Gould & Eldredge 1977 or G&E Strike Back

reading: Gould and Eldredge. 1977. Punctuated Equilibria: the tempo and mode of evolution reconsidered. Paleobiology. 3

I previously talked about an earlier paper where Eldredge and Gould put forward their concept of punctuated equilibria. This was a rather controversial concept at it's time. I think it's gone through the infamous 'stages of truth' series (which weirdly are tied up with Darwin's bulldog Huxley and famed embryologist von Baer as detailed here); first ridicule, then violent opposition, and then the claim that it's trivially and obviously true. 
Or, as the claim was initially stated by Schopenhaur:

"To truth only a brief celebration of victory is allowed between the two long periods during which it is condemned as paradoxical, or disparaged as trivial."

Reading these papers you see a lot of work that seems trivially obvious today: allopatric speciation, local sections don't necessarily represent the wide ranging species, speciation is a process of branching, etc. But some other aspects of these papers haven't had their brief victory celebration yet (species selection and speciation through regulatory gene revolutions).

Gould and Eldredge's major gripe is that gradualism prevents examination of the tempo and mode of evolution, and that in particular gradualism can't be refuted by fossil evidence because the evidence has been heavily interpreted under the gradualistic framework. Punctuated Equilibria, they argue, allows for fair assessment of evolutionary tempos, and once we understand that we can make some inferences about the mode that evolution operates through. They state that if the tempo is punctuational, then the mode is 'speciation' or branching.

By "Tempo", G&E borrow from George Gaylord Simpson (a leader in the Modern Synthesis) and mean variation in rates of evolution between lineages, while "Mode" is the mechanism that produces the variation in rates.
Tempo, once the blinkering effect of gradualism can be removed, can be observed empirically, and punctuated equilibria is  a mode that can be inferred from it. G&E feel that punctuated equilibria will be 'orders of magnitude' more important than phyletic gradualism, and that phyletic gradualism occurs 'hardly ever'. Importantly in this work, the fossil records sometimes failure to display radical morphological change, rather than being a 'failure to record' information is actual information, stasis, they reiterate is data. This is a necessary implication of applying the neontologist's allopatric speciation to the paleontologist's fossil record.

G&E also make it clear that they want two major changes to occur in the way paleontologists do their work (or rather did, since the initial paper on this topic was from more than 30 years ago); 1) evolutionary trends are the result of (higher level) species selection; and 2) workers need to quantitatively study the evolution of entire ecosystems and their members. They particularly want for workers to quantify geographic variation within a population and compare that to stratigraphic variation along and between lineages before anyone can really talk about stasis or punctuation (although they seem to relax that standard sometimes when people make claims of finding punctuation).

They then go through a number of studies that have supported or contra-indicated punctuated equilibria and tests for it, amoung them Stanley 1975 (pdf); Hecht 1974 (a longstanding and well-respected Chair of Biology at CUNY Queens College); Hayami & Ozawa 1975;  Makurath & Anderson 1973; Gingerich 1974, 1976, 1977 (pdf); and Klapper & Johnson 1973 (whom they seem to paint as what I will call 'naive gradualists'). They also take to task workers who publish on evolutionary trends when all they really have are three data points, an original population, and then say one where the individuals are bigger on average, and then a third population where they're a little bigger again, on average. Trends really have to be based on many points, not just a few, G&E go through some basic statistics on why this is so.
G&E also look at a few cases where other authors have calculated the rates of gradualistic change, finding rates along the lines of 10% per million years. This means a fantastically small amount of change per year or per generation, which should just be wiped out by genetic drift. It also implies uninterupted multi-million year long selection pressures, which is terribly odd, and also seems to beg the question of why not select for more change over a shorter span of time?

Speciation Theory

All of this is well and good and most workers today try to follow these recommendations, good papers carry them out fully, examining entire faunas, gathering solid data for statistical analysis, paying attention to stratigraphy, etc, and really great papers make explicit statements about null & multiple working hypotheses and put their theoretical assumptions up front.
Beyond this, G&E start trending into more controversial territory. They enter this territory by way of an opening analogy: 

"speciation is the raw material of macroevolution, and genetic substitution within populations cannot be simply extrapolated [into macroevolution] [...] We therfore challenged [that] change in gene frequency within populations is hte building block of major evolutionary events"

Species selection/species sorting is selection at a heirachical level higher than the cannonical individual, it is something that is still strongly debated today. As they presented it here and in their 1972 paper, marginal species randomly enter peripheral environments. These sub-populations respond especially well to their surroundings in some types of peripheral environments. From this, an overall effect arises, a trend. Lets say a population is exposed to cold conditions at multiple points (and even at many times) along the edge of it's range, and that this species always tends to strongly react to 'cold' by forming thicker fur, bulkier body-types, etc. This means that, of the varieties of this species that are out there, a bunch of them are going to be cold-adapted, and over geologic time-scales, you're often going to have cold-adapted sub-types pop up. There's a good chance that the cold-adapted subtype, simply by the numbers, and through the actions following allopatric speciation, will tend to be successful and replace it's parent variety/sub-species/species. And this process repeats. The "net effect" is a trend in body-type, fur, etc. 
This is a messy idea, and a big problem is, what are G&E saying is actually going on? To continue with the above example, does the trend towards 'cold-types' happen because of differential reproductive success of some individuals within a population, or is it happening because of the success of the species as a whole? How do we distinguish selection for cold-type individuals against selection for cold-type species? 
Also, in the above, and maybe this is just my misunderstanding, the overall environment doesn't need to get cold, the whole lineage can show a trend towards the 'cold-type' while the temperature across the overall range remains the same, the cold-types win out because there's always a bunch of them around and they can expand out of their limited allopatric range. Perhaps that'd be a good test for species selection, a non-adaptive trend that starts out as an adaptation to a local environment (this is the opposite of how Gould often talks about adaptations, with most current 'adaptations' starting off in a different functional context, or iow as 'exaptations', in his coinage).
G&E talk about their ideas about trends and species selection being a necessary/logical consequence of two things, 1) the occurence of punctuated equilibria (itself a logical consequence of allopatric speciation applied to the fossil record) and 2) that the morphology associated with a speciation event is random with respect to the direction of evolutionary trends within the group (their so-called "Wright's Rule").

G&E also wander a bit too closely to the line of '"explains everything" when it comes to punctuated equilibria also. They consider anagenesis to be simply the result of species selection over many, many splitting events (making it something like a trend). But anagenesis is most people's word for "phyletic gradualism". So they appear to be saying that punctuated equilibria can actually explain the very process that it's set up antithetically to, perhaps this dialectic 'negation of negation' is the Marxist-Hegelian spirit at work in Gould. Regardless, a theory that explains everything in the universe is a useless theory (consider the "godidit" idea, even contradicting evidence can be explained as "godidit"). Perhaps, similarly, a theory that can explain everything under it's ambit is a little too good (and by implication we're probably deceiving ourselves and the theory is ultimately wrong or more limited). Along these lines, G&E even claim that punctuated equilibria can now explain events below the species level, especially within asexual species; we already have a theory that explains those phenomena, it's the standard modern synthesis, there's no need to tack on punctuations.
G&E also step into what I think is unfamiliar territory for them, and they really step on it. Perhaps in the 70s it was debateble if they were right or wrong, but their idea that speciation occurs with a 'genetic revolution' (admittedly this is attributed back to Mayr), and that this revolution involves drastic re-structuring of the regulatory parts of the genome, is terribly wrong. When you start writing things like  (and this is actually Carson 1975 that G&E are quoting):

No, not that Carson

"Speciation is considered to be initiated when an unusual forced reorganization of the epistatic supergenes of the closed variability system occurs"

you know you're in trouble: time to take a step back and re-evaluate (see here for a more sympathetic and fuller discussion). 
G&E are particularly wedded to the idea that evolution occurs through variation in the tempo and mode of development, which is why Gould is so interested in heterochrony and paedomorphy. Perhaps Carson's statements were just too 'in line' and tempting with their thinking to prevent them from stepping into this topic. I just want to be clear, this issue of the genetics of speciation was a lively topic for a long period of time, there's nothing 'invalid' about it obviously, it just seems to me the G&E overstepped, widely, by taking a side on this issue in this paper.

G&E, despite getting some things very wrong, got a lot right. In particular, Figure 1 shows how they see punctuated equilibirum, with it's branching pattern of speciation (they call it a 'v' pattern and contrast it with what others call a 'y' patter). Figure 1 (Figure 8 in their paper) is basically what any phylogenetic tree looks like in modern papers today.

Figure 1

Compare this to Figure 2, from Klapper and Johnson 1975, a paper they examine in some detail (and showing the 'y' shapped pattern). Figure 2 is the sort you commonly see in older papers, you very rarely see a phylogenetic tree presented in this manner. 

Figure 2

This is a minor point in some ways perhaps, but it illustrates that G&E were definitely on the right side with respect to branching and clades.

On Punctuated Equilibria

Reading Eldridge and Gould, 1972. Punctuated Equilibria: An alternative to Phyletic Gradualism. in Schopf (ed) Models in Paleobiolgoy. Freeman, Cooper, & Co. 

E&G begin by noting that research is not conducted in a vacuum and that we don't observe data with a "viewpoint from nowhere": we use theory to organize and interpret data. they call this a "picture" rather than a paradigm/research programme/etc, explicitly trying to avoid the longstanding debate over those terms.

Stepping off from this, they claim that most paleontologists hold a conceptual picture of evolution walking along with slow and small steps; they term this "phyletic gradualism" and link it with sympatric speciation. Importantly they feel that paleontologists haven't been keeping up with the mainstream of population biology, where allopatric speciation is (or at least was in the '70s when the paper was written) was all the rage.

So with that hypothetical apparatus in mind (that "picture" influences theory and paleos currently work under a gradualist picture), they consider how allopatric speciation would look in the fossil record. E&G look at two fossil groups and attempt to establish that the data can be explained, and can possibly be more "interestingly" explained, under the allopatric model.

First they consider Poecilozonites, a genus of pulmonate snails from Bermuda. Using different pictures, they can argue in support of allo- or sympatric  speciation. The species under consideration are all subspecies of P. bermudensis that are marked in being paedomorphic; the adults retain juvenile features. A story of gradual cumulative change can be laid out, but when you start including geographic information, more support is seen for allopatry.

Second they review Phacops rana and related trilobite species from Devonian New York--Ohio strata. In particular the discussion focuses on changes in one 'character' (although it's a complex character with many related components, as the 
authors point out), the number of "dorso-ventral files" in the eye.  They find that the mainline species has 18 of these eye-files, and argue that marginal populations arise with variable number of eye-files. These marginal peripheral populations then expand/migrate, overtaking the mainline. This is the allopatric model in essence. In each case of these triblobites there's a reduction of the number of files in the eye (see Figure 1). 

Figure 1 - Hypothetical Phylogeny

I don't really know anything about the eyes of trilobites, other than that they're complex/compound and insect-like, but are not related to insect eyes. Eyes are fascinating structures, famously Darwin seemed to waiver that natural selection could produce something so complex, and who's function seems so reliant on the interdependence of parts. But of course Darwin immediately recognized that the eye could evolve in stages, and he even cited some fossil examples of probable stage. In fact, one would think that that was a lucky accident and that eyes turned up once in a primitive ancestor and have been inherited by all eyed organisms today-- or maybe they evolved twice, one for organisms with eyes 'like ours' and one for compound eye type organisms). But that's not the case, eyes of various sorts have independently evolved many times amoung animals, up to 100 times.
The only other thing I know about trilobite eyes is that their lenses are made of calcite, a mineral. Our eyes lenses are nothing like this, they contain crystallin, which is a protein, not  a mineral (despite what its name might suggest to some) and our lenses are metabolically active.
So this business of "dorsoventral" whatever seemed like it was worth looking into. Trilobite eyes are compound, similar to an insect's, but independently evolved (they're possibly the oldest eyes we have on record). Each facet is made up of a small calcite lens (and other tissues), and a string of lenses is what E&G is referring to as a "dorso-ventral file". There's more to the eye, with the visual unit, capped by a lens, being called an ommatidia. Lines of lenses that run between the dorsal and ventral surface of an eye are called d-v files, and lines of lenses that run horizontally across an eye are simply called rows. The number of files is used in the determination of species within trilobites.

Figure 2 - Trilobite eye structure

Interestingly, the some of the specimens referred to by E&G are from the Marcellus Shale in NY (a source of hydrofraked natural gas).

E&G go on to state that there's an expectation amoung paleontologists that successively higher taxonomic ranks should have progressively more and more taxa within it, they believe this incorrect assumption is a result of the "picture" of phyletic gradualism; as time goes on more and more species are produced. The reality is that there are, infamously, lots of higher ranks that are species poor, so we in some families there are hundreds of genera each containing dozens of species and good sub-species, but often enough we can have Orders with a few monotypic genera. Allopatric speciation can explain this as repeated splitting with either 1) the parent species going extinct and only marginal ones surviving; 2) when geographic isolates adapt through new modes of feeding/motion/protection/etc; and 3) when it involves a small isolated lineage.

Finally E&G address that exemplar of phyletic gradualism, the evolution of long-term (and especially adaptive) trends in a lineage. They feel that allopatry can result in the appearance of  a trend by way of analogy to how random mutation in a population can still result in the overall production of a trend within that population. Selection pressure moves the population in one direction, and something that would eventually become called "species sorting", IIRC, similarly produces the trend at a higher level. They point out a mechanism in a little more detail, relying on something like the genetic and historical constraints of the mainline species tending to result in marginal species reacting strongly and in the same way to particular to similar marginal environments--say, developing thick skin in desert environments; the net effect is an overall trend for the group of species.

You can see a lot of anticipations of Gould's later work on hierarchical levels of evolution in this work, along with some material that, probably through uncharitable readings, was used to charge Eldridge and Gould with being monstrous saltationists.

I have to wonder at some of their examples though and if they really show a signal of allopatry. With Phacops, we see a few marginal populations developing, in these cases through paedomorphosis, in different locations and then expanding over the ancestral range. E&G note that the mainline population is invariant, with 18 d-v files, while the marginals are at first more variable, and then later less variable with a reduced number of d-v files. But why isn't this just a large, general population with variability in the number of d-v files, why consider the variable population to be an isolate?  If you look at any one slice through time, you find a wide-spanning population with variable d-v file numbers, some living in epeiric seas, other in marginal seas, which aren't terribly different environments either. 

Figure 3 - Some notes on the hypothetical Trilobite phylogeny. Red lines 1--3
are samples at a particular time, the green arrow is a possible trendline.

Figure 3 show populations (marked with red lines) with 1) 18--17 d-v files; 2) only 17 d-v files; and 3) 15--17 d-v files. Further, the green line in Figure 3 shows that in epeiric seas as you move through time the number of d-v files changes 18 to 17 to 15, a trend of reduction in this group at this location. The authors posit that migration has occurred  not evolution in place over a long period of time.

 Obviously the justification for allopatry must be in Eldrige's (and others) stratigraphic and geographic work on the group, but it'd be more helpful to have some discussion of that.

One other thing that really sticks out in this work is that E&G are heavily operating within the adaptationist programme, ironic given that Gould is such a critic of that. Whether they're considering allopatry or sympatry, they can find adaptationist explanations for all the features. Perhaps shells became thinner as an adaptation to living in limey soil, or perhaps that was just the result of drift, a meaningless fixation. It's hard to believe that you can have a wide ranging population of Phacops trilobites with something like the structure of the eye varying so much, and that this is the result of selection pressure for the number of d-v files, rather than just meaningless variability in their number. Eldridge, and others, promoted the idea of identifying species within the trilobites by counting (presumably amoung other things) the number of dorso-ventral lines. Perhaps that 'picture' of trilobite evolution coloured his ideas here.