Roman II:
Planning Studies
(Experimental design)
Clam Dancing!?!?!?
The life of a typical garden variety clam would seem to be pretty uneventful. In general, people who study marine invertebrate biomechanics (e.g. clam locomotion) are of the opinion that ocean wave motion erodes the sand around a clam and then propels it toward the beach. However, this simple theory may not hold for northern coquina clams (Donax fossor) in North Carolina. These creatures may attempt to stay near the water’s edge by, in effect, “surfing.” These clams have a muscular tongue-like foot and it may be they use it to propel themselves toward the beach – in effect, surfing – to where the waves break and stir up the plant and animal material that is the clam’s food.
As one can imagine the surf propels both live and dead clams toward the beach, and an experiment is envisioned using the “life status” of the clams as an explanatory variable. The general idea is that if clams have a developed locomotive capability and ride the waves toward shore, the live clams should get further toward shore than the dead clams. The experimental plan is to place the clams in the sand at different locations near the surf and observe their progress toward shore.
The experimental location exists for roughly 100 meters along the beach, and about 20 meters toward shore from the ocean at low tide. The elevation of the shore decreases slowly at a fairly constant rate over the 20 meters. Over the 100 meters the beach ranges from large plant density in the North to a more rocky terrain in the South. Tides come in and go out twice a day, once during daylight hours and once during nighttime hours.
Your task is to design an experiment to address this question: do the coquina clams propel themselves toward the food source? Be particularly clear about how you assign the clams their initial positions and the times of positioning in the sand. Since the clams may not move very far during the experimental period, you will need to be precise in your description of the measurement process. You may assume that identifying the clams with a life status and number, e.g. “L-#” and “D-#”, will not attract any predators, and that the experiment is of short enough duration that the likelihood of experimental mortality is vanishingly small.
Wise Old Owl vs. Bambi Mouse
Rodents in desert communities in the eastern part of Southwestern North America have two fundamental problems: (a) finding food and (b) avoiding becoming food. It may be that these two problems are related; where rodents forage may depend on their “strategies” for avoiding predation. Small rodents such as the deer mouse (Peromyscus maniculatus) generally seem to spend more of their search efforts in spaces under bushes and other desert cover rather than foraging in open spaces where seeds would be easier to locate visually. Ecologists have two theories about why deer mice look for food in different areas.
The first theory is that the deer mouse is able to climb and maneuver in bushes, and searching for seeds there is less tiresome than scampering across open spaces. The second theory is that their foraging strategy is the result of attention to predation risk. This hypothesis suggests that a deer mouse in the open is a proverbial “sitting duck” and therefore would choose to forage in dense vegetation to avoid predation.
An investigator is designing an experiment to evaluate these two theories. A fenced-in rectangular pen approximately 10 acres in size has been constructed in the Nevada desert. The plan is to release deer mice individually into the pen and record the amount of time spent foraging in the open. The pen will contain a rectangular grid with small boxes of seed placed at intervals. These boxes will be placed in the middle of each open space and under each brushy space. That way, an open space is always presented close to a brushy space, and a brushy space is always presented close to an open space. This would suggest to the deer mouse that all foraging options are open at all times. Each deer mouse will forage in the pen for 12 hours, 6:00pm to 6:00am, and the amount of time spent foraging in the open (as opposed to foraging in the brushy parts of the pen) will be measured.
A natural nocturnal predator of rodents in the Nevada desert is the long-eared owl (Asio otus). They are buoyant fliers, gliding noiselessly even when flapping their wings. They maneuver well and can fly through fairly dense brush, low to the ground, listening for prey. Experimental evidence has previously shown that the long-eared owl is an effective predator, especially in the light of the moon. This is thought to be because (a) their prey is easier to see, and (b) the prey cast shadows, which increase the likelihood of being spotted from the air. A sound system has been created with speakers placed in the corners of the pen, creating treatments of “owl-present” and “owl-absent.” In the “owl-present” treatment the owl’s mating calls will be sounded at random times during the evening, suggesting the presence of owls to the deer mice. The set up of the pen is diagrammed below, showing the speakers as owl icons. The locations of the brushy areas are shown as little plant icons.
The investigators expect that if the first theory – brush preference – is true, the deer mice should spend about the same time in the open with the owls present as when the owls are absent. On the other hand, if the second theory – predation risk reduction – is true, there should be a difference in the amount of time spent in the open over the 12-hour period depending on whether owls are thought to be present or not. Because each trial is run only at night for a 12-hour period, only one trial can be run per 24-hour cycle. The experiment will consist of a sequence of 60 trials, and there is some concern that over the length of the experiment (60 days) the moon will provide different illumination over the pen. It is felt that the deer mice, knowing from experience that the owls are better predators during moonlight, might forage differently over different phases of the moon.
Because of the large number of potentially confounding variables, the investigators are not sure how to assign the treatments over the course of the experiment, and have asked you for assistance. You may order any needed random number generator from the Charles River Experimental Design Obligato (CREDO) store at a nominal cost.
Golly, Toto, these don’t look like corn plants!
Canary Island palms (Phoenix canariensis) may live for as long as 100 years. These massive trees grow up to 60' tall, and support a huge crown of over 50 leaves that can reach 18' in length. They may be planted in warm areas in the western U.S. in places such as Arizona, California and Nevada. As an example, Palm Springs, CA, has almost 3,000 palm trees, maintained by the city, many of which are planted along streets to enhance the natural scenic beauty.
With proper care, fertilization and water, Canary Island palms are very hardy – just the sort of tree you would want to live in if you were a bud rot fungus (Phytophthora palmivora). Bud rot fungus is spread from tree to tree by contaminated garden tools, and unfortunately may actually kill the palm tree. This, of course, is not appreciated by those who enjoy the Canary Island palm’s beauty.
Bud rot fungus can be controlled by two methods, chemical and non-chemical. Chemical controls such as mancozeb or basic copper sulfate have been shown to be effective in battling bud rot fungus, but use of chemicals is not advisable in windy open areas. Another strategy for controlling bud rot fungus is the “virgin soil” technique. With this technique, soil in which palms have never grown is placed around the roots of the palm trees, protecting the roots from the fungus. This method has the advantages of being nonhazardous.
The City of Palm Slinky (sister city of Palm Springs) is considering planting Canary Palms along its main street, but has not decided whether to use the chemical or non-chemical treatment. They would like you to design an experiment to see which method is more effective at preventing bud rot fungus. Their main street – where the City fathers and mothers wish to plant the palms – runs for about 5 miles through a variety of neighborhoods, both business and residential, and has varying amounts of traffic at different locations. They are planning on having trees on both sides of the road.
Your task is to design an experiment to compare the two methods of bud rot fungus control. The City fathers and mothers will fund 60 trees for experimental purposes, which of course can be ordered from the Charles River Experimental Design Obligato (CREDO) store at a nominal cost.
Bambi vs. Corn
The typical garden-variety deer (Odocoileus spp.) can be a serious problem even beyond the garden. Deer are thought to cause more damage to agricultural products than any other species of wildlife. In 1993 over $30 million worth of corn was lost to deer just in the 10 largest corn-producing states. In late June and early July corn reaches the silking–tasseling stage of growth. Cornfields are at particular risk during the silking/tasseling stage because (a) the corn is very tasty, and (b) damage during this period seriously reduces yield.
It might be possible for farmers to minimize the amount of damage to crops by using visual and/or acoustic frightening devices during the silking/tasseling stage of corn growth. These devices range from (a) less than distressing to (b) terror inducing.
(a) (b)
One frightening device under consideration is a bio-acoustic device that would mimic animal communication signals such as alarm calls. (An alarm call warns others of possible danger.) These devices typically consist of an infrared detection system that activates an audio component that transmits recorded alarm calls. Researchers have studied the effects of such devices on birds, but little is known of their effect on mammals. If effective with deer, bio-acoustic alternatives would meet with public approval and in addition could be used in both rural and urban settings.
Deer are most active at dusk and dawn, and are frequent feeders at night. They tend to stay close to wooded areas for safety; farms in Iowa are frequently interspersed with wooded areas.
You are to design an experiment to collect data that would allow you to test the hypothesis that sound devices can reduce deer predation on corn. For purposes of this experiment every right-thinking Iowa farmer has volunteered his acreage for use as test plots; you have complete freedom to choose cornfields for treatment and control plots. Not only that, every right-thinking Iowa parent has volunteered their teenagers’ boom boxes (in hopes that you are conducting a multi-year experiment.) These boom boxes came in various brands, with the Aiwa CD player (model CDC-X217) being the most popular. Infra-red animal-activated alarm and distress call systems have been purchased from the Charles River Experimental Design Obligato (CREDO) and the corn yield will be measured by yield monitors linked to Global Positioning Systems on harvesting equipment.
Fly vs. Rainbow
The typical intrepid trout fisher-person uses artificial floating lures called “dry flies.” The general idea is that these flies, artfully made with wool, fur, and feathers to mimic adult prey, are cast onto the waters and confuse the typical Rainbow Trout (Oncorhynchus mykiss) just enough to chomp soundly on the fly.
The optimal design of dry flies is a matter of controversy . The consensus is that flies must be properly presented to the Rainbow before it can be tempted to strike. However, some skeptics claim that the trout is less discriminating than the fisherperson and may not actually be able to tell one dry fly from another. In order to address this important social issue, two commercial dry flies (shown at right) have been selected as distinct treatments.
When fishing with a dry fly, the process is to use a big stick (called a pole) with a small rope (called a line) that has a dry fly affixed at one end. The dry fly is thrown (by a motion called a “cast”) to a certain position in the water. This throwing motion is replicated until a trout chomps (known as a “strike”) on the dry fly.
Rainbow trout are hatched in streams and remain there until they reach 6” – 9” in length, at which time they will travel to lakes or oceans to bulk up before returning to their stream or river to spawn. They tend to prefer clear, flowing waters with turbulence (for oxygenation) and are found in streams with gravel, rock, and sandy bottoms. Three relatively inexperienced Wisconsin fishermen – who by purest chance happen to be the researchers -- have selected a 2-mile stretch of a Wisconsin river suspected to be fully inhabited with Rainbow trout for purposes of this experiment. The river bottom varies in turbulence, depth, texture (sand vs. clay) , and speed over the course of the 2 miles.
Your task is to design an experiment to address the question of whether fish are more likely to bite for one type of fly than the other. You should use blocking to control at least one confounding variable.
A different kind of confidence interval?
When estimating the sizes of bird populations wildlife biologists wear brightly colored clothes during hunting season as a safety precaution. This strategy may have interesting consequences for the estimation of the bird populations for different species. It is possible that some species of birds may be more easily frightened into motion -- and therefore are more visible -- because of the brightly colored vests. What is known as the “species-confidence” hypothesis states that birds prefer colors similar to their own colors, and will avoid individuals with different coloring. If this hypothesis is supported, wildlife biologists wearing orange vests might overestimate the number of -- say -- American robins (Turdus migratorius) because they see more of them in their sample, but underestimate the number of -- say -- Carolina Chickadees (Parus carolinensis) because they fly away sooner, undetected, when they see the orange vests.