.
Welker, W. I., & Welker, J. (1958). Reaction of fish (Eucinostomus Gula)
to environmental changes. Ecology, 39, 283–288.
Wells, M. J. (1964). Detour experiments with octopuses. Journal of Experimental Biology, 41, 621– 642.
Young, J. Z. (1960). The failures of discrimination learning following the
removal of the vertical lobes in Octopus. Proceedings of the Royal
Society of London B, 153, 18 – 46.
Young, J. Z. (1961). Learning and discrimination in the octopus. Biological
Reviews, 36, 32–96.
Young, J. Z. (1965). The central nervous system of Nautilus. Philosophical
Transactions of the Royal Society of London B, 249, 1–25.
Young, J. Z. (1991). Computation in the learning system of cephalopods.
Biological Bulletin, 180, 200 –208.
Received June 12, 2008
Revision received February 6, 2009
Accepted February 9. 2009 !