An Overview—Tips for Searching OWL

 

Background:

            Literature that cannot be later found is information lost to the rest of that science.  There are presently several hundred journals and other serials devoted to ornithology around the world.  Virtually no one person can see and have access to all of that literature.  The purpose of this project is to provide ornithologists around the world with a free, searchable database to the recent literature.  Citations within OWL come from over 1500 journals, reports, and other serials going back to about 1980.  Many of these publications are distributed to a relatively small readership in one part of the world.  Ornithologists can benefit from these papers, but only if they are aware of them.

            From 1983 until 1998, a supplement to the Ibis, Auk, and Emu was produced in hard copy, booklet format.  This was called Recent Ornithological Literature (ROL).  Those supplements have been scanned, edited, and converted into records in OWL.  Approximately 65,000 citations in the present OWL database come from this source.

This conversion process took several years.  Extensive editorial work on these old records was required.  One step was to adjust some of the editorial conventions then used in the old printed supplements that are not useful in an electronic file (e.g., fully spelling previously abbreviated generic names, adding addresses to later records of the same author in the same ROL).  Another step was to spell check all the scientific names and other technical terms to remove the typographical errors; such errors make those terms inaccessible to any future user of the database.  A further step was the indexing and coding the entries to the <current subject topics> and adding additional keywords and secondary subject codes.  The subject headings used before 1998 were much more generalized than those currently in use for the OWL; additionally, those pre-1998 entries were placed only under one topic heading.  These added steps make all of the old material comparable with the more recent data and more useful.

            OWL editors could not review all the original papers for each of those ROL citations during this conversion process.  Many citations had minimal information (author, title, journal citation) and little more.  Where there was some confusion as to the nature of the paper or some indication that a small number of avian taxa might be involved in a particular paper, efforts were made to locate the paper and add the names of the taxa or clarify the subject matter.  Less than 2% of the printed ROL citations received this assistance before being added to the OWL database.

 

Spell Checks (limitations):

OWL data have been spell checked as best as possible.  The names of people and most geographical names and addresses are impossible to verify.  Author names and initials simply cannot be guaranteed 100% as to their validity.  From the various edits, a small percentage of <2% were found to have problems in this area.  The one area that was very carefully spell checked in the data was the scientific names of the taxa (birds, as well as some other animals and plants).  A unique dictionary was created so that binomial and trinomial combinations could be verified.  Over 45,000 combinations were included in this dictionary for spell checking.

            OWL does not follow any particular taxonomy.  OWL accepts all names that have been used in the major taxonomic lists of that region or the world in the past 20­­­­­­–30 years.  Generic assignments have fluctuated.  One authority may consider a form a subspecies, while another considers it a full species and yet a third opinion places it as a synonym of another form.  Thus, forms that may be considered synonyms in one list may be a separate taxon in another.  Accepted spellings (e.g., gender agreement) have also varied over time.  Users of OWL need to know all the recent synonyms for any taxa for which they are searching the database.

 

Record Fields:

            Not all fields found in each OWL record are searchable using the Search function.  Those that are searchable are: Author, Year, Title, Journal, Abstract and Keyword/subject codes.  Address and other fields are not searchable but are part of any extracted record and viewable when Editorial information is checked in the search.  A separate <list of all the fields> provided in an extraction and their use is described.

 

Suggestions for Searches:

            As with most databases, searches can be constructed to yield a variety of results.  Unless otherwise specified or selected, three fields are searched for the search strings:  Title, Abstract, and Keywords.  Although searches might be attempted in a non-English language, no spell checking or other verification has been made on titles with other languages.  An English translation is provided to aid in OWL searches.  The first three searches listed below are the type most frequently used.

 

  1. Taxon/Taxa Search—A scientific name and any synonyms are placed in the Topic of the Search page.  All records that have that name or names are extracted.  The major problem here is that OWL and its predecessors did not list all the taxa discussed in every paper.  Avifauna surveys, censuses, parasite hosts, and similar lists were almost universally not provided in the record.  At the present time, OWL abstractors are asked to provide the scientific names of up to about 15 taxa covered in that paper.

 

  1. Topic/Keyword Search—OWL can be used to find papers dealing with particular topics, regardless of the taxa involved.  <Subject codes> were devised to get around the problem of word choices in the selection of keyword topics.  Word spelling, as well as choice, is also a problem even in English, although thoughtful use of the query syntax can help with this. For example “behav*” will find most of the variants of this word.  The codes try to sort the topics covered into general or specific categories.  Ringing or banding techniques are in E526; egg colors or colours are covered in B710; reproductive effort, breeding success, productivity, and nesting output are covered in code C918.  The term ‘behavior’ can be found in a wide range of topics, from courtship or territorial to migration or foraging.  Although searches for specific keywords are encouraged, the uses of the appropriate subject codes are also to be encouraged whenever feasible.  All of the older ROL records that were converted were originally placed under a single topic heading in the supplement.  Many of those old records could not be fully coded as to subject matter; in such cases, the general B100-series of codes had to be used to indicate that several facets of the bird’s life history are likely discussed but which ones was unclear to the OWL editors.  When searching for topics, try one or more of the subject codes in the Topic search request form and then try the specific terms in a separate search. Fuzzy matching can also be used to find all the Subject Codes in a group – eg B3* will find all the Subject codes starting with ‘B3..’, ie all of ‘Behaviour and Communication’.

 

  1. Specific paper search—In some cases, the search might be to find one or more papers by a specific author or a specific title from a particular journal in a particular year.  That is, some information may be known about a paper but not the full citation.  The year or a range of years might be known.  A particular word in the title or the author may be remembered.  In other cases, the journal in which the paper appeared might be known.  The name of the author is the most problematical; it could be misspelled in a few cases (est. <2%) or shared with another author.  Non Latin characters can also be a problem. We have translated these into what we think are the obvious Latin equivalent, but if you fail to get a hit try a fuzzy match to part of the name that is in Latin characters.

More likely, the junior author name was not provided if in a list of four or more co-authors of a paper in a citation prepared prior to ~1997.  In the old ROL editorial style, page costs drove them to list only the senior author and use “et al.” for the other authors when there were four or more.  Try a combination of known terms in the OWL search form in trying to locate a particular paper.  Note that many authors may cite their previous works in subsequent papers; so a search of more recent works may produce the citation being sought.

 

  1. Geographical or political area searches—OWL has endeavored to provide the name of the country, State or Province in records where such information is important and when known:  censuses, avifaunal surveys, distributional records or summaries, geographical variation, conservation, and similar topics.  As with the scientific names of birds, the spelling or name of many of these places and political units have changed over time.  Both the B100-series of codes and the C300-series have a geographical component.  The B100-series is for life history subjects not covered in other codes used for that particular record, while the C300-series is for distribution, zoogeography and avifaunal history topics.  As will be noted, the one dozen faunal boundaries used in these subject codes are relatively broad.  Although searches can be attempted, the use of specific geographic terms (e.g., river, mountain, valley, city, district, park, refuge) may not result in all the papers in the OWL database that actually cover that target area.

 

 


List of Fields in OWL records

 

The arrangement is the order found in an extraction.  Names of a few fields are altered in the OWL database search results display when Editorial Info is checked; these are given in parentheses.  Content and other notes are included for each field.

 

Recordid (Article #):  A unique number for each record; useful to identify specific records that have any errors or other problems.  There are a number of duplicate records in the OWL database.  At some point in the future these will be located and then combined; for the moment, we would assume database users would rather have two records for the same citation (each with differing recorded numbers and perhaps other information) than have no citation at all.

 

Authorkey:  The author key is a special version of the author field; used in searches for author names.

 

Author:  List of authors in the format of data entered.  Older ROL-style was in typical literature citation format:  Surname, initials, etc., with punctuation.  Current OWL style is the same as the authorkey format.

 

Year:  Year of publication, when known.  If a second year is included in parentheses, the other should be the year the publication actually came out in print.

 

Title:  English or translation in brackets.  In some cases the original language version is provided when available.  If the record is for a New or Renamed or Discontinued journal, the full name of the journal is also included here.

 

Serial:  Name of journal or other serial.  If a non-serial, it can contain a variety of styles or other bibliographic information.  Some records from the same serial may use the full name, while others may use the standard abbreviations used in biological literature.

 

Volume:  May also contain issue number in parentheses, if the issues are paginated separately.

 

Page:  The first page of the paper; it may also indicate the total number of pages for reports or other publications.

 

Fpage:  The final page of the article or other pagination information; some papers were continued over a range of pages, etc.

 

Address:  The address of the author, when available.  In some cases, the address is taken from other records that were in the same ROL issue and may not be accurate.  No proofreading or spell checking has ever been done on this field.  Be aware that most addresses are often out of date after only a handful of years from publication.  If searching for a more current address for a particular author, search for that author's name in a more recent citation.  If the field starts with a set of initials or a name other than the senior author, this is the author who was to be contacted for reprint requests.  If available, the E-mail address of the author is given.

 

Abstract:  If present, this is usually a brief summary of the paper's content provided by the ROL/OWL abstractor or editor.  In a few cases, we have copied the actual abstract of the paper and have so identified it.

 

Abstractor:  The volunteer ROL/OWL abstractor who provided this record for the database.  If the identity of the person is needed for some reason, contact one of the senior OWL editors who will try to reconstruct that information (about 5% of the past volunteers are not known, only their initials).

 

Keywords:  This is a list of subject codes and keywords that maybe useful in subsequent searches of the OWL database.  OWL does not have a standard list of keywords.  See discussion under <Historical Overview>.

 

Source:  The batch of records from which this one citation originated.  The ROL supplement issued in January 1983 was numbered 27; there had been 26 earlier supplements to the Auk.  ROLO and OWLO were records from direct online input.  Subsequent OWL batches continued the numbering of the ROL supplements; these are combined submissions from the volunteer abstractors around the world, which are edited, spell checked and then appended to the OWL database.

 

Inserted (Accession Date):  Date and time for the record being appended to the database; if a revision of the same record replaces an earlier version, this date and time will reflect the later date of that action.  This is very useful if searches are made to exclude records previous to the date of the earlier search; this will then yield any new or updated records.

 

Oktoshow:  A field for internal use as to displaying the record (0) or not (1).

 

Recordstatus:  Degree of edits and other checks that the record has received. 

0 means nothing known that invalidates this record (i.e., its probably ok).

3 means it was entered online, and has not yet been fully checked by an editor.

5 means the data is suspect for some reason; such records usually have much missing information.