Trophic cascades in fragmented forests

Many birds eat insects and spiders. Some of these insects and spider are themselves predators, feeding on critters lower down in the food web. Some of the insects that are fed upon by birds, or other predators, also play important roles in forest, such as munching upon the fresh, green leaves of young trees (here’s a reminder).

Munch, munch, munch. The hungry caterpillar. (photo by Sean McCann, reproduced here with permission)

Munch, munch, munch. The hungry caterpillar. (photo by Sean McCann, reproduced here with permission)

These interactions are ongoing, all the time, in forests around the world. These forests, however, are changing in important ways. Some of them are getting smaller and smaller as humans continue to encroach on the land, via urbanization or agriculture. This results in a ‘fragmented’ landscape. A landscape with small forest patches, perhaps no bigger than your back yard. A landscape with larger forests, perhaps one in which you could get lost in. These forests are themselves connected to each other –sometimes directly by a corridor or hedgerow.

This is the context for PhD student Dorothy Maguire’s research. Within that context, Dorothy tackled a fascinating project, one that was just recently published. In this work, Dorothy and co-authors (including me, an undergrad at that time, Thomas Nicole, and McGill Professor Elena Bennett) put cages around small trees in different types of forests SW of Montreal. The cages (made of chicken wire) were in place to test the effects of ‘predator exclusions’ on the insects and spiders occurring on saplings. The prediction is that if you exclude larger predators, such as birds, this may allow a ‘release’ of other insects and spiders. In turn, this release may have trickle-down effects on an important process occurring in young trees: herbivory. For example, if a predator is more common because it’s not being eaten by birds, perhaps it will eat more caterpillars, which may mean the leaves on trees will be eaten less frequently. In ecology this is dubbed a ‘trophic cascade’. Dorothy did this work in the context of fragmented forests, and she worked in forests that were either small and isolated from other forests, or in forests that were large and connected to other forests. This was done because there’s an expectation that these ecological effects will be different depending on the degree of fragmentation happening on the landscape. For example, insectivorous birds may decrease in abundance in small, isolated patches, which means their effects on insect prey (and perhaps herbivory) may be reduced relative to effects in larger patches of forest.

Dorothy Maguire, working in a forest fragment.

Dorothy Maguire, working in a forest fragment.

During one summer field season, Dorothy and Thomas wrapped up some small sugar maple trees in chicken wire, left some alone as controls, counted insects and spiders over the summer months, and measured herbivory on the trees themselves. As expected, the effects of the ‘cage’ was significant: when you put a cage around a tree, you end up with more arthropods living on those trees. This confirms other papers which report a similar effect: insectivorous birds (and perhaps other vertebrate predators) have a significant, and meaningful impact on the insects and spiders living on trees. Or, stated another way, birds eat critters living on trees, and without these birds, there would certainly be more arthropods around!

Dorothy did not uncover a strong effect on the process of insect herbivory: although more insects and spiders were living in the trees protected by chicken wire, the leaves themselves were not affected. This could be because more insect predators were around, and thus compensating for the lack of birds, and eating just as many herbivorious insects (e.g., caterpillars) as the birds might have eaten.

The lanscape of southern Quebec. Lots of agriculture, some patches of forest.

The lanscape of southern Quebec. Lots of agriculture, some patches of forest.

Scaling up to the landscape context, there were no overall significant effects of the cage treatments in relation to the forest type, nor was the level of herbivory dependent on the landscape context. The general results for large, connected patches were no different than for small, isolated patches. However, the magnitude of the effect was marginally affected by the landscape context for the cage exclusion: vertebrate predator may have a more significant impact in smaller, isolated patches.

As with all research projects, this work resulted with as many questions as answers, which is equally frustrating and fascinating. It’s clear that vertebrate predators are important in these systems, but more work is needed to fully assess whether these effects are truly affected by the degree of forest fragmentation on the landscape. The lack of effects on the process of herbivory itself was equally intriguing – there are clearly many complex interactions occurring on small maple trees. Some of these interactions involve top-down predation events, but there are likely a suite of ‘bottom-up’ effects that are also influencing the system.


MAGUIRE, D. Y., NICOLE, T., BUDDLE, C. M. and BENNETT, E. M. (2014), Effect of fragmentation on predation pressure of insect herbivores in a north temperate deciduous forest ecosystem. Ecological Entomology. doi: 10.1111/een.12166

The effect of insecticides on jumping spider personalities

This post was written by C. Buddle and R. Royaute (a PhD student in the Arthropod Ecology lab).

We are pleased to announce a recent publication from our lab, titled Interpopulation variations in behavioral syndromes of a jumping spider from insecticide-treated and insecticide-free Orchards.  As is traditional in the lab, here’s a plain language summary of the work:

Agriculture has strongly intensified in the last 60 years, causing major concerns the sustainability of biodiversity. Agricultural practices can reduce habitats available for wildlife and also release toxins in the environment through the use of pesticides. Not all organisms living in agricultural fields are harmful, and many predators, including spiders, can help to reduce pest density. We have a relatively good knowledge that the diversity of spider species in agriculture, especially under our temperate latitudes, can help reduce pest damage. However, many of the factors that influence spider predation on pests depend on the outcome of behavioural interactions and we don’t know much about that topic. Spiders are often cannibalistic and aggressive with one another and these types of behaviours may limit their efficiency for pest control. We also need to understand if these aggressive tendencies vary depending on the type of agricultural field considered, a pesticide treated field may favour very different behaviours than one that is managed organically. Another important point is that populations are composed by a multitude of individuals, each with its own behavioural tendencies. Some individuals take more risks when confronted with predators (i.e. they are more bold), others are more active and explore larger areas or consume more prey. These tendencies – often referred to as personality traits – may also be correlated with one another.

In the context of agriculture, this may mean that certain individual spiders may contribute more to biocontrol because they consume more prey, or that certain individuals are more at risk of being in contact with pesticides because they are more active. To understand, how agricultural practices, and particularly insecticidal applications, affects personality and behavioural syndromes in spiders, we focused on the jumping spider Eris militaris, an abundant and charming jumping spider occurring in apple orchards in Quebec. Here’s a lovely photo from Crystal Ernst to illustrate how attractive they are: (thanks, Crystal, for permission to post the photo here!)

Screen Shot 2013-11-26 at 3.34.45 PM

We collected spiders from pesticide-treated and pesticide-free orchards, brought them back to the laboratory, and did a number of behavioural tests on the individuals from the two populations. Compared to the insecticide-free populations, we document that individuals from orchards that did receive insecticides experienced a shift in their behaviours syndromes. The overall shape of this syndrome is multidimensional, but it suffices to say that the correlations among different behaviours (the ‘syndromes’, otherwise known as the ‘personality’) differed depending on where the population came from.

A 'mirror test' - used to study behaviour in E. militaris (photo by R. Royaute)

A ‘mirror test’ – used to study behaviour in E. militaris (photo by R. Royaute)

In sum, the personality shifts that we documented for E. militaris are potentially quite important since the relationships between different behaviours may affect a spider’s ability to be an effective generalist predator in apple orchards. We need to consider how management  (including use of insecticides) may affect specific behaviours, and more importantly, the relationships between the different behaviours.


Royaute, R., C.M. Buddle & C. Vincent. 2013.  Interpopulation Variations in Behavioral Syndromes of a Jumping Spider from Insecticide-Treated and Insecticide-Free Orchards. Ethology. doi: 10.1111/eth.12185

Lunch in the tree-tops for the birds and the bugs

A few weeks ago, our laboratory published a paper in PeerJ (an open-access journal) titled “Vertical heterogeneity in predation pressure in a temperate forest canopy“. This work resulted from a project by former Master’s student Kathleen Aikens. She graduated a little while ago, and although we published one of her thesis chapters in 2012, it took another year to get this paper out, in part because Kathleen and I both become too busy.  Thankfully, post-doc Dr. Laura Timms agreed to help us finish up the paper, and she worked with me and Kathleen to re-analyze the data, re-write some sections, and whip it into shape.

As is now traditional for my laboratory, here’s a plain-language summary of the paper:

Tree canopies, including those in deciduous forests in southern Quebec, are important for many different animals, including insects and spiders. These small, marvelous creatures crawl up and down trees with regularity, feed upon the leaves of trees, feed upon each other, and are food for animals such as birds and bats. Past research has shown that many species of insects and spiders live in tree canopies, and in general, more insects and spiders are found closer to the ground compared to the very tops of the trees. This makes sense, since deciduous tree canopies often need to be recolonized each spring, and tree canopies are relatively harsh environments – they are windy, hot, and often-dry places as compared to the forest floor.  What we don’t know, however, is whether the insects and spiders avoid the tree canopies because they may be eaten more frequently in the canopy as compared to the understory. The objective of this research was to test this question directly, and find out whether insects and spiders are arranging themselves, vertically, because predators may be preferentially feeding on them along this vertical gradient. This is a very important area of study since biodiversity is highly valued and important in forests, but we cannot fully appreciate the status of this diversity without discovering what controls it.


Our mobile aerial lift platform. TO THE CANOPY!

We did this work by using two experiments that involved manipulating different factors so we could get at our question in the most direct way possible. In the first experiment, we made ‘cages’ out of chicken wire and enclosed branches of sugar maple trees in the cages. We did this at the ground level all the way to the tops of trees, using a ‘mobile aerial lift platform’. These cages acted to keep out large predators, such as birds, but allowed insects and spiders to live normally on the vegetation. We counted, identified, and tracked the insects and spiders both within these cages, and in adjacent branches that did not have cages (the ‘control’). By comparing the control to the cage, we could find out whether feeding activity by larger vertebrate predators affected insects and spiders, and whether this differed when comparing the ground to the top of the trees. In the second experiment, we used small pins and attached live mealy worms (larvae of beetles) to the trunks of trees, and we did this in the understory all the way up to the canopy. We watched what happened to these mealy worms, and compared what happened during the day and overnight. This is called a ‘bait trial’, and let us figure out what sort of predators are out there in the environment, and in our case, whether they fed more often in the canopy compared to the ground-level. This second experiment was designed for seeing the effects of insect and spider predators along a vertical gradient whereas the first experiment was focused more on vertebrate predators (e.g., birds).


Munch munch. Carpenter ants feeding on mealworms.

Our results from the first experiment showed that the cages had an effect: more insects and spiders were found when they were protected from predation by birds. Birds are playing a big role in forest canopies: they are feeding on insects and spiders, and in the absence of vertebrate predators, you might speculate more insects and spiders would occupy trees. Our second experiment showed that ants were important predators along the tree trunks, and overall, the most invertebrate predators were found in the lower canopy. Both experiments, together, confirmed that the understory contained the most insects and spiders, and was also the place with the highest amount of predation pressure.  The take-home message is that there is an effect of predation on insects and spiders in deciduous forests, and this effect changes if you are in the understory as compared to the top of the canopy. We also learned and confirmed that insects and spiders remain a key element of a ‘whole tree’ food web that includes vertebrates such as birds, and that predators in trees tend to feed on insects and spiders along a gradient. Where there is more food, there is more predation pressure! Our work was unique and novel because this is the first time a study of predation pressure was done along a vertical gradient in deciduous forests. It will help better guide our understanding of forest biodiversity, and the processes that govern this diversity.

A more detailed discussion of this work is posted on the PeerJ blog.