when referringto this reprint, cite:

Gwfiald, E. Historiographs, librarianship, and the history of science. In Toward a theory oflibrarianship: papers in honor of .&se Hauk Shera, ad. by Conrad H. Rawski (%letuchen,N. J.: Scarecrow Press, 1973), p. 380402.

Historiographs, Librarianship

and the

History of Science

by

Eugene GarfieldInstif ufe j_orScien ti~ic Irs~orma tion

325 Chestnut StreetPhiladelphia, Pa 19106

The study of the history of sciencehas recently taken on new importanceas an academic discipline. It’s easy tosee why. With the pace and complexityof scientific developments accelerating,the need for improved guidelines forscientific research becomes more ob-vious each day. As critical as today’sdilemmas of pesticides, pollution, andnuclear weapons are, applications ofcurrent scientific research could pro-duce far more serious problems in thefuture. For example, the work of mo-lecular biologists could produce a hap-pier, healthier world population or, assuggested by some, it could producesome horrible new weapon of war. 1

Guidelines for scientific research,however, do not simply materializefrom thin air. A solid understanding ofpast developments and the present na-ture of science itself is required fust.Professor Derek de Solla Price claim%“we are getting to the point wherethere must arise a fairly hard academicdiscipline to help understand the ma-chinery that makes science act the wayit does and grow the way it grows.’ ‘z

By studying the history of sciencein a more intensive and accurate fash-ion, we may obtain badly needed in-

sight into such problems as:a The role of science in war and peace.b. The use and misuse of research.c. The inter-relations of science and

technology.d The reciprocal responsibilities of

scientists and society.e, The funding and control of science.f. The determination of future policies

on scientific education.g. The formulation of a public policy

on science in general.With such important questions to be

answered, writing the history of sciencecan no longer be looked on as an ex-ercise to satisfy one’s curiosity. Thisendeavor is too important to be ful-filled, as it has been in the past, as anavocation of scientists. What is neede~as in any other complex activity, is ahighly trained specialist.

There have been some beginningstowards this objective. hr I!)50 therewere only a handful of professionalscience historians in North America,

136

and few sch 00Is offered doctoral oregrams. Today we can count at Last500 scholars in the field, while atleast 25 maior universities offer de-gree program; in the history of science.

Probably the f~st full-time historianof science “was George Sarton of Har-vard who founded h-is, the chief jour-nal of the field. In his early work,Sarton was primarily concerned withprecise chronological reconstructionsof events. In his subsequent work, heused more of a narrativ~ approach andbegan to analyze and interpret causeand effect relationships. Later, Alex-andre Koyr~ of Print-eton attemptedto explain the development of a newscientific concept by examining thework of the scientist against the pre-vailing philosophical and intellectualassumptions of his time. Koyr6, how-ever, did not regard a scie ntist’s outersocial milieu as an important factor inshaping his work. More recently, his-torians of science have begun to stressthe relationship between new ideas andthe outer social order in which theydevelop. a

Whatever the approach, the historianof science will produce useful results.only in direct proportion to the investi-gative and evaluative tools at his dis-posal. The difficulty of amassing thefacts of history is well known. Muchhuman error is injected on the part ofthe historian despite his dedication andrigorous standards. Even with an eventlike the assassination of PresidentKennedy, which was observed by count-less uersons, there still remains doubt,as to precisely what occurred. Writingthe history of science has its own par-ticular difficulties. The motivation andevolution of ideas are frequently omit-ted from scientific writings. Usually,major achievements in Wience are easily

recognized; minor or less heralded con-tributions are difficult to identify andare often overlooked Even rela~ivelyirrmortant events rnav be missed in the, J

plethora of data to be evaluated. It is.not surprising, therefore, that thereare alwavs numerous uncertaintiess in/writing even a fragment of the historyof science.Historiographs

A ne-w tool that tmomises to beltsthe historian of scien~e out of this pr;dicament is the “historiograph,” aterm coined by the Institute-for Scien-tifm Information to describe a graphicdisplay of citation &ta that shows keyscient~lc events, their chronology, theirinter-relationships, and their relativeimportance. Although the technicalfeasibility of this tool has been estab-lished, a~d a good portion of the re-quired citation &ta base is available,much work remains to acquaint thehistorian of science with its availability/and its applications. I feel that this is arole that is most properly filled by thelibrarian. I also feel that if librarianstake advantage of this opportunity,they will be actively participating in afield of growing social significance andwill have taken one more step towardsachieving the dynamic professionalimage they desire.

As early as 1922 E. WyndhamHulme used the term “statistical bib-liography” in his lectures at the Unkver~ity of Cambridge. Hsdme used theterm to describe the process of il-luminating the history of science andtechnology by counting documents. Inlater year% Pritchard used the word“bibliometrics” to describe the quanti-tative asxdvsis of citations. Russianhistorians o; science have su~ested theuse of the term “scientom~trics” forthis type of study,h

137

Key:Rabinowitch. 1941 6. Lawley. 1956. 11. Steiner.1959.

;: Michaelis.1947. 7. Peacocke. 1956. Bradley. 1959.3. Michaelis. 1950. 8, Appel. 1958. :;: Bradley. 1959.4. Zanker. 1952. 9. Appel. 1958. 14.5.

Bradley, 1960.Northland. 1954. 10. Steiner. 1958. 15. Loeser. 1960.

Figure 1. Citation Relationships of a

It was almost by accident, however,that Dr. Gordon Allen triggered theactivities that led to the developmentof the hiatoriograph. In a privatecommunication to me in 1960, Dr.Allen diigrammed the relationships be-tween the citations of a bibliographyon the staining of nucleic acid asshown in Figure 1. Although Dr. Allendid not chink of this diagram in thecontext of a historical tool, my exam-ination of it lead me to form the hy-pothesis that, in most case% a networkdiagram of citation relationships wouldin fact, constitute a fairly reliable “out-

Bibliography on Staining Nucleic Acid.

line” for writing the history of a fieldof science. This belief was further re-inforced in @cussions with Berna~SPrice,G Leake,T and Shryock. 8 I thenpublished an article in A mericms ~s+merstutim9 in 1963 which summarizedmy thoughts on the subject and pro-posed some specWlc applications.Testing the validity of Historiographs

In 1964, the Institute for ScientifKInformation began work on Air ForceOffice of Scientifw Research ContractAF49(638)-1256 to verify whether ci-tation data are useful heuristic tools forthe historian.1 o Essentially, the plan

138

of the study was to construct two net-work diagrams of the history of a fieldof scienc~: one based on the “traditionalhistorical account, and one based on areconstruction of the same history byusing citation relationships. If the’ twodiagrams coincided to any signflcantdegree, it could be concluded that ci-tation data was, in fact, useful in writ-ing the history of science,

To conduct the study it was neces-sary to select a recent important scien-tific break-through which was based onthe cumulation of years of diversescientflc achievement. The dkcoveryof the DNA code was selected as thisevent. The basis for this choice was:(1) the publication in 1963 of Dr. IsaacAsimov’s book. The Genetic Co&.11which describes the major scientific&velopments that led to the laboratoryduplic~tion of the process of proteinsynthesis under control of DNA, and(2) the availability of the Genetics Cita-tion kdex and the 1961 Science Cita-

tion Index e to provide the requiredcitation data.

In our study, Dr. Asimov’s book wasconstrued as the historian’s account ofthe discovery of the DNA code. TheCXX and the SC1o provided most ofthe data by which we would attemptto construct a history of the same top;cby other than expository accounts.

To construct the two historiographs,

w Proceeded as follows: fist we care-fully identified the specii3c papers in-volved in the discoveries described byAsimov in his history of DNA. Theseincluded events explicitly named byAsirnov, as weu as events not explicitlynamed but easily identtiled by his men-tion of such things as date or place ofinvestigation. Forty key events (callednodes) were identified (36 explicit, 4implied) which spanned a period of

about 140 years (1820 to 1962). The40 nodes were then plotted chronol~gically and grouped in broad subjectclassifications such as nucleic acid chem-istry, protein chemistry, genetics, andmicrobiology. Asimov’s book was thenexamined to determine the historicalrelationships between these 40 nodes.The explicit and implicit relationshipswere then diagramed as shown in Fig-ure 2.

An extensive literature search wasthen conducted to identify the specifiipublished works related to each nodedescribed by Asimov. The strictestcriteria were adopted to insure that thereference citations chosen were theones which most definitely correspond-ed to the discovery in question. Itturned out that 17 of the 40 nodes re-presented more than one paper. Thus65 articles were required to cover the40 nodes.

The bibliography of each node ar-ticle was examined to determine theconnections between it and other nodearticles. The 40 nodes were then re-&awn in exactly the same positionsand lines drawn to show the d~ect andindirect citation connections as shownin Figure 3. Finally, the two diagramswere superimposed and the degree ofcoincidence determined. Figure 4 is asummary of the relationships observedfrom the superimposition of the twodugrams.

From this comparison it was de-termined that: (a) enough of the net-work diagrams were coincident so thatit could be concluded that citationdata could be used to develop the his-tory of a field; and (b) significant newrelationships between nodes were iden-tified by citation linkages which didnot coincide with Asimov’s linkages.

139

mR%!s Smctf,ed C“nnec(, c,”

1111181 Impl,ed Conoec I,cm

Figure 2. Asimov’s Specified and Implied Relationships in the History of DNA

140

Key for Figures 2 and 3

i

2::7.8.

1::11.12.13.14.15.16.

:::19.20.

Braconnot. 1820.Mendel. 1865,Miescher, 1871.Flemming .1879.Kossel. 1886.Fischer, Piloty. 1891.De Vries. 1900.Fisher. 1907.Levene, Jacobs. 1909.Muller. 1926.Griffith. 1928.Levene,Mori, London. 1929.Alloway. 1932.Stanley. 1935.Levene, Tipson. 1935.Bawden, Pirie. 1936/37.Caapersson, Schultz. 1938/39.Beadle, Tatum. 1941.Martin, Synge. 1943/44.Avery, MacLeo~ McCarty. 1944.

Over and above these findings, itbecame clear that one picture was in-deed worth a thousand worda. The gra-phic displays of the history made iteasier and quicker to grasp the totalflow of the developmentof the field.Further, they made it possible to tie-in seemingly unrelated events. It was atthis point that we became convincedthat the historiograph would be a boonto the historian.Need for Automatic Diagraming

It soon became obvious, however,that the manual production of historio-graphs would severely limit the usefid-ness of this tool. As the number ofnodes increased in a history, it soonbecame almost physically impossible todraw all the connecting lines, Also, asthe lines increased, the clarity of thediagram &creased.

In 1967 1S1o began an investiga-tion of the problem of the automaticdrawing of network diagrams contain-ing nodes with many interconnections.This work was performed under AirForce Office of Scientific ResearchContract AF49(638)-1 547 and its ob-jective was to identify an existing

21.

;$24.25.26.27.28.29.30.

:;:33.34.

:::37.38.39.40.

Chargaff. 1947.Chargaff. 19s0.Paulirrg, Corey. 1950/51.Sanger. 1951-53.Hershey, Chase. 1952.Wilkins. 1953.Watson. Crick. 1953.Du Vigneaud. 1953.Todd. 1955.Palade. 1954-56.Fraenkel, Conrat. 1955-57.Ochoa. 1955/56.Kornberg. 1956/57.Ho land. 1957/58.

7Jaco , Monod, 1960/61.Hurwitz. 1960.Dintzis. 1961.NovelIi. 1961/62,Allfrey, Mirsky. 1962.Nirenberg, Matthae], 1961/62.

method of automatic diagraming or

develop a new one that would:~ AUOW algorithmic generation and

manipulation of the diagram.b. Allow the display of 100 or more

nodes.c. AUOWthe display of an unrestricted

number of connections betweennodes.

d. Be simple to understand and aes-thetically pleasing.

e. Allow direct printout from a digitalcomputer or plotter.AU known methods of automatic

diagraming were investigated. Theseincluded methods used for:1. Flow charts and PERT chartsl Z-262. Organization chartsz 73. Electronic circuitsz 8-38

4. Tree systemsag~’$o5. Routing systemaal -4’$

The investigation of previous workuncovered no method that satisfied allspecified criteria. Even if all other cr-iteria were met, the dkplays becameinordinately complicated when morethan 10 nodes were involved. It wasthen felt that the solution to this pr~blem was in developing a unique way

141

I8

1

‘n10

SSSSSSSDirect or St fonglv Implmd Cnatoon

wsmtaInd,rect C8t.31ton

Figure 3. Direct and Indirect Citation Relationships in the History of DNA

142

of ordering nodes and/or interconnect-ing lines. Some of the display formatsevaluated and rejected included:a. Linear Array (see Figure 5)b. Cascade Display (see Figure 6)c. Waterfall Display (see Figure 7)d. Rectangular-Node Cascade Display

(see Figure 8)e. Fountain D~play (see Figure 9)

Diagonal DisplayFinally, the concept of “Diagonal

Display” was evolved and refined.aS,aGAs shown in Figure 10, Diagonal Dis-play involves arranging the no&s on a

diagonal with interconnections shownby lines in the areas adjacent to thenodes. In this original conceptual draw-ing, the dots indicate a connection be-tween nodes.

Since the Diagonal Display concepthad been entirely handdrawn up tothis poinL it was now necessary todetermine how this type of plot couldbe drawn automatically. The fwst at-temrws involved the use of a standardcom~uter printer. Although this wasquite satisfactory for many applica-tions, it was felt that even the additionof special characters to the printervmuld not give the printout the over-all visual claritv desired.

With partic~lar concern for improv-ing the clarity of line junctions andcrossover% we then considered severalother types of output devices. Includedin these were:a. Electric lamp displayb. Cathode ray tube (CRT) direct dis-

playc. CRT plotterd. Incremental pen plotter

It soon became obvious that al-though each device could be used inthe system, the best choice for thetype of printout desired would be an

incremental pen plotter. Typically, withsuch a device, digital signals cause thepen to move in increments in lineardirections to any point on a sheet ofpaper. By drawing very smal~ consecu-tive increments, the pen can also pro-duce curved lines.

A number of commercially availableincremental pen plotters were con-sidered, and a CalComp 563 was selec-ted as the most suitable. An appropri-ate program was written and severalvariations of the basic Diagonal Di.+play format shown in Figure 10 were&veloped, with the most significantmod$lcations being the way connec-tions and crossovers of lines werehandled. Most improvements in theselected display format involved theuse of gaps in vertical lines to irdcatecrossovers, rounded elbows to showconnections, and the addition of num-bers at each elbow to show what nodeswere connected at that point.

A printout was generated for the 40nodes contained in the previously dis-cussed history of the genetic code.This printout is shown in Figure 11. Acomparison of Figure 11 with Figure 3(the hand-drawn network of the DNAhistory) will make obvious the in-creased clarity provided by the automatic Diagonal Display.

Another test was conducted to pr-duce a printout for a network with 80nodes. It was concluded that this wasas visually clear as the 40-node print-out. In fact, it is now felt that evenprintouts containing 250 nodes wouldbe equally clear.

To summarize all this, ISI’s workhas shown:1. That citation relationships can be

of great use in writing the history ofa field of science.

143

1960

1955

1950

1940

1930

1900

1820

r/1 ).......(/ ....

...../ ‘...,.

/&v “’~<

— Directcitation connections ————u...Asimov’sspecified historicalconnections——— _ Indirectcitation connections . . . . . . .Asimov’s impliedhistoricalconnections

DATES, NAMES & NODES1935 Stanley-f4 1953 Watson, Crick-27

1820 Braconnot.1 1935 Levene, Tipson-15 1953 DuVigneaud-281865 Mendel-2 1936-37 Bawden, Pirie.161871

1955Miescher-3

Todd-291938-39 Caspersson, Schultz-17 1954-56 Palade-30

1879 Flemmirrg-4 1941 Beadle, Tatum-18 1955.57 Fraenkel<onrat-311886 KOSSC1-5 1943-44 Martin, Synge.19 1955-56 Ochoa-321891 Fischer, Piloty.61900

1944 Avery,Mac Leod,DeVrie%7

1956-57 Komberg.33McCart -20

1907 /’1957.58 1-lo land-34

Fischer-8 19471909

Charga f.21 T1960-61 JacO , Mo”od.35Levene, Jacobs-9

19261950 Chargaff-22 1960

MuUcr-10Hurwitz.36

19281950-51 Pauling, COrey-23 1961

Griffith-1 1Dintz is-37

1951.53 Sanger.241929

1961-62 NoveUi.38Laene, Mori, London-12 1952 Hershey, Chase-25

19321962 Allfrey, Mirsky-39

Alloway.13 1953 Wilkin~26 1961.62 Nirenberg, Matthaei.40

Figure 4. A summary of the relationships observed from superimposition of Fig-ures 2 and 3

144

.<

(,

Figure 5. Linear Display

Figure 6. Cascade Disphiy

Figure 7. Waterfd Display

Figure 9. Fountain Display

1, 1I1 I

B89

to

Figure 8. Rectangular-Node

Cascade Display

Figure 10. Original Diagonal

Display Concept

2. That a graphic representation (hbtoriograph) helps to clarify the re-lationships and give a quick, over-allview of the development of the field.

3. That relatively large-scale histori~graphs can be automatically pro-duced on a practical basis.

ApplicationsManuafly or automatically drawn

historiogtaphs can be used to advan-tage by the historian of science to:a. Reduce a large number of seemingly

unrelated events to a coherent pat-tern.

b. Identify classic papers.c. Identify break-through events with-

in the development of a field.d. Identify the descendant of a de-

velopment.e. Keep track of how often an event

influences later events.

Con textsBy removing the vagaries of manual

drawing, however, ausomatically pro-duced historiographs provide addition-al depth of analysis. For example, be-cause similar relationships will be con-sistently displayed in the same way, cer-tain types of pattern analyses can beperformed. The similarities or differen-ces between the historiograph patternsof one field and those of another couldlead to meaningful insights. This typeof comparison could also apply to his-toriographs generated for sub-fields ofa larger area of scientific endeavor. Pat-tern analysis could also be used to iden-tify unusual chronological spacing be-tween events and to determine howfields of science coalesce or fragment.Automatic manipulation also makes iteasy to emphasize or leave out certainevents in a given history. The historiancan then determine if there were alter-nate paths to subsequent discoveries

Role of the LibrarianAt this point, the librarian may legiti-

mately ask, “What has all this to do withme? of what use will historiographs beto me in my work? Where do I fit in?”Librarians should look on historiographs as tools that rightfully belong intheir domain. Just as much as an indexis a retrieval tool, so too is the histotio-graph. For example, what better waycould there be for the librarian to Se]ectkey papers for a newcomer to a field toread than by examining a pertinenthistoriograph? In the not too distantfuture I can envision a historian ofscience requesting information fromthe library on a specific development.The librarian will then, as he does now,compile an appropriate bibliographyby manual or automatic means. How-

~ever, another step will then take placethat will add a new level of importanceto the assistance a researcher can ex-pect from a librarian. The librarian willthen sit at a computer console (orsome similar device ) and use the bib-liography as input data. He will ask thecomputer to print out or display all orpart of the citation relationships exist-ing in the input bibliography. The li-

Ibrarian will then assist the researcher inidentifying key papers or obscure but

~important papers. The value of thistype of service, especially with lengthybibliographies, is self-evident.

Another important contribution thatcould be made by a librarian with ahistoriograph to analyze, would be theide ntiflcation of a key paper from an-other field that had an impact on thefield of interest. With experience, thelibrarian might also become skillful atrecognizing certain “classic” h“~torio-graph patterns. perhaps then, he will beable to help the historian identify the

1signs of declining or emerging disci-plines.

146

0! Bx.c.””et,,=02

j_l._L-=N IW.L 1SSS

03 “,--- l-,

LOIai“‘:-””’”=.0ss..,Sm06,18c”r.a?,L.O,,M*5~m.mC4 IIr:m’-“b!kd-_G;’”’~~,,;:r.JAu”J-H-*““l..,.l“’‘“II

‘“,,,,.” ,“,

09 t’

,,&E:c..o”j LO”lm.,m.

L-

AL1..lv. ” 1*3Y .

14m...,” ,*M

,, ,,

,,&*;:___

II.Aww..,,.,,,,s,,=

c*srr..s50l4SC””I.TZt.lO.i.-,0,.“,* II 01’8=&”’Ar”M’””’

J*,* ! [ , d-”~-p“.”,,?4 SW”c. !“*]*..

@@o o@

.“.. ” ❑ AC ,.0.. “c . ..., ,..4

~

1,~,11 @@

,,, [ :Ej?q!+$,,:u ““:

(Lines A to E, above, aretinued ad indicated by A Ithe b?ft. )

.*FW.,,o. .“,.. ,.s,

,, m I F~;’””””

,, w I I l’””O;”ss~ I I I 0J’A:y4”’*-

LILI ~,,FRA,”.*1,.CO”,A,,ss>,,,7Z7aamn

Ik

OC”C.*!*ab.,,-‘U33

.0,.,,.6 ,Ym.,ti,“u

l%. . . . . . ,,,7.,,-L“n

u

al

—

con.to E on

— 40N,,ZMW*G “Am”.,, ,“,.,eml

Figure 11. Printout of Computer Generated Diagonal Display

147

The types of analyses described abovecan be performed by using histori~graphs generated from citation datacurrently contained in the Sciertce Ci-tation Index. [n the near future, 1S1will publish, for the first time, the Jour-nal Citation Index. a 7 The JCI will be astatistical compilation that shows howoften other journals cite each of over2000 journals, The JCZ will also showhow often these 2000 journals cite anyof over 25,000 other journals.With thistype of data base available, it will bepossible to draw historiographs thatshow relationships between journalsrather than individual papers. l%is, inturn, will permit several additionaltypes of analyses to be made.

For example, it would be possibleto show the emergence and growth ofa field of science by tracing the historyof iournals in that field, To do this, a.number of journals could be used asnodes in a network. The nodes wouldbe arranged chronologically accordingto when volume 1 of each particularjournal appeared. The citation patternsbetween the journals over the yearscould then be drawn and the relation-ships observed. With this kind of dis-play it would be possible to trace, interms of journal development, how sci-entific fields branched out from olderexisting fields. You could literally tellwhich journals were the “parents” ofother journals. Besides being an aid inwriting the histo~ of science, thiswould also be a useful tool for libraryscience researchers.

Another use for network diagramsmade from ]Ournal Citati”on zna%x datawould be to determine what disciplinesmake up a scientific field. For example,you could start with certain hard-coregenetics journals (such as The A mericun

Journal of Human Genetics, Annals ofHuman Genetics, and Genetics) andhave them represented as a kind ofbulls-eye node in the center of a net-work diagram. Then, you could repre-sent other journals which cite thesejournals or are cited by these journalsas other nodes in the diagram. Thosejournals that had the highest number ofinter-citations with the core journals

% the dia-would be positioned closestgram bulls-eye. Those with a lessernumber of inter-citations would beplaced proportionately further away.Thus, with one diagram, you could geta quick, clear picture of all the disci-plines involved in genetics and the ex-tent of their involvement. Not onlywould this be a help to the historian ofscience, it would help the librarian insuch things as advising the head of ascience department on what journalshe should receive. There is no need toemphasize the beneficial effect theability to give such service will have onthe professional status of librarians asa group.A Project for Students

In conclusion, I would like to pro-pose a project that would be quite ap-propriate for library school students.AI1the preceding discussion of historio-graphs and other types of network d~a-grams presupposes the existence oflarge-scale citation indexes to providethe data necessary to draw the histono-graphs. 1S1 has produced such indexesfor all but two years of the 1960’s andwork is now in progress so that cita-tion indexes for the missing years (1962and 1963 ) will be available by the endof 1971. Our goal is to produce total ci-tation index coverage for the literatureof the 20th century, and steps are beingtaken in this direction.

148

1S1 would like to establish a similargoal for the literature of the 19th cen-tury. In particular, we would like tostart this project by creating citation in-dexes for the rather limited Americanscientific literature of that period. Un-fortunately, 19th-century scientific au-thors had very poor citation prac-tices-very few explicit citations weremade, although implicit citations werecommon. Thus. before 1S1’s citation

rndexing techniques could be applied,explicit citations would have to becreated from the implicit onea. It is thework of creating these explicit cita-tions that I suggest be taken on by li-brary students.

The value of this information tohistorians of science, and perhaps tocivilization, would begin to approachthe magnitude of the respect we allhave for lesse Shera.att

NOTES AND REFERENCES1. Lederberg, J. Science and man: es-

sence of life or death? CurrentContents No. 15, 15 April 1970, p. Ml. Re-printedfrom The WzdringtonPost 28 March1970.

2. Price, D.J.D. Measuring the size ofscience. Proc. Israel Acssd. Sci. Hu-

manities 4:98-114, 1969.3. llackray, A. & Merton, R.K. On dis-

cipline building and the paradoxesofGeorgeSarton.Isis 63:473-95, 1972.

4. Pritchard,A. Statistical bibliographyor bibliometrics. J. Docrsrnenkstio?s

25:348-49, 1969.

5. Bernal, J.D. Private communication,March 1962.

6. Price, D.J.D. Private communication,March 1962.

7. L.cake, C.D. Private communication,August 1962.

8. Shryoclq R.H. Private communica-tion, September1962.

9. Garfield, E. Citation indexes in socimlogical and historical research. Amer.

Documentation 14:289-91, 1963.

10. Garfield, E., Sher, 1.H. & Torpie, RJ.The use of citrstirm data in. writing tJre

history o~sciertce. Philadelphia: Institute forScientific Information, 1964, 76pp.

11. Asirnov, L The genetic co&. NewYork: New American Library, 1962,

187pp.

12. Butler, D.D. & Gatto, O.T. Event-chcain flow charting in au tosate: A

rsew version. The Rand Corporation, RM-472a-PR (DDC No. AD 622744), October1965, 69pp.

13. IBM 7070/7074 Autochart program-mirrg system. IBM Systems Reference

Library File NA. 7070/7074-48, February I1964, 58pp.

14. 1401/1410 Autocoder program logicdirsgrammer. IBM Program Applica-

tion Library File No. 1401-01.4.128, 1964.

15. Stelwagon, W.B. Principks and pro-cedures for the automatic j20 w clrart-

ing program FLOW 2. U.S. Naval OrdnanceTest Station, Research Department, ChinaLake, California, NOTST P 4095 (DM No.AD 637863), August 1966, 63pp.

16. System1360 FLOWCHA R T. IBM Sys-tems Reference Library File No. H20-

0199-0, 1950, 20 pp.17. Haibt,L.M. A program to draw multi-

level flow charts. proc. West. JointComputer Conf. 1959:131-37.

18. Rain, C. & Hain, K. An assto-diagram-mer. (NASA TMX 1317) Nov. 1966.

19. Knuth, D.E. Computer-drawn flowcharts. Comm. ACM 6:555-63, 1963.

20. Scott, A.E. Automatic preparation offlow chart listings. J. ACM 5:57-66,

1957.

21. Sherman, P.M. FLOWTRACE, a com-puter program for flowcharting pro-

grams. Comnt. ACM 9:845-54, 1966.

22. Anderson, H.E. Automated plottingof flow-charts on a small computer.

Comm. ACM 8:38-39, 1965.

23. Daly, J.F. PERT, Core$fructr”cnrpro-ject network analysis. U.S. Naval

School, Civil Engineer Corps, Port Heuneme,California (DDC No. AD 612174), July1964,200 pp.

24. Parikh, S.C. & Jewell, W.S. Decom-position of project networks. Marrage-

rnerrt Science 11:444-59, 1965.

25. Solomon, N.B. Automated methodsin PERT processing. Comprs ters and

Automation 1965:18-55.

149

26. Asselin, T.H. CPM bar chart: A gra- 38. Aarons, M.W. & Goldbe~ M.J. Com-phic CPMschedufe generated by com- puter methods for integrated circuit

puter. Massachusetts Institute of Technology, design. Electro-Technology 5:77-81, 1965.School of Engineering, Research Report 39 Paul. A.J. Jr. Generation of directedR66-4A, FebruaW 1966, 87pp. trees, 2.trees and paths without dup27. Kaitz, M.J. Acorn-computer generated Iicaticm, University of Illinois, Coordinated

organization ch arts. UAZDE Proc., Science Laboratory, Urbana, Report R-241,Sec III 9, North American Aviation, Space January 1965, 45 pp.and Information Division, Douney, Califor- 4~ Benes, V.E. Programming and con-nia, 1964. trol problems arising from optional28. West, L.E. & Caskey, D.L. Topogra- routing in telephone networks. Bell Systerrr

phic simulation us an aid to printed Tech. J. 45,13731438, 1966,circuit board design. Sandix Corporation, 41Albuquerque (Research Report SC-RR-66- “

CalComp plotter unsnarls traffic prm

424), August 1966, 33pp.blems. Digital Plotting Newsletter

(CalComp, Anaheim, California) October/29. Case, P.W. et al. Solid logic design November 1964,

automation. IBM Journal 8:127-40, 421964.

Traffic flow maps plotted mechanical-

30.ly. Public Works 96:117, 1965.

Ceshner, R.A. & Rifkin, A.L. Auto- 43Pipe layouts by computer. Mechanical

mating the production of printed cir- “Engineering. 86:57, 1964.

cuit artwork. Computer Desi@r 5:54-58,1966. 44. Rhoades, E.L. & Milkr, J.R. Com-

31. Grigsby, L.L. & BlackwelI, W.A. De-puter f~ures pipe lengths. PiaFst En-

s~n of a linear mrdtiport network bygineerirtg 19:150-51, 1965.

state-space techniques. Matrix & Tensor 45. Garfield, E. & Sher, LH. Diagonal

Quarterly 16:122-34, 1966.display --a new technique for graphic

32. Kats, S. Methods of synthesis ofrepresen tats”on of complex topological ne t-

Iinear networks in the time domain.works, Philadelphia: Institute for Scientific

Bulr! A cad. Polo.. Sci. 14:107-11, 1966.Information, 1967, 94 pp. (Final Report toU.S. Air Force, Office of Scientific Re-

33. Waxman, R. et al. Automated logic search, Contract AF 49 (638)-1547.design techniques applicable to inte- 46, Garfield, E. & Malin, M.V. Diagonal

grated circuitry technology. Fall Joint Com- display --a new technique for graphicprs ter Conference Proceedings (AmericanFederation of Information Processing So-

representation of network diagrams. Trrrns-actionr of the A nten”can Association of Cost

cieties) 1966, pp. 247-65.

34.Engineers, 13th National Meeting, Pittsburgh,

Burstall, R.M. Computer design of 1969, pp. 222-32.electricity supply networks by a heur-

istic method. Computer Journal 9:263-74, 47” Garfield, E. 1S1’s Journal Citration

1966. lrsrr%xdata base--a multi-media tool.

35. Friedman, I. et al. NAP--networkCurrerstCorztents~, NO.16, 19 April 1972, p.

automatic plotting. Atomics lnterna-5-8. -- Since the article reprinted here was

tional, Los Angeles (AI-64-MEMO-15), Feb-published, the JCI bas become one of 1S1’s

ruary 1964, 118 pp. large group of information products and ser-

36.vices, under the title ISI JoumaJ Citation

Dodd, G.G. Principles and Applica- @orts.tions of Unistor Graphs. University 48 For readers who may not know him,

of Illinois, Coordinated Science Laboratory, “Urbana, Report R-217 (DDC No. 604095),

Jesse Hauk Shera, for many years Dean

July 1964, 47 pp.of the School of Library Science at Case West-

37.ern Reserve University, has been not so much

Sheppard, D.M. et a~ A computer a legend in his own time, as a phenomenon inaided method for checking and mak- the science and at of librarianship. The title

ing monolithic integrated circuit masks.Western Joint Computer Conference PrO-

of the Festschrr”ft in which the paper re-printed here appeared reflects the spirit and

czedings (American Federation of In forma- striving of his prodigiously productive pr~tion Processing Societies), 1966, PP. 1-7. fessional life: toward a theory of librarian-

ship.

150