David K Skelly. American Scientist. Volume 85, Issue 1. Jan/Feb 1997.
Consider life in a freshwater pond. There are clam shrimp, tadpoles, beetles, dragonfly larvae, mosquito larvae, backswimmers, salamanders and dozens of other species, all packed into an area smaller than your living room. Yet this pond is temporary, and in the space of several weeks the entire basin will dry, leaving little indication of a once and future thriving community.
Ponds with alternating wet and dry phases have long fascinated biologists because of the peculiar challenge they present to their inhabitants. The puzzle, of course, is how the so-called aquatic species manage to survive when the ponds dry up. Scientists have observed a number of ways in which different species deal with this paradox. Some insects and amphibians persevere by metamorphosing into terrestrial forms before the pond dries. Other groups lay resting eggs or burrow into the sediment and wait for the pond to refill. Still other species have no resistance to drying at all. Populations of these species become locally extinct with each drying event and must recolonize the pond when it refills.
The initial studies to provide answers to the survival puzzle posed by temporary-pond dwellers also led scientists to explore a more subtle, but no less important, pattern. The composition of the communities within temporary ponds is distinct from community composition within more permanent ponds. In fact, ecologists have realized that the permanence of a body of water exerts an enormous influence on the kind of animal communities that form within it.
In order to understand this pattern, it helps to think about arranging all freshwater, from the smallest puddle to the Great Lakes, along an axis of habitat permanence. When viewed this way, the distribution of aquatic organisms becomes surprisingly ordered. Most species are able to live on just a small section of the gradient using, for example, only ponds that dry up once a year to the exclusion of all others.
There is, in fact, an entire suite of species that appear to have comparable limits on their distributions. Why can’t these species live in more temporary ponds? Why can’t they live in ponds that never dry up? Why do some species coexist while others are segregated in their distributions? Such questions of distribution and community composition are some of the fundamental problems tackled by ecologists.
To get at some of these fundamental questions, I have, for the past 9 years, been studying the distribution patterns of tadpoles. Research on tadpole assemblages has given ecologists some of the clearest indications of how the physical and living components of the world interact to determine the lot of individual species. That lot, we are coming to realize, is strongly affected by both habitat permanence and predation. Through their impacts on individual species, these factors may indirectly determine the composition of communities.
Tadpoles and the Permanence Gradient Like many aquatic animals, tadpoles are faced with a time problem. They must reach a critical stage-metamorphosis, in this case-before their pond dries up. Among North American tadpole species, the developmental period can last anywhere from 2 weeks to 3 years. Obviously, species with long larval periods cannot successfully breed in ponds with short hydroperiods, and a number of studies have shown that species with longer development times tend to breed in more permanent habitats. If maximum development time were the only constraint on the composition of tadpole assemblages, we might expect a gradual increase in species richness as we proceed from ephemeral to permanent habitats, because permanent ponds would be able to include both temporary and permanent pond species. In fact, this is not the pattern that is observed. The most permanent ponds may have just a single tadpole species compared with up to a half dozen in some temporary ponds. This striking difference in tadpole assemblages requires another explanation.
The exclusion of many species from more permanent ponds may be the result of a second relationship with pond permanence: Predators, like their prey, are also distributed along the permanence gradient. In the most temporary habitats there may be few or no predators at all. Those that are present tend to be small. As the time constraint relaxes in more permanent ponds, more and larger predators are found. These include several types of insects, such as larval beetles and dragonflies, as well as salamanders. In fully permanent waters, animals that require water year round-notably fish-become important predators.
Based on the patterns of distributions of tadpoles and their predators, ecologists formulated a model of amphibian-community structure that emphasizes the interplay between competition and predation. Because early drying prevents many predators from living in temporary ponds, it was reasoned that competition between species would be intense and would exert a greater influence on community structure than predation in these habitats. According to this logic, the tadpoles that persist in temporary ponds are the ones that can endure the contest to secure resources and still develop fast enough to metamorphose before the pond dries. In more permanent ponds, predators would reduce tadpole densities, mediating a release from interspecific competition. Species that succeeded in permanent habitats would, according to the hypothesis, be those able to survive in the presence of the predators; competitive ability becomes less important.
Experiments in artificial ponds confirmed the predictions of the competition-predation-gradient hypothesis and suggested that good competitors were poor at surviving with predators. Investigators reasoned that tadpoles, like the invertebrates of the rocky shore, are subject to a trade-off between competitive ability and susceptibility to predators that may drive the pattern of distribution. These ideas made intuitive sense, but they had not been evaluated in natural tadpole populations.
I set out to test the competition-predation hypothesis in a more natural setting. I focused on two species that live in southeastern Michigan: the striped chorus frog (Pseudacris triseriata) and the spring peeper (P. crucifer).
These species are quite similar in most aspects of their life history but differ in their distribution among ponds. Chorus-frog tadpoles tend to be found in temporary ponds that dry each summer as well as in intermediate ponds that dry some years, but not others. Chorus-frog tadpoles are rarely, if ever, found in ponds that remain consistently filled. By contrast, spring-peeper tadpoles are found in all three types of ponds and can be common even in the permanent ponds where chorus-frog tadpoles are absent. Using the logic of the competition-predation hypothesis I would expect chorus frogs, which dwell in temporary ponds, to be superior competitors, whereas spring peepers would be expected to dominate in more permanent ponds because of their superior resistance to predation. In order to assess the potential roles of pond drying, predation and competition in the segregated distributions of Pseudacris tadpoles, I performed a fieldtransplant experiment in southeastern Michigan. I first identified several natural ponds that lie along different points of the permanence gradient. I then assessed the densities of tadpoles and their predators in those ponds. Even though I was performing these experiments in a natural body of water, I wanted to have some way to know exactly how many tadpoles I was starting with so I could determine how they fared by the end of the season. To do that, I placed several fine-mesh cages inside each of the ponds I selected. The mesh of the screen allowed the inhabitants of the cage to experience all of the environmental changes that affected the pond, but was fine enough to prevent most organisms from entering or exiting the cage.
Each cage contained just chorus-frog tadpoles, just spring-peeper tadpoles or both species together. In addition, each predator-treatment enclosure was stocked with the major predators, such as beetle larvae, salamanders and fish, present within the pond at their natural density. I added no predators to control cages. In all, I installed more than 70 cages in six different ponds-two permanent, two temporary and two intermediate. This experiment allowed me to assess the effects of drying, locally present predators and potential competitors on each tadpole species at various points along the permanence gradient.
Two months after I set up the experiment, the results emerged in the form of metamorphosing froglets. I recorded the numbers that survived in each enclosure, their size and the time it took them to reach metamorphosis. These variables are measures of ecological performance. Tadpoles that survive, grow rapidly to a large size and develop rapidly to metamorphosis are more likely to contribute to the next generation than are tadpoles that lag in any of these dimensions. Performance can be compared within a species to determine variation across the permanence gradient or to determine the influence of predators. Measures can also be compared between species to assess changes in their relative abilities to exploit their habitat.
I found that the permanence gradient most definitely affects species performance. Chorus frogs grew larger and survived better at the temporary end, and spring peepers greatly outsurvived and outgrew chorus frogs at the permanent end. By itself this result is important. It could have been that observed distributional patterns were predominantly an effect of adult breeding-site choice. In that case tadpoles from the two species might have been equally adept at living in the different types of ponds. By contrast, my result suggests that, whatever the adult preferences are, sorting takes place during the larval period. In addition, I learned that the influence of predators becomes increasingly important in more permanent ponds. Predators found in permanent ponds have much larger affects on the survival rate of tadpoles compared with the influence of predators in more temporary ponds.
One finding, however, came as a particular surprise. My research showed that interspecific competition exerted little or no effect on the outcome of the experiment. Previous studies in artificial ponds had suggested that competitive effects should be important, particularly where predation is weak. I found that each species segregated into particular ponds even in the absence of interspecific competition. Finally, I had to conclude that for Pseudacris tadpoles, at least, the interaction between competition and predation did not adequately explain the distribution patterns I observed.
A Behavioral Trade-off
My results required an alternative hypothesis. It was clear from my experiment that both pond drying and predation had strong impacts on tadpole performance. Importantly, these impacts fell unequally on the two species. I wanted to determine whether a species was constrained in its abilities to survive and grow such that good performance at one end of the gradient would mean poor performance at the other end. I knew that chorus frogs survived better in temporary ponds partly because they develop more rapidly, reaching metamorphosis up to two weeks sooner than spring peepers. Spring-peeper tadpoles, on the other hand, were not equipped to deal with the timetable of a temporary pond, and I found many of these tadpoles stranded in the bottoms of enclosures in temporary ponds at the end of the season. Chorus frogs fared relatively poorly in the presence of predators.
It appeared as if there were some sort of trade-off between rapid development, at the risk of being eaten by a predator, or slower growth to stand a chance of avoiding predation. Such a trade-off could explain both the natural distributions and the results from the transplant experiment. It would also eliminate the need to invoke competition.
Why might tadpoles be subjected to such a trade-off? As it turns out, there are good reasons to suspect that movement is positively related to feeding rate as well as to risk of predation. The behavior of a typical tadpole includes periods of inactivity interspersed with bouts of rasping and swimming. Rasping is a feeding behavior in which the tadpole uses its tail to steady itself against an object, say a decaying leaf, in order to scrape off food from the object’s surface. Since food fuels growth, any increase in the time the tadpole spends moving to feed should translate into higher rates of growth and development.
Unfortunately for the tadpole, activity may also increase the risk of predation for at least two reasons. First, many aquatic predators sit and wait for their prey. Rather than hunting for food, these predators depend on having their prey stumble into them. The more a tadpole moves around, the more likely it is to encounter a waiting predator. Second, most aquatic predators actually use movement to identify potential prey. Salamander larvae, for example, use their vision to detect moving prey. Dragonfly larvae have mechanosensory hairs that detect pressure waves from objects moving nearby, which means they can strike at a prey without having to see it. Finally, I already knew from earlier work with Earl Werner at the University of Michigan that American toad tadpoles (Bufo americanus) reduce their activity in the presence of a predator.
If this hypothesis is correct, then, the tadpole finds itself in something of a bind. Eating engenders the risk of being eaten. I evaluated in a series of lab experiments whether tadpoles are in fact subject to a behavioral trade-off. To determine the relation between activity and predation risk, I performed a simple experiment in which I exposed different groups of tadpoles to predators. I set up two kinds of situations. In one set of containers, dragonfly larvae were placed in the company of anesthetized tadpoles. The tadpoles in the second set of containers had not been anesthetized and could move around normally. The result was consistent with the hypothesis. Mobile tadpoles were consumed at over four times the rate of the anesthetized animals. In fact, the hungry predators did not even attack tadpoles in half of the anesthetized-treatment containers, underscoring the importance of movement as a cue.
I next evaluated the relation between activity and growth rate for Pseudacris tadpoles. I placed each species in containers that held either a caged predator or an empty cage, for a control. I observed the same thing for each species. In the presence of predators, tadpoles were less than half as active as control tadpoles. As predicted, the growth rates of the more stationary tadpoles dropped correspondingly. Taken together, the results from the two sets of experiments suggest that Pseudacris tadpoles are confronted with a behavioral trade-off pitting growth against risk from predators. Increased activity allows tadpoles to grow faster, but it also invites attacks from predators.
Because the two Pseudacris species tend to live in different kinds of ponds, the trade-off should weigh differently on each. In temporary ponds with few predators and not much time to complete development, it should pay to be relatively active. Rapid growth and development are crucial, and the risks of being active should be relatively small. In contrast, in permanent ponds with more predators, the risks of predation become more important, whereas the benefits of rapid growth diminish. Tadpoles in these ponds should profit from being less active.
I performed a third laboratory experiment to compare directly the behavior of chorus-frog and spring-peeper tadpoles. I gave either high or low food rations to tadpoles of each species, starting with animals that were matched for size. At both food levels, chorus-frog tadpoles tended to be more active and to grow faster than did the spring-peeper tadpoles. The dissimilar behaviors of the two species helps to explain the observed differences in performance in the field experiment as well as the natural distributional patterns of the species. Chorus-frog tadpoles seem to be specialized for life at the temporary end of the gradient, where there is a strong time constraint on development, but where predation pressures are relatively weak. By contrast, slower development and greater resistance to predation make the spring-peeper tadpoles better suited to life in more permanent ponds.
I was encouraged by the internal consistency of the results from the three sets of experiments. In addition, David Smith of Williams College and Josh Van Buskirk of University of Zurich, Switzerland, independently performed similar experiments in rock pools on the shore of Lake Superior. Results from the two experiments corroborated each other well and suggest that we are developing a good understanding of the mechanisms of distribution for these species.
At this point I was curious to determine whether the behavioral trade-off shaping the distribution of species of Pseudacris might also influence the distributional patterns of other species. I tested this hypothesis by comparing the behavior of a large number of tadpole species that live in a wide variety of habitats. Because of its high amphibian diversity, Australia was a natural choice for this research. I performed my studies in the region surrounding Wollongong, New South Wales, where there are over a dozen species within the single genus Litoria alone.
Each Litoria species I studied has a relatively narrow distribution among breeding habitats; some species are found exclusively in temporary ponds, whereas others are found in permanent ponds or in permanent streams. As is the case in North America, the diversity and size of tadpole predators increases with habitat permanence.
I brought tadpoles from seven Litoria species into the lab and assessed their activity. Of these species, three breed in temporary ponds (L. caerulea, L. chloris and L. freycineti), one breeds in permanent ponds (L. peroni), and three breed in streams and rivers, which are the most predator-rich habitats of the habitats I tested (L. citropa, L. leseueri and L. phyllochroa). Based on the Pseudacris results, I predicted that tadpoles from the more permanent, more predator-rich habitats should be less active than tadpoles from the temporary ponds. And this is what I saw. In fact, I found that tadpole species from temporary ponds were, on average, more than twice as active as tadpoles from permanent ponds and streams. As with the Pseudacris species, Litoria tadpoles exhibit behavioral patterns consistent with the trade-off.
These studies suggest that tadpoles are caught in an evolutionary balancing act driven by the covariation between pond permanence and predation, and expressed through traits, such as activity. At one time, some ecologists believed that the force of natural selection would eventually hone a species until it fit perfectly into its environment, whereupon evolution would cease. Today, the view is much different.
Biologists now know that an attribute of an individual-its phenotype-such as the activity of a tadpole, can greatly affect how well it will perform under different environmental conditions. It is also apparent that, from the perspective of this individual, the environment can vary tremendously. No single phenotype could possibly prepare an individual to cope with all of the conditions that it could face. Tadpoles living in ponds provide a good example. A particular pond can dry early in the summer one year and not at all the next. To deal with this variation, individuals can improve performance by assessing the conditions and altering a crucial trait-a phenomenon known as phenotypic plasticity. Plasticity could help tadpoles maintain their performance as the environment varies within a pond and also help a species maintain populations across a wider swath of the permanence gradient.
There is abundant evidence that tadpoles adaptively modify activity. Several species, including some Pseudacris species, are known to reduce their activity in the presence of predators. In a number of cases, the mere chemical trace of a predator is enough to limit tadpole activity, even if the predator has not recently eaten any tadpoles. Behavioral responses can also influence the life-history structure of amphibians. The activity levels, growth rates and the size at metamorphosis of at least two species are reduced in the presence of predators. Tadpoles are faced with two alternatives. They can remain in the water and continue to increase in size before metamorphosing, or metamorphose at a smaller size and move onto land. Lifehistory theory suggests that a poorquality aquatic environment makes it more advantageous for the tadpole to undergo metamorphosis instead of continuing to grow.
The array of responses exhibited by tadpoles has surprised the biologists who study them, but perhaps nothing has been as surprising as the changes in body shape and color exhibited by some species. A growing catalogue of tadpole species actually alters morphological features in response to predators. Andy McCollum of the University of Michigan and Josh Van Buskirk showed that in the presence of predators, the tail fin of the gray treefrog, Hyla chrysoscelis, becomes larger and turns brilliant red. These changes appear to reduce the risk of predation from dragonfly larvae by two means. First, by increasing the tadpole’s ability to swim quickly away from a striking predator, and second, by directing predator strikes away from the tadpole’s vulnerable head region and toward the tadpole’s flashy tail. The costs associated with different tails are still being worked out. However, preliminary evidence suggests that tail shape and color are also traits molded by trade-offs observed in animals living along the permanence gradient.
The variety and striking degree of phenotypic plasticity exhibited by tadpoles might lead one to conclude that tadpoles are able to modify themselves to perform well in any situation. However, despite often parallel responses to the same situations, decreased activity in response to a predator for example, two tadpole species usually also retain differences. Such differences in spite of plastic responses may explain why most tadpoles perform well over a fairly narrow range of the permanence gradient.
There are a few exceptions to the rule. A small number of amphibian species breed in a wide variety of habitats. These generalists, by doing what most species cannot, provide important tests of the behavioral-trade-off hypothesis. One generalist is an Australian species, a small and otherwise undistinguished animal known as the common froglet, Crinia signifera. This species is one of the most widespread and numerous frogs in southeastern Australia. Its call is so common that it can frequently be heard in the background of the soundtracks to Australian films.
The breeding habits of Crinia are remarkable. Its eggs are found in everything from horse hoofprints to sizable ponds and streams. In order to determine how Crinia maintains its unusual breeding distribution, I performed a set of experiments in conjunction with Mick Gregory of the University of Wollongong to examine characteristics of its behavior and life history. This species’s ability to coexist with predators in more permanent habitats stems in large part from its extreme inactivity. Crinia tadpoles typically spend less than 10 percent of their time being active. Most other tadpoles are several times as active. Crinia manages to survive in extremely small, ephemeral ponds by developing rapidly, completing larval development in as little as three weeks.
According to the behavioral tradeoff hypothesis, inactivity and rapid development should be mutually exclusive. Crinia circumvents the trade-off by growing very little and metamorphosing at a small size. In fact, the smallest Crinia metamorphs are among the most diminutive frogs, emerging when they are as small as 20 milligrams-a size at which two or three could sit on your smallest fingernail. In essence, Crinia has reduced its dependence on the aquatic habitat to the point that almost any water will do. For some reason, most other species metamorphose at a much larger size. Perhaps this is related to the difficulties of being an insect-sized vertebrate.
The distributional pattern of an organism is one of its most basic ecological attributes, but the mechanisms that shape distributions are far from obvious. Previous concepts of tadpole distributions had placed a heavy emphasis on the role of interspecific competition. It now appears that tadpoles can maintain their characteristic distributional patterns in the absence of that competition. This is an important insight. However, if amphibian ecologists had stopped with the standard type of field experiments there would have been little success in developing an alternative hypothesis. Evaluating the roles of behavior and other mechanisms have allowed scientists to come to a better understanding of why tadpoles are distributed as they are. In fact, experiments by Earl Werner demonstrate how behavior may contribute to competitive ability. In situations where interspecific competition is important, more active tadpoles from more temporary ponds may be superior competitors compared with less active tadpoles from more permanent ponds.
The results from research on tadpole distributions are likely to apply in other contexts as well. Many freshwater organisms are probably subject to tradeoffs comparable to those experienced by larval Pseudacris. And preliminary evidence indicates that several invertebrate taxa show a similar correspondence between behavior and distribution. These findings offer some hope of a more unified picture of ecological forces to ecologists wearied by ever more detailed studies of particular species in particular places and times. The search for mechanisms may yet yield answers in the quest to find general explanations for complex ecological phenomena. Acknowledgments
This article has benefited from discussions and collaborations with a number of colleagues, including Earl Werner, Gary Wellborn, Mark McPeek and Spencer Cortwright. Harry Ehmann and Garry Daly provided valuable advice during the Australian research. My Australian research was supported by Australian Flora and Fauna Research Program of the University of Wollongong. Comments from Lauri Freidenburg and Michelle Hoffman improved the manuscript.