33 Immeasurable Progress on the Tree of Life

Back

Michael J. Donoghue

548

In listening to the Assembling the Tree of Life (ATOL) symposium

in New York, and in reading the manuscripts for this

volume, I was overwhelmed by the enormous progress that

we have made, over such a short time, on what Darwin so

aptly called “the great Tree of Life.” The word “immeasurable”—

in the dictionary sense of “indefinitely extensive”—

seems to apply perfectly to this situation. But what about the

other, more literal, meaning of the word immeasurable? Is

phylogenetic progress also “incapable of being measured”?

This is the question I want to address. My sense is that there

are many facets of “progress” that matter to us and that we

would like to be able to measure. For some of these we can

devise proper metrics, and we might even be able to provide

concrete numbers. For others, as I’ll argue, we aren’t even

entirely sure what we’d like to measure, and we’re still a long

way from being able to quantify how we are doing.

Let me back up, and ask, What are the ways we might

think about expressing progress—to measure where we stand

now in relation to where we were a decade ago and where

we hope to end up? One possibility would be to tally the

number of known species on Earth that have been included

in bone fide phylogenetic analyses [in December 2003 there

were almost 35,000 species represented in TreeBASE (available

at http://www.treebase.org), but the real number might

be more like 80,000], or maybe even the number that could

potentially be included today if we harnessed all of the data

in relevant databases [e.g., DNA sequences in GenBank (available

at http://www.ncbi.nlm.nih.gov)]. Another possibility

would be to chart trends in the number of phylogenetic papers

published over the years (e.g., Sanderson et al. 1993;

Hillis, ch. 32 in this vol.).

These are certainly interesting measures, and the numbers,

insofar as we know them, certainly do bolster the gutlevel

feeling that we’re making lots of progress. They don’t,

however, capture much about the nature and the quality of

what’s being learned. Maybe we should also be gauging our

coverage of the Tree of Life in terms of the number of major

lineages represented by some reasonable number of exemplars,

or perhaps we should somehow represent the size and

the variety of the data sets that are being analyzed. Or, perhaps

a metric is needed to reflect changing levels of confidence

in the clades being identified. Another worthy measure,

for very obvious purposes, would gauge how many phylogenetic

studies have provided solutions to practical problems.

Success stories along these lines abound—identifying the

source of an emerging infectious disease, pointing the way

toward crop improvement, orienting the search for new

pharmaceuticals, and so on (see Yates et al., ch. 1 in this vol.;

examples of the practical importance of phylogenetic research

are also highlighted in a brochure sponsored by the National

Science Foundation (Cracraft et al. 2002). But how do we

attach a number to such achievements? Patents pending,

perhaps, although this would record only a small fraction of

the successes.

Ultimately, I think we would all like a measure that captures

how phylogenetic studies have affected our understandImmeasurable

Progress on the Tree of Life 549

ing of life—how the living world is structured, how it works,

and how it has come about. At first glance this truly does

seems immeasurable, in the “not-capable-of-measurement”

sense of the word. But on second thought, maybe there is a

reasonably good proxy for this, which takes us back to Willi

Hennig (e.g., Hennig 1966). What if we could faithfully tally

up cases in which traditionally recognized taxonomic groups

had been convincingly demonstrated to be paraphyletic?

Paraphyletic groups are ones that contain an (inferred) ancestor

and some, but not all, of its descendants. In practice,

of course, paraphyly is “discovered” when a phylogenetic

analysis identifies one or more new clades that unite some

of the lineages previously assigned to the traditional group

with one or more lineages placed outside of that group. In

other words, the “negative” discovery of paraphyly is precisely

the “positive” discovery of new “cross-cutting” clades.

Before we think about whether we could actually count

up discoveries of paraphyly, let’s contemplate why this might

be a satisfying measure of phylogenetic progress. First of all,

it’s worth noting that this measure relates how changes in

our knowledge of phylogenetic relationships have affected

the application of taxonomic names, and as such, it can potentially

be assessed everywhere in the Tree of Life, from the

very base out to the tips, without needing to refer to particular

groups or their characters. In this sense, it is a measure without

units. Second, it registers a change in the language that

we use to describe the structure of diversity, which can deeply

(although often quite subtly) influence the way we perceive

diversity, orient our research, and teach. Third, the discovery

of paraphyly has immediate impacts on our understanding

of character evolution. Some characters previously

thought to have evolved convergently are seen instead to be

homologous—to have evolved only once, in the inferred

ancestor of a newly discovered cross-cutting clade. Even more

generally, the recognition of paraphyly allows us to infer a

sequence of evolutionary events, which helps fill in what

appeared to be major gaps between traditional taxa. Often

this is just the information we need to choose among competing

evolutionary hypotheses about how and why major

transitions occurred. In many of the same ways, of course,

such discoveries also help us make sense of biogeography.

Fourth, such discoveries generally change the way we perceive

shifts in diversification, especially by accentuating differences

in the number of species between sister groups.

Putting the third and fourth points together, my guess is

that discoveries of paraphyly will eventually have even more

profound impacts on how we view the connection between

character change and diversification. In particular, I think

we’ll be forced to develop a more nuanced (and more productive)

view of “key innovations.” It will become increasingly

natural to think from the outset about a series of

changes culminating in a combination of traits that ultimately

affected diversification. Rather than simply moving the causal

explanation down a node or two in the phylogeny, this distributes

the causation across a series of nodes and character

changes. Also, increasingly we’ll focus on how apparently

subtle changes early in such a chain rendered new morphological

designs accessible, which in turn enabled the evolution

of the traits that we most often associate with the success

of clades, with ecological transitions, and so forth.

To illustrate these points, let’s look at green plants. Figure

33.1 provides an overview of our present knowledge of

phylogenetic relationships among the major lineages—highly

simplified, of course, and consciously pruned (rendered

pectinate) to serve my purposes (see O’Hara 1992 for a general

discussion of such simplifications). Several widely known

traditional groups are supported as monophyletic in all recent

analyses, including the entire green plant clade (viridophytes),

land plants (embryophytes), vascular plants

(tracheophytes), seed plants (spermatophytes), flowering

plants (angiosperms), and monocotyledons (monocots). A

number of other traditionally recognized groups have repeatedly

been determined to be paraphyletic, confirming suspicions

that they represent grades of organization, diagnosed

only by ancestral features of the more inclusive clades to

which they belong. Specifically, “green algae,” “bryophytes,”

“pteridophytes,” “gymnosperms,” and “dicotyledons” all

appear to be paraphyletic. In each case, one or more new

clades were discovered that linked some lineages traditionally

assigned to the group to related taxa. So, for example,

the streptophyte and charophyte clades (as circumscribed

here; for an alternative, see Delwiche et al., ch. 9 in this vol.)

include lineages that used to be assigned to the green algae

(the Charophyta in the traditional sense) along with the land

plant clade. Likewise, the euphyllophyte clade unites all extant

lineages of seedless vascular plants, except the lycophytes,

with the seed plants, and so on. In the case of the

“bryophytes” and the “gymnosperms,” names were proposed

for new cross-cutting clades (“stomatophytes” and “anthophytes,”

respectively), but recent analyses have cast doubt

on their existence (see Nickrent et al. 2000, Donoghue and

Doyle 2000). Nevertheless, in both cases it remains quite clear

that these traditional groups are paraphyletic (see Delwiche

et al., ch. 9, and Pryer et al., ch. 10 in this vol.).

The impact of these discoveries on our understanding has

been enormous. The most obvious and immediate effect was

on our ability to dissect the evolutionary sequence of events

surrounding the greatest transformations in plant history. For

example, take the transition from living in water to living on

land (see Graham 1993). Before we recognized the paraphyly

of green algae and of bryophytes, this shift appeared to entail

a large number of steps, which we had no real basis for

putting in order. This implied either many extinctions and,

consequently, gaps in our knowledge, or else some sort of

wholesale transformation from one life form to another.

Under these circumstances, alternative theories emerged and

remained viable. What kind of environment did the immediate

ancestors of the land plants live in, and what did they

look like? After all, “green algae” live in saltwater or in freshwater;

may be unicells, colonies, filaments, or more complex

550 Perspectives on the Tree of Life

Figure 33.1. An overview of green plant phylogeny, illustrating progress through the recognition

and abandonment of paraphyletic groups (e.g., “green algae” and “bryophytes”) with the discovery

of new major clades (e.g., streptophytes and euphyllophytes). For references to the primary

literature, underlying evidence, levels of support, outstanding controversies, and additional

evolutionary implications, see Kenrick and Crane (1997), Doyle (1998), Donoghue (2002), Judd et

al. (2002, ch. 7), and chapters 9–11 in this volume. Note that Delwiche et al. (ch. 9 in this vol.; also

Karol et al. 2001) use the name “Charophyta” for the clade here referred to as the streptophytes. The

usage adopted here may better reflect original intentions (e.g., Bremer and Wanntorp 1981) and

subsequent usage (e.g., Kenrick and Crane 1997); in any case, such nomenclatural problems

highlight the desirability of providing explicit phylogenetic definitions for clade names.

Immeasurable Progress on the Tree of Life 551

forms; may or may not have cell walls separating the nuclei;

and so on. And what about the evolution of the land plant

life cycle—alternating between multicellular haploid (gametophyte)

and diploid (sporophyte) phases? In short, the transition

to land largely remained a mystery.

With the discovery of a series of intervening clades (fig.

33.1; Karol et al. 2001; see Delwiche et al., ch. 9 in this vol.),

we’re now able to infer a sequence of events from the first

green plants through the transition to land. We can be quite

certain that their immediate ancestors lived in freshwater,

probably quite close to the shore; had rather complex parenchymatous

construction; and bore eggs (and zygotes) on

the parent plant in specialized containers. Likewise, we can

finally put to rest the debate about the life cycle: the land plant

life cycle originated through the intercalation of a multicellular

diploid phase (through delayed meiosis) into an

ancestral life cycle in which the diploid zygote underwent

meiosis directly to yield haploid spores.

This example is meant only to illustrate the sorts of insights

that can follow the discovery of paraphyly, and so to

justify such a measure of progress. What can we say, then,

about the number of these discoveries in recent years, or

about our expectations in the future? In The Hierarchy of Life

(Fernholm et al. 1989), the last major attempt to take stock

of phylogenetic progress, Gareth Nelson remarked: “Paraphyly,

it would seem, is the most common discovery of

modern systematic research” (Nelson 1989: 326). This may

well be true, but is there a way to put a number on it? Sadly,

aside from asking experts on each major clade to come up

with a list (or an account along the lines of fig. 33.1), we aren’t

really able to do this. We haven’t been keeping track in any

systematic way and, as I will argue, we haven’t developed the

necessary informatics tools.

Let us suppose that we wanted to be able to tally up those

changes in knowledge of phylogeny that significantly

changed our view of the world, and that for this purpose we

wanted to focus on discoveries that changed the way that

taxonomic names are used. Specifically, we would be looking

for cases in which the name of a paraphyletic group had

been abandoned altogether, or the circumscription had been

adjusted so that the name again referred to a hypothesized

clade. These are what might be called “meaningful” taxonomic

changes, to distinguish them from other sorts of name

changes. We would want to avoid, for example, changes only

in the Linnaean taxonomic rank that a group is assigned (e.g.,

a shift from Family to Order). As things now stand, such rank

assignments are fundamentally arbitrary, yet our nomenclatural

codes are intimately tied to them, and in some cases a

cascade of name changes can be required without any underlying

advance in our knowledge of phylogeny. Also, it’s

important to note that quite a few clades are discovered and

named that don’t contradict the monophyly of any previously

named taxon—instead, they resolve bits of the Tree of Life

that were more or less unresolved and to which taxonomic

names had not been applied. The point is that the problem

is not as “simple” as just tracking changes in the names being

used in the taxonomic literature.

What we really are talking about is tracking changes in

the relationship between taxonomic names and hypothesized

clades. If we knew how taxonomic names mapped onto a tree

at some initial time, we could see at a later time how many

names applied to the same clades versus how many no longer

applied to clades but to paraphyletic groups. To do this in

practice, one would need, first of all, a database that recorded

changes in our knowledge of phylogeny. TreeBASE is designed

for this purpose, but unfortunately, it still isn’t used

consistently enough by the authors of phylogenetic papers.

One presumes that this will improve (probably driven by

more journals requiring the submission of phylogenetic data

and results), in which case we will automatically develop the

record we need to make solid tree comparisons over time.

But tracking trees is only one part of the problem. The

other is to understand how names have been used at different

times. Although for some groups of organisms there are

databases that keep track of all the names that have ever been

published (e.g., the International Plant Names Index, available

at http://www.ipni.org), or even of the accepted names

and synonyms (e.g., Species 2000, available at http://www.

sp2000.org), it’s hard to say exactly how these names correspond

to hypothesized clades at any one time, much less at

different times. The problem is that taxonomic names have

not traditionally been defined in such a way that we can be

sure whether they were even meant to refer to clades (sometimes,

mostly in the past, names were knowingly applied to

paraphyletic groups) or, if so, which lineages were intended

to be included (even assuming complete agreement on phylogenetic

relationships). Of course, we could get better about

designating how names are meant to coincide with clades by,

for example, consistently labeling clades in TreeBASE. This

would be a step in the right direction, but it would be even

better to adopt a nomenclatural system in which the connection

between a taxonomic name and a hypothesized clade

needed to be precisely defined at the outset. Here I am referring

to “node-based” and “stem-based” definitions and other

conventions discussed in relation to the PhyloCode (available

at http://www.phylocode.org). Interestingly, taxonomic

names under such a system tend to be maintained in the face

of changes in phylogenetic knowledge, although with a different

composition of lineages. Specifically, the name of a

taxon discovered to be paraphyletic might well be retained

for a more inclusive clade, unless it happened to become

synonymous with a preexisting name. Overall, it is hard to

say how the turnover of names would compare between the

PhyloCode and our traditional nomenclature codes, where

names are neither defined with respect to a tree nor fixed in

terms of content.

The conclusion I draw from the above is that the actual

abandonment of the names of paraphyletic groups is probably

not going to be a very sensitive measure (under either

traditional nomenclature or under the PhyloCode). Names

552 Perspectives on the Tree of Life

can be retained and reconfigured in various ways, and in any

case it would be hard to judge when a particular name had

finally been dropped by the relevant taxonomic community.

In the end, what we really want, regardless of “abandonment,”

is a database designed such that we can identify those phylogenetic

discoveries that change how names map onto trees—

whether a name refers to the same clade at different times or

whether it can be made to refer to a clade only by changing

the content to include lineages previously viewed as being

outside the group. This would be a pretty sophisticated database,

but I see no reason why it couldn’t be developed.

My point is that it’s time we attended to the business of

naming clades and to the informatics issues surrounding the

Tree of Life project. As Hennig stressed, “Investigation of

the phylogenetic relationship between all existing species and

the expression of the results of this research, in a form which

cannot be misunderstood, is the task of phylogenetic systematics”

(Hennig 1965: 97). Progress on the first of these goals—

understanding phylogenetic relationships—has certainly

been impressive. By comparison, progress on the second

goal—expressing the results in a form that cannot be misunderstood—

has been rather pathetic. Much of what we

have learned about relationships has not been translated into

the taxonomic language used to describe the diversity of Life.

And much of what we have learned has not been properly

incorporated into databases, so the effort is effectively wasted.

I hope we have made real progress along these lines before

we take stock again of the Tree of Life.

In summary, at this moment it strikes me that phylogenetic

progress is immeasurable in both senses of the word—

phylogenetic knowledge is expanding at a mind-boggling

rate and we don’t yet have the tools to measure this in the

ways we would like. When we are eventually able to make

measurements of the sort I have described, we will have

achieved something truly monumental. We will certainly

have charted much more of the Tree of Life, but we will also

have changed the language we use to communicate about

biological diversity and, therefore, how we think about the

world. Perhaps most important, we will have rendered this

knowledge widely accessible and prepared it for the queries

that will propel the Tree of Life project to the next level.

“Indefinitely extensive” will have become the only applicable

meaning of “immeasurable.”

Acknowledgments

I am grateful to Joel Cracraft for his leading role in organizing

the symposium and editing the proceedings, and to the other

speakers in the session on plants—Chuck Delwiche, Kathleen

Pryer, and Pam Soltis. I have benefited from discussion of these

issues with Susan Donoghue and Kevin de Queiroz. For their

help with my presentation at the symposium and with figure

33.1, I am indebted to Brian Moore and Mary Walsh. Yale

University, through Provost Alison Richard, generously

supported the symposium and the participation of Yale

students.

Literature Cited

Bremer, K., and H.-E. Wanntorp. 1981. A cladistic classification

of green plants. Nord. J. Bot. 1:1–3.

Cracraft, J., M. Donoghue, J. Dragoo, D. Hillis, and T. Yates

(eds.). 2002. Assembling the tree of life: harnessing life’s

history to benefit science and society. National Science

Foundation. Available: http://ucjeps.berkeley.edu/tol.pdf.

Last accessed 25 December 2003.

Donoghue, M. J. 2002. Plants. Pp. 911–918 in Encyclopedia of

evolution (M. Pagel, ed.), vol. 2. Oxford University Press,

Oxford.

Donoghue, M. J., and J. A. Doyle. 2000. Demise of the anthophyte

hypothesis? Curr. Biol. 10:R106–R109.

Doyle, J. A. 1998. Phylogeny of the vascular plants. Annu. Rev.

Ecol. Syst. 29:567–599.

Fernholm, B., K. Bremer, and H. Jцrnvall (eds.). 1989. The

hierarchy of life. Nobel Symposium 70. Elsevier, Amsterdam.

Graham, L. E. 1993. Origin of the land plants. Wiley, New

York.

Hennig, W. 1965. Phylogenetic systematics. Annu. Rev.

Entomol. 10:97–116.

Hennig, W. 1966. Phylogenetic systematics. University of

Illinois Press, Champaign-Urbana.

Judd, W. S., C. S. Campbell, E. A. Kellogg, P. F. Stevens, and

M. J. Donoghue. 2002. Plant systematics: a phylogenetic

approach. 2nd ed. Sinauer, Sunderland, MA.

Karol, K. G., R. M. McCourt, M. T. Cimino, and C. F. Delwiche.

2001. The closest living relatives of land plants. Science

294:2351–2353.

Kenrick, P, and P. R. Crane. 1997. The origin and early

diversification of land plants: a cladistic study. Smithsonian

Institution Press, Washington, DC.

Nelson, G. 1989. Phylogeny of the major fish groups. Pp. 325–

336 in The hierarchy of life (B. Fernholm, K. Bremer, and H.

Jцrnvall, eds.). Nobel Symposium 70. Elsevier, Amsterdam.

Nickrent, D., C. L. Parkinson, J. D. Palmer, and R. J. Duff. 2000.

Multigene phylogeny of land plants with special reference to

bryophytes and the earliest land plants. Mol. Biol. Evol.

17:1885–1895.

O’Hara, R. J. 1992. Telling the tree: narrative representation and

the study of evolutionary history. Biol. Philos. 7:135–160.

Sanderson, M. J., B. G. Baldwin, G. Bharathan, C. S. Campbell,

D. Ferguson, J. M. Porter, C. Von Dohlen, M. F.

Wojciechowski, and M. J. Donoghue. 1993. The growth of

phylogenetic information and the need for a phylogenetic

database. Syst. Biol. 42:562–568.

Навигация