My life with archerfish chapter 2

Posted on January 2nd, 2013 in Brain Function,Education,Science by Robert Miller

Archerfish deja vu: I revisited my passion for archerfish in 1986, while living in St. Louis. By that time we had 2 young sons and I wanted to repeat for them what I had gone through with these remarkable fish to see if something ignited within them as it had for me. Surely every father wants to recreate some part of his own childhood for his children; at the time I considered it like a solemn parental duty, a kind of epigenetic contractual commitment to a vaguely defined ancestral linkage lined with our ancient past, going back at least as far as electric trains. I decided to fulfill this deep-seeded ancestral urge by creating a second round with archerfish. But while I wanted to repeat the experience, I also wanted to re-create it on a larger scale, which in this case amounted to a much larger aquarium to accommodate more fish and a much fancier superstructure above the tank to serve as the temporary residence for houseflies.  When this idea took root, we had just moved into a larger house with more room, including space for a large aquarium in the basement. The archerfish plan was not a radical departure from a journey we had already started. We had a fairly large aquarium that we acquired when we bought the house as part of the sale agreement with the prior owner and my youngest son had angel fish and neon tetras in a separate smaller tank; as the angel fish grew we would transfer them into the big aquarium. My oldest son had a turtle and a flop-eared rabbit which we had to give away to a rabbit dealer because it consistently bit people and drew blood in the process. It was a pleasant-looking but evil rabbit. Our new house was in Clayton, Missouri, just outside of the city and close to where I had my laboratory at Washington University. For those of you who didn’t catch chapter 1, the archerfish is a truly unique fish, that lives primarily in the brackish waters of India, Australia, the Philippines, and throughout Polynesia; they can spit a volley of water droplets to knock insects off of their perch on low hanging plants or branches and they have uncanny skills for accuracy and success. You can see a video of the archerfish firing its water cannon at a fly, knocking it off its perch and then skillfully jumping out of the water to insure the privacy of its catch. They have sleek V-like bodies which they propel gracefully and efficiently through water. They are fast swimmers and have to be in order to make sure one of their neighbors doesn’t get to the prize first. Although there are several different species of Archer fish, they all belong to the same family of fish Toxotidae, which means bowman or archer. Although one popular story about the discovery of archerfish relates to the story of a British officer in India having his cigarette put out above the water by an archerfish,  in fact the archerfish was first brought to the attention of the West in 1767, a few years before the United States was a country.

Improved environment: The second time around I wanted to improve on the home environment for the archerfish. I acquired a 125 gallon, all glass aquarium and created a brackish water habitat. In the process of establishing the aquarium, I had a machinist friend build an acrylic superstructure that fit tightly onto the top of the aquarium, adding significantly to the height of the overall structure; the superstructure was as tall as the aquarium itself and on top of the acrylic cover, I configured two long lights that stretched nearly the whole length of the aquarium, while leaving space at the ends to screw in mason jars.  At each end of the of the top cover, I carved out fittings for mason jars so that you could screw them in tight.  I called around the pet stores and found one that sold archerfish; I  bought 4 four of them and 2 Mono Sebae, as they also prefer brackish water. The Archer fish began as small 2-3″ fish, but within 6 months grew to 6 ” and the Mono Sebae also grew to be sizable. In addition, living in brackish water seemed to enhance the contrast between the light and dark strips of the archers, who displayed progressively sharper borders between the vertical stripes along their body. In time, they proved to be truly stunning fish, even without the additional attribute of owning their powerful, personal water cannon. With gorgeous, shimmering scales, they presented a striking image of a uniquely adapted fish. One might naturally ask how does a personal squirt gun get encoded into the archerfish DNA?

Demand for houseflies: The Mono Sebae were satisfied with ordinary “flake” fish food, but to get the archerfish performing as I wanted, I had to find a source of house flies—the perfect food for spitting archers, as I had discovered many years ago. I finally located a nearby  barbershop that let me use my mason jar technique that I developed as a kid. I told them it was for a scientific set of experiments, conveying a more adult-like nature to the quest. I was going to use high speed photography with a strobe light to see if the archerfish had their eyes above water when they invoked their water cannon. If they did, it would eliminate the problem of the correcting for the refraction of light at the air water interface, but if the eyes remained underwater when shooting, then the archerfish must learn to correct for the real vs imaginary image above the water (I did in fact have that in mind, but later discovered that the problem had already been resolved as the archerfish learns to spit at the insect, with its eyes below the water in which case it must learn to compensate for the refraction of light by the water which it readily learns to do; see below).

Success: In no time the archerfish were spitting at the flies that I caught and my sons were stunned to see these unusual fish, though I never got them interested in catching flies—that seemed quite  yucky. They periodically brought their friends over to watch the archerfish feeding behavior and, as long as flies were available, the archerfish never disappointed. The fly supply issue eventually turned into a shortage of flies. As one might imagine, fly catching at the barber shop soon become tiresome and it was apparent that this low-yield technique was incompatible with holding a day job. I had to find an alternative method of fly procurement. We were living in an urban environment, so there was limited availability for other kinds of insects and I had already concluded that flies were the perfect food for archerfish–they were good at flying and possessed a fast, visually-based escape system that made them hard to catch, as anyone knows if they try to swat a fly either resting on a surface or flying in the air. Despite the formidable defensive reflexes of the house fly, they proved to be no match for warding off an attack by the archerfish.

Fig 1 The problem with refraction for the archerfish

Fish action: When flies were placed into the plastic cage above the aquarium, all the archerfish were aggressively engaged in knocking them off their perches and devouring them and did so repeatedly  until all the flies were eliminated. When flies were in the plastic chamber it was a feeding frenzy, with fish jumping and jets of water squirting into the plastic dome, until suddenly, once the flies had all been eliminated, things quieted down, as the archerfish settled into their normal swimming behavior, in which they were always moving in the tank, as if preparing for another round. They are a high energy fish. It seemed that the fly was helpless in successfully dodging the fast volley of drops or jet stream of water from the archerfish; it couldn’t see it in time to avoid getting drenched and the drenching power of the water cannon sufficiently wet the wings so that the fly was unable to open them in time to engage in flight before hitting the water. It wasn’t at all clear if the flies even saw the jet stream moving towards them, as the image of the water stream would be of low contrast against the surrounding environmental background—nothing like a bright red fly-swatter. The flies in essence were doomed—they had no natural means of becoming anything other than a meal for the persistent, sure-squirting archerfish. Once the fly hit the water, there was no sense of ownership among the archerfish–no loyalty among hunters; every fly was free game, yet because the fish typically positioned itself in relation to where the fly was likely to land, it was usually the fish that knocked the fly from its perch that devoured the same fly. This was another uncanny part of their motor behavior. Positioning and arching its body to accelerate towards the fly to meet it when it hit the water was another component to the archerfish strategy that typically resulted in success in both knocking the fly off its perch and devouring it with a very fast accelerating movement that was almost too fast to fully visualize. Archerfish have some of the fastest reflex acceleration times every measured in teleost fish and typically jump into the air to make sure they dominate the competition for swallowing the fly as their just reward for knocking it into the water. Although archerfish would take dry flake fish food, they preferred to knock flies from their perches and learned in time to begin squirting water while the flies were still in the just-attached mason jar. Similar to my first experience with archerfish, initially they were shy and went into hiding when I would appear. But, in time, as they learned to associate my presence with a new batch of flies, they aggressively positioned themselves for spitting behavior. When the fly-holding apparatus was first assembled, the introduction of flies through the mason jar did not lead to water cannon activity until the flies emerged from the jar and came into positions where the fish had an easier time of visualizing them. As the archers became more skilled and more competitive, they learned to release their water volleys into the mason jar, revealing a rather amazing adaptability, because vision was distorted by the shape of the mason jar and it also had some regions where the fly could land that were inaccessible for a strike. When archerfish learned to begin firing their water cannons at flies in the mason jar, they positioned themselves almost vertically below the mason jar, virtually eliminating the refractive error problem of the air-water interface which does not come into play for light that is normal to the water surface. There is no evidence to suggest however that positioning themselves vertical to the target improves their accuracy (figure above and see below).

A new source for flies: As I looked for a stable, more convenient source of flies, I was unable to find a local, commercial source. No one in St. Louis ran a fly take-out center or delivery system. So, I turned to Carolina Biological and ordered a fly culture kit, that was complete with fly larvae, culture medium for raising them and a net that fit over a small tub to prevent the flies from getting out, with little glove-like partitions for inserting your hands without exposing the flies to the outside air.  Fly culture was entirely isolated from the outside environment or so I thought at the time. This turned out to be the solution to my problem for a ready source of house flies. Not only did this “kit” produce a continuous source of flies, but I took comfort in knowing that the flies did not provide an unknown contagion that might wipe out the fish, something that might occur with flies from the barbershop where one didn’t know their dietary  history. It took about two to three weeks to generate adult flies, as they went from larvae, to pupae to eclosion (emergence of adult flies). But the kit allowed me to establish a steady supply of flies and by feeding the adults once eclosion took place, the flies grew to be sizable and an even better source of food for the fish.

Fig 2 Image of an archerfish evoking its water cannon. Image superimposed on eye illustrates the cone density of the retina and shows that in the ventral retina, cones exceed 50,000 cones/mm2. Similar color-coded image to the right illustrates ganglion cell density. Note that a small region in the ventral retina has the same ganglion cell density as that determined for cones. This region provides the archerfish with high acuity that rivals that of our own visual system. From Ben-Simon et al in the Journal of Vision (http://www.journalofvision.org/content/12/12/18)

The science of archerfish spitting: Within the past decade, scientists have become increasingly interested in deciphering the neural code that archerfish use to execute their remarkable behavior. This includes detailed behavioral studies, MRI imaging techniques, detailed anatomical studies and electrophysiological recordings. All of these studies have led to an enhanced appreciation of the adaptive skill that archerfish bring to the dinner table. Several interesting conclusions about their behavior have been revealed. First, it has been established for some time that archerfish keep their eyes in the water when they invoke their water cannon (Fig 2). This creates a dilemma for accuracy because the image of the fly will be bent by the refraction of light at the air-water interface and the image of the fly as seen under water will be higher than it actually is, as illustrated in Fig 1. This refractive error could only be eliminated if the archerfish was directly below the target, because light passing into the water from the vertical is not refracted. Therefore if the archerfish aligned itself perpendicularly with its target, the real and refracted images would be the same. I never saw archerfish align themselves vertically with their target, except when they tried to align themselves with the mason jar when I added a fresh source of flies. It is well-known that archerfish have uncanny accuracy and their hit rate is no different up to about 40 degrees of target angle from the vertical. The archerfish learns, presumably through trial and error, how to compensate for the refraction of light. This is only one of the amazing characteristics of archerfish behavior. These animals also have their eyes fairly close together and actually use binocular vision to estimate target depth, just as we do, except the archerfish has the additional burden of learning to adjust and compensate for the refraction of light. Even more remarkable is the presence of a specialized region of the retina that receives light from above the water. This region of the ventro-temporal retina is referred to as the “Area Centralis” which is a rod-free region of the retina that contains the highest density of cones, with three different cone pigments, implying that archerfish have trichromatic color vision, just as most of us do. Even more surprising is the presence of a large number of ganglion cells in that region (the cells that communicate with the brain): with an amost 1:1 cone to ganglion cell density, it means that the archerfish does not degrade the high visual acuity provided by the rich density of cones such that his retina passes onto the brain the high cone acuity without degrading the signal and one can imagine that the signal is further refined by the retinal network, which serves as a powerful neural circuit for enhancing contrast and feature detection of objects in the visible environment.  Archer fish are relatively easy animals to study behaviorally, as they eagerly spit at small images on computer screens above the aquarium and these kinds of experiments have led to the conclusion that archerfish discriminate colors and can distinguish for example red from green, something that requires different cone photopigments, as rods do not provide color information. We are only beginning to understand the many adaptations that make the archerfish truly remarkable in visual-motor skills. The water cannon for ejecting a stream of water is achieved by compressing the gills and shooting the compressed water through a narrow groove in the lower jaw, which helps to accelerate the water stream, just as adjusting the water sprayer on a garden hose can change the force of the ejected water.

Thus far we have mostly considered the skill of the archerfish shooting at fixed targets, targets that don’t move. But recent experiments indicate that archerfish, when confronted with moving targets at different heights above the water can learn to aim their water cannon at moving targets by estimating the delay required by the speed and height of the target. This requires binocular depth perception and the ability to learn and retain target movement information, illustrating once again that archerfish are truly remarkable fish and show an adaptive skill set that exceeds all predictions about their capacity to detect and learn from the behavior of their prey. The inference is that since archerfish generally swim in schools, they may learn from each other how to improve their accuracy for moving targets, in much the same way that that a young male songbird learns to sing the male courting song by listening to his father., although this has not been experimentally verified. If the food source changed for some reason from insects who often perch and are generally stationary targets, to a flying insect population, archerfish can learn to adapt to this new source of food, even though some trial and error would initially be involved. Nothing like a little hunger to accelerate the learning process. Many other features of the archerfish have appeared in a paper by Ben-Simon et al in the Journal of Vision.

The great escape: While I thought that I had conquered the major problems in procuring an adequate supply of flies and felt that the archerfish arrangement was flourishing and healthy, I had not anticipated what would happen next: not too long after I had established my own fly colony, I got a call from my wife, who said in the most stern, but panicked manner “Bob, you must come home immediately, there are flies all over the ceiling of our kitchen and the family room.” I took this seriously and promptly headed out the door and arrived at home in about 15 minutes after the call. I looked up at the ceiling of our kitchen and then in our family room to see a huge density of flies that made the ceiling look like it was decorated by a heavy dose of black pepper. Oddly enough, the flies had spaced themselves as if each fly was allowed a certain level of territory, so they were evenly distributed across the ceiling of the two rooms and some were also in the living room. Understandably, my wife was livid with the situation and immediately thought of health issues related to our sons.  Needless to say, she wanted to know how I was going to get them off the ceiling. My youngest son came home from school, quickly looked up at the ceiling loaded with flies and promptly left to go outside and play. Fortunately, I had a handy shop vacuum suited for this kind of task. The flies were relatively young and, at the moment, they were content to remain on the ceiling, with very little flying, so I was able to initiate the long process of cleaning up by getting on a small step ladder, and using a shop vacuum, began to suck them off the ceiling.  I didn’t count the flies, but my guess was that several hundred were on the ceiling and that implied that most if not all them had managed to escape the confines of my netting arrangement that was designed to keep them contained within a net-covered tub. Somehow the integrity my security system for the fly colony had obviously been breached. This was long before 9/11, so I didn’t think that the security breach might be the result of terrorist activity that could now be invoked for such events. Since 9/11, with such a massive release of flies on might be tempted to call Homeland Security and report a massive fly release with an unknown but highly lethal contagion that could threaten human health. I did find a loosening of the net around the edges of the tub that was the obvious source of the security breech. It didn’t prove too difficult to suck most of the flies with the vacuum cleaner, though I did have a problem with disposing the flies once they were internal to the vacuum cleaner. As anyone with a shop vacuum knows, the captured flies were within the large volume of the vacuum tub, not in a vacuum bag that could be easily disposed of in the garbage. The flies in essence had been transferred from the ceiling of our kitchen and family room to the interior of the shop vacuum, but if I released the lid to the vacuum, the flies would emerge again for round two of the fly fiasco. I decided that the only rational thing to do was take the shop vacuum outside, remove the lid and release the flies, at least those files that survived the suction ordeal. Granted, I was releasing flies into the neighborhood that regarded the fly as an evil pest, so I was very secretive about my actions; I removed the lid and left the canister open for several hours allowing all the flies to escape. I rationalized that the flies would be loaded with shop dust and thus be easy targets for a fly swatter should they entire into private domiciles. From start to finish this process took several hours. Full recovery from this event did not take place. I reduced the scale of my fly keeping project, and managed not to repeat the great fly escape, while keeping enough flies around to continue with the archerfish project.

The final chapter: It was not too long after the fly escape that I was recruited to the University of Minnesota; we moved there the summer of 1988. I was ambivalent about what to do with the archerfish and the Mono Sebae. They had become more like precious household pets and each fish seem to have developed its own personality; I wasn’t sure I wanted to part with them. I made an arrangement with the owner of the aquarium store where I first bought the fish. Our agreement was that I dropped the fish off to his store, but he would not sell them for six weeks. Before the six weeks, if I requested them, he would ship them to Minneapolis, but after six weeks the fish were his. After we arrived in Minnesota, I realized that my new position as Chair of the Physiology Department in Minnesota would be much more demanding than my position at Washington University and it became clear that I wouldn’t have the time to recreate what I had done in St. Louis with archerfish. Sadly, I closed that chapter with archerfish and began a new life with more academic responsibilities. Our myriad of pets condensed into a single dog. To this day, I avidly read the increasing number of scientific papers on the neuroscience of archerfish behavior and continue to marvel at the emerging evidence which suggests that they are avid learners and skilled tacticians for shooting objects of prey into the water with their impressive personal water cannons guided by an amazing adaptive behavior. If Darwin had known about the archerfish retina, he surely would have included it in his writing as an example of natural selection refining the eye to fit a marvelous survival niche for an elegant fish.

RFM

    Print This Post Print This Post

Comments are closed.