HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (22) Citing Articles via Google Scholar Google Scholar Articles by Bouma, J. Articles by Booltink, H.W.G. Search for Related Content PubMed Articles by Bouma, J. Articles by Booltink, H.W.G. GeoRef GeoRef Citation Agricola Articles by Bouma, J. Articles by Booltink, H.W.G.

DIVISION S-5-PEDOLOGY

Pedology, Precision Agriculture, and the Changing Paradigm of Agricultural Research

^a Lab. of Soil Science and Geology, Wageningen Agricultural Univ., P.O. Box 37, 6700 AA Wageningen, The Netherlands

johan.bouma{at}bodlan.beng.wau.nl

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
Precision agriculture (PA) has recently been defined by the U.S. National Research Council as a management strategy that uses information technologies to bring data from multiple sources to bear on decisions associated with crop production. Soil information is important here, but current soil survey data do not satisfy PA requirements. In this paper, the need for soil data in PA is discussed on the basis of Dutch research. Not only operational, but also tactical and strategic aspects are considered. On the operational level, soil data, including parameters derived with pedotransfer functions, support the use of simulation models to quantify dynamically soil water regimes, N transformations, and biocide adsorption. Real time "forward-looking" simulations incorporated in early-warning systems assist in operational decisions on water, nutrient, and crop protection management. Backward-looking simulations, using historic weather data, can be used to evaluate different management tactics for exploratory strategic and tactical purposes. Such simulations should balance production and environmental requirements. At the strategic and tactical level, assembled data on the performance of various farm management systems should be grouped by soil series to build a systematic database, allowing "quick and preliminary" evaluations of the effects of farm management strategies based on experiences obtained elsewhere on similar soils.

Abbreviations: ICT, Information and Communication Technologies • GIS, geographical information systems • PA, precision agriculture • NRCS, USDA Natural Resources Conservation Service

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
FOR DECADES agricultural managers have taken advantage of new technologies to support them in the decision making process regarding farm management and improving the economic efficiency of operations. The extent and rate of development in Information and Communication Technologies (ICT) is opening the way for significant changes in crop management and agricultural decision making. ICT has led to cropping systems in which management operations are executed on a site-specific basis and with an increased temporal precision. This vision is reflected in the concept of precision agriculture, as recently defined by the National Research Council (1997): "Precision agriculture is a management strategy that uses information technology to bring data from multiple sources to bear on decisions associated with crop production".

In Europe, where population densities are relatively high, a broad shift is being made from exclusive emphasis on production agriculture to multifunctional land use in which various forms of environment-friendly agriculture are combined with many other functions, including transport, housing, and recreation. (Bouma, 1997a). This shift, in combination with the accelerating developments in ICT, has major implications for agricultural research and education (National Research Council, 1997). New interdisciplinary approaches are needed to improve the understanding of complex interactions between multiple factors affecting crop growth and farm decision making. Unbiased, systematic and rigorous evaluations of the economic and environmental benefits and costs of agriculture are needed. New developments in precision agriculture need to be evaluated along the same lines.

The advance of ICT is such that we may speak of a paradigm shift in agricultural research and education (National Research Council, 1997). This paper addresses the question as to the implications of this paradigm shift for soil science. Soil science, and more specifically pedology, is uniquely positioned to consider the behavior of different soils in a landscape. However, the National Research Council (1997) concludes that "current soil surveys satisfy few of the soil data requirements of PA. Soil data are not at the appropriate level of detail nor are the indexes required by PA the same as those provided by soil surveys". This paper intends to analyse critically this statement and to illustrate the use of soil survey information at strategic, tactical, and operational levels of PA in the Netherlands. Thus, procedures are presented to initiate a response to the recommendations of the National Research Council ( 1997) for soil survey to define (i) data quality standards; (ii) methods for data collection, testing, and interpretation; and (iii) procedures to access and archive data by private consultants.

	Information and Communication Technologies in Soil Science and Pedology

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
Soil surveys at different scales have been completed in many countries and data have been assembled in geographic information systems. The USDA Natural Resources Conservation Service (NRCS), with local and state agencies and land grant Universities, has been generating soil information in the USA for a 100 yr. Although originally focused on the agricultural use of soil data, the mission of NRCS is now a much broader one: "to help people conserve, improve and sustain our natural resources and environment" (Natural Resources Conservation Service, 1995). Of course, many primary data gathered in the soil surveys cannot directly be used for precision agriculture. Historically, soil characterization procedures had a strong taxonomic focus, leaving soil functional properties unattended. A relation to land use was only established in descriptive terms: relative suitabilities and limitations for major land use types were summarized for a given soil series. These include non-agricultural forms of land use, such as forestry, campgrounds, on-site waste disposal, housing, and roads. Similar developments occurred in other countries. In the Netherlands, the country–wide 1:50 000 soil survey was completed in 1995, providing an extensive soil database containing primary as well as interpretive data (Bregt et al., 1987).

ICT developments in soil science have stimulated the application of geographical information systems (GIS) with advanced possibilities for data storage, data analysis, and visualization. This has led to the development of GIS-based soil information systems (e.g., Soil Survey Staff, 1993) with specific soil related functionalities. As a result, existing soil survey data have become available for different applications and a wide range of users. Computer technology has also enabled the derivation of pedotransfer functions, which relate simple soil characteristics, such as texture, organic matter content, and bulk density to more complex characteristics, such as hydraulic conductivity and moisture retention. In the Netherlands, hydraulic characteristics can be estimated for all soils by pedotransfer functions derived from the Staring Series soil database (Wösten, 1997). These hydraulic characteristics can be used in dynamic simulation models to quantify soil water regimes and crop growth. (Bouma and van Lanen, 1987; Tietje and Tapkenhinrichs, 1993; Wösten et al., 1998, and many others). Such simulations have been used successfully to define N transformations and N fertilization scenarios in the context of precision agriculture, and results obtained by pedotransferfunctions were as good as those based on direct measurements (Verhagen and Bouma, 1998). Use of soil information for PA needs to be defined as a function of questions being raised for different time scales. This will be analyzed next.

	Dimensions of Decision Making

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
Management decisions to be made by farmers can be classified in terms of strategic, tactical, and operational decisions, all of which are focused on realizing a profitable enterprise (Bouma, 1997b). Strategic decisions have a time scope of 10 yr or more and concern issues like the selection of a farming system, such as mixed, organic, or integrated. The choice to switch from conventional to precision farming may be considered a strategic decision as well.

Tactical decisions cover a period of around 2 to 5 yr, corresponding roughly to the time span of a crop rotation. The selection of a rotation scheme mainly involves agronomic considerations. With respect to soils, we may consider soil water regimes, soil tillage methods, mineralization of organic matter, nutrient dynamics, and structural stability as being important soil factors when considering the most desirable crop rotations and associated management practices.

Finally, operational decisions are taken on a day-by-day basis as the growing season proceeds. These include the selection and timing of management operations such as planting, harvesting, fertilizer application, and crop protection measures. Much research on PA focuses on the operational level, even though this is never specified. Moreover, most studies have a traditional backward-looking mode: results are reported and analysed "after-the-fact" while the real challenge is to provide data for decision support systems for the farmer who has to apply adaptive management in a "forward-looking" mode as he faces unknown weather conditions in the days and weeks to come.

Information to support farmers in various decision making processes has been provided by talking to colleagues and through extension services, private advisors, and, more recently, decision support systems on personal computers. Although decision support mainly concentrates on the operational level, strategic and tactical decisions are equally important and farmers are in need of support here as well. Compiling data on yields and environmental side effects of different farming systems for important soil series could be a helpful tool in support of tactical and strategic decision making. For example, a choice for precision agriculture is logical if a given soil series, occurring in a farm, has a high internal variability or if many different soil series are found within a farm. An example for mixed dairy farming in sandy soils of the Netherlands was recently presented by Hack-ten-Broeke et al. (1999) who presented yields and nitrate leaching for five major soil series which were subjected to an innovative management system developed in close interaction with farmers. Thus, "quick and preliminary" exploratory analyses of the effects of different types of management on different soil series, which may occur on a given farm, can be helpful in the selection of the most appropriate management scheme. Exploratory use of simulation models for strategic and tactical analyses, as applied by Hack-ten-Broeke et al. (1999), requires pedotransfer functions that, just like data on soil series, are derived from soil survey data. The operational decision support system for PA requires special discussion.

	An Operational Decision Support System for Precision Agriculture: Outline and Soil Data Requirements

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
Even though research is still in progress, the outline of an operational system for precision agriculture for arable farming in the Netherlands is taking shape. An inventory of components to be included in a decision support system is being made with a primary focus on operational decisions and soil related variability. A forward-looking approach is being pursued, allowing farmers to respond pro-actively to problems expected to develop in the production system, both in terms of production decline or exceedance of environmental threshold values (Bouma, 1997b; Booltink et al., 1996). The system reflects Dutch conditions and consists of several components, of which only the soil-related ones will be discussed below.

Soil Database
A soil database obtained from soil sampling observations (point data), in a grid pattern, over the fields of the farm needs to be established. Once obtained, this data base would be valid for many years; it represents a one-time, strategic investment. The optimal observation density is a function of the spatial variability in the field. It can be determined by analyzing variability in remotely sensed data, yield maps, and exploratory soil surveys using geostatistical techniques (e.g., Verhagen et al., 1995; Van Uffelen et al., 1997). Observations focus on primary soil data that are characterized as part of well established standard soil survey practices, i.e., layer structure, texture, bulk density, and organic matter content (Fig. 1) . Secondary soil data (e.g., soil hydraulic characteristics) are derived for each soil layer through continuous pedotransfer functions. An increasing number of pedotransfer functions is available for many complex soil properties (Wösten, 1997; Wösten et al., 1998; Van den Berg et al., 1997) making costly field measurements unnecessary. This makes a big difference. Creating a soil database for a 120-ha farm costs about $10000, on the basis of commercial survey rates in the Netherlands. This is acceptable for farmers. However, obtaining representative hydraulic characteristics for major horizons in the area by direct measurement would cost at least $30000, which is clearly prohibitive.

View larger version (28K): [in this window] [in a new window]   Fig. 1 Structure and contents of the soil database included in the decision support system for precision agriculture

Real-Time Simulation
Together with meteorological data from an on-farm weather station, soil parameters are fed into a dynamic, mechanistic simulation model quantifying water movement, N transformations, and biocide adsorption. Real-time simulations with the mechanistic model for selected representative soil profiles in each management unit are carried out. Initial mineral N contents in the soil are sampled at the start of the growing season, allowing the model to start daily calculations of crop growth as determined by incoming radiation, temperature, water regime, and soil nitrogen content. Radiation, temperature, and precipitation are measured daily on the farm and are sent by email to the modellers, allowing real-time simulations. The simulations are part of an early warning system, alerting the farmer to near depletion of mineral N supplies in the root zone and thus assisting him in optimizing the timing for split fertilizer application. In calcareous Dutch marine clay soils, N fertilization is most critical; K contents are generally adequate and P is applied following standard rules. If the deterministic simulation models are not calibrated it is advisable to monitor soil moisture and soil nitrogen. Time domain reflectometry (TDR) devices are installed in situ at different depths to provide a time series of soil moisture contents for the major land units. An example of a real-time simulation run for wheat (Triticum aestivum L.) in 1998 is presented in Fig. 2 . Data allowed recommendations for N applications on 18 April and 23 May, well after the time that this would have been considered in practice. The question remains as to the quantity of N to be applied. Booltink et al. (1996) used a weather generator and historic weather records to predict conditions in the remainder of the growing season, resulting in a given yield. They show that , as the season progresses, such predictions become more accurate. The amount of fertilizer to be applied should be a function of this expected yield: lower yields require less N than higher ones.

View larger version (19K): [in this window] [in a new window]   Fig. 2 Optimization of split fertilizer applications (70, 60, and 45 kg N ha-1) using simulated soil nitrogen levels and weekly plant uptake rates

Tillage
Periods can be identified during which soil traffic and tillage are not advisable because of an increased risk of causing compaction or puddling. Threshold values for soil moisture contents in the topsoil have been defined on an empirical basis (Droogers et al., 1996) and are compared with simulated moisture contents.

The system, in summary, consists of a simulation model for water movement, N transformation and biocide adsorption, which is applied to selected representative soil profiles for each management unit on a real-time basis. Characterization of water movement constitutes the core of the system. The applied model presents daily soil-water contents and fluxes which, in turn, serve to derive early warning signals for N depletion, trafficability, workability, and, partly, for control of pests and diseases. A decision support system, developed according to the outline described above, will require detailed information on soil variability within fields. Soil data needed for the system are indicated in Fig. 1. Standard soil surveys (ranging from 1:12000 to 1:20000 in the USA and 1:50000 in the Netherlands) cannot deliver the required level of accuracy. An additional soil database, designed specifically for precision agriculture, will therefore have to be created. However, standard soil survey practices can provide valuable input and expertise relating to strategic, tactical, and operational levels of decision making.

	Soil Series: Genoforms and Phenoforms

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
The efficiency of soil series as carriers of information for tactical and strategic decision making can be improved by distinguishing between major soil management types. Droogers et al. (1997) showed significant differences between a soil kept under organic management for 70 yr and a similar soil, in terms of classification, kept under conventional management. In turn, soils classified to different soil series may in time behave similarly under identical management. Droogers and Bouma (1997) suggested the term genoform for the genetically defined soil series and phenoform for soil types resulting from a particular form of management in a given genoform. We advocate the establishment of a database for well-defined phenoforms, originating from the major genoforms included in Soil Taxonomy. Droogers and Bouma (1997) provided an example for three phenoforms of a prime agricultural soil in the Netherlands, defining yields as a function of N fertilization rates, the latter constrained by a threshold value for nitrate leaching to the groundwater.

Significance of Continued Interaction: Need for Basic Research
The operational prototype Decision Support System for Precision Agriculture, described above, will evolve in time as interaction with farmers, consultants, and researchers continues. Models for N transformation and biocide adsorption will improve through basic research, possibly requiring new soil data. Such requirements should, however, only be defined in close consultation with soil scientists to assure continuity of the system. As basic research in soil science is being challenged by funding agencies, it is important to show that such work is essential to the functioning of decision support systems. These should evolve through a combination of basic and applied research, with a two-way feedback loop in a research chain (e.g., Bouma, 1998). As suggested by the National Research Council ( 1997), and as illustrated by this review, soil survey organizations could provide important services by defining required data, quality standards, methods to collect, test and interpret data, and procedures to access and archive data by private consultants or farmers themselves.

	Conclusions

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
Standard soil surveys, their databases, and sampling protocols can provide valuable information for precision agriculture, even though they do not reflect soil variability to the level of accuracy needed for precision management.

1. An effort should be made to assemble and simulate yield and environmental data for major soil series subjected to different farm management systems. This will allow quick, exploratory evaluations of the effects of different farm management strategies in the context of strategic and tactical decision making.
   
2. The role of soil series, referred to as soil genoforms, as carriers of information, could be improved by further distinguishing between soil phenoforms, resulting from specific forms of management in a given soil series.
   
3. Decision support systems focused on the operational level require a detailed soil database constructed specifically for precision agriculture. The database, including primary soil data and secondary parameters derived through pedotransfer functions, should support the use of real-time simulation models to quantify dynamically soil water regimes, N transformations, and biocide adsorption in a forward-looking mode.
   

National Resources Conservation Service 1995

	ACKNOWLEDGMENTS

 
The research of Dr Stoorvogel was funded by a fellowship of the Royal Netherlands Academy of Arts and Sciences.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 
Contribution from the C.T. de Wit Graduate School of Production Ecology.

Received for publication October 8, 1998.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION Information and Communication... Dimensions of Decision Making An Operational Decision Support... Soil Series: Genoforms and... Conclusions REFERENCES

 

• Boesten J.J.T.I., Gottesbueren B. Testing PESTLA by two modellers for bentazone and ethoprophos in a sandy loam. Agric. Water Manage. 2000;in press.
• Boesten J.J.T.I., Van der Pas L.J.T. Movement of water, bromide ion and the pesticides ethoprophos and bentazone in a sandy loam: description of the Vredepeel data set. Agric. Water Manage. 2000;in press.
• Booltink, H.W.G., J. Verhagen, J. Bouma, and P.K. Thornton. 1996. Application of simulation models and weather generators to optimize farm management strategies. In P.C.Robert et al. (ed.) Precision agriculture. Proc. Third International Conference, Minneapolis, MN. ASA, CSSA, and SSSA, Madison, WI.
• Bouma J. Soil environmental quality: a European perspective. J. Environ. Qual. 1997;26:26-31 a.[Abstract/Free Full Text]
• Bouma, J. 1997b. Precision agriculture: introduction to the spatial and temporal variability of environmental quality. p. 5–17. In CIBA Foundation. Precision agriculture: Spatial and temporal variability of environmental quality. CIBA Foundation Symposium 210. John Wiley and Sons, New York.
• Bouma, J. 1998. Introduction. p. 1–11. In J.J. Stoorvogel, et al. (ed.) Information technology as a tool to assess land use options in space and time. Proceedings of an International workshop, Lima. 28 Sept.–4 Oct. 1997. Quantitative Approaches in Systems Analysis No. 16. The C.T. de Wit Graduate School for Production Ecology and DLO Research Institute for AgroBiology and Soil Fertility, Wageningen, The Netherlands.
• Bouma, J., and H.A.J. van Lanen. 1987. Transfer functions and threshold values: from soil characteristics to land qualities. p. 106–111. In Quantified land evaluation. Proc. of a workshop by ISSS/SSSA. 10 –15 May 1986. ITC Publ. no. 6. ITC, Enschede, The Netherlands.
• Bregt, A.K., J. Denneboom, Y. van Randen, and E. Rose. 1987. Gebruikershandleiding bodemkundig informatiesysteem.Rapport 1945. Stichting voor Bodemkartering, Wageningen, The Netherlands.
• Droogers P., Fermont A., Bouma J. Effects of ecological soil management and trafficability of a loamy soil in the Netherlands. Geoderma 1996;73:131-145.
• Droogers P., van der Meer F.B.W., Bouma J. Water accessibility to plant roots in different soil structures occuring in the same soil type. Plant Soil 1997;188:83-91.
• Droogers P., Bouma J. Soil survey input in exploratory modeling of sustainable soil management practices. Soil Sci. Soc. Am. J. 1997;61:1704-1710.[Abstract/Free Full Text]
• Hack ten-Broeke M.J.D., Schut A.G.T., Bouma J. Effects of implementing newly developed sustainable land use systems for dairy farming on sandy soils in the Netherlands. Geoderma 1999;91:217-237.
• National Research Council. Precision agriculture in the 21st Century. Washington, DC: National Academy Press. , 1997.
• National Resources Conservation Service. Data Rich and Information poor. Washington, DC: A report to the chief of the Natural Resources Conservation Service by the Blue Ribbon Panel on Natural Resource Inventory and Performance Management. USDA, 1995.
• Robert, P.C., R.H. Rust, and W.E. Larson (ed.) 1995. Site-specific management for agricultural systems. Second International Conference, Minneapolis. 27–30 March 1994. ASA, CSSA, and SSSA, Madison WI.
• Sadler E.J., Busscher W.J., Bauer P.J., Karlen D.L. Spatial scale requirements for precision farming: a case study in the southeastern USA. Agron. J. 1998;90:191-197.[Abstract/Free Full Text]
• Soil Survey Staff. State soil geographical data base (STATSGO): data users guide. Washington, DC: Miscellaneous Publication Number 1492. USDA-SCS, 1993.
• Stafford, J.V. (ed.) 1997. Precision Agriculture '97. Papers presented at the first European conference on precision agriculture, Warwick, UK. 7–10 Sept. 1997. BIOS Scientific Publ. Ltd, Oxford, UK.
• Stoorvogel J.J., Kooistra L., Bouma J. Spatial and temporal variation in nematocide leaching, management implications for a Costa Rican banana plantation. In: Corwin D.L., et al. , ed. Assessment of nonpoint source pollution in the vadose zone. Washington, DC: Geophysical Monogr. 108. Am. Geophysical Union, 1999:281-289.
• Tietje O., Tapkenhinrichs M. Evaluation of Pedo-Transfer Functions. Soil Sci. Soc. Am. J. 1993;57:1088-1095.[Abstract/Free Full Text]
• Van den Berg M., Klamt E., van Reeuwijk L.P., Sombroek W.G. Pedotransfer functions for the estimation of moisture retention characteristics of Ferralsols and related soils. Geoderma 1997;78:161-180.
• Van Uffelen C.G.R., Verhagen J., Bouma J. Comparison of simulated crop yield patterns for site specific management. Agric. Systems 1997;54:207-222.
• Verhagen A., Booltink H.W.G., Bouma J. Site specific management: balancing crop production and environmental requirements at farm level. Agric. Systems 1995;49:369-384.
• Verhagen J., Bouma J. Defining threshold values for residual N-levels. Geoderma 1998;85:199-213.
• Wösten J.H.M. Bodemkundige vertaalfuncties bij SC-DLO: State of the art. The Netherlands: Wageningen, 1997 Technical Document, DLO-Winand Staring Centre.
• Wösten J.H.M., Lilly A., Nemes A. Using existing soil data to derive hydraulic parameters for simulation models in environmental studies and in land use planning: final report on the European Union Funded project, 1998. Wageningen, the Netherlands: DLO-Winand Staring Centre, 1998.

This article has been cited by other articles:

				F Baret, V Houles, and M Guerif Quantification of plant stress using remote sensing observations and crop models: the case of nitrogen management J. Exp. Bot., March 1, 2007; 58(4): 869 - 880. [Abstract] [Full Text] [PDF]

				H. Lin and H. Lin Hydropedology: Bridging Disciplines, Scales, and Data Vadose Zone J., February 1, 2003; 2(1): 1 - 11. [Abstract] [Full Text] [PDF]

				B. J. van Alphen and J. J. Stoorvogel Effects of Soil Variability and Weather Conditions on Pesticide Leaching-- A Farm-Level Evaluation J. Environ. Qual., May 1, 2002; 31(3): 797 - 805. [Abstract] [Full Text] [PDF]

				B.J. Van Alphen and J.J. Stoorvogel A Functional Approach to Soil Characterization in Support of Precision Agriculture Soil Sci. Soc. Am. J., September 1, 2000; 64(5): 1706 - 1713. [Abstract] [Full Text]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (22) Citing Articles via Google Scholar Google Scholar Articles by Bouma, J. Articles by Booltink, H.W.G. Search for Related Content PubMed Articles by Bouma, J. Articles by Booltink, H.W.G. GeoRef GeoRef Citation Agricola Articles by Bouma, J. Articles by Booltink, H.W.G.