Sign In. Advanced Search. Search Menu. Article Navigation. Close mobile search navigation Article Navigation. Volume Article Contents Abstract. Feeding habitat quality and behavioral trade-offs in chimpanzees: a case for species distribution models. Steffen Foerster , Steffen Foerster. Oxford Academic. Google Scholar.
Ying Zhong. Lilian Pintea. Carson M. Michael L. Deus C. The idea that movement behavior alone could explain that accumulation at edges has been proposed occasionally e. Some studies have found evidence of accumulation at edges without an associated increase in resources, but have not investigated the mechanism behind the accumulation.
For example, Olofsson et al. Because there is no obvious change in intrinsic habitat quality with distance to the fence, this observation could be consistent with animals accumulating near the fence because of their movement behavior.
However, this movement-based hypothesis has not been explicitly tested. Tenebrionid beetles have previously been used to investigate animal responses to landscape patchiness because of their small size and propensity to move in only two dimensions e. We asked three main questions. First, are individuals equally likely to be in any part of the patch, or is their distribution biased with respect to the edge?
Second, do movement behaviors, measured as step length and turn angles, differ between the edge and the center of the patch? And third, can beetle distribution in a bounded patch be explained simply in terms of step lengths and turn angles, or might additional mechanisms such as edge permeability or individual differences in behavior play a role? The use of simple models in studies of movement ecology can facilitate the generalization of findings across taxa and habitats [ 37 ], but they are rarely tested with empirical data [ 38 ].
To answer the questions we posed above, we first observed the distribution and movement behaviors of individual beetles with respect to their distance from the edge of the experimental arena. We then developed an individual-based computer simulation that mimicked the conditions of the empirical experiment.
The model drew step lengths and turn angles from the distributions that resulted from the experiment and simulated beetle movement using a correlated random walk. We predicted that if step length and turn angles determine beetle distribution with respect to the edge, we would be able to recreate the distributions observed in the empirical experiments in the simulation models. Beetles were housed in a container of dry white wheat flour at room temperature.
The tape differed in texture from the paper, and although it provided no vertical barrier, Morales and Ellner [ 35 ] have shown that T. The arena was marked into seven concentric zones of 1.
The outer zone zone 7a also contained corner zones, each of which was 2. These four corner areas were collectively known as zone 7b. Our zones allowed us to quantify time spent near and far from the edge of the arena in such a way that could be compared to a simulation model. One beetle was observed at a time, and a new experimental arena was used Figure 2 for each trial to eliminate the potential confounding effect of residual chemical cues from previous beetles [ 39 ].
Before each trial began, the test beetle was placed in the center of the arena covered by a transparent plastic vial 2. The vial encompassed all of zone 1 and extended into zone 2; therefore, for the purposes of analysis, zone 1 was incorporated into zone 2. The trial began immediately after the one-minute acclimation period and ended after the beetle had left the experimental arena or after seconds , whichever occurred first.
At the end of each trial, the beetle under observation was weighed to the nearest 0. Beetle condition index was calculated as the residual of the regression of mass on volume estimated as a cylinder [ 42 ].
To determine the distribution of time spent in each zone from the digital videos of the trials, we used an event recorder Jwatcher 1. Time spent in each zone was summed for each individual. To map the paths of beetles, we used motion analysis software Tracker 1.
Each turn angle was recorded as the angle of change in direction from the previous step. We analyzed the first and last 30 seconds of each trial but found little difference in step lengths and turn angles between the two time periods.
We therefore pooled data from both periods in our analyses. We used JMP 9. The conditions of the empirical experiment were replicated as closely as possible—beetles in the model began the trial in the center of an arena with the same dimensions and zones as the experimental arena. We modeled the movement of each individual beetle in our empirical trial as a correlated random walk [ 29 , 43 ].
Beetles in our simulation model took a step each second, and the length of the step and turn angle relative to the previous step were chosen from the distribution of observed movement behaviors of each particular beetle. Therefore, distribution parameters were unique to each beetle.
The observed distributions of step lengths and turn angles were consistent with a normal distribution when square-root and ln-transformed, respectively. As with the empirical trials, simulated trials ended either after a beetle had left the arena or time steps seconds had elapsed. We initially ran two versions of the model. The edge-independent model assumed that beetles did not alter their movement behavior based on their location within the arena.
For this model, distributions were determined for the step lengths and turn angles of each beetle over all zones. The edge-dependent model included different movement behaviors for the center zones zones 2—6 , the edge zone zone 7a , and the corner zone zone 7b according to the observed distributions of step lengths and turn angles in these zones for each beetle.
We combined center zones because we observed little difference in the movement behavior of beetles among these zones analyses not shown. Each version of the model was replicated 50 times for each of the 54 beetles.
The output of each replicate was the number of seconds the simulated beetle was in each of the 7 zones. We employed an individual-based analysis to examine the movement behavior of beetles in our empirical trials and in our models. We calculated the time spent in each zone for each beetle in the experimental study and for both models and transformed , where summed time these values meet the assumptions of parametric tests, including normality, homogeneity of variance, and independent errors.
We used simple linear regression to consider the effect of individual beetle and zone of the arena center, edge, and corner on ln-transformed step lengths and turn angles. We compared the cumulative time spent at the edge zone in the empirical study and in the models using mixed linear models with individual beetle as a random effect.
Finally, to address the influence of additional factors on beetle distribution across the experimental arena, we used multiple linear regression to determine whether beetle traits body volume, and condition index , beetle movement behavior, and the response to edge crosses per encounter explained the proportion of time an individual spent at the edge.
Statistical significance of regression analyses was determined via -tests. For all analyses, we used a significance value of as the criterion for rejecting the null hypothesis. Observed trial lengths varied from 8. The time beetles spent in each zone was not proportional to the area of each zone Figure 3 a. Beetles spent significantly more time than expected at the edge zone 7a: , and corners zone 7b: ,. Conversely, they spent less time than expected in the middle zones zone 4: , zone 5: , and zone 6: ; all:.
Time spent in the innermost zones was proportional to those areas of the arena zone 2: , zone 3: ,. Beetles took smaller steps Figure 4 a with larger turn angles Figure 4 b in the edge zone 7a than in the center zones zones 2—6. Individual differences were evident in that mean step lengths at the edge were positively correlated with mean step lengths in the center Figure 4 a ; , , but there was no consistency within individuals in the relative magnitude of turn angles at the center and edge Figure 4 b ; ,.
The results of our two simulation models, with edge-independent or edge-dependent behavior, differed from each other but neither replicated the distribution of time spent in each zone by beetles in our experimental study Figures 3 b and 3 c versus 3 a. The edge-dependent model resulted in a higher proportion of time being spent at the edge zone zone 7a than the edge-independent model ,. Conversely, less time was spent in the innermost zones zones 2, 3 and in zone 4 in the edge-dependent model than in the edge-independent model.
However, despite differences in magnitude, both models predicted less time spent near the edge in zone 7a and more time spent near the center zones 2 through 6 than was actually observed Figure 3.
Given the above results, we remodeled beetle movement using only the step lengths and turns angles that were most different between the center and the edge observed for any beetle in the empirical trials. As expected from our original hypothesis, the short steps and large turn angles at the edges, relative to the center, resulted in beetles spending most of their time near the edge Figure 3 d. However, this modification to the edge-dependent simulation model did not succeed in replicating the distribution of beetles in our empirical experiment.
To address the influence of additional factors on beetle distribution across the experimental arena, we investigated whether beetle traits, beetle movement behavior, and the response to edge crosses per encounter explained the proportion of time an individual spent at the edge.
Beetle volume and condition did not significantly affect time at the edge both and were removed from subsequent models. Some movement behaviors had small but significant effects on the proportion of time at the edge.
Beetles that took larger steps with smaller turn angles at the center tended to spend more time at the edge Table 1. However, their behavior near edges did not detectably affect the time they spent there, whether measured as the change in behavior at the edge Table 1 or when only edge step lengths and turn angles were used in models , analyses not shown.
The strongest predictor of the proportion of time spent at the edge was the proportion of encounters with an edge that resulted in the beetle crossing an edge Table 1. When the probability of crossing upon encountering an edge was low, the proportion of time spent at the edge was higher Figure 5. Furthermore, the ratio of crosses per encounter was affected neither by body volume , nor by condition index ,. The beetles in our experiment were surrogates for any organism capable of movement.
For beetles born and raised in a container of flour, the arena was likely to seem foreign and exposed. It is reasonable to assume that beetles were motivated to find a better habitat see [ 35 ] but there was no reason to think that they should persist in trying to go in any particular direction. Thus our null hypothesis was that beetles would use the whole arena equally such that the time they spend in a given zone at some distance from the edge should be proportional to the size of that zone.
The taped edge had virtually no vertical dimension and was likely not detectable from any distance. However, beetles encountering this slippery surface were unlikely to cross it to leave the experimental arena. This experimental design allowed us to consider the effect of a semipermeable boundary on the movement patterns and resulting distribution of organisms in a controlled environment on a small, manageable scale. We found that beetles spent more time at edges and corners and less time in interior zones than expected from the area of each zone Figure 3 a.
This occurred despite the fact that beetles were initially placed in the center of the arena. Campbell and Hagstrum [ 44 ] obtained similar results for the congeneric beetle T. As in our study, they found this was due in part to beetles traveling more slowly along edges than farther away from them. The underlying reasons for these distribution patterns often remain unclear because of the sheer number of variables that must be taken into consideration in natural systems.
Our system suggests that an edge alone, without any evident edge-associated changes in resources, can result in greater animal activity at edges. Our study proposed the hypothesis that disproportionate activity at edges could be explained mechanistically by changes in step lengths and turn angles.
Projected climate impacts for the amphibians of the western hemisphere. Conservation Biology 24, 38—50 Malay, M. Peripatric speciation drives diversification and distributional pattern of reef hermit crabs Decapoda: Diogenidae: Calcinus. Evolution 64, — Parmesan, C. Ecological and evolutionary responses to recent climate change. Annual Review of Ecology and Systematics 37, — Sagarin, R. The "abundant centre" distribution: to what extent is it a biogeographical rule?
Ecology Letters 5, — Savidge, J. Extinction of an island forest avifauna by an introduced snake. Ecology 68, — Aging and Its Demographic Measurement. Allee Effects. An Introduction to Population Growth.
Density and Dispersion. Introduction to Population Demographics. Population Dynamics of Mutualism. Population Ecology Introduction.
Population Limiting Factors. The Breeder's Equation. Global Atmospheric Change and Animal Populations. Semelparity and Iteroparity. Causes and Consequences of Dispersal in Plants and Animals. Disease Ecology. Survivorship Curves. The Population Dynamics of Vector-borne Diseases. Citation: Mott, C. Nature Education Knowledge 3 10 Why do species inhabit particular areas, and how are they prevented from dispersing beyond their range limits?
Numerous ecological interactions serve to maintain the geographic boundaries of species. Aa Aa Aa. Abiotic Factors. Figure 2: Range shifting. The ranges of many species have shifted northwards positive values or to higher elevations over the last few decades, and many of these response have been linked to global climate change.
Figure 3: Conservation concerns. If warming trends continue, many species, particularly tropical montane species such as this Harlequin frog Atelopus zeteki may be particularly vulnerable Lawler et al. Habitat, Individual, and Population Characteristics. Figure 4: Patterns of population density throughout the ranges of species.
The abundant center distribution of the indigo bunting Passerina cyanea ; with numerical values representing per transect densities determined from Breeding Bird Surveys. Species Interactions. Some of the most obvious and rapid cases of predators limiting the geographic distribution of their prey occur among invasive species, because introduced predators can quickly decimate endemic prey species Savidge However, the effects of predators on the ranges of their prey species are also highly dependent upon predator feeding strategies.
Among prey populations at or near their range margin, prey densities may be low according to the abundant center distribution, and, consequently, specialist predators feeding exclusively on that prey species may also exhibit reduced densities at the range margins of prey species due to the difficulty of finding enough prey to maintain predator populations. However, generalist predators, those that can consume alternate prey, may remain at higher densities and exert increased predation on prey species occurring at their range limit, potentially restricting the range of prey species to areas of reduced predation.
Figure 5: Competitive exclusion and geographic distribution. Local distribution of Plethodon richmondi shenandoah open circles and Plethodon cinereus cinereus closed circles salamanders in either talus stippled areas or deep soil white areas. Many species exhibit mutualistic relationships that are necessary for their existence; examples include plants that have a single species of pollinator. Under these conditions, the geographic range of one species is inherently linked to that of the other, and the range of either species cannot extend independent of the other.
Even under less strict mutualistic relationships, such as plants that have a variety of pollinators, the geographic range of one species is still highly dependent on that of others. Figure 6: Parasitism at range margins. The prevalence of parasitic trematodes in populations of large pond snails Lymnaea stagnalis increases dramatically when the edge of the range is approached Briers , suggesting that parasitism could contribute to range limitation among host species.
Combined Influences. Although cases may exist where a single factor limits the distribution of a species, it is undoubtedly more likely that combinations of factors act synergistically, antagonistically, or independently of one another in limiting the expansion of species beyond their current range limits. For example, North American birds living at colder, northern range limits require increased food in order to thermoregulate relative to their counterparts at southern portions of the range physiological factor , but at northern latitudes growing seasons are shorter and thus less food may be available throughout the year habitat factor.
Today, one of the most pervasive factors limiting the distribution of species is anthropogenic disturbance, which may impact the ranges of species through habitat loss, alteration, or degradation, pollution, disease, introduction of non-native species, over-harvesting, and global climate change. References and Recommended Reading Andrewartha, H. Share Cancel. Revoke Cancel. Keywords Keywords for this Article.
0コメント