A longer version of this article was originally published in Conservation Sense and Nonsense in March of this year and brought to our attention by CSN editor, Mary McAllister. I asked the author, garden ecologist Nathan Lambstrom, if he would write a condensed version for GardenRant, as I felt it would be of great interest to our ecologically minded readers who witness the interaction of non-native plants with insects in their own gardens, and wish to dive deeper into available research. Here Lambstrom goes back to the data analysis of an oft-cited study with a scientist’s curiosity; and finds that interactions between plants and pollinators are far more nuanced than first reported. It’s worth your undivided attention. -MW
Whether working in a home garden or creating a restoration plan, deciding what to plant is a challenging process. In addition to ensuring that the chosen species will thrive in the given environment, the current crisis facing pollinating insects means that many gardeners also wish to include plants that will provide a high level of support to those insects, particularly native bees and lepidopterans (butterflies and moths) in terms of food and habitat. The importance of native plants in providing that support is well-understood, but the contributions of introduced (non-native, naturalized) plants in supporting pollinating insects is also significant, and we should not discount it.
Primary research has given us better data to help inform our decisions, and also a better understanding of exactly how restricted some host preferences are. But how we read that data matters, and it may be more nuanced than we think.
Which plants can feed the most insects?
The most well-known papers on the subject of lepidopteran-supporting plant species come from entomologist Dr. Doug Tallamy, who has popularized the idea of pollinator gardening with native plants to a wide audience.
One of his earlier papers demonstrating the ecological value of native versus introduced plants is a meta-analysis that was published in the journal of Conservation Biology in 2009 titled “Ranking Lepidopteran Use of Native Versus Introduced Plants.”
After analyzing thousands of available records of preferred food sources of lepidopteran larvae, and ranking them by genus, Tallamy and a colleague concluded that native plants in eastern North America, and particularly native woody plants, support more native lepidopteran species on average than introduced plants or herbaceous plants generally.
I use the associated dataset frequently to guide plant selection decisions and encourage others to do so as well. It shows that many of our beloved woody species can support an amazing diversity of native insects: some genera like our oaks (Quercus), birches (Betula), and maples (Acer) support larvae of hundreds of species of butterflies and moths (Fig. 1).
Fig. 1 – top-ranking woody plants and the numbers of lepidopteran species that can use them as host plants.
A closer look at the data
However, Tallamy’s analysis of the data does not tell the entire story and may be obscuring the value of non-native plants. Here’s how:
In the dataset, plant genera are categorized by origin. If a genus has only native species in North America, it is labeled “native;” if it has only introduced species, it is labeled “alien.” Genera that contain both native and introduced species (oaks, maples, birches, willows, and hundreds of others) are categorized as “both.”
So far so good. Many plant genera in our region include both native and introduced species, and many species that we consider invasive have very close native relatives.
However, for statistical analysis in the Tallamy paper, all genera with both native and introduced species are re-classified as “native” only (see the “origin for analysis” column in Fig. 1).
The reasoning behind this is not clearly explained in the paper, but it has significant implications for our interpretation of the findings.
Since the data were originally collected only at the genus level (not the species level), it is impossible from this dataset to determine whether it is a native or introduced species within a genus supporting lepidopteran species, or whether they both are.
What does this mean in practice?
Consider the barberries (genus Berberis), one of the most common plants labelled invasive on the east coast. This genus includes two introduced species which are quite common on forest edges in disturbed environments (B. thunbergii, B. vulgaris) and one rare native species that is restricted to southwestern Virginia (B. canadensis) (Fig. 2).
Fig. 2 – county-level distributions of our three barberry species in eastern North America (adapted from BONAP)
When native lepidopteran larvae feed on non-native, naturalized barberry species (and the larvae of 11 species of butterflies and moths are known to do so), these observations are counted in the Tallamy paper as “native” because the genus also contains a native species. Thus, the contribution of the introduced barberry to lepidopterans becomes invisible in the analysis (Fig. 3).
Fig. 3 – many of our most abundant “invasive” plants support dozens of native species of lepidopteran larvae
Another example, clovers (genus Trifolium), are known to support 115 native lepidopteran species. There are over a dozen non-native, naturalized clovers throughout the region (quite common in lawns and post-agricultural environments) and two uncommon native ones restricted to the southeast. Categorizing clovers as ‘native’ similarly erases the ecological contributions of introduced species to lepidopterans.
For the purposes of the Tallamy paper, any time an insect uses an introduced plant from a genus that also contains native species, that positive interaction is credited solely to native plants.
A different story
If we more accurately count genera that contain both native and introduced species as ‘both’ instead of counting them as solely ‘native’, a more nuanced picture emerges. (Fig. 4). Re-analyzing the data with those contributions of non-native plants added back in shows us the following:
- Native woody plants still perform better on average, supporting 64 lepidopteran species (61 native) compared to mixed woody plant genera supporting 49 species (46 native). The advantage exists, but it’s much smaller than the difference reported in the original analysis.
- For herbaceous plants, the native advantage disappears entirely. In fact, genera that include introduced herbaceous species support more lepidopteran species on average: 6 species (5 native) for these genera versus 5 species (4 native) for solely native genera.
Fig. 4 – average number of lepidopteran species supported by origin and plant type
The widespread belief that native plants are always dramatically superior to introduced plants is not a reflection of ecological reality. By treating mixed origin genera as entirely native, the ecological value of thousands of introduced plant species is misattributed to native plants, concealing the introduced species’ actual contributions to pollinator support.
Land managers and gardeners using this research to guide their decisions may be removing introduced plants that are, in reality, providing significant support to native insects. When we aim to eradicate naturalized plants based on the assumption that only native plants matter, we may be eliminating valuable resources that insects have already incorporated into their life cycles.
Introduced plants are active participants in ecosystems
Ignoring the contributions that introduced plants make towards supporting imperiled pollinators not only skews our perception of these plants, leading to the commonly held assumption that native plants are the only plants that support pollinators, it causes us to potentially ignore and possibly interfere with the positive contributions that many of these plants, even those labelled invasive, can make.
Primary research has shown us many times that introduced plants, whether in a garden or naturalized in a landscape, can provide food in the form of nectar, pollen, and larval host plants to many of our native bees, wasps, flies, beetles, butterflies, and moths (Sax et al. 2022).
Not to mention the ecological value they can provide in terms of habitat, erosion control, carbon sequestration, bioremediation, etc. For instance, the introduced genus Pyrus (pear), which has no native species in the region, supports over 100 native lepidopterans—more than some native genera.
Insects are often not confined to a single native plant species
Some good news that we often don’t hear is that most of the lepidopterans (butterflies and moths) that need a specific group of plants to host their larvae are limited not to a single species, but to a single genus, a few genera, or an entire plant family.
A good example of this is the black swallowtail butterfly (Papilio polyxenes) native to much of eastern North America. The larvae of the black swallowtail feed almost exclusively on plants in family Apiaceae, the dill family.
Today, the most commonly encountered Apiaceae in most parts of the black swallowtail’s range are non-native garden herbs or naturalized plants (e.g. dill, parsley, fennel, and Queen Anne’s lace). Black swallowtail larvae are able to recognize these plants as food because they are chemically similar to the native plants within Apiaceae that were their historic food source.
Those introduced (non-native, naturalized) plants have become so common they are now the primary host plants for black swallowtail larvae. In fact, in Massachusetts there have been no confirmed sightings of black swallowtail larvae feeding on native species in the Apiaceae family since 2007 (Stitcher 2013).
Information like this is important for gardeners and land managers to keep in mind when making decisions about what plants to keep or remove. Wholesale eradication of naturalized plants like fennel or Queen Anne’s lace could, counterproductively, have a detrimental impact on black swallowtail populations. The abundance of introduced Apiaceae plants is actually good news for the butterfly, and the black swallowtail switching to an introduced food source causes no harm to the native plants since they do not rely on larval feeding to set seed and reproduce. In a rapidly changing climate, it is counter-productive to ignore or dismiss adaptations such as this.
Native plants are still important
These assertions should not be interpreted to mean that native plants do not matter.
Whenever I teach on this subject, I always take pains to point out that native plants are extremely important. We should conserve them, plant them, propagate them, and appreciate them.
But the importance of native plants does not mean that introduced plants have no ecological value. Native plants are extremely important, and introduced plants have ecological value, too.
While I find all of this information extremely useful, and use it to make plant selections, I am opposed to some degree to a utilitarian ranking of plants based solely on the number of insect species they can support. The natural world is incredibly nuanced and complex. Any overly binary system of understanding will never capture all of its beautiful, messy reality.
I believe every plant has value in its own right, and I still grow and appreciate plants, native or otherwise, that support few or no lepidopteran larvae. Many of our native grasses (Bouteloua, Sporobolus, Koeleria), wildflowers (Chrysogonum, Eurybia, Vernonia), and even some of our woody trees and shrubs (Cladrastis, Eubotrys, Itea) support a whopping 0 species of lepidopteran larvae, either native or introduced. (**Please note that the numbers on these specific genera come directly from the original 2009 Tallamy dataset. Further clarification by the author below. –MW)
I do not think that means those plants, or any other plant that supports very few pollinators, have no value, or that we should ignore them entirely. But I do think we can use information like this to re-evaluate how plants that are often vilified are actually integrating into our ecosystems.
Fig.5 – native Virginia Tiger Moth larvae feeding on an invasive Ampelopsis (Porcelain berry) vine
Judging plants by effects, not origins
This is a deeply fascinating and thorny topic, and the more we are able to view plant behavior with curiosity, and an eye towards their effects rather than their origins, we will be better stewards of the ecosystems that are under our care.
I am hopeful that with more information and context we will all be able to make more informed decisions about the management of wild plants, and have a deeper appreciation of the complex, chaotic interplay of plants and animals that is always around us, native or introduced, but wild, nonetheless.
**Clarification 4/27/26: Please note that the cited numbers on these specific genera come directly from the original 2009 Tallamy dataset. As it focuses on eastern North America, host numbers on grasses may be lower than in the Midwest. The Vernonia listing appears to be in error, but more updated claims of larval use on Eurybia are likely assuming use by Symphyotrichum generalists, as there have been no confirmed direct observations that I can find.
I cited these genera from the dataset as they were easily recognizable and could illustrate that a single metric like lepidoptera host numbers is not explicitly linked to nativeness, nor should it be used to justify disregarding or even removing plants based on low numbers in that one metric.
As a point of interest, from the Tallamy dataset and other sources, there are numerous other native plants that have no verified observations of larval host use such as: Itea, Fothergilla, Cladrastis, Xanthorhiza, Pachysanda (procumbens, a native species), and Galax. There are hundreds of common and uncommon native plants that support very few to no lepidopteran species as larval hosts (ferns in particular are very good at fending them off). This does not mean they are bad plants, it means they have evolved good anti-herbivory defenses that no insect has adapted around. We sometimes forget that plants do not want to be eaten! – NL
References:
Cech, R., & Tudor, G. (2005). Butterflies of the East Coast: an observer’s guide. Princeton University Press.
Harris, C., & Ratnieks, F. L. (2022). Clover in agriculture: combined benefits for bees, environment, and farmer. Journal of Insect Conservation, 26(3), 339-357.
Sax, D. F., Schlaepfer, M. A., & Olden, J. D. (2022). Valuing the contributions of non-native species to people and nature. Trends in Ecology & Evolution, 37(12), 1058-1066.
Stitcher, S. (2013). Black Swallowtail Butterfly. The Butterflies of Massachusetts. https://www.butterfliesofmassachusetts.net/black-swallowtail.htm
Tallamy, D. W., & Shropshire, K. J. (2009). Ranking lepidopteran use of native versus introduced plants. Conservation Biology, 23(4), 941-947.
Nice! Thank you!
We need more people like Dr. Lambstrom to counteract the human errors presented in data collected from prejudiced mind-sets toward “alien plants.” These “bad plants” that have successfully integrated into the environment for hundreds and thousands of years are now becoming the scape goat for a messed up world. I hope the tide will turn from militant eradication of “alien plants” to a more reasonable assumption that all plant life exists because of factors leading it to do so. Human extremists will falter as scientists discover the hidden truths regarding how the natural world functions. The participants in the extreme acts of eradication of “alien plant” life may have good intentions, but they are following the erroneous, inadequate dictates of others.
Thank you Kathleen! It’s very encouraging to me to see how many voices are calling for a second look at how we think about, talk about, and manage these plants. It’s given me what feels like a healthier, less combative relationship with the land around me.
Many thanks for publishing Nathan Lambstrom’s article about data manipulation in Doug Tallamy’s meta-analysis about the use of native plants by lepidopteran insects. And thanks to Nathan for making this information available to a broader audience with horticultural expertise.
Since Nathan’s article was originally published on my website, Conservation Sense and Nonsense, I would like explain why I published it.
I encourage everyone to plant whatever they want in their gardens, including exclusively native plants if that is what they prefer. My concern about the preference for native plants is that it often motivates public land managers to engage in projects that attempt to eradicate non-native plants, usually with herbicides. Here is an example of the consequences of a natives-only mindset by a home gardener: https://milliontrees.me/2024/11/01/dana-milbank-the-evolution-of-a-native-plant-advocate/
If you prefer native plants, by all means plant them, but that doesn’t give you the right to kill introduced plants with toxic herbicides on public lands that are owned by and accessible to everyone.
I can see see a lot of good points in the article, but WOW… seeing the sentence “plants that are often vilified are actually integrating into our ecosystems” right next to a photo of porcelain berry was quite shocking… Showing one hungry caterpillar is not good data to redeem that vilified monster. Yes, it is integrating in our ecosystems – in an extremely damaging way and smothering thousands of miles of forest edges in our region. Enough to make me not want to share this article further.
Hello Catherine! I do hope you continued reading the article to the end. I always stress that shifting our perspective around plants like porcelainberry doesn’t mean we can’t work to slow their spread, I have been doing so for decades and it is still a large part of my work and research. I also hope you’ll stay tuned for another article I’m working on around invasive plant rhetoric. It’s become increasingly clear to me, through decades of working with and studying plants like porcelainberry, that our language around these plants as monsters that smother miles of forests is not reflecting the complex reality of the situation on the ground. My next article will be more about this subject specifically, but the science increasingly suggests that plants like porcelainberry are not ecological destroyers, they are symptoms of a damaged ecosystem, and indications that our care is needed. Within the context of this article, though, it is certainly relevant data that 13 species of native butterfly and moth larvae are consistently found on this plant (far more than are found on some native vines like trumpet creeper or Carolina jessamine), and important information to consider when debating how to approach controlling plants like porcelainberry. If we think porcelainberry is ecologically dead and has no redeeming qualities, we might be more likely to accept extreme control measures and accept a certain degree of collateral damage.
Thank you for this article! Looking forward to part two!
it has always bothered me that so-called wildlife refuges do aerial spraying of cancer causing chemicals in the name of native plant restoration. read the label “toxic to fish and wildlife” which means us since we are part of the ecosystem.
I agree Sally. I don’t think there’s reasonable justification for that amount of collateral damage to people and wildlife.
All you have to do is look at a garden full of native and non native plants to see who is visiting. If you follow the basic rules of no chemicals, limit double flowers and hybrids chances are it’s humming with life. We often forget the plant world is always changing and many of the plants we call native have only been here a relatively short time. It worries me when people get overly judgemental about certain plants. All have value in some form especially when put together in an ecosystem. Yes some are thugs but that’s where we need to research the plants more thoroughly. I appreciate Dr Lambstrom providing a more balanced view based on the scientific research out there.
Excellent! ever since the push for native plants emerged, it has been clear to me that over-generalization is a major sales tool for promoting native plants. Thank you for this authoritative article.
Fascinating! Maybe this explains why Tallamy’s research results have not been duplicated, something that is pointed out by Linda Chalker-Scott and others in the Garden Professors group who seem to consider Tallamy more of an advocate than an objective scientist
Thank you Susan! I really admire Linda Chalker-Scott and really appreciate the perspectives on the Garden Professors. Your comment prompted me to revisit some of their articles on this subject and this line sums it all up perfectly for me: “ecosystems are not more complicated than we think; ecosystems are more complicated than we can think.”
https://gardenprofessors.com/native-vs-exotic-not-simple-seems/
“Right plant in the right spot” is my rule. Where the right plant evolved is irrelevant. All those who demand “natives only” in flora and fauna should be prepared to starve.
The article is online and not terribly technical, so I would suggest everybody read it before hopping on the criticism bandwagon: https://conbio.onlinelibrary.wiley.com/doi/epdf/10.1111/j.1523-1739.2009.01202.x
It’s a bit unfair to claim “The reasoning behind this [all genera with both native and introduced species are re-classified as “native” only] is not clearly explained in the paper”. The authors were entirely transparent about what they were doing (see that “both” column right next to the “origin (for analysis)” column?) and explained exactly why they did so: “most host-use records are only specified at the generic level. Our early attempts to compare host use of introduced and native plants in the same genus were thwarted by a lack of information at the species level”. So there was nothing dishonest or nefarious here, no “data manipulation”. They were pretty clear about their methodology and acknowledged its limitations.
Mr. Lindstrom acknowledges that native plants are “extremely important” but goes on to make the odd claim that he grows several native plant genera that “support a whopping 0 species of lepidopteran larvae, either native or introduced” when in fact several of those genera (Bouteloua, Koeleria, Sporobolus, Eurybia, Vernonia) are documented as hosts of numerous Lepidoptera. This focus on Lepidoptera is unfortunate when so many other insects feed on plants (I would direct this criticism at Dr. Tallamy as well, although they address this in the paper too) but the authors go on to note, “It is unclear how many of these species do support Lepidoptera but have been undetected as hosts, or, in fact, do not serve as hosts for any Lepidoptera species.” There’s a lot we still don’t know.
This paper was based not on original research but a review of the literature. As such, its conclusions are tentative and subject to further examination (and yes, even criticism). It would be great if a single study could address every issue, answer every question, without any limitations. But that’s not how science works. So let’s support more work like this and support funding for original research, because that’s the only way we can discover whether a preliminary study like this is correct or not.
Hi John, thanks for your comment. I always appreciate respectful dialogue and pushback on this which I think is sometimes lost in this topic, but is crucial to moving the field forward.
I have a lot of respect for Tallamy and wouldn’t accuse him of data manipulation, and didn’t do so in my article (I think you’re referring to a comment that was not made by me), but I disagree that the paper is straightforward about the reasoning for relabelling “both” genera as “native”. If the purpose of the paper was objective analysis of host use of genera by origin, it necessarily skews the results to essentially delete records of plants like non-native barberry that are known to host lepidopterans. The justification for doing so isn’t clear to me in the paper, but the post hoc defense I’ve heard is that if an insect is using a non-native plant it is likely because it has a close native relative. That is true but does not change the fact that insects are indeed feeding on those non-native plants, which isn’t reflected in the results when the origin is changed.
I agree with your other points, which I now see was not expressed very clearly in my article. My intention when I (clumsily) expressed the importance of native plants outside this one metric was that I still value plants regardless of the numbers of lepidoptera they host. The list of genera that host none came from the dataset in question, and I isolated those to point out that a single metric like lepidoptera host numbers is not only not explicitly linked to nativeness, it also shouldn’t be enough to justify disregarding or even removing plants based on low numbers in that one metric.
When I talk about the importance of native plants I’m thinking of their intrinsic value in their own right as members of wild plant communities we should take care to preserve. I’m also thinking of the rare but important cases where insects have extremely narrow host plant or pollen provisioning plant requirements, like Karner blue butterfly larvae needing Lupinus perennis, or Andrena erigenae, the small native bee, needing Claytonia pollen to feed their own larvae. We should be aware of and plant those specific plants if we wish to support those imperiled insects, not simply plant any native and assume that that is always best. Not to mention value, as you so rightly point out, to other classes of insects, arachnids, birds, numerous other wildlife and uncountable ecological interactions, human cultural and medicinal uses, erosion control, carbon sequestration, storm water infiltration, etc. (most of which are conclusively not traits limited to natives).
I certainly agree also that we can’t make broad generalizations from one paper, which was the central point I was hoping to make with this article. We see in gardening circles sweeping inaccurate statements around nativeness and ecological value that are based in part on way data is interpreted in works like this. Research increasingly shows that nativeness alone is not a good predictor of ecological value, however that term may be interpreted. I think this paper by Tallamy helped to move our field forward in a very important direction, and that it’s worth revisiting foundational papers like this in our search for more nuance.
I really appreciate your thoughtful comment and pushback in the search for that nuance!
Isn’t it fair to observe that the above conclusions should not be surprising, at least with respect to those native and “non-native”/”introduced” plants that are of the same genus? Does it not logically flow that those introduced plants which are part of Tallamy’s study and which are of the same genus as the native plants may very well be genetically similar and even considered related disjunct plants? That is, the introduced plant at one time in its history was either the same plant or closely related when there were land bridges, for example the Bering connection between North America and Asia, but ended up “evolving” in a different, separate region due to various historical forces. Studies have shown that many “conspecific taxa” or closely related species occur on both sides of these historical land bridges. I suppose this may raise the larger question of how important is the “evolution” of a plant species, and the location of that evolution, compared to genetic make-up when evaluating their attractiveness to insects?
Hi Tom. I do think that’s a good conclusion to draw. As a very general rule I think we now know that non-native plants closely related to native plants are likely to support most of the same insects. For the same reason we know that plants from large, diverse lineages will support more insects. Specialization on a rare plant with few living relatives is an evolutionary dead end. There’s a lot of interesting work being down now on insect specialization and the idea of co-evolution generally. It’s incredibly complicated and fascinating but one takeaway seems to be that specialization on a narrow list of plants is a less common, ancestral trait, and generalization is a more recently evolved trait that lends itself well to adaptation when it comes to incorporating novel plant species and genera.
I would also add, though I don’t know that this has been demonstrated empirically, that there seems to be a relationship between insect interactions and the amount of time an introduced plant has been present here. Clovers, corn, dandelions, plantains, and chicory have all been in North America for centuries, plenty of time for native insects to have adapted to their presence and incorporate them into their diets. Not only generalists but even some specialist insects have incorporated naturalized plants into their diets, to the point where they have become somewhat dependent on them. That shouldn’t be interpreted to mean I’m advocating we replace native plants with those non-native food sources, but instead that we have room for both in our ecosystems and should celebrate the ecological connections those plants are making.
Thank you for not only discussing, but also showing the data — it makes my librarian’s heart happy. And, you’re a gifted science writer. Want to write more for Garden Rant? 😉
Thank you Helen! I certainly hope to write more. Very happy to be a part of the community.
So, from the above response I suppose the point is that the evolution of a plant in a particular location trumps genetics, lending some amount of credence to the notion of the importance of “native” plants, at least in so far as attracting insects and perhaps other issues as well. Interesting. Thank you for your insightful article and your thoughtful responses to the comments.
Thanks Tom! The best predictor for insect support seems to be how many species are in the genus, more so than nativeness per se. Insects have more opportunities to adapt to use plants from a genus with a lot of wild relatives, and very few to adapt to species with few wild relatives. Natives with one species in their genus are just as poor in terms of insect interactions as non-natives with one species in their genus. This means a non-native willow will likely support as many species as a native willow, because the insects have already adapted around the willow’s defenses. This is also part of why clovers can feed so many insects: there are over a dozen widespread species, so adaptation to one (or to closely related pea family plants) usually means adaptation to the others, which is good news for insects. The other part of the equation for non-native is residency time. Insect species adapt in a relatively short time because of how quickly they can go through multiple generations. A plant like clover that has been here for centuries has had abundant time for insects to adapt to its presence and begin to utilize it as a food source.