A rejected analogy

Analogies can be useful ways of explaining complicated ideas - but they can also be problematic. Reviewers of a recent paper were having trouble understanding a rather intricate idea we were presenting. Thus, on revision, I attempted an analogy. I liked the analogy and found it helpful but - whether unfortunately or not - it didn't make it into the final paper. It was left on the cutting room floor, so to speak. Still, I kind of like it and so provide it here - with context.

Led by Sarah Sanderson, we recently conducted several studies that tested the hypothesis that populations living in areas with low levels of a limiting nutrient would show compromised performance for traits depending on that nutrient. Specifically, we first tested whether fish populations living in water with very low calcium levels would show reduced levels of calcium in their scales. This work was possible because several native species can be found across an environmental gradient in calcium levels, from "high" in the St. Lawrence to "low" in the Ottawa River (Figure  below).

To cut a long story short, we found NO EVIDENCE that the populations in low-calcium water had lower levels of calcium in their scales. The figure below shows that - for a given change in water calcium - scale calcium levels changed hardly at all.  At one level, this made sense - because calcium is what makes scales strong - and strong scales aid defense from predators and other environmental stressors. Yet experimentally exposing fish to similarly low calcium levels has been shown to compromise various aspects of performance - so how were these fish maintaining high-quality scales?

Our next hypothesis was that these fish really really need high-quality scales - and so prioritize that function when faced with low calcium availability. If so, they might show reduced functionality of other calcium-dependent traits - and so we next looked at various aspects of the skeleton. Again, we found no evidence that fish in low-calcium water show any compromise in this calcium-dependent structure. This result is summarized below, where the results for scale calcium from the earlier paper are in red and the results from various skeleton measurements are in green and blue. 

 

It seems, then, that populations in low-calcium water have somehow evolved to be better at obtaining calcium and/or building calcium-dependent structures - perhaps a classic case of "counter-gradient variation" (which would have to be confirmed using common-garden rearing). Yet, even in cases of counter-gradient variation, something else is often compromised - that is "Darwinian Demons" who are great at everything are unlikely. 

So where are the compromises then? We could always look at other calcium-dependent structures (although scales and skeletons are the most obvious) or perhaps the effort required to obtain calcium in low-calcium water has other costs, such as to growth or survival. But this is could be a wasted effort because ....

And now we come to the alternative idea for which I generated the analogy. Here it unfolds as written out in the draft version. The first paragraph is the analogy that was axed later and the rest of it was also modified for the final published paper. Perhaps you will like it or maybe you will hate it but, regardless, I bet you will remember it - even if only for short time.

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

Perhaps trade-off payments during adaptation are “front loaded”

"We start with an analogy. When purchasing a mortgage for a home, the purchaser incurs an additional cost (beyond the cost of the home itself) in interest payments to the lender (usually a bank) – and most mortgages are structured such that those interest payments are front-loaded. As a result, payments early in a mortgage include a lot of interest payments (on the money loaned by the bank), whereas payments late in a mortgage almost entirely reduce the principal (because the interest was paid earlier). Thus, if one examines mortgage payments early on, this additional cost is apparent – and is a trade-off associated with purchasing a home via a mortgage. Later on, however, examining mortgage payments would suggest this cost was minimal – because the interest payments have become very small. In short, a strong trade-off (interest payments) can be evident early in the process of adaptation (getting a home) but are not evident later in the process (because they were paid early on).

From this analogy, we suggest that the costs of adaptation to a limiting nutrient might be absent in the present because they have been paid via selective mortality in the past. When an environment changes to become more stressful or difficult (e.g., colonizing an environment with limited resources, like calcium), selection for improved tolerance to that stressor is expected to be “hard” (as opposed to “soft”) – that is, by increasing mortality rates (Brady et al., 2019a). This increased mortality represents a cost in the form of “selection load” that stays high until evolution better adapts the population for the new stressful conditions. If population sizes are low during this period, an additional cost can be incurred through inbreeding that exposes “genetic load” (Crow, 1970). Once the population adapts and increases in abundance, however, it has paid those costs of selection (removing maladaptive alleles) and inbreeding (“purging” recessive deleterious mutations); and, hence, might perform better than the ancestral population in both environments. Indeed, populations adapted to stressful environments can show higher fitness (or at least not lower fitness) in all environments – both in laboratory adaptation studies and in field experiments (Reed et al., 2003; Rolshausen et al., 2015).

Applying this last scenario to our study system, adaptation to low-calcium water might have been extremely difficult at first – generating substantial mortality and strong selection. This expectation is supported by experiments that expose naïve fish to low-calcium water (Baldwin et al., 2012; Iacarella & Ricciardi, 2015) and by the failure of Ponto-Caspian invaders to colonize low-calcium water (Iacarella & Ricciardi, 2015; Jones & Ricciardi, 2005; Palmer & Ricciardi, 2005). During this initial period, trade-offs would be expected. During the period of intensive adaptation, selection would tend to remove individuals that showed the strongest trade-offs – or that suffered the most from them. Once the population passed through this period and became reasonably well adapted to the stressful conditions, the result could be a locally-adapted populations able to maintain homeostasis without incurring large costs. Time will tell whether invaders in the system will be able to pass through this same bottleneck.

We are not here suggesting that this selection cost is high every time a high-calcium fish population colonizes low-calcium water. Instead, a long history of native fishes occupying a diversity of calcium environments has probably maintained a pool of standing variation that facilitates rapid adaptation to new calcium conditions. The costs paid during this rapid adaptation would presumably be lower than the cost paid the first time that adaptation proceeded – that is, the first time a high-calcium fish lineage successfully colonized low-calcium conditions. Subsequently, alleles suitable for adaptation to low-calcium conditions might persist within the species as a whole – even when not in low-calcium water. An analog to this situation could be the ability of marine threespine stickleback to adapt repeatedly and rapidly to new freshwater habitats via standing genetic variation in the marine population that persists via gene flow from past and present freshwater populations (Roberts Kingman et al. 2021; Roesti et al., 2014; Schluter & Conte, 2009). We can see considerable value in applying these ideas to other systems where some fishes can occupy a broad diversity of habitats without obvious costs, whereas other fishes cannot."

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