Requirements of Forest Scientists for Literature and Reference Services
Within the professional field of forestry, there exists a considerable body of scientists employed primarily as research workers with governmental experiment stations and large industrial concerns, and as University teachers. These number approximately 1500 in the United States alone and perhaps as many more around the world.
The duties of these scientists are oriented toward the multiple use of forests for wood production, water management, game management, range management, and human recreation. The shortage of actual data on the growth, development, and behavior of both forest plants and animals is such, however, that much of the actual research can be classified as fundamental. Most “pure” botanists and zoologists have preferred to work with organisms easier to handle and control in the laboratory with the result that their direct contribution to an understanding of forest trees and large forest animals has been somewhat limited.
Our forest scientists must be trained in the essentials of botany, but must also be exposed to various aspects of zoology, geology, soils, meteorology, statistics, photogrammetry, and other disciplines. They are almost invariably less specialized than comparable botanists—their closest relatives in the scientific world—but may well claim to be more liberally educated in that usually they have had a greater exposure to a greater range of subjects and of problems. They are not botanists, but scientists in their own right, and may be termed “forest scientists.” Those concerned primarily with the biology of forest trees may be termed “forester-botanists.”
The literature of forest science is both voluminous and diverse. The selection of most scientists from men trained as professional foresters at either the Bachelors’ or Masters’ level has given rise to a scientific body long on competence in the field, but short in competence as library workers. This combination of a
STEPHEN H.SPURR Professor, School of Natural Resources, University of Michigan, Ann Arbor, Michigan.
large literature and a body of scientists poorly equipped to cope with it has created serious problems in forestry research and has tended to limit the scientific productivity of the group.
The present paper is an effort to evaluate the requirements of forest scientists for scientific literature and reference services. The material is largely personal, drawn from the author’s experience in preparing two combination text-monographs and more than a hundred other written contributions in the field, and in founding and editing a number of serial publications.
The place of literature in forest research
An examination of almost any issue of a forestry journal indicates minimum reliance of most forest scientists upon literature survey.
Taking a single issue of the American professional journal strictly at random (Journal of Forestry, April 1951), a check of the ten major articles showed that 3 cited no references, 1 cited only his own earlier work, and the other 6 cited only American writings from essentially the current generation and locality. No historical references, references to other parts of the United States (except to textbooks), and no foreign references were included.
The situation in Forest Science, founded to stimulate scholarly work, is better as far as literature citations are concerned. Again taking a single issue at random (September 1957), about half of the 12 major articles indicate that the authors have made a serious attempt to find out what has been done on their problem by others. About an equal number—but not necessarily the same ones—have looked into the Canadian literature if they were Americans (or into American literature if Canadians). The same number cited one or more European references. One even cited the English summary of two Japanese articles—the only citations in the issue referring to other than American, Canadian, or western European work. Even so, the present writer, who has been editor of Forest Science since its inception, can state that for most of the articles submitted to this journal, one or more highly pertinent references uncited by the authors can be turned up merely by five minutes perusal of standard abstract and bibliographic works.
We may ask whether this failure to consult the literature with sufficient care and perspicacity affects the quality of current research. It does, and to a very marked extent. There is much to be learned from work on ecologically similar tree species and forest types in other parts of the world, from botanists and others who publish on similar subjects in other literatures than forestry, and from competent men of previous generations who were concerned with similar problems.
Without belaboring the obvious, we may note that forest science in western Europe dates from the middle of the nineteenth century, whereas American and Canadian research became significant only well into the twentieth century. The forest types of western Europe and eastern Asia contain many species closely related to ours and growing on soils and in climates similar to ours, thus presenting many comparable problems. The forest literatures in German, Russian, Swedish, Finnish, and other tongues are substantial and largely pertinent to American research.
Even in the United States, the work of a previous generation of foresters is all too often overlooked, and but little effort is ordinarily made to uncover applicable research published in botanical and zoological journals. Articles published in university and museum serials, in local academy of science journals, and in similar series of limited circulation are as apt as not to be completely ignored, especially if written by relatively unknown scientists.
An indication of what can be accomplished through exhaustive literature survey is given in a later section.
The nature of forestry literature
First, however, let us consider and evaluate the present literature relating to forest science.
Relatively little of it is adequately summarized in book form. There are essentially only two complete book literatures in forestry: one American and one German. A number of admirable books on individual subjects have been published in England, France, Switzerland, etc., but one must look to the United States and Germany for texts in most of the aspects of forestry.
Taking American book literature for our analysis, there are currently about two score books in print, largely textbooks for undergraduates published by three New York publishing houses. Of these, certainly less than half are products of men and work that have come into prominence since World War II, and only a portion of these may be said honestly to cover and interpret their fields adequately. In other words, we have in the neighborhood of a dozen or so texts that can be said to be both up to date and comprehensive. These cover perhaps half of the subjects taught in professional forestry schools today.
Monographic treatments in the English language are few and far between. In tree physiology and morphology, we still rely upon a 1929 translation of an earlier German work. In most fields we have not even that. Only one or two American tree species have been adequately monographed in recent years, and only a few segments of our subject matter field.
There is at least one forestry journal published in most countries at quarterly
intervals or better, and a good selection of these are available in a dozen or so American libraries. In the United States, we have had a professional journal for more than a half-century, and a scientific journal since 1955. These are generally available and are generally consulted.
A very substantial amount of scientific material in forestry, however, appears not in books or journals, but in minor serials that are distributed at irregular intervals to limited circulation lists. Most of the research of the U.S. Forest Service, for example, is reported in lithoprinted form by one of the ten regional experiment stations and two territorial research centers. Each station usually puts out a number of such serials, ranging from one-page throw-aways to research summaries in annual reports and more lengthy station papers. To maintain a file of these for the United States alone is impractical for most scientists. Even the library files are generally incomplete as few librarians catalog or otherwise keep careful check of this “ephemeral” literature.
In addition to the lithoprinted output of federal agencies, many additional data are included in similar lithoprinted and mimeographed leaflets distributed by forestry schools, industries, and state agencies.
While it must be acknowledged that the best research eventually is published in bulletin or journal form, the fact remains that a substantial portion of forestry literature is more or less buried in many hundred serials, reproduced by duplication processes other than conventional printing, published at irregular intervals, and circulated to small and often unedited mailing lists.
Finally, a substantial quantity of fundamental research is available only in unpublished theses. Fortunately, these are becoming more accessible through the use of microfilms.
Cases in literature survey
With this background, we may now investigate a number of specific cases as a means of illustrating the requirements of forest scientists for literature and reference services. These are all taken from the writer’s own experience, and represent instances where an effort has been made to evaluate the problem of what can be gotten from scientific literature on a specific topic, and how it can best be obtained.
FOREST AERIAL PHOTOGRAMMETRY
At the end of World War II, the author was active in research involving the application of aerial photographs to the inventory of forests and other natural resources. The program was quite successful in so far as widespread adoption of our innovations and principles was concerned. Most of us concerned with
the topic felt that we were in some sense pioneers opening new horizons of applied science. We were aware, however, of the Canadian use of aerial photographs and aerial sketching for reconnaissance purposes since World War I, and we knew that a few papers had appeared on the subject in various other countries.
In preparing a book on the subject, it was astounding to find several hundred references on forest photogrammetry. Among them was a series of doctoral dissertations written in Germany under the aegis of Dr. Hugershoff at the forestry school at Tharandt in Saxony, which, had the dissertations been available in English, would have greatly shortened discussions, arguments, and field investigations in this country in recent years.
Later, a complete bibliography on the subject was attempted (published by the University of Michigan) in view of the fact that existing bibliographies were noticeably incomplete. Almost no help was obtained from the library catalogs of the U.S. Department of Agriculture, Yale, Harvard, and other major American forestry libraries, and only a few additional references were picked up from major western European libraries (Oxford being the best in this regard). Getting to see many of the items themselves involved major activity in interlibrary loans, microfilm purchases, and foreign travel—activities too time-consuming to be carried out under most circumstances.
It may be amusing to note that the antiquity of our supposedly revolutionary postwar research is indicated by a news item uncovered in a Berlin newspaper of 1887 which relates the experiences of a practicing German forester who attempted to map his forest by taking photographs from a hot-air balloon, and anticipated many of the problems that confronted us nearly sixty years later.
On a subsequent research project at the Harvard Forest, we were confronted with a problem of the statistical correlations between the volume and growth of forest trees on one hand, and the conventional measurements of tree diameter, taper, height, and ring width on the other. Forewarned by the experience with photogrammetry, we undertook a prior analysis of previous work on the subject.
Choosing spruce and fir in the northeastern United States and adjacent Canada for the test, it was found that more than a score of investigators had assayed volume estimates independently, more or less ignoring the efforts of others attempting the same thing. Not only was the duplication excessive, but the labor involved was also so great as effectively to prevent progress in this field for many years.
In preparing a monograph on this subject, during the course of which a
reasonably complete survey of the literature was made, it became apparent that the lack of fundamental research projects based upon a thorough understanding of previous findings had resulted in holding forest inventory techniques to an empirical basis over a period of several decades.
SPECIFIC GRAVITY AND TREE GROWTH
A major controversy in forestry circles has long concerned the effect of growth rate on the specific gravity and other properties of the wood cells. In effect, one group, primarily Americans, have claimed a high correlation, while another group spearheaded by South Africans has countered that there was little or none. In 1950, the present writer initiated an experiment to try and resolve the problem for at least one species of pine in one locality.
At a later date, a routine survey of previous literature was begun. The large number of existing papers turned up prompted an effort to follow the problem back to its beginnings. It was found that Theophrastus clearly implied that the growth rate of trees was used as a guide to the strength properties of wood by the ancient Greeks. A substantial German literature on the subject contributed substantially to an understanding of the relationship. Finally, a detailed study in France around the middle of the nineteenth century was unearthed. The experiments carried on at that time and the conclusions reached had been ignored almost completely by later workers. Nevertheless, when the extensive literature was studied and the experimental data were finally tabulated, it became evident that the conclusions reached a hundred years before in France were still essentially sound. In fact, the whole problem could be resolved quite well by drawing inferences from the data in the existing literature—data that simply had not been carefully studied and digested by many of the later protagonists.
THE NATURE OF THE FOREST COMMUNITY
Although complete surveys of existing literature are possible when a scientist has the inclination and the time, and when the number of previous works are in the tens or the hundreds as in the specific gravity problem, many cases exist when such a survey is impossible. An illustration is provided by the current project of the writer, an effort to integrate conflicting theories and philosophical approaches on the nature of forest communities.
A project of this kind involves at least passing acquaintance with much of the entire literature of forest ecology, as well as with that of certain related aspects of tree physiology, genetics, plant geography, geobotany, soils, and meteorology.
Since 1945 alone, somewhere in the neighborhood of 20,000 items have
appeared in scientific literature around the world dealing with forest ecology or closely related aspects of other fields. All these items should be scanned. From them, perhaps 5000 references are directly applicable to the problem. By winnowing these and being ruthless in rejection, it is hoped to limit the final work to from 1000 to 2000 citations.
It is thus manifestly impracticable to read carefully all the literature relating to this subject that has appeared since 1945. Fully as much again was published prior to that date, including many of the basic works upon which the classical concepts and diverse schools of thought are based.
How to proceed? Reading all the originals is impossible, even were it not for the language barrier in many instances. Bibliographies and catalogs are invariably incomplete and out of date. Even were they not, they do not provide the necessary information. The only solution is some form of abstract or record-retrieval system. The survey (and the resulting book) must be developed essentially from abstracts, supplemented by a good deal of time-consuming library work on the most important items. In choosing these key writings, the scientist must rely upon his imperfect memory and notes of his past reading, and upon exposure to the work of others.
Clearly, studies of this type will continue to be limited in numbers and in quality as long as the scientist must rely upon original sources as the primary basis for his information.
Existing reference facilities
Fortunately, considerable progress has been made in the field of forestry for the systematic compilation of scientific information. Without this progress, the study last referred to would be impracticable even today.
A reasonably adequate and detailed system for the classification of forestry literature has been evolved at Oxford, revised by an international working committee, and published in key languages. This system, a decimal expansion of a single library class, covers many pages, provides for cross-indexing, and is accompanied by alphabetical and other indices.
At least three organizations are making a major effort to furnish bibliographic services of current forestry literature. These include the Commonwealth Forestry Bureau at Oxford, the U.S. Department of Agriculture, and the Federal Forest Experiment Station at Reinbek, Germany.
Of these, the output of the Oxford group is by all odds the most important to the English-speaking scientist. Here are published a quarterly Forestry Abstracts journal, giving more or less detailed abstracts of several thousand scientific papers each year; and a monthly title card service, providing 3-by-5-
inch cards carrying the bibliographic citation and commonly a very brief abstract. The former service provides abstracts of the more important works, while the latter covers in addition a substantial number of minor or quasi-scientific items, but provides but little detail as to the content of the listed items. Both services are keyed to the Oxford decimal classification as to main class and for cross-indexing.
In the United States, Biological Abstracts covers some forestry literature; but its coverage is incomplete, but little subdivided, and poorly cross-referenced. This journal, in the opinion of the present writer, is of relatively slight value or interest as far as its coverage of forestry literature is concerned. As an entree to botanical and zoological literature, however, it is of great value and importance.
Most American forestry libraries do not catalog separately or cross-reference adequately most of the journal articles and minor publications which form the bulk of the world forestry literature. Their card catalogs are thus of relatively little value to the scientist wishing to go beyond the obvious sources. (The forestry library at the University of California is an exception to this.)
There are, however, a number of excellent bibliographies available to forest scientists. The Bradley Bibliography compiled by Dr. Rehder for the Arnold Arboretum provides an almost unbelievable coverage of the literature dealing with woody plants prior to 1900. Furthermore, most of the items may be found in the various libraries at Harvard and Yale (although this in itself requires considerable sleuthing ability). The twentieth century literatures of the United States, Canada, and Germany are adequately summarized in various bibliographies up to relatively recent years. Postwar citations up to the current year may be obtained from the three bibliographic services mentioned above.
It is thus possible to construct a reasonably complete bibliography in a relatively brief time. For instance, American writings on almost any forestry subject may be compiled from the Bradley Bibliography (prior to 1900), the Munns Bibliography (to 1929), U.S. Department of Agriculture and Forest Service sources, and Forestry Abstracts (reasonably complete since 1945 or thereabouts). The problem of where to find the references and how to abstract critical information from them remains.
An approach toward a solution
The present writer has given a good deal of attention to the development of a literature record that will meet the requirements of the forest scientist. The personal system, evolved at the School of Natural Resources of the University
of Michigan since 1952, may serve as an indication of a possible solution in a scientific field where the research workers are still relatively few and the literature is still far from voluminous.
Essentially, the system is a large-card catalog of printed and other abstracts catalogued and indexed in accordance with the Oxford decimal classification. Two copies of Forestry Abstracts are periodically disassembled, the abstracts cut out, and filed in accordance with the classification code printed on them. The major items are pasted on index cards, while the minor contributions are filed by separate decimal classes and geographical locality in envelopes of the same size in the same catalog.
The system was evolved to be kept up to date by the scientist himself and his assistant with a minimum of time spent. An average of 1 to 2 man-days per quarter is all that is required. Reprints are kept in a parallel file (standard filing cabinets), following the identical classification scheme. In the course of doing a large part of the filing himself, and in perusing and filing reprints in accordance with the same system, the author has been able to keep reasonably well informed of current publications in his particular field.
Were the system to be modified to an institutional basis rather than a personal one (and this has already been done at another forestry school), obvious additions and modifications could be made to evolve a sound scientific information catalog. The following points would be considered by the writer were he given the task of developing a centralized reference catalog for forestry literature.
The Oxford classification would be used. It is the most satisfactory of existing systems, has international sanction, and its use avoids the necessity of reclassification of abstracts and bibliographic citations from the major bibliographic centers.
All information would be put on cards, eliminating the use of envelopes for cut abstracts.
The choice of card size poses a difficult question. Standard 3-by-5-inch cards are furnished by bibliographic centers, but are too small to contain abstracts without reduction. The present 4-by-6-inch cards will accommodate most abstracts without reduction, but folding of the longer abstracts on the back side is necessary. Larger 5-by-7-inch cards might prove most satisfactory in the long run.
All abstracting and bibliographic services should be subscribed to—and all pertinent abstracts and bibliographic citations would be separated and either reproduced or pasted on the cards (if not supplied as cards of the proper size). Forestry Abstracts and Biological Abstracts would form the basis for the current
literature records. All available annotated bibliographies would be drawn upon. Simple bibliographic citations would be included in the absence of an abstract.
Reproduction facilities would be developed for the printing of sufficient additional cards for cross-indexing.
Facilities for photographic reduction and reproduction would be developed for the printing of entire short papers on the standard cards. For instance, the large number of one- and two-page “research note” type of articles should be capable of reduction to a single 5-by-7-inch card, thus permitting the inclusion of the entire work rather than of an abstract. In order to permit the simultaneous handling of a large number of cards dealing with a single subject, the writer is inclined to limit the amount of reduction so that the product can be read with the naked eye or under simple magnification.
By following the procedures outlined above, it should be possible to construct and maintain a central literature file for world forestry literature at a very moderate cost. This file would take the form of a single catalog containing abstracts, citations, and photographic reductions of brief articles. With adequate cross-referencing, mechanical retrieval systems would not be necessary. As time and facilities became available—and this would undoubtedly require international cooperation—citation cards could be expanded to abstract cards, and abstract cards in many instances could be expanded to more detailed record cards or even to complete reproduction cards. Personnel trained as scientists in the field should take an active part in the operation, rather than trusting entirely to professional abstractors or librarians.
The advantages of the system are its complete practicability, its low cost, and the fact that it becomes immediately usable once established. The prototype developed by the author over a six-year period is constantly used and has resulted in its users achieving a better comprehension of the literature relating to their problems in less time than previously required for a less satisfactory survey.