Previous Page Table of Contents Next Page

Quality aspects of tradeable cassava products including problems of adulteration
by J.F. Wood


The utilization of cassava as a primary energy source for livestock (particularly pig feeds) within Europe is now well established and provides valuable foreign exchange to cassava producing countries, most of which are third world /LDCs. Within Germany for example, cassava with a 65% starch content is 25–30 % cheaper than maize. Although cassava requires mixing with soya bean meal in the approximate ratio of 4:1 to form a “cereal replacer” its inclusion in feeds has resulted in a drop in maize imports into the EEC from around 15 million tonnes in the mid-1970's to below 2 million tonnes. The main world producers of cassava are Brazil and Thailand. Thailand's exports have increased from about 2 million tonnes in the 1960's, mainly to the EEC, to 9.2 million tonnes in 1988. Under a self limiting agreement (quota) only 5.25 million tonnes can be sent to the EEC each year, the remainder going to the USSR, Japan, the Korean Republic and Turkey. The EEC has similar “voluntary” agreements with Indonesia, China and Vietnam (Schumacher, 1990).

Table 1 summarises imports of cassava into the EEC for the period 1983–89. Import quotas for 1990 are given in Table 2.

It is interesting to note that exports of cassava from the ACP countries in 1988 and 1989 were well short of the 146,000 tonne quota for the year at only 58 000 and 35 000 tonnes respectively. This suggests an increasing use of cassava within the countries of origin as both a food and feed source, rather than a decline in production level.

When considering the subject of the quality of tradeable cassava it is important to place it in the context in which the cassava will be used. This can be demonstrated by comparing typical usage patterns of cassava in livestock feeds between the Netherlands (a high cassava user), and the UK (a high cereal user) both of which are within the EEC (Table 3). The difference in tapioca usage pattern is predominantly a function of distance from the raw material source within the EEC and the proportionate transport costs.

TABLE 1.Importation of cassava into the EEC 1983–89 ('000 tonnes)
Year GATT countries Non-GATT countries Totals
Thailand Indonesia and others (ACP) China and others  
1983 4396 181 54 4631
1984 5355 398 64 5817
1985 4555 507 115 5177
1986 4679 417 300 5396
1987 5510 954 372 6836
1988 5487 883 353 6723
1989 5482 862 218 6562
1990 3119 811 184 4114

TABLE 2.Import Quotas for cassava into the EEC at 1990 ('000 tonnes)
Thailand 5250
Indonesia 825
PR China 350
Other GATT countries* 146
Non GATT countries** 30
PR China (sweet potatoes) 600

* ACP countries e.g. Brazil, Tanzania and Philippines

** Vietnam

Source: Data from COCERAL (1990) and Schumaker (1990)

TABLE 3. Typical raw material inclusion levels in livestock feeds (%)
  pig finisher layer broiler
Holland UK Holland UK Holland UK
Cereals - 54.0 7.0 60.0 22.0 69.0
Cereal by-products 18.0 9.0 15.0 7.0 - 1.0
Vegetable protein 28.0 24.0 35.0 17.0 39.5 14.0
Animal protein 8.5 7.0 7.0 5.0 8.5 11.0
Cassava 37.0 - 24.5 - 20.5 -
Oil and fat 2.5 2.0 3.0 1.0 7.5 4.0
Others 6.0 4.0 8.5 10.0 2.0 1.0

Source: Data from Nijweide 1987


The trade in cassava into the EEC works through the following chain: farmer, chipper, dryer, pelleter, country of origin export broker, EEC import broker, feed compounder. Some of the larger companies have responsibilities within each of these roles with the exception of feed compounding. In other cases cassava chip and pellet production will be through individual local companies. These smaller producers deal with the export brokers whose responsibilities will be to ensure capacity loadings for shipment. Shipments may therefore be of cassava of mixed origins and quality.

The methodology of cassava chip and pellet production has been well documented. In particular the Proceedings of the IDRC interdisciplinary workshop Pattaya Thailand April 1974 on Cassava Processing and Storage. This Workshop highlighted the manufacturing problems within the cassava chip/pellet industries in Thailand.


At the 1974 IDRC cassava conference, Thanh (1974) and others listed the main criticisms from European customers of cassava pellets from Thailand as being:

  1. nutrient quality: minimum starch content was not achieved (minimum 62 %); maximum sand and foreign matter limits were exceeded (maximum 7% raw cellulose and 3% sand); maximum moisture content was exceeded (maximum 14 % moisture).

  2. physical quality: pellets were of poor friable consistency causing excessive dustiness and high meal content with pellets. Bacteria and mould contents were too high.

These criticisms were particularly addressed to “native” products, i.e. those processed by Thai factories using locally manufactured equipment rather than the higher quality “branded” pellets processed by companies using imported pelleting equipment.

In addition, Mathot (1974) indicated that the financial returns to the pelleting companies were sufficiently low to encourage excessively short drying times, absence or non-use of sand sieves, a great deal of mixing in of foreign matter such as ground corn cobs, tapioca wood, tapioca waste (offal of the starch industry), rice bran and other cheap materials of high crude fibre content.

In contrast cassava chips from Indonesia were found to be of good quality being white, well peeled washed and dried.

Since 1974, the regulations on cassava quality within the EEC have been revised, notably to reduce the maximum moisture content from 14 % to 13 %, and to increase the minimum starch content from 62 % to 63 %. The current EEC quality requirements for cassava products (EEC 1979) are presented in Table 4.

TABLE 4. Current EEC quality requirements for cassava products
Feedstuff Description Composition requirement  
Manioc meal Dried and if necessary washed and peeled Starch moisture min 75%
max 13 %
Chips or roots Manioc roots; also products obtained by crushing and grinding Crude fibre Crude ash Ash insoluble in HCI max 5.2 %
max 5.5 %
max 3.3 %
Manioc type 55 meal Unpeeled manioc roots, dried and if necessary washed Starch Moisture Crude fibre min 63 %
max 13 %
max 9 %
Flakes roots products obtained by crushing and grinding Crude aash Ash insoluble HCI max 6 %
max 4 %

Aflatoxin in straight feedstuffs: 0.05 mg/kg.
Hydrocyanic acid in manioc products: 100 mg/kg (maximum content in mg/kg of feeding stuff referred to a moisture content of 12 %) (EEC 1974).

The present state of cassava quality is now considered under the perspectives of the cassava exporter, the import broker and the feed compounder.


For these data I have relied heavily on literature from Krohn and Co, one of the major German cassava traders and refers to activities in Thailand (Krohn and Co., 1990).

The basic process adopted by the pellet producers is as follows:

  1. Chip whole roots to 1–2cm thickness x 2–5cm length

  2. Sun dry 2–3 days with turning to moisture level of 15–18% which is acceptable for pelleting.

  3. Screen chips to remove fine meal and sand.

  4. Pellet with preconditioning at 80–85°C with dry steam at 10 bar for 15–20 seconds.

  5. Cool pellets and reduce moisture to 13–14%.

  6. Screen for fines.

  7. Spray 0.1–0.5% rice bran or coconut oil to reduce dusting.

  8. Bulk store on godown floor.

  9. Barge from port godown to seagoing ship by lighter.

  10. Voyage Thailand to Rotterdam. Approx 30 days via Suez Canal or 42 days via Cape of Good Hope.

If desired the company can arrange for fumigation to be carried out (eg by SGS at 1–2 kg/1000 cuft methyl bromide gas applied for 24–48 hours).


  1. TISI quality standard. The quality standard of Thailand cassava hard pellets has been set by the Thai Industrial Standard Institute of the Ministry of Industries (TISI) whose inspectors control the quality standard of cassava products through regular visits to the factories and drawing random samples from production.

  2. Pre-Pelleting. Prior to reception at a pelleting plant chips receive a visual check for general appearance, normal colour, odour and mould. After this a sample is drawn and analyzed in the laboratory for sand and moisture content.

  3. Production control. At intervals of 2 hours a sample is drawn from the raw material that goes to production, and the finished product. These samples will be analyzed for sand and moisture, and the pellets additionally for meal and hardness.

    A daily average sample is collected by the private independent surveyor, e.g. Société Générale de Surveillance (SGS) for analysis to full TISI standard analysis.

  4. Export control. The sampling and subsequent analysis are carried out by the Office of Commodity Standard of the Ministry of Economic Affairs. The product is checked twice, upon application for an export licence (pre-loading control) and during export loading (loading control). Depending upon the sales contract a second private control organisation licensed by the OCS may be employed.

Although quality control prócedures for centrally controlled processing operations would appear to be more than adequate, such control is not practicable at smaller units where the products of drying and chipping are passed to independent pelleters and brokers who will be handling material from a variety of sources.


Trade quality limits in themselves do not give us a picture of the state of products entering the EEC. For this we need analytical data on cassava imports.

Examination of data from an independent UK laboratory (Moulder, 1990) of cassava samples tested during 1985–90 indicated some important trends in cassava quality parameters (Table 5). The samples were primarily from shippers/import brokers and represent the range in product quality of cassava entering the EEC from Thailand and Indonesia in particular.

i) CHIPS: Indonesia and Java

TABLE 5. Analytical data for chips and pellets 1985 and 1990
  1985 1990
No tests Mean Range No tests Mean Range
Moisture 152 13.5 10.9–15.9 224 11.7 9.6–14.7
Starch 216 65.5 62.5–74.5 225 72.1 66.2–75.5
C Fibre 136 3.1 2.2–4.9 228 3.2 2.1–4.2
Sand 73 1.3 0.2–9.5 157 0.5 0.1–4.2

ii) PELLETS: Indonesia and Thailand

  1985 1990
No tests Mean Range No tests Mean Range
Moisture 4226 12.7 9.4–17.1 1469 10.9 8.7–16.2
Starch 3924 66.0 57.5–74.5 2164 65.5 56.5–74.5
C Fibre 3692 3.9 2.0–10.7 2003 4.3 2.6–9.2
Sand 2979 2.9 0.2–7.2 1676 3.1 0.2–5.7

From approximately 9000 samples of cassava analyzed over the period 1985–1990 in the ratio 1:8 chips: pellets, equal importance was apparently given to the results for moisture, starch and fibre. The clients requests for sand and /or silica were approx 80% as for the other analyses.

The following additional tests on cassava were requested by clients:

aflatoxin 1
salmonella 1
purity 1
insects 11
hydrocyanic acid nil

Data for the intervening years 1986–89 followed a similar pattern to the above and have therefore not been presented. The most important factors emerging from this data are:

  1. There has been a marked reduction in the mean moisture levels of both cassava chips and pellets to a level which is safely below the 13 % maximum set by the EEC legislation. This will no doubt have resulted in lower levels of microbiological and fungal contamination. Regrettably I have no data to substantiate this proposition.

  2. Almost all samples were above the 63 % minimum starch level.

  3. All samples of chips and a high proportion of pellets were well below the maximum limits for crude fibre and sand of 9 % and 5.5 % respectively.

These figures therefore suggest that there has been a gradual improvement in the quality of cassava chips and pellets since the mid 1970's at least in relation to the EEC requirements. Adulteration of cassava pellets with corn cobs, tapioca wood, sand, etc. would appear to be less common occurrences, such practices having been controlled at the point of origin.

However, results for starch fibre and sand ranged widely from the means and clearly indicate that some shipments could result in potentially costly imbalances of nutrient levels in feed formulations if such variability was not accounted for at the formulation stage.


The apparent absence of requests for HCN by shippers indicates that high cyanide levels are unlikely to be found in commercial cassava chips or pellets for use as an animal feedstuff.

This position is confirmed by the findings of Gomez et al. (1984) in studies on cyanide content of cassava during tray drying. These workers found that chopping of whole cassava allowed a very rapid hydrolysis of the cyanogenic glucoside, leading to a fast rise of the volatile free cyanide portion. Within the dried chips, 60–80% of the total cyanide was present as free cyanide. In this trial, the total cyanide level of 50% of the varieties was below the 100 mg/kg level desired by the EEC. These levels would almost certainly have been lowered further had the chips been exposed to the rigours of pelleting.


Aflatoxin B1 when extracted from feed materials shows intense fluorescence under ultra violet (UV) light. Studies to determine the extent of aflatoxin contamination in cassava from the Colombian feed industry initially revealed important levels of contamination due to the fluorescence noted under UV light. However, previous research at CIAT had shown that cassava post harvest physiological deterioration involves the synthesis of scopoletin, an innocuous natural resin which also fluoresces under UV light. In subsequent studies CIAT reported that the fluorescence in all samples of feed grade cassava chips was due to scopoletin and not to aflatoxin (CIAT 1990).

Furthermore, for over a period of 8 years of cassava use in animal feed in Colombia, only two samples showed traces of aflatoxin and these were negligible. This was considered to be the result from good drying practices since aflatoxin producing mould are unable to grow on chips dried to less than 14% moisture content.

Similar results were obtained by Sajise and Ilag (1987). Their observations suggested that cassava was not a good substrate for aflatoxin production. Cassava chips at 8.6–15.5 moisture served as a substrate for A. flavus, but the conditions did not promote aflatoxin formation. In contrast, Wheatley (1984) found that in a survey of cassava in the Philippines, 100% of samples were contaminated with aflatoxin with a mean of 468 ug/kg, though these figures were qualified by saying that high incidence was not always “real” and depended on the analytical method used.

Moulder (1990) has also confirmed the fluorescence from cassava extracts which would at first appearance seem to be due to aflatoxin, but in fact are due to other compounds.

The possibility of aflatoxin being mistakenly measured by fluorescence techniques therefore suggests that further work is needed on this subject.


The data from the shippers analyst is representative of the material selected by the broker for examination. From the feed compounders point of view the results are biassed in the shippers favour and may not necessarily reflect the findings of the end user at the feed mill. The standards set by the EEC are not considered to be buying standards since these will be agreed between shipper (broker) and feed compounder following negotiations between the two parties. Price negotiation after delivery is considered by some cassava users as an undesirable but necessary step, since there is often considerable variation in apparent quality between the results of the shippers analysis and the more detailed analysis of the feed compounder. Starch content, fibre and sand and silica content are of particular importance in price negotiation.

From a users perspective the problems with cassava may be summarised.

  1. Inconsistent physical and nutritional quality

  2. Excessive dust on handling and trans-shipment

  3. Product contamination with sand an similar materials

  4. Presence of undesirable matter such as sacks, string, insecticide envelopes, etc.

  5. Excessive hardness of some grades of pellets causing jamming of intake augers, fractures of machinery bearings etc.

HCN is not considered to be a problem if cassava is bought with a fibre level of less than about 3.5% where HCN levels may be 10–20 ppm (mg/kg). Fibre levels above 5.5% are often associated with undesirable HCN levels of 80–150mg/kg.

Some feed compounders would prefer to buy cassava of the quality consistency of Chinese sweet potato. This material is traded as dried slices and may be sold as Chinese cassava.

The contamination of feed grade cassava with mycotoxin does not appear to be a major problem. Nevertheless, aflatoxin contamination in cassava for food use has been found at up to 1000 ppb (Rensburg et al., 1979) and appropriate precautions against contamination of feed cassava is highly desirable. Bacterial contamination will, to some degree be reduced during cassava pelleting.

These findings indicate that many of the problems of feed grade cassava summarised in 1974 are still prevalent in 1990. For the feed compounder where cassava is included at high levels in the diet, the variability in the starch content of cassava shipments from 65 to 70% will demand reformulations and adjustments within the mill which he may prefer not to make. In spite of these problems the European feed industry has adapted to handling this material without seemingly applying excessive pressure on the importers to improve the physical quality of the product, and above all, provide a product of consistent quality.


When searching for new opportunities for the marketing of cassava it is essential to remember that cassava, as an internationally traded commodity finds its place in the market because of its competitive nutritional value/price structure relative to cereals. To ensure a continued growth in the utilisation of cassava within the feed industry, it must be priced in relation to world market or local cereal prices, while recognising that cassava is not in itself a cereal replacer, but may be considered as such when blended with a supplementary protein source such as soya.

It would be inadvisable for countries which may be considering the expansion of dried cassava production for export to assume that the EEC is waiting to absorb such production under the present pricing structure. In addition, the cassava trade position may be significantly influenced by possible developments from future GATT negotiations.

Improvements in the quality consistency of cassava must be seen to benefit both cassava producer and compounder alike and the costs of improvements in the quality of cassava must result in a concomitant improvement in livestock productivity.


CIAT. 1990. CIAT International 9 (2): 7 COCERAL. 1990. Comité du Com merce des Cereales et des Aliments du Bétail de la CEE. Brussells. Personal communication. Data of 19 October 1990.

EEC. 1974. Directive 74/63/EEC Council Directive on the fixing of maximum permitted levels for undesirable substances and products in feedstuffs.

EEC. 1979. Directive 79/797/EEC: Directive amending the Annex to Council Directive 77/101/EEC on the marketing of straight feedstuffs.

Gomez, G., Valdivieso, M., De La Cuesta, D. and Kawano, K. 1984. Cyanide content in whole root chips of ten cassava cultivars and its reduction by oven drying or sun drying on trays. J.Fd. Technol. 19: 97–102

Krohn and Co. 1990. Tapioca-Kroh nen Pellets K77. March Krohn and Co. D-2000 Hamburg 1.

Mathot, P.J. 1974. Production and export control in Thailand and the marketing in Europe of Tapioca pellets. In IDRC-031e Cassava Processing and Storage. Eds. Araullo, E.V., Nestel, B. and Campbell, M. IDRC, Ottawa, Canada, p. 27–42

Moulder, A.A. 1990. Personal com munication. Salamon and Seaber Ltd, London.

Nijweide, R.J. 1987. The influence of Rotterdam on raw material usage by a Dutch feed compounder and consequential problems and their resolution. Proceedings of Society of Feed Technologists (UK) 27–38.

Rensburg, S.J., Van De, Kirsipuu, A., Continho, L.P. and Watt, J.J. 1979. Circumstances associated with the contamination of food by aflatoxin in a high primary liver cancer area. S. African Medical Journal 19 (22):877-883.

Sajise, C.E. and Ilag, L.L. 1987. Incidence of aflatoxin contamination in cassava (Manihot esculenta Crantz.) Annals of Tropical Research. 9(3): 127–156

Schumacher, K-D. 1990. Liefer abkommen der EG fur Tapioka Kraftfutter 7: 273–275.

Thanh, N.C. 1974. Technology of cassava chips and pellets processing in Thailand. In IDRC-03le Cassava Processing and Storage. Eds. Araullo, E.V., Nestel, B., and Campbell, M. IDRC, Ottawa, Canada, p. 113–122

Wheatley, C. 1984. Aflatoxin in cassava…is it a real problem? Cassava Newsletter 8, 2:2, 14

Previous Page Top of Page Next Page