It should be easy to determine the exact number of scholarly journals that are available online. Surprisingly, it is a challenge. Even how many scholarly journals are published in print isn't easy to calculate. Coming up with these numbers is a tale that information specialists will appreciate.

Ulrich's publishes its serials directory in many formats; the most up-to-date and complete is ulrichsweb.com™. According to Yvette Diven, director, product management, serials, R.R. Bowker LLC, "ulrichsweb.com is the version...that contains the largest numbers of publications pulled from the Ulrich's database…. Ulrichsweb. com contains the active, ceased, suspended, and forthcoming titles from Ulrich's as well as records for titles announced but never published and records for which supplemental research is required in order to confirm the latest information." It is updated weekly. Luckily, ulrichsweb.com searches can be limited to "active" records. This is pretty unambiguous. In mid-November 2003 there were over 180,200 active serial records.

Restricting these to just scholarly journals gets more complicated. Ulrich's has two field designations that might be relevant: "academic/scholarly" and "refereed." Refereed indicates refereed and peer-reviewed periodicals and is assigned mostly on information gathered from publishers. Hence, the problem. Genuine academic publishers understand the process of peer review, but others may not. According to Diven, "On occasion…a publication confuses peer review with the publication [being] frequently consulted by readers of its peers in the marketplace. There is also a misperception among some publishers that a publication is necessarily peer reviewed if it has an editorial board. Ulrich's editors make a concerted effort to clarify our definitions."

Academic/scholarly is a term that publishers pick from a list that describes audience or focus (it also includes Trade, Consumer, and Newspaper). It refers to publications intended for an academic audience or that contain materials with a research focus.

For active refereed periodicals we get just over 21,000 titles; active academic/scholarly weighs in at nearly 43,500. Most of the refereed are within the academic set, so the final answer to active serials (refereed or academic/scholarly) comes to just over 43,500.

How many of these are e-journals? Ulrich's actually provides two digital designations: "online" or "CD-ROM." Online in Ulrich's means full text online but "does not necessarily mean online-only," according to Diven. In fact, most of the online periodicals are also published in print, with only about 4600 e-only publications. Many e-only publications are newsletters, consumer publications, or web-based zines rather than scholarly publications. CD-ROM availability is not included under online, although some titles may have both designations.

Ulrich's lists over 34,500 online, active periodicals of all types (remember, most of these have both online and print versions). CD-ROM is far behind, with only 6500 titles having active CD-ROM versions. As expected, there is some overlap, so the total number for digital periodicals in Ulrich's (either CD-ROM or online and also active) is nearly 37,500.

Restrict that to only online active refereed and you get 11,000; online and active and academic/scholarly or refereed yields over 14,600. CD-ROM may add another couple thousand titles.

It is more difficult to calculate how many of these are science, technology, or medical (STM) titles. Restricting searches to appropriate Library of Congress (LC) or Dewey classification numbers seems the best bet, but even that demands caution. Peter Jacso, well-known author and database reviewer, explains that as of summer 2003 over 80 percent of Ulrich's database records have no information in the LC field. Such "errors of omission can be more dangerous than errors of commission because they are, well, invisible," said Jacso. Luckily, all Ulrich's records include at least one Dewey Decimal number.

Restricting the search by Dewey classification can be time-consuming as fields vary considerably. Many of the online active titles are STM but by no means all. The humanities and social sciences have many online titles as well. On average, in Dewey 500s and 600s, only one-quarter to one-third of active peer-reviewed titles are online. In subfields like astronomy and medicine, the number of titles online is closer to 60 percent.

The American Scientist e-forum (run by Stevan Harnad) recently had a days-long debate on this topic, with some claiming that there are as many as half a million scholarly journals published today. Both librarians and publishers benefit from knowing the true scope of serials titles and online penetration. I can say with confidence that as of the end of 2003, there are just under 50,000 scholarly journals and somewhere between one-third and just over one-half of them are in digital form. One thing I've learned is that these numbers are a moving target and somewhat suspect. Keep checking and keep definitive statements necessarily vague.

Special thanks to Ashley Pillow, who helped verify all of this research.