In the several crazily busy weeks since my last post, I’ve been fortunate enough to have two manuscripts from my dissertation field research go live: one in Ecology and Evolution and one in Oecologia. I’m really glad these two papers were able to hit the airwaves in quick succession because, together, they address many of the central questions I raised in my dissertation. I thought it might be fun to reflect on and summarize those questions a little bit here.
My dissertation was trying to understand what plants (specifically, lowbush blueberry) “forfeit” in order to permit/afford/prioritize production of more reproductive structures. In essence, we would expect that as a plant reproduces more (rightward on the X axis), it would have to do less of at least one other thing (downward on the Y axis) and vice versa. Such a pattern would be called a “trade-off.” Some trade-offs occur because both traits require similar resources and those resources aren’t available enough to permit investment into both traits. Other trade-offs occur because it just doesn’t make sense to invest in both traits at once (cold tolerance and heat tolerance, say). I’m kind of crazy about trade-offs; I think they’re really cool to think about because they connect traits (the physiological characteristics of plants) to ecological and evolutionary pressures that plants encounter. In other words, they suggest reasons why plants look the ways they do!
As the conceptual model I’ve pasted in above illustrates, I had 5 questions about blueberry’s reproductive trade-offs:
- What traits trade off with reproduction, and how diverse are these traits?
- Are these trade-offs always present, or do they get more or less severe as conditions change?
- Is there a “better” way to quantify reproductive investment, if we’re trying to model the severity of trade-offs?
- Is modeling trade-offs as linear too simple an approach?
- Why, mechanistically, do these trade-offs occur? Are they all the product of resource limitations, or is there more to them than that?
To get the complete story, I’d encourage you to check out the papers themselves for all the juicy details, but, in short, here are the answers to these questions as I currently see them (although my thoughts are still evolving even as I write this).
- Of the 20 or so traits I considered, 10 of them showed evidence of trading off with reproductive investment in blueberry (see Figures 1-3 in the Oecologia paper). These included traits related to tissue chemistry (leaf chlorophyll content), phenology (vegetative development rate), fruit ripening rate, reproductive success (reproductive node success rate), and fruit chemistry (titratable acidity), among others. And these were just the trade-offs that were fairly static across years (see #2). So, I would conclude that blueberry’s reproductive trade-offs are pretty far-reaching, even though they don’t touch every single other trait we could imagine.
- Of the 17 traits I measured across more than one field season, several showed evidence that their severity differed significantly by year. But, to me, that’s not really the interesting part—the interesting part is how they differed by year. Many of these traits only showed evidence of significantly trading off with reproductive effort in 2013 and/or 2015. These years were much hotter and drier than 2014, which was very nearly a perfect year for blueberries in Maine (cool, cloudy, and wet). For example, vegetative mass per stem was higher in plots with experimentally reduced reproductive effort than in control plots in 2013 and 2015, but not in 2014. The same was true for the number of ripe fruits produced per reproductive node (See Figure 4 in the Oecologia paper). This is exactly what we’d predict—some trade-offs may only be apparent in years when plants are stressed and they thus can’t “avoid” the trade-offs as easily. This set of results is, perhaps, my favorite set across the two papers because it aligns so cleanly with what theory would predict.
- In studies on reproductive trade-offs like mine that manipulate reproductive investment by removing flowers, the levels of removal (0%, 20%, 40%, etc.) are often used as the X axis when quantifying reproductive trade-offs. I wondered if that might be too over-simplifying an approach. Because I counted all the flowers in all my plots before I removed any (because I’m a sucker for punishment), I knew (roughly) how many flowers remained afterwards, so I wondered if using this value as the X axis value would result in changes to the slope estimates I would get for my trade-off lines. Figure 3 in the Ecology and Evolution paper illustrates the potential magnitude of difference between the two approaches; the latter method produced lines with more negative slopes for 5 of the six traits I considered. Practically speaking, this means trade-offs are probably a bit more severe, on average, than we would otherwise appreciate if we used removal levels instead (not that that’s a bad approach, mind you! Counting all those flowers was really time-consuming! Doing so wouldn’t be possible in many systems).
- Are trade-offs linear changes in one trait with changes in the other? Or are they non-linear? Well, my interpretation from my results is that the answer is a little bit of both. It looks like, in general, trade-offs are approximately linear in the middle of the trait space shown in the conceptual model up above, but they probably turn non-linear towards high levels of one trait or the other. This makes sense, I think; the costs of investing more into a trait probably accelerate at some point, meaning the cuts to other traits would have to get deeper as that happens. Moreover, for traits where this non-linear trend towards the extremes is very stark (like for leaf area in blueberry, potentially), a non-linear curve might be a more informative choice.
- And, lastly, the all-important “why” question! If these trade-offs occur, what causes them? Here’s where things get really interesting—the answer in blueberry doesn’t appear to be limitation for carbon (except for the trade-off between reproductive investment and production of leaf mass) or nitrogen. I found little evidence for either of these hypotheses, which is interesting because most trade-offs are thought to be principally about resources; if a plant could invest more into both traits, it would do so. What I did find evidence of is that some of these trade-offs could simply be manifestations of the classic “quality vs. quantity” trade-off. For three of the six traits I studied, all related to reproductive structures, I found evidence that, when “marginal” reproductive structures near the tops or bottoms of stems were preferentially removed, the traits went up more than when reproductive structures were removed indiscriminately. This suggests, at least, that as more reproductive structures are produced, a greater proportion of them will end up being of marginal quality. That brings down the average quality of the reproductive structures produced as a whole. I dubbed this idea (and related ideas) “sink limitation;” these trade-offs seem to occur because the plant becomes increasingly limited by the quality of the sinks it can produce as it produces more of them in total. Beyond this explanation, though, I believe in hindsight that the most important resource limiting blueberry’s reproductive output may actually be water. I hope someone tests that possibility in the future!
I’m definitely feeling proud this week of what I was able to accomplish with my dissertation fieldwork. I’m hoping others will consider the methods I used and try building and improving on them in their systems; I think there’s plenty of room for others to take what I’ve done and improve upon it. Do you have any questions for me? Is there anything I could have explained better? Post these in the comments, or feel free to contact me directly via email!