Content area
Catalogers, geo-science faculty members, and graduate students from several Rocky Mountain Region institutions were surveyed. The catalogers were asked about their provision of subject access through geologic or geographic names, while faculty and students were consulted about their use of such access points. Strategies for providing optimal access are explored, balancing the needs of users for vernacular or variant forms of names and the sometimes conflicting needs of the cataloging community for adherence to protocols governing the construction and tagging of "correct" forms.
Catalogers, geo-science faculty members, and graduate students from several Rocky Mountain Region institutions were surveyed. The catalogers were asked about their provision of subject access through geologic or geographic names, while faculty and students were consulted about their use of such access points. Strategies for providing optimal access are explored, balancing the needs of users for vernacular or variant forms of names and the sometimes conflicting needs of the cataloging community for adherence to protocols governing the construction and tagging of "correct" forms.
In the December 31, 1909, issue of Science, the secretary of the Geological Society of Washington summed up the dismal state of stratigraphic nomenclature with the declaration, "The present incubus of names is something to be shaken off at the earliest moment."1 While this problem may be less onerous for stratigraphers these days, it continues to be a concern for contemporary librarians and patrons working with geological and geographic name subject terms in library catalogs. While wrestling with the incubus on the home front, we realized we needed more details on what problems exist for catalogers and patrons, how the problems arise, and what we can do about them.
There is much evidence that, in general, subject searching in library catalogs leaves much to be desired and yet is one of the most frequent, if not the most common, methods of gaining access to library collections. In their review of technical services research during 1987, Geraldene Walker and Judith Hudson find that "early studies of card catalog use led to the theory that subject access was of relatively little value to users, but more recent studies of online catalogs revealed a radical change."2 With online catalogs, retrieval of specific information is made even more complicated due to peculiarities of individual systems and the vastly increased number of access points. Research has shown that nearly half of online catalog subject searches retrieve nothing-Judith Adams condemns this as "an appalling state of affairs."3 Lois Mai Chan acknowledges that "online catalog use studies bear witness to the fact that our subject access arrangements fall far short of fulfilling the potential of the online catalog. Improvement is needed-urgently needed-in the light of how fast the MARC database is growing."4
That subject access is increasingly of concern is borne out in recent studies. Chan explores the possibilities of using Library of Congress classification as an alternative approach to subject access by pointing out its unique features, implications, and future work needed to make this a realistic alternative.5 In a comprehensive and notable examination of this topic, Karen Markey consolidates many of the findings of earlier studies sponsored by OCLC and the Council on Library Resources.6
Micco, Smith, Hsiao, and Intaravitak, in attempting to design a prototype database for an integrated system, maintain that the MARC bibliographic format as a tool for "bibliographic record keeping" has distinct limitations for subject access: "It is not practical for periodical literature, and is virtually useless for reference books. Library of Congress subject headings were never designed or intended for postcoordinate searching, nor do they provide systematic subject access as demonstrated in study after study."
In a recent article, Carolyn Frost and Bonnie Dede approach the notion of matching catalog headings with terms used in LCSH.8 Markey proposes in another paper that patron access to Library of Congress Subject Headings in machinereadable form (LCSH-mr) may be a potential breakthrough in subject searching if an interface can be provided to overcome such obstacles as errors, outdated headings, free-floating subdivisions, subdivisions applied on the basis of "model" or "pattern" headings, and the addition of geographic subdivisions to headings, for which free-floating lists are not available.6
James Ross, as one of the few investigators interested primarily in geographic name access, echoes some of these same obstacles as barriers to geographic name access, lamenting "the changes AACR2 has brought to heading structures of all kinds, the existence of pre-AACR2 headings that have not been revised to AACR2 forms . . . the ambiguous and inconsistently applied instructions sometimes found on the LC list . . . and the difficulty of verifying headings in online cataloging or in the online authority file."10 Names for geographic entities are also examined by William Studwell, although he focuses on the differing treatment by LC of corporate jurisdictions and subject place names. He postulates, "Perhaps this dual function of geographic name has been an element of confusion in the development of the system for geographic names as subjects."1 He goes on to suggest that LC adopt the same pattern as has been used for personal names-that is, to use only one form in subject access.
Aside from these studies by Ross and Studwell, the peculiar problems inherent in geographic and geological name access in subject searching have not been specifically addressed in much of the literature to date. This topic merits attention, however, since having an area name will theoretically help to narrow a search, especially in an online environment. People do have trouble refining topical subject searches due to such factors as controlled vocabularies and system designs. When geographic or geologic names, with all the attendant problems of discerning the correct form, are added to search queries, the situation is doubly complicated. Furthermore, a good deal of effort on the part of cataloged may be expended when newly establishing a geographic or geologic name heading according to standards and procedures adopted by LC and AACR2.
In an earlier study on faculty subject searching, Carolyn Frost points out that "the focus of research on subject searching in library catalogs has been primarily on catalog use by students. ... But faculty use of the subject search is also of con,,12
With that in mind, and presented with the vexing problem of establishing geographic and geologic names ("geo-names") in a remote area not much explored by LC's subject cataloged, we began gathering information both from cataloged and users (faculty and graduate students). Three surveys were designed on the basis of our experience, similar to the idea recently expressed by the Association of Research Libraries, Systems and Procedures Exchange Center: "Short, specifically targeted surveys with a limited objective can be very effective means of obtaining information to improve existing services, determine use patterns and/or frequencies, and plan new services, especially regarding use of new technologies." The Center also observes: "In many cases, strictly random samples may not be necessary for libraries to achieve a high confidence level in survey results."13 Surveys developed included: one for cataloged or heads of cataloging departments, one for geo-science faculty, and one for graduate assistants. This final one was restricted to University of Wyoming students, but the first 2 questioned people at selected institutions in the entire Rocky Mountain Region. Response rates for the 3 surveys were 100, 33, and 24 percent. Although the response rates for the last 2 are not overwhelming, the results do provide valuable qualitative information.
CATALOGER SURVEY
We sent surveys to 13 cataloging departments at 12 universities chosen from the University of Wyoming's administrative listing of comparable institutions in the Rocky Mountain Region and to 5 state library cataloging departments in Colorado, Idaho, Montana, Utah, and Wyoming. The mailing was preceded by a telephone call to each institution ascertaining willingness to participate; since this was positive in every case, we determined exactly who was responsible for cataloging operations and would receive the survey. All surveys were returned, either by department heads or principal catalogers involved with name authority procedures. Ten of the libraries belong to OCLC, 6 participate in WLN, 3 in RLIN, and 1 in NOTIS. (One university belongs to 3 networks.)
The questions sent to catalogers were based on problems we have encountered in establishing geo-name subject headings at the University of Wyoming. The foremost one is the problem of discrepancies between forms of names used in searching and forms established in library catalog authority systems. Do these discrepancies adversely affect user success rates in gaining access to pertinent information sources?
Several factors contribute to the problematic nature of finding or setting up such names in libraries where Library of Congress Subject Headings (LCSH) are used. In order to assign MARC subject heading tags 650 or 651 with the second indicator of zero a cataloger is supposed to follow Library of Congress examples or adhere to LC's published guidelines for formulating geographic/geologic names. If the heading has not been established by or for the Library of Congress in the authority file or in LCSH, the next step is to look for an occurrence of the name as a subdivision in LC bibliographic records. Since subheadings are generally not indexed in many catalogs, finding a particular one can be a matter of intuition, diligence, and luck.
In the cases where LC has not used the name in its cataloging, or when a discrepancy is discovered in actual LC practice, the cataloger might establish the heading locally using the same resources LC would use: personal contact with RGN (United States Hoard on Geographic Names), Columbia Lippincott Gazetteer, Webster's New Geographical Dictionary, National Gazetteer of the United States of America, and other standard sources as available. These reference works would be used in conjunction with the specific Library of Congress guidelines for establishing such names in Subject Cataloging Manual, (H690 rev. 11/25/86 and H810 rev. 8/14/85),M together with policy decisions published in issues of Cataloging Service Bulletin.15 The Board on Geographic Names issues Decisions on Geographic Names in the United States quarterly16 and also responds to catalogers' requests to help resolve a problem. BGN staff are extremely helpful in verifying locations, resolving conflicts and providing information recently added to their files. Furthermore, the Library of Congress will respond to catalogers needing a clarification of policy or advice on establishing a local heading. The major problem with the two lastmentioned resources is that initiating correspondence adds another step to cataloging and requires more time than many understaffed libraries can spare.
Whatever the cataloger determines the official form to be and however carefully LC's guidelines are applied in formulating the geo-name for catalog entry, there is always the risk that LC will construct the entry differently in the future. Although LC usually accepts the BGN form of name, it retains the option of rejection.
Another significant problem encountered in formulating geographic or geologic names is that the form of entry prescribed by LC rules may differ markedly from the form appearing on the chief source of information for cataloging or the form in everyday use.
According to Carolyn Frost,
It is widely acknowledged that a major source of difficulty in subject searching is the user's failure to match a search term with the term used in the catalog as an access point. It has also been pointed out often that users are either unaware of the existence of the catalog's source of subject terms-i.e., the Library of Congress Subject Headings List (LCSH)-or find it difficult or inconvenient to use.1.
An example is the feature in Wyoming and Montana known as the Powder River Basin. Although often considered a geologic basin, LC rules dictate that it be established as the Powder River Watershed (Wyo. and Mont.), not Powder River Basin. LC uses "watershed" to designate river basins and drainage basins. While the first form is "correct" according to Section H690 of the Subject Cataloging Manual and has been used in AACR2 LC bibliographic records, it would rarely occur to a patron to construct such a search strategy. To confound the matter, the name could appear on pre-AACR2 cataloging records as "Powder River Basin"; "Powder River Basin, Wyoming"; "Powder River Basin, Wyo."; "Powder River Basin, Wyo. and Mont."; or simply, "Powder River."
CATALOGERS' SURVEY RESULTS
Determining the official name as described above theoretically solves the problem of standardizing geo-name access at the network level but leaves the dilemma of providing access to the variety of names often given to a single place or feature. Bearing in mind the difficulty facing a searcher who knows perhaps only one of the many possible variants for a given heading, and the time-consuming nature of geo-name authority work, we designed a survey to determine:
* how much time catalogers spend on geo-name authority work;
* what kinds of materials need the most authority work;
* what kinds of features are the most troublesome to formulate;
* how often conflicts between local usage versus the Library of Congress' controlled vocabulary are found and how they are resolved; and,
* how catalogers assign subject subdivisions to provide access to the area in question.
This survey reveals that 8 of the 18 departments surveyed spend as much time on geo-name authority work as on other types of authority work, while 5 spend more time, and only 4 spend less time. One provides no comment. Original cataloging requires more geo-name authority work than does comparable adaptive cataloging for 12 departments.
Different types of publications being cataloged demand differing amounts of geoname authority work, with state publications needing the most overall and theses following as a close second. Commercial publications generally take less than either of these. While several catalogers say maps generate a good bit of such work, it is misleading to include them in the ranking, since not all responding departments catalog maps as completely as other materials.
Bodies of water are the type of feature most often needing authority work, but basins, valleys, mountains, and ranges, along with national parks and recreation areas also cause concern. Other features generating authority work forsorne respondents include quadrangles, dams, watersheds, geologic formations, and archaeological sites.
One of the survey questions concerns the frequency of discrepancy between form of name used on the work being cataloged and form of name used by LC or that resulting from application of LC's rules. Only one department feels there would always be a conflict, apparently assuming that a qualifier would need to be added in any case. Another department finds conflicts very frequently, 6 somewhat frequently, and 10 occasionally. As a solution to this situation, 13 departments employ cross-references, and 4 of the 8 with online catalogs provide keyword access as an alternative. Only 2 lack provisions for access to variant forms of geo-names.
Seven departments refer conflicts and inquiries to LC, 4 to regional or state agencies, 4 to their network (WLN), and only 2 to BGN. Some respondents refer to more than one agency, and others solve conflicts internally.
When asked to choose among subject headings they would use in their library for an item on Lizard Gap Basin located in Wyoming (a made-up name), 6 catalogers choose a county-level subdivision (e.g., Paleontolog)'-Wyoming-X County), 11 subdivide by LC form of feature name (e. g., Paleontology-Wyoming-Lizard Gap River Watershed), and 5 subdivide by the feature's common name (e.g., Paleontology-Wyoming-Lizard Gap Basin). Some respondents have marked more than one option, and others have marked no option at all.
In summary, it appears that most cataloged are aware of the need for geo-name access and make some provisions for these needs.
FACULTY SURVEY
In an effort to determine if the frequency of such prescribed terminology has an impact on information access, a survey on geo-name searching practice was sent to faculty members in geolog)' and geography departments at the 12 universities whose library cataloging departments responded to our first survey. We estimated from directories and college catalogs that there were approximately 300 geo-science faculty members at these schools and sent the appropriate number of surveys to each department head for distribution.
The survey was designed to elicit the following information:
* how faculty gain access to the collections for research and teaching (e.g., card catalogs, online catalogs, browsing, database searches, etc.) and which methods are preferred:
* how they use geo-names in subject searching and to what extent;
* how successful they are when looking for materials using a geo-name;
* how often they encounter different forms of headings than they are expecting to find;
* if geo-name access is more important in some specializations or subdisciplines than others; and
* what modifications to current subject access by geo-name is desired.
Of the responses we received, 100 faculty respondents (33 percent of the surveyed group) show considerable interest in geoname access points.
Of these respondents, 94 say they use the card catalog, and 44 search online (an encouragingly high figure, considering that only 8 of their libraries have online catalogs) . Overlap exists here, since several respondents use both card and online catalogs. For the record, faculty also show interest in browsing, along with some use of database searches and assistance from reference librarians. Many "other" methods are listed by faculty respondents-the most usual are periodical indexes and other relevant bibliographies, citations in published literature, and discussions with colleagues.
When conducting a subject search in the card catalog, 3.2 percent of the faculty members always use a geo-name as their sole search element, while 18 percent do so very frequently, 7.4 percent somewhat frequently, 47.7 percent occasionally, and 19 percent never. In the online catalog, 6.8 percent always apply this strategy, 15.9 percent use it very frequently, 20.4 percent somewhat frequently, 43.1 percent occasionally, and 27.2 percent never. (See figure 1.)
Among faculty card catalog searchers, 6.4 percent always include a geo-name in combination with another search element, 31.8 percent add one very frequently, 14.8 percent somewhat frequently, 29.7 percent occasionally, and 10.6 percent never. Of the online searchers, 11.4 percent always adopt this approach, 31.8 percent adopt it very frequently, 27 percent somewhat frequently, 29.5 percent occasionally, and 15.9 percent never. (See figure 2.)
Our next question asked how often faculty searchers find what they are looking for when they search using a geo-name. Only 2.1 percent of the card catalog users say they always find what they are looking for, but 17 percent say their searches are very frequently successful, with 32.9 percent succeeding somewhat frequently, 33.9 percent occasionally, and 2.1 percent never. Of the online respondents, none succeed always in finding what they are looking for, 22.7 percent succeed very frequently with geo-name searches, 36.3 percent somewhat frequently, 31.8 percent occasionally, and 6.8 percent never. (See figure 3.)
The answers to our question about how often catalog entries for geo-names appear in different forms from the users' search terms are hard to interpret. The figures cannot take into account entries which differ from the known form but are never found through alternative strategies. Nevertheless, 1.1 percent of card catalog users say they always find catalog entries in different forms from their usage, 14.8 percent find a different form very frequently, 22.3 percent somewhat frequently, 39.2 percent occasionally, and 1.1 percent never. This suggests that the phenomenon is of concern. The corresponding figures for online searchers are zero percent always, 13.6 percent very frequently, 20.4 percent somewhat frequently, 43.1 percent occasionally, and 6.8 percent never. (See figure 4.)
With all four tables, the respondents sometimes choose more than one selection under card or online catalog for each question. Also, in some cases, no response is given.
Over half of the geo-science faculty surveyed (55) feel access by geographic or geologic name is definitely important in their subdisciplines. Fifteen find it helpful sometimes. Only 10 feel such access is rarely of interest, and 11 see it as definitely unimportant. Nine did not answer this question. Research areas mentioned by those who think it important include mining, hydrogeology, sedimentology, structural geolog}', stratigraphy, palynology, diagenesis, economic geology, remote sensing, mapping, tectonics, geothermal studies, climatology, volcanology, and petrology. Some paleontologists and geomorphologists say geo-name access is useful to them, but others do not find it helpful.
Responses to an inquiry about desired modifications to current subject access by geo-names include a flood of pleas for additional or more extensive cross-referencing, despite the large number of libraries already providing cross-references. More access by specific formation name and quadrangles would please some patrons. (Since 1986, LC has been adding geologic formation names to subject heading lists, a policy announced in Cataloging Service Bulletin 34.)1K Overall, our respondents would simply like to see more geo-names in their catalogs, in the forms that are familiar to them.
GRADUATE SURVEY
A third survey was designed to gather data from other users of the catalogs- specifically, graduate assistants. We chose this population for several reasons. On the premise that some graduate students provide assistance with faculty research, it became apparent that their search methods also needed to be assessed. Additionally, we were curious to see what disciplines rely on geo-name searching at the graduate level, where areas of specialization and research might be more established than those of undergraduate students. We wanted to survey within the Rocky Mountain Region, in order to make a comparison to the previous surveys, a population of students likely to have an interest in the area's geological and physiological features. In order to limit the population size and distance, we elected to survey the University of Wyoming's graduate assistants.
At the time of the survey the University of Wyoming's online catalog contained about 80 percent of the collection, including all post-1978 monographs and approximately 85 percent of the periodicals and serials. It did not represent government document holdings, which were accessible only through standard printed indexes at that time. Searching of the online catalog could be done by author, title, series, subject, ISBN/ISSN, and keyword.19 Keyword searching accessed most fields except imprint, notes, and some serials linking fields and could be enhanced through the use of Boolean operators and truncation. Unconverted holdings were found only in the card catalog, which was separated into author/ title and subject entries and arranged by the 1968 ALA Filing Rules. At the time of the survey, many of the geological materials were not in machine-readable format, either lacking LC copy or still not converted from Dewey; therefore, we might safely assume that researchers interested in geo-names were familiar with the card catalog as well as the online catalog.
Headings in both catalogs have been generally assigned according to LC practice, either taken directly from LCSH or established locally on the basis of LC's guidelines (or lack thereof, since their proliferation has occurred comparatively recently in Cataloging Service Bulletins and in LC's subject manual). Cross-references are used in the card catalog but have not been incorporated into the online catalog.
GRADUATES' SURVEY RESULTS
Our questionnaire was similar to the faculty survey and asked:
* how graduate assistants gain access to the collections for research and teaching and which methods are preferred;
* what their areas of specialization are and whether geo-name access is important for these areas;
* how successful they are when searching by geo-name;
* what terms they would use in a search for information on "hiking on Elk Mountain in Carbon County, Wyoming" in both card and online catalogs; and
* what modifications to current subject access by geo-name in the University of Wyoming library catalogs are desired.
We distributed 604 surveys among the 19 UW departments with graduate assistants. We had 145 responses (a 24 percent return rate). Of the 145 graduate assistants who responded, 49 percent say that they use the card catalog to locate library materials for use in their research and teaching, and 97 percent search via the online catalog. Respondents indicate that they use both the online and card catalogs.
Regarding other methods for locating library materials for use in research and teaching, 78 percent say they use printed abstracts, indexes, and bibliographies. Twenty-six percent use computerized abstracts, indexes, and bibliographies, such as databases on Dialog and RRS. (This seems appropriate since the University of Wyoming charges for computerized searches.) Fifty percent browse the book and periodical stacks. Thirty-seven percent query the reference librarians, while 9 percent use other sources, such as referrals from other students and faculty, citations found in the literature, University of Wyoming's Document Delivery (a journal article delivery service), Current Contents, and the "new periodicals shelf." Once again, the respondents could mark all that applied.
Since an aim of our study was to investigate the importance of geo-names in searching, we centered our attention on the group giving a positive response to the question "Is locating materials by geoname important in your specialization?" This group constituted 20 percent of the total respondents. Looking at this group's responses to the question "Please list your areas of specialization within your major field of study," we found a wide range of areas, from political science through the humanities (such as art, English, and history) to the applied sciences (such as range management, plant science, and civil engineering), to the pure sciences (such as botany and zoology). It is apparent that this finding is in part due to the need for access to particular political jurisdictions as opposed to geologic or geographic feature names, which were the focus of our surveys. Disciplines that were mentioned, such as Latin American studies, international studies, political economy, foreign policy, and studies of social problems, also might be more likely to require access to political jurisdictions. Other areas of concentration given that might require access to a specific feature name or geographical area rather than a political jurisdiction are archaeozoology, faunal studies, engineering geology, soil and rock mechanics, stratigraphy, sedimentology, veterinary parasitology, exercise physiology, and a host of specializations relating to water management and ecology. E. J. Coates in "Significance and Term Relationships in Compound Headings" cites the topics "Geography, Geology, History, Social Sciences, Linguistics, Literature, Ecology, and names of groups of natural organisms" as "significantly conditioned by locality"; therefore, they merit entry under place-name.2"
Of the 20 percent group mentioned above, 68 percent answered on the usage of the card catalog. Of this 68 percent, 35.3 percent find most of the time what they are looking for when they search using a geoname, 50 percent some of the time, and 14.7 percent not very often. Regarding this same 20 percent group, 94 percent answered on the usage of the online catalog. Of this 94 percent, 29.8 percent find most of the time what they are looking for when they search using a geo-name, 47.8 percent some of the time, and 22.4 percent not very often. (See figure 5.)
Of the respondents from this 20 percent who list other methods for locating re- search or teaching materials, 76 percent use printed abstracts, indexes, and bibliographies, while 28 percent use computerized abstracts, indexes, and bibliographies including databases on Dialog and BRS. Forty-eight percent browse book and periodical stacks, 31 percent query reference librarians, and 10 percent list other methods. This last category includes referrals from other students and faculty and citations from published papers. Many use more than one method.
We were interested in what subject terms this same 20 percent group (those who feel that locating materials by geo-name is important in their area of specialization) would use to search in the card and online catalogs for an item about "hiking Elk Mountain in Carbon County, Wyoming." A cataloger using LCSH and LC's subject manual could assign the following subject headings for the catalogs: (1) Hiking- Wyoming-Elk Mountain (Carbon County : Mountain); (2) Elk Mountain (Carbon County, Wyo. : Mountain); and (3) Hiking-Wyoming-Carbon County. As expected, the feature "Elk Mountain" is the most frequent search strategy for both the card and online catalogs. The second most frequent response for the card catalog is "Carbon County." For online searchers, it is a tie between "Hiking" and "Carbon County." The third most frequent response for the card catalog is "Wyoming," whereas for the online catalog it is "Recreation."
In some cases, respondents use "Wyoming" in combination with "Elk Mountain" and "Carbon County." The success rate for finding items using these combinations in the card catalog is greatly affected by the patron's familiarity with LC's policies of direct as opposed to indirect subdivisions and of qualifying non-unique names. Assuming a patron understands that both headings and subdivisions are indexed in our online catalog, the user might also need to be familiar with LCSH terminology, the practice of abbreviating names in headings, and perhaps the truncation feature in the online catalog. In fact, some graduate assistants abbreviate "mountain," "Wyo- ming," and "county" in their responses for this question. The possibilities for missing items during actual searching are readily apparent.
An inquiry about desired modifications to current subject access by geo-name in the University of Wyoming library catalogs brought diverse responses. The most frequent suggestions indicate a desire for more cross-referencing, as well as for access to more library materials via the online catalog. Several students desire the inclusion of additional materials in the catalog, such as maps, journal articles, and governmental periodicals (which have since been added to the online catalog).
Some of the comments are incomprehensible or unrelated to the topic of our survey. One response shows dissatisfaction with our library's practice of binding current journals. Also evident from some of the answers is a basic lack of understanding that both journals and books are broadly accessible in the online catalog, but indexes for contents of articles or chapters are provided for elsewhere, and explicit volume holdings information is found in yet another source (probably the "microfiche" referred to in one of the responses). One comment suggests the consolidation of our card catalog with the online catalog, and another one reads "spend more money on books, periodicals and the like instead of spending it on new cataloging systems" indicates a certain naivete concerning sources and expenditures of library funding, which one might expect. At the same time, however, it is asking a great deal of users to hope they can understand the searching idiosyncracies of two dissimilar systems.
User instruction in the intricacies of searching an automated system is even more important than ever. Not only is each individual system a little different in how things are indexed, but also it is crucial that the user understand the consequences of incorrect data entry (e.g., incorrect spelling or punctuation, entering data in an inappropriate search field, or attempting to search fields not indexed at all).
The difficulties of subject retrieval in machine searching are due in part to the discrepancies of different systems with variant indexing designs."1 Other problems with subject retrieval are influenced by cataloged' choice of headings, coupled with changing practices and cancellations and revisions to LCSH. This is illustrated by a plea from a graduate student: "My only complaint is that you be consistent-put all items dealing with one subject in that subject-e.g. [the terms] Aspen & Populus index [the] same thing, but sep[arate] items pop up in each."
Cross-referencing (as suggested repeatedly) and updating of subject headings at the local level would help to alleviate this problem, but not all libraries can afford these luxuries. Until authority control systems become more widely available and practical for academic libraries, we will have to hope users and reference librarians can cope with the anomalies of our current systems.
RECOMMENDATIONS AND CONCLUSIONS
Even if not all users seek access by geonames, it might be fair to conclude that, for the population within the Rocky Mountain Region, attention given to providing such access by catalogers is well worth the effort. Ours is a region of considerable interest to geologists and geographers. There is substantial literature on the region, and in many cases if catalogers do not provide the access there may very well be none, especially for materials in regional publications. Libraries in other regions of interest to the earth sciences may find themselves in a similar position.
A multifaceted approach taken by librarians, researchers, and designers of information technology would ameliorate the difficulty of providing increased bibliographic access to geologic and geographic names. Catalogers need to provide appropriate and sufficient access points and, together with public services librarians, foster effective searching skills in users. Since access is also dependent on the design of the system itself, librarians should further direct their attention to systems and system-suppliers. Increasing demand for and manufacture of such features as user-friendly interfaces, online thesauri and authority control, and flexible searching capabilities may eventually reduce users' dependence on knowledge of standardized forms.
Cataloging
Catalogers and other technical services librarians can develop an awareness of system design and newer technologies, promoting and implementing the most effective ones. Librarians can also benefit greatly from establishing contact with agencies such as BGN and LC for clarification of policies and conflicts in setting up geo-names.
Further, regarding the form of name for catalog entry, it may be more important at the local level to make variants or commonly used forms accessible than to agonize over what the "correct" form is when it is not readily available from LC or BGN. However, adherence to LC or other national library authority standards for formulation continues to be a priority at the network level. In many cases this may not pose a problem, but there will always be a certain number of situations, such as the basin/watershed discrepancy discussed earlier, in which simultaneous satisfaction of user needs and national standards may create direct conflict. In such cases, each library may wish to arrive at a compromise to connect the needs of users with the available system. The keyword approach to online searching, bypassing the use of controlled vocabulary terms, might be espoused as one solution to this dilemma. Lois Mai Chan remarks in Library of Congress Subject Headings: Principles and Applications:
The power of keyword searching on words in titles alleviates to a certain extent problems with lack of currency or failure to reflect common usage. Many subject indexes rely entirely on keyword display based on article titles or abstracts. Titles of documents that are expressive of content increase the effectiveness of keyword title access. Furthermore, this benefit is not available only to online catalog users; generous use of partial-titleadded entries, to bring potentially useful access terms into filing position, can have much the same effect in the manual environment. Helpful as keyword searching can be, however, few theorists advocate abandoning the use of controlled vocabulary in library catalogs. Keyword access to terms in titles can be an adjunct to systematic indexing, but no more.22
Automated authority control systems and patron-accessible authority files may be a partial solution in the not-too-distant future for larger academic or research libraries. LCSH does not explicitly list all possible unique heading plus subdivision combinations. In an analysis of topical and geographic headings in the University of Michigan Library's catalog to determine degree of match with the 10th edition of LCSH, Carolyn Frost and Bonnie Dede find a higher than expected agreement (88.4 percent) of headings without subdivisions; however, a comparison of subdivisions alone reveals only a 31 percent match of those in LCSH, which might have been higher if a file of free-floating subdivisions were also matched. They also show that "less than 12% of the geographic subdivisions were found in LCSH" and acknowledge that "geographic subdivisions, however, will pose a problem unless free-floating lists of some kind can be developed for these as they have been for topical subdivisions."23
MARC formats provide some alternative tagging strategies. For example, in fields 650 and 651 (for books and serials) OCLC allows second indicator 7 with subfield 2 to define authorities for subject terms other than the national authorities. Locally formulated subject terms are valid on OCLC in fields 690 and 691, with the stipulation that they are "not part of the standard LCMARC formats" for original records.
Despite these provisions, LC subject authority resulting in 650 and 651 fields with the second indicator 0 will be the most attractive option for most libraries. Use of the tags for local subject terms may entail unnecessary complexity in setting up profiles for automated systems, and this can contribute to problems in standardization for shared access when several libraries plan to merge data for union catalog purposes. Use of tags for local subjects also affects the level of programming required. In these situations, the mandate to accept LC forms of subject terms and to formulate original terms by LC rules saves programming effort and implementation headaches.
User Education
Developing user awareness may be the most immediate and potentially most effective means of guiding users through the intricacies of bibliographic systems and library materials arrangement. Such an effort should involve both technical and public services in bibliographic instruction in conjunction with site-specific user surveys to learn more about those interested in geo-name access and how best to serve their needs. It may also be worthwhile for cataloging departments to develop an awareness of which departments at their institutions are most likely to use geo-name search strategies and to direct bibliographic instruction appropriately.
Indeed, enhanced bibliographic instruction would go a long way toward getting users to realize the potential of current library indexing systems, regardless of their particular fields of study. Libraries still having divided systems should be sufficiently staffed to make sure people are using the tools appropriately. Complete conversion needs to remain a top priority. If users' first interactions with the library are successful, there is a greater likelihood they will make subsequent visits.
Technology
Improved technology applied to sophisticated, integrated library systems can complement quality cataloging and bibliographic instruction. Authority control and online thesauri should not be overlooked as beneficial tools for users and librarians alike, possibly with user-friendly interfaces that connect these to the databases. Current experiments with incorporating tableof-contents data into online systems containing keyword access, if successful, may eventually obviate the cataloger's concern with the pitfalls of LCSH.
A certain degree of caution must be observed, however, when expecting users to undertake searching such a file without appropriate training and understanding of the indexing protocols. For example, a keyword query "'Bismarck" might retrieve the following array of results, including author, title and subject entries: Bismarck, Otto, Fürst von, 1815-1898; Bismarck, Herbert, Fürst von, 1849-1904; Bismarck (N.D.); Bismarck Archipelago (PapuaNew Guinea); Bismarck Bange (Papua New Guinea); Grain Transportation Forum, Bismarck, N.D.; Okubo Toshimichi, the Bismarck of Japan; Bismarck-Museum; Bismarck National Bank (Bismarck, N.D.); Bismarck Bank (Bismarck, N.D.); Bismarck family; Bismarck (Ship); and, Bismarck Sea, Battle of, 1943.
Obviously, the main problem in searching without a controlled vocabulary is the retrieval of items unrelated to the desired topic. One way of reducing false drops is a flexible system that enables users to apply Boolean logic, and qualifiers, providing they understand how these strategies work and what the situations demand.
Further Research
Issues such as those mentioned above necessitate more survey research and analyses to find out more about the users themselves-the strategies they employ and why they use them; to what extent they perceive bibliographic instruction to be helpful; what happens in cases of failed searches; how well they understand bibliographic systems and the application of truncation, Boolean logic, and qualifiers; how they react to table-of-contents and other keyword indexes; and whether they are equally adaptable to both controlled and uncontrolled vocabularies.
The topic of user instruction raises several questions, such as whether bibliographic instruction is effective and how it might be more so; what user groups at what levels it can benefit the most; whether automated authority files and user-friendly systems lessen the need for instruction; how to ascertain searchers' abilities and thus direct instruction appropriately; how technical services librarians might contribute jointly towards this effort; and what other means are available to promote user awareness about library systems. Can we guarantee that users find what they need by providing them with enough skills to effectively use library catalogs rather than relying on the sheer persistence of a small percentage of users?
Questions such as these are beyond the scope of this paper but are fertile ground for continued investigation. We hope the potential to better accommodate users can be realized by designing improved systems and ensuring that the confidence in being able to use them is ultimately transferred as automatically as the data itself.
ACKNOWLEDGMENT
This paper represents the fruits of a study conducted over the last two years. A very brief report of the first two surveys appeared in a regional newsletter, Action for Libraries (February and March 1987 issues) . The surveys and resulting study were made possible by a grant from the Office of Research, University of Wyoming and in part by the authors' affiliated institutions. The cooperation extended by the cataloged, geo-science faculty, and graduate assistants who responded to the surveys should be recognized, as should the patience of many library staff members in allowing the authors time to prepare the analysis.
REFERENCES
AND NOTES
1. Arthur Keith, "The Status of Geologic Names," in Philip S. Smith, "The Geological Society of Washington," report of a paper presented at the 224th Meeting of the Geological Society of Washington, December 8, 1909, Science n.s. 30, no.783:975 (Dec. 31, 1909).
2. Geraldene Walker and Judith Hudson, "Research Methodology in Technical Services: The Case of 1987," Library Resources Lr Technical Services 32, no.4:360 (Oct. 1988).
3. Judith A. Adams, "The Computer Catalog: A Democratic or Authoritarian Technologt'?" Library Journal 113, no.2:33 (Feb. 1. 1988).
4. Loi Mai Chan, "Library of Congress Classification as an Online Retrieval Tool: Potentials and Limitations," information Technology and Libraries 5, no.3:191 (Sept. 1986).
5. Chan, "LC Classification as Retrieval Tool," p.181-92.
6. Karen Markey, Subject Searching in Library Catalogs: Before and After the Introduction of Online Catalogs (Dublin, Ohio: OCLC, 1984).
7. Mary Micco, et ah, "Knowledge Representation: Subject Analysis," Library Software Review 6, no.2:84 (March-April 1987) citing George Martin Sinkankas, et ah, A Study of the Syndetic Structure of the Library of Congress List of Subject Headings (Pittsburgh: University of Pittsburgh, Graduate School of Library and Information Science, 1968).
8. Carolyn O. Frost and Bonnie A. Dede, "Subject Heading Compatibility Between LSCH and Catalog Files of a Large Research Library: A Suggested Model for Analysis," Information Technology and Libraries 7, no.3:288-99 (Sept. 1988).
9. Karen Markey, "Integrating the MachineReadable LCSH into Online Catalogs," "Special Section: Online Subject Access," information Technology and Libraries 7, no.3:299-312 (Sept. 1988).
10. James Ross, "Geographic Headings Online," Cataloging h Classification Quarterly 5, no.2:27 (Winter 1984).
11. William E. Studwell, "Subject Suggestions 1: Some Concerns Relating to Geographic Areas," Cataloging b Classification Quarterly 8, no.2:93 (1987/88).
12. Carolyn O. Frost, "Faculty Use of Subject Searching in Card and Online Catalogs," Journal of Academic Librarianship 13, no.2:86 (May 1987).
13. Association of Research Libraries, Office of Management Services, Systems and Procedures Exchange Center, "User Surveys," SPEC Flyer 148 (Washington, D.C.: Association of Research Libraries, 1988).
14. Library of Congress, Subject Cataloging Division, Subject Cataloging Manual: Subject Headings, rev. ed. (Washington, D.C.: Library of Congress, 1985).
15. Library of Congress, Processing Services, Cataloging Service Bulletin 11-12, 15-16, 21-22, 28, 30, 34, 38-39 (Winter 1981-Winter 1988).
16. United States Board on Geographic Names, Decisions on Geographic Names in the United States (Washington, D.C.: Dept, of the Interior, 1963 ).
17. Frost, "Faculty Subject Searching," p.90.
18. Library of Congress, Processing Services, Cataloging Service Bulletin 34:46 (Fall 1986).
19. The University of Wyoming online catalog is a locally designed public catalog and database management system developed with IBM software licensed by Brodart. It has been available to the public since 1984.
20. E.J. Coates, "Significance and Term Relationships in Compound Headings," in Theory of Subject Analysis: A Sourcebook, ed. by Lois Mai Chan, Phyllis A. Richmond, and Elaine Svenonius (Littleton, Colo.: Libraries Unlimited, 1985), p.193-94.
21. Carol A. Mandel includes some graphic examples showing how system design affects subject retrieval in Multiple Thesauri in Online Library Bibliographic Systems: A Report Prepared for Library of Congress Processing Services (Washington, D.C.: Cataloging Distribution Service, Library of Congress, 1987), p.74.
22. Lois Mai Chan, Library of Congress Subject Headings: Principles and Application, 2d ed. (Littleton, Colo.: Libraries Unlimited, 1986), p.352.
23. Frost and Dede, "Subject Heading Compatibility," p.297.
Copyright American Library Association Dec 1989