Chapter 3 - Survey and analysis of post-harvest losses

3.1. Methodological problems
3.2. Loss Assessment and field surveys

3.1. Methodological problems

Study of post-harvest losses covers many aspects, in view of the wide range of (a) products involved, (b) successive operations in the post-harvest system, (c) causes of losses, and (d) pests and other food parasites, not to mention (e) physical, technical, economic and other conditions that aid and abet the action of the agents of deterioration, consequently increasing losses. This has led to the enormous variety in approach and form of analysis found in reports, manuals and other works, depending on the authors' choices and aims.

For example, M. Bourne (1977) focuses on the causes of loss, distinguishing primary and secondary causes, and the places and times of loss; thus he examines biological and microbiological causes first, and is primarily interested in insects, rodents and moulds. Similarly, a recent NRI training manual (prepared for FAO in 1994, unpublished), which does not distinguish between primary and secondary causes, starts by considering pests and micro-organisms, and then moves on to losses during successive post-harvest operations.

G. Schulten takes a more selective approach in a FAO mission report (1982), further differentiating between different categories of produce. Beginning with rice, which he distinguishes from other cereals, he argues that losses for this crop are a result primarily of handling and equipment, particularly at the time of harvesting. He therefore takes a close look at all the manual and mechanical operations from on-field drying through to paddy milling. Only then does he move on to a comprehensive consideration of cereals and pulses in order to analyse loss during storage, distinguishing small-farmer or village storage from warehouse storage, and confining himself to loss caused by insects and moulds. Rodent damage is studied separately.

In the NRI manual, losses are examined in sequence, following the succession of operations in a single system, with special attention to the conditions of cutting and field drying of rice ears, and then to the risks of damage and loss during milling.

In his methodological work, R. Boxall (1986) prefers to distinguish between storage loss and losses in other locations and of other types, leading him to divide his study into two main sections. The first section is devoted to losses connected on the one hand with pre-storage (from harvesting to drying), highlighting problems connected with maize husking, and on the other hand with industrial and domestic processing (milling, polishing etc., and cooking). The second section is devoted to all the losses that can occur during storage, whether caused by insects, micro-organisms or vertebrates.

Most recent reports and handbooks do now pay due attention to the question of equipment in general, and more specifically to small-scale mechanization, particularly as used in threshing and hulling processes. They also take account of some of the financial, technical and economic effects these may have. However, very few pay any significant attention to socio-economic factors (competition between a harvest that is already mature and a new crop to be grown, means of access to fields, means of transport, capacities of transport, availability of labour, seasonal migration, etc.) or socio-cultural aspects (traditional division of labour, dietary habits, cooking methods, etc.).

This brief outline of the non-technical aspects of post-harvest losses reminds us that the post-harvest sector itself is part of a much broader and more complex system, in which the various types of interrelationship and interdependence concern not only the successive operations in the food chain, but also the whole range of human activities, and hence the very workings of society, thereby contributing to existence and the advance of civilization.

3.2. Loss Assessment and field surveys

3.2.1. Rice
3.2.2. Maize
3.2.3. Millet and Sorghum
3.2.4. Pulses and oilseeds
3.2.5. Roots and tubers

3.2.1. Rice

Many of the available works are devoted to rice-growing and its post-harvest system. D. Calverley's recent regional assessment (1994), covering several countries in central and southeastern Asia, is a good example, and gives an idea of the complexity of a summary and/or comparative evaluation, whether one or more products are being considered. Since rice holds pride of place, we shall start with this cereal, so that comparison will focus mainly on working methods, implements and geography. Further, we shall confine ourselves for now to results connected with harvest operations as such: maturity at the time of harvesting, manual or mechanical cutting (or gathering), placing in stacks and bundles, on-field and off-field drying of paddy.

RICE: Losses during harvest operations

Operation		Loss per country
Harvest: too mature,	Sri Lanka		Burma
leading to breakage	0.8%		2.1%
Sickle reaping:		Indonesia
wet season		0.7%
dry season		0.5%

Average loss as a percentage of estimated potential yield

	Thailand	Burma
Harvesting by traditional hand cutting	9.3%	1.9%
Improved (mechanized) harvesting: shoulder power reaper	5.2%	5.4%
reaper-binder	5.2%	5.2%
combine-harvester	1.1%	2. 1%

	Bangladesh	Burma
Field stacking and bundling: it should be noted that this loss in weight and value increases fast if the bundled harvest remains several days in the field, e.g.:	0.6%	0.5%

days on-field in stacks	physical loss	reduction in market value
2 days	0.3%	9.0%
4 days	2.7%	20.0%
6 days	3.4%	34.0%
8 days	9.0%	42.0%

Paddy drying	Bangladesh	Indonesia	Nepal	Pakistan
	2.2%	3.2%	1.6%	0.5%

Given the diversity of the causative factors in this example, it would be hard to draw any general conclusion valid for all the countries concerned (only a selection of these countries was included above).

Nevertheless, this is what Calverley proposes in a comprehensive summary table which not only ignores distinctions between countries, methods, implements or machines, seasons, places, and times taken, but takes all post-harvest activities and lumps them together into five main operations in order to reach a common, general average, or rather two averages, one by simple arithmetical addition and the other by "cumulative" addition. The validity of such simplistic calculations is open to doubt, for the somewhat theoretical results depend on "deductions", as the author himself calls them. Their value lies more in their contribution to our knowledge of trends or their scale and in establishing indispensable statistics. The following table is therefore useful, for it gives averages (in percentages), which can be compared with the results of similar surveys and studies.

RICE: Total post-harvest losses

(Calverley: evaluation of 11 FAO projects in Asia)

Operation	Percentage of losses
	(simple addition)	(cumulative)
Harvesting	0.89%	0.89%
Threshing	0.99%	0.98%
Drying	3.16%	3.10%
Storage	3.74%	3.55%
Milling	4.78%	4.37%
Average	13.56%	12.89%

Doubt may be voiced on the validity of over-simplified results and all-embracing statistics such as these, primarily on the basis of the lack of consistency in the choice and presentation of the technical operations or headings and sub-headings considered. Such a disparity is clearly a result of differences in the research and survey methods adopted, the quality of field teams, the means at their disposal, the priorities set for their work, and the goals pursued. However, it is also a result of differences in levels of social and economic development of the rural and agricultural sector, particularly in terms of improved techniques and the introduction of modern equipment (machines, motors), not to mention the use of new varieties or selected seed, and chemical products for fertilizing crops and treating harvests.

Still with regard to the rice post-harvest system, many studies distinguish six or seven operations or handlings, adding winnowing or cleaning and transport, two essential technical and economic links in the agrofood chain, to the five operations in the above table. Some reports also take specific account of the spread of new techniques, introducing distinctions not found elsewhere.

For example, a three-year study conducted in China divides harvesting into two categories (sickle reaping and combine-harvesting), threshing into another two (pedal and motor threshing), drying and cleaning into three subgroups (sun-drying on bamboo, on cement and with a screen), and storage into three causes of loss (moulds, insects and rats). With ten sub-headings this provides a more detailed and more specific table in which the loss in each sub-operation further explains the average loss for each major operation. The table is given below, with its sixteen results as percentages of production. The six main categories of loss are listed separately (the figures on the left) for greater clarity, and then given as a percentage of total losses (the figures on the right).

Schema of stored-grain ecosystems: a granary, a sack, and a box of processed cereal. Insects can attack the product regardless of the type of storage.

Agents of post-harvest losses on a maize cob.

Post-harvest loss-agents on maize crop.

RICE: Distribution of post-production losses over three seasons (1987/89) in China (Zhejiang) (IDRC study survey)

Operation	Average loss as a percentage of production	Average loss as a percentage of total losses
Harvest	0.85%	5.81%
- sickle	0.43%
- combine-harvester	3.38%
Threshing	1.31%	8.85%
- by pedal thresher	0.80%
- by motor thresher	1.52%
Drying and cleaning	3.47%	23.43%
- in sun on bamboo	3.35%
- in sun on cement	4.10%
- in sun with screen	2.90%
Storage	5.46%	38.86%
- moulds	1.59%
- insects	1.15%
- rats	2.72%
Transport	0.97%	6.55%
Milling	2.74%	18.50%
Total losses (6 operations)	14.81%	100.00%

This table is particularly informative since it not only brings out the main points of loss, but also allows comparison between the various methods of action. It can thus be seen that the two main points of loss are drying and storage, which alone account for almost two-thirds of post-harvest losses; if milling losses are added, this rises to four-fifths. These figures clearly show the priority points for prevention efforts. It can also be seen that rats are the main cause of storage loss, accounting for as much on their own as all milling. With regard to different methods of harvesting and threshing, the combine-harvester causes considerably more loss than sickle reaping and ordinary threshing combined, although it should be remembered that the saving in time and reduction in effort required can largely compensate for, if not justify, the loss in question.

If this table is set alongside Calverley's, the overall results are similar, but there are differences in distribution, particularly with regard to storage and milling. With Calverley the heaviest loss area is milling, while in the Chinese study it is storage. One explanation is certainly the fact that the former study covers numerous countries that are relatively developed technically and that it is an overall average, but another is also more specific causes such as a lack of means of combating pests. Similarly, the level of loss with the combine-harvester in the Chinese study can be attributed to the present lack of sufficient skill in the use of this modern piece of equipment.

The lessons to be drawn from these two tables agree with the analyses and results of the majority of comparable studies and reports on the rice post-harvest system: the operations and handlings subject to the greatest losses are storage and milling, followed by harvesting or threshing and drying. This last operation is obviously a particular problem in forest and humid Savannah zones, where grain has a high moisture content at harvest. Here is an example taken from the report of an FAO project in Africa:

Forest zone: pre-harvest crying	August	September	October
Moisture content of grain	26.0%	20.5%	16.0%
Insect damage	2.8%	2.8%	10.8%
Loss due to birds	---	6.8%	18.2%
Stalks fallen to the ground	---	---	20.8%
Loss of weight	0.9%	2.4%	3.2%

As can be seen, given the high percentage of moisture in the grain and the damp heat of the tropical forest, there can be considerable damage and loss if pre-harvest field drying is extended, especially starting in the third month, when the depredations of insects and birds, added to the percentage of fallen stalks (broken or flattened), can amount to 50% of the harvest, while the moisture content of the grain is still too high for long-term storage (the maximum levels are 14% for paddy and 13% for milled rice).

We can close these observations on rice by noting some figures on storage loss. According to a Kobe University study carried out in five countries in Asia, in four of these the main cause of loss was rodents, and to a lesser degree insects, which agrees with the Chinese conclusions. A World Bank study of commercial paddy storage in Bangladesh also confirms these figures (12-13% of loss caused by insects and rodents, out of a total loss of 19.7%). Finally, a study by the Brazilian Technical Commission for Agricultural Loss Reduction divides post-harvest losses into three major headings, giving the following figures:

harvest:	12.6%)
storage:	7.0%) total: 22.0%
milling:	2.4%)

These broad averages simply give an idea of proportions, and are thus of primarily statistical interest. However, it will be noted that storage on its own accounts for more losses than all harvest activities together (harvesting, threshing, etc.) and that the total of 22% is clearly higher than the 13 to 15% total average losses in the studies and reports previously examined. The same Brazilian commission does, however, reach a result similar to other studies, i.e. an overall average of 15% for post-harvest losses, when considering all cereals nationwide.

3.2.2. Maize

Along with wheat and rice, maize is one of the three major cereals grown throughout the world for human consumption, and the total area under this crop has increased considerably in recent decades. Yields have also risen, thanks to the spread of hybrid varieties and genetic improvements, but it must also be said that these new varieties are biologically more demanding and physiologically more vulnerable than the old land races, as is often the case with seedlings and seed produced by agricultural research.

However, the main reason for this weakness is the twofold fact that maize is a tropical wetland crop and that at the time of harvesting the grain may have a moisture content of over 30%. Immediate harvesting is out of the question unless rapid artificial drying facilities are available, clearly not the case for small farmers in Latin America and Africa, who harvest by hand. The traditional, economical method is to leave the harvest to dry while standing, and wait (for a month, and often longer) until the moisture content of the grain has fallen below 15%. However, in the rainy season, when the relative air humidity is about 90%, as it is in equatorial regions at the time of harvesting, there is no alternative to gathering the damp maize, removing it from the field, and drying it under cover.

In any case, it must be borne in mind that the longer the harvest remains standing, the greater the risks of loss: the wind will break dry stalks, so that ears that are too heavy will fall to the ground; rain will encourage the spread of moulds; vertebrates (birds, rodents, monkeys) will take their share, while insects such as the maize weevil (Sitophilus zeamaïs) or the coffee-bean weevil (Araecerus fasciculatus) will lay eggs in the grain; and lastly many cobs will be eaten or too damaged to be stored or sold, especially in the case of high-yield varieties, which have less developed husks so that the cob is only partially covered.

The following table, taken from the FAO/AGS Bulletin, no. 40, well illustrates the risks of loss during pre-harvest field drying (for three months), which rises from the second month onwards. Two zones (forest and humid Savannah) and four causes of loss (loss due to birds, damage due to insects, loss due to insects, and fallen plants) are compared:

Percentage losses recorded during pre-harvest field drying (Source: FAO/AGS Bulletin no. 40)

	End of August		End of September		End of October
	Forest zone	Humid savannah zone	Forest zone	Humid savannah zone	Forest zone	Humid savannah zone
Bird loss	26.0	26.0	20.0	20.0	16.0	15.0
Insect damage	2.8	1.4	7.8	1.9	10.8	2.1
Weight loss from insects	0.9	0.7	2.4	0.6	3.2	0.8
Fallen plants	+	+	6.8	18.2	+	+

The losses recorded in this table seem heavy, particularly in humid savannah zones at the end of the second month, partly because of fallen plants (almost one-fifth of the total), but also because of depredations by birds, which are heavy and the same for the two types of zone starting at the end of the first month (cumulatively, losses from birds alone amount to about 50% for the three months). It is thus easy to see why circumstances (the level of rainfall, attacks from birds and rodents, etc.) can make it preferable to cut short the pre-harvest drying period, even if this means providing suitable sheltered premises where the drying process can be completed. Further, it will be seen that one heading, moulds, which late rains or a prolonged rainy season never fail to produce, is missing from the above list of losses. The work of a research team in Togo on maize storage in a small-farming context fills the gap (cf. "La production alimentaire et l'agriculture en Afrique", Proceedings of the Scientific Conference held in Lomé in 1986, published by PWPA, 1988). This study analyses the effects of various pre-storage factors on traditional maize preservation, systematically comparing two varieties, one a land race and the other improved, and two harvesting periods, one late and the other early. The following table shows the conclusions of these studies with regard to moulds:

Interaction of variety and period on mould attack as a percentage

Harvest period	Variety		Total harvest periods	Average harvest periods
	Land race	Improved
Late harvest	2.26	1.23	3.49	1.74
Early harvest	2.16	2.93	5.09	2.54
Total varieties	4.42	4.16	General total:	8.58
Average varieties	2.21	2.08	General average:	2.14

These figures suggest that the development of moulds during storage depends much more on the timing of harvesting than on variety, but also that early harvesting can result in more such losses for an improved variety than a land race. This of course depends primarily on the moisture content of the grain at the time of harvesting, as this study clearly shows.

The study presented at Lomé was carried out in close collaboration with farmers, and deserves further attention, not only for its other results but also for its methodological approach. It was a component in a fuller research programme lasting several years, and covered seven months' storage (August 1985 + March 1986), observing the combined effects of three pre-storage factors (variety of maize, whether or not fertilizer was used, and timing of harvesting) on three main groups of parameters: attacks on cobs, loss of dry weight, and yields.

Over the seven months, samples were taken every two weeks (making a total of 13) from 24 well-monitored granaries. Depending on the time of harvesting, the moisture content of the maize ranged between 21.1 and 25.1%; and, again depending on the time of harvesting, once stored, it took 8 to 10 weeks to drop to 13 or 14%.

The results can be summarized as follows:

- average percentages of cobs attacked: these varied between 56.5% and 63.5%, with an overall average of 59.8%; 10% of attacks were due to insects and moulds, but insects were the prime culprits (20 times more damage than moulds); other causes of attack, particularly rodents and birds, were not recorded; (it should be noted that the granaries, set up on the Lomé Agricultural College training farm, were well monitored and protected);

- dry weight losses: these amounted to 9% of harvested and stored maize;

- yields: the post-storage yield of the improved variety was 123% that of the land race; maize grown with fertilizer gave a yield 131% that of maize grown without fertilizer; and early-harvested maize gave a yield 113% that of a late-harvested crop.

This is merely a selection of the extensive information supplied by this study, to be taken with due caution. As said, its interest lies primarily in its concern to explain the method used and the limitations of its work. For example, the following principle is stated with respect to the distinction between early and late harvesting: "A harvest is early when carried out on a field where one stand in five has ears that have tilted downwards; a harvest is late when carried out at least 14 days after the time of an early harvest."

Before comparing these results with those of other studies, it should be recalled that the approach in question had three aspects: (1) the effects of pre-storage factors on traditional maize conservation; (2) preliminary observations on the effects of temperature on the working of traditional granaries; and (3) preliminary observations on the effects of traditional granaries on grain moisture, in function of relative air humidity.

Maize is more difficult to dry than other cereals because of its high moisture content at harvest time, and the same problem arises with storage and conservation. Small farmers therefore often prefer to keep the cobs enclosed in their husks in order to protect them against weevils, even if certain varieties have thinner husks and softer grains that make them an easier prey for insects. On the other hand, in regions with a high rainfall, it is better to store cobs without their coverings, so long as they are treated against insects and ensured sufficient aeration. This is traditionally done in the humid tropics and equatorial regions such as the area bordering the Gulf of Benin, where the cobs are carefully arranged in layers on a circular platform and covered with a thick layer of leaves or stubble stacked like a conical hat. The bases of the cobs are turned outward, so that the gaps between them allow natural ventilation, and once the initial drying is completed, this helps to counter rises in temperature and moisture.

However, climates that are regularly moist and warm are bound to encourage the development of insects such as Sitophilus spp. and Prostephanus truncatus that attack harvests and stocks. The latter, known as the larger grain borer, is particularly formidable because of its proliferation and its remarkable appetite. It has gradually spread throughout sub-Saharan Africa from east to west, especially in areas with a high rainfall, and become a veritable plague over the years.*

* This insect came from Central America and appeared in Africa 70 years ago. Encountering no natural predators, it has adapted extremely well to the African continent.

The many surveys carried out in these regions, notably on behalf of FAO, have shown losses from insects of between 2 and 3% for husked cobs after six months' storage in traditional granaries. The higher losses occurring during threshing as a result of previous damage should then be added to this 2 or 3%. As growers explain, grain damaged by insects in the course of three months' storage becomes worthless, resulting in a loss of as much as 15%. This happens only in regions with no dry season. In more northerly regions with a dry season of about five months (December + April) insects do not appear and proliferate once more until May, when the humidity of the ambient air rises markedly.

The Brazilian study already mentioned, together with some other more general statistics, will enable us to broaden the field of observation and produce closer estimates of the average losses in different regions and with different practices.

For maize, overall verified losses in Brazil, according to the study in question, are 17.7%, comprising 4.4% for harvesting, 7.8% for storage, i.e. 12.2% (a little over two-thirds) for these two operations, while the remaining post-harvest operations account for 5.5% These figures can be compared with those of another study from Latin America (ACOGRANOS), in which two sets of data on post-harvest losses for maize divide the three main operations into two sub-groups, harvesting and shelling, and drying, with the following results:

1st case:	maize production:	3380 kg/ha	2nd case:	maize production:	1640 kg/ha
	% loss			% loss
Harvesting	3.6		Harvesting	5.9
Shelling	0.3		Shelling	0.3
Sub-total		3.9	Sub-total		6.2
Drying	5.9		Drying	3.0
Cumulative total		9.8	Cumulative total		9.2

While these figures do not agree with those from Brazil, the figures referring to harvesting losses are relatively close, the theoretical average of the data in this table being 4.7% (3.6 + 5.9 = 9.5, 2), which is very close to the figure of 4.4% from Brazil. The figures for storage losses are also comparable. Supposing that, apart from harvesting and storage losses, other losses in Brazil, which amount to 5.5% of the total, correspond mainly to drying losses, then the theoretical average of drying losses in the above table, i.e. 4.5% (3.0 + 5.9, 2) is comparable. Unfortunately there is no figure or estimate for storage losses, so that it is impossible to obtain a figure for the whole post-harvest chain to compare with the Brazilian figure of 17.7% At least it can be observed that the difference between the first total in the table, say 10% (9.8% to be exact), and the overall total for Brazil (17.7%), i.e. 7.7% (17.7 - 10), is very similar to the 7.8% storage losses given for Brazil.

Although this does not prove the validity of these statistics, it does indicate their plausibility, which is certainly preferable to misleading generalizations or hasty conclusions, not to mention "creative" calculations and worthless "data".

3.2.3. Millet and Sorghum

Millet and sorghum, which cover a number of botanical families, are short-cycle cereals grown primarily in semi-arid regions. They are usually harvested after the rainy season, so that when they reach maturity there is less danger from humidity than from birds and other field pests, whether rodents or wild or domesticated animals, especially from broken stalks and fallen ears or panicles, not to mention natural husking and thefts. However, there are regions where the harvest is gathered during the rainy season, in which case speedy drying is necessary.

Losses can be serious, depending on how long there is before harvesting and how long harvesting itself takes. They can also be considerable during transport, still often carried out manually. Losses can therefore be heavy, even before the harvest is brought in and stored or sold, as a PPA Project in the Gambia (FAO RAPA 86) notes:

Millet and Sorghum: on-field and transport losses
	On-field loss	Transport loss
Early millet	9.5% (average)	7.4%
Late millet	4.2%	n.a.
Sorghum	4.0%	0.9%

This table shows that millet is more vulnerable in the field and in transport than sorghum, and that early millet is slightly more delicate than late.

The overall losses for the two operations can reach about 5% for sorghum, and over 15%, a considerable figure, for early millet. In the absence here of figures for losses of late millet during transport, on the basis of on-field losses it can be estimated that overall losses of millet would not exceed 5 or 6%, a figure similar to overall losses for sorghum.

Millet and sorghum do not keep well as threshed grain. Traditionally, therefore, they are stored in village granaries in ears or panicles; the women then pound them by hand whenever required, certainly the most economical method of threshing. Flail threshing, which is normally done by a team on a make-shift threshing floor in the village or the fields, pushes much grain into the ground, where it is lost. However, mechanical threshing seems to cause the highest loss because of the length of ears in the case of pearl millet, and the density of spikelets on the rachis in the case of other millets and sorghum. Although the many types of pedal or motorized threshing machine now in use save labour and significantly improve yields, they do not prevent breakages, and they leave grain on ears and panicles.

If certain tests are to be believed (source: John Deere and Class combine-harvester), combine-harvesters can entail heavy losses, between 6% and as much as 30%, partially because of the speed at which these machines move. The two figures for threshing losses recorded by the above-cited Gambia project (this time only for millet) are given below, and clearly show that losses are higher with mechanical than manual threshing:

Millet threshing		Loss
	manual	6.3%
	mechanical	19.3%

Storage of millet and sorghum

Many surveys and studies have been carried out in past years on the methods, conditions and effects of storing cereals, particularly millet and sorghum, in farm or village granaries, traders' storehouses, and public or private storehouses, warehouses or silos. They have considered the variety of products, methods and objectives, without forgetting socio-economic and political changes as these affect the agricultural and rural sector, as well as the international market, and the agricultural and food trade.

In the case of small-farmer or village storage, particularly in semi-arid regions where millet and sorghum are the main crops, these studies agree on the validity of ad hoc methods and the effectiveness of traditional granaries, thus contradicting what has very often been recommended. Summary tables of these very detailed observations are given in an annex, but an idea of their wealth can be gained here from two very different studies, one by the Yaciuk team working particularly in Senegal during the 1970s, and the other, more recent, by the African Studies Centre of the Ecole des Hautes Etudes en Sciences Sociales, Paris, 1987.

Losses during storage (over a period of 30 months) in traditional granaries in Senegal (Yaciuk, 1977)

Level in the granary	Loss
	Millet (pearl) Sorghum
1. Top	3.7%	19.2%
2. Centre	1.4%	4.2%
3. Bottom	1.5%	2.4%

These results are only a small sample from a table in which six types or methods of storage intersect with six columns of observations, but they speak for themselves if it is remembered that they cover a period of 30 months' storage. There is an obvious difference in the performance of pearl millet and sorghum, with the latter showing a much higher loss rate in the top layer in the granary, which boosts the purely arbitrary arithmetical average loss for sorghum (almost 9%) as against that for millet (2.2%).

The above-mentioned African Studies Centre study on grain storage in tropical Africa assembles much information and reflection from a number of sources, particularly FAO. The two pages dealing with the issue of losses are illustrated by a table entitled "Moderate losses for longer storage periods than anticipated".

Below are the results, again for millet and sorghum, for four different countries each, with the storage method indicated in parentheses in each case.

1st Table of G. YACIUK

Region	Losses in %	Remarks and sources of information
Rice West Africa	6-24	Drying 1-2; On-farm storage 2-10; Parboiling 1-2; Milling 2-10
Sierra Leone	10	(NAS, 1978)
Maize
Benin	8-9	Traditional on-farm storage 6 months improved silo storage (NAS, 1978)
Ghana	7-14	(NAS, 1978)
Ivory Coast	5-10	12 months stored on the cob (NAS, 1978)
Nigeria	1-5	On-farm storage (NAS, 1978)
	5.5-7	6 months on-farm storage (NAS, 1978)
Togo	5-10	6 months central storage (NAS, 1978)
Millet
Mali	2-4	On-farm stores (Guggenheim, personal comments)
	10-14	Central stores (Guggenheim, personal comments)
Nigeria	0.1-0.2	On-farm stores (NAS, 1978)
Senegal	see Table 1	(Yaciuk, 1976)
Sorghum
Nigeria	0-37	On-farm stores (NAS, 1978)
Senegal	see Table 1	(Yaciuk, 1976)

Source: Yaciuk and Forrest - 1979

2ème TABLEAU DE G YACIUK

Results in the use of traditional granaries made of woven straw and earth (30-months' storage) in Senegal

TYPE		LEVEL	GRAIN ATTACKED	NOT ATTACKED GRAIN	PERCENT GRAIN ATTACKED	WEIGHT GRAIN ATTACKED	WEIGHT GRAIN NOT ATTACKED	PERCENT LOSS
Millet		1. (surface)	3,522	25,683	12.1	15.1	158.3	3.7
		2 (centre)	1,102	32,021	3.3	4.1	202.8	1.4
		3 (bottom)	1,852	26,318	6.6	10.5	193.9	1.5
Millet 30%		1.	11,476	13,938	45.2	78.0	114.4	7.7
		2.	9,411	13,133	41.7	64.3	136.2	14.2
- Sorghum 70%		3.	5,781	13,427	30.1	36.1	172.2	15.4
Sorghum 70%		1.	1,086	10,129	9.7	17.3	265.5	3.8
- Sand 30%		2.	4,090	6,311	39.3	89.8	193.1	11.1
		3.	3,635	7,201	33.5	60.2	210.2	14.5
Sorghum		1.	7,038	5,209	57.5	37.6	156.0	19.9
		2.	938	10,075	8.5	14.9	313.8	4.2
		3.	1,839	12,771	14.3	33.0	282.7	2.4
Sorghum Bromophos		1.	2,546	10,118	20.1	54.1	267.4	3.9
		2.	1,040	10,890	8.7	19.6	292.7	2.7
		3.	2,850	6,470	30.5	57.4	177.9	8.1
Sorghum not threshed		Woven	523	3,002	14.8	6.1	61.4	6.4
		Mudblock	633	5,206	10.8	7.5	102.6	4.3
Sorghum not threshed		Woven	842	5,538	13.2	10.9	110.4	5.3
- Bromophos		Mudblock	474	7,656	17.2	4.3	135.0	2.8

Source: G. YACIUK - 1977 -

Millet (percentage loss)			Sorghum (percentage loss)
Burkina	(ears)	10	Burkina	(ears)	6
Mali	(ears)	2-4	Northern Nigeria	(ears)	4
Niger	(ears)	3.4-10	Senegal	(grain+sand)	9.8
Senegal	(ears)	2.2	Senoufo/Côte d'Iv.	(ears)	11-12

The differences in percentage are relatively minor between the two products, but are considerable between different countries (varying from 2 to 10% for millet, and from 4 to 12% for sorghum). However, the most striking feature, as the study itself says, is "the wide variations in these figures, resulting from the variety of sources but also the number of parameters that have to be taken into account, so that data can never be strictly comparable." One element is, however, the same in seven out of eight cases in this table, namely the form in which the produce is stored (ears), and this positive factor lends further credibility to the conclusions, while helping to show that threshed grain is no less susceptible to pests than grain in ears, despite the useful technique of adding sand to stored grain.

3.2.4. Pulses and oilseeds

Pulses, bean and cowpea, pea, broad bean and lentil in particular, are part of the staple diet of many countries and have the advantage of adding some protein to diets based essentially on cereals, but are harder to harvest and conserve than cereals. When they are mature, the dehiscent pods can open or burst, so that many seeds fall to the ground; this happens during harvesting, but even more so during transport. They are also more vulnerable to insect attack, particularly from weevils, which lay their eggs on the pods or seeds before they are gathered. These pests are often commonly known by the name of their preferred crop, for example the bean-seed beetle (Acanthoscelides obtectus), the cowpea-seed beetle (Callosobruchus maculatus) and the pea-seed beetle (Zabrotes subfasciatus).

Pulses can have a high moisture content at the time of harvesting, depending on species, variety and climate, as is seen from observations on three varieties of bean (source: the frijoles de la Sierra project in Ecuador):

Bean: moisture content at harvesting according to variety:
Panamito:	19.0%
Canario:	20.1%
Frutilla:	36.5%

Effective drying is therefore important, and this is generally done by exposure to the sun on terraces or in openwork containers; it will obviously be more effective if the beans have been shelled first. According to the above-mentioned study, the moisture content can drop to 12 or 13% after three days in the sun, a level generally sufficient safe storage, as is indicated by the following recommendations:

Maximum moisture content for safe storage of pulses and oilseeds
Species	Moisture content
Bean	14.0%
Pea	13.0%
Soybean	11.0%
Groundnut, unshelled	9.0%
Groundnut, shelled	7.0%

It can be seen that oilseeds, soybeans and groundnuts, require a lower moisture content than pulses. Shelled groundnuts are especially demanding here, which means that they are especially vulnerable. Unfortunately, even such drying is not enough to provide immunity to parasite attack, particularly from weevils. An ad hoc method of combating this has therefore been further developed and considerably improved by modern technology. In this approach, the container is filled to capacity and then sealed hermetically so that the interstitial air becomes so ratified and inert that it anaesthetizes or even asphyxiates larvae and insects. When larger quantities are involved and the containers or storage space cannot be sealed hermetically, treatment with insecticide powder is recommended. Below are some figures for bean and soybean losses, taken from the national Brazilian study previously cited.

Post-harvest losses of beans and soybeans (source: Technical Commission for Agricultural Loss Reduction)

	Total loss	Harvest loss	Storage loss
Beans	15.0%	4.3%	9.0%
Soybeans	10.3%	5.0%	2.7%

Although soybeans are thus less vulnerable than beans, particularly during storage, the two operations of harvesting and storage account for the majority of losses for the two crops. In the case of beans, storage alone accounts for almost two-thirds of total losses throughout the pipeline, whereas in the case of soybeans, harvesting accounts for almost half total losses.

Groundnut

Since its fruit grows in the ground, groundnut is different from other oilseeds. The moisture content of the pods is 40 to 50% when they are dug up, which has to be reduced to 10% before threshing is possible. Natural drying in small heaps can take weeks, during which the insects that have already started their work in the fields will continue their inroads, not to mention the danger of moulds. During storage, Trogoderma will cause more damage than weevils, which normally stay on the surface of the stored groundnuts. Long-term storage therefore necessitates fumigation under tarpaulin or in airtight storehouses.

3.2.5. Roots and tubers

Unlike cereals, tropical root or tuber crops are all fragile, perishable foodstuffs. Harvesting them is a delicate process and must be performed with the utmost care, since it affects all subsequent operations, particularly any action to prevent damage or loss. It is therefore important to avoid cuts and bruising during digging, heaping, gathering and transporting, since such damage leaves the door open to viruses and moulds which inevitably lead to rotting.

Yam

The yam is a tuber crop grown in the humid tropics and also in Savannah regions. There are many species, but the white yam or Dioscorea rotundata, which originated in West Africa, is the species most widely grown for human consumption throughout the world.

The tuber reaches maturity after six to nine months' growth, and then enters a period of dormancy lasting from four to eighteen weeks, depending on species. Yams can be stored during this period of physiological rest. However, preservation is a delicate matter, since the tuber's high moisture content, between 60 and 80%, means that transpiration and respiration continue, even if at a reduced rate. Some figures for the possible ranges of moisture content for the main species of yam will give an idea of the problem here.

Moisture content of the main yam species (source: Coursey 1967, cited by J. Knoth in a GTZ handbook, 1993)

Dioscorea alata	65-73%
Dioscorea rotundata	58-80%
Dioscorea esculenta	67-81 %
Dioscorea bulbifera	63-67%

This is why none of the various storage methods, rudimentary, improved or more sophisticated, can prevent serious losses. The various causes of loss can be placed under five main headings:

- nematodes, primarily the yam nematode (Scutellonema), a particularly dangerous parasite that moves in during the growth period and is therefore present at harvest time;

- rot, caused by moulds and bacteria;

- rodents;

- certain insects, especially the white yam worm and the cochineal;

- tuber budding and respiration, leading to loss of weight and of nutritional reserves and quality; germination often makes yams bitter, while transpiration can result in a weight loss of up to 20% in five months of storage.

Storage losses are generally lower in savannah than forest regions, doubtless because the yam nematode is less frequent in drier zones and the moisture content of the tuber is lower.

It will now be helpful to look at some figures showing the performance of yam in the post-harvest pipeline:

Yam weight loss under various conditions (source: FAO, Accra Workshop, 1994)

State of tuber	Weight loss from respiration (% per day)	Total weight loss (% per day)
		Ventilated cellar	Traditional store
Harvest	0.076	0.25	0.25
Dormancy	0.021	0.17	0.27
Germination	0.068	0.23	0.35

The document makes the following comment on this table: "The weekly average of relative humidity in the cellar varied between 83.9 and 93%. These high levels of humidity combined with relatively low temperatures are close to the ideal conditions (250C and 96%) for damaged areas on the tubers to cure. Moreover, favourable conditions have permitted a reduction in weight loss due to respiratory activity in the tubers."

The same publication contains two small tables comparing storage losses (results presented during the workshop held in Accra in 1994) which give an idea of some interesting experiments conducted in Benin to study and improve yam storage. Here we can combine these into a single table:

Losses recorded in two regions of Benin

Central region	Improved structure (raised or cellar hut)	Traditional structure
Early variety	24.2%	57.3%
Late variety	22.4%	38.4%

Northern region	Raised but	Cellar hut	Traditional structure
Early or late variety	26.8%	20.6%	59.1%

A comment on the economic implications of these figures is also worth quoting: "With respect to the economic results, the study showed that the producer obtained for the early varieties a benefit of 220% in value with the improved storage (against 44% with traditional storage). However, with the late variety the benefit was the same whatever the type of storage. There is a risk that the cost of the improved storage together with the phytosanitary treatment... could prove to be a constraint to the development of this type of storage."

Cassava

Cassava is a supplementary and security crop in equatorial and tropical regions, held in high regard by the inhabitants. However, it is even harder to keep than yam, because the root starts to deteriorate on the second or third day after harvesting and exposure to the light. For family or local consumption, the simplest natural storage method is therefore to leave the root in the ground after it has matured, digging it up as needed.

Over the years many techniques have been tried in an effort to extend the keeping-time of fresh roots. Some of these, tested particularly in Ghana (Gallat, 1994) and Nigeria (Agboola, 1994), have had reasonable results, but the period still does not stretch beyond 15 or 20 days, after which the level of rotting becomes unacceptable (reaching 50% after 42 days and 80% after 63 days).

Research has therefore focused mainly on processing, in order to improve the farm-level techniques used by women, and firstly to detoxify the roots. Fresh cassava contains high levels of hydrocyanic acid, which can lead to poisoning. The initial washing and grating processes are thus intended to detoxify it.

The two main products obtained from the root are cassava flour and gari. In both cases production starts with natural fermentation, but the processing operations are much more laborious for gari than flour. The latter, which is easy to store and market, is consumed mainly in the form of foufou, an African staple. Like foufou, gari is sieved after fermentation, but instead of being formed into lumps and boiled, it is baked. This process is called gasification, and produces a couscous-like granular product.

This consideration of cassava processing can close with some information on the labour time for some of the operations involved in producing gari:

manual peeling: 20-25 kg/hour (peel represents 25% of the weight of fresh cassava);
manual grating: 20 kg/hour (motorized graters can reach 2 t/hour);
traditional pressing and fermentation (to remove hydrocyanic acid): 2-3 days;
end product: 20-25% of the weight of fresh cassava.

(source: FAO, Accra Workshop, 1994)

A scheme of three agroecosystems

A schema of three agroecosystems created and managed by farm households in eastern Ghana and Togo. A Field-crop ecosystem away from the home. B. Stored-grain ecosystem in which corn is stored in a crib. C. Stored-grain ecosystem in which cobs or shelled corn is stored in bags. D. Solar drying of damp shelled corn on a mat or a concrete pad. E. Livestock ecosystem in the home compound where sheep, goats, and chickens are fed with grain that is not normally suitable for human consumption.

CEEMAT: CONSERVATION OF GRAINS IN TROPICAL REGIONS

Table of J. Knoth

Table: Comparison between properties of grains and those of roots and tubers from the point of view of their storage aptitude.

Non perishable food crops	Perishable food crops
Exclusively seasonal harvest, need for long storage	Possibility of a permanent or semi permanent harvest hence prolonged storage can be avoided
Preliminary processing (excluding threshing) of crops to be stored rarely necessary	Transformation in dry products frequently substituting storage of fresh product
Low water content fruits, between 10 and 15% less	High water content fruits, generally between 50 and 80%
Small fruits weighing less than 1 gram	Bulky fruits weighing between 5 grams and 5 kg of more
Very low breathing activity of stored fruits, reduced heat	High or very high breathing activity of stored fruit, hence heat generation especially in tropical climates
Tough fibers, good protection against bruises	Soft, highly vulnerable fibers
Good possible natural aptitude along several years	Easily perishable, natural storage capacity going from a few weeks to several months (varies according to species and varieties)
Storage losses generally, due to external factors (moulds, insects and rodents)	Losses partly due to internal factors (breathing, perspiration, germination), partly due to external factors (rotting, insects)