Sunday, November 29, 2015

Multiple similar populations produced by ecological vicariance, not parallel evolution

[ This post is by Daniel Berner, I am just putting it up. –B. ]

A few years back, Marius Roesti and I started to work extensively on the genomics of adaptive divergence between lake and stream stickleback population pairs from Canada, using genome-wide marker data sets generated by restriction site-associated DNA sequencing (RADseq). Based on this first genomics experience (see http://onlinelibrary.wiley.com/doi/10.1111/j.1365-294X.2012.05509.x/abstract), we agreed that there were two main aspects we wanted to improve in subsequent population genomic investigations. First, we felt that higher marker resolution was needed, because our initial RADseq resolution (based on the standard Sbf1 restriction enzyme) seemed to capture the molecular consequences of divergent selection only on a relatively crude scale. Second, we imagined that insights into parallelism in the genomic basis of adaptive divergence would be easier to obtain by investigating a study system exhibiting parallel evolution at a smaller and thus more clear-cut geographic scale. With these ideas in mind, we decided to start an investigation on lake and stream stickleback within a single watershed, the Lake Constance basin in Central Europe (Fig. A), using higher-resolution RAD methodology. We considered three populations from well-separated inlet creeks to Lake Constance (one of Europe's largest lakes), as well as the lake population itself sampled at two distant sites. The latter proved to be panmictic, so in the end we believed we were dealing with three stream populations, each diverged independently and in parallel from a shared lake ancestor.


Fig. A. Stream stickleback from the Lake Constance basin in their natural habitat. Photo credit Marius Roesti.

Based on this perspective, my group and I started to do population genomic analyses, but somehow the results did not seem to make sense and came with many surprises. For instance, we observed that genetic variation was lower in the lake than in the stream populations, despite the huge number of stickleback that must be living in the large lake.  Also, the highest genome-wide differentiation emerged from a lake-stream contrast and not from a comparison of the geographically isolated streams. This was unexpected because the independent colonization of the streams by founders from the lake should have promoted differentiation among the stream populations at neutral markers. Moreover, and in a phylogenetic tree, the lake population was nested within the stream samples. Finally, inspecting genetic linkage on a genome-wide scale and haplotype structure around single genomic loci under selection revealed that the lake population has been influenced by selection more severely than the stream populations. We thus ended up with an evolutionary scenario we had completely overlooked in the beginning: the lake population must have adapted to its environment after the stream populations formed, and variation among the stream populations in the magnitude of divergence from the lake population primarily reflects to what extent genetic material from the lake population manages to introgress into the streams. We feel this scenario is well captured by the idea of ‘ecological vicariance’, that is, the ecological (as opposed to purely geographical) fragmentation of an initially widespread population (Fig. B).


Fig. B. Ecological vicariance leading to apparent parallel evolution. This process is initiated by multiple habitats becoming colonized by a shared ancestor (in our case a stream-adapted population) (top panel). Next, the connectivity among populations becomes constrained as the core population adapts to its ecologically distinct habitat (in our case the lake; the peripheral circles are stream habitats) (middle). Nevertheless, this ecologically-based reproductive isolation is not complete, allowing for introgression across habitat boundaries (bottom). Depending on asymmetries in population sizes, this introgression might primarily affect the peripheral populations. In our case, the result is variation among multiple stream populations in the magnitude of erosion of the ancestral state (shown by gray shades), mimicking variable progress in parallel evolution among the stream populations.

Hence, the Lake Constance system is appropriate for investigating divergent selection, but inappropriate for studying parallel evolution, because the stream fish (initially considered derived) reflect, to a greater or lesser extent, an ancestral state pre-dating the emergence of the derived lake population. What we learned from this work is that caution is warranted when developing evolutionary narratives in genomics; assumptions should be tested, requiring the combination of extensive analyses including those of haplotype structure around selected loci. If you are interested (there is additional stuff on adaptive chromosomal inversions), check out The genomics of ecological vicariance in threespine stickleback fish.

References

Roesti M, Kueng B, Moser D, Berner D (2015) The genomics of ecological vicariance in threespine stickleback fish. Nature Communications, DOI: 10.1038/ncomms9767.

Roesti M, Hendry AP, Salzburger W, Berner D (2012) Genome divergence during evolutionary diversification as revealed in replicate lake-stream stickleback population pairs. Molecular Ecology, 21: 2852-2862. DOI: 10.1111/j.1365-294X.2012.05509.x.

Sunday, November 22, 2015

Sequential Divergence Across Trophic Levels

[ This post is by Glen Hood; I am just putting it up.  –B. ]

In the fall of 2008, I was wrapping up my research as a M.Sc. student in the Population and Conservation program at Texas State University. I had spent several months narrowing down potential Ph.D. advisors. I was looking for a lab that studied plant-insect interactions and insect speciation. After sending out scores of emails, I had narrowed my search down to five prospective labs.

In November of that same year, I was presenting my research at the Entomological Society of America’s annual meeting in Reno, Nevada. Luckily, four of my five potential Ph.D. advisors were also attending the meeting. As a self-promotional tool, I sent out emails inviting the four potential advisors to attend my talk. My goal was simple—give a talk so great that I would be offered a Ph.D. position on the spot. Unfortunately, I was scheduled to present at 8:15 am on the final day of the conference, which practically ensured low attendance in general and no-shows from all four candidates. In my experience, the night before the last day of a conference is the most social and goes well into the night. My talk about differences in body size and fecundity between the alternating generations of cyclically parthenogenic gall wasps was surely not going to get tired bodies out of bed.

As 8:00 am approached the morning of my talk, a few people filed into the room (i.e., my current advisor, the moderator, and a few early-morning speakers). A few minutes before 8:00, the moderator informed me that the first talk had been cancelled. Great, I thought, even the 8:00 am speaker couldn’t show up for his own talk! I walked out into the hall realizing my plan to make a positive impression on potential advisors had failed. Then I heard someone say, “Glen Hood?” I looked up to see Jeff Feder, one of the potential Ph.D. advisors, surrounded by two oversized luggage bags and one crammed backpack. Of the four labs I had contacted, this was the one I was most interested in, as Jeff studied ecological speciation in Rhagoletis fruit flies. He explained that he was currently on sabbatical in Germany, was briefly in town to give an invited talk, and had to catch a flight in less than two hours. Despite his busy schedule, Jeff said he wanted to make sure we were able to chat before he left. To this day, I still do not know if Jeff was able to catch his originally scheduled flight to Germany. However, in the 10 minutes before I was scheduled to present, Jeff introduced me to the topic that was to consume the next six years of my life.

When I joined Jeff’s lab in the summer of 2009, the groundwork was already laid for what would result in a paper we recently published in PNAS (Hood et al. 2015). That same year, Andrew Forbes, a former graduate student in the Feder lab (now in the faculty at the University of Iowa), was wrapping up his dissertation research. The central theme of Andrew’s research was simple: a major cause of biodiversity may be biodiversity itself. The process, referred to as “sequential” or “cascading” speciation or divergence, has been proposed to help explain a number of diverse patterns including radiations following mass extinctions, and species diversity in the tropics. However, sequential speciation could perhaps be most important for understanding the incredible diversity of plant feeding insects and their parasitoids. The idea is that when plant-feeding insects diversify by adapting to new host plants, they create a new habitat for their insect parasites (parasitoids) to exploit and adapt to. If a parasitoid shifts to the new habitat, it can encounter the selection pressures as its insect host, which could result in the parallel divergence of insect host and parasitoid. However, there were relatively few empirical examples of sequential speciation within insect communities (Stireman et al. 2006, Abrahamson & Blair 2008, Feder & Forbes 2010). The major issue is that analyses of sequential speciation across trophic levels can be complicated by a lack of information about the natural history and geographic context of host shifting. What Andrew and Jeff needed to test the sequential speciation hypothesis was a well-defined system with a well-resolved natural history to directly test whether ecological adaptation can sequentially amplify diversity.

It just so happens that Jeff had spent his entire career working in the perfect system to address these issues. Fruit flies in the Rhagoletis pomonella species complex are a model for ecological speciation via host-plant shifting. In particular, the recent host shift of the apple maggot fly, R. pomonella, from its ancestral host plant hawthorn to introduced, domesticated apples in the last ~160 years, is an example of incipient speciation (i.e., host race formation) in action. To test for sequential divergence, Andrew and Jeff used the Rhagoletis-specific parasitoid wasp, Diachasma alloeum that lays it eggs into the larvae of the fly. Their study showed that populations of D. alloeum attacking hawthorn and apple host races of R. pomonella as well as sister species R. mendax (host: blueberry) and R. zephyria (host: snowberry) had indeed formed genetically distinct host races as a result of specializing on diversifying fly hosts. In addition, the same ecological traits that differentially adapt R. pomonella to their respective host plants and reduce gene flow between diverging populations (host-related differences in the timing of adult eclosion, and host fruit odor discrimination behaviors) are the same barriers that reproductively isolate D. alloeum to their respective fly hosts. Andrew, Jeff, and colleagues published the results in a paper in Science (Forbes et al. 2009).


Fig. 1.  (A) a single sequential divergence event and (B) sequential divergence with multiplicative amplification of biodiversity.

My job as a new Feder lab graduate student was to attack the next, obvious question: How common is the sequential speciation phenomenon in a broader context? The plan was simple: follow the road map set by Forbes et al. (2009) in the Science paper to determine if sequential speciation could not just linearly (one fly to one parasitoid), but multiplicatively (one fly to many parasitoids) amplify biodiversity across the entire community of parasitoid wasps attacking R. pomonella group flies. Jeff, Andrew and I formed a team of evolutionary biologists including then Feder lab graduate student Tom Powell (currently a post-doc at University of Florida), Notre Dame research assistant professor Scott Egan (now faculty at Rice University), a graduate student of Forbes’, Gabriela Hamerlinck (currently a post-doc at the University of Wisconsin), and Jim Smith (faculty at Michigan State) to contribute to the cause.


Fig. 2.  The apple maggot fly, Rhagoletis pomonella, on its native host hawthorn, Crataegus mollis.  Photo credit: Hannes Schuler.

We first outlined a series of conditions modified from Dres & Mallet (2002) and Abrahamson & Blair (2008) that must be met to support the sequential divergence hypothesis. In addition these criteria helped guide our experimental approach. These conditions are as follows:

(1) Shift to a new host resource and multiple host associations occur in close geographic proximity
(2) Host-associated populations form distinct genetic clusters (spatially replicable), but experience gene flow at appreciable rates
(3) Females and potentially males display host preferences and discriminate among alternate hosts
(4) Host choice is linked to mate choice facilitating assortative mating resulting in prezygotic habitat isolation
(5) Host selection and fidelity are under some degree of genetic control and not due solely to maternal, learning or environmental effects
(6) Differences in insect phenologies track differences in host phenologies, resulting in temporal isolation
(7) Insect phenology is under some degree of genetic control, not due solely to maternal or environmental effects
(8) Fitness tradeoffs exist between host-associated populations resulting in migrants and hybrids having reduced fitness

To experimentally address if the conditions of sequential divergence were met in the remaining members of the parasitoid wasps community, Diachasmimorpha mellea and Utetes canaliculatus, we first sampled populations attacking multiple Rhagoletis hosts occurring in sympatry (populations attacking apple, hawthorn, flowering dogwood, snowberry, blueberry and black cherry flies). Given that members of the wasp community are specific to Rhagoletis, this fulfilled condition 1.

First, to test for genetic evidence of sequential divergence we genotyped populations of D. mellea and haplotype C U. canaliculatus for 20 and 21 microsatellite loci respectively. Similar to the pattern documented for D. alloeum by Forbes et al. (2009), both D. mellea and U. canaliculatus showed consistent allele frequency differences between host-associated populations. In addition, in genetic distance networks, populations clustered by host-association, not by geography. This result supports sequential divergence condition 2.


Fig. 3. Host fruit odor discrimination of (A) Diachasma alloeum, (B) Utetes canaliculatus, and (C) Diachasmimorpha mellea. Positive values represent preference for host fruit odor and negative values represent avoidance of host fruit odors in behavioral assays.

In their Science paper, Forbes et al. (2009) concluded that the origin of D. alloeum attacking the apple was not from a host shift from the hawthorn fly but from the blueberry fly. While our results for D. mellea and U. canaliculatus were not as conclusive, our study implies that Rhagoletis and its parasitoids may not always co-speciate in a strict 1:1 follow-the-leader fashion. Wasps attacking different flies in the community appear to be taking advantage of the new niche opportunity provided by Rhagoletis host shifts, not necessarily just the parasitoid infesting ancestral fly hosts. Thus, adaptive starbursts of sequential divergence may be the result of biodiversity radiating from several different origins within the community.

Key features of the fly and wasp life cycles and biology mirror each other, suggesting that the same host-plant related ecological adaptations that reproductively isolate the flies may also isolate wasps. For example, both flies and wasps use the volatiles emitted from the surface locate host plants, and both Rhagoletis and D. alloeum use the fruit as the site for courtship and mating. To test for host plant-related assortative mating caused by habitat isolation for D. mellea and U. canaliculatus, we coupled field observations of mating behavior with tests of host odor discrimination. By making observations of wasps at sympatric sites, we found that, similar to R. pomonella and D. alloeum, both D. mellea and U. canaliculatus mate on or near their host fruit. In addition, in tests of host fruit odor discrimination, wasps prefer the odors emitted from the surface of natal fruit and avoid non-natal odors. We estimated that fruit odor discrimination reproductively isolates D. alloeum, D. mellea and U. canaliculatus attacking different fly hosts by as much as 79%, 88% and 89% respectively, fulfilling sequential divergence conditions 3 and 4.


Fig. 4. The parasitoid wasp, Utetes canaliculatus, searching for its Rhagoletis fly host on a snowberry fruit.  Photo credit: Hannes Schuler.

A common criticism of Forbes’ Science paper was that, unlike the work in Rhagoletis, there was no direct support for a genetic basis for host fruit odor discrimination. To address this issue, we reared D. alloeum originating from blueberry and hawthorn flies in non-natal apple fly and apple host plant environments. We then compared their response to the odors emitted from the surface of their parental host and their novel apple host. As predicted, both hawthorn- and blueberry-origin D. alloeum retained preferences for their respective parental host fruit odors, while avoiding non-natal apple volatiles. While not definitive, the rearing studies support condition 5, a genetic basis for behavioral differences in host fruit odor discrimination during sequential divergence.

The host plants of Rhagoletis fruit at different times of the year. For example, apples ripen 3–4 weeks before native hawthorns in sympatry. Thus, flies must eclose to coincide with the availability of ripe fruit to find mates and oviposition sites. Rhagoletis are univoltine, and live for 1 month as adults. Differences in eclosion timing therefore result in temporal mating isolation. The life cycle of the wasps mirrors that of their fly hosts. Wasps are also univoltine, and live 1–2 weeks. To assess the degree of temporal isolation due to variation in host phenology, we compared the timing of adult eclosion of U. canaliculatus and D. mellea attacking different fly populations in sympatry. We found that eclosion curves differed between sympatric populations of wasps attacking different Rhagoletis, tracking the eclosion times of fly hosts and the fruiting times of their host plants. Coupling the differences in eclosion times with calculations of adult longevity, we estimated that populations of D. alloeum, D. mellea, and U. canaliculatus are temporally reproductively isolated by as much as 75%, 55% and 96% respectively. This supports sequential divergence condition 6.


Fig. 5.  Mean eclosion times averaged across collection sites of adult Rhagoletis attacking different host plants and Utetes canaliculatus, Diachasmimorpha mellea, and Diachasma alloeum attacking each fly host.

To link host-associated genetic differentiation to divergence in life history timing, we tested for associations between microsatellite genotypes and the timing of eclosion for U. canaliculatus and D. mellea. Similar to Rhagoletis and D. alloeum, we found 7 and 12 loci, for D. mellea and U. canaliculatus respectively, that displayed significant genotype or genotype × host effects with eclosion timing. This satisfied condition 7.

Finally, although host-associated fitness tradeoffs have been inferred for several species of Rhagoletis feeding in natal versus non-natal fruit, difficulty in reciprocally transplanting wasps in the lab made it difficult to directly experimentally test condition 8. However, hybrid wasps may display intermediate phenotypes for eclosion phenology and host odor discrimination that suffer reduced fitness for both parental host plant.

In conclusion, we found that sequential divergence can rapidly and multiplicatively amplify biodiversity of entire guilds or communities, as the same host-related ecological adaptations associated with host choice and life history timing cascade from host plant to fly to parasitoid. When combined with the results from Forbes et al. (2009), our study supports seven of the eight conditions we identified as necessary for sequential divergence in D. alloeum, D. mellea, and U. canaliculatus. Our results thus prompt the question: just how taxonomically widespread is sequential speciation and how often does it really contribute to the formation of biodiversity? For organisms such as insects and their parasites that experience and partition resources on a fine scale, the effects of new niche construction may cascade through ecosystems and have an important effect on biodiversity. I hope that our study motivates others to look for patterns of sequential divergence in their own systems.

To this day, I am still not sure what surprises me more – that sequential divergence can multiplicatively amplify biodiversity, or that Jeff risked missing an international flight to talk to a prospective graduate student!

Fig. 6.  Notre Dame graduate student Glen Hood (left; big beard), and professor Jeffrey Feder (right; small beard) rearing Rhagoletis from rotting, infested apples.

References

Abrahamson WG, Blair CP (2008) Sequential radiation through host-race formation: herbivore diversity leads to diversity in natural enemies. Specialization, Speciation, and Radiation: The Evolutionary Biology of Herbivorous Insects, eds Tilmon KJ (University of California Press) pp 188–202.

Drès M, Mallet J (2002) Host races in plant-feeding insects and their importance in sympatric speciation. Philos Trans R Soc Lond B Biol Sci 357:471–492.

Feder JL, Forbes AA (2010) Sequential divergence and the diversity of insects. Ecological Entomology 35:67–76.

Forbes AA, Powell THQ, Stelinski LL, Smith JJ, Feder JL. 2009. Sequential sympatric speciation across trophic levels. Science 323:776–779.

Hood GR, Forbes AA, Powell THQ, Egan SP, Hamerlinck G, Smith JJ, Feder JL. 2015. Sequential divergence and the multiplicative origin of community diversity. PNAS 112:E5980–5989.

Stireman JO, Nason JD, Heard SB, Seehawer JM (2006) Cascading host-associated genetic differentiation in parasitoids of phytophagous insects. Proc Biol Sci 273:523–530.

Wednesday, November 18, 2015

To swim, or not to swim?

[ This post is by Dan Bolnick; I am just putting it up.  –B. ]

To swim, or not to swim?
That was the question
Whether ’twas Nobler for the Fish to suffer
The Slings and Arrows of outrageous Current
Or to take Flight against a Stream of Turbulence,
And by opposing end them: to rest, to swim
No more: and by drifting, I say we end
The energetic expenditure and the thousand Natural selections
That Fish are heir to?  ’Tis a consummation
Desired by some fish.

          – with apologies to William Shakespeare

It was summer 2007 and my graduate student Will Stutz and I had been setting up a large experiment in Blackwater Lake on northern Vancouver Island. For two weeks we (and multiple assistants) had been building large (10 m2) cages – 30 of them, installing them in the littoral zone, and stocking them with stickleback, trout, and sculpin. The goal was to test whether interspecific interactions altered patterns of individual and between-individual niche breadth in stickleback (answer: yes; Bolnick et al 2010). We had wrapped up the construction, and found ourselves with a rare opportunity: a few hours of uncommitted time. Living in Austin, but having a field site on Vancouver Island, I tend to push my crew pretty hard to exploit our limited field time to the fullest, which means little recreation.

Do we head back to the cabin, or go start sampling elsewhere for our secondary project? Or, go exploring? To paddle, or not to paddle? On the spur of the moment we pushed off in my lab’s bulky red canoe, and set a course for the far end of the lake, where I had never been. Blackwater Lake is long and thin, and we had been doing experiments in its north end for a couple of years. The north end is easily accessed from a logging road, and the littoral zone is mercifully free from big underwater tree trunks, the bane of cage-builders like myself. The paddle was pleasant – the weather was perfect, cool and sunny. We passed a pair of loons, but no people; I’ve only once spotted a stranger on this lake (more on him, later).

Will and I pulled up at the far end of the lake, maybe half an hour later. The south end of Blackwater is beautiful and undisturbed, so we climbed out onto a marshy spit of land (a floating mat of vegetation, really) to look around. We found ourselves next to a fast-flowing stream that drained directly into the lake at a fairly high velocity. It was about waist-deep, cold, with a firm sandy substrate and a stiff current that carved a 2-meter-wide channel into the marshy banks covered with muskeg (Fig. 1). We explored upstream a little; it was a nice uniform channel until we hit a beaver dam at 70 meters upstream, above which the ground became still wetter and the stream was slow but still visibly flowing. There were stickleback everywhere in the stream, in both the fast channel and above the beaver dam (See videos of stickleback swimming in a moderate-current stream here: https://youtu.be/2wttxQGhuv4 and here https://youtu.be/nHZh1V7xCSc).


Figure 1. Blackwater inlet stream. Note the pale sandy substrate.

We stood around, enjoying the sunlight, and speculated about this remarkably abrupt lake-stream interface. Until then, I had focused exclusively on lake stickleback, partly out of habit and partly to avoid stepping on the toes of Andrew Hendry, who I knew was studying lake–stream stickleback in this area. Andrew and I had talked about lake–stream stickleback a bit the year before when his crew crashed on my cabin’s floor for a week, providing a bottle of excellent Scotch in payment. I recalled expressing some skeptical curiosity about one of his results, in which he described a very abrupt transition (across tens of meters) in fish morphology when moving from an inlet stream into a lake. I thought that the spatial scale in question was bizarrely small – surely migration from the lake into the stream, and vice versa, should erode divergence at that scale. Unless…  unless the fish themselves avoided switching habitats. Andrew and I had talked about the possibility of divergent habitat preferences accentuating the lake–stream divergence, but he seemed relatively uninterested. So perhaps that was an opening for me to foray into lake–stream stickleback without stepping on toes.

Recalling that conversation, with the Blackwater inlet at my feet, I took stock of the scene: I could see stickleback swimming in the still shallows of the lake, on my right, a few meters away. A few meters to my left, I could see stickleback swimming in a rapid current (Fig. 2).  Did these fish care about that difference in water flow, and if so would they sort themselves into their preferred habitat?  If there was habitat preference, perhaps we could see phenotypic, maybe even genetic, divergence arise not by mortality or differences in fecundity, but by non-random spatial sorting of individuals. Could adaptation proceed by choice, rather than by force?


Figure 2. Blackwater inlet stream (foreground) and the lake (on the far side of the muddy flat peninsula. Fish caught in the foreground are significantly different, morphologically, from the fish on the far side just 5 meters away. At the top left end of the stream, just before it enters the lake, you can see ripples due to a strong current, and you can see eddies in the stream in the mid-ground if you look closely.

Soon after, I paddled back to the inlet with Will Stutz, On Lee Lau, Travis Ingram, and Lisa Snowberg. We trapped in the lake and stream, marked fish with elastomer dye, and released them right where the stream entered the lake (Fig. 3). More exactly, we released them just inside the stream channel so that they had an equal arc of up- and down-stream options. We had fun sitting in the sun marking fish and talking, until we had our first and only visitor on the lake: a guy in his 70’s paddling an open top kayak, with a huge white beard but otherwise buck naked. He pulled up to the shore where we were marking fish, got out of his boat and asked, with genuine curiosity, about what we were up to.  He was very chatty, and stayed talking with us for quite some time. He offered to help, while some of my students squirmed.


Figure 3. (Figure 1 from Bolnick et al 2009 Evolution) showing the mark-displace-recapture experiment layout.

Returning four days later, we retrapped.  Now, I expected to find that lake-native fish went downstream to the lake, and stream-native fish went back up into the stream. But I didn’t expect this: 90% of the fish had returned ‘home’ in only 4 days (Fig. 4). As an aside: would all of them go home eventually? How long does this take? Is this a breeding-season phenomenon, or year-round? This habitat fidelity was observed for both the lake fish, who drifted downstream 1 meter into the lake, and also for the stream fish who disproportionately worked their way up-river (as far as 150 meters, many crossing up past the decrepit beaver dam). We were thrilled, and had the stats and figures and half the paper planned out before the end of the day (Bolnick et al 2009). But the coolest bit had to wait months, until an undergraduate (Claire Patenia) measured morphology of our recaptures. We realized that the 10% who switched habitats were morphologically predisposed to do so: lake-like stream natives, and stream-like lake natives. And, the transition from lake- to stream-phenotypes is so abrupt that one can move a mere 5 meters from the lake upstream and get significantly different phenotypes and microsatellite allele frequencies (Fig. 5). This is 1/8th the median distance that our released fish swam in just four days.


Figure 4. (Figure 2 from Bolnick et al 2009) showing the relative frequencies with which lake and stream natives were recaptured in lake or stream habitats, four days after release.

Since that initial study, I’ve become a bit obsessed with the notion of genotype- and phenotype-dependent dispersal, and the potential for biased gene flow to actually aid rather than inhibit adaptive divergence on small spatial scales (Edelaar and Bolnick 2012, Bolnick and Otto 2013, Richardson et al 2014). The vast majority of population genetic theory presumes that gene flow picks a random sample of alleles from one population, and drops them into a recipient population. Sure, there may be stochastic drift during that sampling process, but in this model gene flow is a homogenizing force, not an adaptive one. That changes when certain alleles are more or less likely to move. Depending on the nature of this bias, non-random dispersal can accelerate and exaggerate adaptation, or it can drive maladaptation (Bolnick and Otto, 2013).



Figure 5.  (Figure 1 from Bolnick & Otto 2012) showing phenotypic divergence between Blackwater Lake and inlet stream stickleback over a small spatial scale.

There’s been something nagging me for a while, however. Although it was satisfying to find evidence for non-random movement of lake and stream stickleback, I still didn’t know why that happened. Something to do with body shape, it seemed, but what? Over a few months, my graduate student Yuexin (‘Kelsey’) Jiang and I developed a hypothesis that stream fish might exhibit ‘positive rheotaxis’ – a tendency to swim up a current, whereas lake fish might exhibit ‘negative rheotaxis’ – swimming down-current.  Where a stream flowed into a lake, this divergent rheotactic response would tend to sort stream fish into the stream, and lake fish into the lake.

To test this idea, Kelsey designed a very nice circular flow tank that allowed stickleback unlimited opportunities to move up- or down-current. She then tested for rheotactic response of lake and stream stickleback (from our original Blackwater Lake study site). For a great video of lake and stream fish side-by-side, see https://youtu.be/7Uwq_sLu4Jk (challenge: guess which is the lake and which is the stream fish group). This video also shows the layout of the circular flow tank. In a paper that just came out recently (Jiang, Peichel, and Bolnick 2015 Evolution), Kelsey’s results provided a nice mix of expected and surprising insights.


Figure 6. Net displacement was the net movement of individual fish up (+) or down (-) current in the flow tank (see https://youtu.be/7Uwq_sLu4Jk for a video example). This was measured for wild-caught lake and stream fish from both the inlet and outlet of Blackwater lake. Lab-reared fish are not shown here, but are discussed in the paper.

First, the expected:  inlet stream fish did a better job of staying put in current. They were displaced an average of 5 m downstream in the experiment, compared to 12–17 meters for lake fish and outlet stream fish (Fig. 6).  Strictly speaking, the inlet stream fish showed less negative rheotaxis than the others. But, this should help them remain stationary despite flowing water, keeping them in their stream habitat. Note that in the absence of a current, there were no differences between ecotypes and no net displacement.

We also expected that not all stream fish are equivalent: inlet stream fish need positive rheotaxis to remain in their home habitat; outlet stream fish need to stay in place or go down-current to remain in their stream. So, the ‘positive rheotaxis’ of stream fish is only true for inlet streams. Which makes sense: in an outlet stream, positive rheotaxis brings you into the lake.


Figure 7 Cumulative displacement was measured as the total upstream path length that fish swam. Although lake (and outlet stream) fish were displaced downstream more than inlet stream fish,  these groups actually tried the hardest to swim up-stream. However, they alternately swam up, then were blown down-stream, then swam up again, leading to a much larger cumulative effort, even though they had less to show for it at the end of the experiment. The stream fish, on the other hand, stayed in place (Fig. 6) with less effort.

 A few surprising insights showed up as well. First of all, we realized that rheotaxis is not a simple trait to interpret. Stream fish actually did a very good job holding their position in current, but they did so with relatively little effort (Fig. 7) and by seeking out low-flow ‘boundary’ areas near the inner part of the circular tank. So they actually avoided current, to stay in place. Lake fish, on the other hand, weren’t very strategic and used the middle of the current. They were tumbled ‘down-stream’, only to swim most of the way back up again. Over the duration of a given trial, lake fish lost a lot of ground down-stream (so did the stream fish, actually, Fig. 6). But the lake fish expended vastly more energy to not-quite hold their place. And they did end up farther down-stream.

Although the rheotactic response of stream stickleback was very striking, we couldn’t recreate it in two subsequent studies. In lab-reared common garden stickleback, everyone showed fairly positive rheotaxis no matter where their parents came from. So either the lake-stream difference in rheotaxis isn’t heritable, or it is heritable but requires prior experience to fully develop.  In wild-caught non-breeding stickleback, the rheotactic difference was also absent.

Our conclusion then is quite intriguing: non-breeding fish, and inlet stream fish, show positive rheotaxis. Breeding lake fish do not. So is the lake condition the derived/induced trait? Perhaps: stickleback are ancestrally anadromous after all, swimming from the ocean upstream into estuaries and rivers and lakes to breed. Perhaps that same instinct persists, but must be actively suppressed in lake fish to prevent their dispersal into the stream?

To really answer this speculative question, we will need to replicate this result more extensively with both marine and multiple lake and stream stickleback, and begin genetic mapping of rheotactic response to understand the mechanisms and polarity of either gain or loss of this swimming behavior. We will be aided in that endeavor by the fact that we have recently found some of the phenotypic traits that contribute to rheotactic response and are heritable. But I can’t tell you more about that yet, as it isn’t published. In fact, I should probably go back to working on finalizing that manuscript right now. But before I sign off, I want to reiterate a few key lessons.

First and foremost: gene flow is an evolutionary effect of individuals’ movement across a landscape (and their ability to survive that movement), and both behavior and morphology can influence that movement. As a result, different phenotypes (or genotypes) may move non-randomly through space. Local adaptation can therefore result from individuals’ movement behavior, rather than differential survival or reproduction (Bolnick and Otto 2013). This poses all sorts of curious puzzles: what are the traits and loci driving non-random movement? What does this directed movement do to eco-evolutionary feedbacks, for instance as predators and prey interact not just through attack rate but also movement? What is the genetic signature of adaptation-via-dispersal? Does this generate peaks of high FST that we might interpret as an effect of natural selection?

Second: read theory, understand theory, and don’t be intimidated by theory, but at the end of the day don’t let theory box in your thinking. Population genetics from the new synthesis onward would have me believe that adaptive divergence isn’t possible within a dispersal neighborhood. But it is (Richardson et al 2014).

Lastly: paddle. When in the field, it is important to take some time to let your mind drift, and to explore your surroundings. As a discipline, we are inspired by biological diversity, but too often we enter the field intensely focused on our own planned project, our hypothesis testing. Many of my own areas of research emerged not from a priori planning, but from these all-too-infrequent mental breaks in which I happen to notice something that catches my interest. So, if you find yourself wondering, “to swim, or not to swim”, or some variant thereon, you’d best put on your mask and snorkel. That’s where the biology is. In particular, that’s where you’ll find something that you didn’t expect, that theory hasn’t trained you to see.

References

Bolnick, D.I., T. Ingram, L.K. Snowberg, W.E. Stutz, O.L. Lau, and J.S. Paull. 2010 Ecological release from interspecific competition leads to decoupled changes in population and individual niche width. Proceedings of the Royal Society of London, Ser. B. 277: 1789–1797.  doi: 10.1098/rspb.2010.0018. PMCID: 20164100

Bolnick, D.I. and S. Otto. 2013. The magnitude of local adaptation under genotype-dependent dispersal. Ecology and Evolution 3:4733-4735. doi: 10.1002/ece3.850. PMID: 24363900

Bolnick,D.I. L. Snowberg, C. Patenia, O. L. Lau, W. E. Stutz, and T. Ingram. 2009. Phenotype-dependent native habitat preference facilitates divergence between parapatric lake and stream stickleback. Evolution 63:2004-2016   doi: 10.1111/j.1558-5646.2009.00699.x.   PMID: 19473386.

Edelaar, P. and D.I. Bolnick. 2012. Non-random gene flow: an underappreciated force in evolution and ecology. Trends in Evolution and Ecology 27: 659-665. doi: 10.1016/j.tree.2012.07.009. PMID: 22884295

Jiang, Y., L. Torrance, C.L. Peichel, and D.I. Bolnick. 2015. Differences in rheotactic responses contribute to divergent habitat use between parapatric lake and stream threespine stickleback. Evolution 69: 2517-2524. doi: 10.1111/evo.12740.

Richardson, J.L., M.C. Urban, D.I. Bolnick, and D.K. Skelly. 2014. Microgeographic adaptation and the spatial scale of evolution. Trends in Ecology and Evolution 29: 165-176. doi: 10.1016/j.tree.2014.01.002.  PMID: 24560373

Sunday, November 15, 2015

Playing with hormones to understand parasite resistance

As my PhD work on the evolution of defense against parasites was coming to an end, I realized that that I had no intention of wrapping up my lab work. On the one hand, there were too many questions I wanted to continue exploring; on the other hand, gathering data in the laboratory was always an enjoyable way to get a day of work started. Also, at this point I was becoming more interested in understanding the mechanisms that drive sex-biased parasitism (see below), so I decided that my last experiment was going to address how hormonal manipulations influence resistance in my study system.

Organisms show considerable variation in their ability to control parasites. Although the causes of such variation tend to be multifactorial, perhaps one of the most prevalent and interesting patterns occurs between the sexes. Given that males and females can show large differences in their morphological, life-history, and behavioural traits, it should not come as a surprise that they can also differ strongly in their ability to reduce or control their parasite loads (i.e. resistance). Indeed, wild populations often show sex-biased parasitism, wherein one sex has higher prevalence and higher average parasite loads. Yet, despite this being a well-known pattern, studies of host–parasite ecology and evolution often ignore sex differences by combining both sexes during analyses. When studies do consider sex differences in parasite loads, they often are unable to identify the mechanisms behind sex-biased parasitism.

There are multiple candidates for the mechanisms driving sex-biased parasitism. One oft-suggested (but rarely tested) explanation for sex differences in parasitism is differential exposure to parasites caused by sex differences in habitat use. Although conceptually reasonable, the generality of this mechanism is undermined by the fact that sex-biased parasitism is often observed even under common-garden rearing in the laboratory. Two other alternative hypotheses suggest that differences in size (which lead to variation in available resources for parasites) or differences in gonadal steroids (i.e., androgens – which can impair immune function) are behind sex differences in parasitism.

In a recently published paper (Dargent et al. 2015) we tested whether androgens influence resistance to infections with the ectoparasite Gyrodactylus turnbulli in the guppy (Poecilia reticulata). Although variation in guppy resistance is explained by various components of their ecology and evolution (e.g., see here or here), this fish also shows sex-biased parasitism in the field and high variation in male signalling traits (which in turn suggests variation in androgen levels), thus pointing to a possible, and previously unexplored, contribution of androgens to guppy resistance. To test this, we experimentally manipulated guppy response to endogenous circulating gonadal steroids. One common problem with experiments that test the role of androgens, in particular testosterone, on host traits is that they increase the concentration of circulating androgens to unrealistic levels – sometimes orders of magnitude above reported field values. To work around this problem, we demasculinised male guppies (with an androgen blocker), to reduce the effect of endogenous androgens while maintaining natural levels of the hormones themselves. Given that a few studies suggest that oestrogen may up-regulate immunity, we also tested for the effect of female gonadal steroids by feminizing males (with a combination of an androgen blocker and exogenous oestrogen).


Photo 1: Monitoring guppy health. Each guppy was individually housed in a 1.8 liter container.

After three weeks of hormone treatments, delivered through the fish food, we infected each guppy with two G. turnbulli and monitored parasite numbers and fish survival for the following ten days, while continuing to apply the hormone treatments. We found that demasculinisation not only decreased parasite abundance but also decreased infection-induced mortality, which suggests that androgens play an immunosuppressive role and may negatively affect guppy tolerance to infection (the ability to reduce the detrimental effects of a given parasite load). Furthermore, we detected no additional role of oestrogens in explaining parasite load (i.e., parasite loads on feminised guppies did not differ from those on demasculinised guppies), nor did we detect an effect of host size on G. turnbulli abundance. These findings suggest that variation in androgen levels is likely an important driver of guppy resistance, and that androgens may mediate fitness trade-offs between male expression of sexual signals and resistance to disease.


Photo 2: A male on the day treatments with the androgen blocker began.

Reference

Dargent, F., Reddon, A. R., Swaney, W. T., Fussmann, G. F., Reader, S. M., Scott, M. E. & Forbes, M. R. 2015. Demasculinization of male guppies increases resistance to a common and harmful ectoparasite. Parasitology 142: 1647–1655.  DOI: 10.1017/S0031182015001286

Sunday, November 8, 2015

How To Teach

I have been marching through “How To” posts in a somewhat sequential (by career state) order, and I intended to next write a third one on “How To Get a Faculty Position.” However, events have conspired to cause me to write an out-of-order post that I had been planning to get to eventually.

This year, at the request of the Awards Committee in our Biology Department, I applied for McGill University’s Principal’s Prize for Excellence in Teaching. Having been chair of the awards committee in the past, I knew how hard it was to convince people to apply for awards, and so I agreed to do so even though I had little expectation of success. I prepared an extensive dossier, collated quantitative and qualitative student comments and ratings, requested letters from past students, and so on. Then, some six months ago, I sent the application off to the university and promptly forgot about it. Amazingly, just a month or so ago, I started to get emails congratulating me on receiving the award. I was quite surprised to have gotten the award as I know that many other professors at McGill are excellent teachers.

Then came invitations from a number of offices at McGill to attend the Fall graduation ceremonies at which the award would be given. As I am on sabbatical in California, however, I had to decline. Just this last Friday, however, I got another request from The McGill Reporter to provide information for an article about the award winners. The request was for me to answer by email a series of questions including the following:

How would you describe your teaching style?
What are your strengths as a teacher? Weaknesses?
What are the biggest challenges of teaching? And the greatest rewards?
What are the most important qualities a teacher should have?
Any advice for students thinking about becoming teachers?

Rather than doing a quick and inefficient job of answering the questions, it made more sense to simply write the post about “How To Teach” that I had intended to do anyway. So here we go – a bit out of order in the “How To” series but suitably timed in other respects (the award is to be presented tomorrow, Nov. 10). For the below points, I have borrowed heavily from the teaching dossier that I originally submitted as my application for the award, which is why it is even more self-referential than previous post.


Inspiration-based teaching

My approach to teaching is centered on the integration of two goals: to challenge students and to inspire their interest in the subject. Perhaps we could call this “inspiration-based teaching” so as to starkly contrast it with classical “information-based” teaching. Information-based teaching fails because most students quickly forget detailed information, no matter how good the teacher. Little long term value thus results from cramming students full of facts, formulae, and figures. This information is, after all, freely available online all the time. (Of course, CONCEPTS can be critical.) It is much more important to challenge and inspire students with an engaging and inspiring narrative. I implement this inspiration-based teaching model through several key elements.

Hands-on learning even in huge classes. Instead of using only images and text on slides to illustrate concepts, I bring the REAL THINGS to every class no matter its size – the examples below come mainly from Introductory Biology, which can have over 600 students. Every lecture is accompanied by multiple physical objects. In one lecture, I brought in a gorilla and human skeleton – no small feat as the former is very heavy. In another, I brought in a measuring tape and spread it out across the entire width of the room to illustrate in a concrete way just how big an 18 m whale shark really is. I even brought in two live snakes, which the students loved (the image shows students clustered at the front of the room after class to see the snakes). In student comments, appreciation of these hands-on opportunities was repeatedly expressed, and here are some edited examples.

Showing off a snake - many students had never been near one before.
I really enjoy seeing some tangible objects, ie skulls of polar bears, snakes, etc.
In the first lecture he actually pulled out a tape measure and showed us how long a shark was, which was really interesting! Also he has this stuffed animal that's a fish, and he used it to explain the different parts of the fish which was really cool.
In addition, in today's lecture he brought out two skeletons of different species of primates. The Gorilla, and the homo sapien. We were able to see first hand the structural differences between the two, and how they affected the way of life of both species.
I loved that he always had something to add to the lecture. He brought in snakes for the reptile section, a fake penguin for the birds lecture and did the drunk dance for the evolution section. Terrific!
love the in-class demos
I love how he bring in objects to show the class (eg. Platypus, gorilla skeleton.
I really appreciated all the material he brought to the lectures.
i enjoyed how he brought in other props for his lectures.

Teaching as performance. I try to make my lectures into performances, perhaps a legacy of my interest in performing arts as a high school student. As the most overt and concrete example, I use a tongue-in-cheek “interpretive dance” to weave all of the evolutionary mechanisms into a single metaphor based on the “Drunkard’s Walk.” If only a single item could be used to encapsulate my teaching, this video would be it. I did the performance in multiple classes for more than 10 years before finally videotaping it and posting it on youtube November last year (2014). In less than a year, it has been viewed 1653 times and a number of other profs had said they either show the video in class or do their own version (of course, the latter is much more fun for students to watch). Numerous students have said to me years later that this was the most memorable lecture of their entire university experience.


 Integrate with research. It is sometimes argued that, at universities, teaching and research are two solitudes, or simply that time invested in one trades-off with  time that can be invested in the other. To some extent these sentiments are true. However, at the least, integrating your own research into lectures helps students see how the material you are discussing is directly relevant to current research at the university. And, of course, lecturers tend to be more passionate and excited when talking about their own research, which the students can really tell. In addition to this approach, I suggest two other routes. First, integrate the published research of other undergraduates at your university into your lectures. I think undergrads are inspired toward both teaching and research when they see the research that has been conducted by the undergrads who came before them. That is, people sitting in the very room where they are now sitting conducted research that is now being used to educate students only a few years later. Second, try when possible to incorporate a research project into the class and, when feasible, encourage students to work toward publishing the projects. For instance, I teach an upper level undergrad-graduate class that has, as one its core elements, the preparation of a meta-analysis (or other project) with an eye toward publication. I have taught the course five times and students (including some undergrads) have thus far (Nov. 8, 2015) published a total of 11 papers that have been cited a total of 693 times. Perhaps it is a bit absurd for a class to have an H-index but so far it would be 9. The point isn’t that this is a great class, merely that teaching and research can work together to mutual gain.


Engage social mediaThe current generation of university students gets much of its information through social media and professors would do well to embrace this technology. Before and after lectures, I tweet relevant topics (#biol111: see examples above) for students, and at least 50 students in the class were immediate “followers.” I have also generated educational videos, including The Adaptive Radiation of Darwin’s Finches (6,793 views this year) and Wooly Bear Caterpillar Cocoon Time Lapse (10,306 views this year). In comments, students repeatedly pointed to these activities as enhancing their appreciation of the course and their “connectivity” with the professor. I have yet to make any social media activities compulsory or testable, and instead use them to connect with the students who are really genuinely interested in the topic.

A few #biol11 tweets.


Have fun

Those are the main elements that I use to form a reasonably successful and useful teaching approach, and I think they will work well in most disciplines and at most universities. However, I don’t assert that this is the only good way to teach, or that even that it is the best way to teach. I am merely outlining one way to teach that students do respond well to in many instances. In reality, I think the key is to be excited and passionate about what you teach, regardless of the specifics of how you teach. Students can tell when you think what you are discussing is cool and they respond well to it. Thus, I guess the key bit of advice that is universally helpful might be, at its essence: HAVE FUN!

----------------------------------------

Previous "How to" posts

Monday, November 2, 2015

Constraints to adaptation: The “oily guppies” of southern Trinidad

[ This post is from Gregor Rolshausen, I am just putting it up.  –B. ]

Arguably, among the most interesting shifts in evolutionary thinking in the past decades is an increased recognition of the limits on rates, directions, and outcomes of evolution. For instance, classic expectations of the adaptive process, such as trait divergence between environments and higher fitness of local vs. non-local individuals, are often not met in natural populations. Effective local adaptation might be hampered by migration and maladaptive gene flow that drags population away from local fitness peaks. Furthermore, instances of (local) maladaptation appear to be particularly pronounced in the context of abrupt and extensive anthropogenic disturbance such as climate change, habitat fragmentation, or pollution. An understanding of the constraints to evolutionary dynamics on the population level therefore needs to become an important cornerstone of ongoing conservation efforts and biodiversity management plans. I here review a recent empirical study from the Hendry lab that investigates constraints to local adaptation in natural fish populations and reveals some interesting contradictions to classical expectations of local adaptation: the “oily guppies” (Poecilia reticulata) of southern Trinidad – a system in which severe crude-oil pollution has a strong impact on the condition of seemingly locally adapted populations.

I first learned about the “oily guppies” at a conference in Berlin in 2010 where Chris Harrod (currently at Universidad de Antofagasta, Chile) presented some interesting results he obtained from guppies that live in oil-polluted streams in the far south of Trinidad. Chris had looked at isotopic compositions of guppies from strongly polluted streams and found that their stable isotope tissue signatures indicated that these fish not only passively tolerate the crude oil, but also seem to actively incorporate it into their metabolism. This was clearly an indication that these fish do not just occasionally “pass through” polluted areas, but dwell in these habitats permanently and, hence, probably show signs of local adaptation to such a severe stressor. Local adaptation to oil pollution – this sounded like a promising research project, not only to me, but also to Andrew, who introduced me to Chris in the first place, and who became my postdoctoral supervisor for the upcoming years of burdensome fieldwork deep down in southern Trinidad.

As soon as (evolutionary) biologists hear the words “Trinidad” and “guppies”, their minds usually wander off to topics such as sexual selection, predation, and life-history trade-offs. All embedded in beautiful northern mountain-range rain forest nerved with pristine crystal-clear streams that provide the researcher not only with ample data but also with refreshing, idyllic... [vinyl scratch sound] Sorry to interrupt, but our story takes place in the south of Trinidad, where things are quite different. Apart from being a Caribbean island with beaches, forests, and Carnival, Trinidad also happens to be located in the southeast corner of the Caribbean plate as part of the eastern Venezuelan Basin – one of the largest oil provinces in the world. Consequently, the Republic of Trinidad & Tobago has been involved in the petroleum sector for over one hundred years undertaking considerable exploration activities, both offshore and onshore (Fig. 1).


Fig. 1.  Impressions from southern Trinidad. (1) Onshore oil rigs and hammer pumps. (2) A tour guide demonstrating local geology. (3) High pressure pump systems at Morne L'Enfer forest reserve. (4) Pipeline running through the forest. 

Particularly interesting – at least from an ecotox / microevolutionary perspective – are inland drillings that take part in remote forest regions, such as the Morne L'Enfer Forest reserve, where companies aim to mine late Pliocene / early Pleistocene source rock formations and near-surface reservoirs. Obviously, these activities entail major surface leakage and sediment contamination, leaving many tributaries of the region's main rivers heavily polluted with crude oil. I am talking about levels of contamination that are not only apparent and widespread (Fig. 2), but that also make the surroundings smell like a ramshackle gas station. It was here where Chris Harrod, together with Dawn Phillip (University of the Western Indies, Trinidad), had first described populations of guppies from oil-polluted streams. And it was here where I started out my field work endeavors to study local adaptation in fish that live in habitats where (almost) everybody else has jumped ship.


Fig. 2.  Pollution impact at our field sites. (1) Morne River – note the oil-soaked black river banks. (2) Field assistant Laura at work – note what the water does to our butterfly nets. (3) Close up of direct leakage into the stream. 

After weeks of on-site exploration and many survey hikes, we settled on two promising field sites: Vance River and Morne River, both running through the Morne L'Enfer reserve, and each one with a polluted and a non-polluted tributary respectively (i.e, four sites in total; Fig. 3). At all four sites we observed plenty of guppies and, opportunely, subsequent population genetic analyses revealed that guppies from different rivers can be considered genetically independent (Fig. 4). Moreover, analyses of morphological divergence among the four study populations revealed that similar habitats harbor similar phenotypes: guppies from independent polluted sites would cluster more closely with each other than with their conspecifics from the clean site in their own river system. This pattern was even more pronounced when we focused the analyses on the cranial region (Fig. 5).

Fig. 3.  Location of our field sites and the two rivers where we conducted the transplant experiments: MR = Morne River, VR = Vance River; np = not-polluted, oil = polluted. 
Fig. 4.  Population genetic differentiation among the four studied populations. 
Fig. 5.  Phenotypic divergence between populations. Shown are warp scores for geometric morphometrics of the cranial region. Along with the four southern populations (MR & VR), we include a northern population (Paria LP, Pa). 

Consequently, with confirmed genetic independence and likely parallel phenotypic divergence between similar habitats, everything looked like the perfect set-up for THE standard approach to test for local adaptation in natural populations: a good old reciprocal transplant experiment. Basically, you take individual fish from each of the four sites and transplant them to all other sites (including one “transplant” within the home site), where you then follow them throughout the experiment to see how they are doing – or, more technically, you measure fitness proxies such as growth rate or survival. If populations are locally adapted, you would expect local individuals to show higher fitness outcomes than non-local individuals from other populations. The classic signature in the data would therefore show crossing reaction lines for the direct performance comparison of locals vs. non-locals (Fig. 6).


Fig. 6.  Classic local adaptation: Fitness distribution of populations A (dashed line) and B (solid line) in their respective native habitats vs. an alternative habitat. 

We did not find this signature in our data. Instead, what we found looked more as if guppies from the polluted environments were doing poorly in all environments. If anything, this was local maladaptation (Fig. 7). What was going on? Well, when your results contradict the expectations that originally led you to set up the experiments, that's where the interesting science starts. Of course, there could have been hundreds of things that went wrong during sampling, transportation, and processing of the experimental fish – as so often in any experimental set-up, especially in the field. Yet, our experiments clearly revealed differences and variation among populations and sites; they also revealed detrimental effects of oil pollution, and we had sufficient numbers of individuals and replicates for each treatment group to draw statistically robust conclusions. But what conclusion to draw from this? At first sight, guppies from the polluted habitats did not seem particularly stressed or unhealthy, and their large abundances in those rivers would indeed suggest that they are locally adapted. But no evidence in our data... well, we needed more data.


Fig. 7.  Performance measures from reciprocal transplant experiments. Instead of crossing fitness curves (Fig. 6), we found that guppies from oil-polluted habitats (Origin.oil) do poorly in both conditions (Environment oil vs. Environment not-polluted) when compared to guppies from not-polluted habitats (Origin.np).

Another field season was arranged to repeat the whole transplant design one more time. This time, instead of focusing exclusively on guppies from the described oil-rich region in the south, our second round of experiments included guppies from two pristine northern streams (Aripo & Paria, LP). While it could well be the case that, in general, guppies from the south had a previous history of oil pollution – even the ones from currently non-polluted sites – for sure the guppies from the north were never exposed to any sort of pollution. If there were any carry-over effects or population history interfering with our results, a comparison to pollution-naĂ¯ve fish should at least highlight any local adaptation signal in our data. This was not the case. Instead, interestingly, guppies from the pristine northern environments were even, to some extent, doing better in polluted-water conditions than the true “pollution-natives” from the south, and when moving “oily guppies” (especially the Morne River population) to clean water conditions, they also performed worse (Fig. 8). Where we expected local adaptation to oil-pollution, we found more or less the opposite.



Fig. 8.  Survival under laboratory conditions (oil-polluted vs. clean) of guppies from polluted southern habitats vs. guppies from clean northern habitats.  (Click to view at full size.)

Now, we have to be careful in speculating about adaptation vs. maladaptation here since we have no data to back up whether the performance differences among our populations have a genetic basis. All fish tested were wild-caught, and thus might have been influenced by plasticity and/or maternal effects. But many examples of adaptive plasticity and maternal effects exist, and given the high selection pressure in our system (mortality), it would still be puzzling to find such an extent of detrimental plastic effects. An even bigger study design would be needed to get to the bottom of this. Another possibility that we did not cover with our study design is that there might be genetic adaptation in early life-stages helping the “oily guppies” to make it to maturity in bad shape (but still functioning), but for the adult phenotypes evolution hasn't come up with a good solution. Moreover, founder effects and/or gene flow from adjacent tributaries might play an important role, but given that our assessment of population genetics and gene flow is currently based on a restricted set of markers, additional analyses with broad-scale genomic tools will have to better inform those hypotheses in the future.

It is, however, equally plausible that crude-oil pollution on the scale present in these rivers is simply too much for the fish to cope with and still be strong and healthy enough to pass the rigorous local adaptation tests. After all, the fish we tested the “oily guppies” against were exposed to oil for a few days only, whereas the pollution-natives face a constant contamination load throughout their lives that might well leave them feeble and vulnerable in their adult life. Interestingly, we did not find many fish species other than guppies in the polluted habitats (Rivulus hartii was there, of course!), nor were there many invertebrates to be found. And indeed, many species are known to have difficulties adapting to highly stressful environments, particularly in the context of anthropogenic pollution and contamination events. Thus, the “oily guppies” could well live on the edge of persistence and, since there is almost no competition or predation, they probably still reach high densities despite poor or sluggish adaptation. Once tested and handled in an experimental set up, the additional stress then pushes them over the aforesaid edge. In this scenario, persistence would not so much depend on the dynamics of adaptation, but more on the fact that the environment is adverse enough to keep away competitors and predators. In addition, we lack information about the exact timing of pollution events. The rivers we studied could have been impacted only very recently, not leaving enough time for the guppies to adapt. Currently, we do not have a final conclusive answer.

Interestingly, around the same time when we were trying to puzzle out why the “oily guppies” from Trinidad were so abundant, yet showed so little evidence of local adaptation, another brave biologist was facing a very similar conundrum: Steve Brady (currently at Dartmouth College) was investigating local adaptation to de-icing salts in amphibians from ponds located near busy roads, where he obtained results pretty similar to ours (you can read his very readable summary of the study at ecoevo-evoeco). Steve's wood frogs from roadside ponds were not only doing worse in non-polluted woodland ponds when compared to woodland populations, they also survived at lower rates in their salt-polluted native habitats – just like the “oily guppies”, they seem to be impacted by the pollutants in a way that causes them to perform poorly in general whenever compared to individuals from undisturbed habitats. Steve, in the search for an appropriate term, called the pattern he found 'deme depression', drawing on the analogous pattern of reduced fitness caused by inbreeding depression.

It goes without saying that our study (as well as Steve's work) on maladaptation is by no means complete, and there are many possible mechanisms that could cause the patterns we find for the “oily guppies”. Yet, I also think that both studies highlight the importance of reporting exceptions to classic predictions, especially in the context of anthropogenic disturbance of habitats. Although negative results, such as the hampered adaptation in “oily guppies“ and wood frogs, are frequently dismissed as resulting from methodological artifacts, it is just as reasonable to think that they actually reflect real biological phenomena that need to be investigated in more detail. We hope that our work encourages further explorations of such exceptions to expected patterns of adaptation in natural populations, especially in the context of stressful anthropogenic disturbances. Understanding the dynamics of adaptation in the wild is of immense practical interest for conservation biology at a time when we are only just beginning to understand our own impacts on evolutionary processes in natural ecosystems. Consequently, our understanding of how and why natural populations sometimes fail to adapt to ongoing environmental changes is an essential target of research that will help to sustain important ecosystem services and inform regulatory decisions. In short, we need not only environmental impact assessments, but also “evolutionary impact assessments” – investigations of the evolutionary consequences of environmental changes on natural populations.

References

Rolshausen, G., D.A.T. Phillip, D.M. Beckles, A. Akbari, S. Ghoshal, P.B. Hamilton, C.R. Tyler, A.G. Scarlett, I. Ramnarine, P. Bentzen, and A.P. Hendry. 2015. Do stressful conditions make adaptation difficult? Guppies in the oil-polluted environments of southern Trinidad. Evolutionary Applications 8: 854–870.  DOI: 10.1111/eva.12289.

Brady, S. P. 2013. Microgeographic maladaptation and deme depression in a fragmented landscape. PeerJ. Full text: https://peerj.com/articles/163/

Predicting Speciation?

(posted by Andrew on behalf of Marius Roesti) Another year is in full swing. What will 2024 hold for us? Nostradamus, the infamous French a...