***This blog post is the long version of the editorial that will be published in the issue 128 of FrogLog. It is linked to an early-view FrogLog article celebrating the milestone***

(Almost) 50 Years of Effort on a Catalogue of Living Amphibians

Darrel Frost, Department of Herpetology, American Museum of Natural History

I was asked by Amaël Borzée to write a history of Amphibian Species of the World. The just-the-facts history of the catalogue is available online (https://amphibiansoftheworld.amnh.org/History-of-the-project-1980-to-2024). So, instead of rehashing that history, this essay is aimed at where the catalogue came from, why it was needed, and where it came from within the rapidly changing technological framework—computers and information management—that occurred during its (and my) development over the last 50 years.  I will also add some conjectures on where taxonomic information management is going.

My father was a computer scientist involved with critically relevant advances in the field of computerization, which gave me a tangential awareness of the early development and trajectories of computer technology. And as a systematist (and taxonomic catalogue maker), I have long been focused on how this technology impacts information flow in systematics and conservation biology. So, the following account is largely based on the memories of someone who is not a computer scientist, but who was aware of the big things happening in that field in the late 1950s through the 1990s.

My first personal experience with computers would have been about 1957–58 when my father was working for the Systems Development Corporation (SDC) inside a nuclear-bomb-resistant building near Kansas City, Missouri. The SAGE (Semi-Automatic Ground Environment) computer within that building covered about 1800 square meters of floor space and the computer memory was in the form of enormous numbers of vacuum tubes and tape-recorder machines. It had something around 256K of RAM (substantially less than the cell phone in your pocket) but still managed to run the NORAD (North American Aerospace Defense Command) missile defense system for the central USA. Programming was done in machine language, this all being before the development of high-level languages such as FORTRAN. Regardless, the functionality of these computers was startlingly advanced. I remember being shown a radar screen where I could aim a “gun” of sorts and click on an incoming object on the screen, such as an airplane or missile. That computer-assisted effort aimed a real-world missile to destroy the intruding object. That’s terrifying to contemplate these days, but when I was 7 it was just so cool. (Given that this was at the height of the Cold War, it is somewhat surprising that families were allowed to tour high-security sites, but it was the 1950s….). Most of my friends’ fathers were World War II veterans. They had no idea that such technology existed in any concrete way and, truthfully, they did not attach a lot of importance to computers beyond vague worries about automation killing jobs. Computers, however, were just part of the adult conversation in my house. But, most mathematical calculations by biologists in the 1950s and early 1960s, at least in the USA, were still mostly accomplished using electro-mechanical Friden calculators. I can still hear their clackity-clack.

By the 1970s, when I was in college, vacuum-tube computers were just a memory and transistors had come into their own, reducing the size and enhancing the speed of computers. (As it was put to me in the early 1960s, computers had to get smaller to increase their speed because electricity could only go so far [about 30 cm] in a nanosecond, so distances had to be shortened to get the most work out of moving electrons.) Most universities had computer centers, with one or more large mainframes. Students could take classes in FORTRAN or COBOL. Major universities started developing computer science programs in the early 1960s so by the 1970s there was at least the capacity for computer number-crunching on most university campuses. Nevertheless, registration for classes, at least for my freshman year (1969), was done with physical class-cards and waiting in long lines, not with any electronic help. Modern computer-driven GIS (Geographic Information System) came into its own in the 1960s and 70s, making our modeled distribution maps of today possible. The logical basis of the current internet was already in existence, with modem connections by phone lines from departments to the mainframe. Although the general public was unaware of it, ARPANET (Advanced Research Projects Agency Network, the Department of Defense’s telephone-line precursor of the internet) was in place. This was a high-redundancy system that shared “packets” of information over telephone lines so that if one intermediate node in the system, like New York City, was destroyed in a nuclear war, the redundancy of the system could compensate automatically and get information to the desired site, such as a missile site near Boston. ARPANET developers also had chat rooms for programmers working on the system. In this period, I remember my father railing about people’s anonymity in online chat rooms allowing them to say evidence-free and irresponsible things—30 years before the existence of the World Wide Web and “influencers.” In 1971, my interest in systematics and taxonomy had become full-blown, and I had a summer job that had me spending every day in the Arizona State University (ASU) library working on a taxonomic catalogue of the mammals of Tunisia for E. Lendell Cockrum, the University of Arizona mammalogist, all paper and pencil and photocopy work. No laptops or internet were available… yet. I did learn how to use hard copies of the Zoological Record and Biological Abstracts and became intimately familiar with the Annals and Magazine of Natural History and other old and important biodiversity journals. As a relatively new university (university status granted in 1958), ASU did not have a lot of antiquarian works, but it was a good place to learn library skills. And I did.

In the professional biological world, science publication was still firmly in the era of printed paper. Bulletin boards (physical cork boards on walls, of course) in biology departments across the USA always had numerous posted advertisements like “I type dissertations. Call XXX-XXXX.” However, by the early 1970s it was also becoming common for people in biology programs to employ computers for data analysis with the result that you would see PhD students dragging multiple kilograms of computer cards across campus from computer center to home and back. That seems laughable now, but the future was clear to anyone who wanted to look. Just over the horizon from public view, information technology and the hardware that supported it were advancing quickly into a new world.

The miniaturization of computers was accelerating. The first time I remember seeing a minicomputer was about 1979 in the offices of the Association of Systematics Collections (ASC) at the University of Kansas. As I recall, the computer was about 2 cubic meters in size, and it supported about five computer workstations and had a great proprietary word-processing program. This was also my first experience with word-processing systems and relational electronic databases, and they have been an integral part of my daily life ever since.

In 1979 there was no ITIS (Integrated Taxonomic Information System) or GBIF (Global Biodiversity Information Facility) and no publicly available internet (at least not in anything like its present form). Wes Sechrest was five. Conservation concerns were already growing fast, however. The IUCN (International Union for Conservation of Nature) Red List was 15 years old but would not go online as a comprehensive database until 2000. CITES (Convention on International Trade in Endangered Species of Wild Fauna and Flora) was 4 years old. GenBank existed but would not go online (ARPANET) until the early 1980s (and not onto the internet as we recognize it until 1994). Physical books were still the standard of publication. With the advent of international regulation of trade in wildlife CITES and ASC had concluded, wisely, that for purposes of regulation, global, comprehensive, and high-qualitytaxonomic catalogues were needed to make sure that the parties were all on the same page with respect to the status and application of taxonomic names. The idea that taxonomic catalogues represented a service to conservation efforts was not new, but this formal recognition of the global need provided new impetus toward their development.

As a PhD student in lizard systematics in the mid-1970s at the University of Kansas (KU: one of the major herpetological research centers in the USA), and with my bibliographic history, I certainly knew the value of taxonomic catalogues, the Zoological Record, Biological Abstracts, and other paper-printed bibliographic tools. And, having typed hundreds of index cards for the herpetology collection at the University of Arizona as an undergraduate part-time worker, I was one hell of a typist. So, in 1979, when I should have been finishing my dissertation on iguanian lizard phylogenetics, I was volunteered by William E. Duellman, my advisor, to compile a catalogue of world amphibians. And, since I have always liked big projects, I agreed, even though this dealt a major hit to my schedule for finishing my dissertation.

Fortunately for me, Mammal Species of the World (Honacki et al., 1982), the first ASC catalogue, was being completed, and I was pulled into the final stages of its development to learn procedures, and how to manage interactions with a world-wide group of contributors, reviews, and members of the oversight committee. I learned quickly that the academic world is deeply infused with many great, generous, and very smart people but with enough narcissists with enormous egos to make life… interesting. And, there was a simple list of species of amphibians kept by Richard Zweifel (the spectacular AMNH curator I subsequently replaced upon his retirement) and lots of older (e.g., Gray, 1850; Boulenger, 1882), regional (e.g., Jerson, 1853) and good but more taxonomically restricted (e.g., Gorham, 1966; Duellman, 1977) catalogues to guide the way.

Finally, after nearly five years of 60-hour weeks of work, sitting on the floor of the KU library going journal by journal, even with the enormous support from contributors, reviewers, and the oversight committee (particularly Alice G. C. Grandison and Marinus Hoogmoed), the physical paper-printed Amphibian Species of the World (Frost, 1985) came out. Was it perfect? No. Was it useful? Yes. Was it a political football? Absolutely. At least some thought it would be a dandy weapon to “stabilize” nomenclature to the detriment of anyone outside of a group considered to be “the right people.” Unfortunately, people who talk this way rarely are talking about promoting progress for the field (i.e., monophyly of recognized taxa), but usually are more interested in personal or institutional prestige. I didn’t like what I was hearing.

Following publication of Amphibian Species of the World, I was initially pulled into the development of Crocodilian, Tuatara, and Turtle Species of the World (which was subsequently transferred to F. Wayne King and published in 1989). I then returned to my dissertation work on iguanian lizards, finished in 1988, and was hired as an Assistant Curator at the American Museum of Natural History in 1990. Although Amphibian Species of the World came at a considerable academic cost to me in terms of my personal academic timetable, it made me intimately aware of the conservation world and the substantial value to society of these people and organizations.

But the amphibian world was not done with me, even though I was a reptile specialist and still am (in my own mind, I suppose). The efforts of the World Congress of Herpetology Checklist Committee (and many corrections and additions provided by me) that were supposed to carry the catalogue forward turned into a paper-published addendum under the sole authorship of Duellman (1993) and seemingly without long-term plans for an electronic future.

It had been clear to me for some time after my exposure to ARPANET that most people did not yet appreciate that the internet (or something like it) was the future. Realizing that paper-publishing was obsolete for taxonomic catalogues (and scientific publishing, generally), I had retained one of several electronic copies of the 1985 amphibian catalogue (in digital form) and, as time permitted, began tracking newly published species descriptions. So, during the period of 1990 to 1999, the catalogue was kept up to date on my office computer insofar as new taxon additions, although the addition of taxonomic comments and such were minimal.

Big advances in computer technology and information management became accessible the public in the 1990s. By 1993 the World Wide Web (internet) was broadly available to the public. (According to Google AI, there were already 600 known web servers and more than 500 websites as of that year.) The AMNH was wired for the internet in 1993, and it was clear to me at the time that an online database was the only realistic path forward for a new edition of Amphibian Species of the World (ASW). The Integrated Taxonomic Information System (ITIS) came online in 1996, providing the USA Federal government with a unified source of animal and plant taxonomy. Then there was a huge development: the Google search engine came online in 1998. Information was now at the public’s fingertips—my fingertips. Things were moving quickly in the information age.

In 1998, I provided an electronic copy of the amphibian catalogue to a collection manager at the Museum of Vertebrate Zoology, University of California–Berkeley, an old friend from KU days, for purposes of collection management. This document became the backbone of what is now AmphibiaWeb. Clearly, I wasn’t the only one who saw the potential of an amphibian catalogue as an online resource for taxonomic and conservation information. But I had plans for developing ASW into something much bigger and more useful than it had been in its original static, paper-printed form without trying to be all things to all people, to compete with conservation resources like IUCN. And I was in a position to put those plans into action. I had a willingness to work, extensive experience with taxonomic catalogues, and the substantial resources of the American Museum of Natural History behind me. Fortunately, the AMNH also has one of the largest, most complete natural history libraries in the world. So, although I was spending a lot of my time as an administrator (Associate Dean of Science for Collections) while also trying to get some reptile work done, I found time between meetings to address one antiquarian volume after another, building complete synonymies book by book. (For those who are curious: it is a lot easier and faster to build taxonomic synonymies volume by volume rather than taxon by taxon, particularly when dealing with antiquarian books, and it is a much quicker way to detect ancient errors long embedded in secondary literature.) There were also people with real database management skills in the building willing to provide support. Moreover, recognizing that paper publishing limited access to biologists trying to work in the amphibian-rich but resource-poor regions of the world, I wanted to promote world-wide access to critical information for previously underserved scientific communities. So, having a free, online database available to anyone with access to the internet was the clear choice for promoting information flow to taxonomists and conservationists worldwide. With the internet, if you knew what the literature was, for the first time in history you could easily get access to it. We had always talked about science not having borders—being international—but for the first time it actually was.

After an enormous effort by David Reddy, Scientific Visualization and Informatics Manager at the AMNH, the first online edition of ASW became available in July 1999. Unlike the paper-published original, the database could be easily searched by taxon, English name (new in 2000), author of the taxon name, year of publication, distribution, and keywords in distribution. In addition, each record included a list of links to additional web resources such as Google, GenBank, or the IUCN Red List. This state of affairs—the database existing in static versions, updated every couple of years—persisted until 2004.

In 2001, Wikipedia came online. HerpNet (NSF funded) came online in 2002 (and was subsequently subsumed by VertNet), providing internet access to collection data for amphibians in a big way for the first time. In 2004, The search engine Google Scholar came online as a resource still unmatched for providing access to scientific literature (and one dear to my heart as my daily morning window into the latest developments in the world of amphibian taxonomy).

Also in 2004, Mark Breedlove replaced David Reddy as the AMNH Scientific Visualization and Informatics Manager, and he moved the ASW database to PERL. For the first time, we provided the Integrated Taxonomic Information System (ITIS) with the taxonomy and synonymies of amphibians, in effect providing the U.S. Federal Government with a standard amphibian taxonomy (serving the same purposes identified by CITES back in 1979). My understanding is that ITIS provides taxonomic information to other organizations, so, in effect ASW provides information to them by proxy. We also started linking ASW to Google, Google Scholar, Wikipedia, and AmphibiaWeb.

In 2005, a simple conversation broadened my horizons with respect to the audience ASW could serve. Until that point, although I had accepted the notion that having a standard set of names for discussion was a good idea, I was focused primarily on the needs of taxonomists. But that year I attended a meeting of Mexican amphibian and reptile experts in Chamela, Jalisco, Mexico, for purposes of constructing IUCN Red List accounts. At that meeting I had a conversation with Simon Stuart, then Head of the IUCN Biodiversity Assessment Unit, that strongly affected my priorities and added a new dimension to the amphibian catalogue. He mentioned that such mundane things as range extensions could be very difficult to find for people doing assessments and asked if I could include such citations/information as I ran into them. Previously I had largely ignored range extensions other than new countries of distribution, but since that conversation I have included meaningful range extensions, although I must admit I still do not include every mention of a collected specimen, particularly if it is mid-range. But that conversation with Simon Stuart made me a lot more sensitive to how people from ecology/conservation backgrounds (not taxonomic ones) look at literature and what is needed to make for good conservation assessments. I started to see the amphibian catalogue in broader terms of needs. I was not (and am not now) interested in competing with any of the more conservation-oriented sources of information (e.g., IUCN Red List) or other specialized websites (e.g., Map of Life, VertNet), but I want to make life easier for them when I can.

About 2006, around the same time as colleagues and I published a very large taxonomic study of amphibians, The Amphibian Tree of Life (Frost et al., 2006), Julian Faivovich asked me to start including reference to frog calls in ASW since they frequently are mentioned in papers but not reflected in titles or picked up by search engines. I started doing so.

Two important resources came online in 2007: the Biodiversity Heritage Library and Biological Abstracts came. International access to literature was enormously expanded. In a very real sense we had arrived in the information age in which we currently reside. It is not the end of progress, however.

In 2012 an administrative reorganization within the AMNH moved the amphibian catalogue into Communications, to be run by Matt Tarr, AMNH Digital Architect, who then moved the database to the eZPublish (a content management system). This meant that updating ASW taxon records, and error correction, was done for the first time on a continuous minute-by-minute basis by Frost rather than as periodic updates managed by the database managers. The user interface was improved. And, frankly, my life got much easier since there is now of team of people working with Matt who bring enormous expertise to the table and I no longer had to manage the unnumerable hidden delimiters in the literature citations. After that point, we ran around 1,000,000 page views a year at that point. This rate was depressed some during the covid epidemic but is now inching back up.

So, here we are. It is 2026. People born in the 1970s mostly do not remember a time before the internet and the ease of access to information that it provides. Digital access to collection records, updating of identifications, and digitization of specimens is lagging, however, a worrisome development. Scientific publications are overwhelmingly available online in PDF format, and few people even bother with reprints anymore. The Amphibian Species of the World catalogue will never again be produced in a static paper-printed version. But it will continue in some form. The number of amphibian species has surpassed 9000 (or shortly will). In 1985, we counted 4014 species. Li et al. (2025) suggest that the ultimate count may be 40,000. I suspect that number is a bit high, but I have no data to reject their model’s estimate, so we shall see.

We have not reached the end of changes in information management and access. To think otherwise is to ignore history. So, what is the future of Amphibian Species of the World? What is the future of taxonomic catalogues? Where is technology going that will affect how data are transferred to the inquiring public? For purposes of this discussion, I suspect that the technological advances going on over the horizon, away from public view, are no less great now than they were in the 1950s. Consider artificial intelligence. AI will change the employment structure of humans on a worldwide scale. It would be surprising if it did not also produce dramatic effects in information management and dissemination and, frankly, in the science of taxonomy. But could it take over the construction of taxonomic catalogues? Of threat estimations for purposes of conservation policy? Of range estimations? Surely the answer is yes in the long run. But, in the shorter run (ca. 20 years), I have misgivings. Computers are great at handling enormous amounts of data and applying analytical tools. They are not known for their judgment and they are no better than the data provided to them. And taxonomic nomenclature is not so much scientific as it is based on usefulness in communication, with guard-rails such as (1) requiring monophyly and (2) being based on good standards of scientific methods. Conservation regulation and priorities depend, among other factors, on (1) crisp taxonomies based on substantial evidence and considerable (2) distributional and (3) ecological data. Factor #1 is probably the easiest of the three to attain in terms of human effort. Distributions can be modeled to a substantial degree, but having more data points is always better, and specimen identifications at all major collections trail the published revisionary literature by years and include many misidentified specimens. This issue could be ameliorated by AI, but not soon. The same applies for sequence data in GenBank. I’ve been impressed by just how few GenBank depositors go back and make necessary changes to the identifications on sequences they have deposited. (I actually was sent a ms for review criticizing me for employing a misidentified genetic sample, identified and deposited to GenBank by the author of the manuscript.) Solid ecological studies are very time-consuming if done well, and some workers spend their working lives in very restricted areas. So, there will always be a tough road forward at the evidentiary end of things, if not the analytical end. If the evidence is incomplete (as it always is) or bad (as it sometimes is) then no amount of analytical power can make the results better than the evidence.

The answer to many of these challenges is, of course, judgment. And this makes me worry. At times I think the amphibian catalogue should be handed off to an international organization to be vetted by taxonomic experts with judgment and good knowledge of the species and of analytical methods. Unfortunately, many older, established taxonomists are too invested in the ways of the past. (And, as we are now in a “regular science” period where no one wants to rock the methods boat, so, frequently, are younger people.) How would an organization select a committee of established (i.e., older) systematists who are willing to change paradigms as necessary from time to time? I still hear the call for “stability” coming all too frequently, not so much from younger people doing the work, but from older workers who just think they should have the last word, particularly with respect to older, enormous and not-very-useful genera. I would love to see a study on mnemonics and how taxonomies should be designed for purposes of maximizing discussion of relationships. The optimum is surely not masses of monotypic genera nor is it genera containing more than about 50 species. But, regardless, in an AI world there is no reason for any nomenclatural confusion. The answer will be only a brief query away on my cell phone.

At this point in time (2026), getting information into the catalogue still requires human effort, primarily from me (generally in the first two hours of the morning while drinking my coffee)—catching a title on Google Scholar, downloading and reading the paper, and entering the required changes and additions into the database. None of this is currently being the capability of AI. As for the underlying science I mostly just follow the most recent literature. I am forgiving regarding methods of alignment and tree-building, although I have my limits. And, I am getting tired of mtDNA-only evidence for species distinction.  Collection of data (specimen examination and genetic sampling and analysis) and the writing of the papers still requires substantial human intervention, although the inability to major tissue and specimen repositories to keep up with taxonomic change will remain a serious problem for the foreseeable future. I can, however, imagine that over the next 30 years that human effort in required by analysis and communication will progressively decline and much of this work will be done directly under machine control.

So, while I think that the future of taxonomic catalogues and the summation of taxonomic publication will rely on human judgment aided increasingly by artificial intelligence, these issues of methods of practice need to be addressed scientifically.

AI is currently not remotely perfect, but it can currently digest papers in any language and insofar as digesting evidentiary controversies, it can only get better. I have also been astonished at how quickly things I type into ASW turn up in Google AI. In a sense, the various AI platforms and I are sharing our morning coffee over the taxonomic novelties of the previous day. So, as far as the amphibian catalogue is concerned, I will stick with it for another few years and see what develops. If AI can get smart enough to articulate worry about species definitions based solely on 16s mtDNA, and if it can recognize the problems caused by insufficient taxon sampling, misidentified specimens, and alignment assumptions, I think we will have arrived. Concern is always the first step on the stairway to progress.

LITERATURE CITED

Amphibian Species of the World (1999–present). https://amphibiansoftheworld.amnh.org/. American Museum of Natural History, New York.

AmphibiaWeb (2000–present).  https://amphibiaweb.org/. Museum of Vertebrate Zoology, University of California, Berkeley.

Boulenger, G. A. 1882. Catalogue of the Batrachia Gradientia s. Caudata and Batrachia Apoda in the Collection of the British Museum. Second Edition. London.

Duellman, W. E. 1977. Liste der rezenten Amphibien und Reptilien. Hylidae, Centrolenidae, Pseudidae. Das Tierreich, 95: 1–225.

Duellman, W. E. 1993. Amphibian Species of the World:  Additions and Corrections. Special Publ., Mus. Nat. Hist., Univ. Kansas.

Frost, D. R. (ed.). 1985. Amphibian Species of the World: A Taxonomic and Geographic Reference. Association of Systematic Collection, Lawrence, Kansas.

Frost, D.R., T. Grant, J. Faivovich, R. H. Bain, A. Haas, C. F. B. Haddad, R. O de Sa, A. Channing, M. Wilkinson, S. C. Donnellan, C. J. Raxworthy, J. A. Campbell, B. L. Blotto, P. Moler, R. C. Drewes, R. A. Nussbaum, J. D. Lynch, D. M. Green, and W. C. Wheeler. 2006. The amphibian tree of life. Bull. Am. Mus. Nat. Hist., 297: 1–291.

Gorham, S. W. 1966. Liste der rezenten Amphibien und Reptilien. Ascaphidae, Leiopelmatidae, Pipidae, Discoglossidae, Pelobatidae, Leptodactylidae, Rhinophrynidae. Das Tierreich, 85: 1–222.

Gray, J. E. 1850. Catalogue of the Specimens of Amphibia in the Collection of the British Museum. Part II. Batrachia Gradientia, etc. Printed by Order of the Trustees. London.

Honacki,  J. H., K.E. Kinman,  and J. W. Koeppl (eds.).1982. Mammal Species of the World: A Taxonomic and Geographic Reference. Association of Systematics Collections. Lawrence, Kansas.

Jerdon, T. C. 1853. Catalogue of reptiles inhabiting the Peninsula of India. J. Asiat. Soc. Bengal, 22: 522–534.

King, F. W., and R. L. Burke (eds.). 2010. Crocodilian, Tuatara, and Turtle Species of the World: A Taxonomic and Geographic Reference. Association of Systematics Collections. Lawrence, Kansas.

Li, X., D. Yang, L. Wang, and J. J. Wiens. 2025. The past and future of known biodiversity: Rates, patterns, and projections of new species over time. Sci. Advances, 11(eadz3071): 1–16.