Integrative Biology: Issues, News, and Reviews最新文献

The conventional wisdom of most ecologists is that herbivores are generally incapable of strongly affecting plant populations in natural communities. Thus, ecologists have largely focused on the role of competition for limited resources but have ignored herbivory as a primary factor determining plant success. Here, we present a quantitative review of herbivore manipulations and find that herbivores do exert important effects on plant biomass—equally as important as those of plant competition. This result should alter the way plant communities are investigated. Furthermore, we find that the effects of invertebrate herbivores are significantly stronger than those of vertebrates; this is in contrast to widely held views. Quantitative syntheses of accumulated studies, such as the one presented here, can provide surprising answers to a broad scope of biological questions. This is especially important in fields lacking a strong theoretical basis, in which generalities are born from empiricism rather than deductive theorizing.

Spatial learning is necessary for most animals to survive in their natural environment. Spatial problems encountered by animals in nature are relatively constant across species, such as going to and from shelter or a food source. Most studies of orientation and spatial learning have been done in birds and mammals, but reptiles are particularly interesting because they have strong links between ecological factors, sensory processes, and behavior. Field studies suggest strongly that snakes can learn and remember spatial tasks encountered in the wild, including orientation, homing, and the localization of mates, shelter and foraging areas. Several sensory cues have been hypothesized to be used by snakes for orientation and navigation, with few direct tests of these hypotheses. A spatial learning and memory task has been developed that is relevant behaviorally to snakes and that they can learn rapidly. This open field escape task uses the natural behavior of snakes to address mechanistic hypotheses about their spatial learning and memory.

Sexual selection by female choice has been defined operationally as a change in male fitness due to variance in the number of mates, and studies of population biology have demonstrated this effect clearly. We argue that sexual selection is a richer phenomenon than this narrow definition implies, involving the interaction of sensory and perceptual mechanisms with signals, all of which can be influenced by an organism's hormonal and experiential milieu. The interaction between signals and receivers is further modulated by physiological processes, social context, and the physical environment. Thus, an approach to sexual selection that integrates mechanistic factors is essential for comprehensive understanding. Merging population biology with mechanistic studies, however, might not be sufficient. This is because the precise forms of the receivers and signals are the products not only of selection on current variability, but also of their evolutionary histories. We cannot imagine how researchers can hope to understand not only the “message” conveyed by a signal but the particular phenotypes that convey and receive this message without reference to history. We urge that our field advance beyond operational definitions and toward an organismal and historical understanding of the processes and mechanisms that underlie sexual selection.

Computers are increasingly being used in undergraduate biology teaching, but often their use seems more trendy than substantive. Simply putting textbooks and graphs on a computer does not accomplish much. The real value of computer-based teaching software lies with teaching the process of science — that is doing experiments and figuring out puzzles. A second major value is the ability to view large amounts of data in novel ways, for instance through 3-D visualization. Here I provide an idiosyncratic tour and review of biology teaching software, along with directions to Web sites that will take you deeper into this world. Computer software can greatly enhance biology teaching, but it is unlikely to replace flesh-and-blood professors or teaching assistants. The computer-in-the-classroom revolution has less to do with technology than with how to teach thinking.

Society invests heavily in science and research aimed at providing guidance on how to manage biological resources, yet the world is filled with too many management failures. Why is this? One reason is that the task itself is so difficult—the environment varies, we never really know the underlying processes that drive population change, and observation errors can be very large when we study populations in the wild. But worse than uncertainty itself is the fact that we tend to underestimate uncertainty. We place too much confidence in our assessment and forecasting models. Fisheries, conservation, and pest control have much to gain by embracing so-called adaptive management. Adaptive management forces us to acknowledge uncertainty, and to follow a plan by which decisions are modified as we learn by doing. Indeed, we can expect little more than continued failures if adaptive management is not adopted in a determined and widespread fashion. © 1998 Wiley-Liss, Inc.

Insect outbreaks have attracted a great deal of attention from ecologists, but an understanding of outbreaks has been elusive. We argue that a major reason for this lack of understanding is that most ecologists focus on single factor explanations, while most outbreaks are probably determined by multiple factors. This focus on single factors is not just due to investigator bias, but seems to be inherent in the major approaches used to study outbreaking insects. Theoreticians have focused on fitting mathematical models to time series of densities; we show, however, that this method is not capable of distinguishing among mechanisms. Field biologists typically rely on experiments that test only one factor at a time, probably due to the difficulty of performing experiments on an appropriate scale. We suggest that a way out of this problem may be to closely integrate models and experiments so that moderately complex mathematical hypotheses may be tested in the field without too great expense.