EPA-RS72-002
August 1972
Socioeconomic Environmental Series
Information Retrieval Capabilities-
National Environmental Research Center,
Cincinnati, Library
ui
CD
-------
EPA-RI-72-002
August 1972
Information Retrieval Capabilities-
National Environmental Research Center,
Cincinnati, Library
M. Friedman
National Environmental Research Center
Cincinnati, Ohio 45268
Project Element 1D2065
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
PREFACE
The library at the National Environmental Research Center,
Cincinnati, is the scientific and technical information
focal point for the entire Environmental Protection Agency
Library system. As such, it provides technical information
back-up to all other EPA libraries, in addition to its
responsibilities to NERC-Cincinnati.
The Cincinnati Library has access to a number of information
centers which maintain over 25 scientific and technical
computerized data bases, and contains over 2.5 million doc-
uments. We are now providing literature searches with our
Datapoint 3300 CRT computer terminals, using a variety of
on-line and off-line data base services. These services
allow the library to provide retrospective searches on de-
mand as well as current awareness programs tailored to the
individual's requirements.
This manual describes these information services in more
detail. Most of the services discussed herein are now
available for usage. Some others will be installed and
operational in a few months.
Future developments in our information retrieval activities
will be designed to enhance the depth and breadth of coverage
and the variety of services provided to EPA staff members
requiring access to scientific and technical publications.
-------
ACKNOWLEDGEMENTS
We acknowledge with appreciation the permission of the
University of Georgia to use the sections on weighting and
logic, some of which is based on the material found in
"Profile Coding and Management Manual for the University of
Georgia Text Search System"
Preparation of the document has been aided by the significant
assistance provided by the entire staff of the NERC-Cincinnati
Library, in particular, Maxine M. Smith, Dottie F. Brofft, and
Mary L. Calkins.
-------
CONTENTS
SECTION
TITLE
PAGE
PREFACE
I.
INTRODUCTION
The Search System
Data Bases
Types of Searches
1
1
4
II.
SEARCH STRATEGY 12
STEP 1. User Interview 12
STEP 2. Concept Identification 15
STEP 3. Search Strategy Formulation 16
A. Subject Coverage 16
B. Data Elements 17
STEP 4. Concept Expansion 17
STEP 5. Profile Refinement 23
a. Truncation 23
b. Weighting 26
STEP 6. Profile Coding 30
Forms 32
Group Number (GRP) 35
Term Weight 36
DEM (Data Element Code) 36
Search Terms 37
III.
SUMMARY
42
Appendix A
Appendix B
Appendix C
Appendix D
APPENDICES
Data Base Description
Price Information
Sample Printouts
Library Organization Chart
A-l
B-l
C-l
D-l
-------
LIST OF FIGURES
FIGURE NAME PAGE
1 NERC-CINCINNATI INFORMATION NETWORK 2
2 Boolean Logic 14
3 CAS Search Guide 20
4 Concept Expansion 21
5 Term Truncation Modes 24
6 Computer Search Information Request 32
7 Profile Coding Logic 33
8 Typical Boolean Expressions 39
9 User Evaluation Form 41
-------
U. S. ENVIRONMENTAL PROTECTION AGENCY
NERC - CINCINNATI LIBRARY
MANUAL ON COMPUTER SEARCHES
I. INTRODUCTION
The traditional library resources of EPA have been augmented
by access to various data bases for computer searching of
the literature. With the ever-increasing amounts of published
literature, it is necessary to have a more efficient means
of access and comprehensive coverage of these sources of infor-
i
mation. Computers capable of rapid screening of material
answer this need and provide a necessary supplement to library
services. The concept of an EPA Information Network was devel-
oped and expanded as exemplified in Figure 1.
This manual is designed mainly for use by librarians in under-
standing computerized searching, and discusses only currently
available or soon-to-be-implemented data bases. Procedures
described herein will be applicable to all present and future
accessible data bases.
The data bases which the EPA library can or soon will be able
to access are located at the University of Cincinnati, Bio-
Sciences Information Service (Biological Abstracts), Battelle
Memorial Institute (AMIC), Lehigh University (Leadermart),
National Library of Medicine (MEDLINE), Informatics Inc.
(Toxicon and Environ), Ohio State University and the University
of Georgia. These information centers and their systems are
described in more detail later in the manual.
1
-------
INFORMATION REQUEST CENTERS
INFORMATION ACCESS CENTERS
Battelle
EPA Regional Offices
EPA NERC's
NERC-Cincinnati Librar
EPA Labs
EPA Headquarters
EPA Contractors
University of Cincinnati
Ohio State
Informatics
•«
NLM
s
University of Georgia
NERC-CINCINNATI INFORMATION NETWORK
FIGURE 1
-------
The University of Georgia has the most data bases, and is usually
accessed for retrospective searches. Lehigh University, Battelle
Memorial Institute, and National Library of Medicine are searched
on-line from NERC-Cincinnati to their respective computers. The
computer terminals for these are located in the EPA Library,
Columbia Parkway, Cincinnati. Access to the terminal is available
by telephoning the EPA Library, using Commercial or FTS lines
(513/871-1820, ext. 492, 493, or 494), by Xerox telecopier (auto-
matic) (513/871-1820, ext. 249), by TWX (810/461-2796), or by
mail addressed to the library. Forms have been provided in order
for the librarian to clearly state the search strategy and selected
terms. These forms should always be completed according to the
instructions provided and sent to the library. Telephone conver-
sation should be limited to supplementing the information on these
forms or for unusual situations.
The NERC Library will supply hard copy or microfiche for retrieved
references if the requesting library cannot obtain them locally.
Appendix A. contains a brief overview of the major data bases
and a listing of the subject areas each data base covers. This
appendix may be duplicated and distributed separately to libraries
for distribution to their subscribers.
This is not to be considered as the final selection. As experi-
ence and changing requirements dictate, new data bases will be
added. Additional systems will continue to be investigated for
possible inclusion. This manual will be revised to show any
changes which may occur during this initial period of development.
-------
Appendix B. contains information on costs for all data bases
to be searched. These costs are broken down to include charges
for each information center and each data base located at the
University of Georgia.
3A
-------
The NERC-Cincinnati library will notify all interested EPA
personnel of additional data bases, search services, price
changes, and new library services as they are developed.
The types of searches provided are:
A. ON-LINE SEARCHING
On-line searching queries data bases instantly
and provides immediate response between the user
and the data base. Data base coverage depends
on the amount of information in the computer
and usually is limited to the last few years.
Besides quick response, on-line searching is
recommended as a way of testing for the presence
of relevant material, or for refining a search
strategy before going further or deeper into a
data base.
MEDLINE is one example of our on-line retrieval system.
Each article which enters the system is indexed under
a number of subject terms. In retrieval, Boolean logic
is used, and terms are linked using "and", "or" or
"and not". Exact terms from MESH (Medical Subject
Headings) must be used. Besides using subject terms,
material may be searched by authors, languages, journal
titles and year.
-------
LEADER is the name of the on-line retrieval system
at Lehigh University. Each document which enters
the Lehigh University on-line system, with its
affiliated noun phrases, is processed by a series
of special programs which:
1. Isolate noun phrases for each document,
2. Isolate the component word stems of each
phrase,
3. Create inverted files to link: word-stems
to noun phrases, phrases to documents,
phrases to associated phrases.
The on-line system has various data bases, based on
the different document sources. Currently available
are document corpora from:
1. Chemical Abstracts - This data base is de-
rived from Chemical Abstracts. Currently,
coverage includes from v. 72, 1970 thru
current issues. Total documents of CA in
system approximately 550,000 (Covers 12,000
journals).
2. Engineering Index (COMPENDEX) - Corresponds
to the published Engineering Index volumes,
and contains mostly applied engineering
references. Currently, coverage includes
from 1970 to date. Total number of documents
-------
approximately 120,000 with abstracts each
10 lines (Covers 2,250 journals).
3. Civil Engineering - Recently initiated
containing select articles pertinent to
civil engineering.
4. MARC - A new on-line Library of Congress
card catalog system is available for de-
veloping computer-produced catalog records.
An English language monograph can be cat-
aloged by using the L.C. card number,
author, or title.
This is primarily a form of current awareness, on-line
system, which provides quick retrieval and can be the
basis for developing a more comprehensive batch retro-
spective search.
Other data bases will be added as soon as feasible.
AMIC (Analytical Methodology Information Center) is a
sample of a new data base. Since it covers water
quality analysis it should prove to be quite valuable
to many EPA laboratories.
-------
B. CURRENT AWARENESS
Current awareness covers either the latest issue
and/or each successive current issue from a data
base and can be used to give a continuous flow of
information from the computerized data base.
The current awareness service available through
the library consists of a profile developed for a
researcher that will be matched against the latest
editions of data bases.
The profile is based on the researchers informa-
tional needs and coded to reflect retrieval terms.
On a bi-weekly schedule this profile is matched
against multi-discipline data bases containing
journal articles and government reports. The output
is printed on cards and sent to the requester.
The cards consist of two parts - the smaller portion
can be sent to the library as an order form and the
larger portion (now a 3 x 5 card) can be filed by
the researcher. A profile can be adjusted at any
time to reflect changing requirements or professional
interests.
A profile can be developed to combine terms, to
weight them, or to eliminate irrelevant output.
The following sample profile shows how terms can be
selected, logic devised and terms weighted. Sample
-------
cards have been selected to show the kinds of
retrieval that can be expected from a current
awareness service.
Co RETROSPECTIVE SEARCHING
Retrospective searching can go back as far as the
user desires or to whenever the data base was
established. Turnaround time for retrospective
searches (known as a "batch mode") may be two to
three weeks due to the size of the files. However,
search results may be forwarded throughout this
period as they are received rather than waiting
until completion of the search.
Search results are usually mailed within two days
after receipt of the tape. Retrospective searches
are scheduled over a two-week processing cycle.
Turnaround time for retrospective batch searches
from receipt of initial profile (search question) to
delivery of final search results may run two to
three weeks. Search results are mailed throughout
this period, however.
Special demand searches on frequency less than this
two-week cycle may be requested. These are accommo-
dated only at higher special rates for which the
requester assumes the entire cost of computer runs.
-------
Each question should be a fully descriptive paragraph
describing the subject area to be searched and any
qualifications, such as particular applications
which are to be excluded. Suggested index terms or
words frequently used in the research area under
consideration are especially helpful. This step will
be explained later in this manual. The search question,
or profile, may be modified by the user, in order to
obtain better search results. Maximum number of
terms which can be processed in one search question,
or profile, is approximately 200, depending on the
length of the terms.
The search result is a bibliography for retrieved
documents. The output includes the title, authors'
names, location of work (if available on the data base),
primary document citation, and all index codes or
index terms available on the data base. Output on
8-1/2 x 11" one-part paper is standard output. Output
on 4 x 6" card stock is available if desired, at
additional cost.
Any combination of these services can be requested as
well as any combination of data bases and years
covered. These searches can be tailored to suit the
users' needs and are extremely flexible. With exper-
ience these searches can produce a high rate of user
satisfaction and relevancy in retrieved information.
-------
In order to obtain maximum results from a computer search,
it is necessary to state the question clearly and to provide
additional information using synonyms and related terms.
Guidelines on developing profiles to serve as keys to unlock-
ing relevant information in the data bases will be discussed.
EPA forms have been provided for reporting this data and
this manual provides detailed instructions for completing
these forms. These EPA forms need not be thoroughly com-
pleted but should include whatever information can be provided.
The library staff at NERC-Cincinnati will develop the profile
or will call the user for more data, if necessary, to complete
the profile. It is important that care be taken in preparing
these profiles as they represent the most significant ingred-
ient in a successful computer search. Each librarian should
maintain a close liason with the user and the Cincinnati
personnel in order to increase his/her knowledge on how best
to use computer searches. The critical role played by the
librarian will be the ultimate test as to the potential worth
of the data base retrieval systems.
Another consideration is that more than one source has the
same data base. Each librarian should decide on which source
and information center can be of advantage — based on years
covered, user's time constraints and costs — before report-
ing any preferences. The NERC-Cincinnati staff will make
every effort to keep costs and turnaround time to a minimum.
10
-------
If new services become available which will offer an
advantage (e.g. better price structure, faster turnaround,
more extensive coverage) then all participating locations
and organizations will be so informed.
Appendix C contains sample printouts that a user might
receive as a result of a computer search from the inform-
ation centers being used.
Appendix D contains a chart of the interim organization of
the NERC-Cincinnati Library and the names of staff members
and their responsibilities.
Study of this manual should lead to an awareness of the user's
and the librarian's role in properly accessing the vast amount
of information now available. If there are any questions,
please do not hesitate in writing or calling NERC-Cincinnati
for assistance.
11
-------
II. SEARCH STRATEGY
The following steps are designed to provide a logical
sequence of events for successful retrieval of information.
The summary provides a one-page outline of these procedures
for quick reference.
STEP 1: USER INTERVIEW
To obtain pertinent references from the computer search, a
clear concise statement of the question is essential. As
the person who will benefit most from the search, the user,
who will have the best understanding of his requirements,
should be consulted. A discussion with him will help
formulate a clear statement of need. He should also be
questioned for other terms related to the search and synonyms
used in the literature with which he is familiar. The result
should be a well-conceived, written statement or paragraph
complete with synonyms arid related terms. If the user does
not prepare this, then the librarian should write it with
the guidance of the user.
One specific point to be kept in mind while recording the
question is the use of the conjunction "and". It is
frequently interpreted as meaning the same as logical "or"
as well as the connective "and". Consider, for example,
the question, "I am interested in the effect of A on B and
C." There are two potential questions represented:
12
-------
1) the effect of A on B OR the effect of A on C, and 2) the
effect of A on B in the presence of C. These questions
imply different logic strategy, so care must be exercised
in clearly stating the proper question. As a general rule,
restrict the use of "and" in recording the question. It
is better to develop short succinct statements than complex
sentences. If possible, several abstracts or references
relevant to the search should be included.
Figure 2 shows some examples of how a search request can be
interpreted as a Boolean logic statement. Although this
table shows only two (2) sets (A and B), more complex logic
may be utilized. For example: A | B -» E signifies the request
for set A or B through E. A & B & -» F signifies A and B but not
F. Try to keep the logic as simple as possible and to use more
search statements. This approach will be far more efficient
for computer operations.
At this time, a brief description of the various data bases
should be given to the user, so he will have an idea which
data base or bases will be most pertinent to his question. The
scientists will be familiar with most of these in their field
such as Biological Abstracts, Chemical Abstracts, Engineering
Index, etc., and can decide which would have the best chance
of giving results.
13
-------
BOOLEAN LOGIC
Logic Statement
Boolean Statement
Venn Diagram
Search for a single
setj(sometimes[ called1
"descriptors" )(e.g., A
Search for the
presence of A and B
A & B
^\^ "i
AlB
Search for anything
excluding set A
Search for anything
containing set A
or set B .
A/B
Searchjfor anything
containing set A orjset
B_but exclude) where
both are present,
A &7B//B &7A
NOTE: The above illustrates the most cotmonly used logic in a search statement.
For best results do not develop more complex statements. In this illustra-
tion the rectangle represents the universal set and the circle represents
the logic set in question. The shaded areas represent the specified set(s)
for retrieval. "AND" is shown as &; J'pR" is shown as |, and "NOT" is
shown as ~i .
FIGURE 2
14
-------
As most data bases cover only recent years, the importance
of retrospective coverage should be decided. Be sure all
searches are limited to the years covered by the data bases,
lest the librarian find she might be forced to commit her
own time for manually supplementing the computer output.
Cost may also be a factor as retrospective searches are more
expensive since charges are usually based on a per volume
or per year basis. A current awareness search covering only
the most recent issue in a data base may be run to test the
profile before requesting a comprehensive search.
STEP 2: CONCEPT IDENTIFICATION
Once stated, the question should be examined to identify
the main concepts. In general, all nouns, including the
objects of prepositions, represent a concept. For example,
the question, "What papers deal with the analysis of fatty
acids from tissues of mammals?" contains four (4) concepts.
A good test is to try to reconstruct the question by con-
verting adjectives to prepositional phrases. If the result-
ing question is logically correct and noun forms of the
adjectives exist, the adjective represents another concept.
Adjectives modifying nouns should be treated as part of
the noun they modify. For example, "All papers concerning
the use of acapella choirs in secondary schools" has true
modifying adjectives.
15
-------
STEP 3: SEARCH STRATEGY FORMULATION
The term "search strategy" is used to mean the number of
search term groups and the Boolean logic relationships
between the search terms. The strategy controls the way in
which search terms will be combined to determine whether
or not a given document is an answer to the profile. Several
factors go into formulating the search strategy which will be
used in the profile. The search strategy depends on the sub-
jective judgment of the profiler. Experience and analysis of
the search results should lead to be best strategy.
A. Subject Coverage
The nature and content of the data base itself will
govern the strategy. If one or more of the concepts
covers all of the subject scope of the data base, then
it can be eliminated from consideration. For example,
the concept "geology" in the question, "All papers
dealing with the geology of the hilltop lake area of
Ontario" is redundant with the subject scope of
Bibliography of North American Geology: the data base
contains only geology by definition. A question which
is to be coded for both CBAC and CA-Condensates may be
coded with a different strategy for each since CBAC
covers only biochemical papers while the CA-Condensates
coverage is very broad.
16
-------
B. Data Elements
The data elements available in the data base will also
affect the precision recall balance; and therefore,
the strategy to be used. A data base such as Chemical
Titles—which contains only titles—must have questions
phrased much more broadly and with fewer concepts than
does a data base which contains abstracts and/or index
terms. Profiles which are to be run against title-only
data bases should seldom include more than two concept
groups. If the question contains more than two concepts,
the librarian, in collaboration with the user, should
select the two most important concepts. With indexed
data bases, especially when the vocabulary is thesaurus-
controlled, the number of groups may be three or four,
or higher, if necessary. Precision of the search and
recall of relevant documents are inversely proportional;
if there are few groups, the recall is better, but the
precision suffers. The user should specify his require-
ments of retrieval;it is then up to the librarian to
construct the profile to meet the user's needs.
*
STEP 4: CONCEPT EXPANSION
*
Concept expansion is a translation of the question from the
language of the user to the language of the information file.
Each concept identified in the question must be expanded to
a list.of synonyms and related terms or replaced by thesaurus
17
-------
terms or codes. Each concept will normally form one group in
the final logic expression with the terms in_ the group linked
by "or" logic. Groups are usually linked by "and" or "not"
logic although there are no restrictions to these logic operators.
Depending on the data elements available in the data base
for search, two- conditions may exist. The data element may
have a completely uncontrolled vocabulary. That is, no
i
restriction is placed on the words which may be used. This
is commonly the case with titles and abstracts where the
choice of words is left up to the author or abstractor.
(Sometimes this approach is called "free-text searching.")
Vocabulary-controlled data elements, on the other hand,
imply use of some authority list or thesaurus for approved
words or terms. Index terms and subject classification areas
or codes are frequently thesaurus-controlled. Expansion of
the concept must be done differently for these two types of
data elements in order to effect retrieval.
For uncontrolled vocabulary data elements, such as titles
and abstracts,,it is necessary to construct a list of .
synonyms or related words for the concept term. In essence,
the librarian with the user's help tries to predict all
possible ways the concept may have been expressed by authors.
In areas with which the librarian is unfamiliar, a basic
text can provide background and general knowledge of the concept.
18
-------
It may also suggest synonyms and related concepts which should
be included in the term list. Search guides, such as the
CAS Search Guide (Figure 3) which contains narrower terms (NT),
related terms (RT), synonyms (SY or S*), and broader terms
(BT), are valuable as memory joggers.
19
-------
January 1967
SEARCH GUIDE
Amino Resins
A-Stage Resins
BT Thermosetting Resins
RT Phenol-Formaldehyde Resins
SY Resols
Abandonment
RT Depletion
RT Life (Non-Biological)
Abatement
RT Damping
RT Dilution
RT Disposal
RT Purification
RT Reduction
Abdomen
RT Gut
Abdominal Dropsy
S* Ascite's
Aberration
RT Distortion
RT Images
RT Interference
RT Lens
RT Reflection
RT Refraction
Abies Sibirica
BT Abis
BT Evergreens
BT Pinaceae
SY Siberian Evergreen
Adietic Acid
RT Rosins
Abiogenesis
RT Biogenesis
SY Spontaneous Generation
i
Abiogenic Synthesis
BT Synthesis (Chemical)
Abiosis
S* Death
Abis
NT Abies Sibirica
Ablating Polymers
BT Polymers
RT Ablation
Ablation
RT Ablating Polymers
Burning
RT Combustion
RT Cooling
RT Melting
RT Vaporizing
Abnormal Growth
RT Hypertrophy
Abnormal Response
RT Sensitivity
Abnormalities
RT Diseases
Abnormalties
RT Irregularity
RT Malformation
Abo Factor
RT Glood Groups
Abo Incompatibility
RT Blood Platelet
Abode
RT Habitat
RT Home
RT Residence
Abrasion
RT Abrasive Blasting
RT Abrasives
RT Erosion
RT Friction
RT Gilling
RT Grinding
RT Metallography
RT Perforating
RT Polishing
RT Rubbing
Figure 3. CAS Search Guide
20
-------
Figure 4 shows the concept "chroraatography" as it might be
expanded for a chemistry search in the uncontrolled vocabulary
data base CA-Condensates. Abbreviations, acronyms, and alter-
nate spelling must all be taken into account if editorial
policies of the particular data base permit them.
Chromatography
Chromatographic
Chromatogram(s)
Chromatog
gas-liquid Chromatography
gas liq chromatog
gas Chromatography
vapour phase Chromatography
vapor phase Chromatography
thin layer Chromatography
paper Chromatography
V.P.C.
VPC
T.L.C.
TLC
G.L .C .
GLC
G.C.
GC
Figure 4 Concept Expansion
The task of concept expansion is similar for thesaurus-
controlled data elements except that all terms must be taken
from the authority list. Instead of predicting the author's
choice of words, the librarian must know the thesaurus and
have a working knowledge of the indexing policies used in the
assignment of index terms. The hierarchy of index terms
must be taken into consideration in formulating the term list.
For example, all papers in ERIC reporting occupational educa-
tion of employees would not be retrieved by the broader term
21
-------
"personnel" since "employees" is also a term. In other words,
there is no hierarchical posting to higher or lower level terms.
Data element types can be mixed during expansion of concepts.
A biological profile might include text terms (title, abstract,
and free index terms), biosystematic codes, cross codes, and
Registry Numbers, many as synonyms for each other. This usually
occurs when a given profile is to be searched against two or
more data bases which have different data elements. Librarians
should use some judgment as to whether it is better to write
separate profiles for each specific data base or to combine the
profiles using data base specific data elements as synonyms.
The search time is proportional to the number of search terms,
and the amount of core required for the profiles is determined
to a large extent by the number of different data elements.
Consequently, the economics of the computer search would dis-
suade the use of large numbers of non-productive terms and/or
data elements. On the other hand, profile maintenance of a
question which is run in several data bases can be difficult
to control if many different profiles must be modified or
updated. The Choice should be based on the degree of similarity
between the profiles which would be written independently.
Remember, too, that the data bases themselves may be dissimilar
enough to warrant different strategies for the same question.
22
-------
During the process of expanding the concepts, it is frequently
necessary to revise the search strategy formulated in the
preceding step. Some concepts are almost impossible to expand
into any reasonable set of terms which will effect retrieval.
In such cases it is usually better to omit the concept from
the coding. This may cause some irrelevant retrieval, but
there is less chance that relevant papers will be missed be-
cause of an inadequately expanded concept. It may also be
better to revise the strategy to exclude irrelevant material
than to formulate the strategy to select only precisely
relevant documents. This negative approach is particularly
ueful when the data base is vocabulary-controlled or has been
classified into subject areas. Techniques for refining both
the profile terms and the search strategy are discussed in
the next step.
STEP 5: PROFILE REFINEMENT
There are several techniques available for refining both the
search terms and the search strategy. These include term
truncation and weighting. It is also possible, where a data
base is available on line, to dp a short search to test the
validity of the selected terms before running a more expensive
and lengthy retrospective search.
a. Truncation
Term truncation is a technique used to retrieve terms
containing word fragments which are common to two or
23
-------
more different forms of the word. For example, the
singular and plural or adjectival and adverbial forms
can be retrieved by a single term. Thus, the term
*OXID* will retrieve DEOXIDATION, OXIDIZED, OXIDATION,
OXIDE, OXIDES, etc. The asterisk (*) placed before
and/or after the word fragment is interpreted by the
search program to mean "accept any character, includ-
ing a blank, as valid in the position shown as asterisk."
Figure 5 illustrates the four possible modes of trunca-
tion and examples of each.
i
Term Explanation
CELL Retrieves terms only if bounded by blank
spaces or punctuation; e.g., CELL, CELL-FREE,
STEM-CELL.
CELL* Retrieves the term followed by any character,
including a blank space and preceded by
blank spaces or punctuation; e.g., CELL,
CELL-FREE, CELL., CELLP1/4JAME, CELLS,
CELLULAR, etc.
*CELL Retrieves the term preceded by any character,
including a blank space and followed by a
blank space or punctuation; e.g., CELL,
CELL-FREE, STEM-CELL, MICROCELL, etc.
*CELL* Retrieves the term preceded and/or followed
by any character, including blank spaces and
punctuation; e.g., CELL, CELL-FREE, ACELLULAR,
MICELLES, CELLULOSE, etc.
Figure 5 Term Truncation Modes
If carefully used, truncation can significantly decrease
the number of individual search terms which have to be
listed.
24
-------
It can also improve retrieval in that word forms which
might otherwise be overlooked will be automatically
retrieved. It can also be dangerous if misused. Word
roots carelessly truncated can cause large numbers of
irrelevant documents to be retrieved. The choice of
appropriate truncated terms is aided by reference to
KWIC indexes in the printed publications or to word
frequency dictionaries available for some data bases.
A good rule-of-thumb is to avoid word roots of three
characters or less with both prefix and suffix trunca-
tion. There are occasions when such small word roots
can be used effectively, but they are rare.
Although there are four possible modes of truncation,
not all modes can be used for all data elements. Each
data element is classified as one of two types: left-
anchored (LA) or free text (FT). The classification
determines the manner in which the data element is
searched. For left-anchored elements, the match always
begins with the left-most character. This implies that
the format and content of the data element are known.
Examples of this type include Registry Numbers, Cross
Codes, authors' names (inverted), and thesaurus-codes.
Only suffix (right hand) truncation (or no truncation)
can be used with left-anchored terms. The only restric-
tion is that the truncation asterisk cannot immediately
follow any special character, including a blank.
25
-------
These are the first two modes shown in Figure 5. The
second class of data elements character strings in the
data base. Examples of this type data element are
titles, abstracts, and free index terms. All four
modes of truncation are applicable to free text fields.
b. Weighting
In a free text search system such as this, not all the
profile terms used in concept expansions are equally
significant with respect to the question. Consequently,
a technique called "weighting" is available as an
optional feature to assign relative numeric values to
the terms. There are two principal uses of weighting,
each of which will be discussed in greater detail.
I
One use is to sort the answers into order of probable
importance. The second use is an extension of the
search logic.
A suggested simplified approach to weighting can be
taken by ranking the selected terms. Give each term
or rank a number, starting from 1 as the most important,
2 the next most, etc. More than one term can have the
same rank of importance. The Technical Reference
Librarian at Cincinnati receiving the request will
interpret the ranking for computer searching. Initially,
a librarian may want to avoid weighting and should treat
26
-------
this aspect of a computer search only as an optional
feature. With more experience the librarians can
prepare profiles involving the more complex approach
described below.
Weights for computer searching are positive or nega-
tive numerical values assigned to search terms.
Numbers must be in a positive or negative five figure
range. In a weighted search, the terms within a group
may have different weights, and some terms may have no
weight assigned. No specific order of weighted terms
is required. A weight of zero (0) is assumed unless
otherwise specified. A threshold weight can also'be
assigned to the question. It is not necessary to code
weight with leading zeros.
Use of weighting for sorting (Threshold weight equal
00000)
Weighting may be used to sort the search results by
assigning relative term weights to the index terms. The
most important words (according to the user) are assigned
the highest numeric values, and less important terms are
given lower values. The sum of the term weights for a
document may be added together to give a question weight
for that document. Output search results are sorted for
27
-------
each user in descending question weight. Thus, the
documents which hit on the most important term(s) will
have the highest weight and will sort first, followed
by the remaining citations in descending order. For
searches in which weighting is used only for sorting,
the threshold weight for the profile is left blank (i.e.,
has a value of zero).
If the threshold weight for the question is zero, then
weighting is calculated non-cumulative. That is, the
weight assigned to each term is counted only once for
each citation, regardless of the number of times the
term occurs in the citation. The results are sorted in
descending order on the question weights calculated as
the sum of weights (non-cumulative) for all terms which
"hit"., The non-cumulative weight is printed on the out-
put. This type of weighting (no threshold weight) is
used exclusively for ordering output.
Use of weighting as logic (Threshold weight not equal
00000)
Weighting may also be used as an extension of the "and",
"or", and "not" logic. "Not" logic, for example, is
absolute. If a document hits on a "NOT" term, the docu-
ment will not be retrieved even if it also contains
desired terms. By use of term weights and a threshold
28
-------
weight for the question, undesirable terms alone will
not cause retrieval, but the document will be retrieved
if desired terms are also present. In other words, a
threshold weight may be assigned to a profile such that
the sum of the hit term weights must be equal to or
greater than the threshold weight.
It should be emphasized that weighting of terms is an
optional and not a necessary technique for developing
search strategies. Only those with some experience or
knowledge of a data base should consider using weighted
terms to any extent. Furthermore, weighted terms should
only be used for current awareness or retrospective
search and not for on-line systems.
For example, if the threshold weight for the question is
not equal to zero (either positive or negative), weighting
is calculated cumulatively for all occurrences of all
terms in the data base. This allows a simulation of
"ignore" logic by judicious use of positively and nega-
tively-weighted terms when truncation would otherwise cause
problems. The cumulative weight is used to test against
the threshold weight to determine whether or not the
threshold weighting is satisfied. The search results are
sorted in order of the non-cumulative weights for all
terms which "hit". The printed weight for the question,
29
-------
however, is the cumulative weight. Consequently, the
search output may not always appear to be in strict
order by question weight.
Negative weighting can be used in place of "NOT"
terms, especially where a few words consistently cause
problems in conjunction with truncation. Consider the
term *ASE which is frequently used in biochemistry for
retrieving enzyme names. However, it also retrieves
words such as "decrease," "increase," "phase," etc.
If the term *ASE is assigned a positive value (e.g., 100)
and the terms DECREASE, INCREASE, PHASE, etc. , are
assigned an equal but negative value (e.g., -100) with
the threshold weight of 100, then a paper reporting an
increase of oxidase still will be retrieved. If only
an undesired term matches, then the question weight
would be negative and less than the threshold weight
necessary to cause retrieval.
STEP 6: PROFILE CODING
Once the search strategy has been formulated, the concepts
expanded to search terms and refinements such as truncation
and weighting noted, the profile is coded for computer search
using the "Computer Search Information Request," EPA 2170-1,
abd tge "Profile Coding Logic" form EPA 2170-2, (Figures 6 and 7).
30
-------
If possible, the current awareness files should be searched
first to test the profile. The profile may need to be
modified before any retrospective searches are made.
If the computer search originates from a laboratory outside
Cincinnati, these two forms "Computer Search Information
Request" form and "Profile Coding Logic" form, together
with sample abstracts, sample references, or any other
material which will help the search are to be sent to:
Environmental Protection Agency
NERC-Cincinnati
Cincinnati, Ohio 45226
Attention: Technical Reference Librarian
31
-------
Page of
COMPUTER SEARCH INFORMATION REQUEST
Date
MAIL TO:
Environmental Protection Agency
NERC-Cincinnati
Cincinnati, Ohio 45226
Att: Technical Reference Librarian
TO BE COMPLETED BY ORIGINATING OFFICE:
LAB NAME
USER'S NAME PHONE NO.
USER'S NO. USER'S 1st REQUEST ( ) YES ( ) NO
LIBRARIAN PHONE NO.
ACCOUNTING SYMBOL
MAILING ADDRESS: _
(to which
material should •
be sent)
DETAILED STATEMENT OF REQUIREMENTS; Please describe, as specifically
as possible, the subject matter for which the search is to be con-
ducted. Define any terms that may have special meaning. If there
are points NOT to be included, please state these. Include language
statement (English and/or non-English). Use additional page if
necessary.
THRESHOLD WEIGHT BOOLEAN EXPRESSION
ON-LINE SEARCH: YES NO BATCH SEARCH: YES NO
DATA BASES TO BE USED AND YEARS COVERED:
AUTHORIZING SIGNATURE DATE
TO BE COMPLETED BY NERC-CINCINNATI LIBRARY:
USER'S PROFILE NO. DATE REC'D. DATE COMP.
LIBRARIAN
(SIGNATURE)
EPA Form 2170-1 32
-------
Page of
PROFILE CODING LOGIC
This form is to be completed whenever the originator desires to
develop his/her own coding logic as a means of expediting the
search. Please attach this to the Search Information Request.
User Name
User No.
Librarian's Name
Date
Please rank the terms in the order of importance.
GRP WEIGHT DEM TERM
Request Reviewed: Date
Signature
EPA Form 2170-2 Figure 7
33
-------
The person preparing the forms should keep one copy of the
request at least until such time as a printout is received.
The other two copies should be sent to NERC-Cincinnati, using
the address printed on the "Computer Search Information Request"
form, (EPA Form No. 2170-1).
In preparing the "Computer Search Information Request" form,
the first part should be completed at the originating office.
The name of the laboratory or organization, the name of the
user, his/her phone number, the mailing address for the print-
out are self explanatory. If a librarian is completing the
form, then the librarian's name and phone number are to be
entered. User number is for the originating organization's
internal use. The appropriate answer for first time usage
should be checked. Accounting symbols refer to the account
number against which the search will be charged. Each re-
questing organization will be required to pay for all external
expenses incurred in the course of answering a search request.
The detailed statement of requirements should be completed
next. If desired, (and only as an option) the user or orig-
i
inating librarian may enter a threshold weight and a Boolean
logic expression. If this option is not exercised, the
NERC-Cincinnati staff will complete the form or will call the
originator for; clarification.
34
-------
The search questions and data base questions must be answered
as fully as possible. The authorizing signature is for use by
the user's division chief, or equivalent, and is necessary
only to discourage ephemeral usage of the retrieval system.
The form can be folded so that the address of NERC-Cincinnati
can be seen through a window envelope. The bottom portion will
be completed by the NERC-Cincinnati staff and one copy will be
returned with the printout. If a series of printouts are to
be sent back, the form will be attached only to the last one to
indicate the completion of the requested search or profile.
The "Profile Coding Logic" form (EPA Form No. 2170-2) is optional
and may be used with discretion. However, if it is completed,
it will greatly expedite the request, as it will be entered into
the computer system with only a check for accuracy and complete-
ness. In addition to saving time and producing faster turnround,
the librarian or user who completes this form will most likely
obtain greater accuracy and less irrelevant output. Thus, the
completion of this form to the extent possible is urged upon each
and every user.
Each search term entry consists of up to four types of data.
The Group Number, labeled GRP on the form, is a four character
string, consisting of the letter G and three digits (e.g. GOOD.
Leading zeros must be coded in all cases. Group numbers are
used to link all terms for a given concept. All terms within
35
-------
a given group are linked by "or" logic by definition.
However, a group can consist of a single term, as well as
many terms. There is no limitation on the number of terms
which may be included in any group. However, there is a
maximum of 240 terms or 1920 characters (whichever occurs
first) which may be used for the entire profile. Group
numbers must be assigned sequentially, starting with G001.
No group number may be omitted.
The term weight can be up to 6 characters. This concept was
previously discussed in this manual. The weight number can
contain both a plus and minus sign or be preceded by a minus
sign if appropriate. The plus sign is assumed and is never
coded. Unless weighting is indicated as being desirable,
the absence of any reference to this feature will result in
all terms being given equal value. Weighting is usually not
necessary for most searches.
DEM represents the Data Element Code. This is a three char-
acter acronym for the data element type which is being searched.
The valid acronyms and their corresponding data element names
will be furnished upon request. The value of using the DEM can
be exploited only after gaining experience with profiling.
In addition to the specific data element codes given in the
chart, there is one general code, TXT, which may be used to
indicate that the title, abstract, and free index terms are
36
-------
to be searched. This will undoubtedly be the most commonly
used DEM. Unless otherwise indicated, TXT will be assumed.
The fourth element is the search term itself. Truncation
is indicated by asterisks preceding and/or following the
characters of the search term. Note that an asterisk can
never appear within a search term, only as the initial or
final character. Ther term may consist of alphabetic,
numeric, or punctuation characters. However, all punctuation
characters will be converted to blanks in the search, both
in search terms and the data base. Punctuation should be
included where appropriate since these characters (hence the
corresponding blank spaces) are counted in the overall term
length. Terms which are identical except for punctuation
(e.g. NMR and N.M.R.) need be coded only once. The maximum
length of a term, including truncation asterisks and imbedded
blanks, is 59 characters. Be sure to use only one term per
line. It is better to use more paper for describing a search
than take the chance of confusing the person entering the
request into the computer.
The Boolean Expression is a statement of the search logic.
It consists of the group numbers, any combination of the
logic operator symbols "&", " ", and "S"1", and parentheses.
"Not" logic is indicated by the two symbols "&" without inter-
vening punctuation. Another convention which can be used
37
-------
are the key words "and", "or", and "and not" enclosed within
periods (e.g., .and., .or., and .and. .not.). Full Boolean
logic operations are permitted. The maximum length of the
Boolean Expression is 180 characters. The "or" operation is
the inclusive or; that is, A B is true if either A or B or
both are true.
Every profile must have a Boolean Expression, even if there
is only one group. Parentheses can be used to improve
readability or to alter the order in which the logical oper-
ations are performed. Without parentheses, the logic opera-
tions are performed in the order "not" equal to "and" greater
than "or", working from left to right for operations of equal
value. For example, in expression (4) in Figure 8, the "and"
operation between G002 and G003 would be performed first,
followed by the "or" operation with G001. Thus, the two ex-
pressions (3) and (4) are equivalent. Expressions (7) and (8)
are different, however. In (7) the "and" operation between
G001 and G002 is executed first followed by the "or" with G003.
Operations within parentheses are executed before all other
operations. The parenthesis in (8) causes the (G002 j G003)
part of the expression to be evaluated first, followed by the
"and" operation with G001. If in doubt about the order in
which operations will be performed use parentheses.
38
-------
(1) G001
(2) G001 & G002
(3) G001 | (G002 & G003)
(4) G001 | G002 & G003
(5) G001 & G002 & G003
(6) G001 | (G002 & G003) & G004
(7) G001 & G002 G003
(8) G001 & (G002 G003)
(9) G001 .and. (G002 .or. G003) .and. .not. G004*
*Although available, this type of logic notation is not ordi-
narily used because of the time and space involved.
Figure 8 Typical Boolean Expressions
The last line on the form labeled "Request reviewed" and date
is for use by the Cincinnati librarian who has reviewed the
request. Feel free to photocopy or Xerox these forms until
the official version is available. The library staff at
Cincinnati is available to answer any questions about enter-
ing a request or completing these forms. Write or call if
help is desired. In return, if it is found that incomplete
information or ambiguities will delay a request from being
answered, a librarian in Cincinnati will call for clarification
rather than simply returning the request.
Upon completion of the computer search, the requesting
librarian should complete the User Evaluation Form (sometimes
called "feedback" as shown in Figure 9). The purpose of this
is to allow NERC-Cincinnati to determine how to improve service
where necessary.
39
-------
NERC-Cincinnati will inform all participants in this network
of any changes that may occur. If the users would like
NERC-Cincinnati to publish information such as successes, prob-
lems, suggestions, search results, etc., please let us know.
As of now the "Library Scene" (New Additions) will carry a list
of the successful searches and will supply copies upon request.
Free subscriptions and additional copies are available upon
request and are encouraged as a way of publicizing these
computer searches.
40
-------
Page of
USER EVALUATION
Lab Name
User Name
User Profile Number
Librarian's Name
Search Question
How do you rank the search? (check one)
(1) Very Good
(2) Good
(3) Fair
(4) Poor
Suggestions and/or recommendations for improvement
Figure 9
41
-------
III. SUMMARY
A typical search strategy should follow these ten (10)
procedures:
1. User Interview to obtain clear statement of search
question with list of synonyms and related terms.
Decide on data bases to be checked and years to be
covered. Include related abstracts or references
if possible.
2. Concept identification.
3. Search strategy formulation.
4. Concept formulation.
5. Profile refinement.
a. Truncation
b. Weighting
6. Fill out profile coding form. Keep copy.
7. Fill out search information form. Keep copy.
8. Keep records by user name or own numbering system.
Allow a maximum of two weeks for return printout
when requesting large batch search services.
9. Check with uer when search is returned.
10. Send in user evaluation form.
42
-------
APPENDIX A
DATA BASE DESCRIPTION
1. Bibliography of North American Geology.
Published since 1887, the B4.bi4.ogia.phy ofi Noith
Ge.otogy was first computer-produced in 1965 for the 1961
annual volume. The kb&tia.c.t& o£ Noith Ame.i4.can Geology
i
began publication in 1966. Ge.ophy&4\c.a.£ Ab4£tac.£&, a
related publication by the U.S. Geological Survey, was
first published in 1929, and ceased publication in
December, 1971, the date through which the B4.bt4.ogia.phy
ojj Month Amet'tca.n Gzotogy will also cease. (For liter-
ature coverage subsequent to 1971, use the GEO.REF data
base)
This bibliography lists publications concerning the
geology of the North American continent, Greenland, the
West Indies and adjacent islands, Hawaii, Guam, and other
island possessions, but not the trust territories of the
United States. Articles by American authors published
in foreign journals are cited if they deal with North
American localities or are of a general nature, but not
if they deal with foreign areas. Articles on North
America by foreign authors are included regardless of the
place of publication, while those of a general nature are
included only if they appeared in North American journals.
A-l
-------
2. Biological Abstracts-Previews (BA-P).
The magnetic tape of B4.otog
-------
assigness; bibliographic citations; and phrases from
the CA issue Keyword Index. There is no accompanying
printed version of CA-Con.de.nAate.&. The service is
designed to provide the user with the capability to
scan rapidly by title and key-word the pertinent con-
tent of some 250,000 new articles and patents each year.
Publication of the tapes began with Vol. 69 (July 1968).
Two volumes of 26 issues each are published per year.
4. Cataloging and Indexing (CAIN).
The initial CAIN tapes (3 reels) designated "CAIN 1969",
contain 42,667 records. The January, 1970, tape records
are included in the CAIN 1969 volume. Approximately
5,500 tape records are produced monthly, depending on the
number of publications received and processed by the
National Agricultural Library.
Approximately 11,000 publications are covered by CAIN.
Tape records are produced monthly and are divided among
18 broad subject areas providing bibliographic data con-
cerning the literature of agriculture and related sciences.
Journal articles, pamphlets, government documents, includ-
ing USDA, state experiment stations, and extension service
i
publications, special reports, etc., received by the NAL
are sources for entry of data on the CAIN tapes.
CAIN contains about 100,000 records to date.
A-3
-------
5. Chemical-Biological Activities (CBAC).
C/ie.mxtca£-8/to£og/cca£ Ac^u/c^e^ (CBAC) is a biweekly tape
service supplied by Chemical Abstracts Service containing
digests of current technical articles related to the field
of biochemistry. CBAC covers the scientific literature
that reports the interactions of organic compounds with
biological systems as well as related metabolism and
in-vitro chemical studies.
About 600 journals are referenced in CBAC per year. Two
volumes of 13 issues each plus Cumulative Indexes for
each volume are published each year, dating from January,
1965. A total of 110,000 digests were publlished in the
first 12 volumes (Jan. 1965 - Jan. 1972); an average of
9100 digests/volume.
The separate printed copy of CBAC ceased in December 1971.
It is contained in the first five section groupings of
Chemical Abstracts, the Biochemistry Sections.
6. Chemical Titles (CT) .
Che.mj.c.at Tj.tf.e.& (CT) , available from January 1962, pub-
lished biweekly, is a current author and keyword index
of recently published articles from approximately 700 of
the world's most active chemical and chemical engineering
journals.
A-4
-------
Articles are input using only words in title, author(s)
name(s) and journal reference. There are 26 issues per
!
year with approximately 5,000 titles cited per issue.
i
7. Computerized Engineering Index (COMPENDEX).
!
C0MPEWPEX is the computerized version of the Eng-cneei/cng
i
Index. The COMPEWPEX data base, containing the input to
the E/t Monthly for each issue from January 1969 to the
present, provides rapid access to over 7,000 abstracts
monthly, reflecting the contents of 3,500 sources of
worldwide! technical information and the work of 88,000
authors, i It covers all fields of engineering and certain
related fields of science and management, pertinent
i
quality research and applications literature; assembles
the various engineering disciplines side by side with
interconnecting cross references.
i
i
Users have the option of searching the complete record
(full text search) or searching selectively on chosen data
elements: Subject heading and subheading, document ID
number, author(s) Ei abstract number, codes assigned to
i
the CARD-A-LERT service divisions, on most records the
access words (which are cross references chosen for the
printed version of COMPENPEX), and on most records the
i
free language terms.
i
Abstracts are included.
A-5
-------
8. Current Information Tapes for Engineers (CITE).
The CITE tapes for Plastics and Electrical/Electronics
were first issued in 1965. The service originally called
the U.P.P. (User Participation Program) became, in 1968,
the CITE service, or Current Information Tapes for Engin-
eering. The Plastics Monthly and CITE service were
continued until December 1969. The E/E Section was dis-
continued in December 1967.
The CITE tapes containing input in Plastics Engineering
of some 30,000 articles covering the years 1965 through
1969 and some 60,000 in Electrical and Electronics
Engineering for the same time period, are for retro-
spective searches only.
9. Educational Resources Information Center (ERIC).
The ERICTAPES comprise the magnetic tape copies of the
Educational Resources Information Center (ERIC) files,
consisting of 30,000 report resumes from 1966 on, 12,000
journal article citations from 1969 on, and a thesaurus.
The Report Resume Files consist principally of resumes of
research reports filed by contractors and grantees on the
results of funded educational research. The Journal Article
Resume Files contain resumes of journal articles on educa-
tional research selected from over 500 education and educa-
tion-related journals. The Thesaurus File is a complete
Thesaurus of ERIC Descriptors from which subject indexing
terms are selected for both report and journal article resumes.
A-6
-------
Under special authorization from the U.S. Office of
Education, these tapes are available only by purchase
from LEASCO Systems & Research Corporation.
10. Environmental Protection Agency (EPA) funded data bases;
a) Analytical Methodology In^oimoit4.on Co.nt.ifi (AMIC) .
The AMIC computerized data base was established in
1971 by Battelle, Columbus Laboratories, Columbus,
Ohio, under contract for the EPA, to inform the
personnel of the Analytical Quality Control Laboratory
(AQCL) and other segments of the National Analytical
Methods Development Research Program of recent publi-
cations on methodology.
Each month AMIC adds to its data base approximately
200 indexed and abstracted current reports on methods
for determining the indentity, concentration, and
ecological effects of pollutants, and measuring water
quality. Subject areas covered by AMIC include biology,
chemistry, instrumentation, microbiology, and quality
control .
The current awareness bulletin, Review* oj
L4.teJia.tu.tie. on knatytj.c.a.1 Methodology and Qu.alj.ty Control,
containing the abstracts is published monthly. There
are now over 3200 documents in the data base.
A-7
-------
10. b) Environmental Information Retrieval On-line (ENVIRON).
ENVIRON is a new EPA system which replaces the
Technical Information Management Planning System (TIMPS)
established in 1970 by the Research Information Divi-
sion, Office of Research & Monitoring, E.P.A.,
Washington, D.C.
This system utilizes a software package STIMS/RECON
which was developed by Informatics, Inc., and will
be run at the computer facility of COMNET (Computer
Network), Washington, D.C. EPA users throughout the
country will have telephonic access to the data bases
at the COMNET facility in an on-line real time
environment.
The data base contains pertinent information on all of
the research and monitoring grants, proposals, projects
and final reports generated by or for EPA. This
includes:
Proposed R & M Projects
Active R & D Projects
EPA R & D Project Report System
Solid Wastes Projects
Water Quality Surveillance Network
Marine Biological Literature
Industrial Waste Literature
Oil and Hazardous Materials (Oil spill data:
TADS (Technical
assistance data
system)
A-8
-------
c) EPA library files.
The library files will consist of references to all
EPA reports and publications, including predecessor
agencies and the journal holdings of the EPA libraries.
The book catalogs of EPA libraries/ as soon as they
are available, will also be added to these files.
d) Solid Waste Information Retrieval System (SWIRS) .
The SWIRS data base covers the period from 1964, in-
cluding both periodical and non -periodical literature
of the world on solid waste management; for the period
1941 through 1963, see Public Health Service Publica-
tion No. 91 and Supplements A through F, which contain
all-inclusive solid waste management bibliographies.
The current awareness bulletin, Acce44>ton
Soii.d Watte. In^otma-tton Re>fUeva£ Sy&te.m, is published
monthly to keep readers abreast of the solid waste
management information being published. There are
over 13,000 abstracted items in this data base, con-
taining a description of the contents of the documents
being added to SWIRS and giving a complete biblio-
graphic .citation.
A-9
-------
11. Geological Reference File (GEO.REF).
GEO.REF is the data base originated by the Geological
Society of America, in cooperation with the American
Geological Institute. The GEO.REF tapes are available
beginning with 1967. The bibliographic information
dates from 1965, however. The printed publication
corresponding to this data base is B/cb-LtogAapfoj/ and
Index OjJ Ge.ology.
This data base covers world-wide literature on geology
and related technologies selected from over 2,000
serials and special publications. Entries are classi-
fied under 21 categories, which include such disciplines
as Geochemistry, Geochronology, Paleontology, Stratig-
raphy and Structural Geology.
12. Geophysical Abstracts (GPA).
GPA data base covers the years 1966 through 1971, when
Ge.oph.y&
-------
For literature coverage subsequent to December 1971, see
GEOoREF data base.
13= Institute for Scientific Information (ISI) Science Citation
i
Index.
The ISI Sc.-ce.ncfc Ci.ta.tion Index tapes (source and citation)
cover the current literature of science and technology
(including the social sciences). They contain biblio-
graphic information about the source articles, and also
about the references cited in those articles, which are
covered in the five sections of Cu.JDie.nt Conte.nt&, approxi-
mately 8,000 current items each week.
Dating from 1964, coverage is comprehensive and includes
all items (other than advertisements and ephemeral
notices) in approximately 2,500 key scientific journals.
File size (1971): 2,538,000 records (source items)
with approximately 410,000 items added per year.
File size (1971); 32,000 items (cited references) with
approximately 5,000,000 added each year.
14. Machine Readable Catalog (MARC II).
MARC is the production of LibJiaJiy o£ Congie-64 cataloging
data in computer readable form. The format was imple-
mented in June 1968. At present, MARC II cataloging data
is limited to English language monographs which also
A-ll
-------
includes government reports/documents, and conference
proceedings. Plans for expansion to cover monographs
in French and German are being studied. MARC is
available on-line and will be used as the basis for
NERC-Cincinnati library providing scientific and
technical cataloging service to any EPA library that
so requests.
A-12
-------
15. MEDLINE
MEP/.INE (acronym for MEPLARS On-Line) is a new service
from the National Library of Medicine. It provides
on-line bibliographic searching of biomedical journal
literature. The data base includes more than 400,000
citations with index terms from 1,138 of the most
important periodicals indexed for Index Med hospitals and research institutions.
A-13
-------
16. National Technical Information Service (NTIS).
The U.S. Gove.inme.nt Repot-ti Announcement (USGRA) and
U.S. Goveinmen-t Repo.t;t4 Index (USGRI) , formerly called
the U.S. Govei.nmen;t Reaea/tcrt and Ve.ve.topme.nt Repo-t/ta
(USGRDR), tapes are issued twice monthly, 24 issues
per year; approximately 50,000 document records are
added per year.
The National Technical Information Service {NTIS),
formerly the Clearinghouse for Federal Scientific
and Technical Information (CFSTI), began making
available magnetic tapes of its U.S. Government
Research and Development Reports (USGRDR) beginning
with the January 10, 1970 issue. With Volume 71,
Issue No. 6, March 25, 1971, the name of the NTIS's
announcement journal changed from U.S. Gove.Jin.me.nt
Re.6eaic.rt and Ve.ve.lopme.nt Re.poit& to Gove.tinme.nt
Repot^tA Announcement (GRA) , and includes technical
publications of the Department of Commerce which are
not research and development reports, as well as the
research and development reports previously being
covered. A similar change was made in name to the
U.S. Gove.finme.nt Re.6ea-t.ch and Ve.ve.topme.nt Repot;t6 Index
(USGRDR-I) , which became Gove.fi.nme.nt Re.polt& Index (GRI)
The USGRA tapes cover new U.S. Government-generated
research and development reports and translations of
A-14
-------
17. Nuclear Science Abstracts (NSA) .
Nu.c.le.0.*. Science Ab&tia.c.tA (NSA) is a semimonthly publi-
cation of the USAEC Office of Information Services,
published by the USAEC, Technical Information Center.
There are 24 issues/year or volume.
A total of 450,000 abstracts were published in the first
21 volumes (1948-1967). The annual volumes now contain
approximately 47,000 abstracts each.
NSA, as a hard-copy publication, began in 1948. The first
computer files available began with Volume 20, Issue No. 12,
(mid-1966) . NSA provides the only comprehensive abstract-
ing and indexing coverage of international nuclear science
literature. It covers scientific and technical reports of
the U.S. Atomic Energy Commission and its contractors,
other U.S.1 Government agencies, other governments, univer-
sities, and industrial research organizations. In addition,
books, conference proceedings, individual conference papers,
patents, and journal literature on a worldwide basis are
abstracted and indexed.
Printed Publications and Aids for using Nuc.te.cUi Science
are:
TID-4552 (Rev. 5), Su.bje.ct Scope. o& Muclecu. Science kb&t>ia.c.tt>
TID-4577 (Rev. 3), Peac/Up^tcue Ca&tlog-cng Guide.
A-15
-------
TID-5001 (8th Rev.), Subject Heading* lUed by tku> and Ca.te.goiy L^
Nuclear
IAEA-INIS-13 (Rev.) IMIS:
the.
the. Indiana. U. Tape^ at IRL - Be^.fee£ej/, May 1970.
TID-4578, Sub^fcC^: Modsifae.*. Gu/ccte
A/SA Entiy and Se.le.c.toi file. fox.ma.tA
TID-85, Report Nu.mbe.ft. Se.Ji-ie.4 U^&d by the. VL kutkoiity L
-------
18. PANDEX - CURRENT INDEX TO SCIENTIFIC & TECHNICAL LITERATURE.
PAWPEX - Cu.tui<>.nt Indzx. to ScUen£c<^c 6
provides a comprehensive data base from 2,100 scientific,
technical and medical journals; 6,000 scientific technical
books, 5,000 selected patents, and 50,000 U.S. Government
technical reports annually. It covers all areas of pure
and applied science, with coverage beginning in 1967. Each
entry record contains manually indexed or manually edited
thesaurus-generated subjects. PANPEX serves research re-
quirements not satisfied by single disciplinary indexes but
provides access to all literature of interest to a disci-
pline as well.
The index' generated by PAWPEX from this data base is
divided into two parts: (1) Subject-entries containing
full title, primary author and periodical reference,
arranged by all significant subject words and subarranged
by all significant secondary words; (2) Author-entries
containing all authors arranged alphabetically (primary
author entries containing full title, all other authors,
and the periodical reference) (secondary author entry
containing name of primary author and periodical reference) ,
The magnetic tape files are prepared by CCM Information
Corporation and issued weekly. The printed hard copy
publication is issued bi-weekly.
A-17
-------
19. Psychological Abstracts.
P&yc.hotog
-------
articles taken from the original papers. SPIN also
contains the reel and frame number of the first page
of each article published in full text in Cut/ien-t
P/i£/4-tc4 M^.cA.OjJo^m (CPM) . Information centers sub-
scribing to both SPIN and CPM can provide a current
awareness alerting service for material chosen to
meet individual interest profiles with the full text
backup for AIP published journals from the CMP
microfilm.
21. Sociological Abstracts.
Soc.4.otoQ4.c.OLt Ab&tJia.c.t&, published under the co-sponsor-
ship of the Eastern Sociological Society and the Midwest
Sociological Society, covers a broad range of sociological
articles in periodicals in various languages. Frequency
varies, but since 1964, the abstracts have been published
9 times a year. The printed abstracts date from 1952.
Currently, the data base tapes include Volumes 13, 14,
and 15 (1963, 1964, and 1965). Tapes are being prepared
for Volumes 11 and 12 (1961 and 1962) and later, Volumes
16, 17, and 18 (1966, 1967 and 1968) will be added.
A-19
-------
22. TOXICON.
The new toxicology information system, entitled TOXICOM,
is an extensive collection of computerized toxicology
information and data. It originates from the Toxicology
Information Program of the National Library of Medicine.
TOXICON provides on-line literature searching and is
designed to serve health professionals working in the
areas of environmental pollution, industrial or occupa-
tional health and safety, pharmacology, toxicology,
medicine, agriculture and other bioscientific disciplines.
The data base contains citations primarily, most includ-
ing abstracts and/or indexing terms, full text state-of-
the-art reports and toxicity data.
Among the data bases included in TOXICOW are:
-The. Haye.A f-lte. on
Includes more than 10,000 citations of reports on
health aspects of pesticides with a limited number
of index terms and abstracts. (EPA, 1940 to 1966)
-HAPAB (Hzatth A^pec^A orf Pe.&ticlde.A kb&ttiac.t Bu.tte.t-in)
Includes more than 8,000 abstracts of reports of the
health aspects of pesticides in humans and animals;
poisoning treatment, pesticide residue analysis and
monitoring.
(EPA, 1966-current)
A-20
-------
-KEEP (Ab&ti
-------
COMPUTER DATA BASES
AGRICULTURE
a. CAIN
BIO-MEDICAL
a. Biological Abstracts
b. BioResearch Index
C. MEDLINE/MEDLARS
d. *TOXICON
.Hayes File
.HAPAB
.KEEP
e. TOXITAPES
CHEMISTRY
a. *CA Condensates
b. *Chemical Biological Activities
c. Chemical Titles
EDUCATION
a. ERIC
b. .Current Index to Journals in Educ
.Research in Education
ENGINEERING
a. CITE
b. *Compendex
10
11,
12
ENVIRONMENT
a. *AMIC (Water quality)
b. SWIRS (Solid Waste)
GEOLOGY
a. Bibliography of North American
Geology
b. GEO.REF.
c. Geophysical Abstracts
LIBRARY SCIENCE
a. EPA Libraries' Holdings
b. *MARC
MULTI-DISCIPLINE
a. ISI Source Index
b. PANDEX
PHYSICS
a. Nuclear Science Abstracts
b. SPIN
RESEARCH REPORTS
a. *NTIS
SOCIAL SCIENCES
a. Psychological Abstracts
b. Sociological Abstracts
NOTE: *A11 or partially available On-Line.
-------
APPENDIX B
This Appendix contains the price schedule based on present
University of Georgia schedules. Lehigh University and
Informatics charge $45 per terminal hour and a small extra
charge for printing, handling, and mailing. MEDLINE services,
both searching and off-line printing, ENVIRON, and AMIC are
free at the present time.
GEORGIA INFORMATION DISSEMINATION CENTER PRICE LIST
July, 1972 - June, 1973
CURRENT AWARENESS SEARCHES
These searches keep the user up to date on the current liter-
ature. Current awareness searches are run whenever current
issue or update tapes are received from the supplier. Search
results are usually mailed within two days after receipt of
the tape.
RETROSPECTIVE SEARCHES
These searches are scheduled over a two-week processing cycle.
Turnaround time for retrospective batch searches from receipt
of initial profile to delivery of final search results will
normally run two to three weeks. Search results will be mailed
throughout this period, however.
B-l
-------
Special demand searches on frequency less than this two-week
cycle will be accommodated only at higher special rates for
which the requester assumes the entire cost of computer runs.
Estimates for special searches are available on request.
SEARCH PROFILES
Information Specialists are available in the Center to con-
struct the search profiles. There is no charge for this
service. The question should be a fully descriptive para-
graph describing the subject area to be searched and any
qualifications, such as particular applications which are
to be excluded. Suggested index terms or words frequently
used and the names of authors who publish in the research
area under consideration are especially helpful. Please
include a telephone number at which the person requesting
the search can be reached for additional information or
clarification if necessary.
Changes to the profile will be made immediately upon receipt
of a request by a user. The changes are not retroactive,
however, for searches which have already been run. There is
no charge for changes to the profile. The maximum number of
terms which can be processed in one profile is approximately
200, depending on the length of the terms.
B-2
-------
SEARCH RESULTS
The search result is a bibliography of retrieved documents.
The output includes the title, authors' names, location of
work (if available on the data base), primary document
citation, and all index codes or index terms available on
the data base.
Output on 8-1/2 x 11" one-part paper is the standard output.
Output on 4 x 6" card stock is available for 2<= per card in
addition to the prices quoted for the standard output on
paper. Abstracts will be printed for those data bases in
which they occur for an additional charge of IOC per abstract.
SUBSCRIPTION INFORMATION
Searches are priced on a cost-recovery basis by the Georgia
Information Dissemination Center. The prices are established
on a per-profile per-search run basis. For current awareness
searches the price is per-profile per-issue of the data base.
For retrospective searches the prices are set on a per-profile
per-volume searched basis. Each data base is searched
separately, so that one profile searched against two data bases
would be invoiced as the sum of the prices for the two files.
i
Partially completed volumes within the current year are billed
at the current awareness price (i.e., the issue price times the
number of issues in the partial volume). Postage and shipping
charges are billed extra for shipments outside the United States.
Users requesting foreign delivery should specify the preferred
method of shipment.
B-3
-------
GIDC Price List, July 1972 - .June 1973
Users are billed by itemized statement monthly. Searches may
be entered or discontinued at any time without penalty.
Subscription discounts, where applicable, will be passed on by
NERC-Cincinnati. In order to obtain the discount the organiza-
tion must guarantee the minimum subscription. Unused balances
are not refundable.
CURRENT AWARENESS SEARCHES
Data Base
CA-Condensates (CA)*
CAIN
Compendex**
Current Index to
Journals in Ed. .(CUE)**
GEO.REF**
GRA (USGRDR)**
Nuclear Sci Absts.
Psychological Abstracts**
Research in Ed. (RIE)**
SPIN**
racts (BA)
ex (BIORI)
(CA)*
Frequency
of Updates
semimonthly
monthly
weekly
monthly
monthly
Price per Profile
Per Issue
$"7
7
7
10
10
quarterly
monthly
semimonthly
semimonthly
quarterly
quarterly
monthly
10
10
5
5
10
10
10
* Odd-numbered or even-numbered issues only may be specified.
** Abstracts available.
B-4
-------
GIDC Price List, July, 1972 - June, 1973
RETROSPECTIVE SEARCHES
Data Base No
Biological Abstracts (BA)
1969 -
Bibliography of North America**
Geology 1961-1968
BIORI (BRI)
1969 -
CA-Condensates-Even Issues
July 1968 (Vol. 69)
CA-Condensates-Odd Issues
July 1968 (Vol. 69)
CAIN
1970 - i
The file prior to 1970 is
treated as one volume
CBAC**
1965-1971
Chemical Titles
1962-1968
CUE **
1969 -
CITE E/E only
1968-1969
COMPENDEX **
1969 -
Geophysical Abstracts **
1966-1970
GEO.REF **
1967 -
GRA (formerly USGRDR) **
1970
NSA
June 1966 -
Psychological Abstracts**
1967 -
RIE **
1969 -
The file prior to 1969 is
treated as one volume
Sociological Abstracts **
1965 - 1967
SPIN**
June 1970 -
.Vols/Year
2
1
1
2
2
1
2
1
1
1
1
1
1
1
1
1
1
1
1
Price per Profile
Per Volume
$ 40
35
80
40
40
70
35
70
35
35
100
35
100
70
70
35
35
50
70
** Abstracts available
B-5
-------
It should be noted, however, that NERC-Cincinnati will be
testing and evaluating other services. Some information
centers, government agencies, and other organizations may offer
their services at reduced prices from time to time, for trial
periods and special data bases. We will endeavor to investi-
gate each and every situation to determine if changes in our
present operations are warranted, or if some of these new
services should be added. For example, a limited number of
searches are available from BIOSIS whereby a search of their
files containing 12 years of references, and their abstracts,
are available for 20 cents a citation. This arrangement is
limited, but if demand warrants, another contract may be
negotiated.
B-6
-------
APPENDIX C
SAMPLE PRINTOUTS
The following represent a typical printout that a user
might receive:
All output that is larger than 8*5 x 11 will be reduced
to that the user need not worry about oversized papers.
Figure C-l is an example from the University of Georgia.
Figure C-2 is a page from Lehigh University.
Figure C-3 represents a current awareness profile; and
Figure C-4 is a sample of three cards that were produced.
C-l
-------
PROFILE # 019840-001 JUN. 02, 1972
ACCOUNT # THRESHOLD WT . 00000
DATA BASE SEARCHED: USGRDR V 72:01-09
GROUP
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
G001
GOO 2
GOO 2
GOO 2
GOO 2
G002
GOO 2
GOO 2
GOO 2
G002
GOO 2
G002
GOO 2
G002
GOO 2
G002
GOO 3
GOO 3
GOO 3
GOO 3
G004
G004
G004
TERM
NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
(G001&G002I
TYPE
TXT
TXT
ESC
ESC
TXT
TXT
EIT
EIT
EIT
EIT
EIT
EIT
EIT
TXT
TXT
TXT
TXT
EIT
TXT
TXT
TXT
TXT
TXT
TXT
TXT
ESC
ESC
ESC
TXT
TXT
TXT
TXT
EIT
EIT
TXT
TXT
TXT
TXT
TXT
G003
WEIGHT
01000
01000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00002
00002
00002
00001
00001
00002
00001
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
00000
JE-.G004
TERM ;
1
WASTE WATER* '
WASTEWATER*
452
187
SEWAGE*
SLUDGE*
SEWAGE ANALYSIS*
SEWAGE BACTERIOLOGY*
SEWAGE FILTERS*
SEWAGE PUMPING PLANTS*
SEWAGE PUMPS*
SEWAGE TREATMENT*
SEWAGE TREATMENT PLANTS*
EXCESS FLOW PLANT*
URBAN RUNOFF PLANT*
WASTE AND WATER
WATER AND WASTE
AUTOMATIC CONTROL* ^
AUTOMAT* A
FEED-FORWARD*
INSTRUMENTATION
ANALOG PROCESS CONTROL*
DIGITAL PROCESS CONTROL*
INSTRUMENTED
COMPUTERIZED PROCESS CONTROL*
732
943
944
DETECTION AND CONTR01 *
COMPUTERIZED CONTROL*
COMPUTER CONTROL*
COMPUTER AIDED CONTROL*
SFWAGF TREATMENT PL ANTS. MODELS*
SEWAGE TREATMENT PLANTS, SI MULATORS*
SEWAGE TREATMENT PLANT MODEL*
SEWAGE TREATMENT PLANT SIMULATOR*
MINE
MINES
MINING
NUMBER OF ANSWERS 3
-------
PAGE 1
DOCUMENTS FROM THE LEADER RETRIEVAL SYSTEM
07/1B/72
PAGE 1 OF 2! DOC 1 OF 301 LEADER DOC 72575.
TITLE CASE FOR CONTINUED STRONG SUPPORT OF BASIC RESEARCH BY THE
FEDERAL GOVERNMENT
AUTH. ANON
JOUR. RES MANAGE V 13 N 4 P 273-9
DATE JULY 1970
SRCE. CMPDX7104 EI19544
PAGE 1 OF 21 DOC 2 OF 301 LEADER DOC 61679.
TITLE CASE FOR CONTINUED STRONG SUPPORT OF BASIC RESEARCH BY THE
FEDERAL GOVERNMENT
AUTH. ANON
JOUR. RES MANAGE V 13 N 4 P 273-9
DATE JULY 1970
SRCE. CMPOX7104 EI19544
PAGE 1 OF 21 DOC 3 OF 30! LEADER DOC 100541.
TITLE INDUSTRY-GOVERNMENT COOPERATIVE RESEARCH FOR WATER POLLUTION
CONTROL
AUTH. HARLIN JR CC
JOUR. AICHE WORKSHOP. IND PROCESS DES WATER POLLUT CONTR V 2. HOUSTON.
TEX P 62-4
DATE APR 24-25 1969
SRCE. CMPDX7108 EI50855
PAGE 1 OF 2* DOC 4 OF 30! LEADER DOC 71966.
TITLE HOW RESEARCH IS HELPING TO SOLVE ASIAvS FOOD PROBLEMS
AUTH. MYERS WI
JOUR. RES MANAGE V 13 N 3 P 229-38
DATE MAY 1970
SRCE. CMPDX7104 EI19008
PAGE 1 OF 2! DOC 5 OF 30! LEADER DOC 61070.
TITLE HOW RESEARCH IS HELPING TO SOLVE ASIAvS FOOD PROBLEMS
AUTH. MYERS WI
JOUR. RES MANAGE V 13 N 3 P 229-38
DATE MAY 1970
SRCE. CMPDX7104 EI19008
C-3
-------
*
*
G01NERC0003
W4 8ENOIT
DISPLAY PROFILE G01NERC0003
OIR AOMIN NAT ENV RCH CTR
0».7J50*******************************
•REPORT*
•PROGRESS
•REPORT*
•SCIEN*
•TECHNI*
LABORATORY
•PROGRESS
•REPORT*
•FACIUT*
•PERSON*
•RESEARCH*
•SCIEN*
CPNTRACUQN.
LABORATORY
PHYSICAL
0.750 2 •IN-HOUSE
0,000 | *NEUR*
6.666 1 •REACT*
0.750 2 •REPORT*
0.750 2 •SYSTEM*
o.ooo i AIR*
0.750 2 COMMUNICATIO
i 0.75JD 2 COMMUNICATIO
'0.750 2 COMPUTER
0.000 1 DEFENSE
0.660 1 ELECTRONIC
0.750 2 FEEDBACK
6.666 1 GENETIC*
0*750 I INFORMATION
0.750 2 INTERIOR
0.750 2 MANAGEMENT
0.750 2 MANAGEMENT
0.750 2 MANAGIHEM
0.750 2 MANAGEMENT
0.750 2 MANAGEMENT
6.750 2 MANAGEMENT
0.750 9 MANAGEMENT
6.666 1 SATELLITE
0,750 2 STORAGE
0.000 0
PLANT
RETRIEVAL
36 SINGLE WORO EQUIVALENTS ARE IN PROFILE
MECHANIZED INFORMATION CENTER
THE OHIO STATE UNIVERSITY LIBRARIES
C-4
-------
FINN KRt MLLER H6
IS YCUK MIS FIT FOR HUMAN CONSUMPTION?
« INFCRMATION).
;WJ 6ENOIT
lOlk ADMIN
!NAT LNV KCH CTR
INDUSTRIAL ENGINEERING
VOL. 3 1*71 NU. 11
TERMS: INFORMATION
WJ BENCH
INFORM AT!
SERVICES
PAGE 18
IFINN KR, MILLER
IIS YOUR MIS FIT; !
JLIEA
iISSUfc il 71
NERCOOQ32062C25 (VOL.3 P.18
3-17-72
CURRtNT AWAKtNESS
IRVlCfc OF OSU LIBRARIES
NERC00032062025
DETACH
HERE
HARVEY P
LIBRARY FOR UNIVERSITY OF riANCl (INFORMATION
SYSTEMS). :
iwj BENCIT
;D1R ADMIN
INAT ENV RCH CTR
BRITISH MEDICAL JOURNAL
VOL. 1 197<: NO. 5
PAGE
I
,'HARVtY P
iLIBRARY FOK UHI
SHEA, CHI, VET
JBMJOA
I ISSUE 5791 72
NERC00032062027 iVCL.l P.54
3-17-72 !
JNERC00032062027
S N FO RM ATI D N CURRENT AWARENESS i ' DETACH
SERVICES A SERVICE OF OSU LIBRARIES !**
TERMS: INFORMATION
WJ JicNCiT
TISHLtR W . . WJ BtfiMOIT
IDIR ADMIN
WHAT IISFCRMA1ICN UUES MtuICiNAL CHEMIST REALLY INAT tNV RCH
IMEEC? PROJECTIONS FOR FUTURE. : -
JOURNAL OF CHEMICAL DOCUMENTATION
VOL. 11 IS?! NO. 3
TERMS: INFORMATION
WJ BcNCIT
|TISHL6R M
PAGE 134 iWHAT INFORMATIO
iCHE, PHA, ENR
iJCHQA
iISSUE 3 71
NEKC00032C62C29 JVCL.11 P. 134
3-17-72
'NeRC00032062029
t h
S
CUKRENT AWARENESS
tJ i^. ZZ
SAMPLE: BI-WEEKLY PROFILE OUTPUT
C-5
-------
APPENDIX D
The following staff members at NERC-Cincinnati and their
responsibilities can be used as a guide to the appropriate
person to contact for help or advice:
MORTON H. FRIEDMAN General services and new systems
DOTTIE F. BROFFT General reference
LILIAN S. BOSWORTH Record keeping, billing and
distribution
MARY LEE BURBAGE Central cataloging services
MARY L. CALKINS On-line, biomedical and profile
analysis
MAXINE M. SMITH Batch, current awareness and
' profile analysis
Do not hesitate to contact any staff member whenever necessary.
This system will work only so long as the librarians and
researchers are willing to participate and make it work.
US. GOVERNMENT PRINTINGOFFICE 1972- 759-546/1013 D —
------- |