top of page
  • Andy Davis

Exciting new study takes a close look at butterfly garden management and effectiveness

Greetings everyone,

A new research project was just published this month by a graduate student that I helped supervise (Ania Majewska, pictured right), and I have a feeling that this is a project that readers of this blog would love to hear about. The topic of the project is the science of butterfly gardening, and in this post, I'll be describing this new paper, and the overall project. Given the importance of this project to monarch conservation, and the immense public-interest in this topic (plus my personal knowledge of this project), this will be a lengthy post.

Let me start with a statement that will knock your socks off - believe it or not, there is very little actual hard science behind the practice of butterfly gardening! That's right, despite all of the books that have been written on butterfly gardening, the blogs, internet articles, facebook pages, etc., I can tell you that there has been surprisingly little scientific study on whether butterfly gardens are beneficial to pollinators! Most of the gardening books and articles you've probably seen have been written based on the observations of gardeners, and very little of this information comes from scientific research (on butterflies). We know this because we've actually scoured the scientific literature on this subject. I'm not saying that these books are not valid, or anything like that, because most of these are written by expert gardeners. And gardeners really know how to make a garden. However, the information in these books regarding the benefits of gardens to butterflies largely comes from observations and assumptions, not from any scientific study or data collection on this topic. So because of this void in our science, we actually still can't say with scientific certainty that butterfly gardens are beneficial for butterflies. I know what you are thinking, because we hear this a lot - but I know MY garden is beneficial, because it has butterflies in it! I know this sounds weird, but technically, that does not mean it is "beneficial". Let me come back to this idea later.

This paradox is where our butterfly garden project comes in. Given that butterfly gardening has become so popular, but yet there is so little research on it, back in 2013 Ania initiated the first-ever experimental butterfly garden research project. She teamed up with an organization called the Wormsloe Institute for Environmental History (link to their website), which at the time was headed by an amazing woman, Sarah Ross. With Sarah's help, and with funding from this group, Ania was recruited to lead this project, and over the last few years Ania has worked tirelessly to collect and analyze a boatload of data, and this new paper is the first of what will (hopefully) be many to come on this project.

Since this is the first time I've described the project, let me explain the experimental gardens first, then I'll move on to the published study. The site is a large tract of land provided to us by Mr. and Mrs. Craig and Diana Barrow. They live on the Wormsloe Historic Property in Savannah, GA, which is a beautiful, historic, plantation-style property that is open to the public. This picture below is the driveway we take to get to the gardens!

In the spring and summer of 2013 Ania established twelve gardens (12 - I know!) on this property for conducting this research. Each of them was identical in size and shape (rectangular), and they were arranged in 3 blocks of 4 gardens. Each of them also contained exactly 128 plants, and the plants were arranged in a grid fashion in the gardens. Can you see what we were going for here? Scientific rigor, and uniformity, was the name of the game. We wanted everything to be the same in each garden so that there would be "replication" in our results. From a statistical standpoint, this is very important. We didn't so much care about how they looked, because they are not really designed for aesthetics (plus, we figured the butterflies wouldn't care how they looked).

Below is a picture of one of our gardens as we were getting it planted. The garden is behind the fence posts (there are deer at the property, so all gardens had to have giant fences around them). All of the plants in the foreground are in pots, waiting to be planted. Let me discuss these plants next.

When we initially started this project, we wanted to replicate how an everyday gardener would go about creating a butterfly garden, so we purchased a lot of these plants from local nurseries. We had some of these plants grown in local greenhouses too. As for which plant species we used, we specifically chose plant types that are known to attract butterflies, and/or ones which are commonly sold in garden centers. Despite their scientific layout, we wanted these gardens to at least contain all of the plants that any typical garden would have in this region. We chose plants based on whether they would be used for nectar, and/or if they would be a hostplant for butterfly larvae. Based on the local butterflies of this region of the country, we specifically chose hostplants for 4 butterfly species - monarchs and queens (milkweed), black swallowtails (which use fennel and dill), and gulf fritillaries, which use passionvine as their hostplant. So we designed each garden to attract adult butterflies and to serve as habitat for caterpillars.

Now here is the interesting part - in half of the gardens (6 gardens) we planted only non-native plants, and the other half contained only native plants to the US. This was to test if butterflies or their larvae react differently to native vs non-native garden designs. If you look closely in the picture above, you can see that a lot of the plants in pots are non-natives - like pentas, gold yarrow, lantana, etc. Note that I'm not going to provide a full list of the plants chosen here, because I don't want readers to get sidetracked by scrutinizing our list and quibbling over which species is native to this region or not. There is a list given in the actual paper, if anyone is keen on reading it. I can tell you that for the caterpillar hostplants, we chose the non-native tropical milkweed for the monarchs and queens, and then swamp milkweed for the native gardens. That's right, we specifically planted tropical milkweed in 6 of the gardens! We are fully aware that this plant has a track record of having negative effects on monarchs, but we also know that this plant is one of the most commonly-purchased hostplants for monarchs by homeowners. So, we opted to plant it in order to study its effects. For the swallowtails, we had a native and non-native fennel, and the fritillaries had native and non-native passionvine.

Next, I'll show a picture of one of the finished gardens, taken with a drone. I think this was a native garden - you can tell because there are fewer blooming flowers. I'll come back to this point later. Each garden has a walking path along the center, and four quadrants, with each quadrant having the same number of plants (though some plants grew larger than others). The plants in this photo all look like they are the same because they are all green, but trust me, there are many different types here.

If it wasn't clear already, I should point out that getting 12 of these gardens up and running was a crazy amount of work. Normally, a homeowner just makes one garden in their backyard, and it's no small task. We created TWELVE gardens! Using over 1500 plants in total! AND, not only did we need to get the plants established and keep them alive, but we had to put up deer fencing around each garden. AND, we had to replace this two years later with a stronger fence, because the deer kept breaking through! THEN we had to install underground piping to each garden for irrigation, and then there is the weeding!

Let me make a point about weeding here, since I brought it up. In addition to the native vs non-native comparison, we also wanted to see if weeding made a difference to the butterflies, so we specifically did not weed 6 of the gardens, and we weeded and mulched the other six. I know this sounds crazy to NOT weed a garden, but we did have some logical rationale to this idea. But first, here is a picture showing what these two weeding approaches looked like as the summer wore on.

The garden on the right is obviously one that was not weeded! So why did we allow this? Well, if you stop and think about it, it actually makes sense from a caterpillar's perspective. Imagine a tiny caterpillar on one of the plants in the left garden. Now imagine a tiny caterpillar on a plant in the right garden. Which caterpillar is more visible to predators? See what I mean? When we began this project we had initially suspected that having more weeds would be beneficial to the caterpillars, so we set out to test this idea. In fact, this demonstrates one of the real strengths of this project, from an experimental standpoint. By having 12 identical gardens, we can easily create test scenarios like this, by establishing different management regimes in different gardens and then monitoring the outcome.

OK, so once we had the gardens established in 2013, Ania began collecting data on them in the summer of 2014. And here, I'm going to start describing results from this first paper that just came out. It was published online this month in the journal, Insect Conservation & Diversity, and a link to it is here. There was also a press release by the university that just came out as well - linked here. This paper was a summary of her efforts to answer one of our first questions about butterfly gardens - how does garden management affect adult butterfly or caterpillar use of gardens? She specifically was looking to see if there was any difference between the native and non-native gardens, and/or the weeded or non-weeded gardens, in terms of how well the butterflies did in the gardens. That is, how many adults were using them, and how many caterpillars there were.

That summer she conducted regular monitoring of both the adults and larvae that were using the gardens, sort of in the style of an MLMP monitoring program. She visited each garden, watched for adult butterflies (all species), and then checked each hostplant for eggs or larvae. This was a lot of work, but it paid off, because it allowed for some very nice results. I'll post just a few of the graphs here. The actual paper has a lot more numbers and charts than you'll see here.

To start with, below is a graph showing how many adult butterflies she saw on any given survey, across the season. This graph shows that by the end of the summer, at any given time there were 7-8 individual butterflies using each garden, and these were from 3-5 different species.

I mentioned that in addition to tracking the number of adult butterflies, Ania also monitored all hostplants for caterpillars on a weekly basis, specifically monarchs, queens, black swallowtails and gulf fritillaries. Below is her graph comparing how many larvae she counted in either the non-native or native gardens.

Look closely at this graph because there are some surprises. For two of the species (fritillaries and swallowtails), there were slightly more individual caterpillars counted on native plants than on non-native plants. Meanwhile, there were more monarch and queen larvae in gardens with tropical milkweed than there were in gardens with native, swamp milkweed (this is not a good thing, as I'll show later).

So the take-away from the two graphs above is that once the gardens were established, they clearly got 'used' by butterflies, both for nectaring and for producing larvae. So does this mean that the gardens are "benefiting" these butterflies? Recall I raised this point earlier. Well, technically, it doesn't. We do know that before the gardens were there, it was just a plain grass lawn (i.e. no butterfly habitat). And after we put the gardens in (surprise), we got butterflies. For most people, this is enough to conclude that the garden is beneficial - there were no butterflies before, and now there are butterflies. However, there is an equally-likely alternative scenario - it might also be the case that these gardens are doing nothing more than drawing in butterflies from the surrounding area and crowding them into one place. In other words, perhaps these butterflies would have already been present in the region, and all we did was draw them into one spot - our gardens. So technically, we can't know for sure if our gardens actually produced more butterflies than what would have been present in this region. But like I said, for most people, it is enough to know that when they put in a garden, they see butterflies using it. Period.

So what about the weeding issue? Well, Ania actually found that weeding didn't matter as much as we thought it would, to both the adult butterflies or to the larvae. For certain species, there were slightly more eggs and larvae in weeded plots, but for others it was the opposite. For adults, it was a similar story. I won't post any graphs on this here, but take my word for it. So overall, it looks like gardeners can rest easy if they have a messy garden - the butterflies don't seem to care!

OK, so now you're thinking, well then what DID affect the numbers of butterflies in a garden? Well it turns out that the gardens that got the most "use" by adult butterflies were those that had the most flowers in bloom. Ania had kept track of the number of flowering plants per garden over the summer and we found that there was a strong correlation with blooms and butterflies (hey, isn't that the title of a magazine?). This makes intuitive sense - adult butterflies need nectar, blooms have nectar, so Bob's your uncle.

About the blooming thing - remember the picture of the gardens with only native plants? Below is a graph that shows the number of plants in bloom per month in either the native or exotic gardens. I'll let others comment on the implications of this result. Note that when we initially designed the gardens we made no special effort of obtain exotic plant types that bloomed a lot.

Now, let me tell you about another part of the story that will really throw people for a loop. There was a secondary experiment that Ania conducted that was really powerful and elegant in its simplicity. The initial butterfly surveys were important for telling us how much "use" the butterflies were getting out of these gardens, but what we also wanted to know was do gardens actually produce more butterflies than you would find in a natural environment. In other words, how many butterflies are grown in these gardens and how does this compare to a natural setting? To address this question, we didn't have a natural setting to compare, but we did have the next best thing, as I'll explain below.

In July 2014, Ania conducted an experiment using monarch larvae to get at this question. She placed potted milkweed plants (swamp milkweed) inside each of the 12 gardens, and on each of these potted milkweeds she placed 2-4 young monarch larvae (1st instar). These were larvae that were obtained from our lab, so we know they were not infected with OE. The potted milkweeds also came from our lab, so we know they were good too. She then placed potted milkweed plants on the lawn outside of the gardens, and in multiple areas (between 5 and 700m sway from the gardens). She again placed 2-4 monarch larvae on these plants as well. Then she tracked the survival of these larvae over the next week. So this experiment essentially allowed her to test if it is advantageous for larvae to live in a butterfly garden, or is it better to not be in a garden at all? Remember, all of the larvae were identical in age (1st instar), and all of the plants they were on were identical. The only difference was that some larvae were living in a butterfly garden, and some were not.

The results of this experiment were extremely interesting. By the time the remaining larvae reached 5th instar, there were more survivors on plants outside gardens compared to inside gardens. Wait, what? No, that wasn't a typo. It wasn't a huge difference but it was statistically significant. Thus, this experiment demonstrated that monarch larvae had higher survival when they were on a single hostplant (placed by itself on the grass), compared to when they were in a "butterfly garden". Think on this for a second. Here we had gone to so much trouble to create these elaborate, irrigated and fenced gardens, with every plant a butterfly could want, but these same gardens actually were riskier environments for butterfly larvae to live in!

Keep in mind that while the survival was worse inside the gardens, not all the monarch larvae in the garden milkweeds died. In other words, the garden plants did "produce" at least some larvae. And furthermore, the regular surveys Ania did over the summer did show that there were lots of larvae of our other target species present in the gardens, so one could argue that despite the low survival, the gardens still produced at least some butterflies. While we didn't go into it too much in the paper, we do have an idea for why larval survival was lower in the gardens. In a word - predators. We suspect that these same gardens that we created to house all of these caterpillars actually served as magnets for caterpillar predators too. Take this photo below as an example, where an insect predator is eating one of our gulf fritillary caterpillars in our gardens. While we did not specifically track their numbers, we suspect that the gardens had a lot of predators that eat butterfly larvae in them. And not just assasin bugs and other insect predators, but birds too. I myself remember seeing bluebirds fly into our gardens, then carry off some of the larvae. I also remember seeing birds go into the gardens and not emerge for a long time (I suspect they were gorging themselves). If you think about it, by putting all of the butterflies and larvae into one place in these gardens, we effectively created an all-you-can-eat-caterpillar-buffet for the critters that eat caterpillars. And this probably goes for all butterfly gardens, especially if the plants are close together in one place.

Now I don't want people getting freaked out about this, and then going out to their gardens with a shotgun or something. I've blogged about this predator issue in the past - it's my feeling that butterfly and caterpillar predators have every right to exist on this planet (and in your garden) as do the caterpillars. In fact, I don't see this predator thing as a bad thing at all. I see this as just nature being nature, and I believe we should just let it be. Where I get annoyed is where people try to have their gardens be "only for monarchs", and because of this they try to squash every other bug in sight that dares use those milkweeds, or that dares eat (gasp) one of the monarch caterpillars. When I talk with such people, I usually try to convince them that these gardens are really like small ecosystems, and that butterfly larvae are part of the normal food chain in a healthy ecosystem. I usually don't get very far.

There was one final point to make with this story, that actually was not in the paper, and it's a doozy. Recall that our non-native gardens had tropical milkweed as hostplants for the monarchs and queens. As part of the monitoring of larvae, Ania brought some of the fifth-instar larvae in to the lab and reared these pupae to eclosion. Nearly every one of the adult monarchs that came from the non-native gardens was infected with OE.

So since this was a long post let me try to sum things up here...

- The experimental butterfly garden project at Wormsloe is a unique and ongoing effort to advance the science of butterfly gardening.

- The first scientific study from this effort, led by Ania Majewska, has documented the degree of use by butterflies in these gardens, and examined how the characteristics of gardens affect their use.

- Butterflies do use these gardens, which some would say is evidence of their benefit (though not from a scientific standpoint).

- The choice of native or non-native plants in a garden can affect some species of butterflies, but not others.

- Garden maintenance appears to have little overall effect on the use of gardens, but there are some minor effects on certain species

- Gardens with the largest number of blooming flowers appear to draw the largest number of adult butterflies.

- Monarch larvae appear to have higher risk of mortality when raised in a butterfly garden, probably because of predators, which also use the gardens

OK, I think that about does it for this blog entry.


Direct link to this blog entry:



The science of monarch butterflies

A blog about monarchs, written by a monarch scientist, for people who love monarchs

bottom of page