Octopus (24 page)

There's another constraint on learning that could have led to these failures to solve the detour problem. The octopuses had to learn that a crab seen through glass was really there, despite the touch information that said the crab was visible but not accessible. Maybe octopuses don't understand what glass is. This must be an unusual piece of information for them, because things visible in the ocean can just be reached out for and touched. Lots of animals don't understand the properties of glass. Birds fly into glass windows, six-month-old infants touch a glass surface over the visual cliff
but don't believe their touch information when their eyes tells them differently. This difficulty with glass could account for problems with getting common octopuses to learn to get a crab out of a glass jar by unscrewing the lid. They can do it, but they don't decrease the amount of time they spend playing around with the lid before they unscrew it. In other words, they didn't learn. Could it be that glass just doesn't make sense to an octopus and so it never learns to use its arms any better with it? Perhaps arm use in octopuses isn't as accessible to learning as it is in vertebrates because of the huge amount of local control of arm movement. Maybe only a little information about arm action trickles up the ventral nerve cords to the brain. The octopus definitely has limitations but we're not yet sure why.

Roland started us asking this question in a new way, one that reveals a lot about how science does and doesn't work. He studied a giant Pacific octopus in the Seattle Aquarium that eventually decreased the amount of time to remove the lid of a glass jar in order to eat a piece of herring inside, and so this animal did learn. In this demonstration, Roland didn't carefully wash his hands, and so chemical cues from the smelly herring were all over the outside of the jar. When we tried this more formally with a crab inside the jar, but smeared herring outside, octopuses steadily decreased the time to open the jar lid, a clear demonstration of learning. We know that when the octopus tries to take a lid off a jar, the jar is hidden inside the arm web and out of sight. Maybe the octopus can't remember what it has seen without a little reminder of chemical cues to keep it working, and information from more than one sense makes a useful combination for the octopuses and for us. After all, we look up the apple pie recipe in a cookbook, reinforced with a great picture, and watch it as we build it. But it's the marvelous smell that reminds us it's time to take it out of the oven and eat it.

While there are limitations on use of information, octopuses can learn a lot about visual and tactile stimuli. In this case, it makes sense to talk about cognition as use of information. Norman Sutherland (1960) set out to test what cues a common octopus used to learn about a visual shape. He hoped to find a simple system that narrowly analyzed incoming visual cues. First he tested octopuses on learning to discriminate vertical and horizontal bars, which they could do easily. They were less good at telling oblique bars from one another but so are mammals. They could tell mirror-image (reversed) figures from one another. Then he tried them on something harder, telling the difference in ratio of edge to area, most simply seen in discriminating V versus W. They had no problem, and bees can do
this too. Octopuses could tell if the same figure was rotated 90 degrees, though this was more difficult for them.

William Muntz thought (1999) that this simplistic approach to octopus vision wasn't working. He decided that octopuses were not just using one simple visual dimension to tell the two shapes apart. He made up two shapes that weren't different in any of the ways that Sutherland said were important for octopuses. When the octopuses discriminated this pair of shapes, he concluded we couldn't use a simple model of shape information processing. Octopuses didn't have a simple and automatic shape analyzer tucked away in the brain. They were, instead, learning what to learn each time they were given a new pair of shapes.

Discussion of what an octopus uses to sort out visual shapes leads to another fascinating aspect of learning, simple formation of a concept or idea. When octopuses were given two useful cues about a shape—brightness and orientation—at the same time, they learned faster than when there was only one cue. Later, testing on only one cue or the other showed that twenty-two octopuses had used brightness and six had used orientation. Children learn to do the same sorting of useful from useless cues. When octopuses were trained on orientation as a cue for a long time and then switched to shape, they took longer to learn it. This switching off, using an unrewarded part of a stimulus, is something humans learn too. Training with a particular comparison helps the octopuses make a difficult comparison too. When they were given a fine discrimination that they couldn't learn, they could master it if they were given more obvious differences and then finer and finer ones. Before we get too impressed, we should mention that bees also can do this. But this skill is not a trivial one, and not what we think of as simple learning. Rather it is attention or selection, learning what to learn and how to ignore other available information that turns out to be trivial.

Another situation—when looking at mirrors—may reveal something interesting about the octopus's use of information. When we look in a mirror, we know that the image we see is ourself. But what does an animal see when it looks in the mirror? Most species tested so far treat the mirror image as another animal, and aggressive males often posture and make gestures at it. James found cuttlefish doing this when they came by the window of the Aquatron tank at Dalhousie University in Nova Scotia. He said it was a good opportunity to get pictures of the aggressive zebra display. Most monkeys do the same, but Gordon Gallup (2002) has used what he calls the
spot test to look for self-recognition. Put a colored dot on a small child's or an animal's face without them knowing. If they look in the mirror and touch the spot on their own body, they know the mirror image is a representation of self. Only the primates pass this test. Even young children don't understand the spot is on them; they need to be two years of age before they touch the spot. Recent work suggests that elephants, with really big mirrors, and porpoises, arguably very smart though marine mammals without hands to use to explore, may pass the spot test too.

It's interesting to wonder whether octopuses would pass the spot test, which isn't a test of vision but a test of self-recognition, and it could be a test for consciousness. With its good vision, an octopus can clearly see the image of the octopus in the mirror. But whether it would know it is itself is hard to tell. Since an octopus does not normally sees itself or guide its arm actions by vision, what kind of information processing would it need to recognize the sight of itself? In addition, octopuses are solitary for much of their lifespan, and those we watched in Bermuda were indifferent to the sight of others of their species. They didn't seem to have a concept of their own species, and even would eat smaller ones. The fact that octopuses will be cannibals if they can, suggests that octopuses don't have an octopus-recognition template in the brain, and this would argue for them not knowing that the mirror image is themselves. But it's an intriguing test not only of dealing with information but also what kind of information matters to the animal. We found that octopuses seeing themselves in a mirror do know that there's something interesting there, but they don't seem to know that it's “me.”

While we know a lot of what an octopus can't do, by finding limits to concepts like learning types, concept formation, and self-awareness, we learn about an animal's abilities. Logically, many animals store information for use in quite narrow situations, such as prey choice or mate recognition. Bees use color information in finding flowers but not in other situations, and they use navigation just to get to flowers, no other time. Bees have pretty small brains. Caldwell, who has studied stomatopod crustaceans extensively, praises their intelligence. Others comment that members of this group, such as crabs, seem to be able to learn cues and use them only in the most apt situations. This selectivity of learning is domain specificity.

Like vertebrates, octopuses gain and use learned information much more flexibly, applying it to a new situation. In particular, consider the octopus's
use of jetting. They start out using the jet automatically for respiration and waste removal and as a means of fast transport. But soon they put jetting to use in building homes, cleaning out sand and mud, and uncovering crabs hiding in the sand. It's amusing that common octopuses jet to repel scavenging fish that take crab pieces from their middens, and it's even more amusing when they use the jet on divers like ourselves when we check the shells they have thrown out. The giant Pacific octopus in our lab used the jet to do the equivalent of bouncing a ball, when she played with the floating pill bottle. Despite limitations, variability is still the hallmark of octopuses' behavior. That's why octopuses are such fun to watch and why there's so much more to learn about them.

10

Sex at Last

M
ating and reproduction in octopuses takes place at the end of their lives, which is the case for most cephalopods. The parents provide no care for the young once the eggs hatch. The males' life work is completed once they mate, and they usually die shortly after. Females die just after their eggs hatch. So for octopuses, mating signifies the beginning of the end.

After octopuses mate and reproduce, males go into a state of senescence—a state comparable to human dementia, in which they don't behave normally. Because octopuses and other cephalopods have short lives by human standards, O'Dor suggested in 1998 that they could borrow the motto of the Hell's Angels biker group: “Live fast and die young.”

Aquarium visitors frequently ask, “Why does the octopus die so young? Look at how big it is, only three years old and already senescent.” Even the largest octopus species has a life span of three to four years at most, and for the smallest species it's six months or less. Some deep-water octopuses may live longer, but everything is slower in the cold depths. Octopuses don't really die young; they die after a full life (unless they get eaten). Their complete life spans just happen to be a lot shorter than ours. It is rather a shame, though, that an intelligent animal like an octopus dies after one or two years. We wonder what they could do with their intelligence if they lived ten or twenty years!

Before octopuses can mate and reproduce, they have to find another octopus to do it with. Attraction by sight, chemical attractants, and visits to previously occupied dens are several ways a male and a female octopus can get together.

The story of Ursula the Octopus provides clues as to how octopuses might find each other and made us believe in the use of chemical attractants by octopuses. When we released this captive-reared giant Pacific octopus from the Seattle Aquarium to allow her to mate in the wild, we made observations on her survival and behavior. There are few such long-term
observations of octopuses in the wild, partly because scuba divers are unable to stay underwater and make observations over long periods, especially in cold water.

A scuba diver had donated Ursula to the aquarium as a very small, about 2-oz. (50-g) female giant Pacific octopus on May 3, 1997. When she got larger, we placed her out on exhibit. She was an aggressive animal, so we named Ursula after the evil sea witch in the Disney film The Little Mermaid. We released her into Elliott Bay in front of the aquarium on March 12, 1999, because she was getting too large for the tank. Although her appetite was still normal, she was approaching reproductive size, 45 lb. (20 kg), and we wanted to give her a chance to mate. She had lived in the aquarium for twenty-two months.

Her release became a media circus. In addition to being watched by some 200 aquarium patrons, Ursula's release was covered by six television news crews, and we gave five radio interviews about her. Her release was featured in three local newspapers and on television's Discovery Channel.

To encourage her to crawl out of her tank, we smeared herring on the lip of the tank and down one side. She stretched a couple of arms up and followed the herring smear with her suckers, leading her to crawl out of her tank into a barrel of water on a cart, amid the screams of about a hundred children who were watching. We then wheeled her to a ramp at the water's edge in front of the aquarium and tipped the barrel so she could crawl into the water.

Aquarium divers with underwater video cameras filmed her release and her first forty minutes of freedom. She jetted about 30 ft. (10 m) to the pilings supporting a portion of the aquarium, perhaps to get out of the light, and stretched out her arms to keep contact with the pilings as she slowly drifted to the bottom 45 ft. (15 m) below. She maintained her bright red color as she drifted down. When she contacted the bottom, she changed to a camouflaging mottled pattern with brown and white colors that matched the background. She came to rest under the end of a suspended log, and was still there when the divers ran out of film and air.

In the succeeding weeks, Ursula didn't go far. Two weeks after the release, a scuba diver found her in a den, which was excavated under a partially buried log at the base of a rock pile under a finger pier at the aquarium, 90 ft. (27 m) deep. This den had previously been occupied by other octopuses. Her den midden at that time was littered with the dismembered
and empty shell remains of several kelp crabs. The den was cleaned out and the crab shells were new, so we knew she was eating well.

Two weeks later, she was still there. And near her was a larger octopus in a new den in a large piece of discarded PVC pipe approximately 90 ft. (27 m) away. Before Ursula's release, scuba divers in the area had not seen any octopuses. Since this larger octopus wasn't in the area before, we assumed he was a male and that he was attracted to Ursula. We confirmed his sex on a later dive. Two weeks later, another male appeared in the area in a different den, so it looked like Ursula was attracting suitors. These observations suggest that a mature female octopus can attract mates by releasing chemical cues.