Introduction

For the last 60 years, scholarly journals have witnessed unprecedented growth, controversy and change. Since the late 1940s the number of scholarly journals has increased sharply, with hundreds of new titles and new topics being introduced each decade. Beginning in the late 1960s and especially since the 1990s, the form of journals has been transformed into digital versions that speed up both the access to and delivery of articles to readers and provide enhanced functionality. Ejournals are now more popular with libraries and readers than their print counterparts, although both forms continue to coexist for many titles. This combination of more titles and more widespread availability, particularly in electronic formats, has engendered lively debates in the library, publishing and scholarly communities, and has kept scholarly journals at the forefront of discussions of the promise and problems with traditional forms of scholarly communication channels.

Profound changes to the look, functionality and publishing rhythms of the scholarly journal are outside the scope of this chapter and are covered elsewhere in this volume. Some of the changes in electronic publishing emphasize the unit of scholarly article rather than the unit of journal title, or even the unit of scholarly article section, table or figure, rather than whole article. In this chapter, the unit of analysis is mostly the scholarly journal title, each of which is made up of many individual articles as well as some non-article material such as letters to the editor, book reviews, calendars of events, etc. It will be up to others to speculate whether or not the journal title, typically made up of issues that include articles and non-article content, will survive and thrive. This chapter is based on the assumption that it will, although with increased digital functionality and the capability of being easily separated into smaller components, such as articles or parts of articles, through search and display. Some characteristics of articles are included here, but only within the context of the journals that publish them.

A historical perspective

To truly understand the growth of journal publishing in the twentieth and twenty-first centuries, a brief history of journal publishing will help, as the history of journals stretches back much further than the middle of the twentieth century, the starting point of many discussions. For a more detailed history of scholarly journals, see Houghton (1975), Kronick (1976), Meadows (1998) and Tenopir and King (2000).

The history of modern scholarly journals is generally traced back to the 1600s, with the first scholarly journal believed by most to be Le Journal des Sçavans (Journal of the Experts), founded by Denis de Sallo in France in January 1665. The first issue shared some characteristics with a modern journal – it had ten articles, some letters, and notes, for example – but it was only 20 pages long and covered information on many different topics. The founding and development of Le Journal des Sçavans corresponded to the budding practice of sharing discoveries and ideas through correspondence between scholars, the rise of scientific societies and the concomitant development and growth of newspapers. Indeed, this first journal was a fresh type of news organ for the scholarly community, including such newsworthy topics as obituaries of luminaries, abstracts of books, reports of legal decisions and summaries of a range of developments on topics or issues thought to be of interest to educated men. The title was later changed to Journal des Savants and published as a literary magazine (Houghton, 1975; Tenopir and King, 2000).

Others point to Philosophical Transactions of the Royal Society of London as the first truly modern journal (Meadows, 1998; Price, 1975). Launched in March 1665 as a monthly periodical, Philosophical Transactions provided a means for members of the Royal Society of London to distribute the results of their scientific experiments and achieve a wider distribution of Royal Society lectures. The first issue is now available in digital form through JSTOR, along with the complete run, and the early issues are certainly more accessible and likely to have been read more in the past few years than they were in their first 300 years.

Whether it was Le Journal des Sçavans or Philosophical Transactions that earned the title of first modern journal, the growing pace of experimental science, and the desire to share ideas and results more widely and efficiently than personal letters, hand copyists or books allowed, meant that by the late seventeenth century the scientific journal was ‘a solution whose time had come’ (Tenopir and King, 2000: 56).

According to Garrison (1934) and Houghton (1975), by the end of the seventeenth century there were between 30 and 90 scientific and medical journals being published worldwide, with a total of 755 journal titles by the end of the eighteenth century. Price (1975: 164) puts the number of titles at ‘about one hundred by the beginning of the nineteenth century, one thousand by the middle, and some ten thousand by 1900’. Journal titles and the professional societies that often sponsored them began to move from broad coverage of all issues that might interest the educated man, to specialization.

Scientific specialization was reflected both in the birth of new journals that covered a narrower or more focused topic and in the ‘twigging’ of existing broad journals into narrower, more specialized topics. Influential thinkers in the history of science point to the growth of specialized journal publishing as a reflection of the development of scientific thought. Kuhn (1970), for example, who described the formation of scientific paradigms and paradigm shifts in his Structure of Scientific Revolutions, stated that the formation of specialized journals (along with the creation of specialized societies and a role in the curriculum for the new ideas) demonstrates the acceptance of a new paradigm in science.

The specialization of science continued through the nineteenth and twentieth centuries, and the importance of journals to scholars as both readers and authors persisted. Derek de Solla Price (1963) famously plotted the growth of the number of journals from 1665 through to 2000 (Figure 6.1). He discussed the exponential growth curve that would lead to nearly a million journal titles by 2000 if the historical rate of growth continued. His calculations showed the number of titles doubling every 15 years, increasing by a power of ten in 20 years, and by a factor of 1000 in 150 years.

According to Price (1975: 169), ‘in the three hundred years which separate us from the mid-seventeenth century, this represents a factor of one million’. Price likens the growth in the number of journals to ‘a colony of rabbits breeding among themselves and reproducing every so often’ (ibid.). He suggests there is something about scientific discoveries that engenders this rabbit-like behaviour, as each new ‘advance generates a series of new advances at a reasonably constant birth rate, so that the number of births is strictly proportional to the size of the population of discoveries at any given time’ (ibid.: 170). Based on this, he concludes that: ‘the number of journals has been growing so that every year about one journal in twenty, about 5 per cent of the population, has a journal-child – a quotient of fecundity that is surely low enough to be reasonable but which must inevitably multiply the population by ten in each succeeding half-century’ (ibid.).

Price’s calculations do not account for discontinued journals, however, and although new titles have undoubtedly arisen more quickly than obsolete titles have been discontinued, the actual growth in total journal titles may be somewhat lower (Houghton, 1975; Kronick, 1976; Meadows, 1998). His disclaimer that he is plotting the ‘number of journals founded (not surviving) as a function of date’ (Price 1975: 166) is also not entirely accurate, as he plots the total cumulative number of journals each year, not just those newly founded each year. He also seems to confuse periodicals with scholarly journals. Price’s assertion that ‘according to the World List of Scientific Periodicals, a tome larger than any family Bible, we are now well on the way to the next milestone of a hundred thousand such journals’ (ibid.: 164), almost certainly includes trade journals and other non-peer-reviewed titles. The following section gives further consideration to such problems with counting.

Price identified the relationship between new discoveries in science and the twigging (or parenting) of journal titles. He also highlighted the relationship between the growth in the number of journal titles, the growth rate of individual journal articles, and the corresponding growth rates of the number of abstract journals. He showed that the number of articles also follows the same exponential growth as the number of scientific journals. As author productivity has remained fairly constant over time, Price speculated that growth in the numbers of scientific journals and scientific articles is closely tied to, and increases in tandem with, the number of scientists.

Price also showed that from 1830 through to the early 1960s about one abstracts journal was created for every 300 journal titles. Abstracts journals are founded due to an impending crisis of information overload: as the number of journal articles increases, scholars cannot hope to read every article that may be potentially relevant in their field. As Price anticipated, this remains a problem. He even called the then new method of ‘electronic sorting’ ‘no more than a palliative and not the radical solution that the situation demands. It can only delay the fateful crisis by a few paltry decades’ (ibid.: 168). Modern ‘electronic sorting’ search engines have solved the problem of finding the ever-increasing amount of scholarly information. As anticipated by Price, the challenge today is finding the best or most relevant information among the myriad of articles.

Recent growth in the number of titles

It is surprisingly difficult to come up with an accurate number of how many scholarly journals currently exist, let alone model how that number has changed over time. Mabe (2003), Morris (2007) and Tenopir (2004) all have struggled with the best way to reconcile the differences in the number of titles estimated by Price, Tenopir and King, and others. If Price’s curve had remained constant, there would have been close to a million titles by the year 2000.

King et al. (1981), Mabe (2003) and Tenopir and King (2000) have shown that the growth in journal titles, journal articles, research and development (R&D) workers and R&D expenditures in the US are all correlated (see Figure 6.3 and later discussion). Although there remains a correlation between the number of scientists and the number of journal titles, as demonstrated by Price, Price’s curve has flattened. The number of scientists continued to increase through the 1990s, but by 2000 we most certainly did not have his predicted one million journal titles. This discrepancy is most likely due to the increase in the average number of articles published each year per journal title (Tenopir and King, 2000).

King et al. (1981) estimated that there were a total of 4447 US-based journals in 1977, compared with their estimate of over 57,400 titles worldwide (Figure 6.2). Using the same method of counting only those titles with a US presence, Tenopir and King (1997, 2000) estimated that the number of US-based journal titles had increased to 6771 by 1995, a figure that is remarkably close to the prediction by King et al. (1981) that there would be 7000 US-based journal titles by 1999. These relatively low numbers did not include any titles that did not have a US editorial office. In today’s international publishing world, this way of counting no longer makes sense.

At the other extreme, Meadows and Singleton (1995) estimated that there were approximately 70,000 to 80,000 scholarly journals worldwide. (Earlier estimates by Meadows [1993, 1998] ranged from 10,000 journals in 1951 to 71,000 in 1987.) These high numbers seem reasonable when compared with King et al.’s 1977 estimate of 57,400 titles worldwide (King et al., 1981) but both may be an over-count if we restrict the picture to scholarly peer-reviewed journals. These estimates most likely include other types of periodicals, such as some non-peer reviewed magazines or trade journals and perhaps even newsletters, in addition to peer-reviewed journals.

Ulrich’s Periodical Directory, published since 1932, is probably the best source for monitoring the number of and growth in periodical titles and offers a level of consistency over time. Ulrich’s 640,000-plus titles include all types of serial, not just scholarly journals. The directory also contains discontinued or forthcoming titles in addition to actively published titles. Searches in the online versions can be restricted to academic/scholarly journals, refereed journals, or titles that are coded as either or both. Searches can also be restricted to active titles, avoiding the problem of counting dead or inactive titles for which Price was criticized. By restricting Ulrich’s November 2011 total to active titles, the total number of periodicals is reduced from more than 640,000 to 327,925. Using Ulrich’s as the basis for monitoring the growth and number of journals has the advantage of being replicable and consistent.

However, although it offers consistency there are also problems associated with relying on Ulrich’s. Almost certainly it does not include every single scholarly journal title available in the world today, particularly those published by small publishers and that are not in English or that do not provide an English-language abstract. Guédon (2008) observes that estimates based on Ulrich’s ‘tend to yield lower numbers’ because Ulrich’s ‘aims at a specific clientele largely made up of librarians from rich countries. It selects what it thinks is of potential interest for potential buyers’ (Guédon, 2008: 43). Morris (2007: 299) points out that Ulrich’s is ‘entirely reliant on the information supplied by the publishers of the journals listed therein. New journals are often not listed immediately. There can therefore be no hard-and-fast guarantees as to the completeness, currency, or accuracy of that information.’ In addition, within the records for the journals that are included in Ulrich’s, Jacso (2001) has documented serious ‘errors of omission’ or incompleteness of fields in some of the journal title records, such as Library of Congress subject headings.

Still, Ulrich’s is the best and most consistent source we have to estimate actual journal growth over time. Mabe (2003: 191) justifies his choice to use Ulrich’s by saying ‘only Ulrich’s attempts to cover all serial publications and to classify them by a number of criteria. Ulrich’s also has the undeniable advantage of being available in a readily researchable CD-ROM format as well as online.’ Mabe used the CD-ROM version for his searches; all of the searches presented below are from the online version, Ulrichsweb (http://www.ulrichsweb.com/ulrichsweb/analysis/). As the web version is completely updated every week, retrospective searches are not possible. All of the older data presented here were collected on the date indicated.

Searches of Ulrichsweb from 2003 through to 2011 show the growth in journal titles over time (see Table 6.1). How many journals there are depends very much on the search strategy (as demonstrated by Morris, 2007), so several Boolean search strategies are presented below. To eliminate one of the above-mentioned problems with Price’s data, in all cases searches were restricted to ‘active’ titles only – that is, those that were currently being published at the time of the search.

It is worth noting that designations in Ulrich’s rely on publishers self-reporting the information accurately. As reported by Tenopir (2004), one Ulrich’s editor has acknowledged that publishers are not necessarily consistent in understanding the difference between academic/scholarly or refereed. According to Ulrich’s FAQs: ‘The term refereed is applied to a journal that has been peer-reviewed. Refereed serials include articles that have been reviewed by experts and respected researchers in specific fields of study including the sciences, technology, the social sciences, and arts and humanities’ (http://www.ulrichsweb.com/ulrichsweb/faqs.asp).

Using a slightly different strategy with Ulrich’s CD-ROM, Mabe (2003) obtained a figure of 14,694 active academic/scholarly and refereed periodicals in the summer of 2001 (that is, active titles coded with both designations). His strategy is closest to row five in Table 6.1 (17,649 in February 2003). The main difference in Mabe’s strategy, however, is the use of the Boolean AND NOT to eliminate titles containing one of several type indicators, including audio-cassette, bibliography, Braille, broadsheet and 20 other similar designations. (Differences between other variations in strategy, or whether the web, Dialog or CD-ROM version of Ulrich’s is searched, are discussed by Morris, 2007.)

Going a step further, Mabe has also recalculated the annual collective growth rate of journals since 1665 using a logarithmic scale (see Figure 6.3 for an updated version of his 2003 chart). Allowing for a few exceptions, such as immediately after the Second World War, his findings suggest that: ‘for most of the last three centuries, the growth rate of active peer-reviewed scholarly and scientific journals has been almost constant at 3.46% per annum. This means that the number of active journals has been doubling every 20 years’ (Mabe, 2003: 193).

Changes in the number of articles and length of articles

On average, journals are publishing more articles each year, whether or not they publish them in issues. Counting growth in journal titles is still important for those who deal in a world of titles and the commodity of the entire journal as a unit, including publishers, scholarly societies and librarians. Journal titles remain important to scholars as well, especially for authorship, current awareness reading and assessment of relevance and quality (Nicholas et al., 2005, 2006; Tenopir and King, 2007). When conducting research or reading for other purposes, however, the article as a unit is generally more important than the journal as a whole. The journal title remains important to readers, however, as an indicator of relevance and quality (Tenopir et al., 2011).

Björk et al. (2008) have attempted to measure how many scholarly articles are published annually and how many of those are available in open access. Working from Ulrich’s and the Thomson Scientific ISI citation databases, they estimate that of 23,750 journals published in 2006, the total number of peer-reviewed journal articles was approximately 1,350,000. (Incidentally, about 19.4 per cent of those articles was available through some variety of open access by early 2008.) They found that non-ISI title journals publish on average 26.7 articles per title per year, while ISI titles publish 111.7 articles per title per year. Tenopir and King (2000) have reported an average of 123 articles published annually per title in US-based science and social science journals. This number excludes humanities journals, but includes social sciences, engineering, medicine and all sciences.

Björk et al. (2008) believe that their calculation of the total number of journal articles published yearly is in line with the estimate by Elsevier (2004) that 1.2 million articles are published every year in peer-reviewed journals by publishers in science, technology and medicine. Björk et al.’s numbers also include social sciences and humanities.

Although co-authorship has increased in most fields over time, the productivity of authors has remained relatively constant, at approximately one article per year per author and approximately two in science (Tenopir and King, 2000). The relationship that exists between the number of R&D workers and the number of journal titles can also be found between R&D workers and the number of journal articles (see Figure 6.4). Although not every researcher publishes, as the number of R&D workers increases so do the numbers of authors and articles (Mabe, 2003; Tenopir and King, 2000).

As part of a National Sciences Foundation (NSF) project entitled Statistical Indicators of Scientific and Technological Information Communication, King Research, Inc. observed characteristics of science journals published in the US. These characteristics were tracked annually from a sample of journals observed in libraries from 1960 to 1977 and 1985, and later at the University of Tennessee, School of Information Sciences in 1990 and 1995 (in 1990 also under NSF contract). The results were categorized by nine fields of science designated by NSF during the first observations. Basic data included number of titles, type of publisher, price, circulation, etc. Sub-samples of issues and articles produced such journal characteristics as issues, articles, article pages, non-article pages, proportion of graphics pages, authors (and their affiliation) and number of citations. ‘Science journals’ is defined as a primary vehicle for communicating information and research results, often employing peer review or refereeing processes to aid in screening and editorial control. They do not include trade journals. Some of these results were published by Tenopir and King (1997, 2000, 2004).

The King Research, Inc. studies found that in 1960 there were 2815 science journals based in the US, with 208,300 articles, or an average of 74 articles per journal. The growth of US science journals is shown in Table 6.2. Between 1985 and 1995, the growth trend in the number of journals based in the US seemed to slow substantially, while the number of articles continued to grow at a healthy rate. It may be that the cost of starting new journals and the financing necessary to fund journals until they become profitable (or at least break even) have encouraged publishers to increase the size of journals rather than start new ones. Typically, it can take as many as six years for a new journal to break even, requiring about US$50,000 in investment just to keep it going during that time (Page et al., 1997; Tenopir and King, 2000).

As the number of articles per journal title increases, presumably the rate of growth in the number of new titles can slow somewhat, remembering that the total growth of journals and journal articles has paralleled the concomitant growth of R&D workers for many years (Mabe, 2003; Tenopir and King, 2000). Another calculation that is relevant to information overload, but may or may not be relevant to the growth in the number of titles or articles, is the change in the average length of the articles published. Bringing up to date the calculations of article length in 1975 by King et al. (1981), and in 1995 by Tenopir and King (1997), and in 2007 by Tenopir and King (2009), Table 6.3 shows that the average length of articles increased by 93 per cent in the 36-year period from 1975 to 2011, with the great majority of that growth coming between 1975 and 1995. Between 2007 and 2011, the average length of articles increased by seven per cent.

Journal article length is somewhat dependent on subject field. Table 6.4 shows that although there have been large percentage changes in length, both positive and negative, during the 12-year period before 2007, the net effect has been a regression towards the mean length across all fields. While in 1995 the difference between the field with the longest article and the field with the shortest articles was 17.2 pages, by 2007 this range had narrowed to 11 pages. By 2011, that range had increased to 11.3 pages.

Online journals

As shown in Table 6.1, many scholarly journals are still only available in print. Of the approximately 30,775 refereed journal titles active in November 2011, 77 per cent (23,832) are listed in Ulrich’s as being available online. This does not mean they are only online – a majority are still available in both paper and e-versions. In 2011, just 14,729 journal titles in Ulrich’s were only available electronically (online or CD-ROM) (Tenopir, 2004). The 77 per cent of journals available in some e-versions includes all subject disciplines. Estimates of online availability of STM titles are much higher – 90 per cent according to publishing consultant John Cox (2007).

The discrepancy between estimates may also be due to factors other than subject discipline, as defining ‘online’ is not as easy as it might be. ‘Online’ in the strictest sense means that the entire journal is available in digital form, including all articles, editorial content and other non-article content. The fully online journal can truly and completely replace any print counterpart, and online versions often contain more articles and more content than a corresponding print version. Online in a less restrictive sense might mean that articles are available online, either directly from the publisher as an online journal or as separate articles within a database of articles.

An even less restrictive definition might mean that the publisher allows open access publishing, so that some articles may be available online from a variety of sources, including subject or institutional repositories, author websites, article databases, etc. This last category, often called the ‘green’ road of open access publishing, includes approximately 11.3 per cent of the 1.3 million articles published in 2006, according to Björk et al. (2008).

The growth of electronic journals

Electronic journal growth has increased exponentially over the last decade. Of particular interest are full-text online periodicals, which have grown dramatically since the early 1990s. The Fulltext Sources Online directory (Glose, 2011), which has tracked the online availability of journals, magazines, newspapers, news transcripts and other periodicals over time, shows a dramatic increase, particularly in the last ten years. Although it includes all types of periodicals and not just scholarly journals, the directory is a useful source when estimating the online availability of periodical titles. As of July 2011, Fulltext Sources Online included more than 47,000 periodical titles that are available online in whole or in part. The growth of entries in this directory is one way to monitor the growth in online periodicals, as growth in the directory can be assumed to mirror the growth of online sources (Figure 6.5). This figure of 47,000 is best compared with the 84,851 online active titles reported by Ulrichsweb (see Table 6.1), which is just 26 per cent of all active periodical titles.

Scholarly journals are much more likely to be available in electronic form. A survey of publishers in 2003 by John Cox Associates, Ltd. estimated that 75–83 per cent of scholarly journals were available online in whole or in part, with the highest percentage in the sciences (Cox, 2004). This same source now gives an estimate of over 90 per cent (Cox, 2007). Although the exact number of e-only journals is difficult to estimate, in part because we still live in a world in which many journals are published in both print and electronic forms, the availability of journal articles in electronic form has obviously increased greatly. Of print and electronic refereed journal titles, 77 per cent of Ulrichsweb’s active refereed titles are online as of November 2011, up from 54 per cent in February 2003 and 65 per cent in February 2008 (see Table 6.1). As at November 2011, the Directory of Open Access Journals (http://www.doaj.org) includes over 7135 titles, up from just over 800 in 2003 and 3400 in 2008.

Predictions for the future

Although the rate of growth has not matched Price’s prediction, the number of journal titles will almost certainly continue to increase for the immediate future. At a fairly steady growth rate of 3.4 per cent (Mabe, 2003) with some ups and downs, journals have shown remarkable consistency in their growth, even as the number of articles and the average length of articles continue to increase for most subject disciplines. Societal factors, such as the decrease in the number of R&D workers or the amount of national spending on R&D, would slow the growth of journal titles. Certainly, this may happen in selected countries, but it is unlikely to occur on the worldwide scale that encompasses scholarship and research. The publication of journals may shift to different nations and new economic models may take hold.

The focus on the digital article rather than the digital journal as a unit for search and retrieval will have some impact on future growth in the number of journals. Article authoring remains constant at between one and two articles per author per year, and the number of authors continues to increase. The pressure to publish in peer-reviewed journals remains high in academic circles worldwide, however. Most academic authors who choose to put their scholarship on their own websites or in article repositories still publish a copy in a peer-reviewed journal. For now, this creates alternative ways to get to the same work, rather than replacing journal titles. According to Fulltext Sources Online, the average number of sources per title is 3.97.

A more likely disruption to the centuries of growth in the number and size of journals is the growth in these alternatives to journals. The electronic distribution of articles, separated from their journal home, is possible and is becoming common. Web search engines serve as a locator for articles housed in subject or institutional repositories. Most of these articles are still tied to their corresponding publication in a journal, but the desirability of this in the future is questioned by some. If journal prices continue to increase, and if the number of journal titles continues to grow, and if readers and libraries perceive they can access a majority of the information needed from repositories, then licence prices (or the willingness to pay the fees of fee-based open access journals) may fall and some journals may not survive.

Whether this is a negative or a positive outcome is debated by many, but one thing is certain – journals have served a purpose to scholars as both readers and authors for over 300 years and their growth reflects the growth of science, scholarship, and research and development. A disruption to their pattern of growth would reflect a major change in the way science is disseminated. The article, as separate from the journal with which it may be branded, is certainly the way that readers often discover what they read when they search from search engines, full-text journal systems or indexing and abstracting services. In these cases, the journal name serves as a quality clue rather than as a starting point for finding articles (Tenopir et al., 2011.)

At present, ejournals mostly are still digital versions of printed journals and as such have still not realized their full potential. Journal and article functionality that breaks the bonds of the printed page has emerged slowly over the last several decades but now seems to be picking up momentum. Engaging the reader through social media is now part of many online systems, with pioneering work done by Public Library of Science (PLOS) (http://www.plos.org). Systems such as ProQuest’s Illustrata™ led the way in disaggregating articles by making parts of articles such as tables and figures searchable and retrievable (Sandusky et al., 2008). Other online systems or web-based journals are also exploiting digital capabilities more fully, including incorporating interactivity through social networks and multimedia. Elsevier has redesigned the appearance and capability of articles, with its Article of the Future developments (http://www.articleofthefuture.com/). JoVE (Journal of Visualized Experiments) combines peer-reviewed video plus text to create a more visual journal (http://www.jove.com). The pace of these changes – which began with citation linking within and between articles, links to external data sets, e-publication of individual articles without waiting for an entire issue, and virtual personalized issues – will almost certainly increase and will greatly influence the future of the journal.

Judging from the past, the journal will remain important even as it is transformed with additional digital functionality. Links to and from data sets, the incorporation of more social networking features such as reader-assigned subject tagging, and the inclusion of automated ways of identifying high-impact or high-quality articles are just some of the features that will be incorporated into journal systems. Most require a substantial investment on the part of journal publishers and authors, which means we will not see every function in every journal. The primary purposes of the peer-reviewed journal – to publish, disseminate and archive high-quality research results – will survive amidst the changes that technology enables in scholarly communication. The number of journals required to do so and the growth rate of new titles may slacken in a fully digital future, as the number of articles per title is not bound by paper limitations.

Acknowledgements

The authors wish to acknowledge the assistance of University of Tennessee graduate students Lei Wu, Liuyan Yang and Amy Love, and in particular William (Ben) Birch and Kayce Gill, who conducted the calculations of journal article length and journal growth.