Standardized methods for the assessment of losses due to insect pests in storage

Contents - Previous - Next



Of the agricultural commodities consumed as food, cereals contribute to the bulk of the world's calories and protein. Unfortunately, a considerable quantity of the world total cereal production is lost after harvest. While losses of lesser magnitude can be expected in developed countries, post harvest losses of grains are frequently high in developing countries where the grain is usually inadequate and the need for conservation is urgent.

Losses of stored grains as high as 30% weight loss and an average of 8.7% during 3 to 6 months storage period has been reported in Tanzania due to the outbreak of Prostephanus/truncatus (Horn) (Golop and Hodges, 1982). Whether the potential for postharvest losses from one country to another is alarming or negligible, a means to come up with a reliable loss estimate is necessary. Although methodology for assessing post harvest grain losses will not in itself reduce losses, sound assesment procedures are essential to post harvest operational programs so that priorities for loss reduction can be determined.

This paper compiles available methodologies that can be applicable in the assessment of losses caused by insects.

Methods for determination of losses due to insects are of four types:

  1. Determination of the weight of a measured volume of grain. In this case the loss in weight in samples taken over a known time period will be the reflection of losses.
  2. Separation of damaged and sound kernels and determination of their comparative weights calculated in terms of the whole sample.
  3. Determination of the percentage insectdamaged grain and is conversion to weight loss using a conversion factor.
  4. Comparison of the Thousand Grain Mass (TOM) between sampling occasions.


A. Sieving

For all methods, it is necessary that the grain sample be sieved to remove dust, insects and other foreign materials that may be associated with the grain. An appropriate size of sieve should be used.

B. 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. This will be the reference sample, from which it is possible to compare all future measurements. This baseline needs to be in the form of 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 contents.

The curve is obtained from analysis and calculation of a baseline sample. In case 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 subsarnples and the measurement required by the appropriate methods applied to each subsamble. Each subsample should then be placed in a jar covered with cheesecloth to prevent the insects entering of leaving and kept for four weeks. After four weeks, the grains must 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 sample(5) with 5% or less damaged kernels should be taken. If the damage is above 5%, assistance from an expert in determining the appropriate correction factor will be needed.

C. 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 moistrue due to intravarietal variations of the local grain(s) then either each lot of stored grain must be treated individually.

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, e.g., if it is 10-18%, this will be 10,12,14,16,18. If it is 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 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 withe a current of air passing over it but protected from insect attack or, preferably, in ventilated oven in shallow trays at a temperature not execeeding 35C. 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 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

For example, if we have a subsample of 1000 9 of grain at 12% moisture content and require it to be at 16% moisture content the calculation is:

Weight of water

= 1000 (16 - 12) / (100 - 16) = 1000 4 / 84 = 47.6 g

This can be weighed out or since 1 g.of water occupies in ml, it can be measued out as a volume. Water is added to the grain with sufficient headspace for thorough mixing. It is left for two weeks to condition, but vigorously shaken daily. For moisture content over 16%, the container should be kept at 5-10 C in a refrigerator to discourage mold growth. At the end of the conditioning period, an accurate moisture content is determined for each subsample. For each subsample, the weight that occupies the volume measure should be determined by filling a test weight container (weight/liter tester or a chrondrometer). The grain should be weighed to the nearest 0.1 9. This should be done three times for each subsample to obtain a mean result.

There will now be five mean weights for each variety at five accurately measured moisture contents. Each of these weight should then be converted to dry weight as follows:

For example, if the volume of grain in the test weight container weighed 800 9 and had a moisture content of 15% then its dry weight is:

Dry weight = 800 x (100 - 15) / 100 = 800 x 85/100 = 680 g

This is done for all subsamples so as to obtain a set of dry weight for each moisture content. A graph is now drawn of the dry weight against the moisture content for example:

% MC 10 12.0 14.1 16 17.9
Dry Weight 700 670 650 620 600

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 can be 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. The Standard Volume/Weight Method

This can be carried out after the preliminary laboratory work for the baseline figures is completed.


  1. Test Weight apparatus
  2. Balance such as a triple beam balance capable of measuring 1.0-1.5 kg accurately to 0.1 9
  3. Moisture meter, calibrated for the type of grain being tested.
  4. A suitable size grain sieve to remove dust and other impurities.
  5. Plastic bags


A well mixed sample, taken from the store is first sieved by locally appropriate methods 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. Then the moisture content of samples are measured.

The weight occupying the volume of a standard container is measured in three replicates and a mean taken. This weight is converted to dry weight using the moisture content and the formula for dry weight.

The graph is used to find the dry weight of the sample at the same moisture content at the time of storage. For example, if the moisture content of the farmer's sample was 12.0% then referring to the example (appendix 1) the dry weight would be 670 g.

The weight loss in the farmer's sample is then calculated as follows:

For example, if our farmer's sample at a moisture content of 12.0% had a dry weight of 600 9 then as the dry weight on the graph for 12.0 moisture is 670 9 the loss would be:

% dry weight loss = [670 - 600] / 670 x 100 = 10.44%

This is the dry weight loss, which by definition excludes moisture content changes.

Sources of Error

  1. Inaccuracies may result from grain samples containing very high levels of damage, where some of the grains may be crusted.
  2. 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.
  3. For rough rice, the effect of moisture content on the dry weight occupying a given volume is negligible. Within the range of 5% moisture, there is no requirement for a predictive graph.

B. Modified Standard Volume/Weight Method When a Baseline Cannot be Determined

There are some situations where the Standard Volume/Weight Method cannot be used without modification. It may also be difficult to obtain reliable moisture content determination 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 rural areas, different varieties of grain are grown under different conditions. 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 the undamaged and the a percentage) between the undamaged and the damaged sample. Conversion for moisture need not be used this in case since the moisture content will be approximately the same.

Sources of Error

1. Unreliable results may be produced if during selection there is hidden internal infestation, preferential feeding and egg deposition by insects in grains of different sizes, and a difference in moisture content.

2. 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.

3. Another way to reduce this error is to take the undamaged sample at random as much 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.

C. The Count and Weigh Method

There are many situations in which a loss estimate is required but where there is only minimal equipment available, the baseline could not be determined before the storage period. In addition, it is sometimes imossible 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.


  1. Balance with a range of 0.5 g to 1.5 kg accurate to 0.1 g
  2. Tallycounter
  3. Plastic bags


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:


U = weight of undamaged grains
Nu = number of undamaged grains
D = weight of damaged grains
Nd = number of damaged grains

Sample Size

A sample size of 100-1000 grains is recommended. 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.

Sources of Error

  1. 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.
  2. Since insects will sometimes select and infest larger kernels, any procedure that compares the individual weight 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.

D. The Converted Percentage Damage Method

This method provides a useful estimate for quick appraisal of losses without needing equipment.

When grains are heavily infested, feeding by secondary pests and multiple infestation may disturb the relations and so lead to an underestimation of losses.

Although the converted percentage damaged method is liable to the same sources of error as modified standard volume/weight method and the count and weight 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.


  1. Tally counter
  2. Plastic bags


One thousand grains are counted at random from the working sample. Then the damaged grain(s) are separated from the sound grains and expressed in percentage using the formula:

This percentage is converted into percent weight loss by dividing it by the conversion factor (c) or multiplying it by 1/c.

The following conversion factors have been estalished in practice where the earlier larval stages develop within the grain eg. Sitophilus species, sitotroga cerealella.

Maize (stored as shelled maize or as ears without husk) % bored grains x 1/8
Maize (stored as ears with husk) % 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


D. The Thousand Grain Mass (TOM)

Basic Principles

When an entire lot of grain is weighed before and after being attacked by insect pests, microorganisms or some other causing agent, the percentage loss of mass is easily calculated by using the formula:


m1 = grain mass before attack
mx = grains mass after attack

"mass" in this context refers to the dry matter weight

A sample taken from the lot in strict accordance with representative sampling principles should possess all the characteristics of the grain in proportion to their ccurrence in the lot at the time of sampling. Therefore, if the lot consists of 40% large grains, 50% medium size grains and 20% small grains, these proportions should be found in representative samples. Likewise if 7% of the grains in the lot are damaged, this percentage of damaged grains should also be found in the representative sample.

It is important that the mass per standard unit of a representative sample should be the same as the mass per standard unit of the entire lot of grain at the time of sampling. A reduction in the value of this unit between two sampling occasins should be proportional to a dry weight loss in the grain lot and should therefore provide a means of estimating the loss.

Mass per standard unit is defined as the dry weight (mass) of 1000 grains calculated from the wet weight and number of grains in a representative working sample by the formula:


m = is the wet weight of the working sample
H = is the percentage moisture content (WB)
N = is the number of grains in the working sample
M = is the Thousand Grain Mass

The formula is obtained by multiplying [m x 1000] / 100 which gives the Thousand Grain Mass on a wet basis, and by [100 - H] / 100 to convert the TGM to the dry basis.

D.1 The Simple TGM Method

When samples for analysis are obtained in a truly repersentative manner both before and after the grain lot has been attacked by loss-causing agents, the loss can be estimated with reasonable accuracy by the simple TGM method.


On each occasion the submitted sample is first weighed and then screened to remove as much foreign matter as possible. Large pieces of foreign matter may: be removed by hand. The partially cleaned sample is then reweighed if data on grain purity is required. Then it is reduced to a working sample of a desired size.

Next, the percentage moisture content of the grain is determined, preferably using a calibrated moisture tester. The working sample is thoroughly cleaned of all foreign matter and weighed with a balance having an accuracy of at least 0.1 gm.

If the submitted sample was obtained some weeks or months after the beginning of the study, the weight loss occuring in that occasion can be calculated by using the formula:


M1 = TGM of the grain at the begining of the study
Mx = TGM of the grain on occasion "x"
N 1 = Number of grains in the working sample

D.2 The Multiple TGM

There are instances in some loss assessment studies that the collection of the majority of samples does not fulfill all the requirements of representative sampling. Thus, while it may be possible to obtain a good representative sample as a grain store is filled, subsequent samples usually have to be drawn while the bulk of grains remains in store.

Even the initial sample may have to be drawn after the store has been filled. Such samples, however carefully taken, cannot be considered to represent the entire contents of the store and cannot, therefore, be used for estimating weight losses relating to all grain by store. This problem can be overcome substantially by multiple TGM analysis.

This method should be used when the grain under study is variable in size and it is known that one or more of the samples cannot be fully representative of the whole contents of the store.


After weighing, the submitted sample is screened to remove as much foreign matter as possible. The partially cleaned sample is then passed through a series of sieves with different aperture widths to divide it into two or more portions according to grain size. (At the beginning of the study a comprehensive set of sieves which will divide the sample into portion of sufficient quantity for TGM determination are selected for subsequent use).

Each portion is weighed and then divided into two working samples: (1) for moisture content determination and (2) for TGM determination. The working samples for moisture content determination are combined for this purpose. The TGM working samples are analyzed separately.

When all analyses are completed the "potential weight" of the grain in each portion of the submitted sample is calculated using the formula:


M1 = is the TGM of the corresponding grain portion at the beginning of the study
Mx = is the TGM of the portion on occasion "x"
Wx = is the actual wet weight of the portion on occasion "x"
Wp = is the "potential weight" of the portion on occasion "x"

The percentage weight loss of the grain in the submitted sample obtained in occasion "x" can be calculated by using the formula:


P1, P2... and x1, x2...represent the various portion of samples

Sources of Error

The naturally occuring variability in grain weights and the variations in the representative nature of samples are the two main factors that may affect the reliability of the TGM method.


Adams, J.M. and G.G.M. Schulten, 1978. Losses caused by insects mites and microorganisms. Paper 83-89 in Post Harvest grain Loss Assessment Methods. K.L. Harris and C.J. Lindblid, eds. Am. Assoc. Cereal Chem: St. Paul MN

Golob, P and R. Hedges, 1982. Study of an outbreak of Prostephanus truncates (Horn) in Tanzania. Tropical Product Institute, London, England 16) 23 pp.

Proctor, D.L. and J.Q. Rowley 1983. The Thousand Grain Mass (TGM): a basis for better assessment of weight losses in stored grain. Page 19-23 in Tropical Stored Products Information. J.A. Mc Earlane ed. Storage Department of the Tropical Development and Research Institute Slough, Berks, England.


Much of the information included here are extracted from the "Post Harvest Grain Loss Assessment Methods" developed and compelled by Kenton L. Harris and Carl J. Lindblad. This manual was published by the American Association of Cereal Chemists in cooperation with The League for International Food Education, The Tropical Product Intitute (England) Food and Agriculture Organization of the United Nations and The Group for Assistance on Systems Relating to Grain After. Harvest.



Fig. 1 dry Weight of grain in test weight containing at different moisture content.

Table of conversion factors to obtain grain weights at 14% moisture content


Practical loss assessment method for storage losses


The FAO Prevention of food Losses (PFL) Project in Bangladesh, funded by the Netherlands (GCPP/ BGD/017/NET)

For each selected storage the following information was collected:

  1. Details of the storage structure, construction, maintenance, etc.
  2. Baseline information about the commodity at the time of storage.
  3. Records of the quantity of grain stored, quantities removed and grain use.
  4. Estimation of quantitative and qualitative loss by analysis of samples.

A series of forms will be used to facilitate collection of these data.


1. At the first visit to a store:

(i) Record a full description of the selected store using Form Sl - DESCRIPTION OF STORAGE STRUCTURE. Complete each section of the form as fully as possible and make a simple sketch of the store, giving dimensions, on the back of the form.

(ii) Weigh the grain into store, or if this is not possible estimate the quantity stored. Enter full details concerning the condition of the grain at storage on Form Sll, the INITIAL SAMPLE RECORD.

(iii) As the grain is loaded into the store collect a sample to provide the necessary baseline information. The sample may be collected in one of the following ways:

  1. By withdrawing small samples from each lot of grain as it is placed in store. The samples are bulked, mixed and divided by coning and quartering to produce a sample for analysis of about 1 - 1.5 kg.
  2. By collecting a 1 -1.5 kg sample from grain already in store, using a grain probe.
  3. By emptying the store, and refilling drawing samples as in (a). This method may be appropriate when a relatively small quantity of grain is stored and when the store is visited for the first time within a few days of filling.

Record the moisture content of the sample (three determinations).

Enter details concerning the initial sample on Form SIT-INITIAL SAMPLE RECORD.

(iv) Pack the sample in a sample bag; enclose a sample label with the sample and secure a second label to the outside of the bag.

(v) Send the grain sample together with the Forms Sl and Sll to the laboratory for analysis.

2. At Subsequent Visits

(i) Each store will be visited at approximately monthly intervals.

(ii) During the visit complete a SAMPLE RECORD FORM (Form SIIl).

(iii) Collect a sample of grain of approrimately 1 1.5 kg.

(iv) Record the quantity of grain removed since the previous visit.

(v) Place the sample of grain in a sample bag, complete two sample labels and enclose one in the bag and secure a second to the outside of the bag. Return sample and record form to the laboratory for analysis.

(vi) As the storage season progresses insect infestation may become serious. When insect infestation is detested, all samples should be fumigated (using Phostoxin in a sealed container) before being dispatched to the laboratory.


1. Procedure for all samples

(i) Register all samples received at the laboratory in the laboratory log book and give each sample a serial number.

(ii) Record the results of the analysis on Form SIV - STORAGE LOSS, LABORATORY ANALYSIS.

(iii) Weigh the sample, clean it by sieving and hand picking. Record the weights of sample and foreign matter.

(iv) Identify and record insects present in the sample. Those insects which cannot immediately be identified should be placed in a suitable container, labelled and sent to the entomologist for identification.

(v) Divide the sample to provide the following subsamples as necessary:

  1. Thousand grain mass
  2. Weight loss from single sample
  3. Quality analysis
  4. Milling yield
  5. Mycotoxin analysis/Microflora analysis

2. Details of specific analyses

(a) Thousand gram mass

(i) Reduce sample to obtain at least three replicates of approximately 50 g.

(ii) Using remainder of sample determine moisture content.

(iii) Weigh each subsample of approximately 50 g. accurately.

(iv) Count the number of grains in each subsample. Record total.

(v) Calculate the thousand grain mass (TGM) as follows:

(vi) Correct for moisture content to 14% as follows:

(vii) Calculate mean TGM

(viii) Determine % weight loss as follows:

(b) Weight loss from a single sample

Under certain circumstances it may not be possible to obtain a baseline sample and so the following method may be used to determine a weight loss. The method may be used as a cross check on the thousand grain mass method of calculating % weight loss.

(i) Reduce the sample to obtain three replicates of 25 - 50 g.

(ii) Determine moisture content of the sample.

(iii) Separate grains into large and small categories using a suitable sieve for the separation. (N.B. If necessary further categories of grain size may be used).

(iv) For each size category of grains, separate into undamaged and damaged fractions.

(v) Count and weigh the number of grains in each fraction.

(vi) Calculate the weight loss by comparing the weight of the sample with the predicted weight of the same sample in the absence of damage:


(i) A SUMMARY SHEET (Form SV) will be used to assemble all relevant data for the calculation of total weight loss. This summary sheet can be completed as and when the data are available. All weights will be expressed on a 14% moisture content basis.

(ii) The total accounted loss will be obtained from the difference between the quantity stored and the total quantity of grain removed (i.e. the final entry in the balance column).

(iii) The weight loss due to insects in any one month calculated from the "loss in sample" result and the quantity (adjusted to 14% M.C.) of grain removed on or about the same day.



QUANTlTY REMOVED IN MONTH = 50 kg (at 14%)


Weight Loss:

= 50 / 95 x 100 - 50

= 52.63 - 50

= 2.63 kg.

Weight loss as % Quantity Stored:

= 2.63 / 270 100

= 0.97 %

The 'monthly' loss figures are summed to obtain the cumulative weight loss due to insects.

(iv) By subtracting the cumulative total loss due to insects from the total recorded loss (i.e. the difference of quantity stored/quantity removed)

the loss due to other causes can be obtained. By reference to field observations this "other" loss may be attributed to rodents, birds, etc.

(v) The quantity of grain discarded as unfit for consumption represents a loss of food and must be expressed as a percentage of the quantity of grain stored. It should be remembered that this discarded grain may also have been subjected to a weight loss (due to insects) at least equal to the loss recorded in the quantity of grain removed for consumption.

(vi) The total loss therefore consists of:

Loss due to insects + Loss due to other causes + Loss of discarded grain ( + some additional weight loss in discarded grain.) = Unaccounted Balance at end of Season



FORM S lll



Form SV

Contents - Previous - Next