A growing group of researchers and practitioners are pursuing the opposite tack. Indigenous knowledge and practices of local farmers are being investigated in the hope that they may reveal ideas that contain "seeds" of adaptive value (Alcorn 1984, Hunn 1985, Rhoades 1990, Richards 1985). Towards this end, Knight (1980) has called for the systematic documentation of traditional farmers' knowledge into "information bank" from which agronomists, extension workers, and other farmers draw enlightenment. This is not to say that local beliefs are always objective and correct. Nonetheless, such repositories of farmers' traditional knowledge will be useful antidote to the posture, much resented by local farmers, of pouring agricultural "know-how" into what are presumed to be empty vessels.
International and national research centers have embarked on programs to collect germplasm of different varieties of staple crops, including rice, wheat, maize, barley, sorghum, millet, potato, and sweet potato. This is motivated by the recognition that, "while breakthroughs in genetic engineering are occurring with increasing speed options for the future are being foreclosed by the erosion of one of the world's most important heritages, the genetic diversity of our crop plants and their wild relatives" (Plucknett et al. 1987). The germplasm collections-in the form of cultivated plants, plantlets in tissue culture, or seeds are preserved for posterity in well-equipped genebanks that are managed by teams of highly qualified experts (Huaman 1988).
However, very little cultural information is included in the collection data. The "passport data" that are routinely collected include such information as varietal name, location and altitude of site, date of collection, and name and institutional affiliation of collector. The information may, but does not always, indicate local name, and rarely does it contain any data on indigenous technologies (Vaughan 1988, Rhoades 1990, personal communication). Thus, one can envision a disconcerting scenario wherein, when the need arises, the germplasm may still be available but the associated technologies, mostly those traditionally relied on by local farmers, would longer be accessible.
This paper argues, therefore, that there is a pressing need for the systematic documentation or memory banking of local farmers' indigenous practices with traditional varieties of staple and supplementary crops. Without this complement, the genetic information preserved in genebanks will be decontextualized the human and ecological forces that shaped their selection will be largely ignored. Essentially what is called for is banking on codes, or banking of codes, for posterity on two fronts: the genetic, which embodies millions of years of biological evolution and the cultural which embodies a long and rough--but often successful history of human adaptation (Fig. 2).
Certain parallelisms can be drawn between a genebank and a memory bank. While germplasm encodes genetic information that has evolved through time as a response to selection pressures, the minds of local farmers who have had considerable experience in growing these crops are repositories of cultural information: coded, time tested adaptations to the environment. With the present trend in agricultural development, genetic information coded in wild strains and traditional crop varieties or landraces is threatened with erosion. Pressure towards more intensive monocultural production favors the adoption of newer, higher yielding cultivars. In the same manner, cultural practices associated with traditional varieties are in imminent danger of being swamped by modern technologies.
The preservation of genetic variability is imperative. But with the homogenizing effect of modernization, variability in technological alternatives and crop choices is becoming narrower and narrower. Farmers' options are increasingly restricted to agricultural packages proffered by extension agents, backed up as they are by loans and other incentives for procuring recommended seeds and other inputs. Unfortunately, this reification may approach a point where the existence of alternatives is hardly recognized.
This problem merits urgent attention. Largely because of the "success" of the Green Revolution, traditional rice varieties are rare in lowland, irrigated areas of the Philippines. Since high-yielding varieties of other crops are presently being developed, traditional varieties of these crops may soon have the same fate. As a corollary, indigenous technology has been replaced by modern technology at a fast pace in many areas. It is highly probable that within the next decade, much of this valuable cultural information generated and refined by long periods of co-evolution will be irretrievably lost. However, just as genetic information coded in the germplasm can be stored, cultural information in the minds of experienced and knowledgable farmers can be tapped, documented, and preserved for posterity.
In the case of drastic and unforeseen changes in the environment, the presence of genetic variability on which natural selection can operate ensures perpetuation of the species, and even of life itself. In the same manner, sociocultural evolution works on the cultural variability that exists within any population in terms of knowledge, technology, and life ways. If this variability is missing in this case, in terms of knowledge systems coding information about agricultural practices then the population has lost its most significant reservoir of adaptive capability.
The motivation behind any germplasm collection is to preserve genetic diversity or to prevent genetic erosion. We should keep in mind, however, that genetic information and cultural information are co-evolving systems, and one major force in the decline of biotic diversity worldwide is human intervention. Since the inception of the Green Revolution, selection, breeding, and culling have significantly reduced genetic variation, culminating in the widespread effort to disseminate and promote only the "best" varieties of each food crop. This trend has reached unparalleled proportions in modern times, with the premium placed on high yield, short maturity, input sensitivity and quick turnover for both varieties and technologies. Complementing genebanks with memory banks of cultural information can provide agriculture scientists and development workers with insights into sustainable technologies used in some regions and with potential application in other regions or in the future.
This brings us to another basic principle behind germplasm collections-that ultimately, accessions must be usable and useful. According to the 1988 Annual Report of the International Board for Plant Genetic Resources, it "seeks to guarantee the long-term conservation of plant genetic resources, but must also seek to ensure that these materials are used." In reality, much of the material remains underutilized. Observing that only a small percentage of the germplasm collection is actually tapped, a plant breeder commented that genebanks run the risk of becoming libraries without readers" (Chujoy, 1990, CIP, personal communication). Could this perhaps be due to the fact that what is stored is often unintelligible to potential users, devoid as such germplasm is of any background information on why it was cultivated, what it was good (or bad) for, how it was managed, and why it might be useful in the future or in other places? Experience with respect to agriculture, particularly soil conservation and integrated pest management, tells us that sometimes the most effective solutions to persistent problems could be those which have been applied by other groups for ages rather than those that scientists concoct in their laboratories (Lightfoot et al. 1987, Fujisaka 1988, Warren 1984).
In view of the above considerations, what is needed is a way to systematically document, store, and retrieve information on cultural practices associated with traditional crop varieties such that potentially useful technologies and varieties will be available and accessible when the need arises. Towards this end, I have proposed a preliminary methodological framework to supplement standard procedures for processing genebank accessions with methods for collecting and storing cultural information (Fig. 3).
In the sweet potato memory banking project, participant observation, a benchmark survey using a loosely structured interview schedule, and not-so-rapid rural appraisal involving transect drawing, agricultural calendar making and variety diagramming, sorting and ranking were all utilized to learn from the farmers of Bukidnon about their practices and preferences with regard to sweet potato: choice of planting area, land preparation, procurement and preparation of planting materials, selection of varieties, timing and methods of planting, pest and weed management, maintenance of soil fertility, harvesting, and storage. Collection, drying, and preservation of herbarium specimens of local varieties were undertaken simultaneously with the gathering of benchmark information and continued as more familiarity with the area was gained. Subsequently, key informant life histories were elicited by informal interviews, guided whenever necessary by probes to encourage them to reminisce about and describe varieties planted and varieties "lost," appropriate soil types, farming implements used, agricultural calendars and rituals, and methods of preparation and storage.
Transcripts of taped interviews recording farmers' perspectives about characteristics of and agricultural practices associated with traditional crop varieties have been produced. In addition, memory bank files of general characteristics, associated technologies, and local evaluations have been stored.
To verify and also to add rigor to the data collection, two other methods were used. First, the triads test was used to get all locally relevant criteria for varietal discrimination and choice. The test consisted of sets of stimuli, each set composed of three sweet potato varieties known in the area. Second, a systematic field and market survey was conducted towards the end of activities to fill up the observed gaps in information pertaining to indigenous technologies and local evaluations and to verify which varieties are still in existence and which ones are simply being recalled. (See Memory Banking Protocol for more detailed recomendations on methods.)
