The Princeton Guide to Evolution

Macroevolutionary Trends

Gene Hunt

OUTLINE

1. Directionality in evolution

2. The scope of trends

3. Trend mechanisms

4. Examples of trend hypotheses

Trend hypotheses suggest an underlying directionality to evolution in which some changes are more probable than others. Such trends can operate narrowly—within a single species—or so broadly as to encompass all life. By tracking the characteristics of species and clades over time, paleontologists have documented trends in many kinds of traits and across many types of organisms. Selected examples are used here to illustrate aspects of trend hypotheses, including their scope, evidence, and expression in the fossil record. Fundamentally, two kinds of mechanisms can generate trends: (1) biased microevolutionary changes within species and (2) differential proliferation of species with different characteristics (species selection). From these two mechanisms, an enormous number of specific trends have been proposed. Assessing these trend hypotheses is important for understanding what generalities apply to macroevolution, and for understanding the long-term trajectory of earth’s biota.

GLOSSARY

Clade. A group consisting of all species that have descended from a particular ancestral species.

Cope’s Rule. A specific trend hypothesis that suggests that body size tends to increase in lineages and clades over time.

Macroevolution. Evolution occurring over very long periods of time or across more than one species, in contrast with the short-term microevolutionary changes in populations from one generation to the next.

Punctuated Equilibrium. A macroevolutionary model in which most trait divergence is concentrated into punctuations associated with lineage splitting (speciation).

Species Selection. The differential proliferation of traits resulting from trait-related differences in speciation or extinction rates. Species selection can generate trends if traits have consistent effects on speciation or extinction rates.

Trend. A persistent temporal change in a characteristic of a lineage or clade.

1. DIRECTIONALITY IN EVOLUTION

Systematic exploration of the fossil record began in earnest in the early nineteenth century. Almost as quickly as fossils could be pulled from the ground and described, paleontologists and others began to wonder what lessons this record held about the temporal trajectory of earth and its biota. Central to these discussions was the issue of directionality: Does the sequence of life-forms reveal repeated or sustained trends, or is it instead a story of fluctuations and catastrophes? This theme emerges in many different contexts, ranging from the specific (did horses systematically evolve to larger sizes?) to the universal (does life proceed from simple to complex?), and it remains an active area of research spanning paleontology, evolution, and comparative biology.

An evolutionary trend can be defined as a persistent temporal change in a characteristic of a lineage or clade. Trends need not be monotonic—reversals in direction are allowed—but they do imply that some kinds of evolutionary change are more probable than others. Given enough time, changes in a preferred direction can accumulate into substantial increases or decreases in the variable of interest. But the defining characteristic of trends is a bias in direction, not the magnitudes or rates of change they yield. Historically, trends were associated with the idea that evolution is progressive—that in some objective sense, evolution improves organisms or species over time. Although attempts have been made to recast progress into terms consistent with modern evolutionary understanding, most workers eschew this term, as it burdens an unambiguous concept—directionality—with a value judgment about whether that directionality is desirable.

Trends can operate over intervals as short as a few generations, but the focus of this chapter is on those trends that span periods long enough to be documented in the fossil record.

2. THE SCOPE OF TRENDS

One can categorize trends according to the biological unit that is traced over time. This unit can be as narrow as a single species or as broad as the entirety of life. The scope at which a trend operates is important, because different kinds of generating mechanisms are relevant when considering evolution at different scales in the biological hierarchy.

Trends within Species

The most focused kind of trend occurs when a species attribute increases or decreases systematically through time. Paleontologists have assessed this kind of trend hypothesis by measuring features of particular lineages through successively younger rock layers. Because the fossil record is dominated by mineralized hard parts (bones, teeth, and shells), paleontological trends usually involve the size and shape of these skeletal elements. Many kinds of organisms have been studied this way, resulting in specific trend hypotheses about directional change in, for example, the sizes of human brains, the shapes of bivalve shells, and the characteristics of mammal teeth, to name just a few.

Charles Darwin thought that natural selection ought to transform species steadily, and he was troubled by the apparent lack of such trends in the known fossil record. Darwin’s solution to this conflict was to argue that the geological record was so woefully incomplete that preservation of these gradual species transformations was unlikely. For many years this view was widely held by paleontologists, only to be challenged in 1972 by the punctuated equilibrium model of Niles Eldredge and Stephen J. Gould. These authors argued that species-level trends rarely appear in the fossil record not because of incompleteness but because such trends truly are rare, at least at the timescales that paleontologists can normally resolve. Eldredge and Gould suggested that most species exhibit little net change through time and that, instead, changes are concentrated into punctuations associated with the splitting of lineages as they form new species. The association between speciation and phenotypic evolution was not read directly from the fossil record but rather was inferred to be a consequence of how new species were thought to form.

The punctuated equilibrium model was controversial for several reasons. Particularly spirited disagreements stemmed from its potential implications for evolutionary processes and for the relationship between micro- and macroevolution. More fundamentally, some paleontologists questioned whether the model was even correct about pattern. These critics argued instead that gradual species-level trends were, in fact, not uncommon among well-preserved fossil taxa. This conflict over pattern proved very productive in that it motivated a series of studies designed to test the frequency of within-species trends. Early interpretations of these studies varied greatly, but recent overviews conclude that strongly directional trends are quite rare in fossil lineages; instead, most species show fluctuating or meandering changes. Presumably, the directional changes that differentiate closely related species accrue over temporal intervals shorter than paleontological sampling can usually resolve (at least 1000 to 10,000 years in most sedimentary records).

Trends within Groups of Species

Trends can also be considered at a broader genealogical scale, usually in groups of related species called clades. Such trends are detected by tracking the mean trait value over the life of a clade, at least for traits measurable on a continuous scale such as size and shape. Trends in qualitative or categorical variables manifest as systematic changes in the frequencies of the different states over time.

Many of the iconic examples of trends in the fossil record have operated within groups of related species. For example, humans and their immediate relatives do not form a single lineage but rather a tree with several major branches and a dozen or more species. The large brains that are so characteristic of modern humans are the result of increases spanning multiple branches of this tree. Thus this trend has occurred at the clade level, albeit at an accelerated rate in humans. The story is similar for that archetype of paleontological trends, horse evolution. Although it is sometimes (wrongly) depicted as a lineal sequence of ancestors and descendants, the net trajectory of horses from small, browsing ancestors to large-bodied grazers plays out over a complex fossil history that spans 55 million years and well over 100 extinct species.

Some trend hypotheses have even broader scope than individual clades. Paleontologists since the 1960s have compiled large databases of where and when different taxa occur, and these data have been analyzed to get a big-picture view of the major features of life over time. One common version of this approach looks over the past half-billion years or so at all durably skeletonized marine invertebrates; this selection of taxa spans many clades that are only distantly related (e.g., mollusks, corals, brachiopods, trilobites), and it excludes subcomponents of these groups (vertebrates, nonmarine, and poorly preserved members of the included groups). There are parallel analyses for terrestrial vertebrates. Earlier, informal versions of these are the source for the familiar statements that link periods of geologic time to taxa common during those intervals. Every child learns that the Mesozoic was the Age of Dinosaurs, and the Cenozoic was the Age of Mammals; other labels involving less charismatic organisms exist but are less often invoked.

In addition to documenting the changing dominance of life’s different players, whole-fauna (or flora) analyses have been interpreted to suggest many large-scale trends in the history of life. One that has received intense scrutiny is the trajectory of species richness through time. Animal diversity is almost certainly quite a bit higher now than it was in the early Paleozoic, but the magnitude, timing, and causes of this increase are all actively debated. Other suggested trends or state shifts through this interval include temporal declines in rates of origination and extinction, shifts in the frequency of different modes of life (e.g., sessile filter feeding versus mobile grazing), increases in the intensity of predation, and increases in organismal and ecological complexity.

3. TREND MECHANISMS

Although the biological context of each trend is unique, fundamentally there are only two kinds of mechanisms capable of generating macroevolutionary trends: (1) biased microevolutionary changes within species and (2) the differential proliferation of species with different characteristics, also known as species selection.

Trends as Accumulated Microevolution

The most straightforward mechanism for producing a trend involves a bias in the direction of phenotypic changes within species. Usually the cause of this bias is assumed to be natural selection favoring one direction of change more than others. When this bias is restricted to a single lineage, a species-level trend is produced; when it applies across groups of related species, the result is a trend across an entire clade.

An example of a trend commonly described in these terms is Cope’s rule, the notion that animal body size preferentially increases within lineages. It has been postulated that larger-bodied individuals have numerous advantages over their smaller conspecifics, including superior physiological buffering, greater ability to subdue prey and defend against predators, and greater success in the competition for mates and resources. While these suggestions are not unreasonable, they can be difficult to test, and quite often the ultimate causes of trends are uncertain. Well-constrained examples often involve climatic drivers because of the abundance of climate-related information preserved in the geological record. Biotic interactions, by contrast, usually leave few detectable traces in the fossil record, and so they are much more difficult to evaluate as putative drivers of trends.

Because paleontological trends typically unfold over millions of years, even the most dramatic trends can be generated by very small biases acting in each generation. One can show, for example, that even the largest evolutionary transitions within single traits in horse evolution can be produced by just a few selective deaths per million individuals per generation. This intensity of selection is very weak—much weaker than genetic drift—and so it is probable that macroevolutionary trends, when driven by natural selection, unfold episodically in response to local and changing conditions, rather than uniformly over time.

In a classic paper, Steven M. Stanley proposed another scenario that can generate a trend via accumulated microevolution. Stanley was interested in explaining Cope’s rule, but he did not assume universal advantages for larger bodies. Instead, Stanley suggested that there may be a lower limit to body size below which evolution is unlikely to explore. For endothermic vertebrates, this lower limit might arise from difficulties in maintaining body temperature at very small sizes, but the proximate biological reasons for these limits would depend on the relevant taxon’s suite of adaptations and constraints. Moreover, Stanley argued that smaller organisms are, on average, less morphologically specialized than larger organisms, and thus they should preferentially found higher taxa. Starting at a small size near a lower boundary will result in an asymmetrical expansion in body sizes: maximum and mean sizes increase as species diverge over time, but minimum size stays stable because it cannot decrease below the lower limit. Thus, a trend is observed, even though increases and decreases are equally likely except in the neighborhood of a boundary.

This kind of trend has been called “passive” because it arises from diffusion-like evolution in the presence of a boundary; “active” or “driven” trends are those characterized by a uniform bias in the direction of evolutionary change. Passive and active trends can be distinguished empirically by their different effects on maxima and minima, and by comparing the dynamics of species near and far from the putative bound.

Trends from Species Selection

Though it might seem counterintuitive, it is possible to generate clade-level trends even in the absence of directional microevolutionary changes. All that is required is that the trait have a systematic relationship with net rates of species diversification. For example, if large-bodied species proliferate more successfully than small-bodied species, then, over time, that fauna will become increasingly dominated by large-bodied species, and as a consequence, that clade’s mean size will increase steadily. This Cope’s rule pattern will hold even when body size increases and decreases are equally frequent within species. This hypothetical trend is generated not by biased transformation of species but rather by biases in their proliferation, a process called species selection (see chapter VI.14). Just as natural selection follows from variation in reproductive success among individual organisms, so does species selection result from variation among species in their propensities to persist and generate “offspring” lineages through speciation. Accordingly, the traits most likely to be important for species selection are those that influence, directly or indirectly, the rates at which species form or the rates at which they become extinct.

Although species selection is not applicable to trends within a single species, most workers agree that it is a plausible mechanism for clade-wide trends. Its empirical importance, however, is much less clear. The main problem is that assessing a trait’s influence on speciation or extinction requires a large set of species of known relationships. But the fossil record is less complete and robust at the species level, and so rates of extinction and origination are usually estimated for genera instead. However, high-quality, species-level data sets are starting to become more common, which should allow for better tests of species selection as a driver of trends. Another way around this obstacle may be afforded by recent methodological advances that permit one to assess the influence of traits on speciation and extinction rates using phylogenies of extant species, with no fossil record required. At present, rather large trees are needed to obtain reliable results, but the explosive growth of molecular phylogenetics has made such trees increasingly available.

At the broadest level, it is worth noting that trends related to the waxing and waning of distinct clades must be driven by differences in diversification among the clades. Sometimes, these diversification differentials are consistent over long periods. For example, the current dominance of flowering plants on land is a consequence of their elevated diversification rates compared with those of other vascular plants. These differences in diversification have persisted since late in the Cretaceous, roughly 100 million years ago. At the other end of the continuum, single extreme events such as mass extinctions can trigger permanent shifts in the dominance of different taxa, as in the shift in terrestrial communities resulting from the extinction of nonavian dinosaurs and other large vertebrates during the end-Cretaceous mass extinction.

4. EXAMPLES OF TREND HYPOTHESES

Cope’s Rule in Mammals

Cope’s rule has been studied quite broadly, but it has been particularly emphasized in vertebrate paleontology. The enormous sizes achieved by many dinosaur lineages have led to suggestions of Cope’s rule for this group, but mammals are the fossil group that has been analyzed most extensively for body size trends. The most comprehensive analysis to date is one by John Alroy, who tracked body mass (estimated from tooth dimensions) in North American mammals over the past 80 million years. By looking at the difference between putative ancestor and descendant species pairs, Alroy was able to demonstrate a bias toward body size increases: on average, descendants were about 9 percent larger than their ancestors. The trend mechanism here is therefore a bias in microevolutionary changes within species, rather than differential sorting among species.

What ultimately explains this preferential size increase within lineages? For a group as diverse and heterogeneous as mammals, it is likely that multiple factors have been important. Fossil horses offer perhaps the best case study because of their richly documented fossil record, especially from North America. The earliest-appearing horses were relatively small, estimated to weigh 30 kg or less, and horses remained at about this size for the next 25 million years or so. Then, average body size among horse lineages increased rapidly following a climate-driven expansion of open, grass-dominated habitats. These body size shifts were associated with changes in other features interpreted as adaptations to grasslands, such as high-crowned teeth (which are better able to withstand high-grit, grassy diets). Thus, the timing of the body size increases, along with their environmental context, is consistent with large-scale shifts in the habitats as the driver of this trend.

Organismal Complexity

Perhaps no trend hypothesis has had a firmer hold on scientists’ imagination than the idea that life has progressed from simple to complex. The fossil record has long been marshaled in support of this notion because the first organisms to appear in the fossil record are single-celled prokaryotes, a form of life that presumably exists near a lower boundary for organismal complexity. Moreover, for its first 2 billion years, the fossil record of life consists entirely of structurally simple microscopic organisms. Evidence for animals, possibly sponges, first appears around 800 million years ago, but it is not until the Cambrian period more than 250 million years later that complex bilaterian animals diversify.

The most persistent obstacle to evaluating this trend hypothesis is pinning down what exactly it means for an organism to be complex. Daniel W. McShea has argued that the criteria used to identify complexity have been loose and impressionistic, sometimes reflecting only gross similarity to Homo sapiens. According to McShea, complexity comes in different forms, but the type most amenable to testing in the fossil record is related to the number and diversity of subcomponents, or parts, that constitute an organism. The more parts and kinds of parts an organism bears, the more complex it is. This concept is made operational by deciding what most usefully constitutes a part in the system under study.

Perhaps the most general attempts to evaluate animal complexity involve counting as parts the number of distinct cell types in different animal groups. Merging these indicators of complexity with information on the first fossil appearances of those groups suggests an increase in maximum animal complexity over time. Animal taxa with few cell types, such as sponges and cnidarians, appear earlier in the fossil record than groups like birds and mammals that today have many different kinds of cells. These data are sufficiently uncertain, however, that it is difficult to establish with precision the timing and pattern of the increase or even if average complexity follows the same trajectory as maximum complexity. In part, this uncertainty stems from ambiguities in enumerating distinct cell types across taxa that vary greatly in how well their microanatomy has been studied; with greater investigation, more cell types will be discovered, and finer categorization of cells into types will be possible. Emerging comparative genomic data, when similarly mapped to the fossil record, may permit more unequivocal tracking of the trajectory of animal complexity through time.

Escalation

The hypothesis of escalation, as articulated by Geerat J. Vermeij, suggests an overarching ecological directionality to the history of life, in which organisms have increased in their ability to acquire and control resources. This trend is said to be episodic rather than uniform over time, but the net result is a biotic environment that has become more and more perilous to organisms as they confront increasingly dangerous predators, competitors, and prey. This trend is detectable in the fossil record by tracking the origin and frequency of adaptations related to this biological arms race, such as arthropod limb modifications for crushing shelly prey and the complementary features of shelled organisms that defend against such attacks.

The scope of this hypothesized trend is broad, encompassing whole faunas of interacting organisms. Biased microevolution and species selection are thought to act jointly to generate the escalation trend. Microevolutionary transitions are inferred to preferentially favor behaviors and morphologies that are more active and energy intensive. Species selection may also contribute to this trend when escalated species radiate, and poorly defended forms preferentially go extinct.

Definitive tests of the escalation hypothesis are complicated by the diversity of escalation predictions and by uncertainty about the relative energy intensiveness of morphologies and modes of life in extinct taxa. Moreover, because increasingly dangerous modes of life are thought to incur steep energetic demands, there might be temporal and spatial refuges—intervals and environments in which resource limitation slows or reverses the normal tide of escalation. Indeed, empirical studies that address aspects of escalation have produced a variety of results, not all of which are consistent in their timing, coordination, or magnitudes of ecological change. Nevertheless, over suitably long periods, many studies have found net positive trajectories in escalation-related measures, including increasing body size (mostly in the early Paleozoic), increasing dominance of active predators and other mobile lifestyles, diversification of shell-crushing predators (especially in the middle Paleozoic and late Mesozoic), increasing frequency of traces of shell predation (drill holes and repair scars), increasing depth of burrowing (presumably an escape from predation), increasing frequency of morphologies interpreted to defend against predators (e.g., narrow apertures in snails), and a corresponding decrease in susceptible forms.