Archive for October, 2014

Handicaps, Honesty and Visibility
Why Are Ornaments Always Exaggerated?

By Claire Asher, on 23 October 2014

Sexual selection is a form of natural selection that favours traits that increase mating success, often at the expense of survival. It is responsible for a huge variety of characteristics and behaviours we observe in nature, and most conspicuously, sexual selection explains the elaborate ornaments such as the antlers of red deer and the tail of the male peacock. There are many theories to explain how and why these ornaments evolve; it may be a positive feedback loop of female preference and selection on males, or ornaments may signal something useful, such as the genetic quality of the male carrying them. One way or another, despite the energetic costs of growing these ornaments, and the increase risk of predation that comes with greater visibility, sexually selected ornaments must be increasing the overall fitness of individuals carrying them. They do so by ensuring the bearer gets more mates and produces more offspring.

Theory predicts that sexually selected traits should be just as likely to become larger and more ostentatious as they are to be reduced, smaller and less conspicuous. However, almost all natural examples refer to exaggerated traits. So where are all the reduced sexual traits?

Runaway Ornaments

Previous work by GEE researchers Dr Sam Tazzyman, and Professor Andrew Pomiankowski has highlighted one possible explanation for this apparent imbalance in nature – if sexually selected traits are smaller, they are harder to see. Using mathematical models, last year they showed that differences in the ‘signalling efficacy’ of reduced and exaggerated ornaments was sufficient to explain the bias we see in nature. Since the purpose of sexually selected ornaments is to signal something to females, if reduced traits tend to be worse at signalling, then it makes sense that they would rarely emerge in nature. Their model covered the case of runaway selection, whereby sexually selected traits emerge somewhat spontaneously due to an inherent preference in females. It goes like this – if, for whatever reason, females have an innate preference for a certain trait in males, then any male who randomly acquires this trait will get more mates and produce more offspring. Those offspring will include males carrying the trait and females with a preference for the trait, and over time this creates a feedback loop that can produce extremely exaggerated traits. Under this model of sexual selection, differences in the signalling efficacy can be sufficient to explain why we so rarely see reduced traits.


However, this is just one model for how sexually selected ornaments can emerge, so this year GEE Researchers Dr Tazzyman and Prof Pomiankowski, along with Professor Yoh Iwasa from Kyushu University, Japan, have expanded their research to include another possible explanation – the handicap hypothesis. According to the handicap principle, far from being paradoxical, sexually selected ornaments may be favoured exactly because they are harmful to the individual who carries them. In this way, only the very best quality males can cope with the costs of carrying huge antlers or brightly coloured feathers, and so the ornament acts as a signal to females indicating which males carry the best genes. This is an example of honest signalling – the ornament and the condition or quality of the carrier are inextricably linked, and there is no room for poor quality males to cheat the system.

Using mathematical models, the authors investigated four possible causes for the absence of reduced sexual ornaments in the animal kingdom. Firstly, like the case of runaway selection, differences in signalling efficacy might explain the bias. Under the handicap hypothesis, ornaments act as signals of genetic quality, so it would be little surprise that their visibility or effectiveness at conveying the signal would be important. Smaller ornaments may simply be worse at attracting the attention of females, meaning that the benefits of the sexual ornament don’t outweigh the costs. Similarly, the costs for females of preferring males with reduced ornaments may be higher than for exaggerated ornaments, because it is easier to find males with exaggerated traits. Again, this could theoretically tip the balance of cost and effect away from selecting for reduced ornaments. An alternative explanation is that the costs of the ornament itself are different for reduced and exaggerated traits. This seems quite likely in many cases, since a large ornament would require more resources to grow. But in this case selection would be more likely to produce reduced ornaments with lower costs! In order to replicate the excess of exaggerated traits we see in nature, reduced traits would have to cost more – much less biologically plausible. However, if large ornaments tend to be more costly, then they may be more likely to be condition-dependent, a key tenant of the handicap principle. Exaggeration may be more effective at producing honest signalling and exaggerated traits may therefore be more useful to females as a signal of quality.

Honest Signals

The results of modelling highlighted two key ways in which exaggerated traits might be favoured by the handicap process. In the model, when exaggerated traits have a higher signalling efficacy or are more strongly condition-dependent, exaggerated traits tend to be more extreme than reduced traits. The model still predicts that reduced traits are equally likely to evolve, just that they will tend to be less extreme examples of ornamentation. The other two possible explanations – higher costs to females that prefer small ornaments or the males that carry them, failed to consistently produce the observed lack of reduced ornaments. Both explanations were able to produce this outcome under certain circumstances, but in other circumstances they produced the opposite effect. Exaggerated ornaments, therefore, may be more common because they are more effective signals that are more likely to be honest.

Based on this and previous work by Dr Tazzyman and colleagues, asymmetries in the signalling efficacy of reduced and exaggerated traits is sufficient to explain the lack of reduced traits in nature. Whether ornaments evolve via runaway selection or the handicap process, asymmetrical signalling efficacy tends to favour exaggerated traits. However, in the case of the handicap process, asymmetries in condition-dependance of the trait may also be involved. These two explanations are not mutually exclusive, and it is likely that in reality many factors are involved.

Importantly, for both explanations and for both type of sexual selection, the models still predict that exaggeration and reduction will be equally likely. The differences emerge in terms of how extreme the ornament will become. Thus, this work predicts that there are many examples of reduced ornaments in nature, perhaps we just haven’t found them yet. This is especially likely if reduced traits that might be less noticeable anyway also tend to be less extreme. Alternatively, there may be other asymmetries not yet considered that make reduced ornaments less likely to emerge in the first place.

Where Next?

The authors suggest some very interesting avenues for future research. Firstly, they suggest that more complex models investigating how multiple different asymmetries may act together to produce sexually selected ornaments will get us closer to understanding the intricate dynamics of sexual ornamentation. Secondly, these models have only considered cases where evolution of the trait eventually settles down – at a certain ornament size, the costs and benefits of possessing it are equal, and the trait should remain at this size. However, in some cases the dynamics are more complex, and traits undergo cycles of exaggeration and reduction. Research into the impact of asymmetries in condition-dependence and signalling efficacy in these ‘nonequilibrium’ models would yield fascinating insights into the evolution of sexual ornaments.

Original Article:

() Evolution


This research was made possible by funding from the Natural Environment Research Council (NERC), the Engineering and Physical Sciences Research Council (EPSRC), the European Research Council (ERC)

PREDICTS Project: Land-Use Change Doesn’t Impact All Biodiversity Equally

By Claire Asher, on 13 October 2014

Humans are destroying, degrading and depleting our tropical forests at an alarming rate. Every minute, an area of Amazonian rainforest equivalent to 50 football pitches is cleared of its trees, vegetation and wildlife. Across the globe, tropical and sub-tropical forests are being cut down to make way for expanding towns and cities, for agricultural land and pasture and to obtain precious fossil fuels. Even where forests remain standing, hunting and poaching are stripping them of their fauna, degrading the forest in the process. Habitat loss and degradation are the greatest threats to the World’s biodiversity. New research from the PREDICTS project investigates the patterns of species’ responses to changing land-use in tropical and sub-tropical forests worldwide. In the most comprehensive analysis of the responses of individual species to anthropogenic pressures to date, the PREDICTS team reveal strong effects of human disturbance on the geographical distribution and abundance of species. Although some species thrive in human-altered habitats, species that rely on a specific habitat or diet, and that tend to have small geographical ranges, are particularly vulnerable to habitat disturbance. Understanding the intricacies of how different species respond to different types of human land-use is crucial if we are to implement conservation policies and initiatives that will enable us to live more harmoniously with wildlife.

Red Panda (Ailurus fulgens)

Habitat loss and degradation causes immediate species losses, but also alters the structure of ecological communities, potentially destabilising ecosystems and causing further knock-on extinctions down the line. As ecosystems start to fall apart, the valuable ecosystem services we rely on may also dry up. There is now ample evidence that altering habitats, particularly degrading primary rainforest, has disastrous consequences for many species, however not all species respond equally to land-use change. The functional traits of species, such as body size, generation time, mobility, diet and habitat specificity can have a profound impact on how well a species copes with human activities. The traits that make species particularly vulnerable to human encroachment (slow reproduction, large body size, small geographical range, highly specific dietary and habitat requirements) are not evenly distributed geographically. Species possessing these traits are more common in tropical and sub-tropical forests, areas that are under the greatest threat from human habitat destruction and loss of vegetation over the coming decades. The challenge in recent years, therefore, has been developing statistical models that allow us to investigate this relationship more precisely, and collecting sufficient data to test hypotheses.

There are three key ways we might chose to investigate how species respond to land-use change. Many studies have investigated species-area relationships, which model the occurrence or abundance of species in relation to the size of available habitat. These studies have revealed important insights into the damage caused by habitat fragmentation, however they rarely consider how different species respond differently. Another common approach uses species distribution models to predict the loss of species in relation to habitat and climate suitability. These models can be extremely powerful, but require large and detailed datasets that are not available for many species, particularly understudied creatures such as invertebrates. The PREDICTS team therefore opted for a third option to investigate human impacts on species. The PREDICTS project has collated data from over 500 studies investigating the response of individual species to land-use change, and their database now includes over 2 million records for 34,000 species. Using this extensive dataset, the authors were able to model the relationship between land-use type and both the occurrence and abundance of species. One of the huge benefits of this approach is that their dataset enabled them to investigate these relationships in a wide range of different taxa, including birds, mammals, reptiles, amphibians and the often neglected invertebrates.

Modelling Biodiversity

Sunbear (Helarctos malayanus)
image used with permission from
Claire Asher (Curiosity Photographic)

The resulting model included the responses of nearly 4000 different species across four measures of human disturbance; lang-use type, forest cover, vegetation loss and human population density. The vast dataset, the PREDICTS team were able to compare the responses of species in different groups (birds, mammals, reptiles and amphibians, invertebrates, between habitat specialists and generalists and between wide- and narrow-raging species. Their results revealed a complex interaction between these factors, which influenced the occurrence and the abundance of species in different ways.

In general human-dominated habitats, such as urban and cropland environments, tended to harbour fewer species than more natural, pristine habitats. Community abundance in disturbed habitats was between 8% and 62% of the abundance found in primary forest, and urban environments were consistently the worst for overall species richness. In these environments, human population density and a lack of forest cover were key factors reducing the number of species. Human population density could impact species directly through hunting, or more indirectly through expanding infrastructure. However, these factors impact different species in different ways, so the authors next investigated different taxa separately.

Birds appear to be particularly poor at living in urban environments, most likely because they respond poorly to increases in human population. Forest specialists and narrow-ranged birds fare especially badly in urban environments; only 10% of forest specialists found in primary forest are able to survive in urban environments. Although the effect was less extreme, mammals were also less likely to occur in secondary forest and forest plantations than primary forest, and forest specialists were particularly badly affected.

Urban Pests
Although many species were unable to exist in disturbed habitats, those species that persisted were often more abundant in human-modified habitats than pristine environments. This isn’t particularly surprising – some species happen to possess characteristics that make them well suited to urban and disturbed landscapes; these are often the species that we eventually start to consider a pest because they are so successful at living alongside us (think pigeons, rats, foxes). These species tend to be wide-ranging generalists, although sometimes habitat specialists do well in human-altered habitats if we happen to alter the habitat in just the right. Pigeons, for example, are adapted to nesting on cliffs, which our skyscrapers and buildings inadvertently mimic extremely well. The apparent success of some species in more open habitats such as cropland and urban environments might also be partly caused by increased visibility – it’s far easier to see a bird or reptile in an urban environment than dense primary forest! This doesn’t explain the entire pattern, however, and clearly some species are simply more successful in human-altered habitats. They are in the minority, though.

Do Reptiles Prefer Altered Habitats?
One interesting finding was that for herptiles (reptiles and amphibians), more species were found in habitats with a higher human population density. This rather unexpected relationship might reflect a general preference in herptiles for more open habitats. Consistent with this, the authors found fewer species in pristine forest than secondary forest. However, upon closer inspection the authors found that herptiles do not all respond in the same way. Reptiles showed a U-shaped relationship with human population density – the occurrence of species was highest when there were either lots of people or no people at all. By contrast, amphibians showed a straight relationship, with increases in human population density being mirrored by increases in the number of species present. This highlights the importance of investigating fine-scale differences between species in their responses to human activities.

Filbert Weevil (Curculio occidentis)

Consistent with previous studies, the traits of species were very important in determining whether a species was found in human-altered habitats. Narrow-ranging species were much less likely to occur in any habitat than wide-ranging ones, but this difference was particularly clear for croplands, plantation forests and urban habitats. The extent of human impact was also a key factor determining the occurrence of species in different habitats. Forest cover, human population density and NDVI (a measure of vegetation loss taken from remote sensing) all reduced the number of species present. Measures of disturbance and species characteristics do not act in isolation – the best models produced by the PREDICTS team included interactions between these variables. Invertebrate numbers were lowest in areas of high human population density and high rates of vegetation loss.

This study is the first step in more detailed, comprehensive analyses of the responses of species to human activities. The power of this study comes not only from it’s large dataset and broad spectrum of taxonomic groups, but also from it’s ability to directly couple land-use changes with species’ traits such as range-size and habitat specialism. The authors say that the next major step would be to incorporate interactions between species in these models – the community structure of an ecosystem can have profound effects on the species living in it, and changes in the abundance of any individual species does not happen in isolation from the rest of the community.

Check out the PREDICTS Project for more information!

Original Article:

() Proc. R. Soc. B


This research was made possible by funding from the Natural Environment Research Council (NERC), and the Biotechnology and Biological Sciences Research Council (BBSRC).

Calculated Risks:
Foraging and Predator Avoidance in Rodents

By Claire Asher, on 3 October 2014

Finding food is one of the most important tasks for any animal – most animal activity is focused on this job. But finding food usually involves some risks – leaving the safety of your burrow or nest to go out into a dangerous world full of predators, disease and natural hazards. Animals should therefore be expected to minimise these risks as much as possible – foraging at safer times of day, especially when there’s lots of food around anyway. This hypothesis is known as the “risk allocation hypothesis”, but it has rarely been tested in wild animals. Recent research from ZSL academic Dr Marcus Rowcliffe showed that the behaviour of the Central America agouti certainly seems to follow this pattern, and highlights the amazing plasticity of animal behaviour.

Central American Agouti
(Dasyprocta punctata)

Foraging, although essential, is always a compromise between finding food and avoiding being eaten by a predator. The aim of the game is to eat as much as you can whilst avoiding being eaten yourself, in order to live long enough and grow large enough to reproduce. Since finding food is one of the most important things an animal has to do, foraging behaviour has been subject to strong natural selection.

The risk allocation hypothesis predicts that prey species should focus their foraging effort at times of day that pose the least risk. So, if your main predator is active during the day, you best forage at night and vice versa. There ought to be some flexibility in this system too, though – if food in your habitat is plentiful, it should be easy to find enough to fill you and there is little need to take any additional risks. Conversely, if food is pretty scarce, you may be forced to take more risks than usual by foraging for longer or at more dangerous times of day.

In a recent study, academics from the Institute of Zoology, London, in collaboration with colleagues around the world, investigated this trade off between food and predator avoidance in the Central American Agouti (Dasyprocta punctata). The agouti’s biggest problem in life is the Ocelot (Leopardus pardalis), who primarily feed on agoutis. Using radio telemetry and camera trapping, the researchers investigated activity patterns of agouti living in areas with lots of Astrocaryum fruits, and those living in areas with less. They were able to generate an enormous dataset – over 30,000 camera trap records of agoutis, with a further 50 individuals radio collared and tracked!

Ocelots are highly nocturnal, and across nearly 500 camera trap observations, Ocelots were almost exclusively observed at night. During this time, agoutis were under a great deal of risk – the predation risk from Ocelots was estimated to be four orders of magnitude higher between dusk and dawn than during daylight hours. The foraging activity of agouties mirrored this – activity was highest during the day, with peaks first thing in the morning and again later in the afternoon. Most interestingly, these patterns differed for agoutis that lived in habitats with abundant fruit and habitats where fruit was sparse. When food availability was high, agoutis took fewer risks, leaving their burrows later in the morning and coming home again earlier at the end of the day. Overall their activity levels were lower, presumably because they didn’t need to forage for as long to find all the food they needed.

The results of this study support the risk allocation hypothesis, and show that animals are able to make complex calculations about risks and benefits based upon environmental conditions and alter their behaviour so as to minimise risks and maximise benefits. Only when food availability is high can agoutis afford to have a lie-in and avoid any ocelots returning home late.

Original Article:

() Animal Behaviour


This research was made possible by funding from the National Science Foundation (NSF) and the Netherlands Foundation for Scientific Research