Sunday, February 7, 2010

The End of Adaptive Radiation?

In a provocative new Bioessays article, Olson and Arroyo-Santos attempt to kill the term "adaptive radation." Olson and Arroyo-Santos are certainly correct when they suggest that interest in adaptive radiation is paradoxical: although it is almost universally regarded as an important (if not the most important) mode of biological diversification, evolutionary biologists have argued for decades about how to diagnose and define it. Olson and Arroyo-Santos suggest that this century-old debate stems fundamentally from the fact that "adaptive radiation cannot be winnowed to any core meaning, because there is no phenomenon in nature to which the term corresponds, simply arbitrary divisions of continua."

The problem with this argument is that it focuses exclusively on the one feature of adaptive radiation that is controversial - namely, whether extraordinary diversification is intrinsic to adaptive radiation. While I agree with Olson and Arroyo-Santos's suggestion that efforts to define adaptive radiation on the basis of the extraordinary levels of diversity are hopelessly ambiguous and arbitrary, I disagree with the fundamental premise that extraordinary diversification, in any form, is intrinsic to adaptive radiation. If we focus on the shared features of modern definitions of adaptive radiation, we find that it can be universally defined as a response to natural selection and ecological opportunity involving divergence of species and associated adaptive features. In this sense, adaptive radiation is analogous to the ‘principle of divergence’ that Darwin introduced in the Origin by suggesting that “the more diversified the descendants from any one species become in structure, constitution, and habits, by so much will they be better enabled to seize on many and widely diversified places in the polity of nature, and so be enabled to increase in numbers.” Perhaps we should replace 'adaptive radiation' with 'principle of divergence,' but abandoning efforts to label the profoundly important evolutionary phenomenon that underlies these terms seems like a bad idea to me.

10 comments:

Brian O'Meara said...

Isn't their major point that adaptive radiation isn't a "thing" -- that is, an actual distinct process in nature -- but rather a category we assign groups of organisms to based on diversification and disparity (see their box 1, figure C)? I think they're right that adaptive radiation is often treated as a pattern to test for (if my special clade has a faster rate of diversification and disparification than its sister (p<0.05), it's an adaptive radiation), in which case it is arbitrary and it would be better to focus on the underlying processes (causes of speciation, factors affecting extinction, diversifying selection, etc.), especially getting parameter estimates for these processes. If we focus on adaptive radiation not as a label for clades but as a distinct process or circumstance (as we might study the effect of glaciation on species or how evolution of mobile predators affected defensive strategies), research moves from testing the appropriateness of the label to understanding the process: do empty niches promote peripatric speciation or do they reduce extinction rates, does lack of competition allow broader intraspecific variation in perch height that is then partitioned between species, etc. People are investigating these questions, using increasingly sophisticated methods and not limiting themselves to just rejecting trivial null hypotheses. But do any of these questions pertain only to clades identified as adaptive radiations?

To ensure "adaptive radiation" remains a useful term, in the same way hybridization, stabilizing selection, allopatric speciation, etc. are, it would be good to start with a description of the envisioned process of adaptive radiation and then think of patterns that would give information about this process' presence and features. It does feel like the definitions generally start with the patterns first. Even in this blog post it is defined as "a response to natural selection and ecological opportunity involving divergence of species and associated adaptive features" -- that is, the pattern outcome ("response", "divergence") not a process. In Schluter's Ecology of Adaptive Radiation book, adaptive radiation is "the evolution of ecological and phenotypic diversity within a rapidly multiplying lineage. It involves the differentiation of a single ancestor into an array of species that inhabit a variety of environments and that differ in the morphological and physiological traits used to exploit those environments. The process includes both speciation and phenotypic adaptation to divergent environments" (pp. 10-11). This describes what a clade evolving via adaptive radiation looks like (fast speciation, high trait disparity, etc.) but not why it has evolved in that way. I'm having trouble figuring out how to describe an "adaptive radiation" mechanism that is not just a mixture of other processes (speciation, adaptive morphological evolution, character displacement, etc.) and that is a common process of adaptive radiation clades and not other groups. Having such a description would undercut many of Olson & Arroyo-Santos' arguments against continued utility of the term -- perhaps someone else can propose a good description.

Glor said...

I agree with Brian about everything except his assertion that the definition used in my post describes a pattern rather than I process. I don't see why the term "response" can't be used to indicate a process.

In any case, what is more important is that we agree that adaptive radiation is a process, albeit a complicated one that includes a variety of subprocesses (i.e, speciation, adaptation, etc.). To me, the feature of adaptive radiation that sets it apart from these other processes is that speciation and adaptive divergence are intimately linked as a lineage evolves to exploit ecological opportunity.

Brian O'Meara said...

Thanks for your comment and the original post, Rich. It'd be interesting to dig into precisely what you mean by "speciation and adaptive divergence are intimately linked as a lineage evolves to exploit ecological opportunity". Does speciation at a "default" rate cut the gene flow connecting populations, allowing at least one to zip to occupy a new niche, and then reduced competition reduce the chance of one of the daughter species going extinct, resulting in higher diversification? Does the presence of empty niches themselves promote speciation in some way? It'd be great to nail down potential mechanisms causing this linkage.

Anonymous said...

Great discussion, and I am itching to read the paper and jump in. However, I am getting ready for a trip to Tanzania that leaves on Sunday, grad student interviews this week, and the final push to get a draft of a big paper that has many long-suffering collaborators...

Thanks Rich for posting this!

220mya said...

I am completely on board with a process-based vs. pattern-based approach to studying adaptive radiation and other comments. But that being said, I have two major concerns:

1) Nowhere in any of the definitions put forth in this post or resulting comments have people included "adaptation"! At its most basal level, "adaptive radiation" is all about an adaptation or suite of adaptations resulting in differential diversification/cladogenesis. This means that to identify an adaptive radiation, you need to demonstrate that some phenotype is an adaptation (see for example Greene 1986 and Padian 1987), then you need to demonstrate an increased rate of diversification relative to clades without that adaptation. Once this has been done, you can then study the underlying process. It's for this reason that I have a problem saying that something like parameter delta in Continuous/Bayestraits can identify adaptive radiations - because it's only looking at a part of the puzzle (you need to determine that the trait is an adaptation first!). Very few recent papers on adaptive radiation actually do the legwork to demonstrate that the trait they are examining is an adaptation.

2) As a paleontologist, I worry that some newer definitions of "adaptive radiation" are only applicable if you have molecular data. This is indeed unfortunate given that the paleontological record is one of the most powerful sources of data for testing macroevolutionary phenomena. Although many molecular biologists conveniently forget it, differential extinction is a bitch if you're trying to correctly recover a deep-time signal with only extant taxa (see for example Turner et al. 2009). We need to make sure that definitions of these macroevolutionary phenomena are applicable across all fields of evolutionary biology.

Greene, H. W. 1986. Diet and arboreality in the emerald monitor, Varanus prasinus, with comments on the study of adaptation. Fieldiana Zoology, New Series 31:1-12.

Padian, K. 1987. A comparative phylogenetic and functional approach to the origin of vertebrate flight; pp. 3-22 in M. B. Fenton, P. Racey, and J. M. V. Rayner (eds.), Recent Advances in the Study of Bats. Cambridge University Press, Cambridge.

Turner, A. H., N. D. Smith, and J. A. Callery. 2009. Gauging the effects of sampling failure in biogeographical analysis. Journal of Biogeography 36:612-625.

Glor said...

@Brian
I don't think the mechanism of speciation is particularly important to adaptive radiation. Some models suggest that speciation results from adaptation (see Schluter) wheres other models suggest that species diverge for other reasons and permit accumulation of adaptive differences (see Lack, Losos, etc.). The key is that the amount of adaptive divergence that is possible in a given group is limited in the absence of speciation (and vice versa)

Glor said...

@220mya
Although I didn't explicitly mention tests for adaptation in my post, I, agree, of course, that this is essential to establishing the occurrence of adaptive radiation. Extinction is definitely an issue and I hope that modern studies will not diminish the importance of paleontological insight on adaptive radiation.

220mya said...

@Glor - My comment was certainly not meant as an indictment of your original post - I apologize if it came off that way! I was mainly reacting to the fact that a sizeable number recent papers on "adaptive radiations" do seem to forget about testing for adaptation in the first place, and this may lead in part to the ambiguity and diversity of definition discussed in the Olson & Arroyo-Santos paper.

Brian O'Meara said...

@Glor: adaptation causing speciation or speciation allowing adaptation are different processes. They might result in the same pattern (positive correlation of disparity and diversity), but if the whole point is that adaptive radiation is a process not just a pattern, it's worth distinguishing these very different processes so you can then test to see which one is acting. Labeling two different processes adaptive radiation because they result in the same pattern is exactly what Olson & Arroyo-Santos warn about. [sorry to keep commenting on this post, but adaptive radiation is an important concept that seems like it could use some clarification, and this is a good venue for that].

Glor said...

I do think this is an appropriate place for this discussion. I think there is some confusion about what a process is. In my view, a process can encompass a suite of possible underlying mechanisms. Wiens makes this point in a recent paper on niche conservatism: "...many well-known processes are actually caused by a variety of different processes when viewed at a finer scale. For example, a variety of genetic mechanisms can lead to speciation (e.g. polyploidy, sexual selection)." In my view, divergence of species and associated adaptations (i.e., adaptive radiation) should be viewed as a process even if this process has more than one possible underlying mechanisms. I think we are on firm ground as long as we can study all of these mechanisms and understand their shared outcomes.