Previous Page Table of Contents Next Page


L. Dempfle1


It is concluded that genetic improvement of trypanotolerant cattle is highly desirable. Human population in that area is growing rapidly, agriculture is intensified, crop production and animal production are becoming more integrated and there are no immediate prospects to use other but trypanotolerant breeds in a sustainable manner.

In a breeding scheme for trypanotolerant breeds there are four traits which are of larger importance: trypanotolerance, milk yield, meat yield and traction. Higher efficiencies in producing milk and meat are always desirable but it is of utmost importance to improve or at least not to deteriorate trypanotolerance.

Taking into account the husbandry system and the facilities available a breeding schemes is proposed which relies heavily on a station but at the same time uses a large fraction of the population as resource. In the program information on trypanotolerance, milk yield and meat yield will be collected on the station. The breeding work on the station follows conventional ways, but trying to reduce generation interval. If found appropriate modern reproductive techniques could be used but it is recommended to consider this aspect at a later stage. The work in the station should be augmented by an extensive screening operation. This screening operation enlarges the effective population size (breeding population), increases selection intensity and could prove very useful if there are major genes.

A continuous control of the efficiency of the scheme is absolutely necessary.

At the moment any proposal can only be preliminary (lack of good genetic parameters, lack of good economic parameters, difficulties in appreciating the feasibility of many measures), thus there is the need for regular readjustment of the plan.

1 Department of Genetics, Institut fur Tierwissenschaften Technische Universitat Munchen, D W-8050 Freinsing-Weihenstephan, Germany


There are a lot of good publications on trypanotolerant cattle (e.g. FAO Animal Production and Health Paper 20/1, 20/2, 20/3; Livestock Production in Tsetse affected areas in Africa, Proc. ILCA/ILRAD, Nairobi, 1987; Immunobiology of African Trypanosomiasis by Roelants and Pinder, 1984), so there is no need nor intention to try to add anything to this long list. However, for the specific topic to be dealt with, the following points are important:

2. Genetic improvement versus improvement of management (e.g. nutrition)

Often it is heard that genetic improvement of livestock is only worthwhile when the husbandry system - especially nutrition - is improved. This opinion might be correct under extremely harsh environments, where the animal needs all the resources in order to survive and reproduce at the minimal level. However as soon as conditions are somewhat improved both genetic and management improvements should go hand in hand. A genetically more efficient animal with respect to milk or meat production can respond much quicker and better to any improvement of the “environment”.

With respect to the economics one has to consider the following differences:

Investment in better feeding brings immediate returns but only as long as better feeding is going on and only with those cows where it is practised.

Investment in breeding earns return only after a considerable time lag (in cattle 5 years and more). However, genetic progress achieved in the nucleus (thousands of animals) can be transmitted with little additional cost to the whole population (millions of animals) and in addition, as long as the environment is not changing drastically and/or natural selection is not working against artificial selection, the genetic progress achieved stays permanent.

As an example if in the nucleus of 1000 animals milk yield is increased by 100 kg, that superiority of 100 kg can be transmitted (with due time lag) to the population of millions of cows and it is then expressed year after year and generation after generation.

Thus investment appraisals of improving the environment and in a very careful way, taking into account time horizon and multiplication effects.

Judged in this way, investments in breeding often turn out to be very economical.

However, as already stressed above, both genetic improvement and improvement of management should go hand in hand. This is especially important since there is ample evidence that animals in good conditions are more able to cope with trypanosomiasis, and, of course, also produce more. Here again, the economics has to be carefully observed in such a way that the options are sustainable in the “field” on a long term basis.

Nutrition in the breeding project (nucleus) is a different story and some remarks will be made later in the report.

3. Designing a breeding Scheme

3.1 Logical Approach

Designing and optimizing a breeding scheme comprises severalsteps which should be carried out sequentially.

Step A: Identify all traits of importance. Determine by how much the utility is changed if we change a trait by one unit. Combine the traits into an objective function, which determines essentially the breeding goal.

Step B: Make a statistical analysis of all traits of importance and determine variances and covariances, heritabilities and genetic and phenotypic correlations.

Step C: Formulate and optimise the breeding scheme taking into account the results of steps A and B, all the constraints and the cost of the breeding operation.

Step D: Begin Breeding!

Step E: Optimise the flow of genetic progress from the nucleus to the multiplier and to the production level.

Step F: Control the efficiency of the breeding scheme. There are two levels of control:

level 1: Determine whether the actions taken are in close agreement with the plan. If necessary re-adjust the operation or the plan.

level 2: Estimate the genetic progress and the time lag.

For the present project “genetic improvement of trypanotolerant cattle” the first four steps have to be done more or less at the same time making bold but realistic assumptions about economic and genetic parameters. These parameters have to be refined as soon as new data are allowing better estimates.

However, this stepwise approach is an excellent framework for the discussion of the various aspects.

3.2 Identification of important traits (Step A)


The main justification for the N'Dama breed (and other trypanotolerant breeds) is the trypanotolerance and perhaps the tolerance against other diseases. If we lose that characteristic we jeopardize any improvement we make in other traits.

An explicit selection for trypanotolerance is, in my opinion, needed. One could argue that animals producing a lot of milk and meat must be healthy and thus, by selecting for these traits, there would be an automatic selection against susceptible animals (correlated genetic response in trypanotolerance). This logic is correct given all animals are under the same challenge. However, in many areas the number of infective bites received daily by each animal is low (at Keneba/Gambia it is estimated to be 0.0041 implying an expected interval between infective bites of 244 days. SNOW. W.F. et al. Annex of “Entomology Programme” 4th Annual Report, 1989) even in areas judged of being of medium challenge.

Without direct selection there is also the danger that given some crosses (with Zebu or exotic breeds) mainly the “lucky” (noninfected) crossbreds are selected since they might produce better with respect to milk and meat on a per animal basis.

What is the appropriate unit of measurement of trypanotolerance and what is the economic value to attach to it? For the time being the exact definition of the trait (scale) has to be left open, the same applies to the economic coefficient, however, PCV (packed red cell volume percentage) seems to be the most appropriate trait.

3.2.2. MILK YIELD (Amount of fat and protein)

If we speak of milk it should always be understood to mean amount of fat and protein produced.

It was my firm impression by visiting the farms that milk is the most important item for the farmer. A similar view is expressed by Agyemang, 1990. The amount of milk offtake can be quite considerable as has been shown by Agyemang et al., 1991. They found a mean milk offtake of 404 kg in each lactation, but with values up to more than 1000 kg. Lactation length was on average 420 days so, roughly, one kg milk per day is taken off on average.

For the milk taken off, there is a market price of about 100 CFA/kg. In many situations half of the selling price is about the marginal profit. This however, is only a very rough estimate. The precise determination of the marginal profit of an additional unit requires a very detailed analysis of the whole production system. I am not aware that any such study is available. Taking half the selling prize is motivated by the figures used for a long time e.g. in Germany.

The milk consumed by the calf has to be estimated (milk consumed is (8 to 9) times body weight gain). For this milk the same weighing (or a little less) is recommended.


Incomes from selling slaughter animals is another important item. As definition of the trait, we could either use body weight at a fixed age (e.g. body weight at 3 years) or daily gain during a given age period (daily gain between 12 and 24 months of age). If we assume that body weight at 3 years is used (a trait which might be highly correlated with adult body weight), then we can make the following calculation:

Selling an animal we get per kg of live-weight about 250 CFA. In order to produce and additional kg of live-weight, additional feed is required. This last point will be effective through a lower carrying capacity. Thus in the absence of better data, half of selling price is taken as marginal profit (125 CFA/Kg liveweight). Here the same comment as for the marginal profit of milk yield is in order.


It is quite clear that traction is another important trait and it is to be expected that in mixed farming it will become even more important in the future. Selecting for size might improve the aptitude for traction to some extent. However, the definition of the trait, the derivation of economic values and the testing for the trait are so difficult that it was decided to ignore that trait for the time being. Additional reason for ignoring the trait was the “feeling” that the economic importance of a genetic improvement of this trait is not so high!

3.3 Economic Values

Production model - Biological Data

Assuming that we have an operation with about 200 heads of cattle and that the whole operation is market oriented. (The size here is essentially only a scale factor, chosen to better visualize the operation):

The following parameters might apply:

Age at first calving3.5 years
Calving interval1.75 years
number of calvings per cow5
Age at slaughter for bulls3 (4) years
weight at slaughter (bulls)300 kg
weight at slaughter (cows)250 kg

Then we have 75 cows (15 cows on average in each age group; more in the younger age groups, fewer in the older age groups due to mortality and culling)


42 calvings per year  - 42 lactations started
- 36 lactations carried through

(with approx. 14 % calf mortality and/or cow mortality)
leading to
18 female calves
18 male calves

Age distribution

0 – 1 year1818
1 – 2 years1818
2 – 3 years1818
3 – 3.5 years-9
3.5 -475
 58+ 138 = 196 heads

Cows represent slightly under 40% of the herd. Income

36 lactations × 400 kg × 100 CFA = 1 440 000 CFA
(for simplicity, a lactation is counted in the year when it is started)

918 males × 300 kg × 250 CFA = 1 350 000 CFA
15 cows × 250 kg × 250 CFA = 937 500 CFA
                                                  3 731 500 CFA ≈ 15000 US$

Increase by one unit of milk:

36 lactation × 1kg × 50 CFA = 1800 CFA

per cow 24 CFA

Increase by one unit of body weight:

18 males × 1 kg × 125 CFA = 2250 CFA
15 cows × 1 kg × 125 CFA = 1875 CFA 4125 CFA

per cow 55 CFA

It can be argued that the marginal profit for an increase in body weight is much lower in cows since maintenance cost over a period of 10.5 years is increased. Also if males are slaughtered later the higher maintenance cost also decreases marginal profit.

3.4 Objective Function

If we have to consider several traits we should combine them in a function such that for each animal we have just one number, which expresses the deviation of the breeding value of the animal from the population mean.

Hi = a1g1i + a2g2i + a3g3i + a4g4i


Hi is the overall breeding value of animal i (aggregate breeding value)

g1i is the breeding value of animal i with respect to trait 1 For short

g1i breeding value of animal i in trait 1 (trypanotolerance)

g2i breeding value of animal i in trait 2 (milk yield)

g3i breeding value of animal i in trait 3 (meat yield)

g4i breeding value of animal i in trait 4 (traction)

The coefficients a1, a2, a3, a4 have to be chosen in such a way that Hi really reflects the utility of the animal i to the farmer. As coefficient aj the marginal profits should be used.

In our case a4 (coefficient of traction) is put to zero, and a2 (coefficient of milk yield) is taken as 25 CFA a3 (coefficient of meat yield) is taken as 50 CFA a1 (Coefficient of trypanotolerance) is chosen in such a way that the trait is at least not deteriorating.

Thus we get:

Hi = a1g1i + a2g2i + a3g3i = a1g1i + 25g2i + 50g3i

4. Genetic Parameters (Step B)


Trail et al. (1991b) estimated repeatabilities, means and standard deviations.

Average PCV0.33 ± 0.10
Mean (healthy animal)34%
σ after treatment3%
σ before treatment5%

In another trail σ was more in region of 2.1%

In another paper Trail et al. (1991c) reported heritabilities for average PCV of 0.35 ± 0.30 and of lowest PCV of 0.48 ± 0.31. Standard errors are rather high and there is an urgent need for more precise estimates.

For any breeding scheme care has to be taken that genetic parameters can be estimated as a byproduct.

For calculations we assume:

h2 = 0.1
σ2A = 0.9
σ2P = 9

and zero correlation between the liability and other traits. The σP is guessed from the various estimates of Trail et al. (1991c) and the h2 is rather guessed from what is known of the h2 of other diseases (e.g. Mastitis etc.). A h2 of 0.35 or 0.48 seems to me highly unlikely for trypanotolerance. If it is really so high for PCV then I would conclude, that the relation between PCV and trypanotolerance is weak.

Milk Yield (Milk offtake)

Agyemang et al. (1991) give:

Mean404.3 kg
σ183.1 kg
Repeatability0.54 ± 0.10

Since no good heritability estimate was found a value of 0.2 is assumed and correlation to the other traits are regarded as zero!

For calculation we assume

h2 = 0.2
σ2A = 6400 kg2
σ2P = 32000 kg2

The h2 of 0.2 is chosen, since it is at the lower end of the range of what can be found in the temperate climates.

Meat Yield

Agyemang et al. (1991) give for post partum cow the following weight

mean225.7 kg
σ28.2 kg

Planchenault (1987) gives the following values for animals at an age of 1.5 years

mean121.23 kg114.29 kg
σ27.04 kg20.55 kg
Fall, A. (1989)
 18 months24 months
mean103.9 kg124.7 kg
σ27.5 kg39.7 kg

For calculation we assume

h2= 0.4
σ2A= 324 kg2
σ2P= 810 kg2

and correlation to other traits are regarded as zero. Since the trait of interest is slaughter weight (age at slaughter assumed to be three to four years) we would need the parameters for that age. The above chosen values are assumed to come as close as possible to the true values.


No values were found in the literature. Since a4 is put to zero, the parameters are not needed for the moment.

4.1 Properties of the objective function and the prediction

Summarizing and simplifying (rounding) the results obtained in the last section, we get (or use)

Hi = a1g1i + a2g2i + a3g3i
    = 2000g1i + 25g2i + 50g3i
H^i = Ii = b1(y1i - μ1) + b2(y2i - μ2) + b3(y3i - μ3)

The variance - covariance matrix of [g' y'] is

a2 and a3 are from the last section, a1 was chosen somewhat arbitrary

σ2H = 8 410 000 σH = 2900

Using the chosen economic weights and genetic parameters and applying standard selection index theory, the index coefficients b', the variance of the index σI2, the correlation between the true and estimated breeding values ■IH and the genetic progress in the total merit (■H) and in the individual traits (■g) are derived.

If on each individual, performance is recorded in all three traits (possible in females) and no other information used:

I = b' (y - μ)

b' = [200 5 20] = a'VgyVy-1
    = [a1h12 a2h22 a3h32] in this special case.

σ2I = b'Vyb = 1 484 000         σI = 1218.2

ρHI = 0.42

Monetary value

ΔH = i ρHIμH = i 1218.2

Δg = i [0.15     26.3     5.32]

The total genetic merit is by definition also the sum of the products of genetic progress in the trait i times the economic coefficient of the trait i. Thus:

ΔH = i[0.15.2000 + 26.3.25 + 5.32.50] = i 1218.2

Since the economic coefficient for trypanotolerance is very arbitrary, it is of value to calculate the worth of the improvement just for milk and meat. This is:

ΔH(milk and meat only) = i[26.3.25 + 5.32.50] = i 923

It might be of interest to know, what is the consequence if we use a1 = 1000 (giving less emphasis to trypanotolerance) or a1 = 3000 (giving more emphasis to trypanotolerance).

For a1 = 1000 we get:

ΔH = iρHIσH= i1102 
Δg = i[0.082 29.05.88]

Direct monetary value (considering only milk and meat)
i[29.0.25 + 5.88.50] = i 1020.1

For a1 = 3000 we get:

ΔH = iρHIσH= i1390.7 
Δg = i[0.194 23.04.66]

Direct monetary value (considering only milk and meat):
i[23.0.25 + 4.66.50] = i 808.2

5. Formulation of breeding scheme (Step C)

In any breeding scheme genetic progress per year is one of the most important items.

The formula for genetic progress as given by Rendel and Robertson (1950) is:

ΔG/Δtgenetic progress per year
ΔtFMgeneration interval in the path “female to male”; Average age of the females when their successful males are born.
ΔtFF, ΔtMF, ΔtMM tolikewise for the path “female to female”, “male female” and “male to male”.
ΔGFMsuperiority of the females selected to have male offspring compared to all females tested.

ΔGFF, ΔGMF, ΔGMM likewise for the other paths.

The genetic superiority in a path can be expressed as:



σH is the standard deviation of the aggregate genotype (of H=a'g)

ρFM is the correlation between the true and estimated breeding values for the females used to produce males.

iFM is the intensity of selection. It expresses the standardised selection differential of the selected females to produce males compared to all tested females.
The value depends on:
• the fraction selected
• the number tested
• the correlation between the selection indices of the various animals.

Rewriting the formula leads to:

In most cases we have:

ρFF = ρFM and ρMF = ρMM

simplifying the formula somewhat further to:

5.1 Proposed scheme

Screening based Open Nucleus Breeding Scheme. This is an ambitious and demanding program but since there are about 5 million N'Dama and another 1,2 million N'Dama × Zebu crosses, quite a large investment seems to be justified.

The program tries to combine the advantages of a centralized breeding station (ability to carry out more complicated tests) with the large field based schemes. In a certain way the station is the backbone of the scheme.

One distinction should be made very clearly: testing of animals should be done in conditions which are as similar as possible to the conditions in the field. However, when the selection decision is taken then we should insure that animals multiply as rapidly as possible (there a calving interval of 2 years e.g. is clearly not desirable).

With respect to feeding, during the test, nutrition should be similar to the anticipated field conditions (what might be common practice in 5 to 10 years time). After the test is finished the best nutrition possible should be applied.

The program is envisioned to be carried out on several (3 to 4) stations more or less independently but following the same protocol and exchanging genetic material from time to time. This reduces the risk of inbreeding and also the risk of a catastrophe (knock out of one center would then not knock out the whole program).

At the start of the program each station begins to work with the existing animals pretending essentially that the older animals are the survivors of the test. That insures that there are no huge investments for animals, at the start (In addition it would be difficult to get superior animals from anywhere). In about 5 years time the scheme is then fully implemented.

Let us assume that the Nucleus has:

180 selected breeding females The average length of use of selected females might be six calvings (including the first).

In the station it should be possible to have an average calving interval of 15 months (longer between 1st and 2nd calving, shorter between the other calvings)

180 breeding females on station
72 male72 femalecalves born
60 male60 femalecalves alive
20 male20 femaleanimals from screening operation
80 male80 femaleanimals available for selection

If cows are kept on average 6 calvings then one sixth have to be replaced each year i.e. 30 cows allowing for some mortality.

Thus we would select the top 30 first lactation cows and discard the other 50 cows (have them available for other purposes).

There are altogether 230 cows on the farm (180 breeding females plus the 50 which are later culled). For these cows we need about 10 bulls (better fewer).

Young bulls (male calves) are all kept till they are used for breeding. Thus at an age of about 2.5 years there are 80 bulls available.

If bulls are kept 2 years then each year the top 5 must be selected.

Thus we have

30females out of 80 females-> iFF = 1.01
5males out of 80 males-> iMF = 1.96

Apart from special cases the same males are also used to produce males, thus iMF ≈ iMM (that could be made more efficient on the expense of more inbreeding).

However, not all females would be needed to produce males.

There the best 5 females of each year (30 females altogether) are more than sufficient to produce the males. We could then even ensure that only the younger cows are used (e.g. first four calvings). However, selection is even more efficient if we can keep all sons of all cows till being used.

Before continuing the description of the breeding scheme the screening operation is outlined.

Screening operation

There are (at least) three arguments in favour of screening:

• If we are dealing with a truly quantitative trait (many loci, each with a small effect) it allows an extremely high selection intensity at a low cost, albeit with a low precision.

• If a major gene is segregating at a very low frequency, there is a chance of finding it and increasing the frequency in the nucleus.

• Selection is based on field performance.

For the first argument, there is little justification to select more intensely than 0.1 %. A fraction selected somewhere between 0.1 % and 0.5 % should be appropriate for most circumstances.

If we want to select by screening 40 animals a year then we should screen about 10 000 to 40 000 animals (taking into account losses etc.).

From experience it is known that it is nearly impossible to buy these outstanding animals. Only the offspring might be bought.

However, in average herds all cows are mated by a more or less random bull. This implies that we would get only half the value of the identified cow. Therefore we must find a way to make that cow pregnant with a bull of our choice.

From discussion with farmers the following procedure might work

Testing of animals

With respect to the size of the nucleus, there exists no firm rule. In a small nucleus chance events are more important and inbreeding can become a problem. The greater the nucleus the higher the genetic progress but with diminishing returns and higher costs. Four centers with 180 cows each, plus screening, seems to be a reasonable size.

5.2 Expected genetic progress of scheme

Generation interval

Since the testing environment should not differ too much from the field conditions it is assumed that selected animals are first used at 2.5 years. Males are used for 2 years and females for 6 calvings (1.25 years calving interval) Then we get the following generation interval (average age of parents at the birth of the replacements):

ΔtMF =(3.25 + 5.25)/2= 4.25
ΔtMM =ΔtMF= 4.25
ΔtFF =(3.25 + 9.50)/2= 6.375
ΔtFM =(3.25 + 9.50)/2= 6.375
∑Δtij =21.25 

Selection intensity

For the selection intensities we get:
iFF = 1.01     iFM = 1.01     iMF = 1.96     iMM = 1.96

Precision achieved

On each animal we have data on:

result of the trypanotolerance test
result of daily gain:
                   and on females we have the
result of the first lactation

For most animals we have these results also from the dam and from paternal halfsibs.

Ignoring (at present) this additional information from relatives except that for males the milk yield of the dam is used.

Then we have for the females:

b' = [200 5 20]

σ2I = b'Vyb = 1 484 000       σI = 1218.2

ρHI = 0.42

ΔH = i ρHIσH = i 1218.2

Δg = i[0.15 26.3 5.32]

Direct monetary value (considering only milk and meat) 26.3.25 + 5.32.50 = i 923 and for the males we get (y2* milk yield of dam)

a', g and σ2H are as before

b' = [200 2.5 20]

σ2I = 884 000       σI = 940.21

ρHI = 0.324

ΔH = i 940.21

Δg' = i [0.19 8.51 6.89]

Direct monetary value (considering only milk and meat)
i [8.51.25 + 6.89.50] = i 557.25

Genetic progress per year in milk and meat (leaving trypanotolerance without a monetary value):

Genetic progress per year in milk (kg milk/year):

Genetic progress per year in body weight (kg bodyweight/year):

5.3 Dissemination of genetic progress (Step E)

While the structure of most herds means that improvement programmes would be difficult to operate, the same structure will provide difficulties for dissemination of genetic progress. It is important that detailed proposals are drawn up, taking into account existing and likely infrastructure to provide a cost effective method of dissemination of improved stock.

A possible structure could be as following:

From 5.1 we have that there are about 80 males available for selection each station and year. As a rough rule of thumb, we should use only the better 50 % (could be somewhat more if timelag is large). Thus 40 males can be used out of which the best 5 are used in the nucleus (also from 5.1). That means that there are 35 bulls per station which can be given to “multiplier herds I”. With four stations there are 140 bulls available for “multiplier herds I”.

By current practice 140 bulls can serve about 3500 cows, which produce per year about 900 bulls. Since there is no testing, all 900 bulls are of equal merit for us and can be further distributed to “multiplier herds II”. If in the “multiplier herds I”, the bulls received from the nucleus are kept two years, then everything is doubled and we have 1800 bulls to distribute to “multiplier herds II”. “Multiplier herds II” could then produce more than 20 000 bulls sufficient for one million cows. To try to organise it beyond “multiplier herds II” might not be feasible and is also not too important.

A “multiplication herd I” is any herd, which uses only bulls provided by the nucleus and uses those bulls only a short time (one or two years depending on the structure). Similarly a “multiplication herd II” is a herd which uses only bulls provided by a “multiplication herd I”. Also the use of the bull is “multiplication herd I”. Also the use of the bull is restricted to a short fixed period.

The whole multiplication structure just outlined can only be a rough drawing, but it stresses the point that this multiplication has to be carefully planned and updated according to the needs.

5.4 Investment appraisal

From the last section we assume that genetic progress per year is 190 CFA on a per cow basis.

We also assume that there are 5 million N'Dama and another 1.2 million N'Dama × Zebu crosses.

We also assume that there are 35 % of the cattle in the category of cows and that one third of the population is influenced by the breeding program (that about 700 000 cows profit from it).

For long term money investment an interest rate (int) of 4% (inflation free!) seems to be appropriate.

The discounted returns per cow of one year of genetic progress is (genetic progress starting next year):

Multiplying it by the number of cows affected (e.g. 700 000) gives

4750 CFA. 700 000 = 3 325 000 000 CFA ≈ 13 300 000 US $

In this investment appraisal no costs are considered. However, we can turn the argument around and can state that we make neither profit nor loss if the cost for one year of selection is 13 300 000 US $.

If the investment is not done for an existing breeding project but starting from scratch, then there is a further delay. If that delay is about 10 years (!) then the return is further discounted and is roughly two third of the figure given.

There is no point in a lot of further elaboration since so many assumptions are very vague.

In my opinion 1 000 000 US $ per year are well justified to spend for such a breeding scheme whereas 10 000 000 US $ would be more difficult to justify.

5.8 Discussion of the proposed scheme

The scheme can be implemented quite easily. With screening the scheme achieves a 15 percent higher genetic progress (and also 15 percent higher discounted returns) than without it. That does not seem to be very much. However, there are strong arguments in favour of screening. In the whole computation it was assumed that the traits were purely polygenic (no major gene involved). In this case screening just adds to selection intensity. If there are major genes then we have a much higher probability of getting hold of them by using screening than by carrying out a more or less closed selection scheme. For me the chance of having major genes and making drastic improvements is high enough to argue in strong favour of a screening based breeding scheme. Also the screening based breeding scheme capitalizes on genetic differences which most likely exist between herds. The possible advantage is not taken into account in the calculation. Also since screening is mainly based on milk yield this selection gain is not taken fully into account in the calculation. Therefore the screening seems to be so promising that it should be tried.


Agyemang, K. (1990). A proposal for a research program to increase milk production of village N'Dama cattle through genetic selection and modification in management practices. Unpublished Ms. ITC.

Agyemang, K., Dwingert, R.H., Grieve, A.S. and Bah, M.L. (1991). Milk production characteristics and productivity of N'Dama cattle kept under village management in The Gambia, J. Dairy Sci. 74, 1599–1608.

Anonymous (1980). Trypanotolerant livestock in West and Central Africa. Vol. 1 - general study. FAO Animal Production and Health Paper 20/1.

Anonymous (1980). Trypanotolerant livestock in West and Central Africa. Vol. 2 - country studies. FAO Animal Production and Health Paper 20/2.

Fall. A. (1989). An Alternative Selection Scheme for N'Dama cattle in Senegal. Ref. No. 030/CRZ/Kolda.

Ferguson, W. (1987). Factors influencing the production potential of trypanotolerant livestock. In: Livestock Production in tsetse affected areas of Africa, pp.187–209. Proc. ILCA/ILRAD, Nairobi, 23–27 Nov.

Hoste, C.H. Chalon, E., d'Ieteren, G. and Trail, J.C.M. (1988). Le bétail trypanotolérant en Afrique Occidentale et Centrale, Vol. 3 - Bilan d'une décennie. FAO Animal Production and Health Paper 20/3.

Jahnke, H.E., Tacher, G., Keil, P. and Rojat, D. (1987). Livestock production in tropical Africa, with special reference to the tsetse-affected zone. In: Livestock production in tsetse affected areas of Africa, pp. 3–21. Proc. ILCA/ILRAD, Nairobi, 23–27 Nov.

Planchenault, (1987). Essai d'amélioration génétique des bovins en milieu défavorable - Etudes et Synthèse 26, 307 p.

Rendel, J.M. and Robertson A. (1950). Estimation of genetic gain in milk yield by selection in a closed herd of dairy cattle. J. Genet., 50, 1–8

Roelants, G.E. and Pinder, M. (1984). Immunobiology of African Trypanosomiasis. In: Contemporary Topics in Immunobiology, 12, 225–274.

Snow, W.F., Raawling, P. and Wacher, T.J. (1989). Estimates of the numbers of trypanosoma-infected tsetse biting cattle in The Gambia. In: Entomology Programme fourth Annual Report. Eds.: Snow, W.F. Rawling, P., Wacher, T.J. and Claxton. J.R.

Trail, J.C.M., d'Ieteren, G.D.M., Colardelle, C., Maille, J.C., Ordner, G., Sauveroche, B. and Yangari, G. (1991a). Evaluation of a field test for trypanotolerance in young N'Dama cattle. Acta Trop., 48, 47–57.

Trail, J.C.M., d'Ieteren, G.D.M. Feron, A., Kakiese, O., Mulungo, M. and Pelo, M. (1991b). Effect of trypanosome infection, control of parasitaemia and control of anaemia development on productivity of N'Dama Cattle. Acta Trop., 48, 37–45.

Trail, J.C.M., d'Ieteren, G.D.M., Maille, J.C. and Yangari, G. (1991c). Genetic aspects of control of anaemia development in trypanotolerant N'Dama cattle. Acta Trop., 48, 285–291.

Previous Page Top of Page Next Page