Competitive position of cassava

Contents - Previous - Next

Looking back on the many uses of cassava, it may be asked: what are the reasons for its rapid introduction and permanent use in some starch-using industries, while in others it has not gained a place of importance? One explanation is found in the unique properties of this kind of starch.

While full treatment of the colloidal behaviour of cassava is outside the scope of the present paper, the more essential of these properties may be summarized as follows. The product is readily gelatinized by cooking with water, and the solution after cooling remains comparatively fluid; jellies or puddings cannot be prepared with it. The solutions, moreover, are relatively stable in that they do not readily separate again into an insoluble form, as is the case with corn (maize) starch and potato starch ("retrogradation").

Various factors of lesser importance have also influenced the position of cassava on the market, mainly in a negative sense. Being produced mostly in developing regions with an unstable economic position, it is not available with the same regularity and predictability as, for instance, corn (maize) starch; it is available in many grades and qualities, which are highly variable, and its price, especially in recent years, fluctuates considerably.

It is possible to distinguish three sectors of the starch-using industries, in each of which cassava occupies a fundamentally different position:

(a) Where irreplaceable by other starches.

In the manufacture of remoistening gums cassava has no competitors for the time being. Attention should be called, however, to the continuous efforts to adapt other starches to the special demands of these industries, both by chemical means as well as by the selection of starch-bearing plants. Mention should be made of the so-called waxy-maize starch, which approaches cassava in many respects.

(b) Where other starches are preferable despite the cost factor.

Some of the more desirable characteristics of other starches may be the result of further processing, as, for instance, in the corn (maize) starch industry. Examples are the thin-boiling, chlorinated and other special starches. Cassava furnishes only a crude starch with a wide range of quality and characteristics.

(c) Where interchangeable with other starches.

In this case, price and marketing conditions are the only controlling factors. Because of severe competition from other kinds of starch, in this field cassava has lost much ground of late.

It can safely be concluded that a market for cassava of all grades will be found for many years to come; however, the possibility of an expansion of its use will depend to a great extent on improved techniques in processing as well as on more efficient methods of marketing the flour.

7. Quality control of cassava products

In the processing of cassava, questions naturally arise regarding efficiency and output; moreover, in selling the product the determination of quality becomes important. These problems can only be resolved by qualitative and quantitative study of the composition of the raw materials and the properties of the finished products. The financial return, especially in large factories, will depend to a certain degree on such control analyses, which in a way are actually part of the processing itself.

Analysis of basic materials

The two important basic materials requiring analysis are the cassava roots and the water used in processing. The best practical qualitative test of these materials consists in reproducing the whole process on a small scale and judging the resulting flour by comparing it with a standard sample or by analysing it according to the methods described farther on. In fact, for judging the suitability of the water available, small-scale processing is the only test of practical value.

Apart from this, since starch is the substance to be isolated, a determination of the starch content in both the fresh roots and the pulp remaining after rasping and wet-screening is necessary for control of the efficiency of the process and in particular for determination of the rasping effect.

Finally, tests for the presence of hydrocyanic acid are necessary owing to the important food uses of cassava.


A random sample of, say, 10 kg of cassava is thoroughly washed to remove the cork layer; then either the whole or the peeled roots are grated or ground. The pulp is washed out over 50-mesh bronze gauze and the flour milk obtained over 260-mesh gauze. When the suspension reaches 3 Brix

(approximately 35 g of dry starch per litre), it is left to settle for four hours. The top liquid is then decanted, and water is added to the settled starch to make a slurry of 10 Brix, which is strained over 260-mesh gauze and left to settle for the second time. After decanting, the starch is mixed with water to a thick suspension (45 Brix) and filtered on a Buchner funnel under vacuum. The moist starch is dried in an oven, preferably in circulating air, commencing at a temperature of 50C, and concluding at 60C. The dried starch is sieved through silk before examination.


This analysis is best carried out with oven-dried pulp, a separate sample of fresh, moist pulp being used to determine moisture or water content, or with a sample of the fresh root material.

1. Quantitative determination of starch content is based on hydrolysis with acid and measurement of the resulting glucose. The weighed sample representing about 2.5 g of dried materials is ground and stirred with 250 ml of water for an hour, after which the insoluble residue is transferred to a vessel along with an additional 250 ml of water. After adding 200 ml of 0.5 NHCI, the solution is boiled under reflux and cooled; then the acidity is adjusted to pH 5 with NaOH, and when the solution reaches 250 ml, it is filtered. The glucose equivalent is determined by an aliquot according to the Munson and Walker method or any other suitable method (as described, for instance, in Methods of analysis of the Association of Official Agricultural Chemists, 1956). The amount of glucose multiplied by 0.93 is taken to equal the amount of starch which was in the aliquot.

2. A short-cut in the analysis of fresh roots is possible by determining the water content rather than the starch content. The method is practical especially where the variety of cassava and the growing conditions may be considered practically constant. The following empirical relation has been established as the result of a series of analyses of roots of four different strains of cassava, all grown on the same soil and during the same period.

Whole root:

Percent starch = percent total dry matter -7.3 = 92.7 - percent water

Peeled root:

Percent starch = percent total dry matter - 6.8 = 93.2 - percent water

Although the constants occurring in the above equations will have to be redetermined for each new set of circumstances, the application of these rules may be of great help in factories lacking chemical personnel and equipment.

3. A simple and inexpensive method for the quantitative determination of starch in cassava roots has been described by Krochmal and Kilbride. During peak seasons cassava tubers are kept in polyethylene bags and stored in a deep-freeze. The frozen samples are sliced and blended with 500 ml of water for five minutes in a blender. The pulp is washed on a sieve with an additional 500 ml of water, and the fibrous material retained on the sieve is thrown away. The washed material is poured into aluminium pans and dried at about 85C for 6-12 hours until a constant weight is attained. The weight of the residue represents the percentage of starch calculated from the weight of the sample.

Rasping effect

The fraction of the starch in the roots which is set free by rasping may be evaluated directly by washing out a weighed sample of the pulp, as obtained from the factory rasper, on a 260-mesh sieve, collecting the starch on a filter, and weighing it after thorough drying. The percentage of free starch thus obtained divided by the total starch content of the pulp gives the rasping effect (R).

The following method for the determination of R may be preferable in certain respects as it obviates the rather difficult direct determination of the free starch by using only the analysis for starch and fibre content of the roots and the waste pulp produced in processing them.

The roots are analysed for starch using one of the methods described previously and for fibre (cellulose). A sample of the waste pulp obtained in processing the roots is dried in an oven, milled and analysed for the same components. Since the waste pulp consists substantially of water, cellulose and starch, the percentage of starch can be deduced from the moisture and fibre content. The same holds for the roots, provided the freshly ground mass is washed out in a filter before analysing it.

If the starch contents of the roots and the waste pulp are sr and sw respectively and the corresponding fibre contents are fr and fw. it is readily seen that the fraction of the starch which remains bound to the fibre (i.e., occluded in cells which have escaped crushing by the rasper) amounts to


and thus the rasping effect is

R = (1- (sf fr / srfw)) X 100 percent


Qualitative test (Quignard's test)

Prepare sodium picrate paper by dipping strips into I percent picric acid solution and then, after drying, dipping them into 10 percent sodium carbonate solution, thereafter drying them again. Preserve these strips of paper in a stoppered bottle. Chop finely a small amount of the roots to be tested and put the choppings into a test tube. Insert a piece of moist sodium picrate paper, taking care that it does not come into contact with the root pulp. Add a few drops of chloroform and stopper the tube tightly. The sodium picrate paper gradually turns orange if the root material releases hydrocyanic acid. The test is a delicate one, and the rapidity of the colour change depends on the quantity of free hydrocyanic acid present.

Quantitative determination (alkaline titration method)

Put 10 to 20 g of the crushed root material in a distillation flask, add about 200 ml of water and allow to stand two to four hours, in order to set free all the bound hydrocyanic acid, meanwhile keeping the flask connected with an apparatus for distillation. Distil with steam and collect 150-200 ml of distillate in a solution of 0.5 g of sodium hydroxide in 20 ml of water. To 100 ml of distillate (it is preferable to dilute to a volume of 250 ml and titrate an aliquot of 100 ml) add 8 ml of 5 percent potassium iodide solution and titrate with 0.02 N silver nitrate (I ml of 0.02 N silver nitrate corresponds to 1.08 mg of hydrocyanic acid) using a microburet. The end point is indicated by a faint but permanent turbidity which may be easily recognized, especially against a black background.

Criteria for quality of flour and starch

For a product like starch, which is used as a basic material in many quite different branches of industry, the value of a certain brand greatly depends on the purpose for which it is intended. Quality in cassava can therefore only be defined with reference to the end use. In each industry using the starch, the starting material will be subjected to certain tests in order to determine whether it is suitable for the process concerned. In some cases, a mere superficial test of purity will suffice (when, for instance, it is used as a filler); in others, more elaborate determinations will be necessary. The value of the product in question will thus vary from case to case, and quality as well as price will emerge as a result of these investigations.

In general it can be said that the more careful and clean the manufacture of cassava flour, the higher will be its value for most purposes and thus its quality. There are, however, some exceptions to this rule, as where a flour of medium purity is preferred to one of prime quality. The medium- and off-quality flours have a market comparable in volume to that of the prime-quality flours.

The analysis of cassava flour consists of a group of selected tests, which together provide the best possible general insight into the usefulness of the material. The analysis comprises chemical determinations, such as those of water, pulp and ash, as well as physiochemical tests for the measurement of viscosity and acidity. On the basis of the results of these tests, quality is usually designated in the form of a grade, that is, a cipher expressing the quality in general or, more specifically, in relation to a certain property. The letters A, B and C thus often denote first, medium and poor quality, each classification being bound to specified limits of the properties investigated.

Cassava producers and industrial users have long understood that a universally accepted system of specifications with concomitant grading based on the results of a number of accurately defined tests would do much to stabilize marketing. The first attempt to draw up such a system of specifications was made in Indonesia between 1930 and 1940; it was based upon a series of qualitative and semiquantitative tests which had long been used in commercial circles. Certificates stating the results of analysis of the flour according to this system, notwithstanding its shortcomings, gave buyers in most countries an adequate idea of the quality of the product.

Technical developments in the countries where most of the end-use industries are centred, particularly the United States, led to much more specific and stringent demands regarding certain properties of the flour. Existing specifications soon ceased to cover satisfactorily the relevant features of the various brands of flour. To correct the situation and to adapt the grading of the product more closely to its various applications, a new system of specifications and grading was drawn up in 1943 by the Tapioca Institute of America (TIA) in close cooperation with most end-use industries. This widely adopted system is given in full in Appendix 1.

The tests to determine the quality of cassava starch include those for mesh size, colour, odour, cleanliness, pulp or fibre content, moisture content, ash, acidity and viscosity of cold flour slurry as well as cooked starch paste. All these tests help establish the grade and therefore the commercial value of the product. The discussion of these quality features will bring out more clearly the significance of the properties involved. For several of these properties alternative methods of determination sometimes give valuable complementary information on the quality of the flour.


This test measures efficiency of bolting. Fine pulp, however, obtained from bolting with disintegrators will pass the screens. In judging the purity of a flour, neither this test nor the determination of pulp under the TIA system is sufficiently precise; it is supplemented in a way by the cleanliness test, and an additional determination of pulp content by hydrolysis would seem advisable.


In this test the brightness or whiteness of the flour is visually compared with that of a "standard," which is a prime-grade flour produced by certain firstclass factories. The result cannot be clearly expressed by enumeration. Besides, the difficulty of procuring flour of standard whiteness which will keep for a sufficiently long time is an important drawback.

With the many excellent modern types of spectrophotometric apparatus available, however, it would not be hard to devise an objective and accurate quantitative expression for the whiteness of the flour as measured by its reflectivity relative to that of a sufficiently durable standard of whiteness (e.g., barium sulfate). In fact, this method has been adopted in Indonesia. Remarkably enough, comparative experiments in which the same flours were judged by the direct visual method and by the objective reflectivity method have shown that the former test is about as accurate as the latter, provided the observer has enough experience with the visual method and has accustomed himself to the standard of whiteness adopted.

As an independent measure of the purity of the flour, dry appearance has lost much of its significance, but it is still a commercial criterion. It is customary in commerce to estimate the number of specks occurring on a flattened surface of the dry flour as an indication of clean processing.


The following is a somewhat elaborate form of the test for specks which indicates the total amount of foreign particles in a sample. Five millilitres of distilled water are added to I g of dried starch. The mixture is stirred, and then 5 ml of 0.7 N sodium hydroxide solution are added and the uniform gelatinized mixture is examined for impurities. The degree of whiteness and clearness depends on the quantity of pigment, dirt and protein present in the starch.


The amount of pulp, or fibre, present in the finished product is of foremost importance in deciding the usefulness of a flour in various applications. The presence of insoluble cellulose is a serious hindrance in almost any industry where solutions of gelatinized starch are needed. Exceptions are the manufacture of corrugated cardboard and of plywood where the fibre is useful to a certain extent as a binder.

In the form given in the specification system under consideration, the test makes possible the determination of small amounts of fibre with comparative ease. The sediment volume measured is somewhat dependent on the fineness of the fibre. The presence of a slight trace of fibre, pulp or other impurity can be detected by microscopic examination of the size and shape of the starch granules.

The actual amount of cellulosic fibre in the flour and of foreign insoluble material can be determined by weighing the residue after a mild acid hydrolysis of the sample. Two to three grams of flour are boiled with 100 mm of 0.4 percent hydrochloric acid for one hour. The liquid is filtered through a weighed filter crucible fitted with filter paper or through a Jena glass filter, G 3. After washing with hot water the crucible is dried at 105" to 110C to a constant weight. One hundred times the gain in weight of the crucible divided by the weight of the test portion is the percentage of fibre and impurities.

It has been found that the determination of fibre content with these two methods runs parallel to a certain extent - that is, a 0.6 percent fibre content by the hydrolysis method corresponds to 10 ml of pulp per 50 g of flour by wet-screening.

A rough estimate of the amount of pulp may be based on the "crunch" of the flour - that is, the sound emitted when a sample, packed tightly in a small bag, is pinched between the fingers. "Crunch" is strong in pure flours, but above a certain pulp content it is lost.


Raw starch suspended in water gives rise to more or less viscous slurries. While in some applications the viscosity of these suspensions may be of some technical importance, the term "viscosity of flour" is generally used for the viscosity of a solution of flour after gelatinization, because it is in that form that flour is used in most industries.

Numerous methods are used to determine this property. They differ in the instrument applied in the actual measurement of the rate of flow of starch solutions, and in the method of preparing the starch solutions to be tested. As the comparative test in the present specification system is rather subjective, a few quantitative determinations which have found wide application in the starch industry (particularly cassava) are described below.

Hot-paste viscosity

This is the viscosity measured after the gelatinization of a sample in hot water. The instrument used is an Engler viscometer, which must be operated accurately, for variations in the preliminary treatment of the paste have an appreciable effect on the result. The Engler viscometer consists of two concentric cylindrical copper vessels. The outer vessel (A) filled with water serves as a heating bath. The inner vessel (C) is gilded inside and marked with three indications (a) on a plane perpendicular to the axis of the instrument. By adjusting screws on the legs of the instrument, the surface of a liquid in the inner vessel can be made to touch these marks at the same time, in this way ensuring an exactly vertical position of the instrument and a controlled level of liquid in the vessel. The outer bath is heated by a ring burner. The vessel (C) can be closed by a lid (D) in which a thermometer (Kl) can be fitted; another thermometer (K2) is placed in the bath. Before measurement the peg (h) is placed in the upper opening of the orifice (b), the actual measuring capillary of very precise dimensions. The liquid to be measured is then poured into the interior vessel to the level indicated by the marks; the lid is put in place, and after a few minutes needed for equalization of the temperature with that of the bath, the peg is removed, and at the same time a stopwatch is started. The flask placed below the orifice is calibrated at 100 and 200 ml; the measurement may be stopped at one of these marks.

FIGURE 39. The Engler viscometer

To determine the viscosity of starch, gelatinization is performed by heating a slurry containing a definite amount of the flour (usually 6 g) in 300 ml of distilled water at a definite rate with constant and not too vigorous stirring. The temperature should reach 100C in six to seven minutes. The solution is then transferred immediately into the inner vessel of the viscometer in the way described above. the water in the outer vessel being kept boiling during the whole determination. The flow of the solution into the flask is started when the temperature has reached 94C during the measurement it usually rises to 96C, a mean temperature of about 95C prevailing during this period. The time needed for 200 ml (or another definite quantity of the solution) to flow out divided by the time needed for the same amount of water at 20C gives the hot-paste viscosity in degrees Engler. In comparing different samples, the same amount of flour (dry weight) must be

AIkaline viscosity

This is the viscosity of a solution of the flour in a dilute solution of alkali. As a practical test it has the great advantage of being carried out at room temperature.

Detailed instructions are as follows. A sample of dry flour weighing 3 g is suspended in 30 ml of water placed in a beaker 600 ml in volume and 7.5 cm in diameter. The suspension is agitated with a mechanically operated stirrer made from a glass rod which is bent into a zigzag at the lower end, giving a stirring surface of approximately 4.5 cm breadth and 7 cm height. The temperature is adjusted to 27.5C and 270 ml of a I percent sodium hydroxide solution is added. Stirring at 200 revolutions per minute is continued for three minutes from the moment the sodium hydroxide is poured in; the mixture is then left to stand in a bath of 27.5C for 27 minutes; toward the end of this time it is carefully poured into the Engler viscometer, and the flow of the solution is started after the 27 minutes have elapsed. Thus, the sodium hydroxide is allowed to act upon the flour for exactly 30 minutes before the measurement of viscosity is started. The quotient of the time of flow of the solution (at 27.5C) and that of water (at 20C) is the viscosity (with sodium hydroxide) of the flour in degrees Engler. The results of alkaline and hot-paste viscosity generally run parallel.

The disadvantage of both the quantitative methods described above is that they assess the viscosity of the paste only at a certain point, whereas for many applications it is important to know the variation in viscosity under specific circumstances over a definite period of time. This objection is overcome with the modern recording viscometers, which allow the viscosity values to be followed during a process using hot paste from room temperature up to about 100C at a rate of heating which is automatically controlled; the viscosity can also be recorded at any desired temperature for any additional length of time.

Recording viscomerers

Most models are built on the same principle and some of them are available commercially. They generally consist of a noncorrosive metal cup fitted on an axis coupled to a strong electric motor, giving the cup a slow rotation at a constant speed. The actual measuring instrument, in the form of a rod with side arms, is fixed on a shaft aligned with the axis of the cup and is suspended in the liquid to be tested. Heating is furnished by the radiation from coils placed in the housing around the cup.

For the measurement of viscosity the cup is filled with a slurry of definite composition at room temperature (25C), and the heating and rotation are switched on simultaneously. The rate of heating is generally controlled at 1.5C per minute. The drag exerted by the contents of the cup on the measuring rod is balanced by a calibrated spring or a counterweight, and the deviation of the rod from its free position is transmitted to a lever with pencil affixed which records the flow resistance on a chart on a revolving drum. After reaching 90C the heating is automatically switched over to constant temperature. The possibility of recording viscosity at falling temperatures is provided by water-fed cooling coils.

The form of the viscogram obtained is more or less characteristic for the starch investigated; in general, the viscosity attains a peak value and then slowly falls off. The maximum value and the other features of the curve serve as a basis for grading. Since the dimensions of the curve depend on the construction details of the apparatus, the viscometer has to be calibrated with standard flours of known quality.


The amount of inorganic constituents present, as measured by the ash content, can be considered an indication of clean processing and, in conjunction with the acid factor (see below), conveys an impression of the quantity of metal ions bound to the raw starch. The colour of the ash is also of interest, as an off-colour indicates the presence of objectionable elements (e.g., brown-red from iron).


Determination of moisture by the oven method, though simple to perform, requires many weighings and is otherwise rather lengthy. Moreover, the results tend to be low in a highly humid surrounding atmosphere such as is likely to occur in tropical regions.

A more rapid method, free from these objections, consists in boiling a sample with xylene (boiling point 135C) and collecting the water driven out in the form of vapour, which separates from the xylene after condensation in a graduated tube.

In the oven method, the starch sample is placed at a depth of less than 2 mm from the bottom in a receptacle with a tight-fitting cover and weighed. With the cover removed' it is heated in a vacuum oven at 105C under a pressure of about 25 mm until a constant weight is obtained. The loss in weight during the heating period is considered equal to the moisture content.

The moisture contents specified for the A, B and C grades in the present system are difficult to obtain in most of the producing areas. The starch leaves the medium-size factories and the bolting installations with a moisture content generally over IS percent, but during shipping moisture declines a few percent. If the moisture content is high, there is a likelihood of mould growth.


This test is of considerable interest to manufacturers of dextrin from cassava. The titration required in the determination of the acid factor is a measure of the acid-binding capacity of the flour. As most dextrins are prepared with the addition of acid, it is understandable that the acid factor should be of prime importance in the conversion of starch to dextrin. It has been found that a smooth course of dextrinization requires an acid factor of between 2.0 and 2.5.

The initial pH, apart from its role as a starting point in the determination of the acid factor' may be used as an indication of the presence of moulds or other impurities in the flour: low pH values indicate deterioration. The amount of acid present is often determined by separate titration of a sample suspended in alcohol with dilute alkali solutions.


Nitrogenous bodies are usually determined by the Kjeldahl method. A sample of about 3-5 g is digested in a Kjeldahl flask by boiling it in 25 ml of concentrated sulfuric acid and about 0.2 g of copper sulfate as catalyst. Boiling is continued for 30 minutes after clarification of the reaction. The flask is cooled, and the contents are diluted in 200 ml of water, with a few pieces of granulated zinc and 25 ml of 4 percent sodium sulfide solution added. About 50 ml of 45 percent sodium hydroxide solution is cautiously run down the side of the flask, which is then connected to a condenser. The outlet of the condenser should extend into a vessel containing a measured quantity of standard acid solution. The flask is gently shaken to mix the contents and heat is applied to distil the ammonia from the reaction flask into the acid solution.

The excess acid in the receiver is titrated with standard sodium hydroxide solution, using methyl red as an indicator. After a determination of the amount of ammonia distilled and calculation of the results as nitrogen, it is common practice to report values as percent protein by multiplying the percentage of nitrogen by the factor 6.25.

Analysis of baked products

These partly gelatinized products are always consumed in cooked form - in soups, puddings and so on. Consequently, their behaviour in contact with cold water and during gelatinization in boiling water is the best criterion of quality. The tests and specifications given below are the result of long experience in the trade of these commodities. They are particularly directed to the swelling capacity and stability of the structure of the pearls in cold water and during cooking. A determination of the titratable acidity is added as a test of durability.


Ten grams of the material are placed in a graduated 100-ml cylinder and the bulk volume is noted. The cylinder is then filled with water, and after standing 24 hours the volume of the swollen beads is determined. A high swelling power is preferable, though definite values are not given.


Three and a half grams of the material are added to 200 ml of boiling water, and boiling is continued for two minutes. After cooling for 30 minutes, the contents of the beaker are heated again to boiling point, with a low flame, and boiling is continued for another 30 seconds. Then the liquid is left to cool for half an hour and transferred to a graduated cylinder. After settling, the total volume of the swollen product is read. Generally, part of the original beads will have disintegrated; the volume of these fragments which form a layer of "slime" on top of the swollen whole beads is noted separately. The market demands rapid swelling with a minimum of disintegration during cooking, and a large volume after the second cooking. Normally' the volume of the whole beads is more than 8 ml per gram of starting material and the slime layer less than 4 ml per gram.


Five grams of the material, obtained by grinding the product in a disintegrator with a 150-mesh sieve, are mixed with 100 ml of alcohol which has been neutralized, using phenolphthalein as an indicator. The mixture is left to stand, with occasional stirring, for 24 hours. It is then poured on a dry filter and 50 ml of filtrate are titrated with N/10 sodium hydroxide, I ml of the reagent corresponding to an acidity of 0.24 percent calculated as acetic acid. Normally, less than 0.05 percent is found.

Specifications for particular uses

Specifications of starch are being developed rapidly and analysis of commercial starches is becoming more and more necessary for both the producer and the consumer. The general specification system just reviewed does not cover all the properties which may be of interest for particular end uses. An instructive example is to be found in the specifications for dextrin from cassava as used in remoistening gum (Appendix 2). Though much depends on the performance of the dextrinization, it may be assumed that only certain varieties of the basic material will meet these requirements. Cassava dextrin is preferred in remoistening gums for stamps, envelope flaps and so on because of its adhesive properties and its agreeable taste and odour.

Requirements surpassing by far the specifications system of the Tapioca Institute of America are found also in the food industry.

Standards, grades and methods of analysis for starch and starch products have been established during the last two decades by the International Organization for Standardization. Some leading food corporations in the United States have also developed their own specifications in order to bring about some uniformity of quality in cassava products (Appendix 3).

Contents - Previous - Next