REVIEW: A quarterly journal on animal health, production and products no. 62-1987

Sub-Saharan Africa

OPTIONS FOR SMALLHOLDER MILK PROCESSING

* F. O'Mahony and J. Peters

   ————————

*F. O' Mahony was head of the Dairy Technology Group and J. Peters is director of research at the International Livestock Centre for Africa, PO Box 5689, Addis Ababa, Ethiopia.

________________________

It is with great regret that we inform our readers of the sudden death of Dr F. O'Mahony, co-author of this article, while on duty for ILCA in southern Ethiopia.

Milk composition
Products for smallholder manufacture
Processing: options for development
Conclusions
References

African smallholder milk producers in rural areas can process surpluses on the farm or in centralized small-scale units. This article describes some products suitable for manufacture at these levels and discusses process efficiency and product keeping quality. Application of the techniques discussed for processing milk produced in rural areas would provide cash income for the small milk producer and would also contribute to dairy development in rural areas not served by structured markets. The technologies described can be transferred rapidly by appropriate training programmes.

Milk is a product of almost every production system in sub-Saharan Africa: pastoral systems, agropastoral systems in the drier zones of the continent, and mixed farming in areas with more favourable climates and a higher population density. Specialized, highly intensive dairy systems can be found in the vicinity of urban centres. In subhumid zones with a tsetse challenge, milk production is very limited. Milk produced in rural areas is usually consumed as fresh or sour milk, and only in the vicinity of markets are milk surpluses processed into dairy products having a longer shelf-life. Organized marketing of smallholders' milk is not well established, and is basically limited to Kenya.

Dairy development in Africa has been hindered by marketing constraints, including poor access to markets in rural areas, low durability of products, absence of a structured marketing system (International Development Research Centre [IDRC], 1984), and unattractive prices to producers where structured marketing does exist.

Past efforts aimed at improving dairy production in Africa have focused on the establishment of large-scale centralized processing plants to meet the liquid milk demand of urban dwellers (Von Massow, 1985). Because of inadequate milk collection systems and unattractive prices offered for locally produced fresh milk, these plants rely on imported butter oil and skim milk powder for reconstitution and recombination to meet the market demand. In most African countries, direct competition between cheap reconstituted milk and locally produced fresh milk has discouraged smallholder dairying (IDRC, 1984).

Churning sour whole milk using a detachable agitator in an experimental trial at ILCA's Debre Zeit Station, Ethiopia

Churning sour whole milk using traditional methods in an experimental trial at ILCA's Debre Zeit Station

Churning sour whole milk in a clay pot churn in the Ethiopian highlands. The churn nests on a mat on the ground—the most common method observed

Churning sour whole milk in a bottle gourd supported on a tripod, Ethiopian highlands

Most countries in sub-Saharan Africa have formal dairy marketing subsystems that cater primarily for urban milk supplies, and an informal marketing subsystem that operates in the rural areas (Mbogoh, 1984). Brumby and Gryseels (1984) stated that in Africa only a small fraction of milk production enters the official commercial sector and many processing plants operate well below their designed capacity. Even in Kenya, a country with an efficient marketing system, it is estimated that only 10 to 15 percent of the total milk production, just over half the amount available for marketing, is sold through the Kenya Cooperative Creameries (KCC), which has a legal monopoly on commercial milk distribution. Lack of small-scale processing techniques is one of the problems encountered at the smallholder level (Brumby and Gryseels, 1984).

Processing milk into stable marketable products such as butter, ghee, cottage cheese, hard cheeses and fermented milks could generate cash income for smallholder producers in rural areas, permit reinvestment in the enterprise, yield by-products for home consumption and enable smallholder milk producers to conserve milk solids for sale or consumption in times of scarcity.

Efficient smallholder milk processing could stimulate increased milk production in the smallholder sector. This article describes some products that can be made by smallholders, the technologies required for small-scale manufacture of these products, and the product options available to smallholders.

Milk composition

People in Africa use milk from cows, sheep, goats and camels. Of these sources, cow milk is the most widely produced and processed. The term "milk" in this paper will thus refer only to cow milk.

The chief solid constituents of milk are fat, proteins and lactose. The concentration of each constituent varies, and is determined by both genetic and environmental factors (Webb, Johnson and Alford, 1983). Table 1 shows the average composition of Bos taurus and Bos indicus milk.

Milk processing exploits one or more of the major constituents of milk solids. Water and minor constituents such as milk salts are incorporated in all milk products to varying degrees. From the commercial point of view, milk fat is the most significant solid component in milk, while milk proteins are most important for human nutrition (O'Sullivan, 1973).

Fat. Fat is the most variable of all milk constituents, comprising a mixture of fatty acids that can be saturated or unsaturated.

Fat is suspended in milk in tiny droplets known as globules, and forms an oil in water emulsion. Fat can be concentrated in cream by gravitational or centrifugal separation, and de-emulsified from either milk or cream by agitation, which breaks the emulsion. Once de-emulsified, fat can then be recovered as butter, which contains fat, some water and non-fat milk solids. Water can be removed from the butter by evaporation, and the clear fat separated from non-fat milk solids to produce butter oil (McDowall, 1953).

Proteins. Proteins are divided into two main groups: casein proteins, which constitute the major fraction and account for up to 80 percent of the total protein, and whey proteins, which account for 20 percent of the total proteins but are of even higher nutritional value than casein (O'Sullivan, 1973).

Casein consists of fractions precipitated from raw skim milk by acidification to pH 4.6 at 20°C. Enzymatic precipitation of casein is used to coagulate milk for cheese-making. Rennin, a proteolytic enzyme extracted from the abomasa of 10- to 30-day-old milk-fed calves, has traditionally been used. A variety of coagulants of animal, plant and microbial origin can also be used for cheese-making (Green, 1977).

Whey proteins are more soluble than the caseins. They are denatured when heated to temperatures above 60°C (Harland et al., 1953), and can be recovered by acid precipitation.

Lactose. Lactose is a disaccharide in milk (Webb and Johnson, 1983). It is usually the predominant solid in milk, although some high-yielding cows produce milk containing more fat than lactose. Lactose is present in solution and is therefore more difficult to recover from milk as an isolated fraction.

Minor milk components. Minor components of milk include salts, enzymes, vitamins and trace elements. Milk has a very high concentration of calcium (Ca) and phosphorus (P). In a total concentration of 132.1 mg/100 ml Ca and 95.8 mg/100 ml P, 32.2 percent of the Ca and 37.9 percent of the P exist in the soluble phase, while in the colloidal phase 61 percent of the Ca and 61–63 percent of the P are associated with the casein (Webb and Johnson, 1983).

Because the colloidal salts are associated with the casein, they are largely recovered with the casein in rennet cheese manufacture. The physical stability of the caseinate is very dependent on the composition of the salt system. Thus the physical state of the salt system affects rennet coagulation time (Fox, 1969a).

Products for smallholder manufacture

Although the manufacture of many dairy products requires a high degree of technical and capital input, there is a range of products suitable for smallholder processing using simple equipment. The processing steps required for their manufacture and the composition and keeping quality of some of these products are given in Tables 1 to 5.

TABLE 1. Process, composition, yield and keeping qualities of dairy products and byproducts

Process	Dairy product	Composition					whey protein (%)	Yield (litre/ kg)	Keeping quality	Reference
		Mois- ture%	Fat (%)	Non-fat solids
				Prot- ein%	Lactose (%)	Ash (%)
None	Fresh Bos Taurus milk	87.2	3.7	3.5	4.9	0.7	0.75	. . .	5 hours	Webb, Johnson and Alford, 1983
	Fresh Bos indicus milk	86.1	5.3	3.4	4.6	0.6	. . .	. . .	5 hours	Ethiopian Nutrition Institute, 1980
Milk separation	Separated milk	90.5	0.1	3.6	5.1	0.7	0.75	1.2	5 hours
	Cream	57.3	36.8	2.2	3.2	0.6	. . .	10.0	1 day

NOTE: ... means data not available

TABLE 2. Process, composition, yield and keeping qualities of fermented milks

Process	Dairy product	Composition					Lactic acid (%)	Yield (litre/ kg)	Keeping quality	Reference
		Mois- ture%	Fat (%)	Non-fat solids
				Pro- tein%	Lactose (%)	Ash (%)
Natural fermentation	Irgo (sour milk)	88.5	2.2	4.7	3.9	0.7	0.9	1.0	10 days	Ethiopian Nutrition Institute, 1980
Sterilization followed by innoculation	Kefir	89.4	2.0	3.5	4.0	0.7	0.6	1.0	10 days	Webb and Johnson, 1983
	Yoghurt	87.2	3.4	3.4	4.1	0.6	0.9	1.0	10 days	Webb and Johnson, 1983
Acid fermentation	Acidophilus skim milk	90.1	0.5	3.5	4.4	0.7	0.7	1.0	1 day	Webb and Johnson, 1983

TABLE 3. Process, composition, yield and keeping qualities of milk fat products

Process	Product	Composition					Yield (litre/ kg)	Keeping quality	Reference
		Moisture (%)	Fat (%)	Non-fat solids
				Protein (%)	Lactose (%)	Ash (%)
Churning sour milk or cream	Butter	16.5	80.5	0.1	0.4	2.5	20.0	10 days	Webb and Johnson, 1983
Moisture removal by boiling	Butter oil, ghee	0.1	99.8	0.1	. . .	. . .	25.0	>3 months	Webb and Johnson, 1983
Heating and spicing	Spiced butter	17.2	81.2	1.3	. . .	2.1	20.0	2 months	Ethiopian Nutrition Institute, 1980

NOTE:....means data not available

TABLE 4. Process, composition, yield and keeping quality of by-products of butter-making

Process	Dairy product	Compostion						Whey Protein (%)	Lactic acid (%)	Yield (litre/kg)	Keeping quality	Reference
		Moisture (%)	Fat (%)	Non fat solids
				Protein (%)	Lactose (%)	Ash (%)	Salt (%)
Milk separation	Separated milk	90.5	0.1	3.6	5.1	0.7	0.8	0.75	...	1.1	5 hours	Webb and Johnson, 1983
Sour milk churning	Butter-milk A	91.5	1.4	3.1	3.2	0.6	. . .	0.75	0.8. . .	1.1	5 days	ILCA Studies
Cream churning	Butter milk B	90.5	0.4	3.6	4.3	0.7	. . .	. . .	. . .	16.0	5 days	Webb and Johnson, 1983
Heating ripened butter milk A	Cottage cheese	76.0	3.0	18.0	—	2.7	. . .	. . .	. . .	6.0	5 days	ILCA Studies
	Whey	93.5	0.4	0.7	. . .	0.8	. . .	. . .	. . .	. . .	. . .

NOTE: means data not available

 TABLE 5. Process, composition, yield and keeping qualities of various cheese varieties

Process	Variety	Composition				Yield (litre/kg)	Keeping quality
		Moisture (%)	Fat (%)	Non-fat solid
				Protein (%)	Ash (%)
Heating milk to 32°C, adding rennet, cooking the curd	Halloumi	35.0	33.0	26.0	3.0	9.5	>60 days
Heating milk to 83°C + acidification	Queso Blanco	46.0	23.0	24.0	3.0	8.0	>60 days
Heating milk to 32°C , adding rennet, picking	Feta	49.0	27.0	17.0	3.0	9.0	>60 days
Heating renneted milk to 35°C, salting	Gibbneh Beda	51.0	23.0	22.0	4.0	9.0	>60 days
Heating salted milk to 35°C , adding rennet	Domiati	61.5	12.0	12.5	4.0	6.0	>60 days

Figure 1 shows the incorporation of the principal milk solid constituents in the milk products listed in the tables. Sour milk and yoghurt contain all the constituents of milk, cream, butter and butter oil, all of which are high in fat. Casein is incorporated with fat and moisture in hard cheeses and can also be precipitated from defatted milk as cottage cheese. Lactose and the whey proteins are more difficult to recover but can be consumed by humans or fed to animals in the form of whey.

Figure 1.The incorporation of the major solid fractions of milk in milk products

Fermentation reduces milk pH, limiting the growth of putrefactive and lipolytic bacteria and thus preserves other milk solids (Kosikowski, 1982). The low moisture content of butter and cheese retards bacterial growth, giving these products greater storage stability than whole milk. Salting of products such as butter and cheese has an additional preservative effect by retarding bacterial growth (Hunziker, 1927; Kosikowski, 1982).

Milk products of good storage stability can thus be made by applying principles of acidification, moisture reduction and salting.

Fermented milk products. Fermented milks are products obtained from good-quality milk subjected to one of a number of controlled fermentations (Fox, 1967). Fermented milks are favoured over fresh fluid milk in certain countries for reasons of hygienic safety, better flavour and texture, and possible beneficial therapeutic effects (Kosikowski, 1977). The Russian biologist Metchnikoff attributed the longevity of the people in the Balkan states to their very high consumption of fermented milk products (Fox, 1967).

Production of fermented milk is a tradition widespread in sub-Saharan Africa. In Ethiopia, milk is allowed to ferment naturally without the addition of a starter. Incremental quantities of milk are accumulated each day in a clay pot of about 22 litres or in a bottle gourd (O'Mahony and Bekele, 1985a), where up to 1 percent lactic acid develops. In cold weather the vessel is placed near the fire to keep the milk warm. The resulting sour milk—irgo— is viscous, has a strong lactic acid flavour, and is consumed as a side dish. If not consumed as irgo, the sour milk is used for butter-making. A similar process is used by the Fulani pastoralists (Waters-Bayer, 1986), and by the Masai (Grandin, personal communication). Borana pastoralists in Sidamo, Ethiopia, make a concentrated fermented milk by removing clear whey from coagulated cow milk; about one-sixth of the total milk volume is removed as whey. The product can reportedly be stored for up to 20 days (O'Mahony, 1986). The procedure of heating milk to sterilization temperatures followed by cooling and inoculation to achieve a controlled fermentation has not been reported from these areas. Heating kills undesirable organisms and also concentrates the milk, giving the fermented product a heavier body (Fox, 1967).

Technology of fermented milk production. All fermented milks have undergone a lactic or mixed acid fermentation. The milk is first heated to the boiling point, and may be boiled for a considerable time. It is then cooled and inoculated with some fermented milk from the previous batch and held at an incubation temperature of 30–38°C for the duration of the fermentation (three to four hours). Higher temperatures cause the whey to separate, while lower temperatures result in slow fermentation that may allow the establishment of undesirable microorganisms. The products can be made in simple containers of suitable size.

Off-flavours described as unclean, putrid and bitter are caused by coliform and putrefactive organisms. Good sanitary procedures and adequate heat treatment prior to fermentation help to preclude the problem.

An acid flavour occurs if the fermentation is too vigorous; wheying off occurs as a result of too-high temperatures during fermentation.

Consideration should be given to the introduction of fermented milks into smallholder processing systems, as they are more stable in storage than fresh milk and have a better product consistency than naturally soured milks.

Milk fat products. Butter. Milk fat is commercially the most significant solid component of milk. Butter and butter oil are the most important products made from milk fat. Butter can be produced by churning either sour whole milk or cream separated from whole milk (see Fig. 2). Sour milk is usually churned in a bottle gourd or earthenware jar (O'Mahony and Bekele, 1985a), but any watertight vessel of the required volume is suitable. Churning efficiency can be improved by the use of a detachable agitator (see Fig. 3) (O'Mahony and Bekele, 1985a). Provided that the churning temperature is held below 16°C, the system enables efficient recovery of milk fat as butter from up to 12 litres of milk in each batch (Bekele and O'Mahony, n.d.). Cooling is achieved by evaporation.

Figure 2. Manufacturing steps for butter and by-products of butter-making

Figure 3. A detachable wooden agitator for increasing the churning efficiency of traditional butter-making

Removing the agitator after butter grain formation. Note the butter grains on the mouth of the churn

Recovery of butter by filtering the churned milk through a muslin cloth

On farm use of a detachable agitator for churning sour whole milk

When making butter from larger volumes of milk (up to 300 litres), centrifugal separation (see Fig. 4) followed by cream churning in a wooden churn (see Fig. 5) is more appropriate. Centrifugal separation concentrates the fat in the cream phase; with proper temperature control (35–40°C) and correct operation of the separator, as little as 0.1 percent of the milk fat remains in the separated milk (Hunziker, 1927). This process produces butter and skim milk and allows for a wider range of processing options than are possible with soured milk.

Figure 4. A cream separator

Figure 5. A wooden churn for cream

After separation, cream is allowed to ferment for two to three days until the serum acidity exceeds 0.5 percent. During this time the milk fat solidifies. Both of these factors contribute to efficient churning, while the acid developed contributes to the butter flavour. The churn should be filled to half of its volumetric capacity with cream. The churn type shown in Figure 5 is normally used to churn up to 20 litres of cream per batch. As churning temperature decreases from 24 to 14°C, churning time increases, as does fat recovery as butter (Hunziker, 1927).

The breakpoint in churning sour whole milk or cream is reached when butter grain forms. The churn contents are then strained through a muslin cloth to recover the butter. Washing the butter in cold water removes residual non-fat milk solids. This practice reduces the off-flavours associated with non-fat milk solids but also depresses butter yield.

The quantity of butter fat extracted from milk and the amount of moisture incorporated in the butter are the two main factors determining butter yield. The standard level of fat in butter is set at not less than 80 percent milk fat, and most developed countries also specify that the moisture level not exceed 16 percent (Webb and Johnson, 1983). While processing losses are unavoidable, these should not exceed 5 percent of the total fat present when churning sour whole milk, and 3 percent of the total fat present when churning cream. Care should also be taken to ensure complete moisture incorporation to 16 percent of the butter by weight.

Fireplace used for heating vessels containing buttermilk for cottage cheese manufacture

Heating buttermilk to 60 °C to precipitate cottage cheese

Cottage cheese is removed upon cooling

Assuming that butter composition is 82 percent fat, 2 percent non-fat milk solids and 16 percent moisture, the expected butter yield from 100 kg of milk containing 3.7 percent fat can be calculated as follows:

• Weight of fat = 3.70 kg
• Production losses of fat = 5 percent
                       = 0.185 kg
• Fat available for butter-making 
                       = 3.515 kg
• Expected yield of butter containing 82 percent fat = 3.51 x 100/82
                       = 4.29 kg

Salting butter to 2 percent increases its keeping quality and yield. The expected yield of salted butter containing 80 percent milk fat, 2 percent milk solids, non-fat, 2 percent salt and 16 percent moisture from 100 kg of milk containing 3.7 percent fat is 3.515 x 100/80 = 4.39. kg..

Butter quality deteriorates rapidly at ambient temperatures. Besides the off-flavours attributed to the non-fat milk solids there are two sorts of off-flavours that develop in the fat itself—hydrolytic and oxidative rancidity (Downey, 1970). Hydrolytic rancidity is caused by lipolysis, while oxidative rancidity is caused by lipid oxidation and occurs in unsaturated fatty acids. Through a series of reactions, a number of products are formed that cause off-flavours in butter. Lipid oxidation is accelerated by the presence of metals (Cu, Fe and Ni), low pH, oxygen, temperature and light (Downey, 1970). It can be retarded by packing butter in light-barrier packages and storing it in a cool place.

Off-flavours caused by hydrolytic rancidity dominate the flavour of traditionally produced butter. Rancid butter is desirable for certain markets, but as the free fatty-acid level increases above 10 percent, the market value of the butter decreases. The development of rancidity can be retarded by salting butter to 2 percent salt by weight, which delays the development of hydrolytic rancidity, presumably by inhibiting bacterial activity. Use of polythene-lined, light-barrier packaging, salting and storing in a cool place are therefore recommended to minimize hydrolytic rancidity in butter.

Butter oil. Butter oil, also known as dry butterfat or ghee, consists of butterfat that is almost completely free of water, protein, milk sugar and mineral substances (McDowall, 1953). In some areas of India, the product is made by direct evaporation of milk or cream (Madan Pal and Rajorhia, 1975). This process has a high thermal energy demand and results in much of the fat becoming entrained in curd particles. A more efficient method is to evaporate the moisture from butter made by churning sour whole milk or cream (Madan Pal and Rajorhia, 1975).

Melting butter by heating it to 60°C in an equal volume of water followed by centrifugal separation gives a good butterfat yield, free from non-fat milk solids and containing not more than 1.5 percent moisture (O'Mahony, unpublished). The residual moisture can be removed by further heating.

According to Madan Pal and Rajorhia (1975), the product yield in butter-oil manufacture should be about 97 percent of the total fat processed.

The heat treatment given during manufacture and the low moisture content in the product prevent the development of hydrolytic rancidity in butter oil. Oxidative rancidity will occur but can be minimized by packing the product in light-barrier, airtight containers. Once packed, the product should be stored in a cool place (McDowall, 1953).

By-products of butter-making. Butter-making recovers almost all the fat in milk but not the protein and lactose. Further processing of milk after removal of fat yields protein in a stable form. Churning sour whole milk gives buttermilk A, cottage cheese and whey as by-products. The main by-products of butter-making based on milk separation and cream churning are separated milk, cottage cheese, whey and buttermilk B. Fresh separated milk can be consumed by either humans or animals, or it can be added to cheese milk. The latter is an attractive option since it increases cheese yield. Separated milk can be ripened and heated to precipitate cottage cheese. The supernatant whey can either be consumed or fed to animals.

In the traditional system, buttermilk A is usually consumed without further processing (Beyene Kebede, 1983). In certain areas of Ethiopia, cottage cheese is heat precipitated from buttermilk A. The process gives good recovery of casein and residual fat and yields a marketable product (Whalen, 1985).

Initial studies by the International Livestock Centre for Africa (ILCA) indicate that whey contains about 0.75 percent protein, implying near complete recovery of casein. An average of 8 litres of buttermilk A is needed to produce 1 kg of cottage cheese. The product composition is 76 percent moisture, 19 percent protein, 3 percent fat and 2 percent ash.

The high moisture content in the product contributes to its poor keeping quality. It has been reported that this is improved in some areas by heating buttermilk A to high temperatures and pressing out as much moisture as possible.

In summary, the on-farm processing of small quantities of accumulated sour milk is flexible, requiring little capital input and attaining a high efficiency in recovery. While the processing of larger milk volumes into butter and separated milk gives the processor more efficient recovery of butterfat and provides more options for the disposal of separated milk, it also requires more equipment.

Cheese varieties. Cheese is known to have been the standard fare of the early Greeks and Egyptians. At present, there are over 2 000 varieties of cheese recorded (Kosikowski, 1982). Cheese is a suitable source of fat and protein (Scott, 1981). Many cheese varieties require considerable technical equipment and skill in their manufacture. An important aspect of cheese-making is the maintenance of starter cultures for controlled fermentation, while for many varieties special storage facilities for the correct ripening of the cheese are required.

The cheese varieties discussed in this section can be manufactured using simple procedures that are well suited to the economic conditions of smallholders in Africa. Some of the varieties are already made in parts of Africa, and resemble either boiled curd or pickled cheeses that can be ripened and preserved under tropical conditions.

The manufacturing procedures for Halloumi, Queso Blanco, Feta, Domiati and Gibbneh Beda cheese are shown in Figures 6–10. The composition of the five varieties tested at ILCA and their expected yield are given in Table 5. Cheese yield is affected by the concentration of fat, casein and insoluble salts in milk as well as by production efficiency and moisture content of the final product.

Figure 6.Manufacturing steps for Halloumi cheese

Figure 7.Manufacturing steps for Queso Blanco

Figure 8.Manufacturing steps for Feta cheese

Figure 9.Manufacturing steps for Domiati cheese

Figure 10. Manufacturing steps for Gibbneh Beda cheese

Halloumi cheese originated in Cyprus, where it was first made from sheep and goat milk (Brumby, personal communication). Its manufacture does not require starter cultures and the fat content of the milk, usually from cows, has not been adjusted by separation. At ILCA, the cheese has been made from rennet and bovine pepsin, and the trials have indicated that the latter does not affect cheese quality or yield.

To manufacture Halloumi cheese, the curd is heated to 70–80°C. This pasteurizes the curd, gives the correct curd texture and denatures proteolytic enzymes. A variety of coagulants can be used to make Halloumi cheese without adversely affecting flavour during ripening. After heating, the curd is removed, sprinkled with dried leaves of Mentha viridis and folded. Halloumi can be consumed fresh or as a salted cheese ripened in whey brine. A whey cheese, Anari, is recovered from the residual whey as a by-product (O'Mahony and Bekele, 1985b).

Queso Blanco originated in Latin America (Kosikowski, 1982), and is so named because it is a white cheese. Queso Blanco is made by acid precipitation of milk solids at 83°C. Milk is normally standardized to 3 percent fat, giving a fat to casein ratio of 1.3:1. Various acidulants such as acetic, lactic and citric acids can be used.

Lemon juice is a readily available acidulant in many parts of Africa. Queso Blanco has been made at ILCA using lemon juice diluted 1:1 with water (O'Mahony and Bekele, 1985b). The lemon juice imparts an acceptable flavour to the cheese and the product yield is good. The low protein content of the whey (see Table 7) indicates good recovery of whey proteins, which coprecipitate with the casein. In addition to the high yield of cheese, cream removed in standardization can be used for butter or gheemaking. Whey can be consumed by humans or animals. Queso Blanco can be consumed fresh or stored for extended periods.

Feta cheese is a pickled cheese, i.e. a cheese in which very high levels of salt are used as a preservative. Feta cheese originated in Greece and is normally made from sheep milk (O'Keeffe and Phelan, 1979). When cow milk is used, the fat content is adjusted to give a fat to casein ratio of 1.3:1. Feta cheese was made at ILCA from unpasteurized milk using the procedure outlined in Figure 8. The cheese is preserved well in brine and developed a desirable flavour over a four-month storage period. The high salt content of the brine ensures good storage stability.

Domiati cheese is one of the most popular cheeses in Egypt (Fox, 1969b). Gibbneh Beda, a variant, is made in the Sudan, where it commands a high market price. The high salt concentration in these cheeses might also prove to be a suitable preservative in countries such as Ethiopia. The five varieties investigated at ILCA exhibited good keeping quality and yield. Initial field studies indicate that the technologies required for their manufacture are adaptable by smallholders and the products acceptable where they were introduced.

Processing: options for development

Milk as it comes from the cow is near body temperature, about 38°C. Bacteria multiply very rapidly in warm milk (Robinson, 1983). Many countries in sub-Saharan Africa have a hot, humid climate for most of the year. Under such conditions, the raw milk spoils easily during storage unless it is cooled or, when possible, treated with a preservative (Korhonen, 1980). Preservatives, are not readily available in rural areas, while cooling systems are not feasible in some areas because of water shortages. In many cases this makes the collection and delivery of milk to a centralized processing plant difficult, as the milk quality deteriorates to below acceptable standards by the time it reaches the plant. The small milk quantities produced, seasonal supply and poor infrastructure make rural milk collection systems uneconomical and difficult to operate. In many rural areas, the stimulation of efficient on-farm processing or village-level processing appears to be a better strategy for dairy development. The choice between on-farm or village-level processing depends on the quantity of milk produced in the vicinity of a village.

In view of the milk volumes available for processing at the smallholder level and the need for small-scale processing techniques, ILCA has tested a number of processing options for small-scale milk producers.

Smallholders' selection of products for manufacturing is influenced by market conditions, milk quantity, available technology and consumer preferences.

Precipitation of fat and casein using lemon juice as acidulant

Transferring salted precipitated curds to cheese mould prior to pressing

Pressing Queso Blanco curds using a lever-action cheese press

Draining residual whey from cottage cheese

Allowing milk to ferment naturally is a suitable option for on-farm processing because the daily milk quantities accumulated are small, the high acidities in fermented milks preserve the other milk solids (Kosikowski, 1982), and churnability is improved (Bekele and O'Mahony, in press).

ILCA studies have shown that as little as 0.1–0.2 percent fat by Gerber analysis remains in the buttermilk left after churning sour whole milk at temperatures of 14–16°C (O'Mahony and Bekele, 1985a). The equipment used to make butter from sour whole milk is locally available. A simple detachable agitator designed at ILCA to improve churning efficiency costs US$5. The butter produced can be sold, salted and preserved for subsequent sale, or converted to ghee for sale or storage.

Residual liquid from butter-making contains all the non-fat milk solids. It is highly nutritious and can be consumed without further treatment. Casein recovered from the liquid can be used as fresh cottage cheese or preserved by producing a salted or low-moisture cottage cheese. The whey can be used for human consumption or as animal feed.

Village-level processing can be successfully undertaken with milk volumes of up to 500 litres/day. The recommended equipment includes hand-driven milk separators and wooden churns. ILCA's experience to date has shown that this equipment is quickly adopted in areas where milk volumes justify its use. Butter of good yield and keeping quality is being produced by milk separation and cream churning at 15 producer cooperatives throughout Ethiopia. Milk volumes of up to 200 litres/day are being processed at each unit. Centralized village processing at this scale where individual smallholders voluntarily bring their milk is now being investigated. It is too early as yet to comment on the effectiveness of this system in Ethiopia, but experience in India (Mogens Jul, 1977) suggests that it could be successful. Small-scale centralized milk processing reduces labour requirements on the farm for milk processing, provides a nucleus for sale of farm inputs and strengthens the marketing capability of individual milk producers.

The manufacture of a variety of cheeses is difficult with small quantities of milk (1 to 2litres/day). However, a village cooperative could make cheese and thus widen its product range. The feasibility of making cheese at village cooperative is now being investigated in the field.

Cheese-making may be an option for rural cooperatives that are far from the main roads and thus have to accept a lower price for their butter, and that do not have any market for local cottage cheese.

The cheese varieties discussed in this article are more stable products than butter. Their manufacture would give farmers greater independence from local traders, who are able to exploit the fact that products such as butter have to be sold quickly. Research will be necessary to assess the market for cheese in rural areas before embarking on cheese production.

A rapid introduction of appropriate dairy processing technologies requires the training of extension agents to work in rural areas. ILCA has held four courses in rural dairy technology for extension agents over the past two years. The trainees were given basic training in milk chemistry, microbiology and in the technologies described above. Follow-up visits revealed that the trainees were passing on their newly acquired skills to farmers and that the equipment was being properly used.

ILCA's Dairy Technology Group is now preparing a dairy technology course for use in national training institutes.

Conclusions

There are several milk processing options available for smallholder milk producers. The technologies required for these options could be applied to process milk in rural areas currently not served by a structured milk marketing system. The rapid adoption of some of these technologies in Ethiopia, where they were tested by ILCA, indicates that milk producers themselves recognize the need for improved techniques. The introduction of efficient intermediate technology for smallholder milk processing in rural areas could contribute significantly to the sustained development of smallholder dairying.

References

BEKELE, E. & O'MAHONY, F. Factors influencing the churnability of sour whole milk. (In press)

BEYENE KEBEDE. 1983. Milk production in Ethiopia. Paper presented at IDRC Workshop on Potential for Small-scale Milk Production in Eastern and Southern Africa, Nairobi.

BRUMBY, P. & GRYSEELS, G. 1984. Stimulating milk production in milk deficit countries of Africa and Asia. ILCA Bull., 19: 2–7.

DOWNEY, W. K. 1970. Butter quality. Dublin, An Foras Taluntais. Dairy Research and Review Series No. 7.

ETHIOPIAN NUTRITION INSTITUTE. 1980. Ethiopian traditional recipes. Addis Ababa.

Fox, P.F. 1967. Fermented milks. Dublin, An Foras Taluntais. Dairy Series, Tech. Bull. No. 6.

Fox, P.F. 1969a. Effect of cold-ageing on the rennet-coagulation time of milk. Irish J. Agric. Res., 8: 175–182.

Fox, P.F. 1969b. Manufacture of Feta cheese. Dublin, An Foras Taluntais. Dairy Series, Tech. Bull. No. 1.

GREEN, M.L. 1977. Review of the progress of dairy science: milk coagulants. J. Dairy Res., 44:159–188.

HARLAND, H.A., COULTER, S.T., TOWNLEY, V.H. & JENNESS, R. 1953. The denaturation of milk serum proteins at temperatures ranging from 180°–290° F. J. Dairy Sci., 36: 568–569.

HUNZIKER, O.F. 1927. The butter industry. La Grange, Illinois.

INTERNATIONAL DEVELOPMENT RESEARCH CENTRE (IDRC). 1984. The potential for small-scale milk production in Eastern and Southern Africa. Ottawa, Canada. IDRCMR 98 E.

KORHONEN, H. 1980. A new method for preserving raw milk—the lactoperoxidase antibacterial system. Wld Anim. Rev., 35: 23–29.

KOSIKOWSKI, F.V. 1977. Cheese and fermented milk foods. 1st ed. Brooktondale, New York, Kosikowski and Assocs.

KOSIKOWSKI, F.V. 1982. Cheese and fermented milk foods. 2nd ed. Brooktondale, New York, Kosikowski and Assocs.

MADAN PAL & RAJORHIA, G.S. 1975. Effect of multiple separation of cream on the recovery of ghee. Indian J. Dairy Sci., 28 (1).

MBOGOH, S. 1984. Dairy development and dairy marketing in sub-Saharan Africa. Some preliminary indicators of policy impacts. ILCA Bull., 19: 8–16.

McDOWALL, F.H. 1953. The buttermaker's manual. Wellington, New Zealand University Press.

MOGENS JUL. Dairy development in India. Wld Anim. Rev., 24:2–8.

O'KEEFFE, A. & PHELAN, J.A. 1979. Cheese varieties. Dublin, An Foras Taluntais. Handbook Series No. 14.

O'MAHONY, F. & BEKELE, E. 1985a. Traditional butter making in Ethiopia and possible improvements. Addis Ababa, ILCA. ILCA Bulletin No. 22.

O'MAHONY, F. & BEKELE, E. 1985b. Small-scale manufacture of Halloumi and Queso Blanco cheese. Appropr. Techn., 12(3).

O'MAHONY, F. 1986. Milk processing by Borana pastoralists. Mission report. ILCA, Addis Ababa.

O'SULLIVAN, A.C. 1973. Milk processing with emphasis on product processing interactions. Dublin, An Foras Taluntais.

ROBINSON, R. K. 1983. Dairy microbiology. Vol. 1. London, Applied Science Publishers.

SCOTT, R. 1981. Cheese making practice. London, Applied Science Publishers.

VON MASSOW, V. 1985. Dairy imports into sub-Saharan Africa and their policy implications. Addis Ababa, ILCA. ILCA Bulletin No. 21.

WATERS-BAYER, A. 1986. Dairy sub-sector of the agropastoral household economy. In Livestock systems research in Nigeria's subhumid zone. Proc. 2nd ILCA/NAPRI Symp., Kaduna, Nigeria.

WEBB, B.H. & JOHNSON, A.H.1983. Fundamentals of dairy chemistry. Westport, Connecticut, USA, AVI Publishing Co.

WEBB, B.H., JOHNSON, A.H. & ALFORD, J.A. 1983. Fundamentals of dairy chemistry. Westport, Connecticut, USA, AVI Publishing Co.

WHALEN, I. 1985. Butter and cheese household utilization and marketing in Debre Zeit. Unpub. summary draft, ILCA Highland Programme. Addis Ababa, ILCA.

• The designations employed and the presentation of material in this publication do not imply the expression of any opinion whatsoever on the part of the Food and Agriculture Organization of the United Nations concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.
• The views expressed in signed articles are those of the authors.
• Information from WORLD ANIMAL REVIEW, if not copyrighted, may be quoted provided reference is made to the source. A cutting of any reprinted material would be appreciated and should be sent to the Distribution and Sales Section of FAO.
• Subscription rate for one year: US$14.00. Rates can be paid in local currencies when orders are placed through the authorized sales agents listed on the back cover. Subscriptions and inquiries may also be addressed to: Distribution and Sales Section, FAO, Via delle Terme di Caracalla, 00100 Rome, Italy.