Sunday, December 29, 2013

Homage to the Island of Misfit Toys, or why do wood frog populations seem locally maladapted to roads?

When I first began my dissertation research—bright-eyed, undaunted, and operating on full nights of sleep—I set my sights on roads, amphibians, and evolution. This was something of a transition from my previous incarnation, which is perhaps best described as a recovering liberal arts enthusiast with a mild interest in birds. Fortuitously, I think, I heeded my dissertation advisor’s suggestion by taking my “eye off the sky” in favor of what most might consider swamps.


A roadside pond primed for wood frog breeding. Due to its proximity, this pool has high concentrations of road salt. Yet, days after this photo was captured, hundreds of adult wood frogs began arriving to breed and lay their eggs.

It was in those swamps that I set about testing the predictions of evolutionary mediated responses to roads and runoff. With unabridged optimism, I set out to conduct a suite of experiments in my first field season. Seeing as how I was a complete newbie—having chosen to spend most of my undergraduate years embracing the youthful drive for renaissance over that of developing actual research skills—I had something to prove.


Renaissance, not research; me on the trombone in 2001.

Naturally, I chose to offset my lack of training with a healthy dose of enthusiasm and a ‘go-big’ approach. This amounted to doubling or tripling most sensible recommendations for sample size (and infrastructure). This conceal-inexperience-with-quantity logic seemed clear to me—if I could not impress my dissertation committee with a deep knowledge of ecology and evolution, I would impress them with my willingness to work hard. In reality, this overambitious plan required me to draggle friends and family aboard my sinking ship (of fool) to complete all the necessary sewing, dremeling, aquarium cleaning, surveying, and to of course keep me somewhat nourished if not adequately caffeinated.


My parents pitching in -- sewing tadpole cages late into the night.

Now before I discuss some actual biology—and hopefully make good on the title of this entry—let me backtrack a bit: why study roads? Well, there are an awful lot of them on the landscape. In the U.S. alone, we have enough paved roads to build a 25-lane highway from here to the moon, with a 12-lane soft shoulder to boot. Roads account for about 1% of the U.S. landscape, about the size of the state of Maine. And roads spur a host of consequences on habitats and organisms, from fragmentation to roadkill to runoff contamination. Among contaminants, road salt takes the cake (or the anti-cake if ferrocyanide is added..[1]). For example, in Connecticut, where I did most of this road-effects research, road salt is applied upwards of 30 tons per lane mile over the winter. That bears repeating: over the course of a single southern New England winter, each mile of road receives 30 tons of salt. That’s roughly the entire contexts of five dump trucks. By extension, over the course of the winter, the equivalent of a dump truck of salt is emptied every thousand feet. Bear in mind, this is per lane of travel.




The point is that that’s a lot of salt. Plus, salt use in deicing has been on the rise. Road salt is so widely and increasingly used that beginning in 2005, the deicing industry became, for the first time ever, the leading consumer of salts. This rather epic application of salts on the landscape has attracted a lot of research in environmental sciences (though for some reason, evolutionary perspectives have seldom been incorporated; I’ll save that for another blog entry). There is also interest stemming from public health as the increasing salinization of freshwater is now impacting private wells, which has raised concern for individuals with hypertension. While the effects of exposure to road salt varies, most organisms living near roads (which is to say, most organisms) do not fare especially well. Consequences can include arrested developmental and growth rates, behavioral modifications, and increased mortality. That is to say that road salt appears to be an agent of selection in roadside habitats. Not to mention, it’s rather tough on my 20-year-old Civic.




So, flashback to 2007 and my newfound interest in amphibians, roads, runoff, and the potential for contemporary evolution. I was investigating road effects in a system of pool-breeding amphibians in northeastern Connecticut, the so-called Quiet Corner of the state. Specifically, I was looking at road effects on the wood frog, a smallish frog that breeds in early spring, when adults abandon their subterranean winter hideouts in favor of ice-cold waters. They do this for one reason: to mate. Wood frogs are explosive breeders. While this term would undoubtedly yield interesting results on Urban Dictionary, here I am referring to the fact that a given pond can go from completely amphibian-free to overflowing with hundreds to thousands of adult wood frogs in a matter of days. It is quite the sight. And then, almost as quickly as it began, the breeding is done. The adults hop off to the woods, leaving behind the products of their reproductive endeavors; this amounts to about 800 externally fertilized eggs per female, glommed together in a single, gelatinous mass.


A wood frog on the move. Photo credit: Geoffrey Giller. geoffgiller.com

Life is tough for any of these newly laid embryos—turtles might come along and nab them up; water levels can quickly drop, leaving eggs high and dry; or temperatures can fall, causing eggs to freeze. But things are even tougher in roadside ponds. Because of the small and shallow nature of these ponds, contaminants can accumulate at high concentrations. For example, I was finding that ponds located about 30 feet from the road were 10% as salty as the ocean. This is not normal; Northeastern CT is far from the ocean. Indeed, ponds a couple hundred feet from roads were virtually salt free. Despite this impressive increase in salt in these roadside ponds, wood frogs and other amphibians are still breeding there. In fact, as I continued my work, I found no difference in the number of adult wood frogs breeding in roadside versus nearby woodland ponds.

So given this strong signal of road salt entering these ponds, the first thing was to figure out whether roadside ponds had any negative effects on the wood frogs breeding there. And the next thing was to figure out whether adaptive outcomes were present. As to the first inquiry, roadside ponds are indeed difficult places to make a living. I found that wood frogs in roadside ponds survive, grow, and develop at lower rates. So, there was evidence that roads were imparting a selective force. Coupled with this, there was good reason to believe that wood frogs were particularly good at adapting. For one, they tend to occupy rather distinct populations at small spatial scales. For two, there was already evidence of contemporary adaptation among populations in this system, albeit not to roads. Specifically, my dissertation advisor had recently reported adaptive responses to dynamic forest canopy regimes that differ across the scale of tens of meters. So given that adaptive responses to natural perturbations were already detected in this system, I expected to find evidence for local adaptation to roads. Indeed, with my excited (read: naïve) sense of vision, I found myself dreaming up the title of the Nature paper that I would write following my experiments, which I imagined would reveal contemporary evolution in an amphibian (a vertebrate, no less!) adapting to roads and runoff… the phrase “Salt saltation” seemed like it would be perfect on the magazine’s cover. Well, that prediction turned out to be squarely wrong. After a host of reciprocal transplants, salt exposure experiments, and field surveys, the data suggested that adaptation to roads for these animals was about as likely as my envisioned Nature paper. Whereas I expected roadside populations to show increased ability to tolerate road adjacency and exposure to road salt, I found just the opposite. Regardless of where embryos from roadside populations were grown—whether in their natal ponds, in nearby uncontaminated woodland ponds, or in various experimental concentrations of road salt—they performed worse than those animals from woodland populations. This was not local adaptation – but what was it? Could it be local maladaptation? And given these negative effects, how did these populations persist at sizes equivalent to those inhabiting the clean, woodland ponds?


Survival (±1 SE) is shown here as the mean proportion of individuals surviving to feeding stage across all experimental units (N = 99). The woodland deme is represented by open circles while the roadside deme is represented by filled squares. The environment in which the animals were grown out is on the x-axis. From Brady SP. 2013. PeerJ. DOI: 10.7717/peerj.163/fig-2

Before we can answer these questions, we should try to define local maladaptation. From a research perspective, local maladaptation has not seen the attention that local adaptation has, so finding a clear definition is somewhat challenging. I think it is reasonable to use local adaptation as a corollary. In this case we would define local maladaptation as a genetically-based process leading a local population to have lower fitness in its local environment compared to that of a foreign population within that same environment. So far, the pattern I found is consistent with this process. But here’s a twist: even when roadside wood frogs were transferred to the woodland environments, they still survive at lower rates. So what would we call this outcome? Well, for lack of a better term, I have been referring to this as deme depression, drawing on the analogous pattern of reduced fitness that is caused by inbreeding depression. Certainly, my description of these outcomes and their significance is underdeveloped here and deserves more discussion at some point. But for now, let’s move on toward accounting for how these patterns might be persisting in this system.

So given the locally reduced survival inherent in roadside populations, we have evidence for local maladaptation. More accurately, we can say that roadside populations ‘perform maladaptively’, but we do not know whether it is maladaptation per se. This is because maladaptation sensu stricto refers to reduced fitness owing to genetics. In the context of roads and runoff (and other human-modified environments), we might expect maladaptive genetic changes to be more common than, say, in natural environments. I offer this suggestion cautiously, as it is based purely on the known diversity of mutagenic contaminants found in roadside ponds. These include heavy metals like platinum and copper from catalytic converters and brake pad wear, and aromatic hydrocarbons (think toluene, benzene, PAHs) from fuel, tires, and the road surface itself. Yet, because the animals in my experiments originated from wild-laid eggs, we cannot say whether there is a genetic component underlying this performance deficit. Instead, this pattern of reduced survival—so-called maladaptive performance—might well be the workings of an experimental artifact. Specifically, it could be that the wild-laid eggs I collected were compromised by runoff the minute they hit the water. In other words, early embryonic exposure might have caused performance detriment in later life history stages. It is certainly plausible to think this might be happening, and so, I tested it. And it turns out that this was not the case: early exposure to roadside water had no carryover effects on performance. Still, environmental effects can be transmitted directly from parent to offspring without the offspring experiencing the aquatic environment first-hand. These so-called maternal effects (or, more broadly, inherited-environmental effects) are widespread in nature and certainly might explain maladaptive performance here. While I found no difference in egg size (a typical maternal effect in amphibians), a colleague and I did find that eggs are pre-loaded with elevated concentrations of methylmercury. The magnitude of methylmercury found in eggs corresponded to that found in the mother. So now we know it is possible for wood frog eggs to be exposed to and accumulate contaminants through their mothers before they are laid into the external environment.

Between inherited environmental effects and these putative mutagenic effects, we have one plausible and one maybe possible mechanism to explain locally maladaptive performance. We could also invoke inbreeding depression and/or drift as explanations, but since these populations do not differ in size, it is difficult to conceive of these processes playing out in the roadside environment but not in the woodland environment. Regardless of the actual mechanisms at play, we still need to figure out how these maladaptively performing populations persist. Could it be that roadside habitat is a verifiable sink, with woodland populations serving as the source? Perhaps. A rescue effect of sorts—wherein woodland animals are leaving to further colonize roadside ponds—could certainly explain why these roadside populations don’t appear to be dwindling in size. But if that were the case, we would expect a more similar phenotype between roadside and woodland populations. In other words, if roadside populations are sustained through the arrival and reproductive contribution of woodland emigrants, wouldn’t the performance traits of the offspring be similar between the two population types? Actually, the answer to that question depends on whether the emigrating population is a random sample of the woodland population. If it is, then yes, we would expect comparable phenotypes. But what if the emigrating group was not composed of a random sample of the population? What if instead the emigrating group comprised some biased subset of the population that was already compromised in some way?

Enter the Island of Misfit Toys. If you’re unfamiliar with the classic stop motion animation version of Rudolph the Red-Nose Reindeer, or if you’ve overlooked some of the finer details of the movie, here’s short a recap: (cue Solsbury Hill background music) a young Rudolph, made to feel reindeer non grata for his glowing nose, runs away from home. In the odyssey that follows, Rudolph meets up with another pariah (Hermey, the elf whose dream is not that of making toys but of mending teeth) and an eccentric silver and gold prospector (Yukon Cornelius). During their adventure, this crew of misfits eventually lands upon the Island of Misfit Toys, where all mal-designed, unwanted toys live out their days.




So perhaps, just as the misfit toys sought solace and companionship on their island, misfit frogs are seeking as much in roadside ponds. Perhaps misfittery loves company? This misfit island hypothesis could be playing out in a variety of ways, all of which would hinge on habitat-oriented dispersal into roadside ponds. For example, in wood frogs, we know that younger adults are relatively poor provisioners of offspring. Thus, misfit frogs might just be younger, and more likely to disperse into inferior (roadside) habitat. Alternatively, perhaps the dispersing animals are weaker (regardless of age), and therefore less capable of securing mates among the healthier animals that breed in woodland ponds, thus seeking companionship among the similarly conditioned animals in roadside ponds.

There are several other processes that might account for the persistence of these maladaptively performing roadside populations. One is that there appears to exist a quantity over quality tradeoff in reproductive output. Namely, while roadside embryos show inherently reduced survivability compared to woodland embryos, roadside females lay more eggs per clutch. So while the chance of survival is lower for a given roadside embryo, each roadside female lays approximately 10.5% more eggs than her woodland counterpart. Therefore, it is possible that this numerical investment helps offset population decline. Another possible explanation for the persistence of this maladaptive pattern is that roadside wood frogs might be adapted to later life history stages. For example, it may be the case that adults are locally adapted to the terrestrial environment. It is further possible that such an adaptation is associated with poor offspring performance, but that adult survival is more important for population fitness.

In the end, Santa promised Rudolph that he would find a home for all the misfit toys. Here, I have no such promise to offer—the fate of the (putative) misfit frogs remains an active area of research. So for now, this story is to be continued.


[1] Ferrocyanide is sometimes added to road salt to prevent it from caking up; hence the unfortunate pun.

Tuesday, December 24, 2013

Epic Wrap Battles of Christmas


I recently published a paper that I wanted to introduce here – but I had been delaying because my posts usually start with some sort of silly introduction that I attempt to segue into something more scientific. I didn’t have the silly intro I needed until yesterday.  Christmas brought it gift wrapped.

My daughter Aspen came home from her last day of school before Christmas quite pleased with herself for having won a class debate. The topic was: “For Christmas, which is better: a single large present or many small presents.” To hear Aspen tell it, the debate started off with about 80% of the students on the many-small side but, after she argued the single-large side, the proportions switched. “What were your arguments?” I asked her. Very seriously, she pointed out that “Big presents are more likely to be expensive and higher quality, and you will appreciate a big present more and spend more time with it and treasure it more.” Fair enough. I am convinced.

Evolutionary biology has also been having a many-small versus few-large debate. The key question has been – at its simplest – whether adaptation is driven by many genes of small effect or by few genes of large effect. The weight of opinion was classically toward many genes of small effect as represented by the field of quantitative genetics and made most stark by Fisher’s infinitesimal model. More recently, however, a perceptible shift has occurred toward the few-large side of the battle. The prime reason has been the recent publication in high-profile journals of various genes FOR this trait or FOR that trait – and, probably even more influentially, repetition as such in the popular press absent the caveats usually tendered in the original publications.

Variation in some traits is clearly influenced by only a few genes (Mendel’s peas being the icon) and the infinitesimal model is clearly unrealistic in its caricatured form (an infinite number of genes of miniscule effect). However, the vast weight of evidence remains – even in this new gene-centered world – squarely in the lap of many genes of small-to-modest effect. In the paper this post introduces, I provide three main reasons, reproduced here in shortened form (and without the examples and citations):

(1) Current genomic methods are strongly biased against genes of small effect. This bias is particularly obvious in candidate gene approaches, which deliberately target just the opposite. A strong bias is also present in linkage and association mapping, where estimation problems arise when relevant alleles are found at low frequency, when not enough recombination has occurred to break up large linkage blocks, when the number of individuals is few, when the number of loci is few, when the effect size of alleles is small, and from the need to assume a high threshold effect size to reduce study-wide type-I errors.

(2) Nearly all studies have sought to explain variance in specific traits, rather than overall adaptation (or ‘fitness’). This distinction is critical because overall adaptation to a given environment will be influenced by many traits. As a result, even the genes explaining high levels of variation in a particular trait might contribute little to overall fitness differences.

(3) Genome scans typically reveal that high-differentiation outliers, presumably influenced by divergent selection, represent 5–10% of the genome—and the distribution of these loci clearly implicates multiple unlinked genes.

The explosion of modern genetic technology has forced a re-evaluation of the many-small versus few-large debate, as well as many other cherished and entrenched ideas. Some classic ideas will stay with us, others will come to rest in the same place as my parent’s 8-track tapes, and perhaps others will end up somewhere in the middle (vinyl! – my brother is getting a record player for Christmas and I want one too).

While writing my book on Eco-Evolutionary Dynamics (nearly done!), I found myself mostly discussing phenotypes for which the genetic basis was unknown. I am not apologetic about this, of course, because phenotypes are the foundation of ecological effects on evolution (phenotypes are directly under selection, whereas genes are only indirectly under selection through their effects on phenotypes) and of evolutionary effects on ecology (phenotypes have direct effects on ecological processes, whereas genes have only indirect effects through their effects on phenotypes). However, lest readers say “Your book should be called Eco-Phenotypic Dynamics”, I figured it would be a good idea to have a chapter that directly addressed the genetics and genomics of Eco-Evolutionary Dynamics. 

While I was preparing the chapter, I received an invitation to submit a review paper to Heredity and decided to convert the developing chapter into that review (it will also appear in the book). The paper was published a few weeks ago. One topic is the few-large versus many-small debate summarized above as an example and the others are summarized in the paper's abstract:

Increasing acceptance that evolution can be ‘rapid’ (or ‘contemporary’) has generated growing interest in the consequences for ecology. The genetics and genomics of these ‘eco-evolutionary dynamics’ will be—to a large extent—the genetics and genomics of organismal phenotypes. In the hope of stimulating research in this area, I review empirical data from natural populations and draw the following conclusions. (1) Considerable additive genetic variance is present for most traits in most populations. (2) Trait correlations do not consistently oppose selection. (3) Adaptive differences between populations often involve dominance and epistasis. (4) Most adaptation is the result of genes of small-to-modest effect, although (5) some genes certainly have larger effects than the others. (6) Adaptation by independent lineages to similar environments is mostly driven by different alleles/genes. (7) Adaptation to new environments is mostly driven by standing genetic variation, although new mutations can be important in some instances. (8) Adaptation is driven by both structural and regulatory genetic variation, with recent studies emphasizing the latter. (9) The ecological effects of organisms, considered as extended phenotypes, are often heritable. Overall, the study of eco-evolutionary dynamics will benefit from perspectives and approaches that emphasize standing genetic variation in many genes of small-to-modest effect acting across multiple traits and that analyze overall adaptation or ‘fitness’. In addition, increasing attention should be paid to dominance, epistasis and regulatory variation.”

I am confident that these “conclusions” have been (if recent emails are any indication), are being, and will continue to be met with agreement by some and derision by others. I hope that folks will take this posting as an opportunity to provide points and counterpoints on these questions. Have at ‘er. Until then – time to finish the rest of the damn book.

My Christmas haircut from the 70s: an attempted mash-up of the 1870s (beard) and 1970s (hair).


---------------------------------------------------------
The paper:
Hendry, A.P. 2013. Key questions in the genetics and genomics of eco-evolutionary dynamics. Heredity 111:456-466. PDF

A related book chapter that Rowan Barrett and I published last year:
Barrett, R.D.H., and A.P. Hendry. 2012. Evolutionary rescue under environmental change? Pages 216-233 in U. Candolin and B.B.M. Wong (editors). Behavioural responses to a changing world: mechanisms and consequences. Oxford Univ. Press, Oxford, UK. PDF

An early many-small versus few-large post on this blog: Fisher is dead, long live Fisher

And, yes, the title is a parody of Epic Rap Battles of History

Saturday, December 21, 2013

Plasticity and genetics in un-invaded refuges

Plasticity and genetics in un-invaded refuges:
Possible consequences for native-exotic species co-existence

One of the greatest contemporary threats to the biotic integrity of native aquatic communities is the invasion and rapid geographic spread of exotic species. Whereas much research is dedicated to understanding factors that control the establishment and impacts of exotic species in non-native habitats, less research has addressed how native species can mitigate the ecological impacts of invaders through evolution. In many cases exotic species can cause the extirpation of native species – but there are also circumstances that allow co-existence of native and exotic species within the same region. In particular, plasticity and maternal/genetic effects may play a special role in enhancing spatial subsidies of dispersing individuals from un-invaded refuges at invaded habitats. In our study, we consider how plasticity versus genetic/maternal effects in native amphipods (Gammarus fasciatus) that inhabit un-invaded refuges possibly contributes to their co-existence with exotic amphipods (Echinogammarus ischnus) at invaded habitats in the upper St. Lawrence River, Quebec, Canada.


Lac St. Louis, QC, Canada (facing west from Montreal), showing distinct ion-rich (left) and ion-poor (right) water masses. The native amphipod persists across this ion gradient, but the exotic amphipod is restricted to ion-rich habitats. Photo credit: St. Lawrence Centre, Environment Canada

We anticipated high trait plasticity in the native amphipod because of high gene flow and high spatial and temporal environmental variation in our study system. Unexpectedly, we detected both plastic and genetic influences on the traits of native amphipods along natural ion gradients. Plasticity was detected in only one trait, post-moult calcification, and three other traits (larval survival, time to first reproduction, and fecundity) were influenced by genetic variation or maternal effects that differed between ion-rich and ion-poor habitats. Overall, we advance a hypothesis for how native trait plasticity and genetic/maternal effects at ion-poor, un-invaded refuges could influence native species persistence in face of exotic invasion along environmental gradients through spatial subsidies. While the most successful invaders tend to be generalists that can displace specialist natives in many cases, our hypothesis suggests that the converse can also be true - generalist natives can persist with more specialized invaders when spatial demographic subsidies are enhanced by plasticity and genetic/maternal effects across environmental gradients. Future work will address the net effect of multiple influences from plasticity and genetic/maternal effects on overall fitness of native amphipods dispersing from ion-poor, refuge habitats at ion-rich, invaded habitats.


Native Gammarus fasciatus amphipod.

We recently published this work:
Derry, A.M., A.M. Kestrup, and A.P. Hendry. 2013. Considering potential plastic and genetic/maternal effects on demographic subsidies: Native species persistence with detrimental exotics. Functional Ecology 27: 1212-1223.

We are looking forward to further developing and testing these ideas on amphipods and other cool creatures in Lac St. Louis in future projects.

— Alison Derry

Monday, December 16, 2013

QCBS Meeting 2013

For the last couple of days, a large number of ecologists and evolutionary biologists working in Quebec gathered to talk about science, conservation and policymaking. This was the annual QCBS meeting (Quebec Centre for Biodiversity Science). As usual, the meeting had lots of young scientists eager to talk about research and potential collaborations – I guess this is why the QCBS meeting always has a casual and friendly feeling. There was no rush between sessions and the program was well organized. If you missed a talk, you could always just approach the speaker during a coffee break or over a beer (and you could also follow the entire conference via Twitter). Since the meeting is fairly local – QCBS is only Quebec-wide – I was happily surprised by the diversity of topics people work on, including invasive species, community ecology, evolutionary biology, and landscape ecology.
One recurrent theme that I noticed was the promotion of open science, including both open-access tools like R and Latex as well as open-access data. Young scientists are more and more preoccupied with how science is conducted and how it is promoted to the public. QCBS is at the forefront of the movement to open up science, and this was obvious during the meeting. One of the plenaries, “The open movement in biodiversity science: tools and data sharing practices”, was given by three young researchers (Geneviève Allard, Scott Chamberlain and Timothée Poisot). The talk was focused on how data-sharing and open access tools could be used to make science more transparent. If you think about it, nobody really replicates experiments nowadays! How are we supposed to be confident in scientific findings if nobody bothers to confirm the results of past studies? When I was an undergrad, I learned that scientific advancement was achieved through replication and validation. But if nobody does those things, how do we advance today? I don’t know the answer, but perhaps open science is a good way to start, by at least allowing us to check each other’s work more informally, and allowing the public more of a view on the process . So go open!
The guest speaker: Prof. Ivette Perfecto

Scott Chamberlain talking about open science and how to go open

Our own Kiyoko Gotanda talking about how predictable is guppy evolution

One final thing that I would like to say about the QCBS meeting is that many of the students that were presenting their work were recipients of QCBS grants that helped them with travel costs and stipends. Since its establishment, QCBS has funded over 120 grad students (as well as many undergrads and postdocs) to travel to conferences and workshops. As one of those 120 recipients, I would like to say “thank you and keep the amazing work!” This support really makes a difference in a graduate student’s career.



Wednesday, December 11, 2013

Species Concepts: Past, Present, and Future

No matter what your study system, whether you're a theoretician or an experimentalist, whether you work with genes or whole organisms – if you do research in ecology or evolution, chances are you're going to have to decide what species concept you're going to employ. When I started my PhD last year, I had to undergo the difficult process of moving from one species concept to the other. I did my MSc in a lab that focused heavily on taxonomy and systematics, and so tended to use the phylogenetic species concept (PSC). In my thesis, I sought to confirm that a species of sea slug (Haminoea japonica), found on the west coast of North America and in Europe, was the same as one found in Japan. Using three gene fragments and the PSC, I determined that the American/European species was indeed the same as that in Japan (the slug had invaded from Japan, most likely with shipments of oyster spat). In this study, the PSC was the most appropriate species concept to use, as we were primarily interested in using the genetic relationships of one population to another for identification purposes, and using the biological species concept (BSC) would have been unfeasible (importation of live invasive molluscs and captive breeding/mating trials of sea slugs both being difficult, if not impossible, tasks) and also irrelevant.

Haminoea japonica

Photo credit: Ángel Valdés

Fast forward to August 2012 and my entry into the stickleback world at the 7th International Conference on Stickleback Behavior and Evolution in Seattle. Of course a major theme at this conference was ecological speciation – the stickleback being a well-studied model of the process – and the implied species concept used was the BSC. This makes sense, of course; those who study ecological speciation aren't interested in defining the species, but rather in the process that creates those species. Using the PSC isn't useful here – just knowing that two things are different doesn't help you understand how they became different. And so, after arguing with Andrew about the merits and demerits of each species concept, I came to accept the BSC.

What I should actually say is that in the context of the work I'm doing now (parallel ecological speciation in threespine stickleback), I accepted that the BSC is the most appropriate species concept. But that's not to say I think that one species concept is inherently more correct than another; it just depends on the question you're trying to answer. Borrowing imagery from the "adaptive peak" metaphor for selection and adaptation, imagine a large mountain with twin peaks. Should such a formation be conceptualized as two mountains, or one? As with the species concept, I maintain that it all depends on your question. If you were stranded on one of the peaks, and were radioing for help, you would want to use a "mountain concept" that classified the two peaks based on their unique characteristics, such as geographic coordinates (analogous to the PSC). On the other hand, if you were a geologist trying to understand why there were two peaks instead of one, you would be most interested in a "mountain concept" that defined the processes by which peaks are made, such as uplift and erosion (analogous to the BSC).


Two mountains or one?

Photo credit: Wendenburg

I don't think I'm saying anything especially significant here, and in any case, books and articles defending one species concept over the other will continue to be published in prodigious quantities, just as they have been in the past.  But maybe for now, in the spirit of the holiday season now upon us, we can learn to accept each other's species concepts as context-dependent, and live in peace and harmony!

Monday, December 2, 2013

Carnival of Evolution #66

Carnival of Evolution #66 is up, with a Doctor Who theme.  Our contribution of the month is Timothy Farkas's post on eco-evolutionary dynamics in Timema cristinae.  It's quite a fun post, so check it out if you haven't already!

The evolution of Dr. Who.  What determines his phenotype over time?  Does the breeder's equation apply?  If so, what is the G-matrix – and is it constant?  Or is all this variation due to phenotypic plasticity?  Could it be some sort of aposematic display, perhaps?  The fourth Doctor's scarf certainly seems like an example of niche construction.  As always, further research is needed.

Sunday, November 24, 2013

Finding God (and other memes) in Paris

I was recently in Paris for a meeting of the scientific committee of DIVERSITAS, where I am co-lead of the core project bioGENESIS. Owing to the annoying tendency of airlines to charge double if you do not stay in Europe over a Saturday night (so they can fleece business travelers who would not do so), I had the happy inconvenience of a couple of extra days to browse Paris. This included my second entrance into Notre Dame de Paris, which got me to reflecting ...

Notre Dame de Paris.

As one would expect of a scientist, I take an evidence-based approach to ideas. Thus, if God were ever to provide concrete evidence of His existence, how could I do anything other than accept it? The closest I ever came to such evidence was on Christmas Day, 2001. I was visiting Paris, seeing all the sights, and taking copious pictures on my Nikon F4 camera. The F4 was the best professional camera Nikon made in the 1990s, and it had served me faithfully and without incident for 10 years. On that Christmas Day, I found myself outside the grand cathedral Notre Dame de Paris. I had never seen it before and it certainly didn’t fail to impress with its grandeur and age (almost 840 years old). Even though it was Christmas Day, they were letting tourists in. How memorable it would be, I figured to enter Notre Dame on Christmas Day 2000 years after Christ’s birth – and so I did. It was spectacular indeed – all the more so because Christmas Mass was underway. The penitent were in the center of the cathedral singing various hymns while the tourists walked around the outside taking pictures. It seemed rather rude to me that folks would be taking pictures, including with flashes, while the faithful were speaking to God, but everyone was doing it and there I was. So I sheepishly brought the camera to my eye, composed the photo, and pushed the shutter button. Nothing happened. Absolutely nothing. Then, of course, I went through all of the problems that might seem obvious – dead batteries, some weird setting, and so on – but I couldn’t find anything obviously wrong, except for the fact that everything was seemingly wrong. The camera was just broken. And it never worked again until I got it back from a repair shop months later. 

Notre Dame de Paris

A pious interpretation of this event might be that God punished my impudence and sent me a message. Or perhaps it was just a coincidence. If, however, the same thing happened on my next visit to Notre Dame de Paris, then perhaps I would need to reconsider. Well, that chance came last week. There I was, standing outside the grand old dame – still just as impressive and now even older, precisely 850 years. Again I entered, again Mass was underway, and again I had a top-of-the-line Nikon camera in hand. I raised the camera to me eye, composed an image, pushed the shutter button, and click. Click, click, click, click … 

God is a meme – an idea that spreads from one mind to another – and it is a very effective one. In fact, a number of social scientists have argued that this meme has not only spread effectively but that it has increased the fitness of individuals carrying it. I suppose this isn’t too surprising given that a tenet of most religions is that reproduction is good (and contraception bad), which presumably increases the chances that the meme will pass effective to more people (the children) and thus spread farther and faster.

Eiffel's meme.

Paris is full of memes. The Eiffel tower is one. I had been to Paris many times but never to the top of the Eiffel Tower – it just seemed too kitschy and touristy. So I had instead spent my time wandering around the just-as-touristy, if considerably less kitschy, Louvre and Orsay museums. Plenty of memes there too, like Greek sculptures copied by Roman and then French sculptors, like Théodore Géricault’s Raft of the Medusa begetting Romanticism, like Titian forever the art of portrait painting – and so on. On this trip, however, I figured I had better also do the tower just to have done with it. So up up up I went, walking as far as possible and then by elevator the rest of the way to the top, which was buried in the clouds. I quickly realized that the Eiffel Tower wasn’t so much a kitschy tourist attraction as a spectacular engineering feat – and displays on the way down showed how the design was much copied immediately afterward – the tower as a meme.

A design much copied.

Religion and engineering marvels might seem admirable or important memes but other memes are just silly – and Paris has no shortage of those either. How about those bridges that are covered in thousands of locks – literally thousands and thousands of them? Who was the first person who put a lock there? Or, perhaps more importantly, who was the first person who put the lock there simply because someone else had put a lock there. And when did someone decide to call these “love locks,” with lovers writing their names on the lock and throwing the key into the Seine. Even governments can’t stop this meme – they keeping cutting them off the bridges in many cities and they just keep coming back.

A new meme and an old meme

As this is an eco-evo blog, I suppose I need to get somewhere with this meme theme. I could perhaps just do the easy “how would memes influence eco-evolutionary dynamics” - but that would lead to rather obvious ramblings about how memes, like any behavior, could dramatically alter ecological dynamics at the population, community, and ecosystem levels – and that this effect might be more dramatic because memes can spread more quickly than genetically based behaviors. But that would just be trying too hard. So instead, let’s just stop to marvel at memes, big and small, important and trivial, good and evil – and also to remember that scientific ideas are memes too: they come, they compete, they fade away. But some are here to stay, such as heliocentrism, relativity, natural selection, hopefully eco-evolutionary dynamics, and – dare I say it – the God Particle. Now that I believe.





Related news?







Wednesday, November 13, 2013

Ecological effects of camouflage evolution in the stick insect Timema cristinae


Shortly after joining Patrik Nosil’s newly formed lab at the University of Colorado, making it a busy research group of two, I was coming to grips with the reality that I needed to gear up for a serious transition from pure community ecology to hard-core evolutionary genomics. This all turned around on a dime, however, after Andrew Hendry’s epic visit to Boulder during the winter of 2009. Details are fuzzy, though I vaguely recall a near-boiling hot tub, and too few beds at the end of a long night. Anyway, despite scalded feet, Andrew’s visit ended with a changed direction for my PhD and a new line of research for the Nosil Lab: eco-evolutionary dynamics!

Independent of my reorientation towards the community ecology of eco-evolutionary dynamics, we were contacted by Ilkka Hanski who, knowing Patrik’s work, was keen to use the stick insect Timema cristinae to evaluate one of his recently developed mathematical models, which makes fine-scale predictions about spatial patterns of local adaptation (Hanski et al. 2011). Timema cristinae experiences strong divergent selection from birds for crypsis on its two dominant host-plant species, Adenostoma fasiculatum and Ceanothus spinosus, causing the evolution of striped and unstriped morphs, each better camouflaged on Adenostoma and Ceanothus, respectively (see figure below, Nosil and Crespi 2006). However, the homogenizing effects of gene flow create complex spatial patterns of local adaptation throughout the landscape (Bolnick and Nosil 2007), setting up an excellent scenario to test Ilkka et al's model. It seemed as though Ilkka’s proposed research would dovetail nicely with the ecological work we were planning to do in Timema come spring, so we began a collaboration with him and his postdoc Tommi Mononen. This collaboration culminated last month with a nice publication in Current Biology (Farkas et al. 2013).

 (A) striped and (B) unstriped T. cristinae, (C) Adenostoma fasciculatum, (D) Ceanothus spinosus
Black bar indicates approximate 1 cm scale 

Rewind to the spring of 2011, when Tommi, Patrik, my labmate Aaron Comeault, and I shared a one-bedroom apartment in Santa Barbara, in which Patrik insisted on sleeping directly in front of the only bathroom, apparently to prevent mid-night toilet visits by the rest of us. Led by Tommi, we trekked into the Santa Ynez National Forest and began to collect data. Using steel fence stakes, flagging tape, string, and a tape measure, we manually mapped (old-school) 186 host-plant patches of Adenostoma and Ceanothus inhabited by Timema, and sampled the bushes by whacking their branches with a stick, recording the abundance of the two morphs on each bush. We also performed two field experiments, one of them in 2012, manipulating the number of striped vs. unstriped Timema on bushes of Adenostoma. After the fieldwork was done, behind the scenes in Helsinki, Tommi and Ilkka scripted away the wee hours of every morning, evaluating their model with our data, and generating new predictions for us to test.

From these studies, we found a very strong influence of evolutionary dynamics in T. cristinae on ecological patterns, both for T. cristinae populations themselves, entire cohabitating arthropod communities, and interactions between herbivores and their host plants. In the mapped metapopulation network of T. cristinae, we found significantly lower local abundances in host-plant patches harbouring populations with high proportions of the poorly camouflaged morph (unstriped on Adenostoma and striped on Ceanothus). Also, as it turns out, the mathematical model performed quite well to predict the spatial patterns of local adaptation in Timema, and made some interesting predictions about patterns of patch occupancy and modified evolutionary trajectories for the Timema. These results were corroborated by our manipulative experiments, which showed lower abundances on Adenostoma bushes stocked with the unstriped morph. Furthermore, we found lower abundance and species richness of cohabitating arthropods on those bushes, as well as lower rates of herbivory from sap-feeders. These community-level results suggested that when birds are attracted to and forage on populations of poorly camouflaged Timema, they opportunistically eat or scare away other arthropods as well, reducing herbivory. Our second experiment involved a bird-exclusion treatment, and supported this hypothesis. 

 Graphical summary of the empirical results.

The various authors of our paper differ a bit in their opinions about what is most interesting about the study, but I will point to three main things. Firstly, eco-evolutionary research, and much ecological research in general I venture to claim, tends to use single types of evidence. I think one of the major strengths of our study is that it combines manipulative experiments with field observations and mathematical modelling: our experiments strongly support the notion that camouflage evolution in T. cristinae causes differences in ecological dynamics, and our observational data show that it correlates with naturally observable patterns. Second, evolutionary ecologists have little sense for how important evolution may be in driving ecological patterns relative to traditionally examined factors. In our observational study, we show that poor camouflage accounts for approximately 7% of variation in population size, comparable in magnitude to the effects of host-plant species (5%) and habitat patch size (14%). For T. cristinae populations, evolution is in this sense “important”. Lastly, our study is among few to empirically examine the ecological effects of evolutionary processes other than natural selection, since we also look at gene flow and founder effects. The consequence is that we have evidence in our system for ongoing effects of rapid evolutionary processes on ecology: year after year, natural selection rapidly increases local adaptation, and gene flow breaks it down, allowing eco-evolutionary effects to persist through time.

To bring it all full circle, Andrew was asked to write a news dispatch by Current Biology (Hendry 2013) to highlight the work he set in motion over three years ago. Thanks for everything Andrew!

Illustration credit: Rosa Marin Ribas


Bolnick DI, Nosil P (2007) Natural selection in population subject to migration load. Evolution 61:2229–2243.
Farkas TE, Mononen T, Comeault AA, Hanski I, & Nosil P. (2013) Evolution of camouflage drives rapid ecological change in an insect community. Curr Biol 23:1835–1843.
Hanski I, Mononen T, Ovaskainen O (2011) Eco‐evolutionary metapopulation dynamics and the spatial scale of adaptation. Am Nat 177:29–43.
Hendry A (2013) Eco-evolutionary dynamics: Community consequences of (mal)adaptation. Curr Biol 23:R869–R871.
Nosil P, Crespi BJ (2006) Experimental evidence that predation promotes divergence in adaptive radiation. Proc Natl Acad Sci USA 103:9090–9095.