DNA barcodes, morphology, and a damned dog
Making species identifiable is not making species known
“Today, bio-informaticians conjure up enormous genealogical trees based on thousands of genomes and thousands of organisms… Their efforts, impressive though they are, are merely technology-driven, rather than concept-driven. This leaves their results almost devoid of any scientific content beyond a simplistic interpretation of the results.”—David Williams, 2020
“Willie Sutton robbed banks because that’s where the money is kept, and taxonomists study morphology because that’s where many of the most interesting evolutionary novelties are found.”—Quentin Wheeler, 2020
Recent, “minimalist” taxonomy has a very different, far less ambitious, aim than traditional taxonomy. For centuries, taxonomists have explored the biosphere on a mission to discover what kinds of living things exist on planet earth, to determine what combination of attributes makes each unique, to make sense of the pattern of similarities and differences among them, to understand the history of transformations responsible for their diversity, and to organize knowledge in a “natural” classification, understood today to mean one that reflects phylogenetic relationships. Minimalist taxonomy, and its primary source of information, a DNA barcode, seek only to make species identifiable. It is the manifestation of a non-taxonomist’s self-serving idea of what taxonomy should be: a set of procedures, as simple, cheap, easy, and routine as it can be made, by which species may be recognized. It seems not to bother users of taxonomic information that divorcing species exploration from testable hypotheses and rigorous studies of characters makes species identifications less reliable and less informative.
An Australian livestock guardian dog protecting its flock of sheep. The relevance of this photo will become apparent later. Source: Wikipedia. Credit: Andy Fitzsimon, Brisbane, “protector of the sheep.” CC BY-SA-2.0
I am all for facilitating the identification of species for the life sciences and society. I appreciate the necessity of accurate IDs, and taxonomy’s long, proud tradition of making them possible. Two things, however, should be kept in mind. First, the most informative identifications are the result of taxonomy pursued to its own high standards; identifications are most reliable when they are based on the application of knowledge—and it is knowledge, not making identifications, that is the aim of taxonomy. Second, taxonomy, done to high standards of excellence, is cognizant of appropriate theories, reliable methods, as much relevant evidence as possible, and rigorously testable hypotheses. Taxonomy does not exist to identify species any more than chemistry exists to identify mystery fluids in unlabeled bottles. IDs are useful applications of knowledge in each case, but they do not reflect the depth, potential, or motives of either science.
When Aristotle squashed a bug and observed that it oozed a liquid other than red blood, like that of a mammal, he wondered why. It was the presence, or absence, of a character, like red blood or a backbone, or a pair of wings, among kinds of organisms that was intriguing to him. To this day, it is diverse and unexpected characters that intrigue, as well as the pattern of similarities and differences among species. For most groups of plants and animals, morphological characters are among the most numerous and obvious ones . You cannot be truly interested in species and not be equally interested in characters.
Taxonomy exists to explore and interpret this pattern. Curiosity to understand characters, their distributions among species, their historical origins, and their precursors are at the heart of taxonomy. In most cases, diverse, novel, and complex characters are what make species fascinating. Simply being able to tell them apart, with, for example, a DNA barcode, ignores that which makes species most interesting and worthy of study.
The challenge to make sense of the diversification of millions of species—and of an even greater number of characters—is undiminished by the availability of DNA data. DNA is useful for identifications and cladogram construction and, eventually, will be woven into the fabric of character analysis by deepening understanding of the connections between genes, developmental pathways, and morphological structures. Regardless of the source of evidence, whether morphological, ontogenetic, or molecular, the motive behind taxonomy, and its prime directive as a science, remain the same: to know, interpret and understand species and characters, and the relationships among them—not to make identifications. Identifications may be the goal of DNA barcoding, but IDs are most reliable and informative when approached as applications of credible knowledge, not as ends in themselves.
Backyard stargazers take pleasure observing well-known celestial objects, from nearby planets to remote constellations. But to arbitrarily limit astronomy to identifying dots of light in the night sky would be an act of scientific sabotage. Given recent, exciting discoveries by astronomers, from refined theories of the origin of the universe to spectacular images returned by the Webb telescope, I pity anyone with so little curiosity, so little sense of wonder, as to think restricting funding for astronomy—or taxonomy—to immediate, practical applications would be a good idea. It would be an affront to the intellectual curiosity that elevates humankind, and an insult to the traditions of science. Astronomy, like taxonomy, exists to discover what exists and the properties of, origin, and relationships among the objects discovered.
Yet, we go along with geneticists and environmental biologists who, selfishly, want to restrict taxonomy to a single source of data so that species might be identified cheaper, easier, and faster. On average, taxonomists describe about 20,000 newly discovered species of plants and animals each year, and there is no end in sight. Among them are spectacular, entirely unanticipated discoveries that fill gaps in, or add unsuspected branches to, the tree of life, that possess characters of unimagined novelty and complexity, and that expand our appreciation for, and understanding of, biodiversity, biosphere and evolutionary history. Only an incurious fool, blinded by their own narrow, selfish wants, would knowingly trade fundamental exploration of species, characters, and history for quick-and-easy identifications.
Since Linnaeus, taxonomists have been exploring, documenting, describing, analyzing, naming and classifying the kinds of living things with which we share planet earth. Even after centuries of spectacular progress, including the discovery of about two million species, their work has only begun. It is estimated that eighty percent, perhaps more, of all species remain unknown to science, and attributes of many named species are only partially described. From brief, one-sentence diagnoses in the time of Linnaeus, taxonomy advanced and matured as a science, raising its standards to synthesize diverse evidence and publish comparative, highly detailed, sometimes lavishly illustrated, descriptions of species, appropriately heavily based in morphology, and most fully developed in the context of comprehensive revisions and monographs. But these advances are now being reversed by DNA-based studies that would return us to 18th-century style, minimally informative, diagnoses based on one source of data. That such minimal descriptions are based on molecular rather than morphological features in no way mitigates the tragic loss of knowledge. No other organized scientific effort exists to decrease, rather than increase, how much we learn.
As if self-imposed ignorance of the attributes of species were not enough, DNA studies seem intent, also, on smothering curiosity about details of the diversification and history of species. Many of the most exciting, spectacular, wholly unexpected attributes of species are found among morphological characters. No one interested in species themselves, or evolutionary history, sees a shift from descriptions to DNA barcodes as progress. A sequenced strand of DNA is qualitatively different from the observation and interpretation of complex and novel structures, like rhinoceros horns or giraffe necks. We could identify the planets in our solar system with precise measurements of their diameters or their distances from the Sun, but without knowledge of their surfaces, composition, temperatures, and atmospheres (if they have one), simply telling them apart would have very little meaning. This, however, is equivalent to the supposed advance being offered by DNA barcoding.
Spilling paint from a bottle onto a canvas isn’t art, it’s an accident. This almost-in-focus picture, showing a single angle of the type specimen of Orgilus marlinricei, is the full extent of the “accidental description” of its morphology by Sharkey et al. (2021). There is no interpretation or discussion of any relevant character information that may, or may not, exist in the photo. A consensus barcode is presented, but this is data, not knowledge. A proper morphological description, in contrast, would include hypotheses related to characters and character distributions arrived at following detailed, comparative studies, of this and related species. The photo seems to be window dressing, as nothing of evolutionary interest or diagnostic value about the specimen is pointed out by the authors. To say this species is now “known,” in any meaningful sense, because it has a name and a DNA barcode, makes a mockery of taxonomy. This may reflect what ecologists think taxonomists are, or should be, doing, but it doesn’t measure up to 18th century, one-line diagnoses—much less standards of modern taxonomy. I wonder what the authors would think of an ecology study done to a comparably superficial level, or the results of an experiment published without being replicated even a single time; neither would pass peer review. This “description” leaves almost all of the work, and most of what may be interesting about this species, for someone in the future to determine and make known. It is far more responsible and informative to describe a species as fully as possible the first time, instead of leaving it for others to mop up after a superficial treatment. When all species in a clade are carefully studied, as in a traditional revision or monograph, it is possible to describe each to a much higher standard, even those known from single specimens. Photo: Sharkey et al., 2021, figure 337.
It is easy to take knowledge of morphology for granted because so many of the most commonly encountered species were described long ago. But keep in mind that we do not yet know anything of the morphology of the vast majority of species, beyond what we can predict, in a general way, from cladograms and phylogenetic classifications. We can predict that any newly discovered vertebrate animal will share the character “backbone,” but we cannot know exactly what its backbone looks like; backbones of an ostrich, whale, and giraffe differ significantly in detail, and details matter.
Limiting taxonomy to DNA data is, at best, like printing a Dewey decimal number on the spine of each book in a great library. Any book may be quickly and accurately identified. But if the pages of every book have been glued together, how valuable is knowing which book is which? Recognizing a book, while learning nothing of its content, of the words and ideas it contains, is setting the bar rather low. I suppose there is some utility for those who only wish to tell books apart and know what to call them, but this mostly vacuous exercise ignores the best reasons for having libraries in the first place. Something very much like this emaciated library will be the legacy of a DNA-based, minimalist taxonomy.
The argument that species are going extinct rapidly and that, therefore, we ought to lower our standards in order to name them more quickly is bizarre. Being able to identify species about which almost nothing is known is as close to useless as taxonomy could be made. Further, as Broad (2025) has argued, successfully, the supposed time savings of DNA barcodes in the place of descriptions of morphology are not substantiated by the evidence. Minimalist taxonomy shows no significant gain in efficiency over descriptive morphology. And this is so with existing taxonomic infrastructure. Were we to invest in modernizing the tools available for storing, accessing, studying, and documenting specimens, and invest in support staff for taxon experts, morphology descriptions could be made far more efficient. One is forced to conclude that the preference for DNA over morphology is indicative of something else: a sad lack of curiosity about species among non-taxonomists; an eagerness to conform to fads and fashions by adopting popular technologies; a selfish willingness to meet one’s own immediate needs at the expense of deep understanding; or, perhaps, it is simply greed, taking the easy path to grant money, instead of the harder one to knowledge. Let me add that, even if morphology-based descriptions took twice as long as DNA barcodes, it would be worth it because they result in far more than twice the information.
Beyond the scientific knowledge, and intellectual satisfaction, derived from studying morphological diversity and its evolutionary history, there is much to be said for being able to identify species by eye, in the field, when that is possible. Recognizing species on sight can be deeply rewarding in itself; just ask anyone birdwatching, fishing, or searching for wild mushrooms. It is extremely useful to ecologists, conservation biologists, and agriculturalists, among many others, as they do their work. And, for taxonomists searching for new species, it is a vital time management tool. A dipterist, for example, does not waste time collecting and randomly sequencing thousands of house flies because Musca domestica is easily recognized by morphology allowing collecting efforts to be focused elsewhere.
As I have suggested, it is impossible to be interested in species themselves, or the history of evolution, and not be equally interested in characters, including, in most cases, morphology. Beyond distinguishing species, and providing evidence of relationships and history, morphology has practical value. This was the point of a joke—slightly updated—told to me years ago by my doctoral advisor, Chuck Triplehorn.
Driving a backroad, a man brought his car to a full stop as a flock of sheep were ushered across his path, with great chaos, in a billowing cloud of dust. As the last sheep passed, the driver called out to the shepherd, “Excuse me, if I can tell you how many sheep are in your flock, may I have one?”
The shepherd, thinking a tally impossible amid so much confusion, agreed.
“You have two-hundred and thirty-seven sheep.”
“That’s amazing!,” said the shepherd, impressed that the man’s count was almost the same as his own. “That’s close enough… take your pick.”
The driver selected one from the flock, put it in the back seat of his car, and prepared to drive away.
“Just a moment,” said the shepherd. “If I can guess your profession, may I have it back?”
Dressed casually for the weekend in jeans and t-shirt, with nothing in his car even remotely related to his work, he couldn’t imagine how that was possible, so he readily agreed.
“You’re a DNA barcoder.”
“That’s amazing! How did you know?”
“Never mind that, just give back my damned dog.”
Further Reading
Broad, Gavin (2025) Minimalist species descriptions: Are they the answer? And if so, what was the question? pp. 42-54 In D.M. Williams & Q.D. Wheeler, eds., The New Taxonomy: A Science Reimagined. CRC Press, Boca Raton.
NASA, Webb telescope images: https://science.nasa.gov/mission/webb/multimedia/images/
Sharkey, M.J. et al. (2021) Minimalist revision and description of 403 new species in 11 subfamilies of Costa Rican braconid parasitoid wasps, including host records for 219 species. ZooKeyszi 1013: 1-665. https://zookeys.pensoft.net/article/55600/
Wheeler, Quentin (2020) An unfinished revolution, Inference 5(2), May. https://inference-review.com/letter/an-unfinished-revolution
Williams, David (2020) Hennig, history, and hubris, Inference 5(2), May. https://inference-review.com/letter/hennig-history-and-hubris