Editor's notebook

What makes a good CJAI submission?

Steve Marshall, March 2014

I’ve now had eight years of editing CJAI and I’m proud to be associated with a journal that has become widely recognized for accessibility, utility and veracity. If you peruse a few recent issues of the Journal you won’t find it difficult to land on an issue that strikes you as a “good” publication ... but why? What makes a good CJAI submission? Here are some of my thoughts on the question.

Perhaps the easiest way to come up with a partial answer is to rephrase part of the question as “what makes an unacceptable CJAI submission?”.  Without naming names, let me describe a few papers that were either rejected prior to review, or that did not survive the review process. A couple of papers have come in that were simple repackaging of keys previously published by the same authors in open access revisions. The CJAI submissions used the same data and the same illustrations as the previously published work, but some were repackaged as matrix keys with lengthy new introductions about the merits and originality of matrix keys. In each case I looked at the original keys and the newly submitted matrix keys, and saw no advantage in the latter. Another rejected matrix key covered only four widely allopatric taxa. This submission also included a long introduction on the merit of matrix keys, but failed completely to provide any justification for publishing a matrix key to four easily distinguished taxa.  Matrix keys have their merits (further discussed below), but we don’t publish them just because some author wants to demonstrate  “innovation” by publishing a matrix key.  Submissions centered on dichotomous keys have been rejected too, some for similar reasons and some because they simply failed to provide any indication of originality.  That might reflect some confusion about criteria for “originality”, so it is worth commenting a bit about a line in the Instructions to Authors that reads “Updates, regionalizations and newly illustrated guides based on older printed works are welcome, especially if they include significant new distributional or taxonomic data”.

I expect that in the long term the great strength of CJAI will by synthetic works by specialists willing to aggregate and deliver their unique expertise in a user-friendly fashion, but in the meantime our encouragement of “newly illustrated guides based on older printed works” opens the door to naturalists, students and other non-specialists to provide new interpretations of previously published keys. Most keys to higher taxa published in the last century are essentially refinements and rearrangements of older keys, with the original authorship of many couplets lost in the mists of time. But that does not mean that merely copying a printed key, adding hyperlinks and adding images constitutes an original contribution appropriate for submission to a peer-reviewed journal like CJAI. The question, then, is what makes a key original? The usual answer I give to students who ask that question as they consider contributing to CJAI is that if they have to ask that question, they are not ready to write a key. Anyone who has spent the time to develop expertise on a taxon by studying the literature, using existing keys and ... most importantly ... curating and identifying the material in a major insect collection, will have a head full (better yet, a laptop full) of new distributional data, shortcuts for identifying the regional fauna, and clarifications of faulty key couplets. In other words, they will be ready to improve and regionalize the existing keys, and of course to supplement them with digital images, hyperlinks and species pages. Several published papers in CJAI illustrate these points well. For example, CJAI 13, a key to the eastern Canadian species of Tabaninae by A.W. Thomas (2011), is largely based on an excellent and widely available key (Teskey 1990). Despite that, Thomas (2011) developed a highly original and valuable contribution by updating the taxonomy, assessing and illustrating variation, providing species pages, adding new couplets and comments, brilliantly illustrating every character or species and providing new distribution maps that compare and contrast new records with distributional data from Teskey. I consider CJAI 13 (and every other published CJAI issue!) to be a “good” CJAI submission but I would consider a mere copy of a previously published key to be a “bad” CJAI submission, even with new photos and a digital format.

CJAI is a peer-reviewed journal, and reviewers do sometimes recommend that a paper be rejected or returned for major revision. Outright rejection usually results from reviewer opinion that the submission lacks utility or originality, and demands for major revision usually flow from reviewer opinion that the submission is useful but incomplete. Reviewers are usually the top experts in the field, and they are appropriately critical of submissions that are not totally up to date and accurate. Authors who are not taxonomic specialists or who have only recently developed the expertise necessary to write a CJAI review often do thorough and excellent work, but I have come to the opinion that such authors are well advised to bring in a specialist (if there is one!) as co-author before completing and submitting the work. We started the journal with the hope that the top specialists would be eager to interpret their taxa for a broad readership, but professional taxonomists have been relatively slow to take leadership in this area. Our experience, however, has been that taxonomists are almost invariably glad to help others who do set out to interpret the taxonomy of their group for a wider audience, and are often willing to contribute significantly to the work if invited to participate as coauthors.  Another characteristic of a good submission to CJAI, then, is that it reflects real and current expertise on the taxon involved, either on the part of the primary authors or coauthors.

Once the criteria of originality, utility and expertise are met, a good submission must meet certain technical standards. Every character and every taxon must be illustrated, and the illustrations must be visible when and where the user needs them. Illustrations must be clear. Text must be clear, unambiguous and as free of jargon as possible, with all technical terms defined and illustrated. Keys should have been adequately tested before submission, and terminal taxa should be linked to species pages. For traditional dichotomous (pathway) keys it is important to keep couplets symmetrical (ie, the two halves of the couplet must be clear antonyms) and it is usually advantageous to use multiple characters in each couplet. Walter and Winterton (2007) provide a set of recommendations for the construction of dichotomous keys that add significantly to these basic principles. Authors using traditional dichotomous keys should follow the guidelines in the CJAI instructions to authors and try to copy the best aspects of the many digital pathway keys already published in CJAI. But what about matrix keys?

A matrix key allows the user to choose from a long list of characters rather than being constrained to a linear series of questions or choices as in a pathway key. Matrix (multiple access, tabular, polyclave) keys have been around a long time (I recall using them to identify plants as a undergrad) but have undergone a resurgence in popularity with the appearance of computer hardware and software able to deal with large matrices of characters and taxa.  In theory, they offer tremendous advantages because the user can choose from a variety of character sets and can therefore work around characters they can’t see, don’t understand, or are missing from the specimen under study. They also allow for sophisticated treatment of variable characters or easily misinterpreted characters, and allow the development of keys from the sorts of large data sets professional taxonomists often compile.  Several software packages are available to assist with the construction of matrix keys. Lucid and XID are two of the best known, and have both been used to prepare excellent matrix keys now published in CJAI. Jason Dombroskie’s matrix key to families, subfamilies and tribes of Lepidoptera in Canada, for example (CJAI 17), is a monumental accomplishment based on careful examination of almost 1400 species. He used the matrix format  (XID software) to accomplish a task that would have been very difficult with a dichotomous key, and those interested in seeing a highly successful matrix key to a huge, traditionally difficult group should examine his key. XID, like other matrix key software, offers several useful options, such as an “analyze” button for efficiently narrowing the search down by listing the most useful characters.  Dombroskie (2011) was obviously an example of a “good” submission to CJAI. So why doesn’t CJAI publish more matrix keys?

I think the main reason we have published relatively few matrix keys is that to develop a matrix key for the kind of taxonomic problem that benefits most from the matrix approach ... many taxa, many characters subject to breakage or misinterpretation, problems with existing pathway keys ... is tremendously time consuming. Given that matrix key software has been widely available for around a decade, there are remarkably few really useful matrix keys available anywhere, and those generally derive from massively collaborative efforts or huge data matrices developed as parts of doctoral theses or similarly large projects.  Given a choice between a matrix key and a pathway key for most taxonomic problems, the pathway key will take much less time to construct, much less time to debug, much less time to edit, and much less time to review. And, all other things being equal, I would argue that pathway keys are easier and faster to use.  Irrespective of these opinions the bottom line is a simple one – a key that is not useful or lacks in originality, content or design will be not be a good CJAI submission, whether it is matrix or pathway.

Literature cited:

Dombroskie, J.J. 2011. A Matrix Key to Families, Subfamilies and Tribes of Lepidoptera of Canada. Canadian Journal of Arthropod Identification No. 17, 19 July 2011, available online at http://www.biology.ualberta.ca/bsc/ejournal/d_17/d_17.html, doi: 10.3752/cjai.2011.17

Thomas, A. Tabanidae of Canada, east of the Rocky Mountains 2: a photographic key to the genera and species of Tabaninae (Diptera: Tabanidae) Canadian Journal of Arthropod Identification No.13, 16 February 2011, available online at http://www.biology.ualberta.ca/bsc/ejournal/t_13/t_13.html, doi: 10.3752/cjai2011.13

Walter, D.E. and S. Winterton. 2007. Keys and the Crisis in Taxonomy: Extinction or Reinvention? Annual Review of Entomology 2007. 52: 193–208 DOI: 10.1146/annurev.ento.51.110104.151054

 

The views in the “Editor’s Notebook” are the personal views of the current editor, and do not reflect official policy of the CJAI

 

--------------------------------------------------------------------------------------------------------------------

The Canadian Journal of Arthropod Identification and cost-effective identification of North American arthropods

“… in the ideal world any insect should be as readily identifiable as lady beetles and butterflies are today. Identification costs would be uniformly low if, for example, there was a central web site with links to user-friendly, richly illustrated, authoritative, regional keys for all adequately known insects. Instead of asking what it should cost to have individual insect species identified again and again, we should be addressing the costs of developing the tools needed to make those individual identifications simple and accurate.”  (Marshall, 2003. The real costs of insect identification. Newsletter of the Biological Survey of Canada (Terrestrial Arthropods) 22(1): 15-18.
http://www.biology.ualberta.ca/bsc/news22_1/opinionpage.htm)

Although most of the World’s arthropod species remain undescribed and thus unidentifiable, generations of arthropod taxonomists have done a relatively thorough job of documenting the North American fauna. Most arthropod species of at least eastern North America, and virtually all North American members of larger and better known arthropod taxa, have been formally described and named.  Despite this relatively advanced state of taxonomic knowledge, most North American arthropods remain functionally unidentifiable except by specialists equipped with libraries, reference collections, and specialized tools. This gap between the completion of a taxonomic infrastructure and its translation into a generally available form is a historical artifact reflecting the cost and difficulty of generating widely accessible identification tools in a pre-digital era. Up until very recently relatively few taxonomists were in a position to produce and reproduce adequate images of all the taxa and characters included in their descriptions and keys. Furthermore, relatively few of those taxonomists able to assemble adequate image libraries could afford to publish them. Inclusion of even a few color images in printed publications was (and is) prohibitively expensive and would in any case be available only to a relatively select readership.  

This situation has changed abruptly over the past few years. Digital photography, even of small structures and specimens, is easy to master and relatively inexpensive. More importantly, the Web now provides a vehicle for efficient, swift and inexpensive distribution of image-rich products to a much wider audience than could ever be reached with print publications.  Newly published taxonomies, especially those appearing in digitally distributed journals, such as Zootaxa, promise to be much more accessible and user-friendly than paper taxonomic publications of the past. In fact, just as no taxonomic revision of the past few decades was complete without the inclusion of a dichotomous key, no major taxonomic treatment today should be considered complete without the inclusion of, or parallel web posting of, a comprehensively illustrated digital key. CJAI welcomes the publication of such keys based on new taxonomic work, but relatively few newly published taxonomies deal primarily with the well-known North American fauna. The important task of rendering the North American arthropod fauna identifiable demands a return to past taxonomies, and those with the necessary specialized taxonomic expertise now face both a responsibility and a great opportunity to “complete” that past taxonomic work by making previously named species accessible through web publication of reviews including photographically enhanced digital keys.

The papers published in CJAI so far demonstrate that this approach can remove the “taxonomic impediment” to the identification of most previously described species, while at the same time flagging and circumscribing problems that still need to be resolved, such as cryptic species complexes or undescribed species. Most, if not all, taxonomists have invested large parts of their professional lives familiarizing themselves with their chosen taxa, and most can easily identify species that fall into their areas of expertise. This expertise can now be easily translated into identification tools for others, and every North American arthropod taxonomist should be availing themselves of this opportunity.

The Canadian Journal of Arthropod Identification is a fully refereed journal and it is now developing into a significant source of expert-authored taxonomic reviews and associated digital identification tools. It is becoming a cost-effective and universally available tool for rapid arthropod identification, meeting user demand while providing authors with a unique opportunity to contribute to a growing body of digital identification tools by sharing their regional taxonomic expertise. We need your help to make the CJAI a major source of authoritative reviews and keys to a wide range of arthropods from North America and beyond.

Steve Marshall,
October 2008