Back to Home Page of CD3WD Project or Back to list of CD3WD Publications

Home - English - French - German - Italian - Portuguese - Spanish
                            POSTHARVEST GRAIN LOSS
                              ASSESSMENT METHODS
                    A Manual of Methods for the Evaluation
                             of Postharvest Losses
                           developed and compiled by
                               Kenton L. Harris
                               Carl J. Lindblad
                         published in cooperation with
                  The League for International Food Education
                   The Tropical Products Institute (England)
            Food and Agriculture Organization of the United Nations
        Group for Assistance on Systems Relating to Grain After-Harvest
                                    by the
                    American Association of Cereal Chemists
                           under Grant AIB/ta-G-1314
        Office of Nutrition, U.S. Agency for International Development
Cover figure: British Crown Copyright. Reproduced with permission of the
Controller of Her Britannic Majesty's Stationery Office.
                     This volume stems from the joint and
                     independent efforts of many who have
                   contributed ideas as well as manuscripts.
Contributors and Authors
     Harpers Ferry, WV, Meeting, September 8 10, 1976
     Slough, England, June 13-24, 1976              
     When world food production is viewed as a system, loss and deterioration
     is seen as a major food-limiting factor. Postharvest loss reduction would
     benefit from reliable loss estimates and cost/benefit comparisons;
     improvements also must be acceptable and feasible to introduce.
I.   Introduction. K. L. Harris and C. J. Lindblad
     Determination of postharvest grain losses requires a blending of, and
     concepts from, several sciences.
II.  Terms of Reference. K. L. Harris and C. J. Lindblad
     A.   Definitions                                                    
         Postharvest, losses, food, insects, microbiological defined.
     B.   Planning: An Overview for Project Administrators. K. L. Harris                   15
         Project planning involves many disciplines and concepts, from national
         priorities to logistics and local cultural values.
     C.   An Overview of the Postharvest System: The Food Grain Supply
         Pipeline. K. L. Harris, W. J. Hoover, C. J. Lindblad, and H. Pfost                19
         Determination of losses should proceed stepwise from understanding
         the overall grain-food pipeline to location of leaks and sites where
         losses are relatively important, can be assessed, and are amenable to
         loss-reducing interventions.
     D.   Preliminary Examination of Specific Problem Points and Making On-Site
         Rapid Appraisals. G. G. Corbett, K. L. Harris, H. Kaufmann,
         and C. J. Lindblad                                        
         Rapid on-site appraisals (30-60 days) are both workable and useful to
         determine feasibility for further investigations and for some inputs,
         and to delineate specific problem points.
III. Social and Cultural Guidelines                     
     A.   The Fact-Gathering Milieu. Allan L. Griff    
     B.   Anthropologic Signposts. C. C. Reining        
         Grain loss does not exist independent of human and social influence.
         Loss assessment and reduction programs need to be seen from within
         the local setting. Cross-cultural sensitivity and understanding are
         essential in planning and executing such efforts. Reminders are given
         on who, what, and how to obtain reliable, useful information on and
         within the social and domestic organizations and in relation to
         individuals. Special attention is given to the role of women.
IV.  Representative Sampling, Interpretation of Results, Accuracy, and
     Reliability. B. A. Drew, with T A. Granovsky and C. J. Lindblad
     Basic statistical requirements for surveys, sampling, probabilities, and other
     concepts required in the assessment of losses are presented.
     A.   Introduction
     B.   Probability Samples
     C.   Detailed Instructions
V.  Loss Measurements as Related to Situations Where They Occur
     A.   Background Information. D. A. V. Dendy, with K. L. Harris
         Losses are discussed as they occur during threshing, cleaning and
         winnowing, drying, parboiling, hulling and polishing, and grinding.
     B.   Guidelines for Performing Studies of Farm Storage Losses. J. M.
         Adams and G. W. Harman
         Evaluation of maize losses in small farms is used to explain loss
         methods development.
     C.  Procedures for Measuring Losses Occurring During or Caused by
         Processing including Threshing, Drying, and Milling of Most Grains,
         but not Maize or Pulses/Groundnuts. D. A. V. Dendy, with K. L.
         Guidelines for studying:
           * Farm-storage losses
           * Total system losses
           * Operator-induced losses
           * Threshing loss with the straw
           * Threshing loss, grain damage
           * Maize shelling loss on the cob
           * Maize shelling loss, grain damage
           * Dryer-induced loss, laboratory method
           * Dryer-induced loss, method for use in mill
           * Batch dryer testing
           * Continuous dryer testing
           * Grinding loss as bran
             Comparison of milling yields by variety
             Comparison of operators
             Comparison of mills
             Due to insect damage
           * Rice milling losses
             Batch process
             One-stage continuous process
             Two-stage continuous process
           * Rice hulling losses
           * Rice polishing losses
VI.  Standard Measurement Techniques                             
     A.   Preamble to the Methodology. K. L. Harris and C. J. Lindblad
         General background of previous work, previously used estimating
         procedures and techniques, standardization of results.
     B.   Losses Caused by Insects, Mites, and Microorganisms. J. M. Adams
         and G. G. M. Schulten
         An explanation of several techniques based either on the weight of a
         measured volume of grain compared with a pre-loss standardized weight
         or on the separation of damaged kernels and the comparative weights
         of damaged to undamaged calculated to the whole sample. Also a
         conversion factor/percent damage method. Weight/unit volume,
         counts and weights of damaged and undamaged kernels, percent of
         damage and weight loss, and conversion factor/percent damaged
         methods are presented.
          * Standard volume/weight method for damage by insects and
          * Modified standard volume/weight method when a baseline
            cannot be determined
          * Count and weigh method
          * Converted percentage damage method
     C.   Losses in Grain Due to Respiration of Grain and Molds and Other
         Microorganisms. R. A. Saul, with K. L. Harris
         Weight loss due to grain respiration is unimportant until the moisture is
         so high that serious microbial deterioration occurs and rejection for
         food use becomes the dominant factor. Tables are given for calculating
         losses based on time, temperature, moisture, and physical damage. A
         formula is given for calculating losses based on weight of damaged and
         undamaged kernels. Rationale and techniques are presented for basing
         losses on locally applied rejects.
     D.   Rodents                           
         1.   General Considerations, Direct Measurement Techniques, and
             Biological Aspects of Survey Procedures. W. R Jackson and M.
             Each rodent ecosystem has features that tend to make it unique.
             Loss evaluations require preliminary investigation to establish an
             environmental and loss perspective as to what features require
             and are amenable to assessment.
         2.   Loss Determinations by Population Assessment and Estimation
             Procedures. J. H. Greaves
             When they can be undertaken, census and food-intake
             procedures will give useful results. Three techniques are
               * Survey for infestation
               * Census trapping and food intake calculation
               * Lincoln-Peterson method for population estimation
     E.   Measurement of Losses Caused by Birds                    
         By brief summary only.
     F.   Moisture Measurement, T. A. Granovsky, G. Martin, and J. L. Multon 
         The measurement of grain moisture is critical for proper assessment of
         weight changes during storage. (See Appendix C for methods.   A
         nomograph is given for calculating weight changes resulting from
         moisture content changes.
VII.  Operations Standardization and Control
      From field observations and sampling through analysis and reporting results.
      the operation requires standardized procedures and written operations
      directions and reporting forms. Supervision and built-in controls are
      A.   Handling of Samples in the Laboratory. T A. Granovsky 
      B.   Operations Manuals and Laboratory Records. T. A. Granovsky, and K.
          L. Harris                                              
VIII.  Application and interpretation of Results                  
       In assessing losses, it is important to plan and follow a system that will
       produce the information required, be it related to traditional patterns,
       proposed interventions, biological parameters, or loss/benefit values.
       A.   The Chronologic Approach: Losses as Reflected by Use Patterns. J. M.
           There is a need to assess losses in grain as they are related to the use
           pattern so as not to base total loss figures on the final condition of
           residual grain.
       B.   Losses and the Economist. M. Greeley and G. W. Harman 
           To the economist, "losses" refer to changes in value, and the magnitude
           of the effort to reduce losses is often dependent on the magnitude of the
           monetary losses. Loss surveys are viewed from this perspective.
       C.   Conversion Into Monetary Values. E. Reusse
           After physical and quantitative assessment, food losses need to be
           expressed in monetary terms. This is necessary to establish a common
           denominator for cost/benefit analysis in which cost (investments in
           potential improvement measures) and benefits (expected reduction of
           food losses) can be weighed against one another.
     A.   Sampling Grain
         1.   Comments on Probing Techniques and Probes
         2.   Techniques for Sampling Bagged Produce. P. Golob
             Examining every grain in a lot is not physically possible. Thus,
             the quality of the whole has to be judged on the basis of a sample.
             The sample must be representative of the individual bag, stack, or
             lot from which it is drawn. Various techniques to obtain
             representative samples from bagged commodities are described
             and discussed. Emphasis is given to problems of probing for
     B.   Tables of Random Numbers and Their Use. B. Drew and T.
         Sample selection by means of randomization is not an unorganized hit
         or miss process to assure that an intentional or unintentional bias will
         not be introduced during sample selection and sampling. Procedures
         for meeting these requirements are discussed and described. A table of
         random numbers is given.
     C.   Moisture Meters
         A review to help the prospective buyer find which of the many meters
         best meets the work requirements. Data sheets are given.
         1.   Guidance in the Selection of Moisture Meters for Durable
             Agricultural Produce. T. N. Okwelogu
             List of meters and characteristics.
         2.   Table of U.S. Department of Agriculture, Federal Grain
             Inspection Service List of Meters Used in the United States and
             Their Manufacturers, April 1978
         3.   French Table of More Recent Moisture Meters with Acceptable

pglxapx0.gif (600x600)

     D.  Assessment of Profitability of Alternative Farm-Level Storages. M.
         An approach is given to evaluating three methods of storage
         improvement for Andhra Pradesh, India. In each case, a cost/benefit
         ratio is determined and compared.
Selected References
                           CONTRIBUTORS AND AUTHORS
  The scope and format of this manual stem from the Technical Advisory Committee
of the American Association of Cereal Chemists and from two meetings. One was held
Sept. 8-10, 1976, at Harpers Ferry, WV. The other was held June 13-24, 1977, at the
Tropical Stored Products Centre, Slough, England. Those present at these meetings are
as much contributors as are those who eventually wrote the individual sections.
  The 1976 meeting was a wide-ranging brainstorming session covering the basic concept
of the manual and getting to the fundamentals of feasibility, format, and scope. It
was a group effort and the benefits stemming from its interdisciplinary makeup cannot
be overemphasized.
  The 1977 meeting was a technical workshop devoted to defining and clarifying
general goals and specific subjects and writing them down. It functioned both as a
group effort and as a vehicle for individual contributions.
  The American Association of Cereal Chemists Committee consisted of Edith A.
Christensen. U.S. Department of Agriculture, Inspection Division, FGIS, Washington,
DC 20250; John H. Nelson. (now) American Maize Products Company, Hammond,
IN 46336; and Raymond J. Tarleton. American Association of Cereal Chemists, 3340
Pilot Knob Road, St. Paul, MN 55121.
  The consulting-editing relationship with Hugh J. Roberts of L.I.F.E. and with Peter
Tyler, Tropical Stored Products Centre, warrant special mention.
  Credit is given to the El Salvador Centro Nacional de Technologie Agropecuaria
(CENTA) for providing field and laboratory assistance in evaluating portions of this
  Participants at the two meetings and authors are given in the lists that follow.
                             PARTICIPANTS AT THE
                           POSTHARVEST GRAIN LOSSES
                               METHODS WORKSHOP
                               Harpers Ferry, WV
                             September 8-10, 1976
J. Mervyn Adams. (now) The Wellcome Foundation, Ravens Lane, Berkhamsted,
     Herts., England
Keith Byergo. Crop Protection, Office of Agriculture, Bureau of Technical Assistance,
     Agency for International Development, Washington, DC 20523
Howard R. Cottam. Consultant, 2245 46th St. N.W., Washington DC 20007
M. G. C. McDonald Dow. Board of Science and Technology for International Development,
     National Academy of Sciences, 2101 Constitution Ave., Washington,
     DC 20418
Maryanne Dulansey. Consultants in Development, 298 West 11th St., New York, NY
Kenton L. Harris. Consultant, 7504 Marbury Road, Bethesda, MD 20034
William J. Hoover. American Institute of Baking, Box 1448, Manhattan, KS 66502
Henry Kaufmann. Cargill, Inc., Box 9300, Minneapolis, MN 55440
Carl Lindblad. Consultant, 1706 Euclid St. N.W., Washington, DC 20009
Floyd E. O'Quinn. 7328 Range Road, Alexandria, VA 22306
Priscilla Reining. International Office, American Association for the Advancement of
     Science, 1515 Massachusetts Ave. N.W., Washington, DC 20005
Hugh J. Roberts. League for International Food Education, 1126 16th St. N.W.,
     Washington, DC 20036
                              PARTICIPANTS AT THE
                                SLOUGH, ENGLAND
                            WORKSHOP ON POSTHARVEST
                            GRAIN LOSS METHODOLOGY
                               June 13-24, 1977
J. Mervyn Adams. (now) The Wellcome Foundation, Ravens Lane, Berkhamsted,
     Herts., England
Bill Andrews. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
Andy Baker. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
Robin Boxall. Institute of Development Studies, University of Sussex, Brighton, Sussex,
Geoffrey G. Corbett. FAO, Via delle Terme di Caracalla, 00100 Rome, Italy
David Dendy. TPI, Industrial Development Department, Culham, Abingdon, Oxon,
Jacques Deuse. IRAT, B.P. 5035, Montpellier, France
Bruce Drew. Pillsbury Company, 311 2nd St. S.E., Minneapolis, MN 55414
David Drummond. Ministry of Agriculture, Fisheries and Food, Pest Infestation Control
     Laboratory, Tolworth, Surrey, England
Rennie Friendship. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
Peter Golob. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
Martin Greeley. Institute of Development Studies, University of Sussex, Brighton,
     Sussex, England
Geoffrey Harman. TPI, 56/62 Gray's Inn Road, London WCIX 81U, England
Kenton L. Harris. AACC/L.I.F.E., 7504 Marbury Road, Bethesda, MD 20034
Noel Jones. TPI, 56/62 Gray's Inn Road, London WCIX 81U, England
Carl Lindblad. AACC/L.I.F.E., 1706 Euclid St. N.W., Washington, DC 20009
Matthias Von Oppen. ICRISAT, Hyderabad, India
Elizabeth Orr. TPI, 56/62 Gray's Inn Road, London WCIX 81U, England
Harry Pfost. Department of Grain Science and Industries, Kansas State University,
     Manhattan, KS 66506
Peter F. Prevett. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
Barbara Purvis. ESHH, FAO, Via delle Terme di Caracalla, 00100 Rome, Italy
Eberhard Reusse. FAO, Via delle Terme di Caracalla, 00100 Rome, Italy
Robert A. Saul. 1412 Martin Road, Albert Lea, MN 56007
Gerard G. M. Schulten. Royal Tropical Institute, 63 Mauritskade, Amsterdam-Oost,
Harlan Shuyler. FAO, Via delle Terme di Caracalla, 00100 Rome, Italy
Philip Spensley. TPI, 56/62 Gray's Inn Road, London WCIX 81U, England
Malcolm Thain. TPI, 56/62 Gray's Inn Road, London WCIX 81U, England
Peter Tyler. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
David Webley. TPI (TSPC), London Road, Slough, Berks, England SL3 7HL
J. Mervyn Adams. (now) The Wellcome Foundation, Ravens Lane, Berkhamsted,
     Herts., England
Geoffrey G. Corbett. FAO, Via delle Terme di Caracalla, 00100 Rome, Italy
David Dendy. TPI, Industrial Development Department, Culham, Abingdon, Oxon,
Bruce A. Drew. The Pillsbury Company, 311 2nd St. S.E., Minneapolis, MN 55414
P. Golob. TPI (TSPC), London Road, Slough, Berks, England, SL3 7HL
Theodore A. Granovsky. Department of Entomology, Texas A & M University, College
     Station, TX 77843
John H. Greaves. Pest Infestation Control Laboratory, Tolworth Surbiton, Surrey,
Martin Greeley. Institute of Development Studies, University of Sussex, Brighton,
     Sussex, England
Allan Griff. 5324 Wakefield Road, Bethesda, MD 20016
Geoffrey W. Harman. TPI, 56/62 Gray's Inn Road, London WCIX 81U, England
Kenton L. Harris. 7504 Marbury Road, Bethesda, MD 20034
William J. Hoover. American Institute of Baking, Box 1448, Manhattan, KS 66502
William B. Jackson. Bowling Green State University, Bowling Green, OH 43403
Henry Kaufmann. Cargill, Inc., Box 9300, Minneapolis, MN 55440
Carl J. Lindblad. 1706 Euclid St. N.W., Washington, DC 20009
Guy Martin. I.T.C.F. Cereal Laboratory, 46 rue de la Cleff, 75005 Paris, France
Jean-louis Multon, Institut National de la Recherche Agronomique, 44072, Nantes
     Cedex, France
T. N. Okwelogu. Produce Inspection Headquarters, PMB 1012, Enugu, Anambra
     State, Nigeria
Conrad C. Reining. Department of Anthropology, The Catholic University, 620 Michigan
     Ave. N.E., Washington, DC 20011
Eberhard Reusse. FAO, Via delle Terme di Caracalla, 00100 Rome, Italy
Robert A. Saul. 1412 Martin Road, Albert Lea, MN 56007
Gerard G. M. Schulten. Royal Tropical Institute, 63 Mauritskade, Amsterdam-Oost,
Manfred Temme. Environmental Studies Center, Bowling Green State University,
     Bowling Green, OH 43403
                            POSTHARVEST GRAIN LOSS
                              ASSESSMENT METHODS
  When world food is viewed in terms of a system of production, distribution,
and utilization, it becomes obvious that in our attempts to improve the system
we have allocated most of our resources to the production component. Distribution
and utilization have been comparatively neglected. But hunger and
malnutrition can exist in spite of adequate food production. They can be the
result of unequal distribution of food among nations, within nations, within
communities, and even within families. Loss and deterioration of available
food resources further adds to the problem. Hence, maximum utilization of
available food is absolutely essential.
  Of the agricultural commodities consumed as food, grains (cereals, legumes,
oilseeds) contribute the bulk of the world's calories and protein. The food
grains system is depicted in Fig. 1, which shows the many points at which

pgl1x2.gif (600x600)

losses of food occur. The reduction of postharvest grain losses, especially
those caused by insects, microorganisms, rodents, and birds, can increase
available food supplies, particularly in less developed countries where the
losses may be largest and the need is greatest.
  In September 1975, the growing international awareness of the need for
reducing postharvest food losses culminated in a resolution of the Seventh
Special Session of the United Nations General Assembly stating that "the
further reduction of post-harvest food losses in developing countries should be
undertaken as a matter of priority with a view to reaching at least 50% reduction
by 1985." Yet, following the Seventh Special Session, an Interdepartmental
Subcommittee reviewed past and current activity and concluded: "There is
no agreed methodology of post-harvest loss assessment. Moreover, loss data
are generally unrelated to the cost of loss reduction. "
  In its interpretation of available information on losses, the Subcommittee
concluded that "there can be no agreed single figure for the percentage of
post-harvest losses on a global scale or even on a national basis. There is
clearly a need for more accurate assessment of these losses, to establish firm
justification for the development and introduction of measures designed to
reduce them where the cost/benefit ratios of corrective measures are favorable."
  The goal of this volume is to provide postharvest grain loss assessment
methods yielding standardized and reproducible results so that effective grain
loss reduction efforts can be undertaken in developing countries. The assessment
information from such a manual may provide essential justification and
motivation for introducing measures designed to reduce grain losses.
  This volume is prepared in large part for use by policymakers who need loss
information both in determining national priorities and requirements and in
bringing their efforts to bear on the small farmer and other small-volume grain
handlers. It is also directed to the individual investigator who seeks a basic
guide in his specific investigations. The manual is aimed primarily at loss
assessment in developing countries.
  Although a methodology for assessing postharvest grain losses will not in
and of itself reduce those losses, the methodology is essential to postharvest
operational programs so that priorities for loss reduction can be determined.
In addition to serving as a much-needed assessment tool, the methodology and
other activities proposed can serve as a means to persuade all concerned that
change is necessary and that effective techniques for reducing losses are available.
Even financial constraints can disappear when priorities are reordered.
  As detailed later in this Preface and in Chapter II, the enormous variability
of local postharvest situations dictates that no complete or definitive loss
assessment methodology for all situations is now possible. Thus, this edition is
not proposed as a final and absolute piece of work. For example, there exists
very little experience which can be drawn from in loss assessment of cereal
grains such as sorghum, millet, teff, and major oilseeds. Judgment will be
required to adapt known assessment methods to those grains and to other loss
situations not dealt with in sufficient depth here. Further, the editors realize
that expansion and refinement of the loss assessment techniques presented in
this manual are desirable and necessary as a continuing process.
  Increasing food production by increasing acreage or yield per acre has been
a readily applied concept while reducing losses to increase food supplies was a
less obvious strategy. This occurred in spite of the availability of a considerable
body of information on postharvest grain losses, and in spite of several
decades of research and development on losses and their control.
  Progress in reducing postharvest food losses requires the identification and
elimination of the constraints to the application of existing technology. The
major constraint may be a lack of finances, but it is equally possible that lack
of knowledge and of trained personnel, as well as political and cultural constraints,
exists. In 1975 an FAO Subcommittee position paper identified four
constraints to the effective use of available technology for reducing on-farm
losses: 1) lack of arrangements for producing the necessary inputs, 2) inadequate
distribution channels for the necessary inputs, 3) lack of purchasing
power or credit arrangements for the farmer to buy the inputs, and 4) inadequate
information to the farmer on how to use the inputs.
  While calling for integrated country programs to address these constraints,
the Subcommittee stressed the need for creating "an awareness throughout
national extension services that on-farm losses are serious and can be significantly
reduced." Postharvest loss reduction intervention must be made, however,
with specific techniques applied to reduce specific losses. While there
may be broad sweeping national needs, not only are the techniques specific,
but they must be applied at specific intervention points. Until data are available
to show the potential gain from the elimination of losses amenable to
reduction, motivation to reduce those losses will not be strong. But aggregate
data reflecting losses on a global or even on a national basis are not really
useful even if it were possible to obtain them. They are singularly unpersuasive
to a farmer, trader, or warehouseman who must lay out his money and time.
  Losses vary by crop, variety, year, pest and pest combination, length of
storage, methods of threshing, drying, handling, storage, processing, transportation
and distribution, rate of consumption, and according to both the
climate and the culture in which the food is produced and consumed. Given
such enormous variability, it is not surprising that reliable statistics regarding
the type, location, causes, and magnitude of postharvest grain losses are not
available. Yet reliable and objective methods for generating them are needed if
priorities are to be given to the reduction of losses. This is needed in regional
and national planning and in motivating those organizations which may fund
loss-reduction programs, and on down to the local level.
  Meanwhile, it is prohibitively expensive and unjustifiable to mount countrywide
assessment studies of losses in the whole postharvest system. As detailed
in Chapter II, an expert judgment is needed to identify the most serious grain
loss points in a country's postharvest food supply system in order to mount
in-depth assessment efforts at those high loss points.
  Stated another way, changes will not be widely accepted until and unless
they are practical for and clearly benefit the individual who is to make the
change. Although losses and savings are far from the only elements which
must be considered in loss reduction efforts, reliable figures can go a long way
in convincing those dealing with grain, and certainly for motivating those
organizations which may fund the loss reduction programs.
  Extent of loss is important, but not all-important. Other factors should be
considered in deciding on the nature of interventions, or whether to intervene
at all: The value of the grain in economic lines; the fact that there will be social
change effected by intervention programs; competition or conflict, or both,
with other national priorities; effect on price stability and similar economic
considerations; the relationship and possible conflict of economic factors that
affect the consumer, grain grower, grain trader, and national balance of payments
mean that interventions need to be subjected to an integrated, multidisciplinary
evaluation and actually field tested within the social and economic
structure before they should be implemented on a broad scale.
  Both "guesstimates"(1) by knowledgeable people and estimates without factual
basis, particularly by people with vested interests, have had a useful role in
the past, will continue to be used in the future, and are especially useful when
timely opinions are needed as to where the more serious losses occur. In using
guesstimates to justify cost/benefit comparisons or to reshape established
practices, however, one needs to recognize the possible bias of the estimator:
Was it put in perspective by a thorough gleaning of the information, was the
judgment based on an in-depth and long-standing knowledge of local or even
country-wide conditions, was it made to reveal some situations and cover
others? It is critical to understand that guesstimates are the type of estimations
that requires the most expert judgment.
  If large area or national survey figures are taken without sufficient regard
for variations in the individual components, these figures may not be useful to
locate specific intervention points.
  Finally, we might ask why, in the face of a need for accurate figures that has
not gone unnoticed over at least two decades, have there been so many postharvest
loss estimates made with obvious biases, and why has a methodology
not been forthcoming from the scientific community?
  As stated above, the guesstimates have served a useful purpose. They have
also been accepted by those seeking national resources and changes as well as
by those allocating international resources. Although the scientific need was
there, the political- and transformation-related requirements did not call for
scientifically derived figures. Now, with increased sophistication and increasingly
limited resources requiring benefit-related priorities, there is a need to
know what the postharvest losses really are. Without such information, it is
impossible to assess needs or to calculate improvements. However, there has
been another factor that has stood in the way of assembling this manual. It
needs to be mentioned, for its recognition is the key to the present status and
ultimate fate of this volume. This factor has been the simple absence of anyone
to do the job.
  Within the L.I.F.E. consortium, the American Association of Cereal Chemists,
under a contract from the Office of Nutrition, Technical Assistance Bureau,
U.S. Agency for International Development, has broken the impasse on
   (1)This term is used to connote estimates with some facts by knowledgeable people.
how and by whom the job was to be done, and it has developed and printed
this volume with the hope that it is a volume to be evaluated, tested, and
improved by actual use in the field. We look forward to the inevitable changes.
Kenton L. Harris
Carl J. Lindblad
August 1978
                               I. INTRODUCTION
                         K. L. Harris and C. J. Lindblad
  This volume is directed mainly to grain loss situations in developing countries.
  Determination of losses to food crops requires careful blending of the concepts
and procedures of several sciences while each is given its necessarily
detailed attention. Nowhere is this more true than in dealing with postharvest
losses to grain. Information gathering ranges from A to Z, and at the outset
emphasis needs to be given to the cultural-social aspects discussed in Chapter
  While many of the methods contained in the manual relate to the evaluation
of damage caused by a single organism or mechanical effect, such selective
attacks rarely occur in nature. Interactions between major causes of losses
must be expected.
  A basic concept of this manual is that it be applied in its entirety. Care needs
to be taken that personal, national, economic, cultural, and other biases do
not generate unwarranted project plans or conclusions. To illustrate, large
influential farmers may want technologies developed to suit their own needs
which may be completely inappropriate for small farmers whose grainhandling
systems are less mechanized or capital intensive, grain storage scientists
may want to continue in their own research area to the exclusion of other
equally important areas, national governments may favor one political region
or group over another, or international development agencies may have their
own priorities.
  There are many ways to produce a list of intervention points. Consideration
could be given to technological improvements that would both cost the least
and prevent the greatest amount of grain losses to the benefit of the entire
country as a whole. However, political, economic, and social priorities need to
be taken into account in locating and identifying intervention points. What is
technologically ideal may be very different from what is practical and feasible
within the actual social, economic, and political environment. A balancing of
technical and social sciences is essential in assessing and reducing grain losses.
  For the purposes of identifying loss points which are critical and amenable
to reduction, this manual uses the pipeline concept to describe the location and
flow of grains. In this way, losses can be viewed individually and in perspective;
however, the pipeline concept is not limited to technical or physical
factors. Social realities come into play and perspective is required to both
understand those attendant social influences and to prevent them from being
blindly introduced as unrecognized bias. The pipeline approach weighs individual
loss points in relative magnitude. Combined with consideration of social
realities which influence amenability to in-depth assessment and loss reduction,
the pipeline concept serves to 1) identify critical loss points for in-depth
assessment and 2) provide a basis for development of improved technologies
for postharvest loss reduction.
  The influence of personal judgment, and therefore bias, cannot be avoided
though the investigator or official may be unaware of its role. The investigator
must also constantly guard against yielding to pressures based on unsubstantiated
assumptions. An example of the consequences of this kind of oversight
is seen in the countless huge, empty, and decaying grain bins installed across
the developing world under incorrect assumptions. They serve to demonstrate
that what is feasible in one situation will not necessarily be successful in
  The compilers of this manual have operated under the well-reasoned opinion
based on some practical experience that interventions to reduce grain loss
are often best channeled to the farmer/producer. There are a number of
reasons for this alignment. One technical reason is that the best form of loss
reduction is early prevention -- grain which is in good condition will deteriorate
more slowly than grain which is, for example, already infested with insects
or poorly dried. Following that logic, to assure good quality food grain
throughout the pipeline, it seems practical and desirable to have it enter the
pipeline under optimal harvesting, drying, and storage conditions. Another
factor is that, in developing countries, much of the grain is stored and consumed
in the rural areas, in large part by farm families.
  A loss assessment study that does not have built into it the strong possibility
and intention of benefiting the situation under study is of no consequence. The
purpose of loss assessment is effective and expeditious loss reduction. Loss
assessment need not and should not be a largely academic exercise.
  Loss-causing damage may not divide into neat, exclusive categories. Moldy
kernels may be insect infested and vice versa. Insects can cause shattering, and
shattered kernels more readily support certain insects. Bits and pieces lost
through holes in bags or in processing may have been produced by too rapid
drying. These and other situations are more the normal than the exception and
need to be duly noted and judgment applied in interpreting data.
  Certain concepts are dealt with in only one section of the manual though
they have applications throughout many facets of loss assessment and reduction.
For example, while the subject of economics is in a separate section, it
has applications throughout the manual. It bears on sampling and how, when,
and where the samples are taken. It bears on the selection of study situations
and how they impinge on each other, and it relates to cultural factors. Similarly,
cultural factors are dealt with in a separate section though their implications
are also pervasive as they bear on sampling, analyses, and the whole
problem of functioning in a system without undesirably changing or destroying
  Early in the preparation of this first edition, an attempt was made to prepare
a manual that could be used by trained and untrained workers alike. This
proved to be impossible. The ideal of writing for those without any background
in grain storage, biology-entomology, food marketing, or the socio-economic
sciences was attempted and abandoned as impractical. The material
is, therefore, prepared for people with at least some pertinent experiential or
academic background.
  One of the important matters not covered in this manual is the matter of
mold toxins. This does not downgrade the seriousness of the mycotoxin problem.
Important as the problem is, this volume is concerned with measuring
losses of stomach-filling grain, not whether its nutritional value has been
reduced. While strongly noting that food contaminated with mold toxins is to
be avoided, as regards mold-caused losses, this manual deals only with such
losses of grain actually discarded for human food because of the presence of
                            II. TERMS OF REFERENCE
                                A. Definitions
                        K. L. Harris and C.J. Lindblad
  This manual deals with food grains, cereals, and pulses and the word
"grain" is broadly used to include all of these. It deals exclusively with the loss
of food from the food chain and largely follows the definitions of Bourne (1).
In it, a working definition of the term "postharvest food loss" is set forth as
given below:
  "POST HARVEST" means after separation from the medium and site
  of immediate growth or production of the food.
    Post harvest begins when the process of collecting or separating food
  of edible quality from its site of immediate production has been completed.
  The food need not be removed any great distance from the harvest
  site, but it must be separated from the medium that produced it by a
  deliberate human act with the intention of starting it on its way to the
  It does not include steps between cooking and eating as covered by Bourne
and agrees with Bourne to "not cover inefficiencies in human metabolism and
utilization of the food." In this manual, however, the pathway ends when the
food grain or the food prepared from the grain, or both, reaches the point
where it is to be finally prepared (cooked) for consumption.
  Three periods of time may be identified during which food may be lost,
  and each period has its characteristic problems, and means of overcoming
  these problems.
    a. Preharvest are losses that occur before the process of harvesting
  begins, for example, losses in a growing crop due to insects, weeds and
    b. Harvest losses occur between the onset and completion of the process
  of harvesting, for example, losses due to shattering during harvest of
    c. Post harvest losses occur between the completion of harvest and the
  moment of human consumption.
  Postharvest intermixes in varying degrees with portions of the maturing-drying-processing
period and often no sharp distinction can be made. Thus,
maize held in the field for drying is also maize held for storage and use. This
manual does not imply that any artificial sharp distinction must be made.
  Harvest and post harvest losses are sometimes combined into a single loss
  because there are some elements of common concern between them. A
  suitable descriptive term for these combined activities would be "post
  production losses". The following schematic representation shows the
  relationship among the various types of food losses:
    1. Preharvest
    2. Harvest        } Post Production
    3. Post Harvest  }
  In addition to Bourne's postharvest grain, this manual includes the ripe crop
remaining in the field, whether standing in its original position or not, for
further drying or holding, or both, until it is brought in or removed from the
growing position, eg, maize drying/storage in much of Latin America.
  "FOOD" means weight of wholesome edible material that would normally
  be consumed by humans, measured on a moisture-free basis.
    Inedible portions such as hulls, stalks, [and] leaves . . . are not
  food. . . . Feed (intended for consumption by animals) is not food [unless
  specifically of interest to the individual assessment exercise].
    The method of measuring the quantity of food in the post harvest
  chain should be on the basis of weight expressed on a moisture-free basis.
  There will be times when information on losses in nutritional units and
  economic losses will also be needed but these should not be the prime
  means of measuring post harvest food losses.
  "GRAIN LOSS," as used in this manual, concerns the loss in weight of
food that would have been eaten had it remained in the food pipeline.
  "LOSS" means any change in the availability, edibility, wholesomeness
  or quality of the food that prevents it from being consumed by people.
    Food losses may be direct or indirect. A direct loss is disappearance of
  food by spillage, or consumption by [insects], rodents, [and] birds. An
  indirect loss is the lowering of quality to the point where people refuse to
  eat it.
    This definition is a people-centered definition. "Food" means those
  commodities that people normally eat and excludes the commodities that
  people do not normally eat. If the food is consumed by people it is not
  lost; if it is not consumed by people for any reason at all then it is
  considered a post harvest food loss.
  Food losses are, at times, simply as they are locally defined or as they locally
occur. For example, grain which is discarded because of discoloration is a loss.
  Processing losses occur when edible portions of food are removed from
food channels by the process or by spillage or breakage from the process. Rice
hulls are inedible. Their removal does not constitute a loss. Rice pieces
diverted from the food-chain are a loss. Rice bran is edible to some, inedible to
others. The handling of each similar situation needs to be clearly defined as it
occurs. Corn cobs or cores are not a loss. The corn seedcoat is removed in
making corn grits. It is not removed in making many other foods. How it is
handled needs to be defined in each appropriate instance.
  Where quality deterioration results in a loss in weight or in the food not
being eaten at all, eg, rejected in the marketplace, the rejected food is a loss. In
this volume, quality is a consideration only as it relates to loss in weight of
food, but how it is handled needs to be defined appropriately in each instance.
  The term "insects" includes true insects (six-legged arthopods) and grain-damaging
  Microbiological losses and microbial losses are used interchangeably to refer
to losses caused by molds, yeasts, and bacteria.
                           Literature Cited
1. BOURNE, M. C. Post harvest food losses -- the neglected dimension in increasing the world
     food supply. Cornell International Agriculture Mimeograph 53 (1977).
                                  CHAPTER II
            B. Planning: An Overview for Project Administrators
                                 K. L. Harris
  Determining agricultural losses involves many disciplines and goes to the
heart of established cultural patterns. Administrators need to recognize the
complexities of what they have to deal with and understand that unless defects
in planning and implementation are overcome, the results will be jeopardized.
While this is an obvious platitude, it is of special importance here since the
nature and quality of the operation can set the stage for the nature and quality
of other programs that may follow in the technical and lay community.
  Without attempting to set forth an administrative manual, the following
details are to be noted:
  1. Project planning, depending on circumstances, may require inputs from,
for example, agricultural economics, agricultural engineering, agricultural extension,
administration, anthropology, biology-zoology, cultivators/grain
owners, education, entomology, food marketing, grain storage science, microbiology,
political science, rural sociology, and statistics.
  2. Revealing the status of the food grain supply may be a delicate matter
that impinges on matters of national and international security, as well as on
local, national, and international commodity markets and on foreign exchange
  3. One needs to be aware of social factors; special village allegiances and
requirements; the role of women, the family, and other groups; and whether
information is best collected by lower-status field-workers, peers, higher-ranking
individuals, etc.
  4. Logistic requirements are imposed by terrain, delineated and undelineated
boundaries; presence or absence of containers, scales, meters, transport;
local customs and work patterns; and training requirements and capabilities.
  5. Assessment work needs to be understood in terms of cultural factors:
local vames and definitions and local social and agricultural systems.
  6. The assessment must relate to local needs -- individual, national, and all
  7. One should be aware of the interrelations between postharvest losses and
  Basic survey operations, schedules, and plans are set forth in Table I and
Fig. 2. The time needed for such a survey will obviously depend on the size of

pgl2x17.gif (600x600)

the country and accessibility of the sampling areas, but the decision on the
selection of farmers must take place before any final work begins so that
sampling visits can start immediately after harvest or any other start-up time.
Modifications to the sampling pattern may be made in the case of crop failures
or similar unavoidable circumstances.
  The nature of the operation -- and planning for the operation -- will
depend primarily on the factors that are to be investigated and how they are to
be investigated. This is the subject of this manual.
  This manual deals with 1) obtaining a planning overview of grain movements,
the grain pipeline, 2) determining what portions of the pipeline should
                                   TABLE I
                           Basic Plan of Operation
   Stage        Weeks                 Activity                      Personnel(a)
Preharvest     6          1  Familiarization with local           CO
                            agricultural structure and
               2          2  Preliminary survey for choice        CO, ES
                            of sampling areas
               2          3  Fact-finding visit to chosen         CO, ES
                            sampling areas for information
                            on storage practices to identify
                            strata and select appropriate
                            method of obtaining farmers
Harvest(b)     up to 4   4  (If required, construction of        CO   laborers
                            experimental stores)
               2          5  Initial visit to selected farmers    CO, Exp, ES,
                            to obtain basic information and      LA
                            baseline samples (also purchase
                            grain for experimental silos)
Postharvest(b) 1-3       6  Examination of baseline samples      Exp, LA
                            in laboratory and check on
                            proposed methodology
               1 per     7   Monthly sampling visits to selected   LA, ES
               month        farmers to collect samples and
                            record consumption patterns
               1 per     8   Laboratory examination of field       LA
               month        samples (and experimental samples)
               7          9  (If required, brief questionnaire    CO, ES
                            survey of other farmers to confirm
                            storage pattern)
               2         10  End-of-season visit to selected      CO, Exp, ES
                            farmers to check consumption and
                            thank for cooperation
  Harvest       4        11  Analysis of results in terms of      Exp
                            loss per sample and integration
                            with consumption pattern
                        12  Preparation of report                 CO, Exp
(a)CO = Country project officer; ES = extension staff; Exp = expert TSPC;
   and LA = laboratory assistant.
(b)Drying, processing, bulking, etc.
Adapted from: Tropical Products Institute, Tropical Stored Products Centre,
Slough, England.
be further investigated both because of the size and nature of the losses and
their feasibility for reduction, and 3) conducting the detailed investigations.
  This manual also stresses the use of existing in-country data on what grains
are produced in what quantities in what regions and consumption patterns.
                                  CHAPTER II
  C. An Overview of the Postharvest System: The Food Grain[\N
Supply Pipeline (Determining the Interrelationship and Relative
                         Magnitude of Losses)
            K. L. Harris, W. J. Hoover, C. J. Lindblad, and H. Pfost
  The flow of grain from its sources, ie, the farm field or import docks, to the
eventual consumer is depicted for the purposes of this manual as a pipeline
with many possible interconnecting pipes and reservoirs. Losses, or leaks, can
occur along the entire pipeline -- during harvesting, drying, transport, storage,
and processing. As presented in the Preface, the purpose of viewing the
food grain supply system as a pipeline is to assign individual loss points (eg,
on-farm losses) relative importance in terms of loss in other parts of the grain
pipeline (eg, transport or warehousing losses). This relative perspective is necessary
to see the importance of the total amount of grain actually lost in any
given point as opposed to the percentage of grain lost which passes through
that point. Failure to obtain such a perspective has resulted in overly high and
low loss figures arrived at by extrapolating from observed losses at specific loss
points without putting those losses into the perspective of the grain moving
through the total system.
  This failure and the need to obtain an overview often apply to expatriates
and others entering a system for the first time.
  One needs to use all possible local information to determine how and when
the grain moves from harvest to consumer, routes for movement and holding
patterns, and where and how processing is accomplished. Most of this information
is known locally.
  Grain does not move in a straight line and uniform sequence from producer
to consumer. Harvested grain can be specially dried and otherwise treated to
go into special household use; some into an even more special seed-grain
storage. This grain may remain there or move out for food or trade under
special conditions influenced by factors such as family, weather, or government.
It may even be replaced by other local or imported grains. A portion of
the harvest may be held for short-term storage, a part for long-term storage,
and the rest sold or otherwise traded off the farm.
  All of these factors, and more, need to be kept in mind in determining where
and what should be tested.
  Delineation of the test sites involves looking closely at general loss situations
and careful on-site evaluations of specific individual sites. Selection of "amenable"
sites (villages, cultivators, markets, transit systems, warehouses) requires
incorporation of many factors. Accessibility must be balanced against
the location being atypical because of proximity to outside influences. Traditionalism
must be balanced against the need for outsiders to be accepted into
the delineated area. Language can be a key barrier, and an absence of direct or
completely competent and trusted lines of communication is unacceptable for
loss survey teams. Sex roles must be considered as to who really does the
harvesting, threshing/cleaning, storing, and marketing of the grain. All parameters
need to be considered, and should cover the entire social, cultural,
physical, commercial, and political setting.
  Even the simple village market has flowing through it all these effects, and
more, so that if there were to be a single measurement it would, in reality,
consist of measurements of many factors, each weighted as to volume.
  Knowledge of actual high-loss and low-loss situations is required in determining
the need for, location of, and types of interventions. However, inordinately
high- and low-loss situations must be put into perspective rather than
giving them overemphasis as has been the case in some instances.
  To further illustrate, out-of-condition grain held by market speculators may
suffer very high losses, say 30%. Taken by itself, this level of loss might
identify grain speculators as a critical focus for improved storage technology
intervention. However, if in fact only 5% of the total grain supply is ever
handled by such speculators who specialize in out-of-condition grain, the real
value of the total losses at this speculator level becomes 30 x 5%, or 1.5%
rather than 30% of the total grain supply.
  A useful investigation of postharvest grain losses requires detailed knowledge
of the entire postharvest food grain supply pipeline. Figures 3 and 4 are

pgl3x200.gif (600x600)

two representations of supply pipelines. Figure 3 emphasizes marketing patterns;

pgl3x20.gif (600x600)

Fig. 4 emphasizes the processing flow through to the consumer. At any

pgl4x21.gif (600x600)

one point, grain or grain products may move out of one pipeline, around
several intervening steps, and re-enter further along in the sequence. Similarly,
movement occurs in both directions. Grain gleaned from the field or from
spillage on a farm or in a rural market can go immediately to a consumer or
may be bartered back into a trade channel. What might be loss to a farmer by
spillage at a local market, or to a transport company, may in reality be a mode
of payment for services rendered at an otherwise unacceptably low pay scale.
  In each country, district, or community area, there exists a marketing system
for food grains. It is imperative that the flow of grain through the various
facets of this marketing system be quantified so as to establish priority points
for observance and measurement of losses, and to subsequently focus attention
on loss prevention programs. Figure 5 shows a quantified flow in which

pgl5x22.gif (600x600)

different grains and oilseeds follow different routes.
  Moreover, different parts of the pipeline have different flow rates. While a
particular grain may be in a storage chamber for some time, it may be in a
milling process for a very short time. The types of losses at those two locations
are different; one is a loss which increases over time, and the other is probably
a one-time loss due to such things as poor physical handling, equipment, or
  To follow the pipeline analogy, the two types of losses occur in the reservoirs
and in the pipes. Once grain has passed through a leaky pipe (eg, a poorly
adjusted grinder), it is not subject to that particular loss any longer. However,
grain in a holding reservoir (eg, rodent-infested bin) is subject to those losses
for as long as it remains there. Loss assessment methods and calculations for
the two types of losses can be quite different.
  This, of course, complicates the task of assessing losses. Separate measurements
are required for the different types of losses that occur due to mishandling
or poor equipment settings, in addition to the biological deterioration
caused by insects, rodents, or moisture or other climatic conditions. Sampling,
tracing, eventual utilization, and testing of overall losses really entail making
and evaluating individual, components in a system and calculating their overall
effects. Moreover, since effective loss reduction interventions need to be
directed to the reduction of specific leaks, it is the individual loss figures that
need to be evaluated, not overall national figures.
Note: With the acknowledged limitation of development resources and perhaps
even greater limitation of available, trained personnel, the pipeline concept
is an approach that is recommended as a means of quickly and inexpensively
focusing on significant losses in the overall system. It is also an effective
procedure for effective resource allocation.
  There is every reason to believe that the presence of a survey in the system
will itself affect the system and the results of the survey. This will not be
discussed other than to note that economic, cultural, and political factors
governing the flow and treatment of grain can be expected to respond to the
survey itself, thus partially skewing the results.
                                  CHAPTER II
    D. Preliminary Examination of Specific Problem Points and
                        Making On-Site Rapid Appraisals
         G. G. Corbett, K. L. Harris, H. Kaufmann, and C. J. Lindblad
  Two of the most critical aspects of postharvest grain loss methodology are
the need to not attempt more than is feasible, and to rapidly seek and identify
for investigation major loss situations that seem both amenable to study and
responsive to improvement through practical interventions. By using a pattern
that has found almost universal application by expatriates from international
and national agencies whether dealing with the most primitive situations or the
most sophisticated, this first appraisal has become accepted as a 30-day exercise.
However, 30 days may be too little or too much time, although this will
only be determined by the complexity of the system and the nature of the
questions being asked.
  As with any investigation, some early judgment is made that the work is
needed and that there is a reasonable likelihood that useful results will be
obtained. After that there is a need to work with local officials in a preliminary
fact-finding canvass of the situation that goes into the entire nature of the
grain pipeline, as explained earlier, and then into individual problems and
their projected solutions. This includes all of the grain movement logistics,
personnel, political and cultural ramifications, etc., that will be called into, or
will force themselves into, the final study.
  It would be well for this preliminary canvass to proceed solely as a prelude
to a larger study, but such will not always be the case when immediate developmental
decisions must be made before detailed information can be made available.
  Interrelated aspects will proceed together during the 30- to 60-day preliminaries:
  Assessment exercises may be undertaken by expatriates to determine losses,
while locals seek to determine how to reduce the losses.
  One task is the probing for specific problem points; the other is the job of
making rapid on-site appraisals.
  One looks ahead to a more definitive investigation; the other comes to
on-site loss and intervention judgments within the rapid-assessment time span.
  In one case we are developing a strategy to conduct a survey; in the other,
the survey and loss reduction efforts may be rapidly under way.
Preliminary Examination of Specific Problem Points
  An initial survey is needed to determine what the problem is and what has to
be done. In the initial survey the best possible information available should be
used to ascertain the order of magnitude of the losses in the whole postharvest
system and to identify the major points and causes of losses. As the loss figures
are evaluated and observed to be accurate or inaccurate, they may serve as
data to evaluate the local system. It is important to obtain information from
people who are knowledgeable of the factors being assessed as well as from
voluble proponents of biased or special interest positions. Already available
reliable information, or lack of it, will help to decide the depth and focus of
the preliminary mission.
  The key element is to identify those problem points that can be adequately
isolated, are likely to yield useful information, and are amenable to study and
loss reduction intervention.
  Few locations in the grain pipeline will be neatly packaged, single-entrance,
single-exit, one-measurement situations. It may be necessary to make measurements
over a period of time, to identify the points at which important losses
are occurring, and to make an estimate from the data and evidence available of
the order of magnitude of these losses. After such a survey (which will probably
reveal the need for longer-term assessment of losses), it will be possible to
define immediate, as well as longer-term actions. At the same time, the cost/
benefit implications for both the operators concerned and the country as a
whole must be considered.
  The composition of the 30- to 60-day preliminary investigation mission will
vary according to the complexity of the grain industry and the local information
and expertise available. At least a grain marketing economist and a grain
storage specialist (entomologist-biologist) should be included plus a processing
specialist if it is anticipated that processing losses at village or industrial level
are important.
  Members of the preliminary mission must have experience in the organization
and operation of the grain industry in developing countries. The social
skills acquired by direct experience are invaluable and essential for the judgments
which must be made during the preliminary survey. As experience is so
critical here, interns would usefully be included in the mission; however, large
missions (more than four) are often hard to accommodate within traditional
social structures.
  The mission will:
  1. Map the pipeline using available government statistics and other inputs
from key informants.
  2. Conduct an initial survey of the postharvest grain sysem to establish who
is handling, storing, transporting, and marketing the harvested crop; what
part of the crop is handled and stored by each operator, and for how long,
including farm storage for self-consumption purposes; and the condition of
handling, storing, and processing.
  3. Review all available data on quantitative and qualitative losses occurring
in the system(s) and identify the major causes and extent of loss.
  4. Prepare an inventory of available storage, transport, marketing, and
processing facilities and assess their adequacy in capacity, design, and condition.
  5. Review the present activities being undertaken to reduce postharvest
losses and list the resources available for these activities from both internal and
external sources.
  6. Design a phased action program to investigate or implement under the
project terms of reference.
  In conducting the preliminary study, remember that grain losses occur in
situations that cause or allow them to occur, and as the losses occur, evidence
is left of what has and is happening and what will probably continue to
  There are many clues to both general and individual aspects of grain losses
that can be disclosed by the rapid assessment of a situation. Knowing that key
elements in insect depredations are moisture, temperature, numbers and kinds
of insects, length of storage, storage sanitation, and insecticide use and other
control practices, one can keep the presence or absence of these factors in
mind and come to some general or specific conclusions based on known scientific
principles. Estimates of 30% losses to maize stored for several months
under humid tropical conditions may be quite reasonable. The same figure
when applied to a cold, dry climate or to grain used up in three months may be
  Many farmers are well aware of these factors. Out-of-condition grain is
often passed along to the local market or government agency. Grain for long-term
storage may be dried, put in better storage, or treated wth a protectant.
Loss-prone varieties may be used first or sold off the farm.
  Some conclusions will be fairly straightforward. For example, if grain goes
into bagged and naturally aerated storage that has evolved within the culture,
reasonably good storage quality may occur. If the same high moisture grain
goes into sophisticated silo storage without the necessary sophisticated drying,
there will be a high potential for loss. Poor sanitation, insects, molds, leaking
roofs, rats, uncleaned bins and bags, high atmospheric humidity, and extreme
temperature variations all affect grain losses.
  Generally, when insect damage is very difficult to find, the weight losses due
to insects are also negligible. One may know what a 250/o loss in maize looks
like in one region and carry this mental picture to other regions and other
situations. The significance of frass, of extensive moth webbing, of adult or
larval insects may be so well known that they automatically lock into a fairly
accurate judgment -- a judgment that may well be sufficient for the experienced
person to come to a general conclusion on the extent of the losses
themselves. This, in addition to contributing to a decision on whether a situation
should be tested or surveyed in depth, may be as much as the situation
warrants, especially if the losses are estimated at around 5%. At this low level,
even an in-depth assessment based on currently known sampling procedures
would probably be subject to an error as large as the loss.
  In short, it is possible to do an overall appraisal based on an expert evaluation
of the system with attention to pertinent parts of the harvest-to-consumption
flow or patterns, and to such loss-inducing and loss-reducing
factors as:
   1. Moisture
   2. Temperatures
   3. Insects, rodents, birds (kinds, numbers, association with the grain)
   4. Length of holding
   5. Local quality and quantity controls
   6. Types of bins and other holding vessels
   7. Sanitation-insanitation
   8. Trading quality factors
   9. Use and nonuse of pesticides
  10. Evidence and nonevidence of grain damage; kinds and amounts
      a. Frass and webbing
      b. Exit holes
      c. Darkened (rotten) kernels
      d. Degermed kernels
  11. Mechanical loss factors
  12. Location in the harvest-to-use pattern
  The need to apply the physical loss parameters and to know what stimulates
or retards losses cannot be overemphasized. Many unreasonable guesstimates
would have been avoided if more attention had been paid to such criteria. Of
course, these same criteria will provide an operational arena for in-depth
assessment and loss reduction.
  Finally, one needs to remember that just as losses do not occur in a vacuum,
neither do loss assessments, and one should expect the presence of a survey -- with
or without an overt attempt to make improvements -- to induce changes.
                           III. SOCIAL AND CULTURAL
  The overall aim of this chapter is to introduce some of the complex
cultural-social-anthropological factors to postharvest grain loss assessment/
intervention activities. The message is made up of a variety of signals that pass
in both directions: from the situation being investigated to the investigator and
from the investigator to the situation. It is a dynamic process.
  In grain loss assessments the need is to find out what the situation was or is.
The investigator wants to affect the milieu as little as possible while he assesses
it. Thus he needs to be in tune with what is happening so that the assessment
will be an assessment of what he sets out to assess -- not of what his presence is
bringing about.
  This chapter is a result of many discussions, not only with Allan Griff and
Conrad Reining, but with many others. Griff, Reining, Harris, and Lindblad,
together with Edna Loose and Maryanne Dulansey, examined, analyzed, and
reasoned the subject many times together. What has resulted is the foundation
statement of Part A and the evocative of Part B. Part A is self-explanatory.
Part B is purposely set forth so as to leave the assessor with many questions
into his own investigations.
                         A. The Fact-Gathering Milieu
                                Allan L. Griff
  It seems obvious that planners and field-workers of grain recovery programs
should be familiar with the social and cultural background of the places where
they are working. But far too often this knowledge is insufficient and incorrect,
and the result can be error and waste. Cultural awareness is no guarantee
of success, but it can help.
  This chapter is but a brief outline of how culture operates, and its place in
the early stages of planning a program. It will raise many questions. It may
slow down some projects until adequate understanding of the people is
achieved. It may improve communications enough to get some projects off a
comfortable and self-perpetuating dead-center. But if one is committed to
tangible results rather than just good appearances and completed missions,
culture cannot be ignored -- rather, it must be understood. Culture is on our
side. Few want grain losses, but only a good understanding of the roles of
social and economic behavior of the people involved (ie, the culture) can make
this a contributory factor and not an adversary.
Culture is not Static Tradition
  First, we must erase the stereotype view of culture as stubborn adherence to
tradition and resistance to change. All cultures contain the seeds of change as
well as the inertia to resist change. This is the basis of cultural evolution.
Changes can and must occur for a society to survive, but they must be opposed
and tested to ensure that they achieve their aim, that the gains are worth the
losses, and that change does not occur so fast that the people cannot adapt to
  In this light, we should realize that what we think is good change, or even
what a country's leaders think is good, is never 100% good. There is a price to
pay for all change, and much resistance arises because the price is too high for
some or just cannot be paid without excess hardship, despite apparent longer-term
  Some people in some countries are used to a logical, scientific sequence of
cause and effect and can thus predict the future, more or less. This enables
them to confidently invest time, labor, and money in the future. It gives a
sense of control.
  But in many developing societies, the people have little control and they
know it. Their plans have been thwarted by natural catastrophe, or by magic,
or by the will of forces distant and far more powerful (including both gods and
central governments). Given the crawling pace of development among the
world's rural poor, we cannot blame them for being a little skeptical about
proposed changes. This is not necessarily blind tradition. It may be healthy
and justified caution.
  And stability itself has a positive value in all societies as it reinforces behavior
by promising future returns for today's behavior patterns. Without stability,
people lose the incentive to keep past social values, as future outcome can
no longer be predicted. The result is an explosive proliferation of values (witness
America and Europe today) and a disincentive to plan for the future at
Evolved Versus Imposed Change
  Many cultural changes have been imposed on people, often suddenly, with
remarkable results attesting to the equally remarkable adaptability and resiliency
of people. Conquerors and rebels have imposed languages, religions,
food habits, and codes of law on other people since prehistoric times. They
have often also brought innovations that were eagerly adopted by the local
people, such as the gun and horse among American Indians, and baseball and
hamburgers in Japan.
  On the other extreme, some changes took many generations to evolve, perhaps
because they were not very important or were not enhanced by political
association, or perhaps the price to be paid for the benefits was high. Where
agricultural innovations were concerned, the risk was often simply too great.
Some people lived and still live too precariously to experiment even if the idea
looks promising.
  Development strategists today are caught in the middle. They do not want to
impose, yet cannot wait for evolution to do the job unaided. So we have
derived an intermediate form of "coaxed" change, in which we decide before-hand
what change is desired. People do indeed want to better their lot, but
may be convinced that such efforts are futile and may be too polite or too
scared to tell us so, or may not even realize why they resist. Therefore, it is a
good idea to look at the recent history of the subject community to see how
changes take place in that community.
Study the Past
  Every group has its own ways of change. They usually are those that minimally
disrupt the effective social order, and are also in tune with the popular
trends as evidenced by past change. Thus, both present and past -- in this case,
related to the economic and interpersonal structure of food storage and use -- must
be appreciated to see what might work and what might not. To this end,
the following questions will be useful:
  1. Has the community made technological or agricultural changes in the
recent past? If so, through what channels were the changes introduced? Were
there models to copy? Key people whose support and influence were critical?
Economic or other incentives? Were the changes mainly imposed, coaxed, or
naturally evolved? Are the changes now an irreversible, integral part of the
culture, or are they artificially supported by current leadership and likely to
revert to original status if the support were removed? (The potential permanence
of a change is as much a measure of success as the change itself.)
  2. Have any change attempts failed in the recent past? What were their
histories and apparent reasons for failure?
  With regard to the "who" questions, the models are particularly important
and simply any model will not do. Certain people will be followed, others
rejected, still others ignored. The one who acts first may not be the real leader;
he may be marginal with nothing to lose by trying or he may be acting under
the influence or command of others. The area of influence is also important -- a
man who can command respect and honor among civil servants may not
count for much among the farmers, or an older leader may be resented by the
young, and vice versa. It pays to learn local history to see how things got done
before, and for expatriate workers it is certainly an error to assume one's own
national patterns of power and influence will apply.
  It is also dangerous to believe everything we are told. Observation of attitude
and even tone of voice may be as important as the actual words said.
Double-checking critical statements is essential; relying on one or two data
points is as inadequate in social science as it is in physical science.
How Do You "Learn" a Culture?
  The most obvious answer is time -- implying that people who have spent
years in a group become expert observers of that group. This is not always
true. Of course, time is necessary, but a competent observer must also know
how to observe, must be himself/herself relatively free from familial or political
involvement that might affect observations, and must be articulate enough
to transmit them to others.
  In dealing with local sources of information, all individuals are not equal.
Some are "balloons" -- innovators who are free to change and the first to do
so, and some are "anchors" -- social-role conservatives who provide and
represent stability. Local landowners and similar elites are often in this class,
while their children may well be balloons as with a relatively secure future they
can afford to be different. This balloon-anchor continuum is a convenient way
to characterize local contacts and ultimately to ensure that one's information
does not all come from one type.
  Just as people's responses depend on their individual characters, they also
often depend on how they view their questioners. Association with the local
government or a donor agency may be helpful in some cases and a handicap in
others, and a strong personality may turn a respondent in many directions. As
an agent of change, an investigator must not imagine himself free from bias
either. Attitudes toward development and efficiency are hardly universal. But
he can try to stand back and put his own values aside for a while, at least while
working, to enable him to learn what makes a host community tick. This will
be necessary to work within it to achieve the goals he has accepted for the
project or, when that is impossible, to get out gracefully.
  Talking to natives or experienced foreigners may be the next best thing to
living for years in a place, but these are not the only alternatives. For some
people, it is easier and better to watch and listen to others without asking
questions. It is certainly less intrusive. Often, a conversation about events
seemingly unrelated to grains and farming will reveal ideas and attitudes which
affect the proposed actions. Economic insecurities, anxiety about family nutrition,
worry about too much centralized control, and local labor problems are
examples of things worth listening to. Reading local newspapers and attending
local public functions where appropriate are useful techniques; beware,
though, of being inadvertently classified with a political party or social class
that is linked with the newspaper or the function. In any case, keeping eyes
open, and perhaps keeping a diary of observations, will pay off. And if your
function and aim become well known, you will receive much useful information.
  In some groups, the very existence of a foreigner implies change and is a
threat to some and an object of economic courtship to others. It is hard for
foreign experts to avoid getting tangled in political games; if we have money to
spend or control, we are obvious objects of interest and concern. In some of
the more cosmopolitan places, however, where agricultural development and
extension work is commonplace, a new face is more easily accepted. Unfortunately,
the very places where acceptance is easier are also those with more
complex and intricate social and economic relationships, so the job is proportionately
more complex.
Culture or Cultures?
  It is convenient but rare to find a homogeneous community with similar
beliefs and behavior. More often there is a continuum of behavior from traditional
to daring, and sometimes a sharp age distinction, separating the younger
people who grew up after World War 11 in an atmosphere of independence and
international communication, from the older generation for whom tomorrow
was expected to be the same as today or yesterday. Sometimes the split is
between urban and rural, or factory workers and farm workers, or on racial or
religious lines, and, of course, there may be more than two groups involved.
The careful observer, then, will not automatically assume "one culture" but
look for signs of pluralism that will help him to identify, classify, and eventually
understand the different attitudes and behaviors of different people.
Advance Preparation
  Much can be learned before ever setting foot on the location to be studied or
assisted. In almost every area of the world, hundreds of observers have already
been there and, consequently, there are hundreds of books and articles telling
about the people and their cultures, ranging in quality from useless to marvelous.
Therefore, it is inexcusable not to study in advance.
  Most field-workers get basic country information through their own agencies,
the host governments, or their own government post descriptions. These
are adequate if they are up-to-date and not too strongly aimed at visiting
businessmen and officials who do not have much contact with the rural people.
A more subtle problem is the definition of a country or region through the
eyes of its own U.S./Europe-educated officials and managers. These people
may ignore basic aspects of the culture because, with good intentions, they
think they are useless blocks to progress.
  More detailed cultural information is available and worthwhile. Some sources
  1. The American Anthropological Association, which has a division concerned
with agricultural development, with names and members keyed to regions
and experience topics. Contact John Bennett, Washington University,
St. Louis, Mo., or Iwao Ishino, Michigan State University, East Lansing,
  2. The anthropology department of the nearest major university. In checking,
you may find a student just back from a year's fieldwork there and eager
to tell you what he knows, or a professor who is a recognized authority. Or the
faculty may know who in other universities would know what you want, as this
discipline is a complex and well-functioning information network in itself. (A
word of warning: "anthropology" outside the English-speaking world sometimes
is narrowly defined as study of physical characteristics and perhaps of
primitive tribes. In these areas, study of the culture-linked aspects or agricultural
behavior may be found in departments of sociology, ethnology, economics,
or in agriculture itself.)
  3. The Human Relations Area Files at Yale University, New Haven, Conn.
which has cultural information on most of the world. You do not have to go to
New Haven to use it, as many other universities have access.
  4. A meeting of appropriate professionals, such as the American Anthropological
Association which meets each year in November, with numerous
speakers, and its subdivision on agriculture meets at that time as well. A
related and useful organization is the Society for Applied Anthropology which
meets in the spring of each year. Details on both groups are available from
their common headquarters at 1703 New Hampshire Ave. N.W., Washington,
DC 20009.
  5. The Society for International Development, an organization of development
professionals -- economists, technical consultants, officials, and field-workers
in aid organizations, and a few anthropologists. Most of the members
have international field experience. There are chapters all over the United
States and Europe as well as in some developing countries. The New York and
Washington chapters are the largest and hold several meetings each month; the
Washington group even has a rural development subdivision. For more information,
contact the North American office, 1346 Connecticut Ave. N.W.,
Washington, DC 20036, or the world headquarters at Palazzo Civiletta del
Lavoro, EUR, 00144 Rome, Italy.
Sources in Developing Countries
  If you are already in the field, it may be difficult to reach many of the
sources noted above. If there is time, you can write to them (offer to pay for
Xerox, book, and airmail costs). But if you have to gather knowledge yourself,
there are still a few things you can do.
  If you have prepared a grain flow sheet (or pipeline chart) -- a diagram
showing the channels and amounts of grain as they move from farm to consumers
-- there will be certain key locations that control movement. Going to
these places and watching who does what can be very useful, if it can be done
without obvious intrusion. For example, watching who buys grain at a central
market will yield information on purchase quantities, which in turn tells us
about home storage. If no money changes hands, there may be a credit situation
which controls purchase.
  Watching harvesting and transportation of grains is also useful, and often
possible in the role of technical expert. But it will help to learn who the
workers are, who owns the vehicles or animals, what happens to spilled grain,
and other such factors. The object is to understand the economic relations
among the people and ultimately to understand the potential effects of any
proposed changes.
  Officials and local counterparts in a grain saving program are certainly
available sources of information, but must be heard with caution. Some are
farmers themselves, or have worked in the grain pipeline for years, but others
may not really know how the majority of farmers and consumers behave; or
they may not want to talk in detail about behavior that they consider old-fashioned
or even embarrassing. We do not wish to imply that all or even most
local officials are devious or misinformed; we warn only against uncritical
acceptance of their descriptions without other indications or feelings that they
are sensitive to and reporting what is going on around them.
  There are many information sources in developing countries beyond the
officials. Many countries have a strong awareness of their own cultures and
have much published research. Appropriate university departments and libraries
as well as government officials can be helpful.
  It is often useful to look at people through the eyes of observant and
articulate members of one's own culture. They can anticipate problems and
reactions, and their advice should be sought. These might include anthropologists
in the field, workers for volunteer organizations, or missionaries.
Key People
  It is important to identify key people who can influence acceptance of
changes, but it is also important to distinguish between apparent influence and
real influence. Some people in important positions may really be servants of
the position and cannot promote certain changes even if they wanted to. (This
also is true in Europe and America.) Thus, personal and logical argument will
be useless and may even embarrass the official who knows you are right, but is
reluctant to explain why he must disagree.
  Some positions of authority are temporary and others permanent, so it is
important to know the system by which people come in and out of power. This
can be quite complex -- in some areas, for example, people move up both
religious and political ladders, switching back and forth in a traditional pattern.
  Many of these traditional systems are breaking down in the face of modern
technology, communications, and other influences. Sometimes a foreign fieldworker
finds himself/herself a symbol of change, with corresponding personal
alignments and antagonisms, even before he ever says or does anything. This is
a hard position to be in and some projects are doomed to failure or dormancy
(a more polite and often more profitable alternative) no matter what the
technical or economic merits of the proposed actions. Even if nothing can be
done, it certainly is good to be aware of such situations and perhaps ask other
colleagues about them on arrival, as part of initial briefing.
The Culture of Development
  The development business has its culture, too, involving both foreign agents
of change and local managers. Everyone has his/her own interests, and it is
reasonable to expect people to act in their own interests.
  It is often easy to blame inaction on a few individuals, or on one class of
people or another, but development is not that simple. In reality, people of all
classes will resist risk, even as they desire growth and improvement of their lot,
if they sense the chance that their status might change for the worse.
  From this need to minimize risk emerge the relations among government
people, local businessmen and farmers, technical experts, and representatives
of foreign and domestic money sources. These relations build, of course, on
existing socioeconomic patterns, and are themselves dynamic, changing as
needed to maintain development money input with minimal disruption.
  In each location, this network is unique, and there can be no fixed guide to
inform the newcomer, but a discerning fieldworker can easily see what is going
on. Observe the social relations of the participants -- who is invited by whom,
who accepts and who can reject, who pays at lunches or dinners, what reciprocity
is expected and what is given, who visits and who stays put, and who
waits for whom at appointments. Watch language cues, too, such as the use of
the familiar verb forms, first names or nicknames, and dialect or slang in
direct conversation.
  In any such network, some people are more free to act than others, and this
degree of freedom should be noted for the people one must work with. In
general, technical experts have more freedom (but less power) than political
officials, young or old people more than middle-aged family heads, people
from another area more than others with local family and business connections.
These are guides, of course, and not rules, and there will be many
  In some places, there are long-standing patron-client relations which keep
subsistence farmers in permanent debt and service, or else maintain them as
low-paid farm workers. To the patrons, anything that may increase the economic
power of their clients -- even a grain use survey -- may be seen as a
threat to the current status, often already endangered by the communications
revolution. Some patrons are very troubled by this; others do not care. They
will all usually cooperate with government and change-agents, and many really
do want their people to eat better if that were possible without disrupting the
entire structure that they feel responsible for maintaining. In fact, where
leaders are sufficiently secure as to be benevolent in deed as well as word, there
is the greatest chance for successful change, as the leadership can then get
things done.
  A special problem is the self-perpetuating project which employs many
people including international civil servants, is government-sanctioned and
supported, and has no place to go if it succeeds. Thus, projects are kept in a
state of incipient success to assure the flow of money and support, as well as
the absence of disruptive change. Seldom is this a conscious conspiracy; more
often it arises from the very nature of the situation.
  Much of this is common knowledge among careful analysts of the development
business. We include it here, though, because it may be useful for fieldworkers
new to development, and also because the interface between fieldworkers
and local officials is an area worth more attention and understanding,
even among the experienced.
What Are We Looking For?
  To understand local behavior with respect to food production and consumption,
observe these areas:
  1. What is the money flow in the food system? What credit system is used?
Are farmers truly independent, or are they dependent through debt, or laborers
on land owned by others? Is there a social reciprocity system that
reinforces a dependence situation? And on whom are they dependent? Can
they afford the extra inputs to invest in new seeds, techniques, or equipment
that would ultimately recover more grain?
  2. What is the belief system of the people regarding food supply? Do they
see it as a purely commercial transaction or are supernatural forces involved?
  3. Do they understand the connection of more food with better nutrition
and health, ie, do they see themselves as having some control over their health?
  4. What are the social connections to securing and consuming food? Is
much food given away, or eaten in larger gatherings, and how would that
affect the costs, risks, and benefits of saving more food? Can social obligations
be used by hungry people to buy food, and thus give more incentive to
grain recovery? Food has many social and personal functions in addition to
nutrition and these should be well understood so that suggested changes permit
continuity of these functions.
  5. What do the people do with extra money? If saved grain is sold for cash,
then the saving may be less critical. If extra money opens problems of taxes or
extra grain opens increased obligations within a social reciprocity system, a
saving may be disadvantageous to the grain owner.
  Other questions and attitudes are explored in Part B of this chapter.
Social and Economic Ecology
  Even with current ecological awareness, it may still be necessary to recognize
the interrelation that exists. The facts of ecology are well known for animals
and plants and the physical environment, but are surprisingly neglected in the
social and economic spheres. There are social and economic ecologies, too,
and the effects of a survey or proposed change are felt in many ways, and
among many people other than those directly involved.
  Social ecology may be linked to economics, if economics is broadly defined
to include all actions that maximize security and the ability to cope with one's
surroundings. People relate to one another, form and break alliances, cooperate
and compete. Some hope only to stay alive, to break even with life, while
others -- more and more as the potential for change becomes known -- try to
improve their levels of wealth, power, and prestige. The individual entrepreneur,
in fact, may well be a role learned from colonials, who brought with
them the idea that work and intelligence (cleverness) can raise a person from
low to high in a lifetime -- a phenomenon previously seen only via miracles
and natural events, not under one's own control.
  To understand social ecology, it is useful to describe levels of wealth and
power within a community and to learn the paths by which people can get
there. Some positions will be very stable, others precarious, and the degree of
stability should be noted as well. Then, the effects of a study or a proposed
change can be cast against this background: What will happen to X if we do
this? Or how does X see this change as affecting his community and his
position? Remember that he may see the exercise from a different vantage
point than that of the investigator.
  It also helps to learn how people define security, what their real aims are,
and whether they understand that they can better their lot without incurring
enemies who now have less. Competition may be based on the philosophy that
if I get more, someone else will get less. Riches breed anxiety in such a system,
and it serves as a device to inhibit excessive differentials.
  Local social customs define associations. Such customs act as social glue to
serve as markers of who belongs where, or who wants to move where, or who
can trust whom, or what set of rules a person is following. Customs can also
define social boundaries to identify different groups within a community.
  Economic ecology can also be viewed in numbers. This is the grain pipeline,
but determined from farmer to consumer, with attention paid to debts incurred
and values received along the line, not only in money but also in
services and promises of services. Prices may be less to one person than another;
that is not always unfair, as it may be the seller's way of repaying a debt
or earning a future favor. Credit is all-important in understanding the pipeline
as the farmer's actions may well be linked to his credit sources and their limits.
  Another socioeconomic factor is visible difference. A man may not want to
do better than the others, at least visibly, if envy is to be avoided. In some
societies, invisible success is tolerated but in others it is betrayal of the common
good, and only a cooperative or communal effort will work, as no one
would be obviously climbing over the others. A knowledge of attitudes toward
envy and success should be useful in planning the scope of proposed changes.
  Outside development processes have reached almost everywhere in the world
and the remembered effects of local involvement have not been universally
favorable or unfavorable. Onset of a new program, either survey or direct
assistance, is an intervention into today and brings with it future concerns. The
investigator will get more done more accurately when he knows the actions and
interactions of the people he is working with, when he recognizes the similarities
and differences among them, and when he knows where they have been
and which way they are going.
                                  CHAPTER III
                         B. Anthropological Signposts
                              C. C. Reining
  The researcher or project manager needs a clear understanding of the cultural
and social setting in order to meaningfully assess grain losses. At its most
basic level this means knowing who does what to the grain, how, when, and
why. It is easy to see that measurements of tangibles should never lose sight of
the people who produce, process, and consume those tangibles. However,
there is a need for understanding the human social and cultural factors which
go far beyond that immediate level and which will dramatically influence the
degree of success of a loss assessment effort.
  Because so often the project managers in grain loss programs are outsiders
to the area being studied, there may be a high incidence of cross-cultural
communication gaps which can impair the progress and accuracy of loss surveys.
However, with careful effort, much can be done to overcome such
cultural perception difficulties. As cross-cultural communication gaps are
likely to occur throughout the span of the project, the effort and time spent in
developing a cultural understanding will more than repay itself in later-saved
time and expense.
  Good social and cross-cultural communication skills will be required in
selecting, training, and supervising field-workers; in determining what questions
need to and can be asked in field surveys, and in ascertaining how to
phrase them for ease of comprehension; in identifying which individuals are
the best informants for specific questions; and in allowing for and putting into
proper perspective potential biases including those of the local farmers, grain
handlers, extension workers, field investigators, and the project manager himself.
Particular objectivity will be needed when local ideas and values differ
from those of the investigator.
  The continual need to balance and blend technically ideal procedures and
approaches with social, cultural, and political realities is a process which will
influence conscious and unconscious cultural values and perceptions. More
than any other discipline or subject area involved in grain loss assessment and
reduction, the sociocultural one lends itself least well to a step-by-step or
procedural treatment in this manual. The cultural observation guides provided
at the conclusion of this chapter should not mislead the reader. No such guide
could be comprehensive. The guides presented here are provided as a tool -- a
thought-provoking means of helping project managers and their personnel to
formulate their own process for understanding the salient aspects of the local
culture and to develop the greatest possible depth of understanding.
  In many circumstances, the limited time available for survey planning will
make invaluable the short-term services of expert anthropological or sociological
assistance. It is assumed that every project would benefit from the assistance
of such expert staff members, although the reality of limited project
funds and personnel will often mean that such professional assistance will be
brief. Where such assistance is not available, a suggested analytical tool for
identifying the human element in the grain pipeline is in following through
each relevant process or stage in the pipeline to trace what might be called the
"grain handlers' pipeline." This can be usefully broken down as to who (age,
sex, and social position) does what, when, where, and why. As the situation is
studied in more depth, critical and subtle elements will become clear, including
who has the decision-making authority and which individuals might be most
and least amenable to changes in their present grain handling and storage
  In spite of recent widespread recognition that women's roles in developing
countries have been largely overlooked, it is useful to emphasize this issue
again here. In subsistence farming cultures, women often perform many of the
tasks in grain handling and storage. Too often researchers and project planners
have failed to see and describe the role played by women. As a result, vital
parts of the intricately interwoven cultural framework have remained unobserved
and unaccounted for, only to be unpredictably changed, alienated, or
harmed when programs are initiated to improve the situation.
  An outsider, defined as any person who does not live in the community,
finds it difficult to find out who does what, why, how, and when. When the
investigator is a man and the major tasks are performed by women, the problems
for an unknowing man can be insurmountable. It is not satisfactory to
ask the men of the village what the women do, how they do it, when they do it,
and why. It is not uncommon to have the men say that a certain task is done a
certain way, and to find out later that their perceptions are off, when the task
is performed by women. In addition to men's lack of awareness about particular
details of women's work, one must add the cultural constraints imposed on
outsiders, particularly those who are men, in communicating directly with the
women. This takes time and carefully selected and well-prepared investigators.
  Female survey workers may be necessary in some cultures to gain access to
women. However, it is overly simplistic to assume that a female worker will
necessarily be more perceptive or reliable than a male in specific women-oriented
investigational work. If there is a severe access problem in outsider
men even being able to talk to women, it may be essential to have female
investigators, although in selecting field investigators, the more perceptive,
imaginative, reliable worker is always preferable, whether male or female.
  When project managers and their field-workers do not speak the same language
and especially when there is a marked difference in their cultural or
social backgrounds, the inevitable communication problems caused by translation
and cultural differences need to be recognized and dealt with. Fieldworkers'
understanding of instructions and the reliability of their observations
must be carefully verified. This verification needs to be done in a number of
  1. Regular personal observation in the field by project managers to check
on workers' methods and reliability.
  2. Rephrase questions and instructions to assure full understanding and
accurate communication between director and workers.
  3. Check several sources of information for cross-checking of observations
and assumptions.
  4. Get to know field-workers' ways of thinking, biases, weaknesses, etc.
  5. Keep to a minimum the number of intermediaries between project director
and the village situation, to minimize communication problems and distortion
of information.
  In summary, it would be hard to overestimate the importance of social and
cultural awareness and understanding on the part of loss assessment project
managers and their personnel. Personal flexibility and willingness to learn will
be great assets in order to gain this understanding. Countless decisions will be
made which draw on this cultural understanding in balancing and adapting the
project's technical needs and scientific ideals with social and cultural realities.
  The following cultural observation guides are intended to help bring to light
salient cultural factors, although no amount of study and instruction will
replace the learning opportunity of direct, personal experience in living and
working in a cross-cultural setting.
1. Social Organization
   a.   Describe the levels of wealth, power, and prestige in the community.
       (Comment: Relations between social classes can have a profound effect
       on handling basic items such as grains.
   b.   Who and what comprises the basic production unit?
   c.   Who and what comprises the basic consumption unit?
   d.   If they are not the same, why is there a difference?
   e.  How do these units form into larger units?
   f.   What are the local names of these units and do they have meanings?
   g.   Which persons or positions are the leaders within each level and how do
       they communicate?
   h.   Who does the harvesting, transporting, drying and other preparation,
       and storing?
   i.   Who removes grain for sale or consumption?
   j.   Who has control of the grain before and after storage?
   k.   What is the relation between producers or producing units and purchasers
       of the grain?
   l.   Are there any legal restrictions on the sale or transport of grain?
   m.   What are the differences in storage of grains intended for sale as compared
       to those intended for home consumption and for seed?
   n.   If there are crops intended entirely for sale, what are the differences in
       responsibilities and in handling?
   o.   What types of occupational specialists are involved in the grain production
       and storage?
   p.   Who obtains the materials for storage facilities?
   q.   Who builds the storage facilities?
2. Domestic Organization
   a.   How large is the usual household and what kinds of relatives does it
   b.   Does it contain any unrelated persons, such as permanent servants or
       temporary laborers?
   c.   Is the household the basic unit or a subunit of production and/or
   d.   How does the household link with the rest of the community?
   e.   What kinds of work are usually done by women?
   f.   What kinds of activities are avoided by, or restricted for, women?
   g.   What kinds of work are usually done by men?
   h.   What kinds of activities are avoided by, or restricted for, men?
   i.   Who makes the decisions about the various stages of production, storage,
       processing, and sale or consumption of grains in the household?
   j.   Can exceptions be made to the rules about who makes the decisions and
       under what circumstances?
   k.   Who does the training in storage techniques?
   l.   What happens to stored grain in the event of death(s)?
   m.   How is transfer of authority made on the death of heads of consuming
       and/or producing units?
3. Cultural Factors
   a.   Are losses permitted because of lack of awareness?
   b.   Are losses felt to be inevitable?
   c.   Are the people concerned about their grain losses?
   d.   What do they think should be done and why haven't they done it?
   e.   Which grains do the people believe store the best or longest?
   f.   Which grains do they believe are hard to store?
   g.   How do they explain the differences in storage characteristics?
   h.   How do they accommodate these differences? Do they have different
       methods? Do they consume some grains more quickly than others?
   i.   How does the availability of other crops, such as root crops, influence
       the storage of grains?
   j.   What are the indigenous materials used to help prevent damage to
       stored grain?
   k.   What do the people see as the tangible causes of damage to stored
   l.   What are felt to be the intangible or supernatural forces controlling
   m.   How do they attempt to influence both the tangible and intangible
       (Comment: There are serious problems of categorization here, both in
       Western and indigenous terms. Often the distinction between "magical"
       and "scientific" becomes blurred, as when a local remedy that is
       felt to have mostly spiritual qualities may, in fact, have demonstrable
       effects on stored grain, while other devices believed to have more direct
       effects do not have any discernible ones. Most preventative practices are
       a blend of empiricism and mysticism.)
   n.   What will be eaten that might have been damaged?
   o.   What are the local guidelines for what should and should not be eaten?
   p.   What is done with spoiled grain? For example, is it fed to chickens or
       other domestic animals?
4. Transition and Change
   a.   Is a need for change or improvement felt by the local people?
   b.   How do they want to change the situation?
   c.   Is their knowledge of desired change sound enough to understand the
   d.   Can they afford the new materials?
   e.   Will they be able to sustain the new equipment and techniques?
   f.   How do innovations get into the community? Are there key positions or
       individuals for introducing innovations?
   g.   What improved procedures have been introduced? By whom? Successfully?
   h.   Have storage systems of various indigenous systems in the same kind of
       environment been compared?
       (Comment: Most communities have had a long time to experiment with
       adapting to their particular setting. It is usually difficult to improve
       upon the local arrangements given the resources available. If introduction
       of new techniques is deemed necessary, it may be more effective to
       consider transfer from a similar indigenous setting rather than from
       Western culture.)
5. Individual Factors
   a.   The local person
       i.    How typical is the person supplying the information?
            (Comment: Often the typical or normal person is too busy to want
            to spend time talking with outsiders. The persons most available too
            often are marginal to the community.)
       ii.   What does the informant see himself or herself getting from the
            (Comment: It is very human to constantly assess any situation to
            maximize the returns. Beware of creating false hopes.)
       iii. What are the biases and interests of the interviewee?
       iv.   Is the interviewee skewing the information to fit the situation as
            (Comment: There is often a tendency to tell the interviewer what the
            interviewee thinks he wants to hear. Misunderstanding is altogether
            too frequent. Consider the difference in response if the interviewee
            thinks there may be a tax imposed on the stored grain, as compared
            to the impression that compensation may be paid for lost grain.
        v.   Are the interviewees saying what should be rather than what is
            actually the case?
            (Comment: It is important to distinguish between the real and the
            ideal. Observe what they do as well as recording what
            they say.)
   b.   The interviewer
       i.    What are the biases of the interviewer?
       ii.   What are the biases and interests of interpreters, if used?
       iii. Are problems perceived from the viewpoint of the interviewer or
            from that of the interviewee?
                         IV. REPRESENTATIVE SAMPLING,
                          INTERPRETATION OF RESULTS,
                          ACCURACY, AND RELIABILITY
              B. A. Drew, with T. A. Granovsky and C. Lindblad
                               A. Introduction
Basic Assumptions
  Every scientific measurement is based on some kind of assumption regarding
the real world about which the measurement is supposed to supply some
information. Conducting a survey to measure average grain losses is such a
measurement and it is based on the following assumptions:
  1. Cultural and economic conditions, level of knowledge of farmers, farming
     practices, varieties grown, and harvesting and storing practices are
     essentially uniform throughout the area to be surveyed. If this assumption
     is to be verified by local observation, one will have to understand the
     cultural milieu. If it is nonuniform in ways that can possibly affect what
     is to be studied, sampling becomes more complicated and the advice of
     experts should be sought.
  2. All grain to be considered is stored in the same manner in units of
     approximately the same size. That is, the largest unit is no larger than
     five times the smallest. If the size variation is greater, then they should be
     sampled and analyzed separately as two or more populations.
  3. Size of farms is uniform to within a factor of 5. That is, the largest farm
     is no larger than five times the smallest farm (in area producing crops for
     storage). Again, if the size variation is greater, then they should be
     sampled and analyzed separately as two or more populations.
  These assumptions limit the survey described to a single stratum. This is all
that can be done using the simple sampling plans outlined here. More complicated
plans should involve the help of experts in sampling as well as in grain
loss assessment.
Uses of Survey Data
  In designing a sample plan it is essential to know the purpose or purposes for
which the results are to be used. For example, one might wish to determine the
calorie losses which are incurred due to parasites, in order to determine
whether to supplement the farmers' diet, or one might wish to determine the
extent of losses in grain held in storage in order to decide whether to treat
it with pesticide. In one case, medical-nutrition concepts are involved; in the
other, grain losses.
  The ultimate use of the results will influence not only the precision and
accuracy which are required, but also what is measured and what additional
data must be collected. Thus, the measurements which are made and the
ultimate use of the results, including the level of loss that is acceptable, must
be decided before the survey is designed.
Editors' note: Given the present refinement of loss assessment methods, it is
generally accepted that  [+ or -] 5% accuracy(2) is the best practical limit which can be
expected (with rational allocation of resources and time against the potential
value of the reduced grain losses). At the same time, where losses are expected
to be 15% or less, a [+ or -] 10% accuracy level could all but obscure any meaningful
information. Where such is expected to be the case, rapid expert assessment
of critical loss points may be economically justified while an extensive in-depth
loss survey is not. For certain economic evaluations, no less than [+ or -] 5% accuracy
can be tolerated for analysis to be meaningful.
Determining Area to be Surveyed
  In making a survey over a large area such as a whole country or region, the
sample population should be divided into parts to reduce the problem to
manageable proportions or to obtain a uniform population. This is called
multi-stage (stratified) sampling.
  In such a situation there are two valid alternatives for sampling a
population. These are: To include in the sample of a population all of its
subdivisions, or to include a random sample of subdivisions of the population.
  Section B presents these sampling methods in detail. The rule for this choice
is to take all the subdivisions when there are only a few, say 10 or less. If there
are more than 10 subdivisions, then as many as are consistent with available
resources should be chosen using random numbers. At such a point knowledge
about the differences between particular subdivisions may make a valuable
contribution to deciding whether to choose all or a sample of subdivisions.
Advice from knowledgable people in this area should be sought.
  Types of subdivision are extremely dependent upon the local situation but a
country (nation) may be divided on political boundaries such as states or into
units based on geographic considerations such as lowlands, uplands, river
valleys, and arid regions. The last division would be preferred when knowledge
or advice is available about the impact of such conditions upon storage losses.
In such a case, resources might be allocated to the various regions in proportion
to the likelihood of postharvest losses.
  The next subdivision might be on the basis of villages or small administrative
or political units. Here the units of the subdivision should be listed and
random numbers used to choose as many units as can be measured with
available resources. Remember that excessive variations in size of storage unit
may require separate analysis of samples as two or more populations.
   (2) In this manual accuracy is expressed in absolute terms. Thus 20 [+ or -]
5% means from 15 to 25%.
  If there are different types of stores within the unit (administrative or political
unit), then each type of store should be considered as a unit in the next
subdivision. It is the last possible subdivision to which this manual refers.
  Accuracy of an assessment of grain losses depends on obtaining a truly
representative sample and making an accurate measurement on the sample. No
matter how accurately one measures a sample in the laboratory, the result will
be of little value if the sample is not representative. It is equally pertinent that
no matter how representative the sample may be, the final result will reflect all
the shortcomings of the laboratory measurement.
                                  CHAPTER IV
                            B. Probability Samples
  The rest of this section will be devoted to methods to ensure a representative
sample and to avoid all sources of systematic error often called bias. If we
always sample the best-looking stack in the field, or the one nearest the house,
or the one the farmer chooses; if we always take samples right by the entrance
into a granary, or where the grain looks good, then we may be putting a bias
into the sample. Even if we try to choose in a way to avoid bias we may
over-correct. If we try to avoid choosing units that are easy to reach, we may
unconsciously choose units that are hard to reach. The only way to avoid bias
is to take the choice out of our hands, to give it to a table of random numbers.
The method is called "probability sampling," and its result is a "probability
A Random Sample or a Representative Sample?
  When establishing a sampling pattern, confusion exists between the terms
"representative sample" and "random sample." Representative sample usually
refers to a "stratified random sample," in which strata are defined and
represented in the sample in proportion to their size in the sampled material.
  If 1) the strata have something to do with the property to be measured and if
2) a random sample is taken within each stratum, the variance of the estimate
may be lower than that of a completely random sample. Both conditions are
necessary, however. The following examples will clarify what is meant by such

pglx490.gif (486x486)

terms as randomization, stratification, random sample, and stratified random
  The two sampling patterns given below are not recommended for use in a
loss assessment survey, but are presented for clarity.
                               Systematic Sample
        A sample is taken every so many units, eg, every 10th bag as it is
     moved from location to location.
        Some problems to be encountered are assumed damage or loss is
     uniformly "normally" distributed, which is rarely true for insect
     populations, the sampling pattern may conform to some inherent
     distribution pattern of the damage, and no random component is
     included and therefore statistical procedures cannot be used.
                         Centric Systematic Pattern
        A sample is taken from the exact center of each unit. If such
     samples are analyzed using parametric statistics and compared to
     samples obtained by the random pattern, results may truly reflect
     what is present.
        All problems present with systematic samples are also present
     with centric systematic pattern.
        The sampling patterns illustrate the advantages of having some
     knowledge about the material to be sampled, and show one way to
     use such knowledge. But when there is no knowledge from which
     strata may be deduced, complete randomization is the only way to
     obtain a representative sample. This applies to each cell or stratum
     in any scheme of stratification. A random sample should be taken
     within each cell or stratum. Otherwise, the advantages of stratification
     may be lost.
Properties of Probability Samples
  This section presumes that a probability sampling plan will be used. The
reasons for this are:
1. With this type of sample one may calculate confidence limits within
   which the actual value of the result is reasonably certain to lie.
2. Generally one may determine in advance how many samples must be
3. This type of sample is guaranteed to be representative.
  The actual value is the value which would be obtained if the loss in every
unit in the area were to be determined.
Observational Units
  The observational unit is the container, location, or process from which a
sample will be removed to determine the loss evident in the sample. This is the
smallest division or unit in which grain is held. It might be stacks in a field,
small silos or granaries on a farm, or woven baskets. It would be a single
basket rather than all of a farmer's storage baskets; it would be individual bags
rather than the whole warehouse. Accuracy of the entire survey will depend on
the accuracy with which the loss is determined on each observational unit.
  To facilitate sampling, the observational unit should be as small as possible.
This makes it easier to get a representative sample since it will be possible to
mix all the grain thoroughly and reduce the sample taken by quartering or
using a sample divider. This may be feasible where the grain is in baskets or in
stacks in the field. In silos or granaries it may not be possible and, unless the
sampling is done with skill, the sample may contain a systematic error which
cannot be removed by any later calculation or analysis.
  When any container is sampled as a unit, the assumption is that the defect,
contamination, or other characteristic to be determined is uniformly or at least
randomly distributed within the unit. As a practical matter such is usually not
the case.
  Insects/mites, moldy kernels, rodent depredation, and insect-eaten kernels
are more usually in location-oriented pockets (see Appendix A).
  With time and money constraints and often with cultural-traditional limits
also imposed, the best that can be done is to design the mechanical sampling so
that the sampled grain will be as representative as practical of both the undamaged
material and the layered or pocketed defects.
  In any study the investigator needs to report what was done and why so that
the significance of the data can be understood by those who will use it.
  Where grain is stored in storage units of variable sizes or types, a person
with competence in statistics should be called upon to help design the sampling
Number of Samples
  To decide roughly how many samples must be taken, two items of information
are needed: the desired confidence limits, ie, the estimate of the overall
average loss within 1, 2, 5, or 10%, and the range of losses to be expected. The
range is the difference (in percent) between the highest expected result and the
lowest expected result.
  With these two items, one can find from Table 11 how many observational
units will be sampled and measured to get a representative sample. If the
number to be sampled is too costly for available resources, the desired confidence
limits will have to be lowered. If the range is underestimated, the number
of samples taken will be insufficient. Therefore, it is generally recommended
to make liberal estimations of the range expected unless the population is
well known.
  For example, as shown in Table 11, if the lowest result that is expected is
25% loss and the highest expected result is 85% loss, then the range is 85 - 25
= 60, and if the desired precision is [+ or -] 5% the sample must include at least 81
units. If a sample of 81 units gives a result of 40% loss, the results should be
interpreted as 35-45% loss (40 [+ or -] 5%).
  The above procedure is calculated on American Society of Testing Methods
(ASTM) Recommended Practice E122-58 and is based on statistical theory.
Other procedures for determining sample numbers which are based on intuition
such as arbitrary numbers and square root samples do not allow specifications
of desired precision in advance.
  Table II is mathematically calculated to assure representative sampling regardless
of total population size. It is based on the range of results expected
and desired confidence limits.
  If the actual number of units is less than the number given in the table, then
all of the units should be sampled.
Preliminary Surveys
  A preliminary rapid fact-finding survey, mentioned in several places in this
manual, is of value in gathering information to assess the homogeneity-nonhomogeneity
of the system.
  Answers to the following kinds of questions should be obtained by the
preliminary survey:
  * Are there large differences in culture? Income level? Farming, harvesting,
    drying, storage practices? Crop and variety grown?
  * In what size unit is grain stored? What is the largest unit found? The
    smallest? How many of each class?
                                   TABLE II
                        Required Number of Samples
                                      Range of Results Expected
                        100     80       60     50    40     30    20    10    5
                        (%)     (%)      (%)    (%)   (%)    (%)   (%)   (%)   (%)
          [+ or -]1%   5,625    3,600   2,025          900          225  
Desired   [+ or -]2%   1,406      900     507          225           57     
Precision [+ or -]5%     225     144       81           36           9         ...
          [+ or -]10%     57       36      21            9            3   ...   ...
Note: This table was derived by standard calculations based on a conservative estimate of
      population-defined standard deviation = range/4.
      Sample numbers in this table were calculated using eq. 1 in Recommended practice for
      choice of sample size to estimate the average quality of a lot or process, ASTM E122-58,
      American Society for Testing Materials (1958).
  * How big is the largest farm (village)? The smallest? How much land does
    each actually cultivate with crops that will be stored? Can you make a list
    of all the farms? Can you locate them on a map?
  * How many storage units of each size class are there on the biggest farm?
    On the smallest? Can you estimate the number on an average farm?
  It may be of value to collect other data in a preliminary survey to facilitate
subdivisions into strata or for other purposes. As the preliminary survey uncovers
separate strata, it uncovers material that needs to be sampled separately
if adequate overall coverage is to be obtained. It is also necessary to look at the
total situation (eg, the subsistence or the marketing systems) and then determine
what elements are to be measured. In other words, what components do
matter? What are the expected ranges of the variables? What should be ignored
as trivial?
  One needs to know all of the possible ways the population stratifies: geographically,
climatologically, politically, and culturally (size of installation,
wealth, mechanization, kinds of storage).
  The pipeline concept (see Chapter II) is a means of sorting out, for example,
situations, locations, economic and political factors. It is a means of focusing
on a situation to reduce the study to a homogeneous stratum.
Designing the Probability Sample
  To design a probability sample, it is necessary to use a method that ensures
that every observational unit in the area to be surveyed has a known probability
to be included. When it is known in advance how many units there are and
where each one is, then a list is made and the units are each given a number in
series from one on up to the total number. Then a table of random numbers
(see Appendix B) is used and those locations whose numbers come up are
sampled and measured until the required number have been done.
  If the number of units and their locations are not known, an estimate of the
total number of units from the preliminary survey can be used to calculate
what proportion of all units to sample. For example, if one wants to sample
200 units and he estimates that the area to be sampled may contain 2,000 units,
then he takes one unit chosen at random for every ten units found. A method
for doing this is to make up lists of random numbers for farms containing
various numbers of units and put them in envelopes for the sampler. When he
comes to a farm that has 51 units, he first numbers each of the 51 units. Then
he opens an envelope labeled   45 to 51" which contains five random numbers
(between 1 and 51 inclusive). He then takes samples from the five units given.
  In sampling farms if the number and location of farms are known, each
farm is given a number and the farms to be visited are chosen with the table of
random numbers.
  Taking samples on a farm which has more than one stack or granary should
also be done at random, taking into account any known pattern of use or any
other known nonhomogeneity. It is best to decide in advance how many units
will be sampled on a farm and to have sets of random numbers of the correct
size in envelopes. Then the sampler can number the units (baskets, stacks)
found, and choose an envelope labeled for that many units that contains the
required random numbers (see Appendix B).
Note: In sampling it is always a good precaution to identify extra sampling
points and to take samples from these sites to replace the inevitable accidents,
dropouts, or loss of sampling sites.
                                   CHAPTER IV
                           C. Detailed Instructions
Choosing Farms or Villages
  All the farms (villages) in the area to be surveyed should be listed and the
number of samples that are required should be determined (see Table II).
  If there are more farms than samples required, and if the farms are all the
same size (within a factor of 5), then
  * Give each farm a number from 1 to as high as necessary.
  * Use a table of random numbers to choose the farms to be sampled. The
    farms chosen may be visited in any order that is convenient.
  * Obtain samples from one observational unit (stack, basket, crib, etc.) on
    each farm. Choose the unit with random numbers after seeing how many
    units there are on the farm.
  If more samples are required than there are farms, and if the farms are all
the same size (within a factor of 5), then
  * Determine (or estimate) how many observational units there are in the
    area to be surveyed. The total number of units is called N and will be
    greater than the number of farms, if several observational units are
    present on each farm.
  * Determine the number of samples necessary from Table 11. This is n. The
    fraction n/N is the sampling proportion.
  * On each farm (or in each village) count the number of observational
    units and multiply by the sampling proportion. The result, rounded to
    the next highest whole number, is the number of units to be sampled.
Sampling on Farm or in Villages
  Labeling of Samples. All samples must be labeled and retain their identity as
to date collected, exact location of source, how sample was obtained, grain
type, variety (if known), time in storage, and type of storage.
  Procedures for Sampling
  Standing Grain in the Field
  * Choose an area (in square meters in broadcast crops or linear area in row
    crops) that will yield 1 to 1.5 kg of shelled grain.
  * Divide the field into units of the chosen area.
  * Give each unit a number starting with 1 and going as high as necessary.
  * Choose as many random numbers from the table furnished as there are
    samples to be taken.
  * Harvest and shell the grain in the unit areas whose numbers were chosen.
  * Package the grain from each unit for transmission to the laboratory.
  In the Field in Stacks (If Each Stack Contains More Than 2 kg of Shelled
  * Give each stack a number starting with 1 and going as high as necessary.
  * Choose as many random numbers from the table furnished as there are
    samples to be taken.
  * Shell each stack whose number was chosen.
  * Reduce the grain by coning and quartering or by using a sample divider
    (see Appendix A) to a sample of 1.5 kg.
  * Package the sample for transmission to the laboratory.
Note: If each stack contains less than 2 kg of shelled grain, choose twice as
many random numbers as there are samples to be taken. Combine the grain
from two stacks into a single sample for transmission to the laboratory.
  When the Shelled Grain is Stored in Baskets
  * Give each basket a number starting with 1 and going as high as necessary.
  * Choose as many random numbers as there are samples to be taken.
  * Reduce by coning and quartering (or use a sample divider) each basket
    whose number is drawn to a sample of 1 to 1.5 kg.
  * Package the sample from each basket for transmission to the laboratory.
  When the Unshelled Grain is Stored in Small Units (Such as Baskets and
Bags). If the grain is stored in small units on the cob, head, or panicle, shell the
contents of the whole unit before coning and quartering to yield a 1- to 1.5-kg
  When the Unshelled Grain is Stored in Large Cribs, Silos, or Granaries. To
sample grain stored unshelled in cribs, silos, or granaries, unload and shell the
entire lot. Then cone and quarter (or use a sample divider) to obtain a sample
of 1 to 1.5 kg. Or unload the grain equally into baskets and then use the
method for unshelled small units (choosing baskets by stratified random sample).
Note: In storage, ears of cob maize or panicles of sorghum/millet and maize
can be labeled randomly as the crib is filled. The farmer can then be asked to
set these ears aside as he encounters them during emptying. Determining an
adequate sample of ears or heads from a crib can be a problem, however. This
procedure should be used only after careful study of its applicability to the
local situation.
  Large Bulk Storage Units, Shelled. Obtaining a representative sample from
a large bulk container is difficult. Ideally the grain would be transferred into
another container in such a way that samples could be obtained from the grain
as it falls into the new container. A container small enough to be handled
easily should catch the entire falling grain stream until it is full or passed
through the entire stream and the caught grain placed into a larger sample
container. This procedure would be repeated at frequent, regular times
throughout the transfer.
  When all the grain has been transferred, the sample that has been collected
may be reduced by coning and quartering or by using a sample divider to 1 to
1.5 kg for transmission to the laboratory.
  If it is not possible to sample the grain during a transfer, then a probe may
be used. It is recognized from research results that a probe sample is not
representative (see Appendix A). When probe sampling is used a note should
be made of that fact in the final report. In using the probe, an effort should be
made to reach every part of the storage container. Several times as much grain
as is necessary for the final sample should be taken and then reduced by coning
and quartering or by using a sample divider. Samples should be taken with the
probe in at least the positions shown in Fig. 6, using a compartmented probe

pgl6x57.gif (426x426)

that samples at all levels.
  Mass Storage in Bags. Obtaining a representative sample of a large mass of
grain stored in bags can only be done if every bag is accessible. To sample such
a store requires that one chose enough random numbers and then move the
grain one bag at a time to a new location diverting bags for sampling corresponding
to the random numbers. The diverted bags should be sampled, preferably
by coning and quartering the whole bag or putting it through a sample
divider to obtain 1 to 1.5 kg of sample for the laboratory. The remainder can
be returned to the bag and to the store.
  A less satisfactory alternative is to obtain a sample from each randomly
chosen bag by probing. A probe long enough to reach diagonally from corner
to corner of the bag should be used and the bag should be probed on both
diagonals and in enough other locations to obtain 1 to 1.5 kg of grain from
each bag.
  It should be noted if every bag is not available to be sampled so the result
will refer only to those bags that were accessible. The bags sampled should be
chosen by assigning numbers to those that are available and using a table of
random numbers to choose the bags.
  Sampling procedures should always be reported, especially when the sampling
is suspected to be nonrepresentative as in the case of stacked bags,
unshucked or unshelled grain heads and cobs, and when there are visually
observed concentrations of insects or mold, or both.
                         V. LOSS MEASUREMENTS AS
                       RELATED TO SITUATIONS WHERE
                                THEY OCCUR
  Many, if not most, postharvest losses occur as a result of externally applied
adverse factors, as when insects, rodents, and birds consume the grain. Other
losses occur while, or because, the grain is in an otherwise useful state or
process. Losses are often sustained while the grain is being threshed. These
losses are brought about by (deficiencies in) the threshing process.
  Grain must be transported from farm to urban centers. During this process,
bags or vehicles may leak and grain is lost along the way. The transporting
process is useful; it also may result in losses.
  In this section, measurement procedures are dealt with as they relate to the
process the grain is undergoing. The techniques for analytical-type testing not
given herein are in Chapter VI.
  Processing losses are affected by prior induced quality factors such as checking
and cracking rice and corn, and a methodology should put such factors in
  Methods are not given for all the procedures needed to determine prior-to-processing
damage that brings about subsequent losses during processing.
Also methods are not given for all processing damage that causes losses during
further manipulation.
                       A. Background Information
                   D.A.V. Dendy, with K. L. Harris
  Two basic concepts are used in this chapter. One is to measure the situation
(usually output) of a given operation and to compare it with an ideal (hand or
special machine) operation. The other is to measure losses by weighing the
various food, feed, and other streams and making direct calculations of what
does not end up as food.
  Whether the loss is waste is not a matter that depends on methodology. Bran
can be waste, feed, or food, independent of loss-assessment methodology.
  What results as food may be compared to total food value, to food obtained
by the best possible process or best possible commercial process, or even by an
experimental process. The methodology simply needs to be set up to make the
required measurements.
Shelling of Maize
  Stripping of maize grain from the cob is known as shelling. Losses occur
wherever mechanical shelling is not followed by hand-stripping of the grains
remaining on the cob. Certain shellers damage the grain, making insect penetration
easier and subsequent storage losses higher.
  Losses occur during threshing by spillage, by incomplete removal of grain
from stalk, or by damage to grain during threshing. They also occur after
threshing due to poor separation of grain during cleaning or winnowing.
  Incomplete stripping usually occurs in regions of relatively high labor cost at
harvest time, where the method of threshing leaves some grain unthreshed but
labor is too expensive to justify hand-stripping. Workers in Malaysia observed
that 1.13% of paddy was lost by falling outside the threshing tub; it was also
noted that up to 11.7% was left on the straw.
  Certain mechanical threshers have cleaning equipment designed for only dry
grain. A wet season's harvest, eg, of paddy, will clog the screens and grain will
be lost with leaf and broken stalk.
  Use of oxen for threshing paddy provides a trodden straw said to be more
easily digested. If the threshing floor is muddy or cracked, grain will be lost.
  There can be a 5% increase in cracked and broken kernels after combine-harvesting
paddy compared to hand-harvesting and hand-stripping.
Cleaning and Winnowing
  Cleaning is customary before milling. At the home, hand-cleaning is a combination
of hand winnowing with hand removal (eg, of stones); losses can be
very low when carefully done or high when siftings are allowed to scatter on
the ground or winnowing done with the same result. With correct equipment,
losses should be low in mills, but equipment undersized for the quantity of
extraneous material, such as dirt, will cause losses of grain by removal with the
dirt or by the dirt being carried forward into the milling stages. Loss assessment
is difficult as losses are usually low; high losses are spotted by operators
and the extraneous matter is recleaned.
  Two losses are frequently caused by drying: removal of grain and portions
of grain from the drying system, and damage to the grain leading to a subsequent
  Grain which is dried in yards, on warehouse floors, or on roads will be
partially consumed by birds and rodents. Wind, either natural or from passing
vehicles in the case of road drying, will blow some grain away. Although very
little grain is removed on vehicle tires, damage by vehicles may cause subsequent
losses. Mechanical dryers may cause damage leading to removal of parts
of the grain (such as bran) from the system either in the air flow or in subsequent
cleaning operations.
  The principal loss-factor occurring during drying is caused by kernel cracking
("checking") of grains such as rice, which are eaten whole. Usually the
greatest damage occurs through re-wetting which happens when grains of
different moisture content are mixed in a dryer, and when rain or dew re-wets
grain in a yard. The damage is manifested as broken grains during milling,
especially in the polishers.
Primary Processing (Milling)
  This includes all processing operations carried out on grain in the home or
mill, such as cleaning, parboiling, hulling, de-branning, grinding, and separating
(classifying). Secondary processing (cooking, baking, fermenting, extruding)
is excluded; such losses as occur are usually unavoidable, being intrinsic to
the process and preventable only by a change of process -- more a subject for
the sociologist than technologist.
  In the home and small mill, grain processing is effectively a batch process in
which relatively small quantities of grain are processed by one or more operations
and the product collated, then brought together for sale or other processing.
In large mills, the processes are continuous and loss measurement is
performed periodically by sampling product streams. All of the pre-milling
history affects the fate of the grain during milling.
  Though easily quantifiable losses of soluble materials occur during parboiling
of paddy, these losses are more than offset by the improvement in nutritional
value of the kernel.
Hulling, Polishing, Especially Rice Milling
  Removal of the outer coats from a grain may take place in one or more
stages. For paddy rice, red sorghum, and oats, considerable mechanical effort
is needed to remove these layers. Any weakness in the kernel, caused previously
or inherent, will manifest itself at this stage. Even with grain in perfect
condition, only the best process with correctly set machinery will yield an
out-turn of whole polished grains approaching 100%. In the case of rice,
broken grains command lower prices and finely shattered material ceases to be
human food. Some leaves the mill in the husk (fuel or waste), but most with
the bran (feed). Bran removal may be considered a loss. With the consumer
demanding rice with a high degree of polish, the loss at that stage must be
measured and then changes made to keep the losses to a minimum. It has been
noted that even a 1% increase in yield of whole grain rice can result in huge
increases in national resources.
  In some processes such as wheat milling, removal of an edible part of the
grain, eg, the germ, is deliberate and desired by the consumer. Whether this is
a loss depends on the terms of any particular study. However, mechanical
losses of desired ground products frequently occur, often caused by maloperation
of the process or worn equipment. Common processes are pounding in a
mortar, grinding between stones or toothed steel plates, and the complex
Hungarian system for milling wheat into flour.
  Whether the separation of edible from less desired products is done in the
home (eg, winnowing hulls and bran from rice) or mill (eg, sieving flour from
bran), complete separation is rarely achieved. With rice, it is difficult to separate
the more finely broken grains from bran, and with wheat, flour adheres to
bran and special equipment is used to remove most of this as flour.
  Processing of mixtures that are nonuniform because of such factors as
hardness and softness of kernels, size (length, plumpness, etc.), and moisture
content difference is itself a cause of losses.
                                   CHAPTER V
                   B. Guidelines for Performing Studies of
                            Farm Storage Losses(3)
                         J. M. Adams and G. W. Harman
  1. An inter-disciplinary team, comprising at least a storage technologist and
an economist, is necessary. The team should arrive in the area early enough
before harvest to enable it to plan effectively, to select fieldwork areas, to train
and brief enumerators, and to conduct necessary trial runs.
  2. The sampling frame for investigations on both technical and economic
aspects should be determined and stratified. Areas chosen for fieldwork
should be as representative as possible of traditional practices, both preharvest
and particularly postharvest. (See Chapter IV.)
  Information on the technical aspects of losses should be obtained by:
  1. Collecting the necessary baseline data on the moisture content, damage,
and bulk density (bushel weight) of the commodity immediately prior to storage,
and recording any procedures involving selection or treatment of the
product for storage.
  2. Recording the quantity of the commodity placed in storage.
  3. Recording the date on which some of the commodity is first removed
from the store. Thereafter samples of the commodity should be taken at
monthly intervals. The sampling method used should be pre-tested, prior to
large-scale use, for its acceptability to both the investigator and the farmer.
  4. Collecting information on the rate of consumption of the stored commodity
over the storage period. This should be done on each sampling visit.
  5. Analyzing the samples to obtain estimates of loss and applying these to
the consumption pattern to obtain an estimate of loss over the complete storage
period. Weight of a standard volume of grain corrected for moisture
content changes should be used to assess losses in samples when regular sampling
is performed. If this is not possible the formula method may be used to
estimate losses within individual samples, but with less accuracy. (See Chapter
  6. Setting up simulation stores, if necessary, which are under the control of
the investigator and simulate the farmers' pattern of consumption. The commodity
should be accurately weighed in and out of the store. Care should be
taken that the grain placed in these stores is of the same quality and selected in
the same way as that placed in the farmers' stores.
  Information on economic aspects will be obtained:
  1. By a questionnaire survey on a once-only basis, conducted with a representative
sample of farmers.
  2. On a regular basis from farmers from whom grain samples are taken, if
this is part of the research, and from official sources.
  The questionnaire survey should be evolved in three stages:
  1. A basic outline following on-site discussions.
  (3)Adapted from J. M. Adams and G. W. Harman. The evaluation of losses in
maize stored on a selection of small farms in Zambia with particular reference
to the development of methodology. Trop. Prod. Inst. Rep. G109 (1977).
  2. A trial run (see below).
  3. A final revision. The questions to be asked will depend on the objective
of the survey, the potential ability of the interviewees to respond, and the time
and staff resources available to the research team.
  The questionnaire should be sectionalized as required by the study. The
following is a guide to some but not all of the main subject areas:
  * General. Farmer's status, household size, measurements of wealth (cattle
    ownership, alternative employment, size of farm), credit facilities and
    usage of.
  * Cropping. Crops grown, area, and disposal/storage.
  * Principal grain crop(s) production. Varieties grown, seed source and
    costs, use of fertilizers and insecticides, drying and pre-storage activities.
  * Storage. Quantity stored, form in which stored, number and type and
    structure of stores, cost of stores and store materials, labor for building
    and maintenance, age of stores, potential life, pre-storage and in-store
    treatments, dates of first and last removals, frequency and quality of
    grain removed, site of removal from the store, usage of grain removed.
  * Storage losses. Cause, severity, usage of damaged grain.
  * Marketing. Sales of grain which is never stored, quantity, variety sold,
    reasons for sales, grade/price made, buyers, transportation.
  * Buying. Quantities bought, form (grain, meal, etc.), frequency, price,
    source, usage.
  It is important to emphasize that the above are broad outlines only. Each
situation may require some addition or deletion and all situations will require
precise framing of the questions to be asked. These six criteria should be
  1. do not ask unnecessary questions; limit the number and complexity of
questions so that each interview is completed in 30 to 40 minutes maximum.
  2. as far as possible, frame the questions so that the answer is yes or no.
  3. have a trial run and revise or eliminate difficult questions.
  4. avoid sensitive questions if possible and seek local advice as to which
questions are sensitive. (It is, however, surprising how many seemingly sensitive
questions can be asked and will be answered if correctly phrased and
properly put, emphasizing the importance of enumerator training.)
  5. train enumerators thoroughly, work with them through their initial field
operations, and spot-check their activities at intervals.
  6. consider the feasibility and advisability of moving enumerators between
areas and strata both as a check and as a stimulus on the individuals' performance.
  This questionnaire survey will probably be asked of a larger sample of
farmers than the one from which samples of the grain are drawn for analytical
purposes (assuming that the latter is part of the study involved). Nevertheless,
all of the latter should be asked the questionnaire survey; their actual activities
on grain removal can be observed in practice and comparisons of observations
and statements will provide a valuable check on farmers who are involved in
making statements in the questionnaire survey only.
  Economic information should be collected on a continuing basis from
farmers. If, as is likely, it is necessary to undertake a program of regular
sampling of farmers' stored grain, regular visits should be made to collect
economic information of usage patterns, quantities and prices for sales and
purchases, time required for store building and maintenance work, and cost of
materials used.
                                  CHAPTER V
            C. Procedures for Measuring Losses Occurring During or
         Caused by Processing Including Threshing, Drying, and Milling
              of Most Grains, but not Maize or Pulses/Groundnuts
                        D.A.V. Dendy, with K. L. Harris
  Processes may be continuous or batch. In the former, samples of input and
output should be taken at regular and measured intervals. The amount (1, 5,
or 10 min) of production taken from various lines in the system can be weighed
to give the quantity of stock carried in that line in proportion to other lines.
Samples may be taken in the usual way from the bags of grain entering the
process and bags of product(s) leaving. Overall mass balances must be
measured and converted to standard moisture content or to dry weight.
  Two fundamental methods are used: measurement of total system (mass
balance), and comparison with a standard.
  Measurement of total system. The loss itself may be weighed. The optimum
process gives zero loss. Examples are threshing (loss on stalk) and maize
shelling (loss on cob). In some cases the loss itself cannot be measured, but the
input of grain and output of products can be weighed, the difference being the
loss. In other cases, loss will be a comparison of the traditional or commercial
system as against a perfect hand-stripping standard.
  Comparison with laboratory standard. Comparison is not against a perfect
(100% recovery) standard but with an optimum standard, usually taking each
unit operation (stage) separately. Although this method is not ideal, if the
standard of comparison is adequately described, the comparison will produce
useful information.
  It is important also that unit operations (eg, hulling and polishing) subsequent
to that under consideration (eg, drying) be investigated or that information
be obtained on the entire flow in the best possible and most standardized
Sampling (see also Chapter III)
  Sampling procedures are simple for batch processes such as are carried out
in small mills and homes. If a loss of material is looked for, then a weigh-in
weigh-out procedure will be adopted. Where a lowering of quality is suspected,
a sample should be taken before the process and put through a parallel but
optimum process (eg, in a laboratory mill) to compare the products. In continuous
systems, the unit operation (stage) can be scrutinized while representative
samples of substrate are taken at regular intervals before and after. The
condition of the inputs and outputs is determined by laboratory examination.
The amount (weight) of the outputs is obtained by comparing the total weight
of the streams over a fixed period of time so that the comparative amounts of
grain going to food, feed, waste, etc., can be determined. For example, in a
continuous flour-milling operation, weights taken over a 1-min period of
flour, bran, shorts, and dust will show what proportion goes into each product.
If dust is 0.5% of the flour + bran + shorts, and dust is used for fuel
while flour, bran, and shorts are all food, then the loss in this stage is 0.5%.
  Where losses depend on operator efficiency, there will always be the problem
of deciding whether the operator is working normally or at an enhanced
efficiency to impress the assessor. The tester must gain the operator's confidence
and impress on him that it is not he who is under scrutiny.
  The following examples can be used as a guide for other unit operations.
THRESHING LOSS 1: Unstripped Grain (Loss With the Straw)
  A suggested method is as follows. Random samples of bundles of cut grain
are chosen and threshed by the customary method. The threshed grain (sample
1) and straw are retained. Directly supervised labor hand-strips every grain
(sample 2) from and out of the straw. The two grain samples are then hand-winnowed
carefully to bring hand-stripped and mechanical material to the
same quality. The good grain is weighed, moisture content measured, and the
weights converted to a standard moisture content.
  It is important to examine the two samples and estimate as accurately as
possible (eg, by hand sorting of a representative subsample) the proportion of
useful quality grain. Note and record unfilled, immature, or green grains that
would be rejected during subsequent processing. Then the total of these plus
extraneous matter should be determined and the estimated total weight subtracted
respectively from the main threshed sample and the hand-stripped
material. The good hand-stripped grain would normally be lost, and the loss is
the percentage ratio of this to the total good grain, hand-stripped plus normally
  Losses due to scattering and spillage, which may occur with certain threshing
procedures, would be evaluated separately by recovering scattered or
spilled grain from known or controlled amounts of threshed grain or by weigh-ins
and weigh-outs if these are known or can be determined.
THRESHING LOSS 2: Damage to Grain
  The method to be followed for estimating grain damage during threshing is
basically the same as that for any other processing stage: One must standardize
all other processing steps leading to the final product and do the threshing by
the normal (local) method and by an optimal method which will give maximum
yield of undamaged grain.
  As with estimating loss with the straw (threshing loss 1 above), the estimator
selects random bundles of cut grain. These are randomly divided into two lots
of approximately equal weight. The methodology consists essentially of
weighing initially and at the end to compare the traditional (or any other
processing procedure) with a processing procedure that gives 100% recovery.
Lot 1 is threshed in the manner under evaluation. This may include a final
hand-stripping, depending on local custom. The threshed grain, including dry
hand-stripped, is bulked.
  Lot 2 is hand-stripped carefully and bulked. (Note: Subsamples of each lot
may be taken if laboratory equipment is available.) The separate samples are
processed carefully to avoid loss or damage through the locally used processing
system (cleaning, parboiling, drying, or milling) if this is a batch system in
which the samples can retain their identity. The products are then analyzed for
broken grains and damaged grains. This is especially important for rice, which
is desired as a whole grain, and grains such as red sorghum which undergo a
two-stage grinding system wherein bran or husk is first removed from the
whole grain before grinding.
  If local labor is available, separation of whole from broken grain may be
performed by the local method (eg, hand-winnowing): The out-turn of whole
grain is calculated and the results for threshing (by one or more local methods)
compared with those for hand-stripping.
  If the identity of the samples would be lost by processing through the local
system (large dryers or large continuous mills), then subsamples should be
taken and processed in the laboratory.
MAIZE SHELLING LOSS: Loss on Cob or Core
  The method used is basically the same as for threshing: Random samples of
cobs are taken and the grain is shelled by the method under test. All the grain is
collected and weighed and a sample taken (sample 1). The grains left on the
spent cobs are hand-stripped and weighed and a sample taken (sample 2).
Moisture content of the two samples of grain is measured with a moisture
meter and, if necessary, an adjustment made to the weights. The percentage
ratio of the hand-stripped grains to the total is the percent loss. The two
portions of grain must be kept separate for the next loss assessment, grain
  Losses of insect-damaged, mold-damaged, or stored grain may be different
from the losses without such added factors. It is therefore necessary to define
the situations being measured and the condition of the grain. For example,
losses during the shelling of maize may actually be due to the release of frass
(insect chewings, excreta, cast skins, insects and insect fragments) at the time
of the shelling process, or the intentional removal of weevils or musty grains
(see next section).
  Many mechanical and hand shellers cause damage to the maize kernels
which can result in a loss of food.
  Shelled grain from the previous loss assessment, but not the hand-stripped
material, is sampled and a representative subsample of at least 200 grains
obtained. These grains are examined visually for cracks and scratches, and the
number of damaged grains counted and the total expressed as a percentage. It
is important not to count insect-damaged, mildewed, or shrivelled grains, only
damage caused by the sheller. To check this, a parallel sample of cobs should
be carefully stripped by hand and at least 200 grain samples also examined. An
example of the use of these methods is given in Fig. 7.

pgl7x70.gif (600x600)

  In this section the grain under consideration will be raw paddy rice, though
the methodology can be applied in principle if not in detail to other grains and
to parboiled paddy. The method is based on that used by a TPI team in
Malaysia and was used to compare three drying methods: 1) yard (sun), 2)
batch (Lister), and 3) continuous.
1. Yard (Sun) Drying.
  The method for dryer-induced losses based on a laboratory milling operation
may be performed in a mill yard, on the highway, or in the farmyard.
(a) Method for Use in a Rice Milling Laboratory on Small Samples
  As the bags of one variety of paddy arrive at the yard, they are sampled (see
Chapter IV) and blended. The composite or bulked sample (of about 1 to 1.5
kg) is then dried carefully. ((4)"Carefully" means dried in a laboratory
dryer with forced air convection
at 1.5[degrees] to 2[degrees]C above ambient air so as to bring the samples to
an equilibrium moisture constant (ie, about 14%) in not less than 36 hr.)
Meanwhile the paddy will be dried in the usual way
and, when dry, rebagged for storage prior to milling; a further sample of
about 1 to 1.5 kg is then taken. The two samples (before and after drying) are
placed in cloth bags and, as soon as possible after sampling, are dried carefully4
down to approximately the same level of moisture. A small flatbed dryer
with a flow of air only slightly (1.5[degrees]C) above ambient is suitable. Drying to
around 14% moisture content should take 6 to 12 hr. After a further three to
five days to equilibrate (stabilize), the samples are checked for the exact moisture
content and milled.
  The best procedure is to use a standard laboratory mill (huller plus cone).
Each process should be done in a standard way and in accordance with
manufacturer's instructions. The rice will be separated from husk and bran in
the laboratory mill. Whole and broken grain proportions are then measured by
separating on a hand trier (indented tray) or a small rotary trier (indented
cylinder) and weighing.
(b) Method for Use in Mills
  If a laboratory mill for small samples is not available or if the data are
required for mill use, the following procedures can be used: Large samples (1
to 2 kg) are taken from representative bags being emptied onto the drying
yard, so that the total bulked sample weighs at least 25 kg. This sample is then
dried carefully(4) in a small batch dryer (as above). A sizable sample of the dried
paddy from the yard is also obtained and the two samples dried and equilibrated
as for small samples. If parboiling is customary, it should now be
performed in a standard manner, suitable to the variety and district. The
samples are then milled in a small commercial mill of local type (Engleberg,
cone, "modern" )and the total product collected. Many small mills that operate
on a toll basis are suitable for this purpose. The product is separated into
whole and broken grains. If possible this should be done on a separator (some
small mills have these and will provide the product fractions already separated).
Alternatively, local labor may hand-winnow to separate. The fractions
are weighed and the out-turn of whole grain calculated as before (a).
Note: While it may be inconvenient to deal with large samples, use of a
commercial rather than laboratory milling system ensures that the results are
directly applicable to the local situation.
2. Batch Dryer.
  Samples are taken from at least four places near the top and four near the
bottom of the drying bin with good distribution across the bin area. Samples
must be taken as the paddy is entering the bin (6 to 12 in. from the bottom) and
just before the bin is fully charged.
  Samples are taken from approximately the same sites as the bin is emptied.
Each sample is kept separate in a cloth bag and is not blended with the other
samples. There will thus be at least eight samples before and eight after drying
for each batch. The samples are dried uniformly and carefully(4) on a laboratory
dryer as for (a) above, stabilized three to five days, milled on a laboratory mill
as in (a) above, and the results tabulated. It is important to compare drying
damage on samples from each part of the bin; that at the bottom is frequently
overdried and that at the top is frequently re-wetted by transfer of moisture
from below, with consequent high breakage during subsequent milling.
  The mean figures for brokens for input and for the batch-dried paddy
indicate the average damage caused by the drying process. As a guide to
maloperation, the differences between brokens obtained from samples of
dried paddy from different parts of the bin are important; the mean figures for
a whole dryer are not.
3. Continuous Dryer.
  With a continuous dryer, sampling of input and output is performed periodically.
Samples (1 kg)(5) should be taken every 15 min over a period of at least
1.5 hr. Larger dryer output may require larger samples. If input is varying,
sample the same grain in and out of the dryer. ((5)Appropriately larger
samples must be taken if a small commercial-type mill rather than a
laboratory unit is to be used, ie, sample size must be matched to test
  As with batch dryers, it is better if the samples are kept separate. Samples in
cloth bags are placed, as soon as possible, in the laboratory dryer (see 1.a).
When dried to 14-16% moisture, the samples are kept for three to five days
before laboratory milling. The proportion of broken grains should be constant
if the wet paddy is of constant quality and the dryer is running consistently; the
difference between the mean figures for input and output samples gives a
measure of the damage caused during drying.
GRINDING LOSS AS BRAN: Comparative Assessment by Weight
  Grains such as wheat, maize, and sorghum may be ground in stone mills, in
mortars, or in steel plate or steel roller mills. If the objective is not only to
provide a flour or meal but to remove bran, the optimum milling will remove
all the bran and leave all the endosperm (inner part) of the grain as flour. The
separation of bran from flour is usually done periodically during the grinding;
sieves of cloth are frequently used. Winnowing (air classification or purification)
may also be used. The bran and other offals will usually be used for
animal feed. The problem in assessing the yield of desired product (flour) is
that of comparative weighing of various mill fractions over measured time
periods. Quality of flour (eg, amount of bran) also may be a factor.
   Standard procedures have been evolved for milling wheat on an experimental
mill, but this equipment is extremely expensive and of little use for other
grains. The methods proposed below may be used to compare the yields of
acceptable flour derived from different varieties of the grain or to compare the
performance of different operators, and to obtain information on other factors.
1. Comparative Measurement of Milling Yield by Variety
  The method selected for milling must be that which is used locally. The
ultimate test is milling yield; whatever losses occur must be measured by a
standardized procedure.
  A number of different operators (eg, women if they are the traditional
operators) are required, each with a mill (querns or hand-cranked plate-mills)
of the same type and size.
  A portion (about 5 kg) of each variety is given to each operator. Each
sample is then milled by sieving or winnowing the product to obtain a flour or
meal considered by the operator to be of the usual standard desired in the
community. The total weights of grain, flour, and bran are weighed, samples
are taken in sealed bottles for laboratory moisture content measurement by
oven-drying; and the weights are converted to 15% moisture content basis (or
dry weight basis).
    Weight flour (15%)
    _________________  = extraction rate (milling yield).
    Weight grain (15%)
  The average of the milling yield for any given variety obtained from different
operators is calculated. Provided that the operator yields for each variety
are similar, the method will give an indication of practically attainable milling
yield. This same procedure can be run on a commercial mill.
2. Comparison of Operators
  With the above procedure (1), a series of milling yields is obtained for a
given variety of the grain for a number of operators. If the products obtained
were all acceptable to users, the operator attaining the highest yield can be
employed to improve the communities' out-turns of edible flour or meal.
3. Comparison of Mills
  The procedure of (1) is followed with any one variety to compare the milling
yields (extraction rates) for a series of mills.
4. Insect Damage
  A constant volume of each grain sample is weighed and milled by a standard
milling process and input-to-output of food and nonfood product measured.
Insect-damaged grain will give a lower yield of flour than undamaged grain.
  There are many different milling systems in use, but these may be classified
as being either one- or two-stage, and either batch or continuous. In the first,
the hulling and polishing are carried out in one machine; in the second, separately.
One-Stage Batch Processing (eg, Engleberg Type Huller)
  The bag of dried paddy to be processed is sampled and the sample of about
0.5 kg placed in a sealed bottle or plastic bag. The bag of grain is weighed and
the moisture content of the grain measured. The paddy is then processed
through the huller and the product collected in the customary way. A representative
sample of the product is taken. Subsamples (100 g) of the input paddy
are then milled on a laboratory mill. The product is separated into husk, bran,
and polished rice, and the rice is separated on a hand trier (indented tray) or a
small rotary trier (indented cylinder) into wholes, halves, and points. The
sample of mill product is separated likewise. The relative proportions of whole
grains and total grain are compared; the efficiency of the commercial mill can
then be related to that of the optimal laboratory mill and the relative loss
One-Stage Continuous Processing
  As the paddy flows from the hopper or storage bin into the hopper of the
huller, a sample (about 100 g) is taken every minute for 10 min. A sample of
the product flowing from the output side of the huller is sampled, again a
sample of 100 g is taken every minute, beginning about 0.5 min after the first
input sample has been taken. The two bulked samples (labeled "in" and
"out") are taken to a laboratory and there analyzed by the same procedure as
for the batch process.
Two-Stage Continuous Processing
  As typical of this system, the "modern" rice mill consists of rubber roll
shellers and a series of cone polishers with, perhaps, a finishing brush polisher.
Separations are carried out at each stage and after each polishing (usually at
least two, frequently four). Skilled operators judge visually the product quality
at each stage and also the effectiveness of the separation of product from
by-product. Quantitative estimates of machine effectiveness may be measured
by sampling on the input and output sides of any machine or battery of
machines, processing the input sample by a standard optimized laboratory
method, and comparing products for yield (out-turn) and quality (percent of
whole grain).
Hulling (suggested basis for a method)
  Many mills have two hullers in parallel and some will have a "return huller"
for the 10% or so of paddy unhulled in the first pass. It will not be possible to
sample the whole product of the huller system, as the material passing back to
the return huller has already been through the first huller unit and has been
separated from brown rice and husk. Samples must therefore be taken at the
entry and exit to each individual machine; if the mill possesses three hullers,
each must be sampled separately.
  Representative samples (250 g) are taken from the flow of paddy to the
huller on a regular basis (eg, every minute), and from the product as it flows to
the first separator (likewise every minute) for about 10 min. It is important to
obtain a truly representative sample of product; once it has reached the chute
leading down to the separator some separation can occur. If possible, the
sample should be taken immediately below the rolls.
  The well-mixed samples are subsampled for triplicate laboratory testing; the
paddy is milled in a laboratory sheller.
  The products from the plant and laboratory mills are then examined quantitatively.
The ratio of weights of total brown rice gives a measure of the
effectiveness of the hulling attained in the plant compared to that in the
laboratory. More important is the comparison of the ratio of weights of broken
to total grains of brown rice. If the plant huller is giving a higher proportion
of broken grains, then wear or a wrong setting on the rollers should be
Polishing (whitening)
  Whether one is endeavoring to measure losses over the whole polishing
system or for each machine, the method to be used will be the same: As with
other unit operations, samples are taken of the feed to a machine or series of
machines and of the product therefrom. The sample of brown rice should be
milled carefully in the laboratory to the same degree of milling as that of the
machine(s) in the mill. The out-turns of whole grain are measured and compared
and the loss in the mill assessed.
Note: Whether it is, in fact, possible to set up such a loss evaluation system
remains to be seen. The principal difficulty lies in using a laboratory polisher
in one pass to give the same degree of milling as the battery of polishers in the
plant and yet also give minimum breakage.
                           VI. STANDARD MEASUREMENT
                        A. Preamble to the Methodology
                             K. L. Harris and C. J. Lindblad
Definitions((6)See also Chapter II, Section A.)
  There is a need to define certain terms and concepts before proceeding to the
working methodology.
  This effort deals with removal of food grains from the direct human food
chain which, especially in developing countries, is the fundamental energy
(calorie) basis of the human diet. The rice weevil consumes rice when living in
the kernel. If the kernel is weighed before and after it is bored, it will have lost
weight. If the larva or adult is still present when the kernel is eaten, less weight
is lost. No consideration is given to a proportional change, if any, in protein
accompanying the feeding. The relevance of the insect presence depends on its
fate. If it is cleaned out it is loss; if it remains as food, it is weighed as food.
Whether insects are eaten or whether the frass is sifted out or falls from bagged
grain is sometimes fortuitous, sometimes purposeful. It varies with the season,
with the culture, with hunger, or plenty. While the decision to eat may be more
socioeconomic than scientific, use or nonuse as food in the specific situation is
the controlling factor in these procedures.
  In this manual pilferage is not considered to be a loss. It is a transfer of
ownership as is spillage when it is used as sweepings in lieu of, or in addition
to, wages.
Fungal Damage
  It is anticipated that the quantification of weight loss when the loss is due to
fungal damage will depend on local practices in the use of the damaged material.
People accept or reject damaged kernels as local custom and hunger
dictate. One purpose of this manual is to set forth standardized procedures so
that measurements in one country can be compared with measurements made
elsewhere. Therefore, in each situation acceptance-rejection limits should be
defined in terms of a widely used language. Despite such difficulty, judgment
limits based on information obtained from interviews must be quantified.
Processing Losses
  Grain removed from the direct human food chain is a loss. Thus milling
losses that become animal feed would appear as a loss although reentering
down the pipeline with a reduced calorie and, perhaps, improved nutrition
input. This "feed" as against "food" use needs to be recognized and
described in any situation where it is a factor.
  This manual generally accepts Bourne's (1) definition of postharvest as the
point at which grain, separated from the plant stalks or root, is bundled for
field drying or placed in a container in which it is moved or held, or both. It
can extend earlier, however, to include the time during which the mature crop
is held in the field for storage or drying.
  This manual does not cover losses in food after it reaches the point where it
is being prepared for cooking or for direct consumption, even though there
can be serious losses in the hands of the ultimate user. In the United States, for
example, this may be the most important site. However, estimates and prevention
of these losses are so dominated by cultural habits and preferences that
in-depth anthropological inputs are required which are not usually within
grain loss reduction biology-technology.
Separation From Other Factors
  This report anticipates that grain losses will be considered in isolation from
other food-availability factors in the areas studied. It is proposed that there is
no present need for guidelines that include such sophisticated concepts as how
the availability of fish and meat influences the losses, and need to control
losses, in staple grains.
Rapid Laboratory-type Procedures
  None of the shortcut tests such as presence of numbers of adult insects,
amount of frass, or insect emergence holes are sufficiently accurate when used
alone for anything more than loose approximations. "Loss" should be a measurement
of actual grain Substance removed from the food chain. Techniques
for basic statistical concepts are covered in a separate section.
  How to quantify losses has been the subject of detailed investigations by the
Tropical Stored Products Centre, England, and has been assessed by the
Group for Assistance on Systems Relating to Grain After Harvest.((7)The
acronym GASGA now stands for Group for Assistance on Storage of Grain After
Harvest.) Papers
listed in the Bibliography at the end of this section give a definitive appraisal of
these losses. From these review papers, from the original published material,
from discussions with acknowledged experts, and from first-hand field and
laboratory experience come the following conclusions on techniques for
measuring losses:
  All of the U.S. Food and Drug Administration-generated procedures are too
time-consuming, require a laboratory setting, require difficult-to-standardize
judgments, are on too small a sample size, or have too variable a relation to
grain weight loss to warrant use in determining grain losses. These are the exit
hole test, acid fuchsin egg-plug test, berberine sulfate fluorescent stain egg
plug test, gelatinzation with sodium hydroxide, and examination for internal
insects. Radiographic (X-ray) examinations require expensive laboratory-based
apparatus, and are time-consuming and difficult to standardize. The
Ashman-Simon Infestation Detector has similar liabilities.
  Examinations for insects on the surface of the grain, weighing insect frass
(dust from insect chewings and excrement), and various procedures to visually
detect damaged grains and count and/or weigh them have been given field
trials in developing countries. There is a positive correlation between damage,
insects, and frass with some loss quantifications possible and the 1970 IBRD
report suggests their use in making rapid assessments.
  Some confusion exists concerning the application of these procedures in
quantifying actual losses. Their use in actual test situations and positive correlations
to weight losses have been taken by some to indicate a practical degree
of precision to routinely determine weight losses. Such is not the case. They
cannot be so used unless the biological and physical characteristics of each
assessment situation are completely understood. If lots of grain have the same
histories, then their frass-to-loss relations will be similar and may be used to
survey them all on a comparative basis. However, if some have been moved
(and frass is lost), or some have lesser grain borers (produce much frass), or
some have weevils that make exit holes and some have moths that hold their
frass in webbing, or the surface insects have been removed from some lots and
not others, then any standardization between lots, regions, grains, and countries
becomes a new scientific investigation, not subject to rapid comparisons.
  However, all of these procedures are of value in a rapid visual and discussion
appraisal of a situation to come to a personal judgment. Their precision
as indicators of actual losses depends on the expertise of the user. This is
discussed in Chapter 11, Section D.
Rapid Judgment-Based Procedures]
  As these estimates with some facts by knowledgeable persons have discovered
immediate and urgent needs that could not be met in any other way, they
have served many purposes. However, as they have been simple guesses or
preconceived opinions for special purposes, they have no validity as determiners
of losses. True guesstimates have a valid role in reaching rapid judgments
that may suffice for some purposes or precede more accurate evaluations.
Biased Estimates
  Although not germane to the present effort, the practical effect of many of
the biased figures should not be underestimated. Many have been used to draw
forth budgetary support for grain storage and marketing research, build storage
structures of sometimes useful value, draw international attention to
sometimes real and sometimes imaginary needs, and build local and national
stockpiles that have both fed people and wasted grain to the ravages of biological
and physical factors.
Traditional Local Estimates
  These are especially useful in getting one's bearings on local situations.
Interviews should not be passed over lightly. They need to be done with care,
as discussed elsewhere in this manual, assessing the point of view and biases of
the giver of information, what the figures are based on, and local meanings of
such basic terms as "loss" and "percent."
  When reinforced by on-site observations or measurements, such estimates
may be especially useful in obtaining a picture of local conditions, extrapolating
to larger areas, and seeking out specific examples and situations. There are
times when local people can make quite accurate comparisons between conditions
found in grain as it goes into and is taken out of storage and on actual
wastage to insects, birds, and rodents.
On-Site Expert Judgments
  While this type of rapid appraisal can be used only by experts to assess
percentage or weight losses, its use should not be underestimated.
  In making such judgments, one needs to consider how local conditions
affect the physical and biological potential for losses. For example, transport
in damaged bags or makeshift wagons with visible spillage indicates an obvious
loss situation.
  Dry conditions spell trouble for insects. At 12% moisture or less, grain
insects have a more difficult time feeding and reproducing. By 10% there are
serious living problems, and if there is evidence of an arid 6 or 8%, then grain
losses to insects are minimal.
  Absence of visible insects or damage after six or eight weeks of storage is a
good indication that there will be few insects for the next few months also.
  The habits of many rodents are well known. Whether stores are open or
closed to them, and whether harborages or needed water are available can be
readily ascertained.
  Losses to rats can be predicted from the nature of the local ecological
system. The problem may be more difficult with mice and other small rodents.
  Short-term storage, good sound bagging, well-constructed transport vehicles,
strict weigh-in/weigh-out control with accompanying records, the use of
insect, rodent, bird, and fungal control procedures, and low temperatures all
point to minimal losses. Low or high temperature can be of overriding importance.
Rice harvested in September in temperate climates may go into natural
cold storage before insects make even a minimal start. Grain held under metal
roofs or in bags in the sun at over 55[degrees]C will have no active insect losses.
  On the other hand, while high moisture, active insect, rodent, and bird
depredations, and visible mold or heating from microorganisms clearly indicates
trouble and potentially heavy losses, the extent of the losses is determined
with difficulty by even an expert.
Production and Consumption Figures
  Production and consumption figures have often been suggested as a means
of assessing losses, the difference between what is produced and what is consumed
being loss. Unfortunately, accurate figures at either end of the system
are available only in the most sophisticated and developed situations, and the
approach is of small practical value in many developing nations and local
developing-country locations.
  Changes in volume and weight due to moisture need to be explained. Grain
harvested at 21% moisture dried to 15% by mechanical means or aeration has
lost weight but not food value.
  Measurement of moisture changes requires the use of meters or drying
ovens. Weight changes need to be determined by sensitive devices. Use of
moisture meters and scales or balances requires such devices and a degree of
expertise in their use that may necessitate some basic training. Moisture meters
are discussed in Appendix C.
  Overall statistical concepts are presented in Chapter IV. It seems reasonably
safe to anticipate that 75% confidence limits of [+ or -] 5% would, for the present,
be as much, or perhaps more, than can be generally expected. However, as yet,
there is no fixed gauge as to what constitutes reasonable accuracy. The amount
of method variation that may be expected to occur in different commodities,
ecological zones, parts of the harvest-to-consumer pipeline, and types of damage
by different individual or mixed types of losses are subjects that require
clarification in and before any survey appraisal. The first field appraisal
should bear these and other factors in mind, particularly as the desired confidence
limits influence the duration and expense of the assessment.
                               Literature Cited
1. BOURNE, M. C. Post harvest food losses -- the neglected dimension in increasing the world
   food supply. Cornell International Agriculture Mimeograph 53 (1977).
ADAMS, J. M. Storage loss assessment techniques, a biologist's view. Trop. Stored Prod. Centre
ADAMS, J. M. Report on post harvest loss assessment in durable produce, with particular
   reference to methology. Trop. Stored Prod. Centre (1976).
ADAMS, J. M. A guide to the objective and reliable estimation of food losses in small scale
   farmer storage. Trop. Stored Prod. Inf. 32 (1976).
ANONYMOUS. GASGA Seminar on Methodology of Evaluating Grain Storage Losses. Trop.
   Stored. Prod. Centre (1976).
HARRIS, K. L. Evaluation of grain storage losses. Report of the International Bank for Reconstruction
   and Development (1970).
                                  CHAPTER VI
            B. Losses Caused by Insects, Mites, and Microorganisms
                      J. M. Adams and G. G. M. Schulten
  Insects are a major cause of postharvest grain losses. By boring within the
kernels and feeding on the surfaces, they remove food, selectively consume
nutritive components, encourage higher moisture in the grain, and promote
the development of microorganisms.
  Methods for detection of internal insects have been summarized earlier in
this chapter. Methods given in this section are for determination of losses to
the grain itself and are of three types:
  1. Determination of the weight of a measured volume of grain (see Methods
A and B1). In this case the loss in weight in samples taken over a known time
period is a reflection of losses caused by insects or microorganisms, or other
factors. Judgment as to cause of the loss is a second and necessary step in the
  2. Separation of damaged and sound kernels and determination of their
comparative weights calculated in terms of the whole sample (see Method B2).
  (In both 1 and 2 above, it is usually necessary to obtain a baseline sample of
the condition of the grain at the beginning of the test period or to conduct tests
to estimate the baseline condition in order to determine the real losses at that
point in the pipeline.
  3. Determination of the percentage insect-damaged grain and its conversion
into a weight loss using a multiplication factor (see Method B3). (This method
also gives an approximate figure for use in preliminary surveys.
  In all of the methods, prior to analysis, the grain sample should be sieved or
winnowed, or both, to remove dust and insects. Use the sieve and sieving/
cleaning technique commonly used by the local farm/merchant/consumer for
removal of such fractions that would be normally discarded as inedible prior
to further processing.
Determination of the Original Condition of the Grain
  Since the weight-to-volume method is based on differing weights for different
levels of loss, it is necessary to obtain a baseline point, by sample or
calculation, from which it is possible to compare all future measurements.
This baseline needs to be in the form of a curve covering all of the grain/
moisture conditions to be found in the particular grain situation because some
grain volumes change significantly, and most often regularly, at varying moisture
  The curve is obtained from analysis and calculation of a baseline sample.
Determination of the baseline condition is essential so as to have a fixed
reference point with which to compare losses incurred during storage. if it is
not possible to obtain this sample until after storage or the process under study
has already begun, a visibly undamaged sample should be taken and analyzed
as early as possible. This should be split into three replicate subsamples and the
measurement required by the appropriate methods 1, 2, or 3 applied to each
subsample. Each subsample should then be placed in a jar covered with
muslin, to prevent insects entering or leaving, and kept for four weeks. At the
end of this period, the jars should be examined for insects and damage. If
there is no damage in any jar, then all three replicates can be used to calculate a
value. If there is damage in one, this must be discarded; if two have damage,
both are discarded; and if there is damage in all three, then take the sample(s)
with 5% or less damaged kernels. If the damage is above 5%, assistance will be
needed from an expert in determining the appropriate correction factor.
Method for Baseline Determination
  A sample of approximately 5 kg is either taken from every farmer's store if
they are being treated as individual case studies or, if there are distinct grain
varieties under study, a representative sample of at least 5 kg is taken for each
variety, assuming that they are fairly homogeneous. If any of the varieties is
not uniform (does not have a standard weight-to-volume variation with
changes in moisture due to intravarietal variations of the local grain(s)), then
either each lot of stored grain must be treated individually or expert advice
must be sought.
  This large sample is sieved in the laboratory. The bulk sample is subdivided
into five replicate subsamples. The moisture content of a representative subsample
is measured. The range of moisture content which might be expected in
the field over the storage season is determined either from locally available
data or by approximation (a normal range that fulfills most purposes is
8-18%, depending on climatic conditions). The weight/volume relationship is
taken over the range as follows: the range is broken down into five equal steps,
eg, if it is 10-18%, this will be 10, 12, 14, 16, 18. If small, perhaps 1%, steps
such as from 8-12%, this will be 8, 9, 10, 11, 12%. One subsample will have a
moisture content near to one of these figures and the moisture contents of the
other subsamples will have to be changed either by drying or wetting, as
follows, to cover the range.
  Drying down to a moisture content. This should be done with the grain in a
shallow layer either in a warm, dry place with a current of air passing over it
but protected from insect attack or, preferably, in a ventilated oven in shallow
trays at a temperature not exceeding 35 [degrees]C. Its moisture content should be
checked at regular intervals by allowing a sample to cool and measuring its
approximate water content. When it has reached the required moisture content,
it should be placed in a sealed container to cool and the moisture content
should be measured accurately. As a rough guide, a small sample of known
weight can be placed on a dish in the oven and its loss in weight checked.
  Wetting up to a moisture content. This requires addition of a calculated
weight of water to the grain to bring it up to a required moisture content. The
weight of water required is given by the formula:
         Weight of water to be added (g) = weight of grain
     x Required % moisture content - initial % moisture content
                    100 - required % moisture content
  For example, if we have a subsample of 1,000 g of grain at 12% moisture
content and require it to be at 16% moisture content, the calculation is:
      Weight of water = 1000 16 - 12           4
                             --------  = 1000 -- = 47.6 g.
                             100 - 16         84
  This can be weighed out or, since 1 g of water occupies 1 ml, it can be
measured out as a volume. Water is added to the grain in a sealed container
with sufficient headspace for mixing, and mixed well. It is left for two weeks to
condition, but vigorously shaken daily. For moisture contents over 16%, the
container should be kept at 5 [degrees]-10 [degrees]C in a refrigerator to discourage mold
growth. At the end of the conditioning period, an accurate moisture content is
determined for each subsample.
  There are now five subsamples of grain at different moisture content for
each variety. For each subsample the weight that occupies the volume measure
(test weight container) should be determined by filling the container (see Fig. 8)

pgl8x86.gif (600x600)

according to the instructions provided with the apparatus and then pouring
out the contents and weighing it to the nearest 0.1 g. This should be done three
times for each subsample and a mean result obtained.
  There will now be five mean weights for each variety at five accurately
measured moisture contents. Each of these weights should then be converted
to dry weight as follows:
     Dry weight = weight of grain x 100 - % moisture content
  For example, if the volume of grain in the test weight container weighed 800
g and had a moisture content of 15%, then its dry weight is:
     Dry weight = 800 x 100 - 15          85
                        -------- = 800 x ---  = 680 g.
                           100           100
  This is done for all subsamples so as to obtain a set of dry weights for each
moisture content. A graph is now drawn of the dry weight against the moisture
content, for example:
     % m. c.       10.2          12       13.9          16        17.8
     Dry wt.        700         680        650         620         600
  From this a reference line can be plotted of dry weights as determined by
measuring the actual moisture content and test weight at the time a test is
made. This graph is then used throughout the rest of the sampling period to
represent the dry weight of sample at any moisture content as if it had not been
damaged in store.
  A curve must be made for each variety or area-cultural situation (see Fig. 9).

pgl9x87.gif (600x600)

Loss Measurement Procedures
METHOD A -- Standard Volume/Weight Method for Damage by Insects and Microorganisms
  After preliminary laboratory work for the baseline figure, the measurements
can be made in the field or laboratory.
1. Test weight apparatus for obtaining a standardized volume of grain.
2. Balance, such as a triple beam balance, capable of measuring 1.0-1.5 kg
   accurate to 0.1 g.
3. A moisture meter capable of measuring to 0. 1 and calibrated for the type
   of grain being measured.
4. A suitable size of grain sieve for the removal of insects, dust, and any other
   material that would normally be removed prior to further processing.
5. Plastic sample bags and a liquid fumigant such as [CCl.sub.4] to retain samples
   for examination at a later date.
  A well-mixed sample, taken from the store, is first sieved by a locally appropriate
method and the weight of sievings are counted as a loss if they are not
used locally or calculated back to the weight/volume if they are used.
  The moisture content is measured.
  The weight occupying the volume container is measured. This is repeated
three times and a mean taken. This weight is converted to dry weight using the
moisture content and formula for dry weight (see derivation of Fig. 9).
  The graph is used to find the dry weight of a sample at the same moisture
content taken at the time of storage. For example, if the moisture content of
the farmer's sample was 12%, then referring to the example, Fig. 9, the dry
weight would be 680.
  The weight loss in the farmer's sample is then calculated as follows:
     % of weight loss - dry wt. from graph - dry wt. in sample
     --------------------------------------------------------- x 100
                        dry wt. from graph
For example, if our farmer's sample at a moisture content of 12% had a dry
weight of 600 g, then as the dry weight on the graph for 12% moisture is 680 g,
the loss would be:
     % dry weight loss = 680 - 600 x 100   80 x 100
                         --------------- = -------- = 11.8%.
                               680            680
This is the dry weight loss, which by definition excludes moisture content
Sources of Error
  The standardized method of obtaining the volume attempts to eliminate
variations in packing, but with grain samples containing very high levels of
damage, some of the grains may become crushed and lead to inaccuracies
especially with small grains that may be sieved or winnowed out or crushed so
that their insect- or microorganism-caused emptiness is not detected. In this
case they may have to be picked out and losses otherwise estimated. Conversion
factors change in the course of the storage period from high to low, due to
increased severity of damage to the already-damaged grains.
  The admixture of an insecticidal dust to shelled grain increases friction
between grains and will reduce packing and hence the weight per unit volume
will be less. Therefore, the weights for treated grain must not be compared
with weights obtained for untreated grain.
  For paddy, the effect of moisture content on the dry weight occupying a
given volume is negligible, so within a range of 5% moisture there is no
requirement for a predictive graph.
  Rice (as distinct from paddy) would best be measured by out-turn of the
  Lumps of, or otherwise webbed-together, grain can add weight. However, if
the lumps are picked or sieved out by local custom, they should also be picked
out and the kernel loss estimated.
  Since little is known about methods for determining losses in insect-damaged
millet which, in effect, are hollowed shells, and since no procedures
have been satisfactorily described for picking out and weighing of insect-infested
millet, this grain presents a real problem not yet resolved by this
current methodology.
METHOD B1 -- Modified Standard Volume/Weigh Method When a Baseline Cannot be Determined
  The standard volume/weight method as described under METHOD A is
presently the most reliable method of loss determination. There are, however,
situations where this method cannot be used without modification. It may also
be difficult to obtain reliable moisture content determinations in some cases.
  It is often necessary to make loss estimates in the middle of the storage
period when no baseline has been previously determined. It also frequently
occurs that in a rural area different varieties of grain are grown under different
conditions, such as with or without fertilizer, or on poor or good soils. This
may affect the size of grains and, consequently, the volume/weight ratio.
  Application of insecticide dusts may also affect the settling of the grains in
the standard volume and increase the volume occupied by the grain.
  Because of these various conditions, a separate baseline may have to be
determined for each individual farm or storage situation. This is often impossible
to achieve between harvest and storage.
  The standard volume/weight method should be used but an artificial baseline
should be prepared by selecting undamaged samples from the grain
present in the store at the time of loss determination. The loss is the difference
in weight (expressed as a percentage) between the undamaged and the damaged
sample. Conversion for moisture need not be used in this case since the moisture
content will be approximately the same.
  Experience with this modification of the method is still limited. For maize
ears stored with husks, it is possible to select a number of undamaged ears, to
shell these, and to use the grains to determine the baseline. With other grains,
it may be more difficult to obtain an undamaged sample, especially in cases of
heavy insect infestation.
Sources of Error
  Unreliable results may be produced if, during selection of undamaged grains
from the stored grains, there are hidden internal infestation, preferential feeding
and egg deposition by insects in grains of different sizes, and a difference
in moisture content.
  To overcome the problem caused by hidden infestation, the same procedure
for obtaining an undamaged sample as indicated for the normal standard
volume/weight method can be followed.
  Insects do not often feed or oviposit on grains at random but, depending on
species, they may show a preference for smaller or larger grains. There is then
the risk that in selecting undamaged grains, a particular grain size may be
selected which is less liable to infestation than grains of another size. Grain size
obviously affects the volume/weight ratio. When undamaged ears are selected,
there is the possibility that smaller ears (with smaller grains) may be unintentionally
selected, since smaller ears are less often infested than larger ears due
to a better husk protection. The only way to reduce this error is to take the
undamaged sample as much at random as possible. In addition, a sample must
be taken which is larger than necessary and, after good mixing, only a part of
the sample should be used for baseline determination.
  When the baseline and field sample are taken from the same part of the
storage structure, it is not usually necessary to determine moisture content
since differences between the two samples are likely to be small. The weight
difference between the two samples represents the actual loss. If there is doubt
about the homogeneity of the moisture content of the grains in store, the
method for baseline determination should be used.
  Insects prefer moist grains rather than dry ones. This behavior may cause
the baseline sample to be drier than the field sample. When this is suspected to
be the case, the method for baseline determination should be followed. When
this is impossible, the error can be reduced as much as possible by taking large
random samples.
METHOD B2 -- Count and Weigh Method
  There are many situations in which a loss estimate is required but where
there is only minimal equipment available and the baseline could not be determined
before the storage period. In addition, it is sometimes impossible to
determine a baseline for the standard volume/weight method because too
many grains have been damaged.
  This is essentially a method that takes a sample, separates it into undamaged
and damaged portions, counts and weighs each, and calculates the percentage
weight loss. It assumes that the undamaged portion is totally undamaged.
  Used for unshelled and mold-damaged grains, it provides a useful means of
estimating loss at moderate infestation levels with a minimum of apparatus.
1. Balance with a range of 0.5 g to 1.5 kg accurate to 0.1 g.
2. Tally counter.
3. Plastic bags and a liquid fumigant such as [CCl.sub.4] to enable retention of
  The grains are separated into undamaged and damaged categories, the latter
being separated according to cause. Grains in each category are counted and
weighed. The resultant data may be substituted in the formula below:
     % weight loss = (UNd) - (DNu)
                     ------------- x 100
                      U(Nd + Nu)
where U = weight of undamaged grains,
     Nu = number of undamaged grains,
      D = weight of damaged grains,
     Nd = number of damaged grains.
Sample Size
  Experience with this method is still limited. A sample size is recommended
of 100-1,000 grains. Besides its simplicity, the method has the advantage that
damage by different species of insects, such as Sitophilus, Sitotroga, Ephestia
spp., and Rhizopertha, can be measured. The method may also be used to
determine damage caused by termites, rodents, and birds.
Sources of Error
  Hidden infestation results in an underestimation of loss because grains that
have lost weight are included in the undamaged portion. When the grain is
heavily damaged, it may become so broken as to lead to counting errors.
  At low levels of infestation with the insects selecting larger or otherwise
nonrandom grains, the method is not dependable. At very high levels of infestation,
kernels may be so destroyed as to be not measurable. For example, in
maize ears at low infestation, often only the grains at the top of the car are
damaged because they are incompletely protected by the husks. These grains
are often the smallest of the ear. The only recommendation to reduce this error
is to take large samples.
  Since insects will sometimes select and infest larger kernels, any procedure
that compares the individual weights of kernels may result in a negative weight
loss finding. The selection of internally infested kernels and their inclusion and
weighing as undamaged can also result in negative loss findings unless care is
taken to recognize and account for these samples.
  A preference of insects for moist grains may confuse the relation between
weight loss and damaged grains as well. To reduce a possible error arising
from this behavior, the grains could be dried to the same moisture content.
METHOD B3 -- Converted Percentage Damage Method (For Use in Field or Laboratory)
  This method is suitable for insect damage only and provides a useful estimate
for quick appraisal of losses without needing equipment. It can be easily
used by unskilled but trained personnel.
  When grains are heavily infested, feeding by secondary pests and multiple
infestation may disturb the relation exit/weight loss and so lead to an underestimate.
Therefore, when possible, it is preferable to determine the conversion
factor instead of using those factors indicated below. It will be obvious that
the conversion factor can be ascertained in a sample at any time after the
sample has been taken as long as the sample is properly stored.
  When losses have to be measured in a large number of samples, originating
from cereals which were stored for about the same period of time and under
similar conditions (eg, regional surveys), at least some samples should be kept
for determination of the conversion factor.
  Although the converted percentage damage method is liable to the same
sources of error as the modified standard volume/weight method and the
count and weigh method, it has given very good results in practice.
  When earlier-mentioned methods cannot be used, it is recommended to use
the converted percentage damage method rather than guessing. With this
method, weight losses in cereal grains and pulses are determined in a slightly
different way.
1. Tally counter.
2. Plastic bags and a liquid fumigant such as [CCl.sub.4] to fumigate samples when
   determinations are done at a later date.
  The number of damaged grains is counted in the sample and expressed as a
percentage. This percentage damage is converted into weight loss by means of
approximate conversion factors as indicated below. This factor can be determined
separately for each individual situation or established factors can be
used. This loss determination is only applicable when the damage has been
largely done by insects which leave a clear exit hole in the grain (Sitophilus,
Sitotroga, and Bruchidae).
  A random sample of 100-1,000 grains is taken and the number of bored
grains is counted. This can be done immediately or within a few days after
sampling. When there are too many samples to be counted, it is recommended
to store each sample in a sealed plastic bag to which some liquid fumigant has
been added.
  The percentage of damaged grains is calculated with the following formula:
     Number of bored grains
  ------------------------------  x 100 = % bored grains in sample.
  Total number of grains counted
  This percentage is converted into a percent weight loss by dividing it by the
conversion factor (C) or multiplying it by 1/C.
  To determine the conversion factor, a random sample of 100-1,000 damaged
grains is taken which contains 10% or more damaged grains. The percentage
weight loss is determined with the count and weigh method, and the conversion
factor is calculated as follows:
      Number of bored grains
      ----------------------  = conversion factor.
      Weight difference in %
  The following conversion factors have been established in practice where the
larval stages develop within the grain, eg, Sitophilus species, Sitotroga cerealella:
       Maize (stored as shelled maize
              or as ears without husks)    % bored grains    x   1/8
       Maize (stored as ears with husks)   % bored grains    x   2/9
       Wheat                                % bored grains    x   1/2
       Sorghum                              % bored grains    x   1/4
       Paddy                                % bored grains    x   1/2
       Rice                                 % bored grains    x   1/2
  In pulses several well-defined exit holes may be found in one bean or pea.
When infestation is not too heavy, it can be assumed that each weevil consumes
about the same amount of food for its development. Therefore, in the
case of pulses the number of exit holes is counted and not the number of bored
beans (peas). For determination of the conversion factor in pulses, the same
procedure is followed as for cereal grains but the damaged sample must consist
of beans (peas) with one exit hole only. The conversion factor indicates the
number of exit holes which equals a weight loss of 1%.
  In the field sample, the number of exit holes has to be counted in 100-1,000
beans. This number is divided by the conversion factor and the percentage
weight loss obtained.
  A known conversion factor for cowpeas when bruchids are the cause of
damage is number of exit holes in 1,000 grains divided by 200.
                                  CHAPTER VI
         C. Losses in Grain due to Respiration of Grain and Molds and
                             Other Microorganisms
                       R. A. Saul, with K. L. Harris
  A mass of grain can be considered as a living organism that feeds on itself. It
is made up of the individual seeds which are hosts to the many microorganisms
of fungus, yeasts, and bacteria. It loses or gains moisture depending on its
moisture content and the ability of the surrounding air to absorb or release
moisture (relative humidity). For example, maize at 12% moisture in air of
75% relative humidity will gain moisture until it reaches 15%. If the grain
moisture gets high enough the grain will sprout. At lower moisture levels the
seed is essentially dormant and has a very low and rather constant rate of
  Microorganisms can grow under lower moisture levels than grain. They take
moisture from the air and use it for their metabolism. Yeasts and bacteria
require an atmosphere of 95% relative humidity or higher, while fungus grows
under conditions as low as 75% relative humidity.
  The rate of growth of the microorganisms is dependent on temperature as
well as moisture. Also, the extent of physical damage to the kernel is a factor
which influences the rate of growth.
  Growth of the microorganisms and the seed is at the expense of the seed dry
matter. The rate of growth is reflected in the rate of dry matter loss. When
quality is reduced to the degree that the grain is rejected, there is an additional
loss of quantity.
  Weight losses due to respiration of the grain itself are unimportant until the
moisture is so high that serious deterioration by microorganisms occurs. In
other words, when there are serious quantitative losses due to respiration, the
quality has so deteriorated that total, or kernel by kernel, rejection for food
use becomes the dominant factor, not losses in weight due to respiration. At
this point, determination of losses involves an appraisal of amounts of grain
rejected for food use.
  The conclusion must not be reached that if there are no changes in weight
that the grain is free of mold' toxins. Toxins are a separate matter. When
suspected they must be determined by special tests.
  Thus, there are two types of losses. One is the loss due to grain being
converted by microorganisms to carbon dioxide and water. The other loss
occurs when the grain (in its entirety or as individual kernels) is rejected as
food. Such rejection can occur because of an obvious discoloration or odor, or
because of the more technical knowledge or implication that harmful substances
(mycotoxins) are present. In the latter situation, one must determine
the amounts of grain rejected for food use.
  Any visual survey by locals or outsiders on what an individual rejects or
accepts becomes a difficult assessment. It needs an input of all of the principles
of measuring subjective values, bearing in mind that bias is eliminated with
difficulty and that all elements of bias are probably not completely known.
Loss Measurement by Standard Table Based on Time, Temperature, and
  It is the nature of molds, yeast, and bacteria to reduce organic material to
simpler organic compounds or even to its inorganic form. That is, molds decay
the grain and, if conditions are favorable for the growth of mold, then they
will destroy the grain.
  Long before the grain is completely destroyed, it is made useless as food
because of the musty odor, discoloration, and possibly formation of toxic
substances. In fact, this will occur by the time 1 or 2% of the dry weight has
been destroyed.
  The rate of loss of dry matter due to mold growth depends on, in order of
importance, grain moisture content, temperature, and amount of physical
damage to the grain.
  Although, as stated earlier, yeasts and bacteria grow at moisture levels lower
than those required by grain, a high moisture environment of 95% relative
humidity or higher is required for growth. Grain in equilibrium with this
relative humidity will be about 22% moisture, depending on the temperature.
Rice and maize are often harvested at this moisture content, but most other
grains and seed crops are harvested at lower moistures. Mold, however, can
grow under these and even somewhat drier conditions. Mold growth stops
below conditions of 70% relative humidity. Safe storage moisture content for
grain will be below that in equilibrium with 70% relative humidity. Some
molds can grow very slowly in grain at temperatures below freezing of water,
but at temperatures of 54.5[degrees]C their growth is stopped. Table Ill shows the rate
of dry matter loss in relation to temperature and moisture, and shows how
much weight loss may be expected to occur in undamaged grain at given
moistures and temperatures. As seen in the Table, grain at 25% moisture and
15.5[degrees]C will lose 0.0312% of dry weight per day. Thus, in 60 days the loss will
be: 0.0312 x 60 = 1.87%. By this time the grain will be obviously out of good
                                   TABLE III
                     Rate of Dry Matter Loss in Undamaged
              Grain as Related to Grain Moisture and Temperature
                                         % Loss per Day
   ([degrees]C)       15% m.c.(a)      20% m.c.     25% m.c.     30% m.c.
       4.5             0.0003           0.0033       0.0098       0.0173
      15.5             0.0010           0.0106       0.0312       0.0553
      26.5             0.0034          0.0338       0.0994        0.1766
      38.0             0.0101           0.1074       0.3165       0.5622
(a) m.c. = moisture content.
Notes: Oilseeds will not necessarily follow this table. Mechanically field-shelled (combine-shelled)
maize will regularly contain approximately 30% damage and Table IV will apply. Below 15%
moisture-caused losses will be inconsequential.
  Damage to the seed coat of a kernel creates a more favorable condition for
mold growth. Physical damage is defined as any break or rupture in the seed
coat of the grain. Physical damage is associated with shelling or threshing and
is also caused by insects and rodents. It can be pronounced in corn mechanically
shelled at high moisture levels. Small grains such as wheat and rice would
have very low levels of damage due to harvest but insect damage should be
considered. Table IV shows the factor by which the rate of loss for undamaged
grain in Table Ill is multiplied to estimate the rate of loss for damaged grain.
Thus, if the loss were 1.87% as calculated above and the grain had originally
had 10% damaged kernels, then 1.87% must be multiplied by 1.30 and the loss
would come to 2.43%.
  Tables III and IV apply to the first 1 or 2% loss of dry matter. The rate of
loss will increase with time as the molds grow and multiply; however, the grain
will generally be rejected as food by the time 2% loss has occurred.
  Moldy grain may be unevenly distributed in layers or pockets associated
with high moisture from leaks, condensation, and insects. In such cases, it is
necessary to measure separately moisture and temperature in these pockets and
in nonmoldy portions of the grain.
Loss Measurement by Weighing Damaged and Undamaged Kernels and
Calculation of Loss
  The sound and moldy kernels should be counted and weighed and the average
weight determined.
            % weight loss = (UNd) - (DNu) 
                            -------------- x 100
                              U(Nd + Nu)
where U = weight of undamaged grains,
     Nu = number of undamaged grains,
      D = weight of damaged grains,
     Nd = number of damaged grains.
  Samples taken from stored grain may contain kernels from portions significantly
stratified as to mold and moisture (insects also), and in the calculation
of losses it may be necessary to allow the samples to reach a moisture equilibrium
before weighing. Internal insect-damaged kernels may be present in both
                                   TABLE IV
                       Physical Damage Modifier on Rate
                              of Dry Matter Loss
          Physical Damage                             Modifier
           (% by weight)
                 0                                     1.00
                10                                     1.30
                20                                     1.67
                30                                     2.17
the sound and moldy portions and may need to be considered. Experience has
shown that, if as much as 1% is insect infested, infestation will be visible as
insect emergence holes when about 500 g of grain is rapidly examined for
insects or insect damage. This examination is conducted by passing a small
amount of grain at a time across a well-illuminated surface, and rolling or
turning the kernels while searching for emergence holes. The 500 g can be
examined in about 5 to 10 min. Such an examination should reveal some, but
not necessarily all, of the holes.
Loss Measurements by Comparison of Weigh-In and Weigh-Out
  Losses will be measured from start of storage until grain is removed from
storage. The method to use for measuring loss should be based on changes in
unit weight (test weight). As mold destroys dry matter, it will reduce the unit
weight of the grain.
  To use this method, a baseline for each storage unit needs to be established
by sampling the grain when it is put into storage and measuring the unit weight
from this sample, which becomes the basis for estimating loss from future
samples from that storage.
Respiration-Induced Losses That Result in Grain Being Rejected as Inedible
  Any measurement of weight loss due to respiration of microorganisms
would be dominated by a quality loss which would make the individual kernels
so bad they would be picked out and thrown away (or fed as feed), or the lot
would be rejected.
  Therefore, the methodology is to determine what is locally not used for
food. This needs a survey technique. The survey will measure a level that
depends on a subjective measurement that will vary with time, place, and
hunger. In surveying, comparative or permanent use of the data requires that a
sample or photographic record, or both, be kept of what the rejection levels
were during the particular survey.
  Experience has shown that grain may be rejected as inedible when there is
about a 20% loss in weight due to mold damage. The level at which this occurs
is highly variable and subjective. It varies by socioeconomic levels, by local
beliefs and customs, by the degree of hunger, the season and what is available,
by whether one is a seller or buyer, and by the difference between common
practice and a demonstration for the outsider. There is, however, no definitive
answer to the problem of how to obtain a realistic appraisal of actual conditions
of use.
  The appraisal must fit the local use situation as set forth in the following
  1. Consult with the person regularly making the decision.
  2. Take care that the sex of the person interviewed is the same as that of the
person regularly making the decision.
  3. Take care that age and other social status situation is as practiced.
  4. Take care that outside pressures are not applied:
     a. To be more careful.
     b. To be less careful.
     c. To demonstrate special sight or odor skills.
     d. To impress a husband (wife), headman, outsider, etc.
  5. Take care that location, light, time of day, and utensils are normal.
  6. Consider using internal checks such as replicate samples or repeat samples
on other days.
  7. If for home use, buying, selling, or market grading, have the total situation
appropriate to that decision.
  8. Recognize the importance of interviewer-related bias.
     a. Standardize the approach.
     b. Consider the use of one interviewer throughout the survey.
     c. Consider the use of identical sex, age, and size of local individuals
        throughout the survey.
  Additionally, individual and/or local and/or seasonal or yearly standards
with national and international criteria suitable for area-to-area, year-to-year,
and country-to-country understanding and standardization may be compared
but require detailed expert consultation.
  As stated earlier, in any level-of-rejection evaluation, it is imperative that
the level be in a form that can be preserved in photographic or other standards,
so that there will be a record of what the levels were and what was rejected.
  This transfer from local decision to technical standard is one for an expert
grain grader. Such a person can transfer the subjective criteria to an area-wide
survey of the frequency of occurrence.
  The standardized grading approach could be undertaken from the beginning
with an experienced grain grader who would use the information obtained in
the field to establish a standard of grading which could then be used to train
the laboratory technicians. If a sound field basis of judgment were established,
it could be uniformly and accurately applied by the trained technicians. This
would remove that bias which results when relying on a farmer's judgment. It
would also reduce interview time and the time to sample the farmer's store. It
would be necessary, however, to either have the experienced grain grader
present in the field and laboratory during the entire time of the first year's
sampling, or to maintain the samples as evaluated by the farmer in a way that
the grain grader could use them to establish the rejection standard. The second
approach would be preferred since the seasonal effect could be observed and a
realistic average would be more easily obtained. From this, then, would develop
a grain standard which would allow grading of any sample of grain on
the basis of edible or unedible, and therefore an estimate of the loss of grain as
food within the area.
                                  CHAPTER VI
                                  D. Rodents
        Part 1. General Considerations, Direct Measurement Techniques,
                  and Biological Aspects of Survey Procedures
                          W. B. Jackson and M. Temme
  Food losses to rodents are acknowledged to be great, but quantification of
this diversion from human food supplies is less than satisfactory. Literature on
rodent depredations to food (both pre- and postharvest) recently has been
summarized by Jackson (1). Lack of adequate data and appropriate survey or
sampling techniques was recognized as a prime deterrent in obtaining adequate
estimates of loss.
  Most data of local or national postharvest losses result from bureaucratic
guesses. Studies are rarely undertaken, although extrapolations are sometimes
attempted. (See Jackson [11 for detailed analysis of this problem.) While many
of the figures quoted in government reports may be correct, they usually
cannot be documented.
  Most of the surveys which have attempted to obtain data must be suspect,
such as a "felt loss" survey among Indian stored-grain merchants who reported
that monthly losses (from all pests) ranged from 1.7 to 3.75% of their
stocks. Another report notes that 1.7% of sacks holding cacao beans in a
Nigerian warehouse were opened by rats and "estimated" that 10% of the
stored product was damaged. Estimates by different investigators of postharvest
losses to rodents in India range from 2.5-5.9% to 25-30%, and even
higher, and annual village losses in India were "estimated" from 2.3 to 3.3
metric tons.
  A few small-scale studies have provided some statistics. A 1975 study in one
godown in India over 11 months showed losses of 1,400 kg of food grains due
to 200 rats. Some rodents hoard food, 3 kg having been found in a single
burrow; but the time required to amass such a volume is generally not known.
Other estimates of burrow hoards have been as high as 15 kg.
  Most efforts at rodent-damage assessments have been focused on crops
under field conditions; however, even in the most recent summary of methods
(2) only sugarcane is cited as having an acceptable survey tool. Suitable techniques
for field assessment of damage to rice also have been developed and
field tested in the Philippines.
  It is evident that one cannot turn to an existing body of knowledge for
obtaining an accurate measure of postharvest losses. It is acknowledged that
"the usual method of estimation is to blame vertebrate pests for all losses that
cannot be accounted for in any other way." FAO, in assessing its role in
reducing food losses, indicated that no agreed methodology existed for assessment
of losses from pests generally. Present resources available for the necessary
engineering, biological, and statistical studies to develop and evaluate
procedures in each country were deemed inadequate. However, GASGA and
several FAO projects are now devoting effort to this concern. The program at
the Vertebrate Pest Center, Karachi, is of particular interest, but no working
reports are known to be available.
Field Losses
  Postharvest losses often are assumed to start with some manner of storage,
though it must be recognized that crops that are shocked or windrowed in the
field for drying may well have rodent infestants and that these rodents can
cause local damage and then be transported into storage sites. Assessments
may be made in the field directly (usually involving a sampling technique) or
with indirect procedures.
  If comparable fields without rats can be found, weight or volume differences
in the ultimate harvest would provide a good estimate of rodent losses -- if
fungus, insect, bird, or large mammal depredations were not involved or
were assessable. Techniques developed for assessing bird damage to maize
utilize counts of individual kernels destroyed, the length of kernel rows eaten,
or simply the proportion of the ear damaged. Separation of primary from
secondary involvement also is necessary. For example, insects or smut may be
able to invade the maize ear when the husk has been penetrated by bird or
rodent activity.
  If the crop is left in sheaves or stacks in the field for a time, serious damage
may be caused by rodents. This damage can perhaps be measured by comparing
grain losses and contamination in the damaged portions with sheaves
and stacks that were protected from rodents.
  Threshing yards are known to be sites where considerable rodent damage
and loss can occur. Comparing pre-threshing harvest estimates with grain
finally used may ascribe losses to the wrong operational sector, however.
Storage Losses
  Direct determination of actual losses is one approach, although the total
volume of stored products usually cannot be examined due to time, manpower,
or financial limitations, so a sampling technique must frequently be
used. Obviously, moisture losses and damage from insects, fungi, birds, or
other pests must be assessed separately.
  Changes in quality of stored food can be important. Loss of germ by selective
feeding markedly reduces the value of maize. Urine, fecal, or hair contamination
of stores may provide a disease potential (eg, Salmonellosis) and alter
the aesthetic valuation, and hence the market price, of the product.
  Unlike insects, which often are distributed throughout the grain stores,
rodents will be at the periphery of bulk storage and often nonrandomly distributed
through bagged or boxed products. This complicates any statistical
approach to sampling and assessment. One approach would be to examine all
susceptible products by inspecting each bag or container incoming and outgoing
for rodent damage (and urine by ultraviolet light if this form of contamination
is of concern). Contents of each damaged unit would require detailed
examination to determine actual loss. The remaining portion may be judged
satisfactory for use, convertible to animal food (at lower market prices), or
unsuited for any use. Operationally, products stored in certain structures or
sections of structures known to be without rodents could be omitted from such
  To ascertain rodent damage and contamination in the total contents of bulk
storage units, such stores can be sampled around their perimeter to determine
incidence of droppings and gnawed kernels, but this is likely to be most difficult
because of inaccessibility of this layer.
  Sampling schemes extensively used for assessing grain quality, especially in
transit, will be satisfactory for determining rodent infestation or contamination
only if the period of transit is relatively short, the load is well mixed, and a
large active rodent population is not present. Allowing loaded boxcars to stand
on a siding for several weeks permits invasion from local populations, but
damage is likely to be peripheral and not detectable by probe samplers.
  Indirect determinations of losses involve learning the sizes of infesting rodent
populations. If the rodent population can be censused or estimated, their
daily food consumption (and contamination) could be extrapolated as an estimate
of the loss. The techniques used to estimate population size require
statistical assumptions that cannot always be met, although some simple techniques
that can be utilized to determine population numbers in most storage
facilities are described in Chapter VI, Section E.
  The now classic techniques used to census rat populations in New York and
Baltimore (3, 4) require calibration for each environmental complex of concern.
Even so, this may represent the most practical approach. Essentially the
rodent activity in evidence (droppings, runways and burrows, gnawed food) is
evaluated by one team and the population size estimated on the basis of these
signs. After this a second team determines the actual rodent population by
intensive trapping. When the population estimates of the first team are in
essential agreement with the capture determinations of the second team, the
first team continues through the area with sight surveys and consequent population
estimates. Unfortunately, this calibration process is lengthy and must be
repeated whenever different species or different environments are encountered.
Its adaptation to village or godown environments has not been specifically
demonstrated, but as long as the areas of rodent activity are discernible,
its application should be possible.
  Some attempts occur at popularizing estimation of rodent numbers by assuming
the rats seen during daylight hours represent a scientific proportion of
the total population. Unfortunately, such procedures are without experimental
backing. Furthermore, rats with a larger home range and daily need for water
may be more rapidly observed than mice that remain hidden within their food
  On a limited basis, direct and total counts of a population may be obtained
in a circumscribed area and losses estimated by calculating the food eaten by
the population. This involves trapping, marking of individual animals, and
direct observation. This tends to avoid difficulties with widely varying movement
patterns and nonrandom distribution of animals but is very demanding
of time. This requires some judgment as to migrations in and out of the area,
amount of grain as against refuse eaten, etc.
  One traditional estimation technique employs census baiting. By ascribing a
given quantity of a placed bait eaten to a rat, the population can be estimated.
However, where high quality food is stored and thus competes with placed
baits, the competition and the neophobic responses (of rats) are likely to result
in serious underestimates of the actual population. Mice, with very limited
home ranges, often cannot be estimated with such a technique when they are
infesting food-storage facilities.
  If the population has been satisfactorily assessed, an attempt can then be
made to estimate the corresponding losses, or at least the losses caused by the
predominant species, for it is rare for only one species to be involved.
  A minimum estimate can be made by multiplying the daily consumption of
an individual by the number of individuals in the population. Consumption is
related to the liveweight of the animals. Mean daily consumption varies with
the nature of the foodstuff and especially with its nutritive value. For cereals,
the following amounts of grain can be used: For Rattus norvegicus, 20-25 g,
Mus musculus, 2.5-3.5, Mastomys natalensis, 8-10, and Bandicota bengalensis,
  If no experimental data are available, daily consumption can be estimated at
1/10 of the mean liveweight of the species.
  In addition to the grain eaten by rodents, there are partially eaten grains
which are unfit for human consumption. Decisions on discarding such grain
will vary with the season, with the abundance of any particular harvest, with
local and national mores, etc. Thus, losses need to be on the basis of actual
discards, not what should be discarded according to aesthetic and health consideration
(see Chapter VI, Section C).
  One very real concern is for the process of obtaining accurate data. Catching
rats and then releasing them (for Lincoln Index estimates) is difficult to explain
to a farmer suffering from rodent depredations. Probably such an approach
should be reserved for government facilities where research can be
conducted without intrusion into personal rights. Yet studies ought to be done
in housing units, local godowns, and small shops or markets. Residents and
owners must have confidence in the investigator and must be able to see some
direct benefit to themselves for their cooperation, such as removal of rats or
better storage conditions. Without the full support of local peoples, the data
derived from study programs are likely to be another set of "estimates" that
are not well grounded.
  Pragmatically how much damage or loss occurs from rodent infestations is
less important than getting to the sanitation, construction, and control techniques
that will result in more stored foods being available to people. But to
justify and evaluate rodent management programs, cost/benefit ratios have to
be determined. Herein lies the reason that such documentation needs to be
Summary of the Problems
  Each component in handling and transportation of foods following harvest
must be evaluated separately.
  * In-field losses lend themselves to direct appraisal (weight loss, kernels
    damaged) and use of sampling techniques.
  * Transportation from one site or field to another may enclose rodents
    within a food supply. Especially if the vehicle is relatively small and the
    time great, losses can be of real consequence. Determination of weight
    loss, especially after damaged or contaminated portions are removed,
    can be made directly.
  * Local storage -- either in the home or in local godowns -- is the fate of
    most grain, and these sites are the most vulnerable to substantial losses.
    Direct measurements (weight/volume) of depredations are most readily
    done, but interpretations must be integrated with local environmental
  * Bulk storage, because of larger volumes involved, is likely to have less
    damage proportionately. The ability either to determine numbers of
    rodents or to assess the damage itself is more limited, however. If the
    grain is bagged or containerized in some way, damage to specific containers
    and their contents can be determined. Contamination especially
    is of concern in bulk storage, since the mixing of a small quantity of
    contaminated or infested grain with a large quantity of clean product
    results in a total lot of contaminated product.
  * Economic (and aesthetic) thresholds for food damage and contamination
    need to be established (5). Efforts at sampling become increasingly
    costly at lower infestation and contamination rates.
Methods-Oriented Summary
  The problem of postharvest losses to rodents resolves itself into three aspects:
1) Losses due to the removal of corn, sorghum, and millet in which grain
is eaten from cobs, heads, or spears; 2) losses to threshed or shelled grain; and
3) losses caused by contamination in which the contaminated grain is discarded.
(Losses due to rejection by the users is discussed in Section C of this
1. Losses to Ears or Heads of Corn, Millet, and Sorghum
  Measurements consist of estimating the percentage of grain removed from
the heads, shelling, and weighing undamaged heads of the same size, and
calculating losses by percent or actual weight loss.
  Samples may be taken so as to be representative of the lot as a whole if the
damage is distributed throughout the lot. When damage is located in a particular
portion of the stack, pile, or windrow, sampling needs to be representative
of that situation (see Appendix B) with an estimation of the proportion of the
whole that is so affected.
2. Losses to Threshed or Shelled Grain
  Problems of sampling bagged or bulk grain are of three types: a) Those in
which before and after weights are available or may be obtained; b) those in
which bagged grain with and without damage may be weighed and compared;
or c) those in which no actual comparative weights may be made of the grain
itself. These procedures are amplified below:
  a. In many market, transport, and warehousing situations, the grain has
been previously weighed. Reweighings will give the amount lost to rodents, if
this is the only source of change. This can be a laborious and costly task,
however, and usually an estimate must be made using one of the procedures in
the following two paragraphs.
  b. Comparison of weights of undamaged and damaged bagged grain: Rodents
often concentrate their feeding and nesting in fairly well-delineated areas
of bagged grain storage. When this is the case, damaged bags may be weighed
and compared with the weight of undamaged bags taking appropriate care to
obtain representative samples of the bags if weights before loss are not available.
When the individual bags have already been weighed, direct and actual
losses may be readily obtained.
  c. Overall losses to grain in storage: Most often serious rodent losses occur
in relatively long-term storage or in a long-established marketing or warehousing
situation where grain is present under a stabilized pattern. With long-term
storage, local rodents may be found out of the store, moving in for feeding
and subsequent habitation. They will live in the stored grain if undisturbed and
if water is nearby. Rodents in markets where there is a permanent supply of
grain moving in and out of the storage will usually be living nearby, in holes in
or under the floor, between walls, or in burrows, moving into the grain for
food, and to nearby sewers, drains, or sinks for water. In these cases, losses
involved are estimations of the rodent population, and the food loss is calculated
on the basis of the number of rodents x time x food consumption.
  Some simple methods suitable for general use of rodent population estimation
are given in Chapter VI, Section E. Rodents, however, are known for
their diverse feeding habits and their food intake may not be limited to the
grain supplies.
  Specific field studies, preferably integrated with insect-loss evaluations,
should be undertaken to quantify rodent losses in selected environmental situations.
Typical sites might be small community or commercial godowns, individual
farm storage structures, kitchen or household storage, and field drying
or curing operations. Effects of different environmental regimes and different
rodent species need to be considered. Whenever possible, association with
existing FAO, EPPO, CARE, or binational programs would have obvious
  At the village household level, direct measurements of loss contamination
could be made on a daily or short-term basis. This requires measurement of
foods purchased or taken from stores and analysis of amounts actually available
for later consumption. Rodent populations could be evaluated by estimating
sign or intensive removal trapping. Such an effort would require exceedingly
good cooperation of village residents and merchants and great honesty
on the part of all participants.
  For small godowns the most satisfactory measure is the comparison of input
stores to those taken out at a later date. This involves measurement of the total
stores and evaluation of contamination. For larger godowns, this requires use
of sampling techniques. Rat populations have to be determined by trapping or
census feeding. Because of the inaccessibility of many areas, use of sign probably
would not be satisfactory.
  Considerations in evaluative efforts (6, 7) should include: obtaining known,
estimated, or "felt" losses from owners, occupants, or merchants; evaluating
structure for harborage and infestation potential; quantifying the rodent sign;
evaluating daily/weekly/monthly/annual grain-handling procedures and sanitation
practices; monitoring incoming and outgoing products to determine
depredations; accounting for hoarding activities (eg, burrow excavation); and
segregating losses from moisture decrease, insects, birds, and fungi, and determination
of primary causes of loss.
                               Literature Cited
1. JACKSON, W. B. Evaluation of rodent depredations to crops and stored products. EPPO
     Bull. 7(2): 439 (1977).
2. FAO/CAB. Crop loss assessment methods. FAO Manual, Commonwealth Agr. Bureaux,
     Slough, England (1971).
3. DAVIS, D. E. The rat population of New York, 1949. Am. J. Hyg. 52(2): 147 (1950).
4. DAVIS, D. E., and FALES, W. T. The rat population of Baltimore, 1949. Am. J. Hyg. 52(2):
     143 (1950).
5. TAYLOR, T. A. Major problems affecting productivity of cereals -- the pest problem. In:
     Agr. Res. Priorities for Economic Development in Africa, Abidjan Conf. 1968. NAS-NRC
     Publ. 2: 175 (1968).
6. ANONYMOUS. Group for Assistance on Storage of Grain in Africa Seminar on the Methodology
     of Evaluation Grain Storage Losses. Trop. Stored Prod. Inf. 24: 13 (1973).
7. ADAMS, J. M. A guide to the objective and reliable estimation of food losses in small scale
     farmer storage. Trop. Stored Prod. Inf. 32: 5 (1976).
BROWN, R. Z. Biological factors in rodent control. U.S. Public Health Service Training Guide
EVERARD, C. 0. R. Some aspects of vertebrate damage to cocoa in West Africa. Proc. Conf. on
    Cocoa Pests W.A.C.R.I. (Nigeria), p. 114 (1964).
FELLOWS, D. P., and SUGIHARA, R. T. Food habits of Norway and Polynesian rats in
    Hawaiian sugarcane fields. Hawaii. Plant. Rec. 59(6): 67 (1977).
FERNANDO, H. E., KAWAMOTO, N., and PERERA, N. The biology and control of the rice
    field mole rat of Ceylon Gunomys gracialis. FAO Plant Prot. Bull. 15: 32 (1967).
FOOD AND AGRICULTURE ORGANIZATION. Reducing post-harvest food losses in developing
    countries. ACPP: Misc./21:15 pp + annexes (1975).
FRANTZ, S. G. The behavioral/ecological milieu of godown bandicoot rats -- implications for
    environmental manipulation. All India Rodent Seminar, Siddhpur (1975).
HOPF, H. S., MORELEY, G. E. J., and HUMPHRIES, J. R. 0. (eds.). Rodent damage to
    growing crops and to farm and village storage in tropical and subtropical regions. Centre
    for Overseas Pest Research, Trop. Prod. Inst. (1976).
KRISHNAMURTHY, K. Problems of food grain storage. Training Manual. Asian Productivity
    Organization, APO Project TRC/IX/73, Tokyo, 81-84 (1974).
PRAKASH, I. Rodents and their control. Post-harvest prevention of waste and loss of food
    grains. Training Manual. Asian Productivity Organization, APO Project TRC/IX/73,
    Tokyo, 185-192  1974).
SANCHES, F. F. et al. Annual Report. Rodent Research Center, College, Laguna, Philippines
SPILLETT, J. J. the ecology of the lesser bandicoot rat in Calcutta. Bombay Natural History
    Society (1968).
                                  CHAPTER VI
                                  D. Rodents
                   Part 2. Loss Determinations by Population
                     Assessment and Estimation Procedures
                                 J. H. Greaves
  Direct measurement of postharvest grain losses to rodents is difficult. As
explained in Part 1, the losses to rodents have to be distinguished from losses
to birds, spillage, and pilferage, and, in the fields, from shedding or preharvest
losses. Therefore, to determine the loss to rodents, all of these other losses
must be identified and measured separately. Weight losses due to other pests
and to changes in moisture content must also be measured and considered. In
addition, specialized studies in the ecology of the rodents may be required.
Thus, direct assessment of losses to rodents is complex, and can rarely be
contemplated except as an aspect of a multidisciplinary research study.
  In contrast, techniques for the estimation of rodent populations, developed
by specialists in the fields of rodent control and small mammal ecology, are
well established. Clearly, the extent of grain loss to rodents depends on the
distribution, size, and species composition of the rodent populations involved.
Simple versions of established population assessment techniques can therefore
enable the ordinary competent biologist with a little specialized training to
derive loss estimates which, though indirect, will be based on objective data
and, though approximate, will generally be of the correct order of magnitude.
  The methods proposed here are intended primarily for use in grain stores.
They may also be considered for use, if intelligently adapted, in fields during
the immediate postharvest period and in threshing yards. They are unsuitable
for use where the grain, prior to threshing and still attached to the straw or
haulm, is either stored in large compact stacks or on vehicles during shipment.
The aim of the methods is to estimate the weight of grain consumed by rodents;
related losses attributable, for example, to contamination, health hazards,
and damage to sacks must be evaluated by other means.
Personnel and Training
  The work, including all practical operations such as placement, setting, and
checking of traps, should be performed by zoology graduates, preferably with
some experience in the fields of rodent control, grain storage, or small mammal
ecology. They must possess or first acquire various skills in order to carry,
out the following operations competently:
  a. Identify the rodent species, and distinguish adults of the smaller species
     from juveniles of the larger species.
  b. Identify and evaluate signs of rodent infestation.
   c. Set traps.
  d. Handle live rodents.
  e. Keep field records of the high standard required for investigative work.
These skills are best acquired on the job under the guidance of an experienced
specialist. The basics may also be learned in a week or so of laboratory
and field training at an institution specializing in rodent control and ecology,
in which case it will be necessary to add a further self-training period of 2-4
weeks in which to practice and improve the newly acquired skills in an operational
Selection of Study Sites
  The methods given in Appendix B should be employed. Frequently it will be
found that appropriate government departments maintain registers of farms,
premises of licensed grain traders, etc., which can greatly facilitate selection of
a representative sample of study sites.
METHOD A -- Preliminary Survey of Infestation
  A preliminary survey of the study site must always be made in connection
with the two detailed techniques to be described subsequently (METHODS B
and C). In addition, a systematic survey of a random sample of sites can, by
determining the incidence of sites on which rodents are present and have access
to grain, make a valuable contribution to an overall assessment of the rodent
problem. It is emphasized, however, that the METHOD A survey procedure
will lead to a valid estimate of the quantity of grain lost to rodents only if it is
followed up with either METHOD B or C.
1. Electric flashlight/torch.
2. Tracking powder (talcum or finely powdered chalk). A glass jar with a
   perforated lid provides a convenient means of dispensing the powder.
3. Clipboard and record sheets.
  Two visits will be required. On the first visit record the following information
on a record form:
  a. Date of survey
  b. Address of store
  c. Commodities stored and quantities (by weight)
  d. Nominal capacity of the store (by weight)
  e. Date of inward shipment
  f. Expected date of outward shipment
  g. Estimated annual turnover (by weight)
  h. Brief description of the storage structure and conditions of storage
  i. A sketch map of the store (made on the back of a form) showing important
     features and the location of the stored grain.
  Inspect the site thoroughly for signs of rodent infestation, including burrows,
excreta, smears, footprints, damage to the commodity or structure, and
places where rodents may enter the store. Record these signs on the sketch map
as they are found. During the inspection, whether or not signs of infestation
are found, lay tracking patches approximately 200 x 300 mm at intervals
along the walls of the store and beside the stacked grain, especially around
corners. The tracking patches should be laid at the rate of approximately one
per 50 tons of grain, except that in stores of less than 250 tons, not less than
five patches should be laid. The tracking patches should be entered in a numbered
sequence on the record sheet and their positions indicated on the sketch
  The second visit should be made the next day and the presence or absence of
rodent tracks on each tracking patch recorded. Usually it will also be both
useful and possible to record whether any tracks found were made by large or
small rodents (rats or mice) or by rodents of both sizes. It will not normally be
permissible to conclude which species is present until several trapped specimens
have been identified.
  A simple estimate of the incidence of infestation may be calculated when a
random sample of stores of a single type has been surveyed, as follows:
    Percent of stores infested = No. of stores infested
                                 ---------------------- x 100    
                                 No. of stores surveyed
  Percent standard error = [square root](% stores infested x % stores not infested)
                                                   No. of stores surveyed
METHOD B - Trapping to Extinction
  In principle, if a complete census of the population is made by trapping all
the rodents that have access to the grain, then the feeding capacity of the
population, and hence the current daily grain loss to rodents, can be estimated
by multiplying the number of rodents by their daily food requirement, since it
may reasonably be assumed that rodents with access to stored grain will use it
as their primary food source. The method is suggested for use in stores with
populations of up to 200 rodents; this would include a fairly heavily infested
store holding up to 500 tons or larger, more lightly infested stores. For larger
infestations an alternative technique for population estimation (METHOD C)
is advocated.
  The following equipment is needed in addition to that specified in
1. 200 snap traps (rat size; striking bar 70-80 mm long).
2. 200 snap traps (mouse size; striking bar 40-50 mm long).
3. Spring balance (100 x 1 g).
4. Spring balance (500 x 5 g).
5. Blackboard chalk for marking trap locations.
6. Bait (see later).
  First make the preliminary survey (METHOD A). The objective is next to
trap out the population as rapidly as possible and in a period not exceeding 21
days; to achieve this, the bulk of the population should be caught in the first
week. The correct siting of traps is helped by knowledge of the movement
patterns of the rodents. Much will be known from the preliminary survey, but
it is essential to increase and update this knowledge while trapping is in
progress by the temporary placement of extra tracking patches, which should
be renewed regularly. The tracking patches will also show, by the absence of
tracks, when all of the rodents have been caught.
  A large number of traps must be used, at least equal to the supposed size of
the rodent population and preferably exceeding it by a factor of 2 or more.
They should be distributed at intervals of 1 m or less in all places where the
presence of rodents is suspected. Each investigator should be able to deal with
about 100 traps daily. Place the traps in a systematic sequence (called the "trap
round"), numbering and entering each placement on the record sheet and
chalking up the trap number boldly nearby to make it easy to locate on
subsequent visits. The bait should be of a sticky consistency such as peanut
butter, crushed fruit (banana, oil palm pericarp, or melon), or sweetened
dough, and should be pressed firmly into the bait hook so that the rodents
cannot simply lift it off but are induced to exert some lateral or downward
force on the release mechanism while getting the bait. Succulent baits are often
particularly attractive to rodents in the dry environment of a grain store and it
may be worth changing the type of bait used after a few days. The traps should
be set as finely as possible.
  Each day check the trap round and record the species and body weight of
each rodent caught for each trap. Every trap, whether it makes a capture or
not, must be freshly baited and reset each day and, if judged to be advisable,
its position adjusted so as to increase the chance of making a capture. Where
both large and small rodents are present, concentrate first on trapping the
larger rodents and, as their numbers decrease, gradually switch to using the
smaller traps.
  It may sometimes happen that though the vast majority of rodents are
trapped, a few recalcitrant individuals evade the efforts made to capture them.
The size and species composition of this residual population, provided it is
very small, can often be estimated from the frequency and size of footprints on
the tracking patches. Such estimates and the evidence on which they are based
should always be clearly stated.
Grain Loss Assessment
  The primary data which should be reported are the numbers and body
weights of each species of rodent trapped. The data for each species should be
divided into two body-weight classes: 50 g or less, and more than 50 g. The
biomass (sum of the body weights) of each weight class should then be obtained
for each species. The estimate of the daily grain loss attributable to each
species is obtained by multiplying the biomass of the rodents in each weight
class by a factor representing the daily grain requirement of a rodent in that
weight class, and then adding together the two products.
  Preferably the daily grain requirement of each species of rodent in the two
weight classes should be determined (as a proportion of body weight) for the
commodity and country in question by measuring the actual amounts consumed
by representative samples of captive rodents in cages. Where facilities
for this are lacking, however, it will generally be adequate to base the calculation
on an assumed grain consumption equivalent to 7% of body weight for
rodents weighing more than 50 g and 15% of body weight for rodents weighing
50 g or less. The estimated daily grain loss attributable to species "A," for
example, would then be (0.07a + 0.15b) g, where a = biomass (g) of rodents
of species A weighing more than 50 g, and b = biomass (g) of rodents of
species A weighing 50 g or less.
  The total estimated daily grain loss is then readily determined by adding
together the estimates for the different species, and should be expressed both
as an absolute amount and as percentages of the amount of grain in the store
and of the nominal capacity of the store. If it can be assumed that the rodent
population was reasonably stable, then the loss over a period of time can easily
be calculated. Estimates of the annual loss expressed as percentages of the
amount of grain actually stored, of the nominal capacity of the store, and of
turnover are usually of particular interest.
METHOD C - The Lincoln-Petersen Method of Population Estimation
  This method (1) is based on the following principle: First a sample of animals
is caught alive, marked, and returned to the original population. When a
second sample is then taken, the number of marked animals in the second
sample has the same ratio to the total number in the second sample as the
number of marked animals originally released has to the total population.
Since both the number of marked animals originally released and the proportion
of marked animals in the second sample are known, the size of the total
population can easily be calculated. The application of this principle to estimating
rodent populations involves making several assumptions about the
behavior of the populations. In practice the two most important of these
assumptions are that 1) the duration of the study is sufficiently short that no
significant change occurs in the population, and 2) the chance of capturing a
rodent in the second sample is independent of whether or not it is marked. In
the typical grain storage situation, the first assumption may be satisfied by
completing the study in a period not exceeding 21 days. The second assumption
may be satisfied by using live-capture traps for the first sample and snap
traps to collect the second sample, since the behavioral responses of rodents to
the two types of trap are relatively independent of one another.
   The following equipment is required in addition to that specified for
1. 100 live-capture traps (rat size).
2. 100 live-capture traps (mouse size).
3. Simple restraining devices to hold live rodents for marking (see later).
4. 2 pairs of dissecting scissors.
  Two types of live-capture trap are suitable. These are the funnel-type,
multiple-catch trap with a horizontal counter-poised door operated by the
weight of the rodent as it approaches the holding compartment, and the single-catch
trap with a door-closing mechanism operated by a treadle. Live-capture
traps actuated by a bait hook are not recommended. Live-capture traps for
mice should be made of sheet metal or of 7 mm or finer wire mesh. Specialist
advice should be taken if there is any doubt about the suitability of the trap
designs available.
  First complete the preliminary survey (METHOD A). The operation is next
carried out in two stages.
  Stage 1 lasts 10 days during which the aim should be to capture, mark, and
release as many rodents as possible. Distribute, bait, and set the live-capture
traps, recording the trap round as in METHOD B. An average density of one
rat-sized and one mouse-sized trap per 9 [m.sup.2] is suggested. Fresh bait (eg, soaked
grain or fruit) must be provided daily. One investigator should be able to
service 50-100 traps. Every morning, each newly caught rodent must be
marked by clipping off the middle digit of the right hind foot. To do this, the
rodent should be transferred from the trap to a cloth bag where it is restrained
gently, while the mouth of the bag is opened to give access to the foot.
Alternatively, larger rodents may be restrained in a cylinder or cone made
from chicken wire, while mice may be grasped directly with the forefinger and
thumb by the loose skin over the neck, either straight from the trap or after
first transferring them from the trap to a box or bin 500 mm deep. Newly
marked rodents should be released at the point of capture and their numbers
and species recorded beside the trap entry on the record sheet. Previously
marked rodents should be released at the point of capture without making any
additional record.
  Stage 2 also lasts 10 days during which the objective is to snap-trap as many
rodents as possible, using the procedure described under METHOD B. The
body weight, species, and presence or absence of a mark should be recorded
for each rodent trapped. In accordance with conditions, a lower trap density
may be permissible; however, for the purpose of making satisfactory population
estimates it is desirable to recover at least 20 marked rodents of each
species in Stage 2.
Population Estimates and Grain Loss Assessment
  The primary data which should be reported are:
  * The numbers of each species marked in Stage 1.
  * The numbers of marked rodents of each species trapped in Stage 2.
  * The numbers of unmarked rodents of each species trapped in Stage 2.
  * The species and body weight of each rodent trapped in Stage 2.
  * The population estimate (P) for each species as P = an/r where a =
    number marked in Stage 1, n = total number caught in Stage 2, and r =
    number of marked rodents caught in Stage 2.
  The estimate of daily grain consumption is obtained as before, except that it
is necessary to determine the weights and relative sizes of the two body-weight
classes by reference to the sample of rodents trapped in Stage 2. Thus, where in
the absence of data from captive rodents it is assumed that the daily grain
consumption figures for animals greater than 50 g and for smaller rodents are
respectively 7 and 15% of body weight, the daily grain loss attributable to
species A will be:
                       P [0.07ab + 0. 15 (1 - a) c] g
where P = the population estimate for species A,
      a = the proportion of rodents of species A of body weight greater
          than 50 g,
      b = the mean body weight (g) of rodents of species A weighing
          more than 50 g, and
      c = the mean body, weight (g) of rodents of species A weighing 50 g
          or less.
(The parameters a, b, and c must be calculated from the sample trapped in
Stage 2.
  If the population estimate, P, is unsatisfactory owing to fewer than 20
marked rodents of the species concerned having been trapped in Stage 2, then
the data for two or more species may be pooled to give a combined estimate.
The estimate of total daily grain loss should be expressed in the various ways
suggested under METHOD B.
                               Literature Cited
1. LE CREN, E. D. A note on the history of mark-capture population estimates. J. Anim. Ecol.
   34: 453 (1965).
                                  CHAPTER VI
                  E. Measurement of Losses Caused by Birds (8)This
brief summary was excerpted and added to by K. L. Harris from Estimates
of Bird Depredations to Agricultural Crops and Stored Products by W. B. Jackson
and S. S. Jackson, first presented at the Colloquium on Crop Protection Against
Starlings, Pigeons, and Sparrows, European and Mediterranean Plant Protection
Organization, Jouy-en-Josas, France, Oct. 18-20, 1977.)
  This section recognizes that there is scarcely any line between grain held in
the field for maturing and drying and grain held for maturing, drying, and
storage. The storage portion of the cycle is intertwined with both the drying
and holding requirements. At times grain, chiefly maize, sorghum, and millet,
may be held for extended periods in the field prior to harvesting for storage or
direct to the table use. Some of the most serious grain losses occur at this stage
when losses to Quelea spp., parakeets, and blackbirds have assumed disastrous
proportions; however, losses are rarely quantified.
  It is often difficult to relate specific birds to designated damage or losses.
Feeding patterns may be irregular or overtap; insect outbreaks, drought, or
flood may alter expected patterns; fungi may enter as a secondary factor
related to bird damage; and the measurement techniques, themselves, may be
tedious and exacting. Comparisons of damage to benefits, whether off-season
removal of weed seeds compensates for food losses, effects of intensive mono-culture,
the mutually destructive breakage or cutting of heads by mammals
and birds, and other matters all complicate loss assessments. Losses to piled
and bagged grain are often observed but rarely if ever quantified, and birds are
usually more readily accepted than rodents as part of the environment.
  While losses are real, satisfactory methods of determining losses have seldom
been available or used. The most intensive statistical efforts have been on
blackbird damage in the United States. These have used the detailed row-centimeter
measurement technique and visual-toss estimates as summarized
  Row-Centimeter Measurements (used on maize). The number of damaged
and undamaged ears in a row (15-100 ft) are counted. On damaged ears, the
average lengths of damaged and undamaged kernel rows are measured to the
nearest, approximately, 2 or 3 mm. These lengths are converted to losses per
area, eg, tons/hectare. Less exacting are simple measurements of the portion
of ear damaged, which may require some arbitrary averaging if the damage
pattern is not symmetrical.
  Visual-Loss Estimates. This technique is usable on many different crops,
but observers must be trained and their procedures calibrated for each crop.
This is a much more rapid technique, since counting is not specifically required.
Damage-level criteria (5, 10, 20, 40%) are established and workers
trained by repeated tests to distinguish visually between these levels of
  Losses to stored bagged or bulk grain can best be measured by before and
after weights over a period of time. The kinds and numbers of birds and how
much time they spend on the grain should be noted. These figures can then be
used in estimating losses in similar situations elsewhere.
                                  CHAPTER VI
                            F. Moisture Measurement
             T. A. Granovsky, G. Martin, and J. L. Multon
  Accurate measurement of grain moisture and its variations is critical for
proper assessment of weight losses during storage. Changes in moisture content
are accompanied by changes in weight and volume and need to be recognized
as separate from actual grain losses. Frequently, the weight of moisture
gained or lost by grain may exceed weight losses induced by insects, rodents,
birds, or fungi. Moisture changes are merely the gain or loss of water; the
others may alter food quantities or qualities. Therefore, measuring the moisture
content of grain is an extremely important operation from three standpoints:
  1) Technology: Knowledge of moisture content is needed to efficiently determine
and manage the harvesting, drying, stocking, and processing operations.
it is also essential for assessing and controlling postharvest losses insofar
as the action of water governs deterioration phenomena.
  2) Analysis: To compare the results of analysis with a fixed basis (dry
matter or standard moisture content). In particular, assessing the weight of a
stock of grain and making loss determinations requires accurate knowledge of
the moisture content.
  3) Marketing: Commercial purchasing and sales contracts often stipulate an
upper limit for the moisture content not to be exceeded.
  Samples should be analyzed as soon after being obtained as is practical.
Since grain can gain or lose moisture rapidly, all samples not immediately
tested should be retained in air- and moisture-tight containers and not exposed
to undue temperature variations.
  It is necessary to emphasize how important it is for all measurements to be
made with thoroughly standardized procedures. The International Association
for Cereal Chemistry has called attention to various procedures for measuring
moisture content.
  Moisture measurements depend on two fundamental baseline procedures.
These procedures determine what water shall be classified as free moisture in
the grain and, hence, is the water that is dealt with in a percentage moisture
determination and will be the basis of the reading given by a moisture meter.
Table V summarizes international approval of the two types of baseline

pgl4x120.gif (600x600)

  An in-depth discussion of the comparative values of the "fundamental reference
methods" versus the "practical reference methods" is not within the
scope of this manual. They do, however, involve highly specialized apparatus
and conditions (see ICC standards in Table V).
Use of Meters (See also Appendix C)
  The amount of grain necessary for determining moisture content will depend
on the testing method used. Some methods are portable and enable
determinations in the field. Other methods are laboratory-based and may
require a constant power supply and chemical agents. Selection of a meter will
depend on where the determinations are to be made. In general, use of a
moisture meter is encouraged, especially one which is both portable, enabling
on-the-spot moisture determinations, and rugged enough to withstand transport
from locale to locale. Aspects of proper adjustments and sensitivities of
each meter should also be considered when making a selection. The data in
Appendix C are pertinent in deciding what meter an investigator will select,
amount of grain needed, speed of operation, and accuracy of each. In any
case, manufacturer's directions in using the meter should be followed.
  Moisture meters require periodic calibration, the frequency of which will
depend on the meter and conditions of its use. Often it can be checked against
samples especially prepared and packaged for this purpose. In other cases, it
may be taken to a central laboratory for comparisons with a meter reserved for
this purpose, for comparison with control samples, or for comparisons with
results by standard oven-dry methods. To re-set, manufacturer's directions
should be followed.
  As a general rule, all field or laboratory determinations should include at
least three and preferably five replicates in an effort toward greater validity.
Consistency in handling and preparation of samples for moisture content
determinations is indispensable.
  The percentage loss or gain in weight by the grain may be derived from the
average initial and average final moisture contents. The nomograph (Fig. 10) is

pglx121.gif (600x600)

employed as follows:
  1. Lay a straight edge so that the initial and final moisture content values lie
along this edge.
  2. Read the percentage gain or loss in weight off the right-hand bar.
  For instance, if the initial moisture content of a sample is 12.5% and the
final moisture content value obtained is 16.5%, then the percentage weight
gain is about 4.8% (represented by dots). Conversely, if the initial value is
20% and the final value only 150/o, then a 5.9% weight loss has been realized
(represented by dashes).
                            VII. OPERATIONS
               A. Handling of Samples in the Laboratory
                            T. A. Granovsky
  When a sample arrives in the laboratory from the field, it should be in a
sealed moisture-proof container and at ambient laboratory temperature when
opened. This will require proper preparation and care of samples in transport,
field to laboratory, in addition to prompt attention upon arrival by the laboratory
  During handling in the laboratory, each sample must retain its identity as to
location, data collected in the field, grain type, variety, and time in storage at
all times.
  As each sample enters the laboratory, it should be handled as per the sample
flow and by the procedures indicated below:
Sample Flow
If Moisture Content was Determined in the Field
  1. Sample enters laboratory.
  2. Sample collection data recorded on laboratory data sheet.
  3. Whole sample weighed (grain, dust, insects, dockage).
  4. Grain sieved: insects are recovered and placed in 70% alcohol; dust is
weighed, if necessary, and discarded.
  5. Weight-to-volume vessel properly filled and weighed.
  6. Grain from weight-to-volume vessel recombined with rest of original
field sample, repeated five times, and averaged.
  7. Sample (1 kg) is divided into a series of 8-32 subsamples.
  8. Five subsamples are randomly selected for tests on losses induced by
insects as per other instructions (see Section B, Chap. VI).
  9. Other subsamples may be used as needed in tests on losses induced by
microorganisms/respiration, aflatoxin, etc.
  10. All data derived during loss analysis should be recorded on the data
record sheet (Fig. 11).

pglx1240.gif (600x600)

If Moisture Content is to be Determined in the Laboratory
  1. Steps 1 to 5 are as above, but after weight of the weight-to-volume vessel
has been recorded, the moisture content is determined before the sample is
recombined with the rest of the original field sample. Weighings and moisure
content determinations are repeated three times each and averaged separately.
Steps 7 to 9 are then finished.
  2. The total weight of grain, dust, insects, and dockage should be determined
for the whole sample as it arrives from the field. This figure, and all
subsequent data, should be recorded on separate data record sheets for each
sample. A suggested partial sample sheet for data derived in the laboratory is
presented in Fig. 11.
  3. The grain is then sieved to separate off insects and dust (depending on the
characteristics of the debris, use No. 10 or No. 25 sieve and solid botton pan).
Insects should be placed into bottles containing 70% alcohol, labeled as to
origin, and identified as required.
  4. The weight-to-volume vessel, Fig. 8, should be properly loaded, filled,

pgl8x86.gif (600x600)

sliced, and weighed. This is repeated five times and a mean is taken. After each
weighing, the grain should be recombined with the original field sample and
remixed before another sample is removed (see Section B, Chap. VI).
  5. The 1-kg sample is divided into 8-32 subsamples by using a recognized
method such as a sample divider or by coning and quartering (Appendix A). It
is suggested that the subsamples be placed into individual pre-marked containers
to facilitate their manipulation. As noted in Appendix A, subsamples
may vary somewhat in size (number and weight of kernels) depending on the
commodity and the conditions under which the grain was produced.
  6. Five subsamples (replicates) are then selected at random for subsequent
tests on losses induced by insects. See Section B, Chap. VI, for measurements
of losses caused by insects.
  7. Other samples may be used as needed in tests on losses induced by
microorganisms/respiration, aflatoxin, etc.
  8. All data derived during loss analysis induced by insects, microorganisms/
respiration, rodents, birds, and physical losses should be recorded on the data
record sheet (Fig. 11).
                              CHAPTER VII
             B. Operations Manuals and Laboratory Records
                   T. A. Granovsky and K. L. Harris
  In the conduct of any survey there is absolute need for an operations manual
that describes how the survey is to be managed to ensure that the purposes of
the project will be performed. Operations manuals can be in any useful format,
but should specify duties of each employee and operation. Such a manual
is designed for internal use by operating personnel.
  Depending on the complexity of the operation, the manual may be divided
into subsections for on-the-spot use in specific operations. If an operation is
large enough to involve a payroll, there should be a division under corresponding
functional headings for the purchase of supplies, travel, field observations
and sampling, laboratory analyses, and reporting and tabulating results.
  A complete compendium of what to include in an operations manual is
beyond the scope of this work; however, guidance for a field and laboratory
operations manual is given below:
  It is imperative that all procedures for information-gathering, sample collection
and transport, sample examination and reporting, and collection and
tabulation of results be tested in dry runs before the actual information gathering
gets under way. This period of preparation is used to give a final assessment
of the quality of the written directions, on training or the need for
additional training, and on the suitability of individual people, procedures,
and forms for the job. Make changes as required.
Field Controls
  1. Once the sample collection and field observation sites, system, and criteria
are established, these same parameters need to be recorded on paper in
terms suitable for the user.
  2. Sample collection should be explicitly set forth as to where, when, and
how -- with no room for deviation.
  3. Use of alternative procedures, when permitted or applicable, reporting of
inability to sample or make observations, reporting of broken containers, lost
samples, and miscounts all need to be explicitly detailed.
  4. There need to be observation reporting forms, sample collection forms
and labels, packaging and shipping forms, and supplies where required.
  5. Triers (see Appendix A) and other technical devices and supplies need to
be provided (bags, preservatives, clasps) and their use completely described
(see below).
  6. Where, when, how, and how much sample is taken needs to be explicitly
set forth. How to operate triers, how much preservative is to be added, how to
get samples to the laboratory, and speed and route of sample shipments must
be established, set down on paper, and controlled.
  7. Use of moisture meters, scales, or balances and any special devices needs
to be explained stepwise in complete detail, as well as their care and maintenance
and checking for malfunction.
  8. All reports of all observations and collections are to be on pre-numbered
forms or in numbered-page notebooks furnished by the project. All entries
should be original and in ink or ballpoint pen with no erasures or data-recording
on other slips of paper. All pre-numbered pages and forms must be
accounted for with no forms discarded.
  9. All entries are to be made directly into the notebook as each measurement
is made. Supervisors should check on this immediately upon arriving for
va surveillance visit.
  10. Any confusion, or lost or broken sample, should be reported to the
immediate supervisor without fear of reprisal or penalty.
  11. Suggested data record forms are presented in the following figures: Fig. 12,

pglx130.gif (600x600)

a sample field observation form; Fig. 13, a sample collection form; and

pglx131.gif (600x600)

Fig. 11, a sample field/laboratory data sheet for maize.
  12. Supervisory field controls require careful monitoring by several varied
techniques, such as scheduled and unscheduled supervisory visits, discussions
with various employees and subjects being investigated, and comparisons with
automobile logs and daily expense logs and diaries. These techniques can be
part of the supervisory operations manual and can be kept as checklists.
Laboratory Controls
  1. Once the sample analysis procedures are established, they need to be
recorded on paper in terms suitable for the user.
  2. Analytical techniques must be followed to the letter. No alternative procedures
are permitted unless expressly authorized in the operations manual.
  3. All needed equipment must be provided and maintained in working order
using a recorded maintenance and calibration record.
  4. All reports of all tests are to be on numbered forms or in numbered-page
notebooks furnished by the project. All entries should be original and in ink or
ballpoint pen with no erasures or data-recording on other slips of paper. All
pre-numbered pages and forms must be accounted for with no forms discarded.
  5. All entries are to be made on-the-spot as the results are obtained. Supervisors
should monitor this very carefully.
  6. Any confusion, mistake, mixup, lost sample, damaged container or sample,
or spoiled sample should be reported to the immediate supervisor without
fear of reprisal or penalty.
  7. Figure 11 is a sample reporting form.
  8. Analytical controls require careful monitoring by several varied techniques,
such as scheduled and unscheduled supervisory visits, generally observing
operations if the analyses are being done close to headquarters, and
comparisons with daily logs and diaries. A supervisor should know the analytical
procedures. By watching the operator, the supervisor will form a dependable
judgment as to the analyst's expertise and working habits.
  9. Analytical operations require the use of internal controls, such as seeded,
or pre-set, standardized control samples sent through the analytical procedures
with or without the analyst's knowledge, duplicate samples analyzed at different
times by different analysts, and supervisors who can check or repeat grain
separations and other analyses.
  10. All instruments require regular calibration, especially moisture meters
and balances for grain loss work.
  (a) Moisture meters usually can be calibrated against a standardized meter
      in a national or international institute. Standardized held-in-glass or
      otherwise sealed samples may be obtained from well-known institutions
      for use in calibration. For periodic use this is more practical than using
      oven-drying moisture determinations.
  (b) Balances and scales need to be checked against a special set of weights of
      known value. Frequency of checking depends on accuracy requirements
      and the usage to which the balance is subjected.
Reporting Results
  1. All results should be on numbered forms or in numbered-page bound
  2. Results should be submitted on a regular basis, and should be checked
and otherwise followed as the work proceeds. To allow them to accumulate for
an end-of-project or delayed review is to lose an opportunity to find and
contain sources of error.
  3. Decisions on interim reports and keeping the staff informed of the data
need to be resolved on an individual project and person basis. In some cases,
being aware of what is happening and working toward overall goals will maintain
and improve work equality, although it could introduce bias.
  4. Standard terminology of weight loss should be followed. This manual
        ow - cw/ow x 100 = % loss
where ow = original, weight on a dry weight basis,
      cw = current or final weight on a dry weight basis.
  Other formulas, such as those in Chapter VI, Section B, where direct differences
cannot be calculated, may have to be substituted.
                         VIII. APPLICATION AND
                       INTERPRETATION OF RESULTS
                    A. The Chronological Approach:
                  Losses as Reflected by Use Patterns
                              J. M. Adams
  In making grain loss estimations, it is important to relate losses to the
pattern of grain consumption. If grain is left untouched throughout the storage
period, the total loss over the season can be obtained by accurately
weighing all the grain in and out of the store and comparing the totals. This
does not, however, indicate the relation between loss and time, ie, when the
loss reached a peak or whether it was related to a particular part of the season.
If at the time of removal the estimated loss is 10%, then this represents the
total loss over the storage period. In most cases, however, grain is removed at
intervals during the storage period and each quantity removed will have been
exposed to deterioration for a different length of time and will have suffered a
different degree of loss.
  If a measurement of the quantity removed is available, then estimates from
samples covering the removal period and pattern may serve to cross check with
the total loss as well as showing the pattern of loss.
  If, as often happens on subsistence farms, the amount removed is quoted in
volume terms (eg, tins), then the volume removed will be the same whether or
not the grain is damaged but the weight will be different. In this case, the
weight of grain that occupies the farmer's measure should be recorded carefully
at the beginning of the storage period. For each subsequent removal of
grain, this weight can be reduced by the percentage of loss estimated from the
appropriate sample. If samples are taken at monthly intervals and the dates of
removals are known, an approximation can be made by applying the estimated
loss to removals two weeks either side of the sampling date. To obtain the total
loss, all individual losses can be summed.
  Where removals are roughly estimated, the loss may be obtained by calculating
the percentage of the total quantity stored which was removed at each
sampling date and applying the percentage loss to this. The resulting losses are
then summed to produce an overall percentage loss, as in Table VI.
  When stored grain is regularly removed for household use, weight loss may
be measured by taking, or having the user set aside, a sample from, or taken at
the same time as, the portion withdrawn for use. The household may be
provided with an equivalent amount of grain in exchange for the test samples.
                                   TABLE VI
                 Relation Between Weight Loss and Consumption
                                        Months in Store
                 1     2        3       4         5         6        7        8
removed, %      10    10       10      10       10        10        15      25
Weight loss
in sample, %     1     2        3       5         8        12       18       25
Weight loss as %
of total stored  0.1   0.2      0.3      0.5     0.8       1.2      2.7     6.25
Cumulative weight
loss as % of
total            0.1   0.3      0.6      1.1     1.9       3.1      5.8    12.05
  This is an actual use-weighted loss of 12.05% compared with a loss of 25%
(as measured in month 8 of Table VI) if only a single, final visit had been made
and there was no allowance made for consumption (see Fig. 14). Line A of the

pglx136.gif (600x600)

top of Fig. 14 represents a farmer who holds a quantity of grain in store for
sale when the price is high and does not remove any until the date of sale, when
the store is completely emptied. Line B represents a subsistence farmer who
regularly removed grain from the store for family consumption. The total loss
in weight suffered in case B is considerably reduced because a decreasing
proportion of his total stored grain is exposed as the level of loss increases with
  The same procedure may be adopted in relation to nutritional loss, bearing
in mind that damage may cause greater losses in preparation of food where
soaking of the grain is involved. It may also be used to evaluate quality loss in
terms of money. For seed grain, the loss is the drop in germination from the
time of storage to the date the seed is required and is simply the difference
between the percentage germination recorded on the two dates.
                             CHAPTER VIII
                    B. Losses and the Economist
                      M. Greeley and G. W. Harman
  To the economist, storage losses refer to changes in the value of grain which
occur as a result of any physical change while it is in store. Alterations involving
biological changes normally reduce its value and thus involve an economic
cost. Losses may also occur during marketing and, to the extent that waste and
unintended physical alterations take place, during the primary processing of
Setting Terms of Reference
  The economist evaluates loss by assessing the cost or sacrifice borne as a
result of its occurrence. Since losses can occur at various points in the marketing
pipeline and will, if significant, have consequences for individual stores-owners
and consumers, merchants, marketing boards, etc., and to the country
as a whole, it is essential to define from whose viewpoint the assessment is to
be made. In this guide, concentration is centered on the consequences of losses
for stores-owners at farm level in developing countries.
  An attempt should be made to approximate the magnitude of the value of
losses before time is spent on trying to reduce them. If this value proves to be
low, expenditure of appreciable resources on reducing losses may not be justified.
Even when it is established that losses are sizable, consideration should be
given to the relative desirability of their reduction compared to alternative
investments. If the purpose is to increase the quantity and quality of grain
available to users, there may be other more practical and cost effective ways of
achieving this end. Examples of possible alternatives to improving storage are
measures to stimulate the use of fertilizer to increase production of grain and
changes in the marketing system to encourage stores-owners to store less by
making grain/flour available for their purpose at a fixed price throughout the
storage season. On the other hand, there are situations where no alternatives
are available, and food lost equals people starved. These situations are difficult
to resolve on an economic basis.
  In assessing the practicability of improving storage, it is essential to have the
improvement tested by stores-owners since this may reveal unanticipated problems.
One important aspect of this testing will be to determine if storage
owners are sufficiently motivated to undertake improvements in their storage
methods. Factors affecting acceptability and utility of storage improvements
are not all predictably quantifiable and require practical testing before their
benefits can be accurately assessed. National priorities may meet individual
needs and vice versa. Distribution of potential benefits should also be taken
into account since these will vary appreciably according to the type of improvement
proposed and the point in the marketing pipeline at which it is
Nature of Losses
  The physical alteration and diminution of grain in store will affect its weight
or quality. Both changes will alter its value and should be assessed separately.
Change to the nutritional worth of grain may be regarded as a particular type
of quality loss. Such losses are only relevant to the owner if they affect the
price of grain that is sold or are of sufficient size to reduce the value of grain in
other ways. An example is a reduction in an owner's capacity to work which
may occur through eating grain which has suffered a nutritional loss.
  Losses may involve other economic costs by necessitating expenditure to
reduce them and by affecting the timing and, therefore, the price of grain that
is sold. Major factors influencing the economic consequences of loss are:
scarcity of grain, the extent of seasonal price fluctuations, the time at which
the impact of loss is felt, the proportion of a crop that is stored, the extent of
the premium on better quality grain, and the opportunities for using damaged
grain in other ways.
Collection of Data
  The objectives of data collection are to ascertain, by examining the behavior
of procurers, handlers, and stores-owners, the consequences of losses incurred
by them; and, if the level of losses justifies changes in the system of storage, to
assess the likely costs and benefits of such changes.
  The exact nature and amount of information to be collected will depend on
the circumstances in each situation and the time at the researcher's disposal.
The basic minimum data necessary for reliable evaluation are as follows:
  1. Use of stored grain, preferably throughout the whole storage season
  * amount consumed by stores-owner and his dependents
  * amount sold; price obtained
  * amount used for other purposes such as for seed, feedingstuffs, making
    beer, payment of wages
  In obtaining these data, attention should be given to the reasons for usage at
a particular time, interrelations between usage and type of store, effects of a
stores-owner having more/less grain available, influence of the variety/type of
grain on its use, and the effect of varying degrees of physical damage on usage.
If regular visits are not being made to an owner, data will need to be collected
on the time at which grain from the store was exhausted and the consequences
of this for the owner. In examining usage, grain sales to any marketing authority
should be distinguished from those at village level since the price received
will probably be different.
  In addition to the amounts of grain used in various ways, the pattern of
usage should also be noted in relation to other stores which the owner may
possess. For example, is grain taken out of one store until empty or is it taken
out of more than one? Why?
  2. The marketing system
     * method of operation
     * factors determining prices received, particularly timing of sale and
       quality of grain (including any statutory regulations applicable)
     * influence of variety/type of grain
  3. Behavior of stores-owners
     * motivation for growing grain
     * degree of knowledge of losses
     * measures (if any) taken to eliminate loss
     * capability and motivation for adopting any suggested improvements
       in storage
     * work undertaken off farm; its nature, timing, and remuneration, ie,
       is the store the major source of food grains?
  4. Stores/storage practices (existing and as would occur when suggested
improvements are included)
     * materials used in construction, quantity and price
     * time taken to collect materials and build or improve the store
     * season at which constructed and alternative work at that time (on and
       off farm)
     * expected life of improved or traditional store
      * insecticides used, quantity and price
  5. General
     * purchases of grain, reason, timing, amounts, prices
     * type/variety of grain grown/stored
     * cost of seed
Methods of Collecting Data
     * published reports and economic data
     * discussions with those having detailed knowledge of the behavior and
       practices of stores-owners
     * questionnaire surveys.
  Training of field staff who will be conducting, or assisting in conducting,
questionnaire surveys should receive close attention to ensure that they thoroughly
understand the questions to be asked and the reasons for them.((9) This matter
was debated at Slough and no real consensus obtained. Some
felt that information gatherers should not understand their questions
and that the most reliable information was obtained when information was
gathered in a fixed mechanical manner.) If at
all possible, all field staff should be accompanied on initial visits, and periodically
thereafter, to ensure that questions are put without bias. Questionnaires
should be tested experimentally on a sample of participants before a complete
survey is made so that misunderstood questions can be rephrased or removed.
Use of Data in Evaluating Losses
1. Weight Loss
  The value is obtained by pricing the weight loss according to the use to
which the lost grain would have been put and the effect of its loss on the
stores-owner. For example, if the grain would have been consumed by the
owner, its replacement cost as food would normally be used; similarly, if sold,
its sale price, and, if used for seed, its cost of replacement.
2. Quality Loss
  This may be assessed by adopting a standard of quality and measuring loss
as the difference between this standard and that of the grain in the store. The
relevant standard will depend on the intended use of the grain but often it will
be that set by a marketing authority. If no such authority exists, an attempt
must be made to examine how the grain usage is affected (if at all) by the
existence of differing qualities. The standard which affects its usage should
then be adopted.
  The economic cost of the quality loss will be represented by:
              Lq = [V.sub.s] - [V.sub.a]
where Lq        = value of quality loss,
      [V.sub.s] = value of grain if it was all of a standard set,
      [V.sub.a] = value of the quality of the grain in store when used.
  Quality loss of grain intended to be used as seed is especially serious. If the
stores-owner does not realize that it is damaged, it may be planted and result in
a lower rate of germination. This loss is assessed as the difference between the
value of the crop expected from the undamaged seed and that which would be
produced from the damaged seed.
3. Indirect Loss
  This is the cost of any insecticide or other treatment used by the stores-owner
to minimize his losses.
4. Nutritional Loss
  This can be valued in the same way as quality loss by adoption of a standard.
Since this method is liable to a high degree of subjectivity, the reasons
for using a particular standard need to be clearly stated. In some cases, nutritional
loss will not reduce the economic value of grain to an owner; for
example, it may not, taken by itself, necessarily reduce its sale price.
5. Other Losses
  Stores-owners may suffer other economic costs due to losses, but the valuation
of these will be specific to particular circumstances and it is not possible to
provide more than the general principles of valuation already outlined.
Further Points to Note on Evaluation
  1. Valuation should be based on the time when the impact of loss is felt by
the owner. This will not necessarily be at the time when the loss occurs. This
factor will be of particular importance in cases when the price of grain fluctuates
appreciably during a storage season.
  2. In arriving at a final loss figure, the value of damaged grain in any
alternative or secondary use should be considered. For example, if grain intended
for human consumption was damaged and, therefore, used to feed
cattle, the loss suffered by the stores-owner would be:
              Ln = Lf - Lc
where Ln = net loss,
      Lf = value as food,
      Lc = value as feedingstuff.
  Summation of the different types of economic costs occurring as a result of
physical loss will provide an estimate of the total economic impact of losses.
Such estimates should be related to the "wealth" of the stores-owners concerned
since losses of the same value will affect poorer stores-owners to a
greater extent. In this respect, care should be taken in quoting average values.
Use of Data in Assessing Improved Method(s) of Storage
  The benefits of a system of storage are assessed by a comparison of the costs
involved with its output as measured by a valuation of grain leaving the store.
Improved storage may be reflected by a reduction both in weight and in quality
losses per unit of storage cost. The value of any additional amounts of grain
made available by a reduction in weight loss should be based on the use to
which this extra grain would be put. The value of the reduction in quality
losses is obtained by grading grain stored in both the normal and improved
manner as it leaves the store using a common standard. The amount of qualitative
benefit will be:
               Qb = Vi - Vu
where Qb = qualitative benefit,
      Vi = total value of grain leaving improved store,
      Vu = total value of grain leaving unimproved store.
  In assessing the reduction both in weight and in quality losses, it is necessary
to ascertain the level of these before improvements in storage are made. Care
should be taken that the figures obtained are representative since there may
exist appreciable variation between different seasons and stores. The costs
involved in adopting a particular system of storage may be divided into those
of materials and labor used in constructing the store and of any treatments
applied to the grain. The cost of any purchased inputs, including labor, will be
the actual amount paid. Any time spent by the stores-owner or his family on
constructing the store or treating the grain should be priced at a theoretical or
imputed wage rate. The rate used will normally reflect the wage being offered
in a type of occupation similar to that in which the stores-owner is engaged.
This rate should be taken only as a general guideline. The objective in using
any particular one is to express the cost (if any) to the owner of the time which
he and his family spends on storage by the value of the time given up on its
alternative use. In some cases, materials used to build a store will not be
purchased but gathered from fields or woods. The cost of these free goods in
evaluation should be that of the time spent in obtaining them.
  In assessing the cost of time, attention should be given to the seasonal
pattern of agricultural activity and also to the fact that the value of time at a
particular period may differ between different stores-owners according to the
amount of land and labor at their disposal.
  The three main methods of relating costs of benefits are by means of a ratio
(cost-benefit ratio), a rate of return, or by comparing the additional benefits
from taking a particular action with the additional costs incurred. The last of
these approaches is particularly suitable where the changes to an existing system
of storage are relatively small. The rate of return concept is more suited to
situations in which changes to the system of storage is extensive and sizable
capital investments are involved. Where the rate of return concept is used, the
value of grain removed from a store will be expressed as a percentage of the
store's cost. Finally, but importantly, if benefits gained over a period of years
are being compared with costs incurred at a point in time, they must be
discounted using a suitable rate of interest. The spread of benefits within the
total period is a significant factor in this procedure.
                             CHAPTER VIII
                     C. Conversion into Monetary Values
                                 E. Reusse
  After having been physically and quantitatively assessed, food losses have to
be expressed in monetary terms. This is necessary to establish a common
denominator for cost-benefit analysis, in which cost (investment in potential
improvement measures) and benefits (expected reduction of food losses) can
be weighed against one another.
  Thus, if a farmer can reduce his storage loss from 8 to 4% by means of
fumigation, and the fumigant plus amortization of plastic sheetings amount to
$3 per 500 g, then 1 kg of grain must be worth more than 15 cents to warrant
the investment. If a rice miller can raise the extraction rate of paddy rice from
63 to 66% by additional installations (including rubber rollers) and additional
controls by qualified technicians, together increasing milling cost from $2 to
$2.50 per 100 kg, then 1 kg of rice must be worth more than 17 cents to make
the improvement financially feasible. While the financial value of the rice to
the miller might be only 15 cents per kg, the economic value for the national
economy of the country concerned may be much higher, as it is when the rice
gained through the advanced milling technique can serve the substitution of
imports, thereby freeing valuable foreign exchange.
  The question is how to determine the value of a unit weight of grain. The
financial value can be one value for the individual innovator (the farmer,
trader, or processor, whether private, cooperative, or state enterprise) and a
differing economic value for the economy as a whole. The viewpoint from the
individual enterprise sphere is also referred to as the micro-economic consideration,
as opposed to the macro-economic one taken from the viewpoint of the
national economy.
  Food losses occur principally at three different levels: farm, wholesale and
processing, and retail. These levels are linked by transport. The gains in time-,
form-, and place-utility added to the food product at and between the various
levels, carrying those essential inputs as storage, transportation, processing,
packaging, financing, risk-bearing, and logistics decisions, add value to it. The
cumulative value added in the postharvest system for storable food crops in
developing countries generally amounts to between 50 and 100% of production
cost, depending on distribution radius and degree of processing involved.
  In a competitive marketing system, the value added is reflected in the market
price received for the food product at the various levels of the process. A
typical postharvest cost-price structure for rice might be as shown on the chart
on the following page.
  It follows that the physical loss of 1 kg of rice in the form of paddy occurring
at the farm level in financial terms represents only 57% of the loss of the
same quantity of rice, after milling, at urban retail level. It is, therefore, vital
to value a food loss at the farm gate or market price prevailing for that stage of
processing and for that geographical area where it occurs. For transport-inflicted
losses, the market price at point of destination would apply; for
milling losses, the price for the milled product would apply.
                                Cost per kg      Cost per kg
                                paddy            milled rice
                                                            (at 66%
                                                           ext. rate)
  farm gate value                     10
    + transport                         1
  rural assembly market value         11
    + bagging, transportation, etc.    1.5
  provincial market value             12.5
    + milling cost                     1.5
  milled rice (in terms of paddy)    (14.0)                    21
    + bagging, transportation, etc.  ( 1.6)                     2.4
                                      ----                     ----   
  urban wholesale market value       (15.6)                    23.4
    + packing and other retail cost  ( 1.8)                     2.7
                                      ----                     ----   
  urban retail market value          (17.4)                    26.1
  urban retail market price            ...                     27
  Since market and farm gate prices are subject to seasonal fluctuations, when
working at a national level, annual average prices should be used. To eliminate
abnormal annual crop situations, the average over the past three years may
best be taken. An inflation factor, however, should be added, if necessary,
since implementation of any remedial measures will usually be delayed.
  So far we have discussed the financial valuation of food losses typical for
micro-economic consideration; let us now look at a few major situations where
under macro-economic consideration the financial, price-based valuation has
to be corrected or substituted by an economic valuation. As the examples will
show, these situations typically arise because of government intervention in the
price structure:
  1. The situation of subsidized producer or consumer prices
     a. Subsidized producer (farm gate) prices: for economic valuation the
        subsidy element has to be eliminated (downward correction of financial
     b. Subsidized consumer prices: same applies, but upward correction of
        financial values.
  2. Overstated official foreign exchange rate of national currency: In such
situation domestic price development lacks close correlation to world market
prices. This fact has little relevance in a closed food economy, ie, where the
country is neither a regular exporter nor importer of the staple food crops
(products) in question. In an open food economy, however, where food losses
are affecting the foreign exchange intensive marginal area of export surplus or
import substitution, those losses have to be valued at the average annual FOB
export or CIF import price, respectively, under application of a shadow rate of
the foreign exchange involved in converting to national currency values,
shadow rate being understood as the rate expected to prevail under conditions
of free floating exchange rates. The FOB or CIF value thus established in
national currency has to be deflated by the transportation cost between the
geographical area where the field losses are occurring and the seaport. This
would include the simplifying assumption that, in most developing countries,
consumption of imported staple foods is concentrated in geographical areas
near ports of importation.
  3. A shifting area between financial and economic valuation is entered
when food losses in government reserve stock and price stabilization schemes
have to be valued. Since the selling price of those stocks in most cases is related
neither to market value nor internal cost price calculation, the value applied to
stock losses should at minimum reflect the full unit cost of operation, including
accumulated storage cost over the recycling period which may extend over
two to three years.
                              APPENDIX A
                             SAMPLING GRAIN
             1. Comments on Probing Techniques and Probes
  a. In this volume the terms trier, probe, thief, and spear are used interchangeably.
  b. A compartmented grain trier should be used that will reach the bottom of
the container with each compartment 15 cm long (see Fig. 15). Noncompartmented

pglx150.gif (600x600)

grain triers should not be used to sample grain.
  c. In probe-sampling a bin from the top, the probe or trier should be
inserted in the grain at an angle of about 10 degrees from the vertical, with the
slots closed. The probe should be opened while the slots are facing upward.
While the slots remain open, the probe should be moved up and down so that
all openings may be filled. The probings should be emptied onto a sheet and
coned or quartered or mechanically divided to sample size.
  d. In bag sampling, the trier should be inserted from a corner diagonally
across to the farthest corner.
             2. Techniques for Sampling Bagged Produce (10) Adapted
from Trop. Stored Prod. Inf. 31: 37 (1976).)
                               P. Golob
The Importance of Sampling
  Quality is an important factor which dictates the value of a commodity. It is
judged by the overall appearance of the produce and will be adversely affected
if there are holes in the grains caused by insect attack, discolored grains from
mold damage, shrivelled grains, cracked and broken grains from bad handling,
or rodent hairs and droppings.
  Infestation by stored product insect pests before harvest is common so that a
consignment may enter a store having a low-level infestation. Depending on
climatic conditions, the pests can multiply rapidly and greatly damage the
crop. Thus it is of vital importance that the infestation be detected as early as
possible, preferably before storage begins. The crop must be inspected and
sampled as it is unloaded from lorries or railway trucks before it is stacked for
  As a commodity deteriorates during storage, it loses value. For the government
of an exporting country this can mean a loss of foreign exchange. For the
subsistence farmer the losses result in less food to eat. Poor storage conditions
can aid the increase of insect and mold populations and bad storage structures
can allow the entry of rodents. It is, therefore, important to continually check
stored produce in order to monitor changes.
  For practical reasons it is not physically possible to examine every grain in a
consignment. Thus the quality of the whole has to be judged on the basis of a
sample. The sample must be representative of the individual bag or stack from
which it is drawn. In this Appendix the various techniques which may be used
to obtain representative samples from bagged commodities are described and
their limitations are discussed.
Sampling From Stacks (See also Chapter IV)
  The principles of sampling from stacks apply to all types of stacking situations
whether in a large warehouse or godown, a ship, a train or lorry, in a
trader's store, or a farmer's crib. In practice, however, it may not be possible
to put all the principles to use due to the accessibility of the stack.
  Consignments of produce can be divided into sectors on the basis of location.
For example, in a ship the commodities may be segregated in different
holds, wherein each hold can be regarded as an individual sector in terms of
climatic and other physical influences. Similarly each boxcar of a train might
be regarded as a single sampling entity. Each sector must be identified and
sampled individually.
  As the conditions within each sector may fluctuate as much as those affecting
the total consignment, it is important to obtain samples which are representative
of the sector from which they have been drawn. Each sector itself can
be stratified and samples must be drawn from all areas within each sector, ie,
from the top, middle and bottom, left and right, center and periphery. Removing
samples from these strata should be performed at random.
  Twenty-four sampling points from a cuboidal stack should provide an accurate
representation of the stack. However, taking samples from as many points
as this for all but the largest stacks is uneconomical and unwarranted, as fewer
sampling points will give as accurate a pattern. Five sampling points are recommended
for wagons and lorries of up to 15 tonnes, eight points for up to 30
tonnes, and eleven points for containers up to 50 tonnes, as shown in Fig. 16.

pglx151.gif (486x486)

Number of Sacks From Sector (See also Chapter IV)
  The above recommendations are inappropriate for sampling stacked bags
because they regard the stack as a two-dimensional structure. They take no
account of the difference between top layers in a container and lower layers
where dust and insects would tend to accumulate. They disregard any possible
changes that affect one side of the stack rather than the other and they ignore
the fact that stacks are accumulations of individual units that can be separately
  Practical experience has shown that the optimum number of samples to be
obtained from a large consignment (over 100) of sacks is given by the square
root of the total. Jelier (1) suggests that for sectors of 10-100 bags, 10 bags
should be taken at random and for up to 10 bags, each bag should be sampled.
Thus, from a lorry having perhaps 100 sacks the sample would consist of 10,
which would represent all areas of the stack.
  Bags drawn from the stack using the above rules constitute the initial sample
which should be taken randomly but at the same time should be representative
of the whole stack. In practice, when obtaining the initial sample from a small
stack as found on a lorry, it is not possible to sample entirely at random. The
structure and size of the stack determine from which areas bags must be
chosen, so that the number of bags from which the random choice has to be
made is relatively restricted and may only be three or four bags.
  In many cases it is not possible to sample randomly from all sectors of very
large stacks. Only by breaking the stack down would most bags become available.
Thus only a relatively small area can be sampled. Effort should be made
to get to bags in the middle. To do this, several layers of bags at the top of the
stack should be removed and a bag in the sixth or seventh layer obtained for
observation. This practice in no way utilizes randomized searching for initial
The Bag as the Sample Unit
  Sampling of the commodity within the bag must be random so that every
grain has a chance of being picked. Many of the sample-taking procedures are
not random but tend to be haphazard, resulting in having human bias. With
haphazard sampling, such as using a spear or trier, every grain does not have a
chance of being picked as just a portion and not the whole bag is the sample
Methods of Obtaining Samples From Sacks
1. Spear Sampling
  Bag sampling with a spear or trier is practiced throughout the world. There
are many types and variations of sampling spears, the commonest of which is
illustrated in Fig. 17. Bag spears are usually cylindrical in shape and between

pglx152.gif (486x486)

40 and 45 cm in length with a diameter of 2.5 cm, except at one end which is
drawn to a point.
  The tube is open on one side to allow grains to fall into a collecting channel,
which passes back along the length of the spear and opens out through the
handle. This type of spear is used for collecting large particled material, such
as maize grains or coffee berries. Other types of spears may be of similar
design but narrower for collecting smaller grains such as wheat and sorghum
(Fig. 17B) or simply open-grooved lengths of metal attached to a handle (Fig. 17C).
  The spear has several good features; it is cheap, simple to use, and is a quick
way of obtaining grain from bagged produce. The tip of the spear is pushed
into the bag, and the body with the open side face down is inserted for the
required distance. Grain is collected in the channel by twisting the spear so that
the open side is turned upwards. On withdrawing the spear from the bag, the
grain is tipped out of it into a container. If the spear is inserted into the sack at
an angle, with the point uppermost, grain entering it can pass straight into a
container without the spear being removed, so that a large sample can be
obtained (see Fig. 18). Generally six or more samples are removed from each

pglx153.gif (486x486)

sack to make up a primary sample.
  Because of its widespread use, the faults of spear sampling are usually
disregarded. However, the disadvantage of the spear is so fundamental that it
negates most of the results obtained upon analyzing samples collected by this
method. When a spear is inserted into a sack either horizontally or at an acute
angle, only a very small volume of the sack commodity is sampled, ie, precisely
that material that actually falls into the spear cavity. The sack is not sampled
randomly; the grains picked depend on the haphazard method used to insert
the spear into the bag (see Fig 19).

pglx154.gif (600x600)

  Many elements of stored crops (such as protein and vitamin contents) are
generally constant throughout a single sackful of produce or any variations
that do occur are insignificant. Produce moisture content and insect numbers,
however, may not be constant throughout the bag. Insects, in particular,
distribute themselves neither uniformly nor randomly. They are often found in
pockets associated with the dust or meal material at the bottom of the bag or in
areas of local heating and wetting.
  Producing information on insect numbers in a sack using a spear sample can
lead to erroneous conclusions and be totally misleading, either overestimating
a population or more frequently underestimating it. Examples of the way in
which this could occur are shown in Fig. 19. In Fig. 19A, a large population of
insects crawling on the bottom of the bag could easily be missed by spear
sampling; it is difficult to sample very close to the bag fabric, top and bottom.
Observing or missing a population such as this could influence the decision to
treat the commodity to eradicate the infestation, resulting in heavy losses of
produce. In Fig. 19B, small pockets of two or three insects could by chance be
picked up by a spear sample. Six insects in 100 kg of maize may not require
eradicating if the produce is not going to be stored for long periods. However,
six insects in a 500-kg sample is equivalent to 1,200 individuals in a 100-kg bag
if randomly distributed, whereas there may be less than ten in the whole bag.
  Thus spear sampling can produce grossly misleading results and should be
avoided. A compartmented probe (Fig. 15) should be used whenever a probe

pglx150.gif (600x600)

sample is taken. Compartmented probes are available in bag size, as in Fig. 15,
or in larger sizes for probing deeper piles in bins, wagons, etc.
2. Coning and Quartering
  Sampling at farmer and trader level requires a procedure that is simple,
cheap, and accurate. Coning and quartering is such a method.
  When a bag of commodity is opened and the produce is tipped onto the
floor, the contents naturally assume the shape of a cone. By shovelling material
from the periphery of the cone to the apex, while circling the periphery,
complete mixing and randomization of the produce will occur. This mixing
needs to be done for 3 to 4 min at least five times round the circumference.
Division of the bulk into halves and then quarters using a flat piece of wood or
quartering irons produces four samples of very similar properties. From a
100-kg bag, each sample would be 25 kg, too large to be useful. By further
subdivision, using the same coning and dividing procedure, each quarter can
be divided into 1/8th, 1/16th, 1/32nd, etc., subsamples.
  Sampling error by coning and quartering is about 10%, which is much more
accurate than spear sampling. This method is time-consuming, however, and
can only be used when a small number of bags require sampling. For continuous
sampling at marketing board or export level, the produce flow sampler can
be used.
3. Sieving
  The three techniques described above comprise methods by which small
quantities of material can be removed from the bulk for analyses or inspection.
An estimate of dust content or insect number in a sack can best be
obtained by using a sieve. Unlike the methods discussed above, a sample
representing the whole sack is not obtained. Instead the commodity is divided
on the basis of particle size. Smaller particles, including insects, pass through
the sieve mesh whereas large particles pass over it and are returned to its bag.
  A type of bag sieve is shown in Fig. 20. The produce is tipped into a hopper

pglx156.gif (600x600)

located above the sieve mesh. On oscillating the mesh by a simple handcranked
gear mechanism, the produce flows out of the hopper and over the
mesh surface. The bulk of the produce passes back into the sack, and dust and
insects are collected in a tray slung below the mesh. The mesh size can be
altered as required depending on the particle size of the produce being sieved.
Tests have shown that more than 90% of all dust and insects is removed using
this apparatus, the recovery of insects being independent of the population
Apparatus for Sample Reduction
  Sample reduction can be performed by coning and quartering (see above) or
by using specific apparatus designed for this purpose. Generally, this equipment
divides the sample into halves which then have to be passed repeatedly
through the divider until a workable sample is obtained. Such a divider is the
Boerner divider (Fig. 21).

pglx157.gif (540x540)

1. Boerner Divider (Conical Type)
  This is a gravity mechanical divider which works on the same principles as
the produce flow sampler (see below). The produce flows out of a hopper and
around a cone but, unlike the PFS which takes a single sample of up to 12% of
the total, the Boerner simply divides the total in half. Instead of the four
sampling points of the PFS, the Boerner has a series of channels around the
periphery of the cone. As the commodity flows into the channels, it is directed
into one of two collecting points. The direction of flow of the channels alternates
around the periphery so that every other one directs the flow into the
same collecting pot. The Boerner is an accurate method of sample division.
2. Box Divider
  A simplified version of the alternate-channel separation is the box divider
shown in Fig. 22. It is less expensive than the Boerner, more easily transported,

pglx158.gif (437x437)

less subject to damage (and when damaged more easily repaired), and does
almost as accurate a job as the Boerner. In using, care should be taken that the
slot widths remain uniform and are not bent out of position.
3. Motorized Divider (Centrifugal Type)
  In this divider the seed falls into a shallow rotating cup from which it is
flung into a chamber divided into two or more outlets at the bottom. An
example of this type is the Gamet divider (Fig. 23). In dividers of similar

pglx159.gif (486x486)

design, the grain may be delivered from a rotating spout over a number of
containers or over a cone with adjustable dividing blades at the bottom which
may be arranged to separate off any desired fraction.
4. Produce Flow Sampler
  The produce flow sampler (PFS) (Fig. 24) is a device designed by the British

pglx160.gif (600x600)

Tropical Stored Products Centre for taking samples from whole bags of grain.
The produce is tipped into an upper hopper which has an opening at the
bottom. The opening is closed by a bung until sampling commences. On
removing the bung, the produce flows down and around a cone and, because
the apex of the cone is placed exactly under the center of the hopper opening,
the flow of produce is equal all around the cone. Samples are separated from
the main flow at four points at the base of the cone, the points being spaced
equally around its periphery. The bulk of the produce is recollected in a sack
attached by hooks to the bottom of the collecting funnel. Sampling time is 20
sec for a 100-kg bag. Size of the samples can be altered by changing the vent
that covers each sampling point.
  The PFS was originally designed for sampling bags as they were off-loaded
from lorries before the produce went into store. For this purpose, the PFS is 8
ft high but the length of the legs can be lowered if required. All flowable
commodities can be sampled using the device.
  Tests on the accuracy of this method have been performed using bags of
produce containing a small percentage of grains stained with a dye strategically
placed at different parts of the bag to simulate pockets of defective produce.
With groundnuts, for example, containing 5% dyed kernels, the percentage of
stained nuts in the samples ranged between 3.4 and 6.0% in 15 tests, and for
maize and wheat which had 1% dyed grains, the recovery range was 0.3-1.5%
in 30 tests. Thus accurate recoveries were obtained.
  The PFS method of sampling is accurate because, unlike spear sampling, the
whole bag is sample unit and the sample is obtained randomly, each grain
having a chance of being picked.
  Samples obtained from bagged produce must be both representative and
random of that produce. Sampling using a spear is not random and does not
result in a representative quantity of produce being taken. At farmer or trader
level, coning and quartering do provide accurate results and samples of similar
quality at marketing board or export level are best obtained with the PFS.
Sieving, although not strictly a sampling procedure, can give accurate estimation
of surface insect population. Subdivision of primary samples must be
random and the Boerner, box, and Gamet dividers fill this function. However,
the equipment for subdivision of samples is relatively sophisticated and is not
always available. It may be more practical and almost as reliable to reduce
samples by coning and quartering.
                           Literature Cited
1. JELIER, G. Sampling of grains, milled products, starch products, and potato starch. Int.
      Assoc. Cereal Chem. ICC Standard 101 (1970).
                              APPENDIX B
                           B. Drew and T. Granovsky
  Sample selection by means of randomization is not an unorganized hit-or-miss
process. It is a rather formal protocol-dictated process to assure that an
intentional or unintentional bias will not be introduced during sample selection
and sampling.
  A random sample means that each and every unit (ears, plants, baskets,
row, farm) in a population has an equal chance of being selected. It means that
the selection of "good looking," or "typical," or "some of the good ones and
some of the bad ones," or those within a convenient distance will be avoided.
To select on such bases neglects the principle that each sample should have an
equal chance of being selected. Any such selection, therefore, introduces bias.
Random selection usually means that randomization must be done by the
project planners and supervisors although it may be accomplished at the working
level and situations may be classified by the state of knowledge into 1)
where information about the size of population to be sampled is available
before field-workers are sent out, or where field-workers are competent to
randomize, and 2) where information about population sizes is not available in
advance and field-workers are not competent to randomize.
  In either situation, the only way to select at random is by a table of random

pglx164.gif (600x600)

numbers. Any other means simply will not give the total randomization that is
provided by a table of random numbers.
  A table of random numbers (see Table VII) should be used by a fixed
procedure determined in advance. To do this, one should know in advance
what units and how many are to be taken as the sample: ears of corn, bags of
grain, farms lying on map coordinates, etc. The procedure for taking the
sample also needs to be established in advance.
  1. Plan the selection of elements to be sampled in advance. Decide what is
to be selected: rows, bags in piles requiring predetermination along a three-dimensional
grid, bags as they are moved for sampling, etc. Decide how many
of these units are to be taken for the sample.
  2. Number the units in any convenient way starting with 1 and going as high
as necessary.
  3. Use the table of random digits. Start at any point in the table and proceed
to read off pairs of digits in any direction - up, down, sideways, diagonally.
  4. Write down the pairs of digits as they occur. Skip any numbers that are
repetitions, or that are bigger than the total number of units numbered in
step 1.
  5. When you have written down the number of units to be taken in the
sample, stop.
  6. Sample those units whose numbers have been listed.
  7. Each time the table is used indicate the starting pair of digits by circling.
Do not start at the same place again.
Cases Classified by Situation
  1. Where units can be numbered in advance:
     Cribs on a farm
     Baskets in a building
     Houses in a village
     Stacks in a field
  2. Where units are encountered sequentially:
     Bags being unloaded from a truck or boat
     Farmers coming to market
     Farms located along a road
  3. Where units can be designated by coordinates:
     Map coordinates
     Three-dimensional (a pile in a warehouse)
Special Instructions for Map Coordinates
Map Coordinates Method 1 (Preferred Method)
  Number every grid point on the map. Leave out grid points that are inaccessible.
Choose pairs of random digits as given earlier. If there are more than 100
grid points, follow the same procedure but use triples of random digits.
Map Coordinates Method 2 (Alternative Method)
  Consider the vertical (north-south) coordinates to be units to be sampled.
Number them from 1 up and use random numbers to choose as many coordinates
as are needed in the sample. In this case do not skip repetitions.
  Then consider the east-west (horizontal) coordinates to be the units. Number
them from 1 up and use random numbers to choose as many coordinates as
are needed. As each coordinate is chosen, pair it with the next unused one of
the N-S coordinates from the first set. Repetitions are only skipped if they are
paired with the same N-S coordinate.
                              APPENDIX C
                            MOISTURE METERS
                                Part 1
               Guidance in the Selection of Moisture Meters
                   for Durable Agricultural Produce((11) Adapted from
Trop. Stored Prod. Inf. 21: 19 (1971).)
                                T. N. Okwelogu
  The market for moisture meters is both specialized and growing, and there is
a need for special attention to the selection of meters. The manufacturer aims
to reach as many possible users as he can, while the prospective buyer wants to
know about as many meters as he can before investing in any model. Over the
years 1966-70 enquiries about moisture meters have been received at The Tropical
Stored Products Centre at the rate of approximately 100 a year. These
enquiries have varied from wanting to know if a particular meter had a supply
address in the locality of the enquirer, to seeking advice on what meter should
be used for a specified purpose.
  This statement is not a treatise on moisture meters, but an attempt to help
the prospective buyer dealing with durable agricultural produce to determine
which moisture meter best meets his requirements.
                          Sources of Information
  The three principal sources of information available to the prospective users
are 1) newspapers, magazines, and journals, 2) manufacturers' brochures, and
3) organizations in a position to give unbiased information about moisture
  Some newspapers, magazines, and journals, which occasionally contain information
about meters, include the Financial Times, Electronic Age, and
Power Farming. While manufacturers are always helpful in supplying data
about their own range of meters, information about a wider range of meters
will be more likely obtained from organizations having unbiased interest in
these instruments. Examples of such organizations are 1) Tropical Stored
Products Centre (Tropical Products Institute), Slough, England, 2) Grain
Storage Department, Pest Infestation Control Laboratory, Ministry of Agriculture,
Fisheries and Food, Slough, England, 3) National Institute of Agricultural
Engineering, Wrest Park, Silsoe Beds, England, and 4) Grains Division,
Agricultural Marketing Service, U.S. Department of Agriculture, Agricultural
Research Center, Beltsville, MD 20705. Articles on moisture meters sometimes
appear in the publications of these and similar organizations.
  Tables VIII and IX give details of some available moisture meters, particularly

pgl81680.gif (600x600)

used. These details are based on information provided by the manufacturers of
the meters.
  With every piece of information, it is important to ask the question: Is this
information sufficient for a decisive opinion to be formed about the meter?
Where the answer is no, further enquiries should be made.
Factors to Consider in Making a Choice
  It can be seen from Tables VIII and IX and in Parts 2 and 3 of Appendix C
that there are several meters for any specific purpose. For satisfactory selection,
the following factors should be carefully considered:
  1. Meter types and their implications.
  2. Characteristics of the commodity.
  3. Requirements of the work for which a meter is sought.
  4. Business considerations.
Principles and Implications of Meter Types
  Most manufacturers indicate the principles upon which the action of their
meters is based. An appreciation of the implications of such principles will be
of considerable value in deciding which of several meters will be the most
suitable. The meters commonly used with durable agricultural products fall
into five groups, according to the principles of their action:
  1. Those involving chemical interaction between calcium carbide and the
     product water, with the evolution of acetylene gas, the pressure of which
     is subsequently measured.
  2. Those involving heat-drying of the product, the attendant loss ascribed
     to evaporated produce water (Fig. 25).

pglx174.gif (600x600)

  3. Those involving measurement of electrical conductivity (or resistance) of
     the product, since the value of this property is related to the moisture
     content, within a suitable range of moisture contents (Fig. 26).

pglx175.gif (600x600)

  4. Those involving measurement of the dielectric constant of the product
     (or capacitance of the electrical system of which the product is a component),
     since the value of this property changes with the moisture content,
     within a suitable range of moisture contents (Fig. 27).

pglx176.gif (600x600)

  5. Those involving measurement of that atmospheric relative humidity
     which is in equilibrium with the product moisture, since, under equilibrium
     conditions, there is a definite relation between the moisture content
     of a product and the ambient relative humidity (Fig. 28).

pglx177.gif (600x600)

  Although it is tempting to try to list the advantages and disadvantages of the
meter types, this approach is ineffective in providing buyers with adequate
guidance. For example, although many resistance meters require a ground
sample, use a small sample, or test products with a relatively short range of
moisture content, there are others in the same group which do not require the
sample to be ground, which can test large samples (by using probes on whole
sacks), or have an extended range of operating moisture contents. There are,
nevertheless, certain outstanding group features to be noted: Heat-drying
methods require a suitable source of power supply or fuel, which may not be
available. Methods based on the evolution of acetylene gas require regular
supplies of fresh calcium carbide, which is not a safe commodity to handle by
post, because of the risk of explosion. Meters measuring the intergranular
relative humidity require, first, a knowledge of the relation between the
produce moisture content and the relative humidity of the intergranular air;
secondly, a periodic check on their calibrations; and thirdly, in some cases,
large quantities of produce which must have remained undisturbed for some
time prior to testing.
  The electrical meters are faster and, in the main, less demanding on calibration
checks, but require skilled servicing. Also, they give less reliable readings
outside the middle region of the range of moisture contents for which they are
calibrated. The accuracy of the probe-type electrical meters is affected by
variations in the pressure exerted by the produce on the electrodes, while the
consistency of the readings of those meters which measure the dielectric constant
is affected by inconsistent packing of the sample in the test chamber.
  Attention has been focused above on the less favorable features of the meter
groups mainly because they are more likely to be overlooked. Information on
the merits of any meter will not normally be difficult to obtain, and Tables
VIII and IX show the relative merits of the meters discussed in the present
Characteristics of the Commodity
  The commodity to be tested imposes a number of limitations, and these
must be taken into account when considering the use of any meter. Perhaps the
best way to do this is to answer questions such as the following:
  First, is the chemical nature or any normal pre-treatment of the produce
likely to interfere with the use of the meter? For instance, meters measuring
electrical conductivity may not be suitable for produce, like salt-fish, which
will become highly conductive when damp. Again, for commodities like dried
egg or milk, a heat-drying meter may not be suitable.
  Second, is the moisture content to be measured outside the range for which
the meter is calibrated? For example, very few electrical meters are known to
be suitable for a product such as tea whose moisture content is normally
required to be below 5%, that is, outside the range of moisture contents for
which most electrical meters are calibrated.
  Third, is the milling property of the produce incompatible with the effective
use of the meter? For example, commodities such as macadamia nuts, palm
kernels, copra, and cashew nuts are not amenable to grinding.
  Fourth, are the unit size and shape of the produce likely to affect the
efficient use of the meter? Construction of the meter may be such that it
cannot be pushed into floury or powdery produce without hampering the
measurement of moisture. Again, larger products like cocoa beans, unshelled
groundnuts, cashew nuts, and pieces of illipe nuts (Shorea spp.) will present
packing problems with some meters.
  If the answer to each of the above questions is an unqualified no, then the
meter may be considered suitable for the product. But a yes answer can make
all the difference between a meter being chosen or rejected. In such a case,
steps should be taken to see what, if anything, has been done to solve the
problem, either by the manufacturer or by someone else.
Nature of the Situation Needing a Moisture Meter
  In a summary of this kind, it is not easy, even if it is possible, to cover all the
situations where the use of a moisture meter may be desired. However, such
situations are likely to fall into one or the other of the following categories:
  1. Knowing whether grain is at the right stage for harvesting.
  2. The processing (eg, drying, milling, or storage) of foodstuffs.
  3. Bulking or packaging for storage.
  4. Commercial transaction, where moisture content is part of the basis for
  5. Produce inspection including loss estimates.
  All the above situations require moisture meters which are not fragile, which
are consistently accurate within limits acceptable for the particular purpose,
and whose performance is little affected by the operating conditions of space,
temperature, pressure, light, dust, or wind. They also require, to a certain
extent, meters that are simple to operate, portable, and capable of taking
remote measurements, as with probe-electrodes, or stem hygrometers, or that
samples be taken of the material for laboratory testing.
Operational Considerations
  The purpose for which the use of a meter is usually contemplated is twofold:
to increase or improve productivity (that is, the flow of goods and services),
and to ensure economical operations. The usefulness of the meter can be
improved by employing one which can give results rapidly; for which spares
and facilities for servicing or calibration are easily available; and which does
not depend on sources of operating power that run out, break down, or
become short in supply (eg, battery, mains supplies, gas, paraffin, and other
  Economy of operation implies keeping to a minimum both capital and
operating costs or increasing the return to unit cost. Additionally, even though
it may have been purchased for a specific grain, the wider the range of commodities
that a meter can test, the more flexible and economical may be its
total use. Likewise, the less destructive a test is, the less will be the incidental
loss of material caused by the use of a meter. Although this kind of loss may
appear small, it must be realized that its magnitude will depend on how much
produce is damaged at each test, and how many times such tests are performed
on a given product.
  Few meters, if any, can win the top position in every conceivable area of
consideration, and there is no magic formula for choosing a meter. Where a
choice has to be made, however, all known factors need to be considered. This
implies having adequate information about as many meters as possible, and
then carefully checking the meter descriptions against the requirements.
  The buyer must have a knowledge of the commodity to be tested and the
accuracy required of a determination of its moisture content; the availability
of the meter, and the cost of operating it; the conditions under which the meter
will be operated; the ease of obtaining spares and facilities for servicing or
calibrating the meter; and the type of power supply required and available.
When a provisional choice has been made, it is often advisable to obtain the
meter on loan for trial before buying. This will make it possible to verify
certain claims which may not be possible otherwise. For example, the buyer
may discover that the meter does not give as "precise, error-free, and effortless
moisture measurement" as he was made to expect. He may discover, too,
that although the meter is calibrated for rice, it in fact needs a different
calibration for his own type of rice.
  Choosing a moisture meter must be approached from both commercial and
technical aspects, and requires a critical appraisal of many variables.
                                  APPENDIX C
                                    Part 2
       Table of U.S. Department of Agriculture, Federal Grain Inspection
         Service List of Moisture Meters Used in the United States and
                      Their Manufacturers, April 1978(a)
                      Principle of
 Name of Device         Operation            Manufacturer or Distributor
American Moisture     Infrared heat-      American Farm Equipment Co.
Tester -- Model       ing -- direct       340 E. Main St.
M-20                  reading             Lake Zurich, IL 60047
Apollo Microwave      Loss on drying      Apollo Microwave Products
Laboratory            using microwave     6204 Official Road
                      energy              Crystal Lake, IL 60014
Auto-aquatrator       Karl Fischer        Precision Scientific Group
                      Method              3737 West Cortland St.
                                          Chicago, IL 60647
Brabender, C.W.       Thermobalance       C.W. Brabender Instruments, Inc.
Rapid Moisture                             50 East Wesley St.
Tester                                     South Hackensack, NJ 07606
Brown-Duvel           Distillation        Burrows Equipment Co.
Moisture Tester                            1316 Sherman Ave.
                                          Evanston, IL 60204
                                          Gerber Industries
                                          P.O. Box 1387
                                          Minneapolis, MN 55440
                                          Seedburo Equipment Co.
                                          1022 West Jackson Blvd.
                                          Chicago, IL 60607
Burrows DMC-700       Dielectric          Burrows Equipment Co.
                                          1316 Sherman Ave.
                                          Evanston, IL 60204
                                          Dickey-john, Inc.
                                          P.O. Box 10
                                          Auburn, IL 62615
Burrows Moisture      Capacitance         Burrows Equipment Company
Recorder                                   1316 Sherman Ave.
                                          Evanston, IL 60204
Burrows Safe          Capacitance         Burrows Equipment Company
Crop III Moisture                          1316 Sherman Ave.
Tester                                     Evanston, IL 60204
Burrows Model 400     Capacitance         Burrows Equipment Company
(Radson) Moisture                          1316 Sherman Ave.
Meter                                      Evanston, IL 60204
Buhler MIAG Rapid     Thermobalance       The Buhler Corporation
Moisture Tester,                           P.O. Box 9497
Type MLI-400                              1100 Xenium Lane
                                          Minneapolis, MN 55440
Cera-Tester           Capacitance         A/S N. Foss Electric
                                          Slangerupgade 69
                                          DK 3400 Hiller[phi]d, Denmark
                      Principle of
Name of Device        Operation           Manufacturer or Distributor
Delmhorst             Conductance         Delmhorst Instrument Co.
Moisture Detector                          607 Cedar St.
                                          Boonton, NJ 07005
Dickey-john DJ1S      Dielectric          Dickey-john, Inc.
                                          P.O. Box 10
                                          Auburn, IL 62615
Dickey-john           Dielectric          Dickey-john, Inc.
Forage Moisture                            P.O. Box 10
Tester                                     Auburn, IL 62615
Dickey-john           Dielectric          Dickey-john, Inc.
GAC-II                                    P.O. Box 10
                                          Auburn, IL 62615
Dickey-john           Dielectric          Dickey-john, Inc.
GAC-III                                    P.O. Box 10
                                          Auburn, IL 62615
Digital Moisture      R.F. Capacitive     Diversified Engineering, Inc.
Meter Model DM/6      Measurement         2022 Sledd St.
                                          Richmond, VA 23220
Grain Quality         Near IR             Neotec Instruments, Inc.
Analyzer                                   2431 Linden Lane
                                          Silver Spring, MD 20910
Higropant             Conductance         National Instrument Co., Inc.
Moisture Meter                             4119 Fordleigh Road
                                          Baltimore, MD 21215
Humidimetre           Dielectric          Cedem, Division Instrumentation
Digital HD. 2000      Fully Automatic     Agricole Et Alimentaire
                                          33-5 rue Jean Baptiste Charcot
                                          92400 Courbevoie, France
Insto-I Moisture      Dielectric          Dickey-john, Inc.
Tester                                     P.O. Box 10
                                          Auburn, IL 62615
Insto-II Moisture     Dielectric          Dickey-john, Inc.
Tester                                     P.O. Box 10
                                          Auburn, IL 62615
KF-4B Aquameter       Karl Fischer        Beckman Instruments, Inc.
System                Method              Scientific Instruments Div.
                                          P.O. Box C-19600
                                          Campus Dr. at Jamboree Blvd.
                                          Irvine, CA 92713
KPM Aqua Boy          Conductance         Chatham International Corp.
MS-I                                       P.O. Box 377
                                          Larchmont, NY 10538
Koster Crop           Heating             Koster Crop Tester, Inc.
Tester                                     4716 Warrensville Ctr. Rd.
                                          North Randall, OH 44128
Marconi Moisture      Conductance         Marconi Instruments
Meter Type                                 100 Stonehurst Court
TF-933C                                    Northvale, NJ 07647
Mettler LP 11         Infrared            Mettler Instruments Corp.
                      thermobalance       20 Nassau St.
                                          Princeton, NJ 08540
                      Principle of
Name of Device        Operation           Manufacturer or Distributor
Model G8R or          Radio frequency     Moisture Register Company
Model G9              dielectric power    6934 Tujunga Ave.
                      loss factor         No. Hollywood, CA 91605
Moisture Teller       Heating             Harry W. Dietert Company
Model 276                                  9820 Roselawn Ave.
                                          Detroit, MI 48204
Motomco Moisture      Capacitance         Motomco, Inc.
Meter -- Models                            267 Vreeland Ave.
919, 840, and 430                          P.O. Box 300
                                          Patterson, NJ 07513
Ohaus Moisture        Infrared heating     Ohaus Scale Corporation
Determination         and balance         1050 Commerce Ave.
Balance                                    Union, NJ 07083
Optical Moisture      Infrared            Anacon, Inc.
Analyzer              absorption          P.O. Box 416
                                          Burlington, MA 01803
Pier Moisture         Infrared            Neotec Instruments, Inc.
Analyzer              Reflectance         2431 Linden Lane
                                          Silver Spring, MD 20910
Protimeter Grain      Conductance         Cosa Corporation
Moisture Meters                            17 Philips Parkway
                                          Montvale, NJ 07645
Quik-Test             Dielectric          Dickey-john, Inc.
Moisture Tester                            P.O. Box 10
                                          Auburn, IL 62615
Schenk Moisture       Capacitance         Schenk Moisture Engineering
Monitor               and/or              R.R. 7, Box 78
                      Conductance         Vincennes, IN 47591
Semi-Automatic        Thermobalance       Haake, Inc.
Moisture Tester                            244 Saddle River Road
                                          Saddle Brook, NJ 07662
Skuttle Moisture      Conductance         Skuttle Manufacturing Co.
Meter                                      Electronic Division
                                          Canfield, OH 44406
Steinlite             Electronic          Seedburo Equipment Co.
Moisture Tester       Impedance           1022 West Jackson Blvd.
                                          Chicago, IL 60607
Super-Conti           Capacitance         A/S N. Foss Electric
                      Automatic           Slangerupgade 69, DK 3400
                      Recording            Hiller[phi]d, Denmark
Super-Matic I         Capacitance         A/S N. Foss Electric
                      Print-out           Slangerupgade 69, DK 3400
                                          Hiller[phi]d, Denmark
T & M Vacuum          Infrared vacuum     Townson & Mercer, Ltd.
Moisture Tester       thermobalance       Scientific Equipment
                                          Beddington Lane
                                          Croydon, England
Technicon             Near Infrared       Technicon Industrial Systems
InfraAlyzer                                511 Benedict Ave.
                                          Tarrytown, NY 10591
                      Principle of
Name of Device        Operation           Manufacturer or Distributor
                                          Dickey-john, Inc.
                                          P.O. Box 10
                                          Auburn, IL 62615
Universal Moisture    Conductance         Burrows Equipment Company
Tester                                     1316 Sherman Avenue
                                          Evanston, IL 60201
902 Moisture          Phosphorous         E.I. DuPont Co.
Evolution             Pentozide           Instrument Products Div.
Analyzer                                   Quillen Bldg.
                                          Concord Plaza
                                          Wilmington, DE 19898
No. 1210 Froment      Mechanical          N.J. Froment
Moisture Tester       Plunger -- 9-volt   P.O. Box 758
                      Battery             Trenton, Ontario
(a) This list was compiled from manufacturers responding to a USDA inquiry. Inclusion does not imply U.S.
Government endorsement; omission does not imply disapproval.
                              APPENDIX D
                          FARM-LEVEL STORAGES((12) This appendix is
abstracted from a paper given in Coimbatore, Tamil Nadu,
India, 1976, to a national meeting of engineers working on postharvest technology.
The final report referred to is the IDS/IGSI Crop Storage Project report
submitted to the Government of India in 1978.)
                                  M. Greeley
  There have been relatively few attempts to assess the private profitability of
alternative farm-level storage improvements. Yet without this evaluation there
is no basis for choosing between alternative technologies.
  The exercise below illustrates an approach to evaluating three important
methods of storage improvement for Andhra Pradesh, India. In each case, we
ascertain a benefit-cost ratio for each rupee invested by determining how many
rupees are gained through grain saved by improving storage methods.
  It must be emphasized that we are concerned here mainly with explaining the
approach and that, for example, the levels of losses due to different causes
given here are rough and are presented only as examples.
  The three storage improvements, all designed by a local grain storage institute,
  1. The domestic metal bin, manufactured by Andhra Pradesh State Agro-Industries
  2. the improved platform for the outdoor gade (bamboo basket); and
  3. the improved base for the puri (large circular paddy-straw rope structure).
  Improvements to the gade and puri are both designed to prevent access to
rodents and groundwater migration. The puri is not fumigable but the gade
can be fumigated successfully once a mud and dung coat is applied. The
project has built over 30 gade improvements and 10 puri improvements. To
make comparisons easy, all calculations are based on storage of one 75-kg bag
of paddy. We are using loss-levels by cause((13) Comparison between the gade
and the metal bin is unaffected by the relative
importance of different causes because all three types of losses can be prevented
in both. This is not true for the puri where fumigation is not possible. It may
also be true that the importance of different causes of loss varies significantly
between unimproved gades and unimproved puris as well as there being variation
in the total percentage of losses, but the purpose here is to describe the method.
The actual results are secondary, though it could be said that the improved
gade-metal bin comparison is more realistic than comparing either one of these
with the puri.) in the traditional stores of:
rodents, 2%; insects, 2%; and molds, 1%, assuming that the maximum saving
possible through storage improvement is 5%.
  In addition, other values required are:
  1. Initial construction costs both of the structure and the improved base/
  2. Annually recurring costs.
  3. The price of paddy.
  4. The effective life of the structures.
  5. A discount factor.((14) A discount factor is a simple concept. It gives
the relation between
future cash flows and their present value. Asked to choose between a gift of
Rs 100 now and Rs 100 in ten years' time, we would all choose Rs 100 now. To
be willing to give up Rs 100 now, how much money would I require to be given
in ten years' time? This depends on how much extra money I could earn in ten
years with the Rs 100 invested, which in turn depends on the rate of return.
This depends on the rate of interest. The discount factor works like a
compound rate of interest. The value now of a Rs 100 in ten years' time is the
amount of money I would have to invest now in order to have Rs 100 in ten
years' time at a compound rate of interest. If I invest Rs 32 at a 12% rate of
compound interest, its value in ten years' time is just under Rs 100; so the
discounted present value of Rs 100 in ten years' time in this case is Rs 32. In
valuing future costs or benefits to obtain their present value, we divide by a
discount factor (the inverse of multiplying by a rate of interest). After one
year an investment is worth P (1 + i), that is, the principal sum (P) plus the
principal times the rate of interest. This sum which we call [P.sub.1] divided by
(1 + i) equals P. Looking at the change after one year helps to understand the
role of the discount factor. Rs 100 now at a 12% rate of interest equals $100 +
Rs 12 after one year (100 + 100 x 0. 12 = (P + P x i) = 112). We write this
formula as P (1 + i). To find the original (present) value of that Rs 112 which
we can call [P.sub.1], we simply reverse the process. Instead of multiplying
by (1 + i) we divide by (1 + i). The present value is
       112                    [P.sub.1]
     -------- = 100, ie, P = ---------
     1 + 0.12                  (1 + i)
Similarly, to reach the present value, after two years, we divide
by [(1+i).sup.2] and after three years by [(1+i).sup.3]. The
value now of Rs 100 in 10 years' time is    100
                                         ------- = Rs 32
                                         [(1 + i).sup.10]
(where i = the proportionate rate of discount. In this case, the rate of
discount is 12%, i = 12/ 100 = 0.12). We have assumed a discount rate of
12% simply because it is one used in some national planning exercises and it
may reflect not too misleadingly the rate of return in alternative forms of
  These values are given below:
  Metal Bin. Currently priced at Rs 341 excluding transport and with a capacity
of 10.5 bags, the cost per bag is Rs 32.5. Excepting fumigation there are no
annually recurring costs and no platform costs. All three causes of loss are
  The gade is a basket-type structure usually made from bamboo. Its cost
depends on its capacity. Payments from the farmer to the basketmaker is in
kind (not cash) at the rate of 2 kg of paddy for every 40 kg of capacity. To
calculate the money value of a kind payment, we assume a price of Re 1 per kg
of paddy. The cost of a 75-kg capacity structure, that is, one-bag capacity, is
then equal to Rs 3.75. The cost of the improved platform is Rs 5.1 per bag.
Total initial cost is therefore Rs 8.85.
  The cost of the new mud coat each year is given as Rs 0.5 (based on an actual
amount of Rs 8 for a 16-bag structure which is about average). The other
annual recurring cost is fumigation. Total annual recurring cost is therefore Rs
  Improved Puri. The cost each year of the structure construction is approximately
Rs 0.80 per bag after allowing for reuse of the straw. The structure is
rebuilt each year. The cost of the improved base is Rs 4.2 per bag. Insect losses
are not preventable because fumigation is not possible.
  The life span of all permanent structure/platforms is conservatively estimated
as 15 years.
  The cost of fumigation (1 EDB ampule) is assumed to be Rs 0.75; one
fumigation only is given at the time of initial storage.
  Loading/unloading and cleaning costs are excluded since the puri is completely
rebuilt each year and is loaded in the actual process of construction, but
the estimated labor costs of loading (inseparable from construction) are
roughly the same as for the other structures.
  The price of paddy is assumed to be 1 Re per kg. The discount rate is
assumed to be 120%. It is assumed also that no credit has been taken to
purchase any of the structures so no loan or interest payments are due.
  The costs are as above and the benefits over the 15 years' life of the structures
are measured by the grain saved:
             Rodents 2% = Rs 1.5 undiscounted
             Insects 2% = Rs 1.5 undiscounted
             Molds 1% = Rs 0.75 undiscounted
  From the totals at the bottom of Table X the discounted benefits/cost are

pglxtabx.gif (600x600)

the money benefits/costs divided by the discount factor over a 15-year period.
  The discounted benefit-cost ratios are as follows:
             Metal bin 25.58:38.23     = 0.67:1
             Improved gade 25.58:18.40 = 1.39:1
             Improved puri 15.35:10.49 = 1.46:1
  The importance of discounting is shown in the case of the metal bin. Without
discounting the benefit-cost ratio is 1.29:1 (51.25:43.75), which implies
that for every rupee invested, a return of Rs 1 29 paise can be expected,
whereas after discounting we obtain a return of only 67 paise, for a loss of 33
paise. We must emphasize again that the loss-levels given are assumed only for
convenience in illustrating the approach.
  The same approach can be easily adapted to include additional factors such
as risks of fire, flood, and theft or the use of different prices for (a) different
uses of stored grain, or (b) different removal patterns. An important additional
factor very relevant in some states now for the metal bin is the cost of
credit. Further refinements can be introduced by examining how sensitive the
results are to changes in the parameters (eg, different price levels). Indeed, this
is an important exercise if the values used are at all uncertain. Some subjective
factors such as the preference for a modern metal bin or contrarily the reluctance
to switch from a traditionally used structure are more difficult to incorporate.
  In this exercise we have ignored the question of actual storage requirements
based on production and disposal patterns. If a farmer wishes to store 100
bags of paddy, then the theoretical choice would be between 1 puri, 4 gades
(average size of our improved gades is 25 bags though individual gades up to
160 bags exist) or 10 metal bins. Space constraints and possible scale economics
(which have been ignored by using average costs) then become relevant;
both factors work in favor of larger unit capacity structures. However, it is
also likely that, all things being equal, the percentage of losses is inversely
related to size. In other words, the potential gross benefits from improvements
to small structures are greater. The list of additional factors is by no means
exhaustive; particular regions, particular crops, particular use patterns, etc.,
will require giving different emphasis to one or another factor but these can be
incorporated as needed and still allow meaningful comparisons through the
benefit-cost ratio.
  Finally, we should note that a parallel approach can be used to estimate
"social" benefit cost ratios from an extension program for storage improvement
though this involves including (a) additional costs of the extension program
and the associated administrative overheads, and (b) a set of prices that
reflects real social values rather than using direct market prices.
                          SELECTED REFERENCES
ADAMS, J. M. A bibliography on post-harvest losses in cereals and pulses with particular reference
    to tropical and subtropical countries. Trop. Prod. Inst. G 110 (1977).
ADAMS, J. M., and HARMAN, G. W. The evaluation of losses in maize stored on a selection of
    small farms in Zambia with particular reference to the development of methodology. Trop.
    Prod. Inst. G 109 (1977).
ASIAN PRODUCTIVITY ORGANIZATION. Training manual: Post-harvest prevention of
    waste and loss of food grains. APO Project TRC/1X/73, Asian Productivity Organization,
    UNIPUB (1974).
van BRONSWIJK, J. E. M. H., and SINHA, R. N. Interrelations among physical, biological,
    and chemical variates in stored-grain ecosystems; a descriptive and multivariate study.
    Ann. Entomol. Soc. Am. 64(4): 789 (1971).
BROWN, R. Z. Biological factors in rodent control. U.S. Public Health Service, Communicable
    Disease Center Training Guide - Rodent Control Series (1960).
CHRISTENSEN, C. M. Storage of cereal grains and their products. Am. Assoc. Cereal Chem.: St.
    Paul, MN (1974).
CHRISTENSEN, C. M., and KAUFMANN, H. H. Grain storage: The role of fungi in quality
    loss. Univ. Minn. Press: Minneapolis, MN (1969).
COTTON, R. T. Insect pests of stored grain and grain products. Identification, habits and
    methods of control (out of print). Burgess Pub. Co., Minneapolis, MN   1963).
    marketing short course outlines (in English-Spanish-French). A. Fundamentals. B. Grain
    inspection and grading. C. Handling, conditioning and storage. D. Sanitation. E. Marketing,
    operations and management. Mimeo. Dep. Grain Sci. and Ind., KSU, Manhattan
GREIFFENSTEIN, A. C., and PFOST, H. B. Moisture absorption of bulk stored grain under
    tropical conditions. Res. Rep. No. 6. Food and Feed Grain Inst., Kansas State Univ.,
    Manhattan (1974).
HALL, D. W. Handling and storage of food grains in tropical and subtropical areas. FAO Agric.
    Devel. Paper No. 90 (1970).
IDRC. Rice postharvest technology, ed. by E. V. Araullo, D. B. De Padua, and M. Graham.
    International Development Research Centre, Ottawa, Canada (1976).
LINDBLAD, C., and DRUBEN, L. Small farm grain storage. Action/Peace Corps, Program and
    Training Journal, Manual Series No. 2, Washington, DC, or Volunteers in Technical
    Assistance, Vita Publications, Manual Series No. 35E (1976).
MONRO, H. A. U. Manual of fumigation for insect control (2nd ed.). (In English-French-Spanish.)
    FAO Agric. Devel. Paper No. 79 (1969).
MUNRO, J. W. Pests of stored products. Hutchinson & Co., Ltd.: London (1966).
PEDERSEN, J. R., MILLS, R. B., PARTIDA, G. J., and WILBUR, D. A. 1974. Manual of
    grain and cereal product insects and their control. Dep. Grain Sci. and Ind., Kansas State
    Univ., Manhattan (1974).
PHILLIPS, R., and UNGER, S. G. Building viable food chains in the developing countries.
    Special Report No. 1. Food and Feed Grain Inst., Kansas State Univ., Manhattan (1973).
PINGALE, S. V., KRISHNAMURTHY, K., and RAMASIVAN, T. Rats. Food Grain Technologists'
    Research Association of India, Hapur (U.P.), India. Kapoor Art Press, Karol Bagh,
    New Delhi, India (1967).
RAMIREZ, G. M. Almacenamiento y conservacion de granos y semillas, 2a impresion. Compania
    Editorial Continental, S. A. Mexico, Espana, Argentina, Chile, Venezuela (1974).
SINHA, R. N. Uses of multivariate methods in the study of stored-grain ecosystems. Environ.
    Entomol. 6(2): 185 (1977).
SINHA, R. N., and MUIR, W. E. Grain storage: Part of a system. Avi Pub. Co.: Westport, CN
WORLD FOOD PROGRAM. Food storage manual. Part I, Storage theory. Part II, Food and
    commodities. Part III, Storage practice. Commodity and Technical Index. Prepared by
    Tropical Stored Products Centre, Ministry of Overseas Development, Slough, England
WYE, A. J. Selected bibliography on improving farm storage. Trop. Stored Prod. Inf. 21: 13
Accuracy, 45, 77
Baseline procedures, 83, 119
Bias, 1, 7, 39, 45
Boerner divider, 149
Bourne, 11
Confidence limits, 45
Consumption, relation to loss, 135
Cost/benefit analysis, 1, 145
Cowpeas, 83
Culture, 29, 39
Damage assessment, 101, 109
Data-record sheet, field and laboratory, 123
Domestic organization, 39
Drying losses, 59, 67
Dust, effect on volume, 83
Emergence holes, 77, 83
Estimations, 1, 77
Expert judgments, 77
Food defined, 11
Frass, 77
Fungal damage defined, 77
Grain pipeline, 15
Grain storage, 109, 187
Grinding losses, 67
Guesstimates, 1, 77
Household losses, 77, 101
Interventions/intervention points, 1
Leaks, 19
Loss-causing factors, 25
Loss defined, 11, 77, 139
Loss measurements, 67
Loss points, 19
Loss reduction, 7
Losses (cleaning and winnowing, drying, par-
 boiling, hulling, polishing), 59, 67
Losses due to microorganism growth, 95
Maize losses, 59, 67, 83
Map sampling coordinates, 163
Market price/market value, 145
Meters, moisture, 119, 129, 167
Moisture measurements, 119
Moisture meters, 119, 129, 167
Moisture-proof container, 123
Moisture reference tests, 119
Mold toxins, 7
Mycotoxins, 7
Paddy losses, 59, 67, 83
Pattern of loss, 135
Pilferage defined, 77
Pipeline concept, 7, 19
Postharvest defined, 11, 77
Postharvest price structure, 145
Post production, 11
Preharvest, 11
Probes, 149
Probing techniques, 149
Processing losses defined, 77
Pulses, 83
Questionnaire, 63
Random numbers, 163
Random samples, 45, 163
Rapid loss appraisal, 25
Rate of microbial growth, 95
Representative samples, 45
Resource allocation, 19
Rice, 83, 187
Rodents, 101, 109
Row-centimeter measurements, 117
Sample dividers, 149
Sample reduction, 149
Sample selection, 163
Sample sieving, 149
Sampling, bag, 149
Sampling, stack, 149
Sampling methods, 49, 149
Sampling stratification, 49
Sampling survey design, 49
Social factors, 39
Sorghum, 83
Storage, 187
Threshing losses, 59, 67
Time, effect on loss, 135
Tradition, 29
Triers, 149
Unit weight changes due to mold, 95
United Nations General Assembly, 1
Visual loss estimates, 117
Volume/weight determination, 83
Weight loss determination, 83, 95, 119
Wheat losses, 59, 83