Drawing on my work in evolutionary ecology, wilderness ethics, and machine learning, I reflected on the tension between our desire to intervene and our limited ability to forecast long-term ecological outcomes. Using examples like the Chernobyl exclusion zone—where many species are thriving in the absence of people despite nuclear contamination—I argued that ecological recovery is often less about precision intervention and more about restraint. We discussed how machine learning can help us simulate alternative futures and understand potential tradeoffs, but that ultimately, the most powerful conservation tool may be humility. More wilderness, not more control, might be the best way to meet the uncertainties ahead.

Listen to the episode here or wherever you get your podcasts.

Here’s my talk from the conference, “Evolution of Intrinsic Rates: Can adaptation counteract environmental change?“:

Rats have traveled around the globe alongside humans. Where we make our homes, they are happy to make their own, too. Unfortunately, they can make terrible neighbors, especially when they become vectors for diseases.

The rats that live in the slums of Salvador, are to blame for the widespread of emergence of leptospirosis infections that poses a serious health risk. In response, the municipality planned a multi-year rat-eradication project.

The bad news is that it is nearly impossible to fully eradicate rats from large territories. After just a short time, the populations tend to rebound, either because some rats were missed and then repopulate from within or because new rats migrate from outside and colonize the area. Understanding how rats repopulate after eradication efforts is important for deciding how best to proceed with future management strategies.

To test which of these scenarios was at play in Salvador, my collaborators trapped rats before, during, and after the eradication campaign in three contiguous geographic regions within the slums. We looked at the genetic relatedness of the populations across the three time-points. If the rebound populations were more similar to each other post-eradication, it would suggest that the populations were recolonized from source populations outside. However, if the regions became more genetically distinct post-eradication, it would suggest that the rebound populations were seeded from the few local genetic lineages that persisted.

We found that the populations showed distinct genetic differences immediately after the eradication effort, suggesting that remnant rats from the original populations had repopulated from within. Those difference persisted after 7 months.

In addition, we looked at the genetic signatures of population expansion and contraction. We found that post-eradication populations had pronounced reductions in genetic diversity. As animals were removed from the population during eradication efforts, it created a bottleneck. Since the new populations arose from the few remaining rats, only those few genetic variants left over persisted.

If enough genetic diversity is lost, and is not replaced by migrant mice from outside the original population, it might eventually lead to inbreeding and reduced fitness. If that is the case, it means that creating serial bottlenecks in the population through eradication efforts could be an effective way to weaken the rat populations to manageable levels, even if the overall population number rebounds in the short term.

I think my collaborator Dr. Jonathan Richardson (@JRichardson_44) did a nice job of succinctly outlining our findings in this Twitter thread:

I'm very excited about our new paper out today on the significant genetic impacts that accompany an urban eradication campaign of rats in 3 areas of Salvador, Brazil, w/ @georgisil_ & many Twitterless colleagues. The TL,DR version is in short thread below:https://t.co/WLqmQ1UJkD

— Jonathan Richardson (@JRichardson_44) June 6, 2019

Speciation starts small. The path of divergence starts with minute change in just a handful of basepairs among the billions of bases in a genome. Overtime, these small differences create feedbacks that propagate further divergence.

That’s the idea of sympatric speciation, the emergence of two species out of one. I’m most interested in the very first steps of this process, but catching this kind of small-scale divergence is hard, because the difference are small and easy to overlook. We call noticeable differences within a species polymorphisms. In most cases, we notice polymorphisms that we can see. For instance, it is hard to miss difference in color morphs of chimpmunks. Many of the classic examples of polymorphisms are visually detectable: industrial melanism in moths, beak size in Galapagos finches, stripe and color pattern in Timema stick insects, etc.

 

View this post on Instagram

 

It’s been way too long since I celebrated #TaxidermyTuesday All of these specimens are the same species of Eastern chipmunk (Tamias striatus), but two of them have genetic quirks that make their coats either all black (melanistic) or all white (leucistic). #biology #naturalhistorymuseum #naturalhistory #specimen #taxidermy #chipmunk #zoology

A post shared by Andis A. (@azandis) on Mar 6, 2018 at 3:54am PST

But many polymorphisms cannot be detected visually. For instance, polymorphism in dart frogs is often detected by the chemical composition of their skin toxins. Divergence in birds is often diagnosed by the sound of their calls. I’m in the middle of a collaboration to detect possible divergence in clover populations within and outside of cities via cyanogenesis.
The reliance on our human sight to detect polymorphisms is a problem because human sight, while impressive, is very narrowly limited. There is no reason for us to expect that the most relevant variation (from an ecological or evolutionary perspective) in a population would fall within the range of our senses. For instance, there is a whole vision of the world in the ultra-violet frequency that we cannot see. Flowers that all appear a uniform color to you and me can exhibit huge variation in ultra-violet light. Since birds and bees see ultraviolet, it stands to reason that the strongest selection for divergence would occur in this frequency, completely unknown to us.

All of this led me to think about how we, as eco-evo scientists, can start to shift our paradigm to think about the world through the senses of the organisms we study, which would give us a much better ecological picture of the selection pressures maintaining polymorphisms. The first step, is to take one small step outside of our anthropocentric sense and see the world through the eyes of animals. Thanks to the Yale Center for Teaching and Learning, I was able to design an interactive digital project that helps take those first few steps along a Proustian walk.

I’m excited to announce that my paper “A new, noninvasive method for batch marking amphibians across developmental stages” is now published at Herpetological Conservation and Biology.

This project originally grew out of frustration that no methods existed to be able to track amphibian larvae through metamorphosis into adulthood. This is a key bit of information needed in understanding amphibian ecology and asking questions like “How many tadpoles survive to adulthood?” “What percentage of frogs return to the same pond?” “Do tadpoles tend to school with kin or strangers?” etc. And ultimately this information is integral to estimating dispersal kernels and defining microgeographic variation.

Originally, I was looking into using radioactive isotopes to mark tadpoles, but just as my digging made that method seem intractable, I came upon a paper (Mohler 2003) that used a calcein fluorochrome solution to mark salmon.

Calcein binds to calcifying tissues like bones and scales. Although it had been tested in fish and bivalves, it had never been trialed on amphibians or any terrestrial applications. My study demonstrates that this technique is extremely promising for herps and solves a major limitation of marking amphibians. Below are the pertinent figures from the paper and supplemental materials, and also a couple extras.

Video of a wood frog tadpol approximately 24 hours after administration of calcein label.

This technique allows for both short and long term labeling. Short-term marking is detectable throughout the entire integument for 3-4 days and is visible in internal structures for up to 20 day in tadpoles marked within 28 days of metamorphosis. Labels are most useful for long-term marking (over 146 days) across metamorphosis when applied within 10 days of metamorphosis with 99% detection rate. If marked within 16 days of metamorphosis, the detection rate falls to 90% and sharply declines if tadpoles are marked earlier in development.

Check out the paper for more info. And also check out my poster on the project.

Not GFP!

A lot of folks ask me if this is technique is similar to the GFP (green fluorescent protein) that Shimomura, Chalfie, and Tsien discovered in the 60s (Tsien 1998). The answer is, no. GFP is a gene that can be introduced to animal genomes to induce production of a growing protein originally derived from jellyfish genomes. GFP is a genetic technique and so must introduced in germline or other stem cells. In contrast, calcein is a molecule that binds chemically to calcium. This means that calcein can be administered to any tissue for immediate fluorescence without interacting with the genome.

Thanks

This project required keeping lots of tiny frogs in the lab for almost 8 months, which turned out to be a major cleaning and feeding project a couple times a week. I couldn’t have done this work without loads of help from our lab manager and undergrad researcher–they deserve glowing medals for their service.

I especially would like to thank my brother Wes for spending part of his vacation playing scientist and helping me set up the experiment.

References:

Mohler, J. W. (2003). Producing fluorescent marks on Atlantic Salmon fin rays and scales with calcein via osmotic induction. N. Am. J. Fish. Manage. 23, 1108–1113.

Tsien, R. Y. (1998). The green fluorescent protein. Annu. Rev. Biochem. 67, 509–544.

But it turns out, when you take the time to look closely, evolution is taking place all around us, fast enough for us to see and measure. What’s more, evolution may even happen faster around us since we humans tend to create novel and often extreme selection environments that encroach into natural habitats (or entirely new habitats for those species that hitchhike into citeis with us). Johnson and Munshi-South (2017) reviewed the growing list of studies uncovering rapid evolution of wildlife in response to urbanization.

I showed the article to my partner (Baylaart.com) who used it as a theme for her most recent piece which is comprised of organisms subject to contemporary urban adaptations. Both the article and the painting are exceptional work that I’ll walk through below.

When faced with new environmental challenges, populations of organisms are faced with two options: move or adapt. (The third, extreme alternative is extinction). Urban wildlife populations are either residual populations that existed prior to urban development or new populations that colonized after a city emerged.

A classic example of urban adaptation is industrial melanism in the peppered moth (Biston bitularia) (Kettlewell 1955, 1958). As soot poured out of cities during the industrial revolution, the light colored polymorphism of the peppered moth offered poorer camouflage than the black, soot-colored morph. As a consequence, the dark morph came to dominate urban populations. As industry cleaned up its act, the trend reversed.

We can see the effects of urbanization in modern cities, today. White-footed mice (Peromycus leucopus), a North American native, in urban settings are marked by much less genetic variation, and therefore, lower evolutionary potential (Munshi-South et al. 2016). On the flip-side, this reduced variation could be the result of selection sweeping all unadapted alleles from the population, leaving a more genetically homogenous population as a result of the evolutionary process. Even animals that look unchanged may have evolved subtle adaptations. For instance, blackbirds (Turdus merula) in cities exhibit a molecular level difference in a gene associated with harm avoidance behavior compared to their natural brethren (Mueller et al. 2013).

Some species have adapted so well to urbanization, that they are almost synonymous with cities world-wide. For instance, the German cockroach (Blatella germanica), Rock dove (Columba livia), and Norway rat are veritable mascots of cities. Urban roaches and rats have evolved resistance to pesticides (Booth et al. 2011; Rost et al. 2009), and roaches in cities have even evolved an aversion to glucose in response to selection by sugar-baited traps (Wada-Katsumata et al. 2013). Rock doves in the cities have evolved defenses, not to human extermination attempts, but to predation by city-dwelling falcons (Palleroni et al. 2005).

Some populations precede city creation. Red-backed salamanders (Plethodon cinerus) in Montreal, Canada managed to persist as the city was constructed around them, but their populations have been isolated genetically, resulting in low variation (Noel & Lapointe, 2010). Across the Atlantic fire salamanders (Salamandra salamandra) also had a city (Oviedo, Spain) built atop their population starting over a millennia ago and managed to persist despite severe restriction in gene flow (Lourenco et al. 2017). Animals don’t need hundreds of years to adapt to new development, though. Water dragons (Intellagama lesueurii) inhabiting newly established city parks built as recently as 2001 in Brisbane, Australia have developed genetic difference in body shape and total size (Littleford-Colquhoun et al. 2017). Similarly, a small population of dark-eyed juncos (Junco hyemalis) established in the 1980s have been thoroughly studied (due largely to their location on the UC San Diego campus) and found to have evolved shorter wings, shorter tails, and alternate plumage pattern in just the past few decades (Rasner et al. 2004; Yeh 2007).

Other species, like the common wall lizard (Podarcus muralis) and striped mouse (Apodemus agrarius), seem to have been able to adapt to the development of Trier city in Germany (for lizards (Beninde et al. 2016)) and Warsaw, Polans (for mice (Gortat et al. 2014)) and now disperse through the city architecture in a similar way to their natural environment. Cityscapes can offer very similar (although much more angular) habitat to an organism’s natural habitat. For instance, Anoles (Anolis cristatellus) perch on the vertical trunks of trees in the forest. The flat walls of building in Puerto Rico offer a similar niche; however, artificial walls tend to provide less grips for lizards. In response, urban Anoles have evolved longer limbs and stickier toepads to cling to homes and businesses (Winchell et al. 2016).

Small animals are not the only wildlife subject to urban impacts. The movement of many large mammals, such as bobcats (Lynx rufus) (Serieys et al. 2014), are restricted by roadways, despite our best efforts to promote connectivity. In some cases, large and mobile animals are able to break into the new habitats afforded by cities. Red foxes (Vulpes vulpes) colonized the city of Zurich, Switzerland less than two decades ago and in that time their urban populations have exploded (Wandeler et al. 2003). While the original urban populations were likely established by just a few intrepid foxes, now that the urbanite populations are large enough, they have established genetic connectivity with their rural counterparts, essentially extending the larger population’s range to include the city.

In addition to landscape alterations that reduce gene flow and incur selection, urban settings can actually increase genetic mutation rates (which is ultimately the raw substrate for evolutionary adaptation). As an example, the rate of mutation in herring gulls (Larus argentatus) that nest in a heavily industrialized site in Ontario is double their less urban counterparts likely due to exposure to toxic chemicals in the environment (Yauk & Quinn 1996).

Let’s not forget that animals are not alone in their evolutionary response to urbanization—plants have also demonstrated adaptations to cities. Clover (Trifolium repens) in urban populations have evolved a reduction in cyanogenesis, a process that makes the plant less palatable to herbivores but in trade-off leaves the plant less tolerant of freezing temperatures (Thompson et al. 2016), which makes sense since there aren’t many large grazers wandering city streets and urban settings tend to act as heat islands.

In order to reduce seeds falling on infertile concrete streets and sidewalks, Holy hawksbeard (Crepis sancta) a weed in Montpellier, France have evolved to produce less dispersing seeds (Cheptou et al. 2008). In addition, the urban plants evolved an increase in photosynthesis and larger flowers (Lambrecht et al. 2016). Virginia pepperweed (Lepidium virginicum), is a common weed in many U.S. cities. A study of genetic divergence between urban and rural settings found that urban plants were more closely related than the more geographically proximal rural populations (Yakub & Tiffin 2016). In addition, the city pepperweed had developed a different shape and growing season to rural plants.

It’s clear that almost anywhere you look in cities, wildlife is evolving in response to our presence. This bestows us with a massive responsibility and indebts us to an ethic of conservation, not of species themselves, but to preserve as much unencumbered wild habitat as possible (for instance, as designated Wilderness). Where saving wild space is impossible, we must work on mitigating the effects of our urban infrastructure (for instance).

Literature cited:

Beninde, J., Feldmeier, S., Werner, M., Peroverde, D., Schulte, U., Hochkirch, A., et al. (2016). Cityscape genetics: structural vs. functional connectivity of an urban lizard population. Mol. Ecol. 25, 4984–5000.

Booth, W., Santangelo, R. G., Vargo, E. L., Mukha, D. V., and Schal, C. (2011). Population genetic structure in german cockroaches (blattella germanica): differentiated islands in an agricultural landscape. J. Hered. 102, 175–183.

Cheptou, P.-O., Carrue, O., Rouifed, S., and Cantarel, A. (2008). Rapid evolution of seed dispersal in an urban environment in the weed Crepis sancta. Proc. Natl. Acad. Sci. U. S. A. 105, 3796–3799.

Gortat, T., Rutkowski, R., Gryczyńska, A., Pieniążek, A., Kozakiewicz, A., and Kozakiewicz, M. (2015). Anthropopressure gradients and the population genetic structure of Apodemus agrarius. Conserv. Genet. 16, 649–659.

Johnson, M. T. J., and Munshi-South, J. (2017). Evolution of life in urban environments. Science 358.

Kettlewell, H. B. D. (1955). Selection experiments on industrial melanism in the Lepidoptera. Heredity 9, 323.

Kettlewell, H. B. D. (1958). A survey of the frequencies of Biston betularia (L.) (Lep.) and its melanic forms in Great Britain. Heredity 12, 51.

Lambrecht, S. C., Mahieu, S., and Cheptou, P.-O. (2016). Natural selection on plant physiological traits in an urban environment. Acta Oecol. 77, 67–74.

Littleford-Colquhoun, B. L., Clemente, C., Whiting, M. J., Ortiz-Barrientos, D., and Frère, C. H. (2017). Archipelagos of the Anthropocene: rapid and extensive differentiation of native terrestrial vertebrates in a single metropolis. Mol. Ecol. 26, 2466–2481.

Lourenço, A., Álvarez, D., Wang, I. J., and Velo-Antón, G. (2017). Trapped within the city: integrating demography, time since isolation and population-specific traits to assess the genetic effects of urbanization. Mol. Ecol. 26, 1498–1514.

Mueller, J. C., Partecke, J., Hatchwell, B. J., Gaston, K. J., and Evans, K. L. (2013). Candidate gene polymorphisms for behavioural adaptations during urbanization in blackbirds. Mol. Ecol. 22, 3629–3637.

Munshi-South, J., Zolnik, C. P., and Harris, S. E. (2016). Population genomics of the Anthropocene: urbanization is negatively associated with genome-wide variation in white-footed mouse populations. Evol. Appl. 9, 546–564.

Noël, S., and Lapointe, F.-J. (2010). Urban conservation genetics: Study of a terrestrial salamander in the city. Biol. Conserv. 143, 2823–2831.

Palleroni, A., Miller, C. T., Hauser, M., and Marler, P. (2005). Predation: Prey plumage adaptation against falcon attack. Nature 434, 973–974.

Rasner, C. A., Yeh, P., Eggert, L. S., Hunt, K. E., Woodruff, D. S., and Price, T. D. (2004). Genetic and morphological evolution following a founder event in the dark-eyed junco, Junco hyemalis thurberi. Mol. Ecol. 13, 671–681.

Rost, S., Pelz, H.-J., Menzel, S., MacNicoll, A. D., León, V., Song, K.-J., et al. (2009). Novel mutations in the VKORC1 gene of wild rats and mice–a response to 50 years of selection pressure by warfarin? BMC Genet. 10, 4.

Serieys, L. E. K., Lea, A., Pollinger, J. P., Riley, S. P. D., and Wayne, R. K. (2015). Disease and freeways drive genetic change in urban bobcat populations. Evol. Appl. 8, 75–92.

Thompson, K. A., Renaudin, M., and Johnson, M. T. J. (2016). Urbanization drives the evolution of parallel clines in plant populations. Proc. Biol. Sci. 283. doi:10.1098/rspb.2016.2180.

Wada-Katsumata, A., Silverman, J., and Schal, C. (2013). Changes in taste neurons support the emergence of an adaptive behavior in cockroaches. Science 340, 972–975.

Wandeler, P., Funk, S. M., Largiadèr, C. R., Gloor, S., and Breitenmoser, U. (2003). The city-fox phenomenon: genetic consequences of a recent colonization of urban habitat. Mol. Ecol. 12, 647–656.

Winchell, K. M., Reynolds, R. G., Prado-Irwin, S. R., Puente-Rolón, A. R., and Revell, L. J. (2016). Phenotypic shifts in urban areas in the tropical lizard Anolis cristatellus. Evolution 70, 1009–1022.

Yakub, M., and Tiffin, P. (2017). Living in the city: urban environments shape the evolution of a native annual plant. Glob. Chang. Biol. 23, 2082–2089.

Yauk, C. L., and Quinn, J. S. (1996). Multilocus DNA fingerprinting reveals high rate of heritable genetic mutation in herring gulls nesting in an industrialized urban site. Proc. Natl. Acad. Sci. U. S. A. 93, 12137–12141.

Yeh, P. J. (2004). Rapid evolution of a sexually selected trait following population establishment in a novel habitat. Evolution 58, 166–174.

The darwin is a great unit if you are interested in long-term macroevolutionary change, but sometimes it falls short for certain questions. For instance, what if we wanted to compare the evolutionary rates of a tree and its insect pest? One limitation of the darwin that Haldane noted was that it doesn’t account for generation times. Over the order of millions of years, it probably is a wash, but if you are interested in short time windows, say 1000 years, generation times matter, especially when comparing a tree with 100 year generation time and an insect that reproduces annually. For this purpose, Philip Gingerich proposed a unit he dubbed the haldane.² The haldane is similar to the darwin, but the rather than directly comparing mean trait values it compares trait values scaled to their standard deviation and rather than measuring time in millions of years, it measures time in generations. Here’s how it looks:

The haldane is good unit for change over time, but what if we are also interested in divergence over space? For instance, what if we are interested in comparing rates of evolutionary change in populations of fish in a river channel? We would expect there to be more divergence in population at the headwaters and the outlet than between populations just a few stream-miles apart, so how can we account for that? Richardson et al. proposed a unit for just this question in 2014.³ They suggested that there is a radius in which we would expect gene flow to disallow divergence. This is the dispersal kernel, depicted in their paper. Within that radius we would be surprised if trait values differed between populations, but outside we might be less surprised. Therefore, they scaled their unit to a distance ratio with that radius. To keep with tradition, they named their unit the wright after Sewall Wright, one of the fathers of population genetics. In their formulation, the difference in trait means is compared as in the darwin, but it is scaled to the pooled standard deviation and the distance ratio.

And this leads me to my current adventure in evolutionary analysis: how to best visualize the pairwise divergence of multiple populations with overlapping dispersal kernels? In my case, I have phenotypic traits for 15 population. I would like to visualize the pair-wise difference in traits with respect to their geographic distance. To further complicate things, I’m not entirely sure what distance to assign for a standard dispersal kernel.

My strongest idea so far has been to display a pair-wise matrix with wright values in the lower triangle and geographic distance in the upper triangle. I ran an ANOVA on trait means by location and used the multiple Tukey HSD post hoc comparison p-values to assign a shading to significant divergence in the lower triangle. In the upper triangle, red shaded values indicate distances less than a 1000m dispersal kernels. The outlined squares in the lower triangle indicate significant divergence within a dispersal kernel, or microgeographic divergence as defined by Richardson et al. 2014. Since I’m not sure what an average dispersal distance might be, I ran the same analysis for 500m, 1000m, and 2500m (only the 1000m analysis is shown). To help illustrate the matrix, I’ve include a nearest neighbor joining spatial plot for each dispersal kernel distance (500m, 1000m, and 2500m).

There is a lot of information digested into this one plot. So I’d be happy to hear any thought on how I could improve it.

¹ Haldane, J.B.S. (1949). Suggestions as to quantitative measurement of rates of evolution. Evolution 3:51–56.

² Gingerich, P.D. (1993). Quantification and comparison of evolutionary rates. American Journal of Science. 293-A, 453-478.

³ Richardson, J.L. et al. (2014). Microgeographic adaptation and the spatial scale of evolution. Trends in Ecology and Evolution. 29(3), 165-176.