In traditional pastoral systems, camel milk is mainly used for feeding calves and for human consumption. Two quarters of the udder are usually selected for milking and segregated with ropes, while the calf suckles the other two quarters (Ramet, 1987; Ramet, 1989a; Ramet, 1994a).
Milk for human consumption is usually drunk immediately after milking. It can also be consumed as fermented milk made by natural lactic souring over several hours in a skin or clay container. The fermented milk may sometimes be separated by vigorous shaking; the acid skimmed milk is drunk and the butter used for cooking or cosmetic or medicinal purposes (Yagil, 1982).
The processing of camel milk into cheese is said to be difficult, even impossible (Dickson, 1951; Gast, Maubois and Adda,1969; Yagil, 1982; Wilson, 1984). It is surprising that although the majority of pastoral systems have produced at least one type of cheese, no traditional methods exist for making cheese from camel milk. This might be explained by local cultures which allow the consumption of camel milk only as drink and exclude the possibility of trade. It is also possible that the highly perishable nature of cheese in hot desert climates has not been conducive to creating trade between isolated communities.
In addition to these cultural considerations, it appears that camel milk is technically more difficult to process than milk from other domestic dairy animals. A bibliographic review indicates that in the Ahaggar region and the Sinai peninsula only a few rare cheeses are manufactured by acidic separation and heating of milk proteins (Gast, Maubois and Adda, 1969; Yagil, 1982). These products seem to have the characteristics of perishable fresh cheese with a high moisture content. Shelf-life may be increased to several months by air and sun drying (Abeiderrahmane, 1994). It is noted that these cheese types do not come under the standard definition of cheese which results from the simultaneous action of a milk clotting enzyme and lactic souring (Ramet, 1985b).
Action of clotting enzymes on camel milk.Most attempts to make cheese from camel milk have revealed major difficulties in getting the milk to coagulate. Initial field attempts increased the rennet concentration compared with that usually used for clotting cow's milk by 50 to 100 times (Gast, Maubois and Adda, 1969; Wilson, 1984). More recent attempts confirm that the rennet coagulation of camel milk is two to four times slower than for cow's milk treated under the same conditions (Ramet, 1985a; Farah and Bachmann, 1987; Ramet, 1987; Mohamed and Larsson-Raznikiewicz, 1990).
This specific behaviour has been observed with most of the clotting enzymes used for coagulation. Significant differences in the inhibition of clotting activity related to the origin of the enzyme have been noted, however. Several observations (Ramet, 1985a; Ramet, 1990) have shown that bovine pepsin coagulates camel milk well. Calf rennet and the clotting enzyme extracted from Mucor mieheihave an effect similar to but lower than bovine pepsin. Chymosins of genetic origin and proteases of Endothia parasitica ave the lowest effect (Figure 2).
Milk clotting trials made under similar conditions, using either milk reconstituted from low-heat powdered milk (pH 6.65) or fresh raw camel milk (pH 6.55), have demonstrated a noticeable improvement in clotting camel milk compared with cow's milk when calf rennet, Mucor mieheiand Endothia parasiticaproteases and genetic chymosin were used. With bovine pepsin, the clotting time was five times shorter in camel milk (Table 10). This unique behaviour of pepsin could be explained by its higher affinity for camel milk and its limited activity at a near-neutral pH.
These different affinities, which depend on enzyme source, could be partly explained by the incidence of environmental factors (pH, temperature, ionic strength, etc.) regulating enzyme activity. The main origin of the disparity in the clotting effect of the different milk-clotting enzymes could more probably be the presence in camel milk of specific protease inhibitors and/or a particular casein micelle structure limiting access of the protease to the kappa casein substrate. These hypotheses are yet to be confirmed.
More generally, it must be stated that some nomads in the Sahara and Sinai seem to be able to make cheese using parts of the stomach of the desert rabbit as a coagulating agent (Gast, Maubois and Adda, 1969; Yagil, 1982). This stomach contains pepsin (Lebas, 1991). More recent work carried out in Egypt (El-Abassy, 1987; El-Batawy, Amer and Ibrahim, 1987) shows that the pepsin produced from the stomach of the adult camel is just as good in terms of activity and stability. However, this work did not deal with the general ability of the enzyme for cheese making. It appears that using the stomach of young camels for making camel cheese has never been investigated or tried, which is strange. Moreover, no work has mentioned the actual enzyme composition of the camel calf stomach.
A further distinguishing feature of the Mucor miehei nzyme when used in weak concentration in camel milk is to cause partial inhibition, as shown by the non-linear relationship between the clotting time and the inverse of enzyme concentration (Figure 2). This phenomenon, previously observed in raw cow's milk, probably originates from the enzyme reacting with the whey proteins. As a consequence, in practice the enzyme quantity has to be slightly increased. The inhibitory effect disappears when the milk is heat treated under high time-temperature pasteurization conditions (Ramet, 1985a).
Curd formation and rheological properties. any observations of making cheese from camel milk point to the difficulty in measuring the early stages of coagulation. An empirical appraisal of the physical properties of milk during the liquid to gel transition phase is not easy because of the persistence of a diffuse, curd-like pseudo-gel. Further build up of the coagulum is slow and weak (Ramet, 1985a; Farah and Bachmann, 1977; Ramet, 1991; Ramet, 1994a). The gel texture is characterized by low elasticity and high fragility. Moreover, the fragility of the curd is increased where acid fermentation occurs (Ramet, 1987; Ramet, 1994a). On a practical note, this rheological development indicates the need to increase the speed of coagulation in order to avoid making the curd too weak to withstand the mechanical action used in draining.
This unique rheological behaviour has been traced by empirical methods (Gast, Maubois and Adda, 1969; Ramet, 1985a; Ramet, 1987; Mohamed and Larsson-Raznikiewicz, 1990; Ramet, 1994a) and confirmed and quantified by instrumental methods. Figures 3 and 4 show examples of measurements made by gelograph and turbidimeter (Farah and Bachmann, 1987; Ramet, 1990; Bayoumi, 1990).
The relationship between milk composition and clotting ability.Influence of casein composition. he limited ability of camel milk to be coagulated by enzymes is probably largely due to the composition of the casein micelles. Some recent research has shown that the kappa casein, representing the micellar fraction which reacts with the clotting enzymes, has a different electro-potential from cow's milk, which causes lower electrophoretic mobility (Farah and Farah-Riesen, 1985; Jardali, 1988; Mohamed and Larsson-Raznikiewicz, 1990; Farah, 1993; Larsson-Raznikiewicz, 1994.)
This unusual behaviour indicates a very specific casein micelle composition characterized by a low proportion of kappa casein. Relevant data are listed in Table 3, which indicates that the average content of kappa casein in camel milk from various sources rises to only about 5 percent of total casein, compared with 13.6 percent in cow's milk (Jardali, 1994). Camel milk casein also differs in terms of micellar size (Table 4). Instrument measurements showed that the mean diameter ranges from 280 to 325 micrometres, about double the 160 micrometres in cow's milk (Farah and Bachmann, 1987; Jardali, 1988; Farah and Ruegg, 1989; Jardali and Ramet, 1991).
It is important to emphasize that seasonal variations in the composition and size of casein micelles have also been found in cow's milk. These are caused by the variable effect of environmental factors such as temperature and feed availability. For example, a noticeable variation in the diameter of the micelles from 150 to 250 nm was observed in bulk milk collected in the eastern part of France. During the hot season, the micelles are larger and lower in kappa casein. The same milk had a reduced ability to coagulate compared with winter milk. The clotting time with rennet was longer and the firmness of the curd significantly reduced. In the cold season, on the other hand, the micelles were richer in kappa casein, coagulated faster and produced stronger curd (Ekstrand, Larsson-Raznikiewicz and Perlmann, 1980; Niki and Arima, 1984; Scher, 1988).
Adding rennet to camel milk causes a proteolytic reaction which can be traced through the development of the quantity of non-protein nitrogen. The trend of the curves shows that hydrolysis is similar in camel milk and cow's milk, although the percentage of kappa casein is quite different (Farah and Bachmann, 1987; Mehaia, 1987).
It appears that the secondary reaction of the clotting process in camel milk, which corresponds to the aggregation of casein micelles, occurs in a definite sequence. It has been observed by electronic microscopy that in cow's milk a homogenous network of micelles existed after a time corresponding to 80 percent of the visual clotting time. In camel milk, the aggregation of micelles occurs later and the network is softer and less dense (Farah and Bachmann, 1987). It seems that the reduced ability of micelles to polymerize is the result of the weak capacity of the substrate to link calcium bonds with particles. The large-sized micelles are known to be lower in calcium than the smaller ones (Scher, 1988). Measurements made elsewhere during the hot season showed that the content of colloidal calcium bound to the micelles in camel milk was much lower (35 percent of the total calcium) than in cow's milk (65 percent) and that the total calcium content was also much reduced by water restriction (Yagil and Etzion, 1980; Yagil, 1994).
The major role of calcium in the coagulation process is corroborated by the fact that controlled enrichment of camel milk with ionic calcium drastically reduces clotting time and reinforces the gel strength more than in cow's milk under similar conditions (Ramet, 1985a; Ramet, 1987; Farah and Bachmann, 1987; Jardali, 1994; Ould Eleya and Ramet, 1994).
Influence of total solids.It is known that the rheological properties of curd also depend closely on the total solids in the milk and are improved as total solids are increased. The components of the dry matter behave differently during clot formation. The casein content has the major role: the higher it is, the stronger the formation of the micelle network. Fat is not active in gel formation. Fat globules are caught in the casein matrix, where they decrease clot rigidity. At a similar fat percentage, the curd is much weaker in the presence of small fat globules than with larger ones. The soluble substances do not act directly on gel formation; they only modify the viscosity of the whey located in the interspaces of the curd.
Analyses show that the dry-matter content of camel milk varies according to the origin of the milk (Table 2). Similar variations exist in the fat and protein contents. Generally, however, the fact that the amounts of these components are lower than in cowmilk explains the lower rheological quality of camel milk curd. Such adverse effects occur most when animals have restricted access to water. It has been observed, for example, that total solids can fall from 14.3 to 8.8 percent, protein from 4.6 to 2.5 percent and fat from 1.3 to 1.1 percent (Yagil and Etzion, 1980; Yagil, 1994).
A third cause of weaker curd rigidity in camel milk is the small size of the fat globules, which are between 1.2 and 4.2 micrometres instead of the 1 to 10 micrometres in cow's milk (Dong Wei, 1980; Knoess et al. 1986; Farah, 1993).
Milk composition and acid fermentation ability.The acid coagulation of camel milk is governed by lactic acid bacteria which originate either from the raw milk or from the external inoculation of lactic starters (Ramet, 1985a). The ability of camel milk to acidify is, in turn, dependent on several compositional factors which interfere with bacterial growth.
Milk may be considered a medium favourable for microbial growth with a near neutral pH, a high water activity and a large variety of nutritive substances facilitating the proliferation of cells including lactic acid bacteria. Lactose is the nutrient of prime importance. Although its content in camel milk may vary greatly depending on feeding and watering conditions (Yagil and Etzion, 1980; Yagil, 1994), it appears that lactose availability is always satisfactory, even in cases of strong acidity. There are no studies of nitrogen nutrition that assess the precise requirements of lactic bacteria in relation to the specific composition of camel milk.
On the other hand, a bibliographic review indicates that raw camel milk contains several antimicrobial agents that can limit microbial growth to a higher degree than in milk from other domestic animals. Significantly high levels of lysozyme (Barbour et al, 1984; El Sayed et al, 1992; Farah, 1993) and vitamin C (Kon and Cowie, 1972; Knoess, 1979; Yagil, 1982; Yagil, Saran and Etzion, 1984) are reported. More recently, the antimicrobial activity of other natural proteins such as lactoferrin, lactoperoxydase and immunoglobulins was studied (Monnom et al. 1989; IDF, 1991; El Sayed et al. 1992; El Agamy,1994). Each of these antimicrobial agents possesses a selective spectrum of activity against specific strains of bacteria and viruses.
As a major consequence, when fresh raw milk is allowed to sour, a bacteriostatic period is observed for the first few hours after milking. This lag phase is greater in camel milk (four to six hours) than in cow's milk (two to three hours). Acid development rates are slower throughout the incubation period (Ramet, 1985b; Ramet, 1987; Gnan et al. 1994a). After camel milk has been heat-treated using thermizing or high pasteurization conditions, partial inhibition persists because the antimicrobial factors could be rather more heat-resistant than in cow's milk (Ramet, 1994a; El Agamy, 1994). Another reason for the reduced acid production rate appears to be related to the higher buffering capacity of camel milk compared with cow's milk (Rao, Gupta and Dastur, 1970; Ramet, 1985b; Ramet, 1987; Farah and Bachmann, 1987).
Formation and rheological properties of lactic gels. hroughout the course of acidification of cow's milk, progressive neutralization of the electric charges of the casein micelles occurs, leading to the emergence of the curd. The coagulation point occurs earlier when the acidity and temperature are high (Veisseyre, 1975; Ramet, 1985a). In camel milk, it is difficult to detect a similar development because the formation of the clot is slow and unstructured and resembles a flock rather than a precipitate (Ramet, 1985b; Ramet, 1987; Farah and Bachmann, 1987).
The ability of curds to drain is directly dependent on their rheological properties, which develop throughout the hardening phase, taking into account the development of firmness and elasticity.
The extreme weakness of camel milk curds causes the destruction of the casein network if physical treatment applied at cutting and moulding is not done carefully and slowly. If these conditions are not observed, a significant portion of the dry matter of the milk is not retained in the cheese but lost in the whey. Recovery is then limited to about 30 percent, whereas it increases to about 50 percent for cow's milk and 68 percent for sheep's milk under similar manufacturing conditions (Ramet, 1990).
The draining of curd made from camel milk is characterized by rapid syneresis compared with cow's milk. Figure 7 shows the large difference during wheying off measured in a curd obtained chiefly by acid coagulation (Ramet, 1987). This development appears to be a consequence of the low water retention capacity of the gel because of its rather limited casein content. Another factor is that the hydration of camel milk casein micelles is reduced by the low kappa fraction, which is very hydrophilic, and by the restricted surface area relative to its high volume (Jardali, 1988; Scher, 1988; Jardali, 1994). It should be noted that similar significant relationships have been observed when seasonal variations in the composition of cow casein micelles have been accurately measured. In the hot season, micelles are larger but lower in kappa casein and the resulting curd has a weaker water-retention capacity than in the cold season (Scher, 1988).
The fact that the acidification rate is slower in camel milk appears to have no adverse effect on wheying off. It must be stressed, however, that under these conditions the protective effect of acidity in preventing the spread of spoilage organisms is delayed. It is thus necessary to make the cheese under especially hygienic conditions.
The composition of camel milk whey is characterized by higher total solids than cow's milk, at 7.0 and 6.5 percent, respectively, whereas the dry-matter content is often lower in camel milk whey (Ramet, 1987; Ramet and Kamoun, 1988; Kamoun and Bergaoui, 1989; Ramet, 1994a). It has been emphasized that fat content is particularly high, reaching three to four times the value measured in whey from cheese made under similar conditions from cow's milk - 0.3 and 1.3 percent, respectively. This concentration is equivalent to more than 60 percent of the content of the milk (Ramet, 1989b; Mohamed, 1990; Ramet, 1994a). The small size of fat globules and the fragility of the casein micelle network are the cause of these losses.
The whey from camel milk cheese is identified by its white colour, compared with the greenish whey from cow's milk cheese (Ramet, 1989b; Ramet et Kamoun, 1988; Mohamed and Larsson-Raznikiewicz, 1990; Ramet, 1994a). This property of camel whey is probably the result of a concentration of small particles (proteins, fat globules) which, through complex diffraction and refraction phenomena, cause the white colour. Another reason could be the low concentration of riboflavin in camel milk (Webb, Johnson and Alford, 1974; Farah, 1993).
Little information is available on the ripening of cheese made from camel milk. What is available is based on experimental production carried out using small quantities of milk. The first commercial production started recently in Mauritania in a new, purpose-built camel cheese-making facility (Ramet, 1994b). Trends may thus be observed but final conclusions cannot yet be made.
Results from sources in Tunisia (Ramet, 1987), Saudi Arabia (Ramet, 1990) and Mauritania (Ramet, 1994b) show that the taste of fresh camel cheese is highly satisfactory. The smooth texture and sharp taste of the curd were well liked by a tasting panel. Similar results were observed by most of the panellists for soft cheese with a total solids content of 35 to 45 percent at the end of draining. However, some judges trained in the sensory evaluation of soft cheese made from cow's milk have noticed a rougher texture in camel cheese. The somewhat chalky structure is probably a result of the reduced fat content of the cheese because of high fat losses in the whey and the weak water-binding capacity of camel milk curd. The sensory profile of soft camel cheese is very similar to that of low-fat soft cheese made from cow's milk. A similar crumbly, granular texture has been also found in semi-hard and hard cheese (Ramet and Kamoun, 1988; Mohamed and Larsson-Raznikiewicz, 1990; Ramet, 1994a). The last observation confirms that the cheese becomes less smooth when the fat and water contents decrease.
Another defect is that greasy, sticky cheese curd has sometimes been noted. The cheese tends to adhere quite strongly to the tongue and palate while it is being chewed. No explanation of this is known but it seems that some properties of the camel cheese fat, such as the high level of short-chain fatty acids and their significantly high melting point (Abu-Leiha, 1987; Abu-Leiha, 1989; Farah and Ruegg, 1991; Mehaia, 1994) may be related to the phenomenon.
Temporary bitterness has been noted in some soft and semi-hard cheeses (Ramet, 1987; Ramet and Kamoun, 1988; Ramet, 1994a). The defect is detected mainly after the cheese has been swallowed. This perception of bitterness is delayed because the receptors sensitive to bitter molecules are located at the back of the tongue.
The probable origin of the bitterness in camel milk cheese has not been clearly determined. It is known that bitterness in dairy products may be caused by factors such as alkalis of ingested plants, salts of external origin - mainly calcium and magnesium chlorides - and carbonates or bitter peptides generated by casein hydrolysis. The most likely cause is those proteolytic residues which accumulate when the pH of the cheese is low and a high residual proteolytic activity from the clotting enzymes remains in the curd. The fact that it is necessary to overdose the clotting enzyme to speed up coagulation of camel milk indicates the last possibility as the origin of the bitterness.
Producing whey cheese by coagulating the soluble proteins in camel milk whey is more difficult than with cow's milk whey, at least when traditional methods are used. When camel milk whey is heated, aggregates of denatured proteins begin to form at temperatures between 72 and 80°C (Ramet, 1987; Mohamed and Larsson-Raznikiewicz, 1990; Ramet, 1994a). However, the particles remain very small and isolated and do not come together during further heating as in cow's milk. When left at ambient temperature for ten to 16 hours, three distinct phases occur: an upper floating layer composed of water, proteins and fat; an intermediate layer made up of clear whey; and a weak white precipitate at the bottom. The separation of the upper part by traditional simple filtration is ineffective. Alternatives to this process, such as acidification with lactic and citric acids, addition of calcium and sodium chlorides or addition of 30 percent of acid camel milk, do not improve collection of the particles (Ramet, 1987; Ramet, 1990). The only way to separate is to use a centrifuge, which allows a watery concentrate to be recovered with about 16 to 22 percent total solids (Ramet, 1990).
The unique behaviour of camel milk whey compared with cow's milk whey could be explained by differences in the composition of the soluble whey proteins and their higher heat stability (see The main constituents on p. 3). It has also been noted that when cow's milk is strongly heated, a reaction occurs between beta-lactoglobulin and kappa casein that makes the formation of large aggregates easier (Zittle et al. 1962). Absence in camel milk of a protein similar to beta-lactoglobulin and low kappa casein content could cause this different behaviour. Finally, it is possible that the high amount of fat in camel milk whey could have some adverse effect on the surface properties of the whey protein particles, leading to their dispersion.
Given the very high fat content of camel milk whey, the question arises whether it is possible to make butter from it. A review of the literature indicates that making butter from camel milk whey has been controversial for a long time. Many nomads do not produce butter from pure fresh camel milk (Dickson, 1951; Wilson, 1984), whereas some authors report that butter is produced under good management conditions (Yagil, 1982). Research has confirmed that butter making from camel milk whey is feasible but more difficult than with cow's milk whey (Farah, Streiff and Bachmann, 1989; Ramet, 1990).
The difficulties seem to stem from the properties of the fat globules, which are generally small with a thick membrane (see The main constituents on p. 3). For these reasons, the mechanical resistance of the fat globules is probably strengthened, which results in a long churning time of about five hours when milk is directly processed without prior centrifugal concentration of the fat (Ramet, 1990). If the agitation of whey is carried out after increasing the acidity to pH 5.0, churning time is reduced to one to two hours.
The concentration of the fat emulsion into cream by natural creaming or by centrifugation was found less easy than for cow's milk because of the small size of the globules. To obtain 20 to 30 percent cream fat, it is necessary to double centrifugation. This leads to a significant reduction in churning time, which promotes the occurrence of butter grains. The time falls to between five and 45 minutes, depending on temperature, fat content and cream acidity (Farah, Streiff and Bachmann, 1989; Ramet, 1990). Acidifying the cream makes churning faster but lowers fat recovery in butter (Figure 8).
A feature of the composition of camel milk fat is its low short-chain fatty acid content and high proportion of palmitic and stearic acids. This results in high melting and solidification points compared with cow's milk: 41.4 to 41.9°C and 30.5°C for camel milk and 28 to 32°C and 22.8°C for cow's milk. It was shown earlier (see Summary of butter-making technology on p. 8) that temperature is important to balance the physical state of the fat. The major role of temperature is confirmed by the fact that formation of butter grains does not occur at 10 to 12°C, which is the usual churning temperature for cow's milk cream, and that over 36°C the butter yield begins to fall. The best conditions for making butter are 25°C for a 22.5 percent fat cream with a churning time of 11 minutes (Farah, Streiff and Bachmann, 1989).
The sensory profile of butter made from camel milk is conditioned by its very white colour (Farah, Streiff and Bachmann, 1989; Ramet, 1990), which probably results from a high amount of non-fat components such as proteins linked to the fat globules, and considerable retention of buttermilk by capillary action (Ramet, 1990). The butter is greasy and sticky when eaten or cut with a knife (Farah, Streiff and Bachmann, 1989; Ramet, 1990). The flavour is neutral and unlike butter made from cow's milk.
The foregoing remarks on butter processing are only hypotheses as to the feasibility of making butter from camel milk whey. It is obvious that the fat in whey is more adulterated than in milk or cream as a consequence of the physical and chemical processes applied during the different stages of cheese making. It appears that churning times and fat losses in buttermilk are more important than for fresh cream. For the same reason, the taste and keeping qualities of camel whey butter would also be less satisfactory.
Trials have shown that camel milk whey may be used to make acidified drinks. These drinks have an excellent nutritive value because of the presence of essential amino acids, lactose, lactic acid, vitamins and minerals. The taste properties of whey are well known: it is sweet or slightly acid depending on the level of acidity. These dominant flavours can be masked if a milky taste is to be avoided by adding concentrated juices from acid fruits. Because of the opaque colour of whey and the possible presence of a whey protein precipitate, it is better to use cloudy juices that contain pulp, such as citrus fruits. The low pH of these juices gives the whey a characteristic refreshing taste; the additional acidity is protection against development of most spoilage organisms. Consumption of the product should be within two to three days. Additional preservation by pasteurization is necessary for longer-term storage.